Understanding Single-Firm Behavior : Exclusive Dealing Session

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.



2 and










19 9:30 A.M. TO 4:00 P.M.

24 Reported and transcribed by:

25 Susanne Bergling, RMR-CLR




3 Chief, Economic Regulatory Section

4 Antitrust Division, Department of Justice

5 and


7 Assistant Director

8 Bureau of Economics, Federal Trade Commission


12 Morning Session:

13 Jonathan M. Jacobson

14 Howard P. Marvel

15 Richard M. Steuer

16 Mary W. Sullivan

17 Joshua D. Wright

19 Afternoon Session:

20 Stephen Calkins

21 Joseph Farrell

22 Benjamin Klein

23 Abbott (Tad) Lipsky



1 C O N T E N T S


4 Introduction

5 Presentations:

6    Jonathan M. Jacobson

7    Howard P. Marvel

8    Richard M. Steuer

9    Mary W. Sullivan

10    Joshua D. Wright

11 Moderated Discussion

12 Lunch Recess


15 Introduction

16 Presentations:

17    Stephen Calkins

18    Joseph Farrell

19    Benjamin Klein

20    Abbott (Tad) Lipsky

21 Moderated Discussion

22 Conclusion



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

2 - - - - -

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

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

5 Commission. My title is Assistant Director for

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

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

8 Regulatory Section at the Department of Justice,

9 Antitrust Division.

10 I am going to be leading the morning session,

11 and Dan will be leading the afternoon session, and

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

13 hearings, I am going to cover a few housekeeping

14 matters.

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

16 detention if you -- the BlackBerries and any other

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

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

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

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

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

22 guide you.

23 Third, a safety tip particularly for visitors.

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

25 which they actually did yesterday, please proceed calmly


1 and quickly as instructed. Dan and I will keep

2 everything calm and orderly. If we must leave the

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

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

5 of people running to a gathering point where you can

6 await further instructions.

7 Finally, we request that you not make any

8 comments or ask questions during the session. Thank

9 you.

10 Okay, today's session concerns exclusive

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

12 all the various topics involving vertical restraints and

13 vertical contracts. It has been an active area of

14 economic research and an active area of antitrust as

15 well. We are honored to have assembled a distinguished

16 panel of practitioners and professors who are very

17 knowledgeable in the issues we are going to tackle

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

19 the morning and then one later in the afternoon.

20 I will just briefly introduce the panelists for

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

22 little more detailed introduction as each speaker takes

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

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

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


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

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

3 Professor of Accountancy at George Washington

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

5 Assistant Professor of Law at George Mason University

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

7 Professor of Economics in the Department of Economics at

8 Ohio State and also Professor of Law in the Michael

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

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

11 Wilson Sonsini Goodrich & Rosati and a Commissioner of

12 the Antitrust Modernization Commission.

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

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

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

16 biographical description of each of the speakers as

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

18 Department of Justice web sites.

19 Our first speaker is Richard Steuer, who is a

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

21 in the practice of antitrust law, including litigation,

22 mergers and acquisitions, intellectual property

23 licensing, franchising and e-commerce. Richard has

24 written a book and several articles on antitrust law

25 which have appeared in various journals throughout the


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

2 Antitrust Committee of the Association of the Bar of the

3 City of New York.

4 Richard?

5 MR. STEUER: Thanks, Joe.

6 In baseball they say you can learn a lot by

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

8 have been able to observe a great deal about exclusive

9 dealing and in various contexts, both in litigation and

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

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

12 try to present what I have learned about exclusive

13 dealing would be to go through those articles and

14 briefly outline what it is that I have learned from

15 watching.

16 The first one was an article on "Exclusive

17 Dealing in Distribution," focusing on how exclusive

18 dealing works when you are talking about selling to

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

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

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

22 dealing to be virtually per se unlawful. The Standard

23 Stations case in 1949 introduced the rule of

24 quantitative substantiality. Then the major case of

25 Tampa Electric in 1961 brought in qualitative


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

2 reason approach with the Beltone case from the FTC in

3 1982, Jefferson Parish in the Supreme Court in '84, and

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

5 we get from California Dental.

6 Now, what I have found is the level of

7 distribution really matters in assessing the impact of

8 exclusive dealing. What we are measuring with exclusive

9 dealing -- why exclusive dealing is different from other

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

11 of competitors than anything else. Exclusive dealing is

12 interesting among the vertical restraints. This is the

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

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

15 are looking at is something quite different than in

16 vertical resale restraints where the restraint is on

17 reselling rather than purchasing. Exclusive dealing is

18 a restraint on purchasing, not on selling.

19 So, the level of distribution could be

20 wholesalers. One wholesaler can reach every retailer in

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

22 Retailers are chained to a location typically, although

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

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

25 chains, but basically they have a universe of consumers


1 that they reach. Wholesalers are a little bit

2 different, because foreclosing wholesalers does not mean

3 that you are foreclosed from reaching retailers.

4 Foreclosing retailers may or may not mean that you are

5 foreclosed from reaching end users. Reaching end users

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

7 exclusive dealing arrangement tying up 10 percent of end

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

9 Type of product is important. Shopping products

10 are products for which consumers will go from place to

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

12 that each dealer only has one brand does not necessarily

13 have as much of a foreclosure effect, because consumers

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

15 go and continue shopping, looking at other brands at

16 other dealers.

17 Convenience products, on the other hand, include

18 impulse products, products that a consumer is more

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

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

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

22 about the nature of the retail operation, a customer

23 going to that retailer is going to buy whatever brand

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

25 more considerable impact.


1 Another variable that is important to keep in

2 mind is alternate channels of distribution -- what is

3 sometimes called intertype competition -- and there was

4 a rather classic book that Palamountain published in

5 1955 on that. Today, the variation in intertype

6 competition is richer than ever with the rise of the

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

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

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

10 getting the product distributed.

11 Another possibility is simply establishing new

12 distributors. Is it more efficient, is it more

13 competitive, to have competitors with other brands

14 establish their own distribution networks than just

15 piggyback on the existing distribution network and

16 possibly compromising the amount of vigor with which the

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

18 you better off having one brand at each reseller and

19 having them competing against one another?

20 Foreclosure is measured in many, many antitrust

21 defenses. There is a measure of foreclosure for

22 monopolization, for attempted monopolization, under

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

24 recently had an opportunity to study what the different

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


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

2 The interesting thing is "foreclosure" is a term

3 that is used throughout the antitrust lexicon, but it

4 has a different meaning with each substantive offense,

5 and that is important to keep in mind.

6 The procompetitive effects when you are going

7 through distribution: Combating manufacturer-level free

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

9 were talking about in a case like Sylvania where one

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

11 one manufacturer free riding on the efforts of another

12 manufacturer, and exclusive dealing, by keeping other

13 manufacturers out of a particular wholesaler or

14 retailer, prevents that.

15 Of course, stimulate distributors. If the

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

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

18 in measuring how valuable that is, there is a

19 distinction between commodities and differentiated

20 products. With a differentiated product, there is

21 something more for the dealer to explain, typically,

22 about the features of the product. With commodities,

23 that is probably less so.

24 Stimulating suppliers. Exclusive dealing also

25 stimulates suppliers to put more time and effort and


1 money behind their channels of distribution, because

2 they know that other brands are not using the same

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

4 worry about divided loyalties where they are wasting

5 their effort.

6 Protecting trade secrets is similar. To the

7 extent that a manufacturer is providing trade secrets to

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

9 retailer or wholesaler is carrying other brands, it can

10 use that kind of information for the benefit of the

11 other brands.

12 Quality control as well is something that can be

13 controlled more directly with exclusive dealing where

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

15 particularly true where retailers or wholesalers are

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

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

18 control quality with other brands in there.

19 Resale restraints. There is a lot of talk and

20 we were talking earlier about whether there is going to

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

22 Some of these same considerations also go into the kind

23 of resale restraints we looked at in a case like

24 Sylvania, customer restraints, territorial restraints,

25 resale price maintenance, but those are all restraints


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

2 they do not apply in the same way.

3 The next thing I looked at ten years later was

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

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

6 started with single products. In the simplest case,

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

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

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

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

11 Sixth Circuit that the plaintiff won on essentially a

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

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

14 packaging.

15 The important thing to know in these cases is

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

17 These are discounts to induce exclusive dealing. It is

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

19 saying if you buy 50 percent of your requirements from

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

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

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

23 much foreclosure as exclusive dealing, and in many

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

25 fine.


1 However, sometimes it is essential for the buyer

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

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

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

5 commodity product. It was a fairly undifferentiated

6 product, but it was differentiated in certain quality

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

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

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

10 manufacturing. When a second company came along and was

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

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

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

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

15 much higher price.

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

17 the customers needed they could not buy from anyone

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

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

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

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

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

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

24 the market would have to replace all of those lost

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


even though it was not really a different product,

2 analytically, it almost was a different product, because

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

4 other brand.

5 A little like bundling. Bundling is almost

6 easier to see, because there are different products in

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

8 have because they are patented in some cases. Sometimes

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

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

11 bundling, because I know you have another program

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

13 on bundling.

14 The last thing I looked at was, who is

15 instigating exclusive dealing, and should it make a

16 difference? And particularly, "Customer-Instigated

17 Exclusive Dealing." There are mixed motivations on how

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

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

20 they would like to assure that there are plenty of

21 suppliers, because they would like to have alternatives,

22 and they want to play one supplier off against another

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

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

25 requirements contract not only means I will buy


1 everything from you, but the seller promising I will

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

3 a requirements contract and there are not enough other

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

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

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

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

8 survive.

9 Resellers, it is somewhat similar. In the short

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

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

12 disappear. They only provide competition to you. In

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

14 necessarily perpetual, the day may come when you would

15 like to have some options with other brands that could

16 supply you.

17 Now, why would a customer want exclusive

18 dealing? The most obvious reason is to induce lower

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

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

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

22 price.

23 Another reason is to assure a dependable supply,

24 and that is the requirements contract. Another is to

25 assure quality, in that it is expensive to qualify


1 suppliers in certain very technical industries, and you

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

3 assuring uniformity is important. There is a case

4 involving auto racing where it was felt to be important

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

6 level playing field among competitors. And achieving

7 logistical efficiencies. In some settings, just having

8 fewer suppliers is going to wind up lowering expenses.

9 Now, how do you find an appropriate legal

10 analysis where it seems that the buyer has instigated

11 the exclusive dealing? The supplier's objectives often

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

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

14 impact on competition -- will exclusive dealing

15 foreclose other suppliers from having customers or

16 having distribution? Another is to achieve

17 distributional efficiencies.

18 The reseller's objectives are the ones we just

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

20 and there are mixed motives about how strong a reseller

21 wants other brands to be.

22 The end user's objectives are a little bit

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

24 pricing, may have concerns about delivery, quality,

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


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

2 one supplier is really in favor of weakening other

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

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

5 So, what is the right analysis? When should

6 courts second-guess buyers for instigating exclusive

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

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

9 certainly when the buyer has a demonstrable motive to

10 eliminate competition at the supplier level so that it

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

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

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

14 instigating exclusive dealing.

15 Let me just conclude by saying I hope this quick

16 snapshot has highlighted some of the very many

17 differences that exist among exclusive dealing

18 arrangements. All of us as lawyers and economists are

19 always searching for those unifying principles that make

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

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

22 some of these variables that we have just been talking

23 about really do make a difference to the analysis.

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

25 (Applause.)


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

2 on time, perfect.

3 Our next speaker is Mary Sullivan, who is an

4 Assistant Professor of Accountancy at George Washington

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

6 of Chicago, Department of Economics, and taught

7 marketing at Chicago Graduate School of Business from

8 1987 through 1997. While at Chicago, she conducted

9 research on industrial organization and marketing

10 issues, such as slotting allowances, brand names and

11 trademarks.

12 In 1997, Professor Sullivan left academia for

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

14 she worked on a variety of antitrust matters and served

15 as Assistant Chief of the Competition Policy Section.

16 In 2004, she joined the Accountancy Department

17 at George Washington University, and as many of you

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

19 leading economics journals.

20 Mary?

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

22 by thanking the DOJ and FTC for inviting me to

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

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

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


1 they might charge me a slotting allowance, although in

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

3 charge one unless you charge it in advance.

4 Nonetheless, I will try to stay on track.

5 Slotting allowances and payola are two allegedly

6 exclusionary practices that receive different regulatory

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

8 whether the different regulatory treatment is warranted.

9 Slotting allowances and payola are similar in

10 many respects. They are basically the same practice

11 used in different settings. Slotting allowances are

12 payments made by manufacturers to retailers for stocking

13 new products. Payola consists of payments made by

14 recording companies to radio stations or DJs for playing

15 a particular piece of music. Both practices have

16 promotional effect. They serve to increase demand by

17 providing exposure to the product or music to consumers.

18 In each case, there is a scarce resource that

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

20 slotting allowances and airspace in the case of payola.

21 For both types of fees, there are concerns about

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

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

24 have talked to industry participants, you will learn

25 that these practices are widely believed to be


1 exclusionary, and the potential exclusionary effect is a

2 major motivating factor in the regulatory scrutiny that

3 each of these practices has received.

4 Now, oddly, despite their similarities, the

5 practices receive different regulatory treatment.

6 Slotting allowances are not regulated by the FTC. In

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

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

9 with attention both to likely competitive harms and to

10 likely procompetitive effects. So, they take a basic

11 rule of reason approach.

12 Alternatively, the FCC does regulate payola.

13 According to the FCC regulations, payments are

14 prohibited unless an announcement of the endorsement is

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

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

17 regulations, the major recording companies have recently

18 settled investigations brought by Elliott Spitzer, as

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

20 well known about these settlements is that the terms of

21 the settlements are more restrictive than the FCC

22 regulations, with payola completely banned in most cases

23 even if an announcement is made of the endorsement.

24 Now, given over the past few years we have

25 learned a lot about slotting allowances, both in terms


1 of the economic theories and in legal challenges, I

2 thought it would be an interesting exercise just to go

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

4 some insight as to why payola has received different

5 regulatory treatment and whether this makes sense.

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

7 the theories of exclusion. Can theories of exclusion

8 explain slotting allowances and payola? Now, there are

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

10 There are the popular theories or notions of exclusion,

11 and then there are the economic, sort of rigorous

12 economic theories of exclusion.

13 The popular theory of exclusion, according to

14 these theories, the payment of the fees increases the

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

16 increased entry cost may exclude manufacturers,

17 particularly small ones, and many of the complaints are

18 of this nature.

19 However, this so-called theory cannot really

20 explain exclusion. It is fairly well accepted that

21 auctioning scarce resource results in efficient

22 allocation, and unless something in the auctioning

23 process reduces the number of slots that are available,

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

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


1 someone will give the product financing in order to

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

3 consider this a valid theory of exclusion.

4 The other class of theories are the economic

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

6 the purpose of this talk are Farrell 2001 and Shaffer

7 2005. Now, without going into much detail at all about

8 these theories, all these theories share the feature

9 that you need to have a contractual provision for the

10 retailer to actually exclude a competitor in return for

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

12 retailer is reducing the number of slots available for

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

14 one important conclusion that I take away from these

15 theories is that simply paying a slotting allowance is

16 not enough to cause exclusion.

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

18 at the evidence, what do we know about slotting

19 allowances and payola, and ask the question whether the

20 evidence is consistent with the Farrell/Shaffer type

21 theories of exclusion.

22 In the case of slotting allowances, the answer

23 is sometimes. Occasionally slotting allowances are

24 accompanied by a contract to reduce the shelf space

25 available to competing manufacturers which could weaken


1 them and potentially exclude them. According to the

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

3 are fairly unusual, but they do occur.

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

5 evidence that exclusionary contracts are being used with

6 payola. The evidence that I have seen suggested that

7 recording studios are simply trying to use payola in

8 return for getting the radio stations to play their

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

10 exclude a popular song of a competing recording studio.

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

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

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

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

15 the Spitzer settlements, you will see that the evidence

16 he collected was quite thorough. What I conclude from

17 this is that according to the economic theories of

18 exclusion, payola is very unlikely to be exclusionary.

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

20 evidence from the courts to see what the courts say

21 about slotting allowances and exclusionary effects.

22 This is not really intended to be a comprehensive review

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

24 do is I looked at two legal challenges to slotting

25 allowances that are both important, have been very


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

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

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

4 cases.

5 One of the quotes from the Gruma case is

6 particularly revealing. In this case, the Court said,

7 "Some of the plaintiffs' losses are due to a

8 'self-inflicted' wound -- they chose not to compete for

9 shelf space."

10 Now, in this case, the plaintiffs were small

11 companies, small tortilla manufacturers who were

12 complaining that Gruma, the large manufacturer, was

13 buying up all the shelf space and giving it unfavorable

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

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

16 shelf space.

17 Now, in the Reynolds Tobacco/Philip Morris

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

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

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

21 somewhat different situation, because Reynolds, the

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

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

24 case, the Court concluded that the Philip Morris program

25 that involved the payment of slotting allowances


1 increased industry competition.

2 Okay, so if the theory predicts that payola is

3 unlikely to be exclusionary and the courts have ruled

4 that slotting allowances are an efficient means of

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

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

7 different regulatory treatment than slotting allowances?

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

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

10 stations should select music on the basis of public

11 interest rather than the radio station's commercial

12 interest. This view highlights the difference between

13 slotting allowances and payola.

14 The FTC and the courts see slotting allowances

15 as a valid and efficient means of allocating shelf

16 space, but the FCC believes payola results in an

17 allocation of airspace that is not in the public

18 interest apparently because it allows the radio station

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

20 this make sense?

21 Another way of asking that is, will regulating

22 payola cause radio stations to select music that is in

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

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

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


1 play both with and without payola.

2 Now, if payola is banned, radio stations are

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

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

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

6 they are going to earn all of their profits from

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

8 they are going to select music that appeals to people

9 who buy the advertisers' products.

10 Now, if payola is permitted, radio stations earn

11 revenue from both advertising and payola, and this may

12 cause the radio stations to change their selection of

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

14 who buy records and play less songs that appeal to

15 people who buy advertised products. It is not obvious

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

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

18 the radio stations choose what music to play on the

19 basis of what maximizes its profits.

20 So, I have several conclusions from this. The

21 first conclusion from the analysis, from this exercise,

22 is that it seems highly unlikely that payola will

23 exclude promising music. This argument of exclusion

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

25 Second, regulating payola will not help achieve


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

2 without regulations, radio stations will design

3 playlists to serve their own commercial interests. This

4 is unavoidable.

5 Third, prohibiting explicit payment for radio

6 airspace will not make competition for airspace

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

8 going to be competition for it. The competition will

9 take a different form. To the extent that recording

10 studios can find loopholes in the regulation, then there

11 will be little effect on the regulation on what is

12 played.

13 So, my own personal conclusion from this is that

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

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

16 to needless enforcement costs.

17 Thank you.

18 (Applause.)

19 MR. VITA: Thank you, Mary.

20 DR. SULLIVAN: No slotting allowance?

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

22 DR. SULLIVAN: Okay.

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

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

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


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

2 Professor Wright's research focuses on the law and

3 economics of the competitive process for product

4 distribution, including slotting allowances, category

5 management, exclusive dealing and other contractual

6 arrangements. He has published in numerous journals.

7 Professor Wright received his Ph.D. in economics

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

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

10 managing editor of the UCLA Law Review.

11 Joshua?

12 MR. WRIGHT: Thank you.

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

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

15 less background talking about what they are, since that

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

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

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

19 slides and papers from the other panelists, both

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

21 the afternoon.

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

23 use a slightly different definition of slotting

24 arrangements than Mary used and define the contracts as

25 per unit time payments made by manufacturers to


1 retailers for shelf space. There is a couple of

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

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

4 distinction between per unit tying payments and

5 discounts for slotting contracts, and it is an important

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

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

8 I just referenced on the economics of slotting

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

10 a distribution contract to be a per unit tying payment

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

12 issue for the next 19 minutes.

13 What else we know about slotting is that they

14 cover both new products and established products. So,

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

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

17 risk imposed on the retailer by giving shelf space to

18 some unproven product. We see slotting allowances on

19 those products as well.

20 What else we know is that they increased, there

21 was a spike in the prevalence and the magnitude of

22 payments somewhere between 1981 and 1984, and over the

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

24 products covered and the magnitude of payments has

25 continued.


1 So, the anticompetitive theories of slotting,

2 first, before I try to explain a procompetitive

3 rationale for shelf space contracts. We see slotting

4 contracts used by manufacturers with small market

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

6 the normative time for these agreements are between six

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

8 are not significant economies of scale in manufacture,

9 one of the conditions that drives the anticompetitive

10 theories in the literature. And also, the

11 anticompetitive theories have a difficult time

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

13 middle of the 1980s.

14 In terms of the procompetitive story for

15 slotting allowances, there are really two important

16 economic questions with respect to slotting fees, and

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

18 right? The first economic intuition one might have is

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

20 competitive retail market, supermarkets, et cetera, why

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

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

23 that is supplied for different products like we let them

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

25 for the shelf space?


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

2 discussion today, is we see sometimes that these

3 contracts include exclusivity provisions, unlike the

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

5 70 percent of the shelf space, give me a space to sales,

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

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

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

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

10 interesting questions, again, the form of the payment

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

12 skip so I can focus on exclusivity.

13 So, the answer provided by Ben Klein and myself

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

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

16 contracts solve this pervasive incentive incompatibility

17 problem where the retailer does not want to supply the

18 joint profit maximizing level of promotional shelf space

19 under the conditions where the supply and the shelf

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

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

22 space allocated for spices. Well, supplying additional

23 promotional shelf space to spices does not induce a

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

25 this retail outlet because they have given 90 percent of


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

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

3 incentive incompatibility problem solved with a separate

4 contract under these conditions.

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

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

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

8 fundamental point here is that for many products, and

9 differentiated products, we have manufacturers with a

10 large profit margin. So, the manufacturers, the

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

12 minus the marginal cost of manufacture, is large

13 relative to the retailer's incremental profit, whether

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

15 For a number of products, this is generally the

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

17 on the optimal level of shelf space, promotional shelf

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

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

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

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

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

23 incremental profits are large for the manufacturer and

24 not taken into account by the retailer.

25 Now, we can make the same argument with respect


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

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

3 least, allowing the manufacturers to set the retail

4 price, and competition between retailers is sufficient

5 to get an optimal jointly profit-maximizing price set

6 but not the jointly profit-maximizing level of shelf

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

8 wrong ends up being the question.

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

10 talking about price competition, you see here we have

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

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

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

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

15 products. But we have this offsetting effect induced by

16 customer switching. So, the intuition here is that

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

18 got this switching effect, so the quantity response

19 faced by the retailer when it changes the price has

20 these two different components.

21 One, when it reduces the price or increases the

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

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

24 inter-retailer competitive effects, right? So,

25 consumers may end up switching stores when we are


1 talking about price decisions or at least are more

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

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

4 So, the key point and argument here is that

5 because promotional shelf space does not involve large

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

7 consumers switching on a number of grocery products. My

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

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

10 there is some exclusive space granted for dog collars,

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

12 nobody switches the stores because there is one dog

13 collar versus two, okay?

14 And because we have this idea that there are

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

16 we have this incentive incompatibility problem, right,

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

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

19 relationship be approximately equal. The big difference

20 is this elasticity from the retailer's perspective of

21 the shelf space effect, right?

22 And so this is all to illustrate the point that

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

24 this incentive incompatibility problem is pervasive, and

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


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

2 the distinction here is not just because of price and

3 nonprice competition, okay? There are elements of

4 nonprice competition where there are inter-retailer

5 effects because all consumers value the service.

6 So, the supermarket provides a free parking lot.

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

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

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

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

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

12 consumers value some nonprice services, then they will

13 induce some switching, that for those services, the

14 incentive incompatibility problem is solved. The

15 retailer will supply those because consumers are all

16 willing to pay.

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

18 promotional shelf space is to give some sort of

19 effective, targeted discount to the marginal consumers

20 who are sensitive to allocations in the shelf space,

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

22 shelf space, and there is a substantial marketing

23 literature which demonstrates sometimes some really

24 surprising results about how large the effects can be in

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


1 shelf space allocation.

2 So, in these fairly general circumstances, the

3 disparity in margins and the small inter-retailer

4 switching effects from the supply of promotional shelf

5 space, the manufacturer wants more shelf space than the

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

7 some separate contract where the manufacturer pays the

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

9 solve this incentive incompatibility problem.

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

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

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

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

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

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

16 and the retailer has some incentive to not perform.

17 It can provide less than the contracted-for

18 level of space. It can otherwise violate the implicit

19 contractual understanding between the manufacturer and

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

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

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

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

24 exclusives in shelf space contracts.

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


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

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

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

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

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

6 not just buy the shelf space. They specify a

7 percentage. They specify a full exclusive. They

8 specify limits on the placement of rival products.

9 So, there are a number of procompetitive

10 rationales for exclusivity terms in these contracts, and

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

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

13 sort of shelf space contracting model, is that an

14 exclusive can help facilitate performance of the

15 contract, right? The retailer pockets this money and

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

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

18 it can efficiently define exactly what the manufacturer

19 is purchasing. Purchasing all of the shelf space,

20 detecting cheating becomes easy. The other thing it

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

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

23 promotional shelf space, and this intensifies the

24 bidding process between the manufacturers for the shelf

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


1 antitrust analysis, a good thing for consumers, because

2 these shelf space payments are passed on to consumers,

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

4 payments.

5 Quickly, so I can end here, category management

6 contracts are just a form of limited exclusive, where

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

8 of the space is the retailer delegates the function to

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

10 this in circumstances where consumers' demand for a

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

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

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

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

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

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

17 agreement, okay?

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

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

20 facts and the tortious behavior and lots of bad stuff

21 going on, there is some bothersome language in the

22 opinion about imposing a standard on category managers

23 that is tougher than the standard on monopolists using

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

25 dealing can make economic sense in these circumstances


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

2 demonstrating an anticompetitive effect before we engage

3 in any sort of balancing under the rule of reason

4 analysis.

5 I think I went over, sorry.

6 MR. VITA: Not too bad.

7 (Applause.)

8 MR. VITA: Thanks, Josh.

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

10 Professor of Economics in the Department of Economics at

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

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

13 vertical restraints is very well known. He has written

14 on a variety of different topics, including resale price

15 maintenance and exclusive dealing, and I know those

16 papers have appeared in some leading economics journals.

17 Howard also has advised the Japanese

18 International Trade Ministry, had a post in

19 telecommunications, the Federal Trade Commission and the

20 National Association of Attorneys General law on

21 vertical restraints issues. In addition, he has served

22 as an expert in vertical restraint matters for a number

23 of firms.

24 Howard?

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


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

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

3 for today is exclusive dealing.

4 It is obvious that exclusive dealing is a very

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

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

7 and Haagen Dazs has had the exclusive dealing in their

8 distribution contracts, car dealers typically have it,

9 there is exclusive dealing in beer distribution. It is

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

11 anything about it. You know, any business format

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

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

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

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

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

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

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

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

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

21 ride upon.

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

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

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

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


1 practiced exclusion, if you had exclusion in your title

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

3 was essentially a per se violation.

4 Now, exclusion there does not mean foreclosure.

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

6 and that is very different than keeping the firm totally

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

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

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

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

11 restraints is. It is that vertical restraints create

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

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

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

15 fruits of their actions will end up being frittered away

16 as other people take advantage of them, okay?

17 So, the idea behind vertical restraint is that

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

19 exclusive territories, for example, create a property

20 right for customers that a particular distributor or

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

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

23 customer generated, how do I guarantee the customer gets

24 generated? I protect the rights to that customer for

25 the guy who actually did the work?


1 Resale price maintenance is very similar. There

2 is a property right for the services that the

3 distributor provides, and Josh talked about how this

4 sort of works in slotting as well, like exclusive

5 dealing, that creates a property right for customers

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

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

8 Richard included in his discussion from the 1983 paper,

9 they all have that characteristic, that the supplier is

10 doing something to pull in customers and those customers

11 are being protected through exclusive dealing by -- from

12 some sort of bait and switch approach.

13 Now, the problem with exclusive dealing and what

14 makes it more serious and more of a worry than

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

16 supplier is creating a property right for somebody else.

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

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

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

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

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

22 right.

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

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

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


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

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

3 Aspen Skiing, Aspen Skiing and Aspen Highlands

4 cooperated to develop the Aspen market as a destination

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

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

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

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

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

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

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

12 elsewhere.

13 If you have a patent holder who has accessories

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

15 may decide to engage in exclusive dealing to try and

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

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

18 expert witness in a matter in which I thought exclusive

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

20 clear that these are anticompetitive so much as fraud or

21 contracting problems, but they are problems.

22 Okay, so the basic exclusive dealing story is

23 simply that the manufacturer invests in a product or a

24 reputation that brings in customers, if the manufacturer

25 confers upon its customers -- its customers onto dealers


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

2 dealer for a particular manufacturer, then customers

3 say, hey, that dealer is essentially certified as

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

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

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

7 got a better deal for you.

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

9 customer cost, the cost of generating the customers has

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

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

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

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

14 those customers wants to switch them to some other

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

16 circumstances in which you only charge for the customer

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

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

19 with royalties in business format franchises, right,

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

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

22 the royalty is generated, okay?

23 So, the dealer can avoid this particular charge

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

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


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

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

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

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

5 firm X's dealer, okay?

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

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

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

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

10 be a slogan for the hearing aids manufacturers who were

11 engaged in exclusive dealing, and they were going out

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

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

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

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

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

17 got a better deal on another hearing aid.

18 Now, the interesting evidence on this is that

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

20 manufacturers who used exclusive dealing, take them out

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

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

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

24 companies had agreed not to engage in exclusive dealing

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


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

2 counterfactual, what would happen if the practice were

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

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

5 screwed it up.

6 The manufacturers in the hearing aids case did

7 not recognize the role of exclusive dealing themselves,

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

9 the other manufacturers of hearing aids did, and they

10 ended up dead in short order, okay?

11 Now, after the Chicago explanation came out,

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

13 famous paper by Aghion and Bolton sort of launched the

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

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

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

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

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

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

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

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

22 write a contract between ourselves that has a penalty

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

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

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


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

2 works, but I'll be done.

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

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

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

6 the existence of this contracting penalty clause, okay?

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

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

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

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

11 okay?

12 There is a second set of theories that are

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

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

15 train leaving the station contracts. The train is

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

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

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

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

20 All of these theories require contracts. No contract,

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

22 game theory counter-revolution.

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

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

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


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

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

3 Chicago writers -- post-Chicago writers long ago

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

5 that in many contexts exclusive dealing can be deployed

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

7 allows the defendant to reap gains from the arrangement

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

9 agree, okay? True. Absolutely.

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

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

12 to happen. I do not have the slightest disagreement

13 with that.

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

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

16 appreciate the potential reply that it is impossible to

17 say that a given practice "never" could injure

18 customers. A creative economist -- there are creative

19 economists -- could imagine unusual combinations that

20 would cause injury in the rare situation, but antitrust

21 law applies rules of per se legality to practices that

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

23 Yes, Chicago.

24 Okay, but then we also have this statement the

25 literature on anticompetitive exclusive dealing, so


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

2 producing "possibility results" in simple settings to

3 counter Chicago School arguments. It is possible that

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

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

6 some of the possibilities, but the possibilities take

7 contracts, okay?

8 Problems are possible, and the problems involve

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

10 foreclosing a particular set of dealers. It means

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

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

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

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

15 The default rule in these cases is going to

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

17 exclusion could be bad, what will happen is that

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

19 factors suggesting the presence that we might have one

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

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

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

23 find that possibility means exclusion, becomes the

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

25 Exclusion plus dominance will equal violation. That is


1 where we were before. One minute.

2 Beltone, forget them, okay?

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

4 minute? The first possibility is that all of the

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

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

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

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

9 Then, we ought to require some notion that there

10 might be something wrong in this market in the sense

11 that there be a showing of foreclosure, and success

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

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

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

15 should never be considered the equivalent of

16 foreclosure.

17 But if you get to that point where you have

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

19 that foreclosure is a real problem in this industry in

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

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

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

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

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

25 find it very difficult to prove that the efficiency


1 benefits that you are claiming are really present.

2 With that, we will be done, okay?

3 (Applause.)

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

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

6 Wilson Sonsini Goodrich & Rosati, where he practices

7 antitrust law and has taken a lead role in many

8 significant antitrust matters over his 30-year career.

9 Among other cases, Jonathan was lead counsel for

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

11 monopolization case.

12 Jonathan was appointed by Congress in 2002 to

13 serve on the Antitrust Modernization Commission, which

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

15 considering several changes. He also is the editorial

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

17 chaired a number of ABA antitrust section committees.

18 He has written and edited numerous articles and books on

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

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

21 Sense For Exclusive Dealing."

22 John?

23 MR. JACOBSON: Thank you.

24 I also want to express particular thanks for

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


1 that is entirely appropriate, although I would comment

2 that in exclusive dealing cases, I have never

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

4 always been defense representation so far.

5 So, let's talk about exclusionary conduct and

6 exclusive dealing in particular. There are lots of

7 different exclusionary conduct devices, and these

8 hearings will cover most of them. I actually think

9 ripping your competitor's racks off the shelves is

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

11 the dialogue, but that is one example of exclusionary

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

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

14 know, major significant consumer benefits.

15 Exclusive dealing is in the middle, and it

16 presents a real challenge, because what makes exclusive

17 dealing potentially harmful is the very same mechanism

18 that makes the arrangement efficient and may lead to

19 lower prices for consumers.

20 So, what are the consumer benefits? I think

21 Richard went through them and I will just go through

22 briefly, but basically the distributor, if we are

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

24 distributor focuses his or her attention on the

25 supplier's product and becomes a more effective


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

2 supplier has an incentive to provide the distributor

3 with information and displays and all sorts of that

4 stuff without concern of free riding by competing

5 suppliers.

6 So, these benefits are very important, but they

7 are possible only because the arrangement is exclusive,

8 denying rivals access to the distributor's capabilities.

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

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

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

12 world phenomenon -- that the exclusive can deny the

13 rivals access to customers or supplies and have the

14 effect of driving their costs up and rendering them less

15 effective competitors, less effective constraints on the

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

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

18 allow the supplier to increase prices to consumers as a

19 result of the weakening of that competitive constraint.

20 So, the question is, how do we evaluate

21 exclusive dealing and quasi-exclusive dealing

22 arrangements in light of these simultaneous benefits and

23 harms?

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

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


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

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

3 overall approach to exclusionary conduct. One, we do

4 want to prohibit behavior that leads to the creation or

5 expansion of significant market power. We want to be

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

7 hearings, to avoid deterring procompetitive conduct. We

8 want to have rules that businesses can understand and

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

10 illegal. And we want to provide the courts with

11 sufficiently clear rules so that they can tell in the

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

13 So, for exclusive dealing, we have applied these

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

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

16 respect Richard's qualification of that and take the

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

18 first sort of modern formulation of the rule of reason

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

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

21 on foreclosure is unfortunate, and my basic point of

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

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

24 exclusive dealing case, what the plaintiff must show to

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


1 the efficiencies, is to raise prices or otherwise harm

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

3 major exclusive dealing cases over the last ten years,

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

5 consistent with that kind of paradigm.

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

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

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

9 abandon Section 2 enforcement entirely, but that has led

10 a lot of conservative thinkers and some more mainstream

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

12 determine whether there is a universal test for

13 examining exclusive conduct, and at some level we have

14 been searching for the universal rule ever since Learned

15 Hand's decision in the Alcoa case.

16 I would commend to all of your attention an

17 excellent article in the Antitrust Law Journal a few

18 months ago by Marc Popofsky, that having a

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

20 to practices as diverse as predatory pricing, refusals

21 to deal, ripping your competitors' products off the

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

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

24 will.

25 The main area of disagreement is the extent that


1 we need extraordinary screens to ensure that

2 procompetitive conduct is not deterred. The sort of

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

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

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

6 avoid deterring procompetitive conduct at the cost, many

7 recognize, of allowing the occasional illegal behavior

8 to go through.

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

10 allowing questions from the audience, because Greg

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

12 trepidation, although he and I have had a few

13 discussions on this subject, that I challenge the no

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

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

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

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

18 practice is not exclusionary for purposes of Section 2

19 unless it would make no economic sense for the defendant

20 but for the tendency to eliminate or lessen competition.

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

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

23 vertical conduct.

24 If you look at the certiorari brief filed by the

25 Justice Department in the Trinko case and the briefs


1 filed in the Court of Appeals in the Dentsply and

2 American Airlines cases, the Justice Department has

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

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

5 argued with some vigor by the Department of Justice.

6 One of the issues I have with the no economic

7 sense test is that it is fundamentally the Areeda Turner

8 predatory pricing pricing test in new garb. Areeda

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

10 urged that predatory pricing not be condemned unless it

11 is below cost with a likelihood of recouping the lost

12 profits through the market conditions that will result

13 from the predatory pricing scheme. And their test was

14 acknowledged and stated by them to be an extraordinary

15 test reserved exclusively at that time for price

16 cutting, because price cutting is so rarely harmful and

17 so extraordinarily important to our economy that we want

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

19 purely on the side of allowing the defendant to win

20 rather than the plaintiff to prevail.

21 Now, there have been efforts starting with the

22 article that Janusz Ordover and Bobby Willig put out a

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

24 regularly to other forms of exclusionary conduct, but in

25 general, we have been asking ourselves the question


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

2 purposefully extraordinary test -- and it was designed

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

4 it to other types of exclusionary conduct?

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

6 no economic sense test really does make no economic

7 sense, and I say that because exclusive dealing

8 arrangements make economic sense precisely because they

9 lessen competition by rivals for the affected business.

10 So asking that question tells us nothing about whether

11 the arrangement is procompetitive or anticompetitive.

12 Exclusives are usually associated, even in

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

14 exclusives are usually associated with real efficiencies

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

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

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

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

19 determination that the exclusive makes economic sense

20 for the defendant.

21 But the way in which those efficiencies are

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

23 exclusion. So, the judicial audience, the business

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

25 This arrangement makes no economic sense to me unless I


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

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

3 that is you apply a different test.

4 So, exclusive dealing is also interesting and

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

6 under some scenarios there need be no period in which

7 profits are sacrificed during the course of the

8 exclusive dealing arrangement. You can have

9 simultaneous exclusion and recoupment.

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

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

12 Justice Department and Federal Trade Commission to file

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

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

15 decision, and if you really want to read something

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

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

18 Texas Supreme Court, and it involved exclusive

19 promotional agreements with retailers, not exclusive

20 dealing arrangements, but exclusive promotional

21 agreements.

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

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

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

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


1 exclusives required the most prominent displays in the

2 stores and also exclusive ads.

3 In return for this, Coke provided very

4 significant lump sum promotional payments and deeply

5 discounted wholesale prices. So, the result was to

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

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

8 you accepted the market definition in the case. The

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

10 and it was not seriously disputed that the level of

11 promotional activity resulted in overall lower prices in

12 the marketplace for carbonated soft drinks as a whole.

13 Now, the exclusivity in that case, the

14 agreements, made economic sense only because the

15 exclusives made more -- made things more difficult for

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

17 pay thousands of dollars to a supermarket for a

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

19 you expect that the reduced price would be something

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

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

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

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

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

25 figure out some other way to sell soft drinks?


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

2 this kind of exclusivity could fail an incautious

3 application of the no economic sense test, but

4 appropriately, the majority upheld the agreements under

5 the rule of reason because there was no showing that

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

7 Now, I will very briefly talk about Microsoft,

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

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

10 doing was virtually costless. Leaving Internet Explorer

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

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

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

14 that this conduct makes economic sense and is,

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

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

17 we have a good precedent there.

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

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

20 that the focus that we care about in antitrust generally

21 and in exclusive dealing cases as one piece of that

22 overall puzzle is does this behavior injure consumers?

23 Does it raise prices? Does it otherwise injure

24 consumers and the benefit of the bargain that they are

25 going to receive?


1 The no economic sense test asks that we bypass

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

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

4 directly. The shortcut, which if applied incorrectly

5 can lead to very questionable results, is not a

6 necessary route. It does not protect competitive

7 conduct any more than a careful application of the rule

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

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

10 that basis.

11 Thank you.

12 (Applause.)

13 MR. VITA: Thank you, John.

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

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

16 minutes past, and we will reconvene.

17 (A brief recess was taken.)

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

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

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

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

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

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

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

25 going first.


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

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

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

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

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

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

7 you worry more about exclusive dealing the larger the

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

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

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

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

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

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

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

15 associated with exclusive dealing with little or none of

16 the benefits?

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

18 example. Those were not contracts at least of any

19 duration in that case. Microsoft basically told Dell

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

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

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

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

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

25 that looking at the economic analysis of exclusive


1 dealing and at the places where the game theoretic

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

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

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

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

6 find another example of a circumstance in which you say

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

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

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

10 the clever economists that Easterbrook was talking

11 about, to come up with the explanation.

12 I think I probably could for Microsoft as to why

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

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

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

16 in Dentsply, one of the things that was interesting

17 about that case was that the Justice Department seemed

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

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

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

21 investments. I bought so many inventories from these

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

23 stuck with the inventories, and the alternative

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

25 those inventories to tide you over while you are trying


1 to convert customers, right?

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

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

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

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

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

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

8 that case at all.

9 And again, it is possible to imagine

10 circumstances in which a manufacturer exerts or creates

11 a property right for itself to take advantage of

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

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

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

15 property right and take that right away from the other

16 guy?

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

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

19 problem. It becomes an antitrust problem only if you

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

21 the sidelines unwilling to invest because they are

22 subject to this misappropriation of their up-front

23 investments.

24 So, I can imagine circumstances under which that

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


1 them in this sort of standard exclusive dealing context.

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

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

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

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

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

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

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

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

10 know the facts.

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

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

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

14 MR. STEUER: It may be that lawyers and

15 economists do not always define "contract" exactly the

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

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

18 example, a monopolist had 100 percent share of the

19 market and came up with a discount schedule that

20 basically made it advantageous for customers who needed

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

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

23 factory, it was facing the daunting challenge of having

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

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


1 contract.

2 It was similar to a Colgate relationship that

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

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

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

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

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

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

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

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

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

12 that has really impressed me about the cleverness of the

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

14 coming up with sort of contract-based explanations for

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

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

17 NicSand, and those are matters in which the Justice

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

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

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

21 in Lepage's.

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

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

24 collateral damage, because there was a real problem with

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


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

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

3 trying to do is induce the discounter that you are

4 dealing with -- and this is particularly true for

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

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

7 to increase consumer welfare even though it is going to

8 increase prices or it is going to increase economic

9 welfare.

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

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

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

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

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

15 have the efficiency effects that are pretty significant,

16 and I think that cases like those may just be

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

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

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

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

21 offering you this really good deal to carry the broader

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

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

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

25 these are really tough questions.


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

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

3 more interesting, because it was real competition for

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

5 that case.

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

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

8 you would like to pose to the other speakers before

9 we --

10 MR. WRIGHT: I have one.

11 MR. VITA: Yeah, go ahead.

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

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

14 Jonathan had mentioned that he --

15 MR. JACOBSON: Ripping competitors' racks off

16 shelves? Yeah.

17 MR. WRIGHT: So, I think you either

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

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

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

21 arguing about in the paper is the economic analysis of

22 category management contracts, giving a procompetitive

23 explanation for why, under some conditions, the retailer

24 may want to delegate to the manufacturer the

25 responsibility of the shelf space allocation decisions.


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

2 What the point is about the decision in Conwood

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

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

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

6 would be bad --

7 MR. JACOBSON: Makes no economic sense?

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

9 the United States Tobacco employees sat out in the

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

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

12 point is about whether or not there is anticompetitive

13 effect and whether or not there are any foreclosure

14 effects and whether or not the conduct was sufficient or

15 likely to generate anticompetitive effects.

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

17 I will use someone else. Professor Hovenkamp, in

18 Antitrust Enterprise, using the testimony in the record,

19 estimates the distribution cost increase as something

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

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

22 is that there is this other question about whether or

23 not there is a likelihood of anticompetitive effect and

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

25 conduct, we should be asking the question.


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

2 DR. SULLIVAN: Ah, no.

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

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

5 propositions up and get some reactions from the panel,

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

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

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

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

10 Jefferson Parish Hospital District Number 2 versus Hyde,

11 1984, and the statement is, "Exclusive-dealing

12 arrangements are analyzed under the rule of reason."

13 Let me just pose probably a simple question to

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

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

16 in that case accurately summarize the law regarding

17 exclusive dealing? Richard and Joshua, Jonathan?

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

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

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

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

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

23 the product and relationship, all the things that I

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

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


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

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

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

4 development of the case law.

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

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

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

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

9 law is fairly clear about this, certainly under Section

10 1, and I think Microsoft and Dentsply, properly read,

11 import this analysis into Section 2. The greater the

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

13 impairment of rivals you are generally going to require

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

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

16 motivating the inquiry.

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

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

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

20 agree on, we thought we would start simple.

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

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

23 MR. WRIGHT: I third the motion.

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

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


1 think there are exclusivity arrangements that should be

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

3 are exclusivity arrangements that are always or nearly

4 always procompetitive and are thus appropriate

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

6 thoughts on that, you can step in.

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

8 case, four Justices saying that exclusive dealing

9 arrangements with multiple retailers are illegal because

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

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

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

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

14 to a per se rule on exclusivity here.

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

16 they are basically going to be low market share safe

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

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

19 the whole market is tied up with exclusives as in

20 Standard Stations? Do we really look just at the

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

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

23 question.

24 MR. VITA: Anybody else?

25 MR. WRIGHT: Sure.


1 MR. VITA: Josh?

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

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

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

5 think in addition to the point about safe harbors for

6 exclusives that do not foreclose some significant share

7 of distribution, sort of foreclose trivial shares of

8 distribution, then that is an appropriate place for a

9 safe harbor.

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

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

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

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

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

15 recognize this is subject to probably more debate, may

16 also be appropriate for safe harbors.

17 MR. STEUER: Some courts have misapplied the

18 term "exclusive dealing" to both exclusive selling and

19 exclusive buying. There is almost a safe harbor for

20 exclusive selling other than those rare arrangements

21 where one dealer has the exclusive for every brand there

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

23 In terms of real exclusive dealing, exclusive

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

25 out of Jefferson Parish, talking about 30 percent.


1 There are some other contexts where 20 percent is surely

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

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

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

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

6 percent exclusivity, that becomes more difficult, but

7 Jefferson Parish for practical purposes has introduced a

8 quasi-safe harbor of about a third.

9 MR. VITA: Okay, Howard?

10 DR. MARVEL: That is an awfully small harbor,

11 but on top of that I wanted to ask you about the

12 exclusive -- the exclusive -- which side did you put it

13 on, seller is --

14 MR. STEUER: Exclusive selling and sometimes it

15 is called an exclusive distributorship, "You will be my

16 only dealer in the State of Maryland" or something like

17 that.

18 DR. MARVEL: Yeah, but then turning that around,

19 how do you regard an agreement extracted by a dealer

20 like Toys 'R Us from seller -- a seller where he says,

21 you know, don't sell to my rival the same product that

22 you are selling to me. Is that okay?

23 MR. STEUER: It can be. Again, if it extracts

24 that from every manufacturer, that becomes increasingly

25 a problem. If Toys 'R Us were to enter into an


1 agreement with one manufacturer for one product and says

2 "I want to be the exclusive seller of this product," it

3 is rather limited what the impact is. In fact, I think

4 the decree that was finally negotiated specifically

5 provides for some limited exclusivity like that.

6 But if one chain were to become powerful enough

7 to sign up as the exclusive seller of all the toys for

8 all the major manufacturers, obviously everybody else is

9 frozen out, and I think there actually have been a

10 couple of examples like that.

11 DR. MARVEL: So, in Toys 'R Us, what happened,

12 if I recall, was that the Seventh Circuit of all people

13 said that the Toys 'R Us arrangement was not okay, and

14 that is because Toys 'R Us did have this sort of

15 monopoly position in the toy business, and it was

16 unassailable -- because of their unassailable position,

17 they really needed to protect the other poor souls like

18 Sam's Club from the depredations of Toys 'R Us. So --

19 is that right?

20 MR. JACOBSON: Well, another way to --

21 MR. STEUER: Well, Sam's Club or consumers. I

22 mean, the classic example, there was a wholesaler on an

23 island, I think St. Thomas, that was the sole

24 distributor for, it turned out, every single brand of

25 liquor, so that it basically created a bottleneck and


1 had monopoly at the distribution level, and to the

2 extent any of these examples approach that almost

3 textbook model, then you have a situation where

4 consumers really do not have other options at which to

5 shop for those particular products.

6 DR. MARVEL: So, is it an advantage to consumers

7 when Toys 'R Us contemplates getting out of the toy

8 business?

9 MR. JACOBSON: Because of WalMart? Look, there

10 were a lot of things going on in the case. One of them

11 was that the facts supported a finding of a horizontal

12 arrangement that was facilitated by Toys 'R Us, and I

13 think that is what concerned Judge Wood most --

14 DR. MARVEL: Right, absolutely.

15 MR. JACOBSON: -- in terms of the significance,

16 but looking at it purely on a vertical basis, at the

17 time there was a credible theory that it was raising

18 prices. Even though Toys 'R Us had a 20 percent market

19 share nationally, there were pockets of the country

20 where the share was in the high 40s, low 50s, and where

21 they were a must-have retailer for Mattel and Hasbro and

22 those other toy stores, and the result of this was that

23 the real, you know, the real discounters were cut off by

24 it, and you could make an arguable case that consumers

25 were paying higher prices as a result.


1 So, it was not -- it is not a crazy case. I

2 think it is a tough case, but I do not think it was a

3 crazy case.

4 DR. MARVEL: Well, I brought it up because it is

5 a tough case, but it is not a crazy case that what they

6 were doing was actually in the interest of consumers.

7 In fact, to have reasonably broad distribution of the

8 lines of the toy manufacturers -- and, of course, we

9 have also seen that not only has Toys 'R Us gone

10 belly-up and KB Toys and FAO Schwartz, but also the toy

11 manufacturers are rapidly fading into the sunset. Maybe

12 that is because only one Tickle Me Elmo was -- one

13 variety of Elmo was sold every Christmas at Sam's Club,

14 maybe not, but it does not appear that that industry is

15 a model of good health, and it may possibly be that that

16 is because a vertical restraint that was contributing to

17 not the monopoly behavior, but the good health of the

18 industry, was expunged.

19 MR. JACOBSON: Well, it may also be that our

20 analysis of monopsony power/buyer power is in its

21 infancy and that we really do not understand the

22 ramifications of WalMart, and I think that is the larger

23 issue, and I do not think anyone has a good answer to

24 that.

25 DR. MARVEL: I think that is right, because if


1 you look at Conwood, for example, and what Josh was

2 talking about, the Conwood case seems to me to have

3 turned in part upon the, shall we say, hyjinks of the

4 UST representatives who were trashing the Conwood

5 racks --

6 MR. JACOBSON: Right.

7 DR. MARVEL: -- but what it really turned on was

8 what was going on at WalMart, and that was a different

9 tale entirely. They wouldn't dare trash the racks at

10 WalMart, and so it kind of conflated those two things.

11 I mean, I have come up with a number of sort of

12 hair-raising anticompetitive activities that firms used

13 to engage in, and it is easy to come up with these

14 things, but that one is tough, because you start

15 conflating these things, and then you get a decision

16 that is made more on emotion than on what the economics

17 of it are.

18 MR. VITA: Let's go to the next slide, Brandon,

19 and let me just again read this, but this discussion

20 that Howard and Jonathan have been having I think sort

21 of leads into this next proposition and some of the

22 questions surrounding it. Let me just read it.

23 This is a quotation from Posner's Antitrust Law,

24 Second Edition, 2001, and in that book, Posner says, "I

25 propose the following standard for judging practices


1 claimed to be exclusionary: In every case in which such

2 a practice is alleged, the plaintiff must prove first

3 that the defendant has monopoly power...all the

4 plausible cases of exclusionary practices involve

5 defendants that have monopoly power."

6 And so let me pose two questions, two related

7 questions, you know, should monopoly power be a

8 requirement for challenging an exclusive dealing

9 arrangement under Section 1 of the Sherman Act and

10 Section 3 of the Clayton Act, and related to that is,

11 can exclusive dealing involving a non-monopolist result

12 in substantial lessening of competition?

13 And I think you two were already starting to

14 discuss that. Let me see if anybody else wants to have

15 any thoughts on that. Richard, Mary, Josh?

16 MR. STEUER: Well, clearly I think one of the

17 toughest areas is that space between 33 percent and 50

18 percent, because when you get above -- where you are in

19 the realm of Section 2 cases, the legalities change. I

20 know this means nothing to economists, but it certainly

21 does in terms of where you can get into court and

22 whether you can stay there.

23 The Microsoft case is an interesting example,

24 because there, in terms of browsers -- and I don't want

25 to dwell on this one case -- but certainly the share at


1 the time the case was brought was very low, and that may

2 explain why there was talk about monopoly power in

3 operating systems, but if you look at it purely as a

4 Section 3 type case and not searching for monopoly

5 power, but even at a low market share, was there a

6 danger -- an anticompetitive effect from the types of

7 exclusivity that was being entered into? Purely on the

8 numbers, you would say, no, the share is much too low,

9 and come back when it gets higher, but we all know where

10 that ended up.

11 MR. VITA: Well, let me ask this, and this may

12 be a question more for the economists, although the

13 lawyers are free to jump in, too.

14 Can we articulate or identify necessary

15 conditions in the downstream market that -- conditions

16 that are necessary for the exclusive dealing arrangement

17 to have an anticompetitive effect? Are there certain

18 things that have to be there before we have any ability

19 to infer anticompetitive consequences from an exclusive

20 dealing arrangement?

21 Josh, got any thoughts on that?

22 MR. WRIGHT: Sure. One -- I mean, let me make

23 sure I understand -- I understand the question.

24 MR. VITA: Yeah.

25 MR. WRIGHT: So, when you say competitive


1 conditions in the downstream -- you know, the downstream

2 market, so I am envisioning a manufacturer with

3 exclusive deals to a retailer --

4 MR. VITA: Think about that, that's a good

5 scenario.

6 MR. WRIGHT: That's what I would think of as an

7 example.

8 MR. VITA: Yeah.

9 MR. WRIGHT: I mean, substantial foreclosure

10 on -- I mean, the sort of well-known conditions from the

11 literature are that substantial foreclosure of the rival

12 so he can't achieve minimum efficient scale is a

13 necessary condition of most of these models, if not all

14 of these models, and so I think that that is -- you

15 know, in the legal analysis, we can have certainly, you

16 know, in the economics literature is a necessary but not

17 sufficient condition, and, you know, we know in the

18 cases, there are cases that end up on both sides. We

19 have a large foreclosure share but no liability because

20 of short duration or entry conditions or some such, and

21 so I think it is appropriate to use foreclosure as a

22 necessary but not sufficient condition.

23 MR. VITA: What about things like scaled

24 economies in the downstream -- when you talked about

25 scale economies, you were thinking about upstream, but


1 what about downstream?

2 MR. WRIGHT: So, in downstream, you can have --

3 there are cases where if you have large economies of

4 scale in distribution, you get -- you can have these

5 exclusionary effects as well.

6 MR. VITA: I mean, if there weren't substantial

7 scale economies downstream, or maybe some other factors

8 as well, do you think it would be possible in the kind

9 of long run or medium run for exclusive dealing

10 arrangements to have an anticompetitive effect? I mean,

11 why wouldn't -- you know, because if you don't have

12 substantial scaled economies and/or sunk costs at the

13 retailing level, why can't the -- supposedly the

14 foreclosed manufacturer get around the --

15 MR. WRIGHT: Right, so if you have -- at the

16 retail level you have -- I am going to frame this a

17 slightly different way, but if you have -- even if you

18 have the manufacturing scale economies but the retail

19 level you have free entry condition, then you are going

20 to have retailers who will re-align the supply

21 contracts, new entrants into the retailers who will

22 re-align the supply contracts, and so you need it at

23 some level, and the theory is you can do it with

24 economies of scale at the manufacturer level, but if you

25 have free entry at the retail level, I think that is


1 another problem for the exclusionary dealings.

2 MR. VITA: Jonathan, you looked like you might

3 have had something to add there.

4 MR. JACOBSON: No, I actually agree with that,

5 but it led into one of my sort of favorite topics in the

6 space, which is let's not talk about foreclosure,

7 because if we look at the percentage of distribution or

8 retail outlets foreclosed without examining entry, for

9 example, we may get a large number that's meaningless,

10 and that is why I think we are a lot better off if we

11 get rid of the word "foreclosure" and think about the

12 impairment of the rival, because that is the mechanism

13 that is going to lead to the consumer harm, not the

14 foreclosure, as such.

15 Foreclosure is a part of the analysis, but I

16 think it is only part of the analysis. You have to look

17 at the broader picture. Clearly there have to be

18 impediments to entry downstream.

19 And incidentally, I would agree with Posner's

20 book depending on the definition of "monopoly power."

21 You know, I think if you change it to market power, I

22 think, you know, a lot of people would subscribe to it.

23 I certainly would.

24 DR. SULLIVAN: Yes, I have one comment to make

25 on the -- following up on Josh's comment about free


1 entry in the retailing level. I agree that if there is

2 free entry in retailing, this is problematic for

3 theories of exclusion, because the excluded manufacturer

4 can more easily go to one of the new entrant retailers

5 to obtain distribution, but on the other extreme, if you

6 have, say, a monopolistic retailer, then I think that

7 the exclusive dealing arrangements, it is very hard to

8 prove that they would be harmful just because of the one

9 monopoly rent problem. So, I think you need to -- there

10 may be more potential for harm from exclusion in the

11 more intermediate market structures.

12 MR. VITA: Okay. Brandon, let's move on to the

13 next slide.

14 Here's another -- this is yet another quotation

15 from Justice O'Connor in Jefferson Parish Hospital

16 District Number 2 versus Hyde, and the proposition here

17 is, "Exclusive-dealing arrangement 'may be substantially

18 procompetitive by ensuring stable markets and

19 encouraging long-term, mutually advantageous business

20 relationships.'"

21 Let me put a couple of questions out. You know,

22 what are the -- empirically, what kinds of efficiencies

23 do the panelists perceive to be most likely to be most

24 significant in one of these exclusivity arrangements?

25 And think about this, you know, are there efficiencies


1 that are sometimes discussed maybe in the academic

2 literature in connection with exclusivity arrangements,

3 but in all likelihood, really aren't likely to exist or

4 likely to be very important empirically in real cases?

5 So, let me put that out there. Anybody --

6 DR. SULLIVAN: Yes, I will take that one just in

7 the sort of specialized area of slotting allowances. In

8 the academic literature, people make a big deal out

9 of -- one of the efficiencies of slotting allowances is

10 that it signals the product quality to retailers of

11 manufacturers' new products in cases where product

12 default is uncertain, and based on a lot of the

13 empirical studies that have been done by people in

14 marketing, that is simply not one of the efficiencies

15 that pops up, and I think the reason is there are quite

16 a few tools that manufacturers use to introduce their

17 products in addition to slotting allowances, and that

18 just -- so, I would feel comfortable ruling that out as

19 an efficiency, although there are plenty of other

20 efficiencies involved in slotting allowances.

21 MR. VITA: Howard?

22 DR. MARVEL: One of the cases that Richard

23 mentioned is the first nuanced case of exclusive dealing

24 I think was Beltone, and I think it is fair to say that

25 if there had not been some very un-nuanced evidence in


1 that case, that Beltone would have gone down in flames,

2 because by the time Beltone came up before the

3 Commission, its four principal rivals in that particular

4 channel that it was involved in had all met their

5 demise, and so Beltone was left as the monopolist --

6 thank you very much, FTC -- and at that point, they

7 didn't really have a good explanation for why they were

8 engaging in the exclusive dealing that they were

9 engaging in, but -- and so I don't see how they really

10 could have prevailed in that case unless there was this

11 evidence that was pretty clear that the companies that

12 had to give up the exclusive dealing practice had gone

13 belly-up.

14 So, in some ways John's paper talks about how

15 there probably is not a case that you can find where you

16 cannot determine that there are some advantages, but the

17 real difficult problem is to figure out how important

18 they are, and that is an incredibly difficult trade-off.

19 It is very hard to measure these things.

20 MR. VITA: Let me ask a follow-up on that point.

21 What significance, if any, should be given to observing

22 a challenged exclusive dealing arrangement in a similar

23 but somewhat more competitive market? So, you know,

24 that is sometimes an argument you make or you hear,

25 that, well, you know, this particular arrangement must


1 have some competitive benefits, because we see it over

2 here in these other markets that are structurally

3 competitive and where there is no plausible

4 anticompetitive theory of harm. How much -- how

5 powerful are those arguments and what weight should they

6 be given?

7 MR. JACOBSON: I think it is a much more

8 powerful argument if a small company is doing it than if

9 a large company is doing it in the same market. I think

10 looking at comparable markets and saying exclusive

11 dealing works efficiencies there, therefore they must in

12 this other market, really depends on how similar the

13 markets are. I would not make that leap without, you

14 know, a good deal of comparability evidence.

15 MR. VITA: Josh?

16 MR. WRIGHT: A related point, I mean, the nature

17 of the exclusive deal to facilitate some sort of

18 contract or performance, in the slotting example, again,

19 where the contract is over some sort of form of

20 promotion, and you see this a lot in exclusive dealing

21 cases where the underlying relationship between the

22 manufacturer and retailer relies on some sort of

23 promotional effort of the retailer and, in fact, is

24 contracted for, but the nature of performance in these

25 different markets varies a great deal, whether we are


1 talking about putting a product on an eye-level shelf

2 space or giving a product demonstration or some other

3 form of promotion.

4 So, the contracted-for conduct varies so much

5 market to market, I think the best you can make out of

6 seeing exclusive in a more competitive but different

7 market is sort of one of a cautious inference that we

8 generally know that exclusives can be procompetitive,

9 which I think there is not much disagreement on anyway.

10 MR. VITA: Okay.

11 MR. JACOBSON: I have a question for Mary. If

12 we renamed it payola, from payola to music leaders or

13 retail music program, do you think we would get a

14 different result?

15 DR. SULLIVAN: No. I think the people at FCC

16 and Elliott Spitzer would figure it out in a second.

17 DR. MARVEL: Why don't we call grocery store

18 slotting allowances payola?

19 DR. SULLIVAN: Well, I think we could, and one

20 thing you could do --

21 MR. JACOBSON: Because we would like to win the

22 cases.

23 DR. SULLIVAN: -- if the FCC regulated slotting

24 allowances, they would require the cashier at the

25 checkout counters to tell the customer each time he or


1 she was buying a product for which a slotting allowance

2 had been paid, then say, do you still want to buy it?

3 MR. WRIGHT: Well, as funny as that is,

4 California had proposed at one point -- I think it is

5 still kicking around in committee --

6 MR. JACOBSON: No, it was killed.

7 MR. WRIGHT: It was killed now?


9 MR. WRIGHT: Senate Bill 582, which would have

10 made -- it would have been illegal for -- essentially a

11 retailer would have to tell Pepsi exactly what Coke was

12 paying in terms of its promotional allowances, in terms

13 of the slotting fees, and if you conceive of these

14 things, these payments, as I do, as part of the

15 competitive process, I mean, this is a statute that is

16 a -- it is, you know, a legislatively enforced

17 collusion, right? And so it is silly, but, you know,

18 not silly enough to write down in a bill.

19 DR. MARVEL: Was it going to be the California

20 Raisin and Coca-Cola board? Is that --

21 MR. JACOBSON: It was proposed by a coalition of

22 the same people who represented the plaintiffs in the

23 Gruma case and the Harmar case. I mean, it was serious,

24 and it did get some traction, but it got killed fairly

25 early on in committee.


1 MR. VITA: Okay, let's move on then. The next

2 proposition is from Dennis Carlton from his article in

3 the Antitrust Law Journal, "A General Analysis of

4 Exclusionary Conduct and Refusal to Deal -- Why Aspen

5 and Kodak Are Misguided," and Carlton's proposition is

6 as follows:

7 "In the presence of scale economies, exclusive

8 dealing can be a way of depriving Firm 2 (or its

9 distributors) of the necessary scale to achieve

10 efficiencies, even though, absent the exclusivity, Firm

11 1 and Firm 2 would both be large enough to achieve

12 efficiency."

13 So, two related questions for the panel. One,

14 do you agree with Dennis' statement, and secondly, other

15 than its potential to deprive competitors of scale and

16 the resulting effect on prices, are there any other

17 theories of harm from an exclusivity arrangement that

18 should be the subject of antitrust concern?

19 DR. SULLIVAN: I will try the second question.

20 There is a theory -- and this is one I referred to in my

21 presentation -- by Greg Shaffer, a 2005 theory, and he

22 had a theory of exclusion in which scale economies did

23 not play a role, but what was going on is the retailing

24 segment was very, very competitive, and essentially

25 retailers, without exclusion of a manufacturer, would


1 earn almost no profits because their segment was so

2 competitive, and they could easily be coerced into going

3 along with an exclusivity deal that would exclude one of

4 the manufacturers because it simply would increase the

5 industry profits, and he developed conditions under

6 which this was true. One might argue that that would be

7 fairly unusual, but it -- you know, it is there.

8 MR. VITA: Anybody else? Dan, did you want to

9 add something?

10 MR. O'BRIEN: I would just like to ask, Mary,

11 following up, in that kind of a theory, if a

12 manufacturer could secretly get to a -- get with a

13 retailer, okay, assuming that everybody else was being

14 coerced into this exclusive with the manufacturer, and

15 negotiate something on the sly, wouldn't they be able to

16 undercut what, you know, the monopoly price that was

17 presumably being set by the other guys?

18 DR. SULLIVAN: I think so, and I think there was

19 something in particular about the nature of the game

20 that Greg set up that allowed him to get this outcome,

21 so I agree that might be -- it might not be that

22 problematic in reality.

23 MR. STEUER: There are a lot of assumptions in

24 here obviously. It makes a huge difference whether the

25 exclusivity is with end users and for how long. If this


1 is simply competition for the contract, clearly if one

2 manufacturer can get exclusive arrangements with the

3 bulk of the end users and freeze out the other, that is

4 going to have a profound impact, but if the second

5 manufacturer can survive long enough to go and bid the

6 next time and try to get the contract back, that is very

7 different. We do have some situations in defense, for

8 instance, where there is only going to be one winner of

9 these contracts. They are always exclusive, and yet you

10 do have some back and forth bidding as long as both

11 companies can survive. Here, I presume the assumption

12 is, with economies of scale, that there is a danger that

13 one of the companies disappears off the face of the

14 economic map.

15 MR. JACOBSON: I think this identifies a case --

16 there are certainly exceptions, as Richard points out,

17 but I think this identifies the exclusive dealing case

18 that you ought to worry about, you know, if these

19 conditions are holding, this is the case you ought to

20 worry about. There may be other cases you ought to

21 worry about. There may be cases where this is not a

22 problem because it is competition for the contract, but

23 in terms of our analysis, this is where I think we

24 should focus most of our resources.

25 I would add that this is an excellent article,


1 although Aspen I do not think was misguided, although

2 that is debatable, and Kodak was clearly correctly

3 decided.

4 MR. VITA: Whoa.

5 MR. O'BRIEN: Okay, we will try -- if we have

6 got time -- we have got time for one more, I think. Oh,

7 one more on this? On this proposition or one more

8 proposition?

9 MR. VITA: One more proposition, I think.

10 MR. O'BRIEN: I think we can go a little longer.

11 MR. VITA: All right, Brandon?

12 Okay, this next proposition is from Herbert

13 Hovenkamp, Antitrust Enterprise, 2005, and I will read

14 it.

15 "Exclusive dealing is a rule of reason offense,

16 requiring a plaintiff to show that the defendant has

17 significant market power, that the exclusivity agreement

18 serves to deny market access to one or more significant

19 rivals, and that market output to consumers is lower (or

20 prices higher) as a result."

21 A couple of questions for the panel. As to

22 significant market power or some indicator of

23 significant market power, is there or should there be a

24 safe harbor? And does anybody have an -- you know, it

25 says here in my notes that Jonathan in his writing


1 suggests courts apply a 40 percent market share safe

2 harbor, and if that -- you know, is that actually true,

3 and does anybody have an alternative minimum requirement

4 that they would prefer?

5 So, let me put those two out, those two

6 propositions out there and see what the panel thinks.

7 MR. JACOBSON: Well, I generally agree with what

8 I said.

9 MR. VITA: Glad to hear that.8

10 MR. JACOBSON: I think this is a pretty good

11 quote. I think "market access" needs a little bit of

12 definition, because I do not think you need -- this was

13 one of the other questions that we had talked about

14 before the program -- I do not think you need total

15 foreclosure. Again, I think the test needs to be the

16 degree of impairment of rivals. So, as long as denying

17 market access is read in that context, I think this is a

18 pretty good analysis.

19 I think 40 percent is a pretty good rough

20 screen. I think Richard's correct to point out that

21 Jefferson Parish is a 30 percent number, but it does not

22 say anything about a screen here or there, but if you

23 look at the subsequent cases, you are not going to find

24 any where the defendants have liability with less than

25 40 percent unless you consider Toys 'R Us an exclusive


1 dealing case, and there, you know, there were

2 extenuating circumstances given the horizontality of the

3 agreement.

4 MR. STEUER: And the term in here "significant

5 rivals" is significant, because it really raises the

6 question, who should have a cause of action here? At

7 some point, if there is ample competition in a market

8 and there is exclusive dealing going around, there may

9 be some marginal players who claim that they are being

10 excluded, and those can be emotionally appealing cases

11 in terms of jury appeal, and yet in terms of what the

12 actual effect is on the market, it may be very marginal

13 indeed, and there are not very clear tests right now as

14 to who should be able to bring a claim.

15 MR. O'BRIEN: If I could follow up with that,

16 John, earlier you had said that one of the areas in

17 which there was an agreement, you listed four points,

18 one of which was we want to prevent the enhanced -- you

19 know, practices that enhance market power. I am

20 wondering if you would agree with the last part of this

21 proposition, which is that plaintiffs have to show to

22 successfully bring a case that market output goes down

23 and/or prices go up.

24 MR. JACOBSON: Well, I think what he means is

25 that market output is likely to go down, and if you show


1 there is a significant enhancement or creation of market

2 power, I think you have done that. So, I do not think

3 this is inconsistent with that proposition.

4 MR. VITA: Okay, let's move on then.

5 This next proposition is from United States

6 versus Microsoft, the D.C. Circuit en banc decision.

7 The quotation is as follows:

8 "If the monopolist's procompetitive

9 justification stands unrebutted, then the plaintiff must

10 demonstrate that the anticompetitive harm of the conduct

11 outweighs the procompetitive benefit."

12 A couple of questions, and again, this may be a

13 little more for the economists, but anybody can step in.

14 First of all, does economics supply tools that

15 would assist courts in making this kind of assessment,

16 and do courts have the ability to apply these kinds of

17 tests?

18 Let me stop right there and see what the

19 reaction is from the economists on the panel.

20 DR. MARVEL: How about no?

21 MR. VITA: Say again?

22 DR. MARVEL: Do the courts have the tools? No.

23 MR. VITA: Actually, the proposition was, can we

24 as economists supply tools that courts could use? I

25 mean, what kind of analysis, if any, can we provide that


1 will allow noneconomists to make the kind of

2 determination that the Court called for in this case?

3 DR. MARVEL: I think that you really need to be

4 very careful about if you show anticompetitive harm, it

5 is pretty clear that you have got anticompetitive harm,

6 then I guess once you have gotten to that point, unless

7 convinced that the procompetitive benefits you are

8 trying to demonstrate will be easily enough measured and

9 ready available in such a way as to make it possible for

10 the courts to do the trade-off, I just think they are

11 awfully hard to prove what they are.

12 So, if you can really show that somebody is

13 locked out by the nature of the arrangement -- and that

14 means from the market, that does not mean from the

15 channel that the manufacturer in question controls, but

16 from the market as a whole -- then it is going to be

17 hard to do this trade-off, but if you have got the

18 anticompetitive harm and people are absolutely convinced

19 that it is there, then I think that that might be

20 enough.

21 MR. VITA: Yes, Josh?

22 MR. WRIGHT: Well, I think in this particular

23 quote, we have to -- there may be differences with

24 respect to what economists can do before and after -- in

25 the first and second clauses, right? The economist


1 might have tools to supply with respect to understanding

2 a monopolist's procompetitive justifications. Something

3 we can do is understand why we might see exclusives,

4 understand why conduct might be procompetitive, and the

5 conditions under which those explanations are likely.

6 That is something we can do and should be doing.

7 It is a lot tougher, the challenge of doing the

8 balancing is much tougher, and I guess the part that is

9 not in this quote is that the first step of requiring

10 the plaintiff to show the likelihood of some

11 anticompetitive effect is also an area where economists

12 can contribute by explaining the conditions for

13 anticompetitive effects are either satisfied or they are

14 not.

15 MR. O'BRIEN: Do you want to follow up, John?

16 MR. JACOBSON: I mean, this is what my article

17 is all about, so I do not want to leave this one

18 untouched.

19 A, most cases do not reach the level where you

20 need balancing. The number of cases where you really

21 need to balance it are few and far between. Usually the

22 case will fail because a prima facie case of

23 anticompetitive effect will not be shown. If that is

24 shown and the defendant shows a significant

25 justification, usually the plaintiff gives up at that


1 point. So, it is a very rare case that requires

2 balancing.

3 But if balancing is required, I think we need to

4 do it, and to say -- to throw up our hands and say it is

5 too complicated is just completely the wrong answer. We

6 do it every day. This building is filled with people

7 doing that in merger cases. It is done at the Justice

8 Department in merger cases all the time. This is

9 exactly what we do. So, to say that we are not going to

10 do this, it is too complicated, we might as well just

11 get rid of antitrust, because this is the guts of what

12 hard antitrust cases are all about, and we not only want

13 to do this, but we have to do it. This is one issue I

14 feel very strongly about.

15 MR. O'BRIEN: So, I wanted to follow up with

16 Howard, and, John, you may want to chime in on this,

17 too. You are concerned that if we can establish that

18 there may be an anticompetitive effect, that it is often

19 very hard for defendants to come in and argue, well, no,

20 in fact, there are efficiencies and that they offset the

21 anticompetitive effect, and I --

22 DR. MARVEL: No, what I am saying is that if you

23 can really show anticompetitive harm and --

24 MR. O'BRIEN: That may or may not be offset by

25 efficiencies, okay, so that is what I am saying. It may


1 or may not be offset, and what I took you to be saying

2 was that --

3 DR. MARVEL: That would make it really tough

4 for -- once you have a compelling demonstration of

5 anticompetitive harm -- and that is compelling for me,

6 not for you --

7 MR. O'BRIEN: Right.

8 DR. MARVEL: -- then I am not so sure that -- it

9 reminds me of the original merger guidelines when they

10 did not allow efficiencies as a defense, and I do not

11 think that that was absolutely nuts. So, if there is a

12 strong demonstration of anticompetitive harm -- and that

13 is not just locking up a channel, that is locking up the

14 market -- then I am not sure how much balancing I want

15 to do at that point.

16 MR. O'BRIEN: I see.

17 MR. JACOBSON: It is a rare case, Dan, it is a

18 rare case where you need to do this, but there can be,

19 at least in theory -- I will tell you, I have never seen

20 one -- but there can be one, at least in theory, where

21 the effect of the exclusives is to create a market

22 structure such that the defendant can raise prices to

23 some extent.

24 However, there may be sufficient dealer focus as

25 one traditional efficiency or other effects that overall


1 output of the product is increased. Think about your

2 resale price maintenance cases, the same -- it is the

3 same type of analysis, and if you can show -- first of

4 all, the burden is on the plaintiff, not the defendant,

5 but if the defendant can put in evidence to say that

6 notwithstanding the price increase, we are going to have

7 a significant overall market output effect that is going

8 to be procompetitive, I think you have got to entertain

9 that defense, and then I think you have got to see

10 whether that is true at the end of the day. Is the net

11 effect going to be to increase output or not?

12 MR. O'BRIEN: I guess I -- I am sorry.

13 DR. MARVEL: I think maybe if I can go, John's

14 point, I think part of the disagreement with -- the

15 implicit disagreement here is in my determination of

16 what constitutes an anticompetitive effect, because I

17 certainly would not agree to that parenthetical remark

18 that Hovenkamp had that said that prices are higher than

19 they would have been if the restraint was taken away.

20 Well, you cannot do that, because all of these

21 explanations talk about setting up a property right that

22 allow you to get a return on your investment which could

23 very well take the form of, you know, if you shift up

24 the demand curve, you are going to get a higher price

25 and greater output. If you get more output, end of


1 story. If it is a higher price, that does not really

2 tell you much of anything, and so that is I think part

3 of what we are -- we may be agreeing, somehow have a

4 different setup.

5 MR. O'BRIEN: So, following up on that, Howard,

6 I am curious how you feel about something like the no

7 economic sense test as a way to, you know, ask is there

8 a plausible efficiency rationale and, you know, maybe

9 short-circuit this balancing some.

10 DR. MARVEL: Sorry, but I -- it hurts me, but I

11 would have to agree with John on that one. I do not

12 like the test.

13 MR. O'BRIEN: Okay. Why don't you like the

14 test?

15 DR. MARVEL: I think your explanation is that

16 there is always economic sense in these practices, and I

17 think that that is right, that there will always be some

18 plausible argument that could be made. Unless we are

19 talking about gunning down your rivals or some such,

20 anything short of that, you are probably going to be

21 able to come up with some plausible argument on behalf

22 of that.

23 MR. JACOBSON: One convert, not a bad morning.

24 MR. VITA: Well, with that, then, we will bring

25 the morning session to a close. I would like to thank


1 the panelists. This was a really great discussion, and

2 I think everybody got a lot out of it. So, thanks very

3 much.

4 (Applause.)

5 (Whereupon, at 12:19 p.m., a lunch recess was

6 taken.)




2 (1:31 p.m.)

3 MR. O'BRIEN: Okay, let's get started. Well,

4 welcome to the second exclusive dealing panel of the day

5 in what is part of our ongoing series of public hearings

6 on single-firm conduct. My name is Dan O'Brien. I am

7 the Chief of the Economic Regulatory Section at the

8 Antitrust Division, and I will be moderating this

9 session along with Mike Vita, who is the Assistant

10 Director in the Economics Group, the Bureau of Economics

11 at the Federal Trade Commission.

12 The Department of Justice and the FTC are

13 jointly sponsoring these hearings to help advance the

14 development of the law concerning the treatment of

15 unilateral conduct under the antitrust laws.

16 Transcripts and other materials from the prior sessions

17 are available on the DOJ and FTC web sites, and I just

18 wanted to advertise that upcoming panels include a panel

19 on bundled loyalty discounts on November 29th, obviously

20 a practice that is somewhat related to exclusive

21 dealing, which is the topic for today, and then there is

22 a panel on misleading and deceptive conduct on December

23 6th.

24 So, today's session concerns the law and

25 economics of exclusive dealing. It was 40 years ago in


1 the Brown Shoe case that the Supreme Court made a very

2 strong statement against exclusive dealing, asserting

3 that it conflicts with the central policy against

4 contracts that take away the freedom of purchasers to

5 buy in an open market.

6 Since that time, the treatment of exclusive

7 dealing by the courts has changed fairly dramatically

8 over time, and the economics of exclusive dealing has

9 progressed, identifying both procompetitive and

10 anticompetitive aspects of the practice depending on a

11 range of circumstances.

12 We have a very distinguished group of panelists

13 here this afternoon to talk about these developments and

14 the current state of affairs from both the legal and

15 economic perspectives. My goals from today's panel are,

16 first, to highlight some areas hopefully where there is

17 some consensus on the effects of exclusive dealing and

18 how to treat it, but also maybe identify questions that

19 remain unsettled so we can have some consensus about the

20 questions that need to be addressed as we move forward.

21 So, before introducing the panelists, I just

22 wanted to thank my colleagues at the FTC and at the

23 Antitrust Division, particularly June Lee and the

24 economics staff at the Antitrust Division and Joe

25 Matelis in Legal Policy. The two of them together did a


1 lot of the work in putting together this panel.

2 The organization of the panel is going to be as

3 follows: We have four panelists. They will give

4 presentations of approximately 15 minutes. Then we will

5 take a short break. Then the panelists will have a few

6 minutes to respond to the other presentations if they so

7 desire, and then we will have a moderated discussion,

8 and we can go until around 4:00 p.m.

9 So, the order of the panelists, in case people

10 are wondering, will be Steve Calkins first, Tad Lipsky

11 second, Joe Farrell and then Ben Klein. So, let me

12 introduce Stephen Calkins. He is our first speaker.

13 Stephen Calkins is Professor of Law and Director

14 of Graduate Studies at Wayne State University Law School

15 where he teaches courses and seminars on antitrust,

16 trade regulation, consumer law and torts.

17 From 1995 to 1997, Steve served as General

18 Counsel of the Federal Trade Commission. Steve lectures

19 widely throughout the U.S. and abroad, most recently in

20 Europe and New Zealand. He has authored many

21 publications on competition and consumer law and policy,

22 and he has served on the editorial boards of well-known

23 journals in antitrust.

24 Stephen?

25 DR. CALKINS: Thank you. Thank you for the


1 introduction. What was not said is that I am actually

2 the most novice of all the people who are speaking here

3 today. I mean, you go over everybody else, and they

4 have been an expert witness in one or more of the

5 leading cases, they have litigated one or more of the

6 leading cases. Richard Steuer, in the previous session,

7 got up and proceeded to point out that he had published

8 three articles specifically on exclusive dealing. I

9 have never been an expert witness on exclusive dealing,

10 I have never litigated, I have never done an article

11 about exclusive dealing, as such, you know, we are

12 talking about somebody who is just not in the same

13 ballpark. So, with great humility, let me just tell you

14 that I am trying to sort out my own thinking and to

15 learn from all these geniuses.

16 To do that, we need to start somewhere, and so I

17 found one interesting case that I thought I would begin

18 just looking at a little bit, and here is a court

19 opinion that talks about how exclusive dealing "may well

20 be of economic advantage to buyers as well as to

21 sellers, and thus, indirectly of advantage to the

22 consuming public," and these advantages may often

23 explain why there are exclusive dealing contracts, and

24 if you wanted to go and understand whether they were

25 harmful or beneficial, you would look at a series of


1 tests.

2 You would look at "evidence that competition has

3 flourished, despite use of the contracts," or you would

4 look at the conformity of the length of their terms to

5 the reasonable requirements of the field of commerce, or

6 you would look at the status of the defendant as a

7 struggling newcomer or an established competitor or the

8 defendant's degree of market control, and you would go

9 through all this sort of stuff, but the opinion goes on

10 and says that to do this would just be extremely

11 difficult and to sort everything out would be an immense

12 challenge and, using words very similar to sort of the

13 basic sort of Areeda Hovenkamp mantra, we need to have

14 tests that are administerable by courts, we need to have

15 rules that can be enforced without wasting a lot of

16 societal resources on hopelessly complex litigation that

17 can't lead to any predictable outcomes, and so for

18 reasons of administrative efficiency, exclusive dealing

19 contracts should almost all be illegal, because this was

20 the original Standard Oil/Standard Stations case with

21 those thoughtful observations about the procompetitive

22 benefits of exclusive dealing, but the conundrum, the

23 difficulties, of litigating.

24 So, when I sat down and took a look to start my

25 sort of thinking about this and went back in time, I


1 said, golly, what an interesting beginning place, and I

2 then decided to pull out key dates in exclusive dealing

3 history, and we began with the classic Supreme Court

4 cases, which have been reviewed a little bit in the

5 morning session, and I will not mention them except that

6 Standard Oil you know, Brown Shoe was just referenced,

7 the classical Supreme Court cases were certainly

8 important moments in exclusive dealing history.

9 That led us to the key year of 1977 when all of

10 antitrust, as we know it, changed with Continental TV,

11 and then along came Robert Bork and the antitrust

12 paradox -- actually, along came all of the Chicago

13 School -- but Bork in particular is associated with

14 exclusive dealing, because he said so emphatically that,

15 by golly, there is only one monopoly profit. Exclusive

16 dealing cannot increase a monopolist's monopoly profit,

17 and so, therefore, "if Standard finds it worthwhile to

18 purchase exclusivity, the reason is not the barring of

19 entry but some more sensible goal such as obtaining the

20 special selling effort of the outlet," emphatically

21 saying that one cannot increase the profit of the

22 monopolist, and so there must be a procompetitive

23 justification, and those Supreme Court cases were just

24 dead wrong, a really clarion call for a different way of

25 looking at exclusive dealing.


1 As mentioned in the previous session, that call

2 was picked up first in the courts or the adjudicative

3 bodies in the Beltone Electronics opinion, where the

4 Court specifically relies on Bork and the antitrust

5 paradox to take a different approach to exclusive

6 dealing, the Federal Trade Commission, leading the way

7 to a new day of exclusive dealing decision-making, even

8 if we learned in the last session at the cost of having

9 sacrificed four of the five competitors, but

10 nonetheless, having led the way, that was followed

11 shortly thereafter by Jefferson Parish. Of course, it

12 is always cute, we refer to the Jefferson Parish

13 exclusive dealing holding, and it wasn't a holding at

14 all. It was part of the concurrence of Justice

15 O'Connor, but we all think of it as the holding from

16 Jefferson Parish where she emphatically said exclusive

17 dealing is judged more permissively than tying, it is

18 rule of reason, and "exclusive dealing is unreasonable

19 restraint on trade only when a significant fraction of

20 buyers or sellers are frozen out of a market by the

21 exclusive deal."

22 And since then, if you look at things that have

23 happened and you sort of parade through the exclusive

24 dealing cases that we know, which I throw up on the

25 screen in front of you or I throw up more of the


1 exclusive dealing cases that we know or I throw up more

2 of the exclusive dealing cases that we know, the one

3 great unifying principle is, of course, that the

4 defendant always wins. There are a few exceptions, but

5 overwhelmingly, the judicial treatment of exclusive

6 dealing ever since Beltone Electronics came down has

7 been that defendants win these cases, and you can find

8 support in the case law for all sorts of pro-defendant

9 propositions, with exclusive dealing being strongly

10 presumed to be legal if there is a market share of less

11 than 40 percent, if the restraint is of less than a

12 year, the contract is of less than a year, if the

13 contract is easily cancellable, if we do not have a

14 complete and total foreclosure, see the words in

15 Jefferson Parish, if there are no entry barriers, and

16 on, there are probably other ones as well, a whole

17 series of different principles, standards under which

18 defendants have won these cases, and that's a whole lot

19 of the exclusive dealing story, and then there is the

20 "but" part of the whole thing that makes our life

21 slightly interesting here.

22 There are three things to mention. The first,

23 the post-Chicago literature, I have reason to suspect,

24 although I did not look at his slides, that Joe Farrell

25 will reference a little of this, and it can be done in


1 all sorts of wonderful mathematical sophistication. I

2 think of the lesson as a common sense story of

3 collective action.

4 There was recently a case that Tad knows dearly,

5 the Coca-Cola case just decided by the Texas Supreme

6 Court. I do not know anything about the facts of that

7 case, and I have no opinion on the case. I do not know

8 what happened down there, but one of the things that

9 allegedly happened was that Coca-Cola paid retailers not

10 to allow 7-Up in its stores, and if you think about that

11 for a minute, you know, it sort of sets out the

12 collective action story very crisply. Why would a

13 retailer agree not to carry 7-Up when it knows that if

14 in the long run there is no 7-Up, that is probably bad

15 for retailers? And the answer is, of course, that if a

16 payment goes to a single retailer, that single retailer

17 can collect the payment knowing that its excluding of

18 7-Up is not really going to make a difference in the

19 long run, and you do not have all the retailers getting

20 together and agreeing that they will resist Coca-Cola,

21 because that would be illegal under the antitrust laws,

22 and so each separate retailer looking at its individual

23 self-interest can quite reasonably say, I will agree not

24 to allow 7-Up in my store, even though in the long run,

25 that is against the collective interests of all of them,


1 and it is because of that kind of a collective action

2 problem that exclusive dealing can sometimes harm

3 competition in the long run because one can have an

4 exclusive dealing arrangement that helps someone today,

5 with all the benefit going to that one entity, in the

6 long run, there is harm, but the harm is shared widely,

7 and so, therefore, you have a mismatch between the

8 benefit of the harm, a collective action problem, and

9 therefore, mischief can be worked.

10 Two cases have come along that have sort of set

11 out the -- sort of the other ways of thinking about

12 exclusive dealing, being Microsoft and Dentsply. People

13 in this room know those cases far more than I do, but

14 just mentioning a couple of points quickly, Microsoft

15 is -- you can find several different points in the

16 Microsoft opinion on exclusive dealing. This is one

17 where the District Court had said that there must be

18 complete and total exclusion before there is a

19 violation, and the Court of Appeals wrote that "even

20 assuming the holding is correct," and went on to say

21 there could still be a violation, thereby suggesting

22 that that holding may not be correct.

23 It went on and said there could be a violation

24 because there is a different standard under Section 2

25 than under Section 1, and even if something might be


1 lawful under Section 1, it could be unlawful when

2 engaged in by a monopolist. The Court asked rather

3 tough questions about the justifications for the

4 practices going on there, specifically saying that with

5 respect to one practice, where 14 of the 15 top Internet

6 access providers had contracts to work only with

7 Microsoft, the justification was to keep them focused on

8 Microsoft's product, "which is to say it wants to

9 preserve its power in the operating system market, that

10 is not an unlawful end, but neither is it a

11 procompetitive justification," thereby raising nice

12 questions about the difference between a benefit to the

13 seller and a benefit that qualifies as a procompetitive

14 justification.

15 Also of interest to the Microsoft case is we had

16 a very economically sophisticated court unable to resist

17 quoting some language indicating subjective intent.

18 "Kill the cross-platform Java by growing the polluted

19 Java market," so on and so forth, finding some comfort

20 in the words that business people had used to describe

21 what they were doing, and then finally being troubled,

22 even though we did not have total exclusion. So, we

23 have a whole series of interesting points that come out

24 of the Microsoft case.

25 In the Dentsply case, what did we have in


1 Dentsply? You had something where you had an at-will

2 contract, and yet the Court of Appeals said that was not

3 reason for the defendant to prevail, because

4 realistically, wholesalers are not going to give up $22

5 million in sales in order to pick up $200,000, and so an

6 at-will contract does not really give a new entrant

7 realistic access to the market. So, also, there was

8 talk about monopoly maintenance as a separate kind of

9 problem, and once again, we had reference to subjective

10 intent evidence.

11 So, where am I at that point in terms of, as I

12 end, little lessons that I draw from my sort of going

13 over things, and they are very tentative, because I

14 really have not thought these things through all the

15 way. I am learning, okay, but tentative things that I

16 might throw out as propositions.

17 One, it should be possible for a short-term

18 contract or contract that is cancellable still to be

19 found to be unlawful. It should be possible for there

20 to be illegality without total exclusion. Section 2

21 standards should be tougher than Section 1 standards.

22 It does not make sense to take all of the teaching of

23 Section 1 exclusive dealing cases and then import them

24 unthinkingly into the world of Section 2. If you have a

25 firm with a 75-80 percent market share and entry


1 barriers and lots of power, it ought to be tougher than

2 on a smaller, less powerful firm.

3 I hesitantly think that it is -- this will not

4 be popular with some of my panelists -- sometimes it is

5 interesting and possibly informative, if done very

6 carefully, to look at subject intent evidence to help

7 you sort through these difficult things. Clearly it

8 makes sense to scrutinize the procompetitive

9 justifications that are being offered up in a case that

10 otherwise looks troubling. The classic procompetitive

11 story is that the manufacturer has expended resources to

12 bring a consumer into the store who will then be bait

13 and switched off to another product. Well, you know, do

14 the facts fit that story or not? In Dentsply, the Court

15 thought they did not fit that story but went on to try

16 to really sort of sort through what is the

17 justification. It should not be enough just to say it

18 is a nonprice vertical restraint.

19 I personally would not think that one should

20 require a plaintiff to prove that prices have increased.

21 I mean, think again about your classic exclusive dealing

22 situation would be something where we are trying to

23 cause problems in the future. Go back to my Coke paying

24 to have 7-Up not around. The reason to do that is so

25 that things will be better for Coca-Cola in year two or


1 three or four or five, and one can have a lessening of

2 competition without prices today being affected. The

3 hard question here is the long-run competitive effects,

4 though, can't be a complete defense to say that current

5 prices have not gone up.

6 So, also we would say that the legal standard

7 really does matter in these cases. Going back to

8 previous sessions that you have had, you heard a lot

9 about the no economic sense test in the last session.

10 Another standard that can make a big difference in

11 exclusive dealing cases is whether you choose to adopt

12 the Posner "Exclude an equally efficient firm" test.

13 Were you to adopt that, which I would not favor, that

14 would make it much harder for a plaintiff to win an

15 exclusive dealing case.

16 And finally, in closing, pretty much on time, it

17 is interesting as you survey the landscape that there is

18 a whole lot of theory, not a great deal of empirical

19 evidence, and so I hope that this program, if nothing

20 else, inspires some people to go out there and get their

21 hands dirty and bring forth more empirical evidence.

22 Thanks very much.

23 (Applause.)

24 MR. O'BRIEN: Okay, our next speaker is Tad

25 Lipsky. Tad is a partner at Latham & Watkins and a


1 former Deputy Assistant Attorney General at DOJ. Tad's

2 30-year legal career has been devoted mainly to

3 antitrust, and it spans virtually every facet of

4 competition law.

5 From 1981 to 1983, Tad served as Deputy

6 Assistant Attorney General at DOJ under William Baxter.

7 Following government service, Tad developed a broad U.S.

8 and international antitrust practice, successfully

9 managing a variety of important antitrust matters.

10 As chief antitrust lawyer for the Coca-Cola

11 Company from 1992 to 2002, Mr. Lipsky conducted and

12 supervised competition matters before courts and

13 antitrust authorities in the U.S., Canada, the EU, EU

14 Member States, and dozens of other jurisdictions. He is

15 a frequent author and speaker on antitrust topics.

16 Tad?

17 MR. LIPSKY: Thank you very much. Until a few

18 moments ago, I had forgotten how stupid it was to follow

19 Steve Calkins to the podium, because he knows more about

20 whatever he speaks about than anybody else and expressed

21 his interesting views so trenchantly and with such great

22 humor that that is a very tough standard, but I will do

23 my little bit and see if we can find something to agree

24 on. I think we can find a few things to disagree on,

25 and we will see where it goes.


1 Exclusive dealing is a very elastic label. It

2 applies to a lot of different kinds of things. We have

3 already heard mention of the fact that tying, certain

4 kinds of bundling and price discounting can have effects

5 very similar to exclusive dealing, and therefore, when

6 you talk about exclusive dealing, you also need to be

7 considering a bunch of its very, very close relatives,

8 and so we are talking about implicitly, at least, a very

9 broad category of business conduct and competitive

10 phenomena.

11 Now, on the plus side, for our policy evaluation

12 of exclusive dealing, it has never been a per se

13 offense, which is a very good thing. It is a little

14 like saying, well, in Eastern Europe, they have a little

15 better luck re-adopting capitalism, because they were

16 capitalists within living memory, whereas in the old

17 Soviet Union, in the heart of Mother Russia, that was

18 not the case, and so there is no great body of learning,

19 there is no familiarity in the culture, and similarly,

20 with exclusive dealing, although it is true that back in

21 the Standard Stations days and when we were dealing with

22 the International Salt comment, that under Section 3 of

23 Clayton, you could condemn exclusive dealing either if

24 the defendant had market power or if there was not an

25 insubstantial amount of foreclosure, that is coming


1 within an eyelash of saying it is per se, but we never

2 quite got there.

3 There was always a little bit of procompetitive

4 culture left in exclusive dealing, and so -- as a matter

5 of fact, even in the dark ages, between the decision in

6 Schwinn, all vertical agreements are illegal per se,

7 until the release from bondage in 1977 with Sylvania

8 taking the nonprice verticals out of that category, I am

9 not aware of any decision going whole hog and saying,

10 well, that because of Schwinn, now we have to say that

11 exclusive dealing is per se. Even in those dark days,

12 we never had a rule for exclusive dealing that said

13 basically shoot on sight.

14 So, now, having escaped per se condemnation, I

15 think it was easier for exclusive dealing cases to kind

16 of re-absorb the economic learning, to talk about

17 procompetitive justifications, to insist upon genuine

18 proof that the process of competition had been

19 obstructed before liability could be imposed. We went

20 from Standard Stations, we went to Tampa Electric, which

21 basically said, well, even quantitative foreclosure does

22 not really give us the story that we want to hear when

23 we are talking rule of reason. And so, in effect, this

24 evolution is kind of a testament to just how thoroughly

25 the microeconomic analytic approach has been absorbed in


1 the antitrust enforcement industry, the enforcement

2 agencies, the courts, counselors, what have you, and

3 this is all very much to the good. This is as it should

4 be.

5 But one result of this emergence into the more

6 full-blown consideration of justifications and actual

7 competitive effects is that the role of market power and

8 monopoly power have been pushed to the fore, and for

9 most kinds of exclusive dealing claims, you need to have

10 market power or monopoly power at one level in order to

11 have any kind of a plausible theory of restraint, and so

12 now it has become a topic that is addressed more under

13 the Section 2 standards than under Sherman 1 or Clayton

14 3, and that is fine. So, that focuses, to the extent

15 that these issues come up under the Section 2 rubric,

16 that focuses you on monopoly power, because it is a

17 required element of proof in every Section 2 case, or in

18 an attempt case, of course, the reasonable likelihood of

19 monopoly power being attained -- and it also means

20 that -- it really brings us down to I think the main

21 discussion, the main subject of discussion, which is the

22 definition of monopolizing conduct, and, of course, that

23 is a much broader area, and let's see what light we can

24 shed on the exclusive dealing aspect.

25 Well, one of my colleagues, Steve Calkins, has


1 already alluded to the fact that if you look at

2 exclusive dealing cases, there are not many in which

3 plaintiffs win, and it is interesting that some of those

4 cases are really not Section 1 or Clayton 3 cases

5 anymore, they are Section 2 cases, oddly enough, in

6 which the decision-maker for one reason or another

7 failed to condemn exclusive dealing under Sherman 1 or

8 Clayton 3, but only under Section 2, and that would

9 include U.S. v. Microsoft, Lepage's v. 3M, sort of in

10 the margins of exclusive dealing, one of those forms of

11 bundling, and then we have heard about U.S. v. Dentsply.

12 Now, within the broader debate about legal

13 standards for monopolizing conduct, exclusive dealing I

14 think is more or less kind of a classic example. What

15 do we have to go on when somebody is challenged for

16 their conduct under Section 2? Well, we have Grinnell,

17 we have Aspen, exclusion on the basis of something other

18 than efficiency; we have Image Technical Services, not

19 the part that everybody has had seminars about and

20 talked about for years and years and years, and Salop

21 said this and somebody else said that and it is

22 post-Chicago -- no, it is pre-post-Chicago -- okay, it

23 is post-modernist Chicago, but the point is there is a

24 second part of Kodak versus Image Technical, which say

25 what you will about the tying part, the first part of


1 the Supreme Court opinion, there is that second part

2 that makes some extremely broad characterizations of

3 what it takes to -- broad and vague characterizations --

4 of what it takes to prove monopolization. That part of

5 the opinion was so good that when Image Technical got to

6 go back and have its trial, it did not even bother with

7 all the hard post-Chicago stuff in the first part. It

8 just relied on that great language in the second part of

9 the opinion. So, it is really a question of

10 deconstructing and coming up with a monopolistic conduct

11 standard that can be applied sensibly to the generality

12 of these cases.

13 Now, I will put all my cards right on the table

14 and say I am not one of those who says there is

15 salvation to be had in taking the vague language of

16 Grinnell and the vague language of Aspen and the vague

17 language of the second Section 2 part of Image Technical

18 versus Kodak and trying to put some kind of a

19 microeconomic overlay on it, whether it is no economic

20 sense, profit sacrifice, exclusion of equally efficient

21 competitor. I think all of those things can come in

22 very handy. I mean, if you see a monopolist doing

23 something that causes it losses, you are entitled to

24 inquire, is it an eleemosynary motive, was it a mistake,

25 or was the monopolist taking money and paying for


1 something, and was it a competitive restraint? So, I do

2 not want to suggest that those concepts are useless, but

3 I think they are not going to get us the distance to a

4 standard under Section 2 for judging exclusive dealing.

5 As a matter of fact, I am prepared to say that

6 as a general matter, any standard that is simply stated

7 and purported to apply to the generality of exclusive

8 dealing cases cannot possibly give you specific enough

9 guidance to decide any particular case. This is just

10 one of those situations where we are kind of stuck with

11 the dilemma that Steve referred to in the initial part

12 of his remarks when he was quoting from Standard Station

13 saying, well, you know, we would love to consider all

14 these justifications, but, you know, it would not be an

15 administerable rule of law, there is nothing you could

16 do with it. Therefore, we are going to have a per se

17 rule based on quantitative substantiality.

18 It is a little bit like my favorite footnote in

19 Topco. Remember United States v. Topco, which was a

20 horizontal case, and it was a bunch of independent

21 grocers who had banded together and had arranged to have

22 their own private label line of products to offer in the

23 grocery store so they could compete with A&P Ann Page

24 and Safeway's whatever, and the District Court had said,

25 well, this is very procompetitive as a rule of reason,


1 case dismissed, and the Supreme Court said, oh, no, oh,

2 no, when you are talking about a horizontal restraint --

3 and it was a territorial restraint in that particular

4 case -- what the Supreme Court said is you don't

5 consider all that stuff, it is per se, and then they

6 dropped a footnote that said, well, look, if Congress

7 would like to adopt a rule of reason for this kind of

8 restraint and send the courts off into the wilds of

9 economic theory -- that's the exact phrase they use in

10 that footnote in Topco -- Congress can go to that, but

11 we are not going to, per se illegal, next case. So, we

12 have got a similar situation here.

13 Exclusive dealing could be good, could be bad,

14 depends on a lot of different factors, very hard to

15 formulate a different -- a reformulation of a general

16 standard that is going to apply in all circumstances,

17 and so I have very little faith in any such

18 reformulation. I think we are just stuck, you know,

19 courts do what they do. You have got a difficult area

20 where it is hard to make a judgment. Actually, as I

21 think as I am going to talk about toward the end of my

22 remarks, which will be soon, what I am basically saying

23 is if the courts find it difficult to take such an

24 amorphous standard and apply it to this practice, what

25 we have to have is better courts.


1 Now, we have mentioned that defendants almost

2 always win. So what? So what? I have no great faith

3 in the numerology of one loss statistics. The real

4 question is whether anticompetitive conduct gets struck

5 down in these cases and procompetitive conduct is

6 exonerated, and by that standard, as I read the same

7 cases that Steve has obviously read -- and he has

8 probably spent a lot more time reading them than I have

9 and has read a lot more cases as well -- but I find it

10 very difficult to say that something is seriously awry.

11 I have cases where I would disagree with what is

12 going on, but there are two cases in the -- well, I have

13 talked about Microsoft, U.S. v. Microsoft, Lepage's v.

14 3M, U.S. v. Dentsply. I have listed -- have I listed in

15 my -- well, anyway, three cases I could name where the

16 defendants won, three recent important cases where the

17 defendants won, Pepsico versus Coca-Cola, this is the

18 New York case affirmed by the Second Circuit where

19 basically the Second Circuit said you do not get a trial

20 on the proposition that the reason quick-service

21 restaurants do not buy Pepsi-Cola is that Pepsi-Cola

22 cannot figure out a way to deliver the syrup to the

23 restaurants. Whatever reason there is for the relevant

24 market shares in quick-service restaurants for

25 carbonated soft drinks, it is not that Pepsi-Cola could


1 not figure out a way to get its product delivered.

2 Omega Environmental versus Gilbarco, I do not

3 know any more than what you, the average case reader,

4 knows. I had no involvement with that case. Then we

5 have Harmar Bottling, which is, again, a case that I do

6 know something about. I am not sure the facts bear the

7 characterization that Steve was giving it. I do not

8 want to get into a cat fight with him over that, but I

9 will just say that I think the result in that case was

10 correct, and so of the cases I know, of the cases I have

11 read about and tried to understand, I do not think you

12 can say that defendants are winning in cases where they

13 should not win.

14 So, you know, we need to figure out a way to

15 assess exclusive dealing efficiently, and basically, as

16 I say, my message is there is some exclusive dealing

17 that is good, some exclusive dealing that is bad.

18 Harmar took about 14 years to tell one from the other,

19 and my main message is that there has got to be a way of

20 getting to an efficient resolution of these cases much

21 more quickly. As a matter of fact, I would consider

22 whether -- I might regret this if it became a sound

23 bite, but if there is a sound bite I would give you,

24 let's have the antitrust enforcement mechanism, let's

25 adopt as a policy objective, that in the area of


1 exclusive dealing, we want to reduce the duration and

2 the expense of deciding whether exclusive dealing in a

3 particular case is good or bad. Let's reduce the

4 duration and expense by an order of magnitude so that a

5 Harmar, which took 14 years to litigate, takes, say, 14

6 months to litigate.

7 Now, in this column, I have very high praise for

8 the Ann Bingaman suit against Microsoft which resulted

9 in the 1994 consent decree. I know that there was some

10 investigation prior to the time that the DOJ got the

11 file in that case, but I remember being incredibly

12 impressed for two reasons with that effort. Number

13 one -- well, other than feeling that the result was

14 right. It was a consent decree, but I think it did the

15 right thing.

16 Number one, it was about exactly one year

17 between the time that the Department of Justice got the

18 file in that case and the date that the decree was

19 entered, and number two, it was a very specific,

20 targeted form of relief. It was a doable form of

21 relief. So, if you can do an exclusive dealing case

22 that quickly and come up with a result that concrete in

23 a year, it forgives almost any other defect that you can

24 find in that case, because on that time scale, you can

25 correct for your mistakes. You can, you know, do in


1 year two what you failed to do in year one, or vice

2 versa. So, litigation efficiency is an extremely

3 important consideration, and we ought to figure out ways

4 for a great increase in litigation efficiency.

5 One minute, that is exactly what I need.

6 So, here are some ideas for enhancing the

7 efficiency of this process, and I think a lot of the

8 tools are already at hand. Daubert, it has already been

9 used in an exclusive dealing context. Let's have more

10 of it. Let's make sure that expert testimony is forced

11 to go through and survive a plausibility test, the

12 Daubert standard. Let's make sure that the plausibility

13 formulation in Matsushita and Brooke Group, even though

14 that is relative to predatory pricing, a plausibility

15 test should also be applied to other types of antitrust

16 claims, including exclusive dealing, help filter out

17 losing claims early, and focus remaining claims on all

18 phases for the remainder of the litigation, so you are

19 not carrying forward speculative theories and going

20 through the wasteful discovery and legal motions and so

21 forth that that involves.

22 Second, expand the use of neutral expert or

23 expert panels, and I want to emphasize here, it is not

24 just in a strict Rule 706 sense, in other words, an

25 expert witness providing economic testimony to a judge


1 in a matter in litigation, like Fred Kahn's testimony in

2 the New York versus -- the Nabisco Brands case. That

3 was a very effective use of a 706 expert, but we need

4 ways to bring specialized knowledge about antitrust

5 cases, discovery, theories, the nature of the market, we

6 need to put those resources at the service of the courts

7 that are having these exclusive dealing litigation

8 things litigated before them.

9 And the last one I won't go through due to the

10 shortness of time, but the Manual for Complex Litigation

11 does contain a few things about antitrust, but perhaps

12 of the ideas that we could expand, the sort of helpful

13 guidance, the identification of issues, the suggestion

14 of efficiency-enhancing methods of resolving complex

15 litigation, expand it specifically in the area of

16 monopolization and exclusive dealing for the use of the

17 courts.

18 So, just to sum up, I do not think that our

19 exclusive dealing jurisprudence is in crisis. I kind of

20 like where the law is. Some exclusive dealing is good,

21 some exclusive dealing is bad, it is not per se legal,

22 it is not per se illegal, but if we could reduce the

23 time it takes to tell the difference between good

24 exclusive dealing and bad exclusive dealing by an order

25 of magnitude, I think that would be a very worthy goal


1 for the antitrust policy.

2 (Applause.)

3 MR. O'BRIEN: Thank you, Tad.

4 Okay, our next speaker, shifting gears to a

5 couple of economists, is Joe Farrell. He is Professor

6 of Economics at the University of California, Berkeley,

7 and he is a Fellow of the Econometric Society, former

8 editor of the Journal of Industrial Economics and former

9 President of the Industrial Organization Society.

10 Currently he's the senior consultant for Charles River

11 Associates.

12 Joe's published widely articles on a broad range

13 of topics in industrial organization and microeconomics,

14 including exclusive dealing. He has substantial policy

15 experience as well, having served as Chief Economist at

16 the Federal Communications Commission from '96 to '97

17 and Deputy Assistant Attorney General for Economics at

18 the Antitrust Division from 2000 to 2001.

19 Joe?

20 DR. FARRELL: Well, I am an economist. I am

21 going to talk about economics for a few minutes, and

22 then I am going to talk about the law. I feel all right

23 about this because I hear a lot of lawyers talking about

24 economics.

25 So, economics for the most part in antitrust


1 analysis has focused on the question, what should we do

2 if we knew really quite a lot about the case, okay? And

3 in the area of exclusive dealing, I think a bland and

4 very fair summary of economics in this area is both

5 efficiency and anticompetitive effects and explanations

6 of exclusive dealing are very possible, and on both

7 sides of that, the analysis is really quite subtle, and

8 I am going to spend a few minutes on this. In terms of

9 the efficiency explanations, I am going to focus on the

10 investment incentive theory, which I think Ben Klein is

11 also going to talk about a form of. In terms of

12 anticompetitive effects, I am going to talk about what I

13 think is the leading example, though not the only

14 example, of an economic structure to understand

15 anticompetitive effects of exclusive dealing.

16 So, in terms of the investment incentives, you

17 will often hear it said that exclusive dealing is

18 efficient if you have to motivate relationship-specific

19 investment or some such phrase as that, okay? As far as

20 I know, the state of the art in the economics literature

21 on these arguments is the article by Elias Segal and

22 Michael Whinston in the Rand Journal, 2000. They start

23 out by showing that in what appears to be quite a

24 general model, relationship-specific investments, that

25 is, investments that have no value outside the


1 relationship, are not -- repeat, not -- an efficiency

2 rationale for exclusivity.

3 They then continue to show that investments that

4 are not in that strict sense relationship-specific, that

5 have a spillover to deals between the customer and the

6 potential entrant, might or might not be an efficiency

7 rationale for exclusivity. It depends on quite a number

8 of things. It depends on who is doing the investment.

9 Is it the buyer or the seller? It depends on how it

10 spills over. Is it a complement or a substitute with

11 the efficiency of potential deals between the buyer and

12 an entrant? It depends on the bargaining structure

13 between the buyer and the seller. It depends on what is

14 the nature of any investment by us absent the exclusive

15 dealing. And that is all within their model. If you

16 step outside that model, it also depends on whether

17 their model sort of applies or sort of does not apply.

18 So, I am going to leave you for the moment with

19 the thought, how is a court likely to be able to

20 disentangle all this in addressing an asserted

21 efficiency rationale along the lines of investment

22 incentives?

23 Now, what about the other side of the courtroom,

24 divide and conquer exclusion, Rasmussen and Ramseyer and

25 Wiley, 1991, corrected, beefed up and radically improved


1 by Segal and Whinston in the American Economic Review,

2 2000, show that exclusion can profitably and harmfully

3 work against end users; however, although I think that

4 is very well understood and accepted, the fact is their

5 models involve buyers who are end users.

6 In most cases that I am aware of, exclusive

7 dealing is not a deal struck with end users. It is a

8 deal struck with retailers or distributors or someone

9 else intermediate in the value chain between the

10 manufacturer and the end users. That makes a lot of

11 difference.

12 So, interestingly, a year or two ago, there

13 appeared to be economics literature, two broadly

14 parallel articles, papers, one by Fumagalli and Motta,

15 which I believe has been published or is about to be

16 published in the American Economic Review, and one by

17 John Simpson and Abraham Wickelgren, and within the last

18 24 hours, I have learned about other articles by Yong

19 and Shaffer that may be somewhat along the same lines,

20 and both of these articles address the question, how

21 does the RRWSW theory of anticompetitive exclusive

22 dealing change when you recognize that the buyers in the

23 model, in practice, should be replaced by buyers who are

24 not end users?

25 Well, there are two forces, okay? One force is


1 that intermediate buyers, nonfinal buyers, actually do

2 not care that much if the price goes up or stays high,

3 provided it goes up or stays high to all of them,

4 because then it gets passed through downstream, okay?

5 How much that is true depends on the details of the

6 market structure and so on, but that tends to be true.

7 That lowers their resistance to things that maintain

8 monopoly upstream relative to what it would be if they

9 were end users. So, that you would expect would make

10 anticompetitive exclusive dealing easier.

11 Another force, however, is that if you have a

12 nonfinal buyer who holds out and does not sign the

13 exclusive deal, then an entrant can come to him and say,

14 "Aha, I will give you a lower price than all your tied

15 up rivals will be getting. You can expand. You and I

16 can meet my scale requirements, and you will make a

17 bundle of money." So, that dynamic potentially makes it

18 harder to have anticompetitive exclusive dealing.

19 Well, Fumagalli and Motta found conclusively

20 that it went one way, and Simpson and Wickelgren found

21 conclusively that it went the other way, and which way

22 Yong and Shaffer come out, I do not know yet. Which of

23 them is right and when? Well, I attempted to diagnose

24 this in my Antitrust Bulletin article last year. My

25 attempted diagnosis is that it depends on whether in


1 that last situation where you had one hold-out buyer,

2 the incumbent is then able to or does adjust the price

3 that it charges the tied buyers. So, I believe

4 Fumagalli and Motta assumed that it does not, and

5 Simpson and Wickelgren assumed that it does, or maybe it

6 is the other way around, okay?

7 When I put this tentative diagnosis to one of

8 the four economists -- and I will not say which one --

9 the response I got was, "Ah, that is interesting, I am

10 not sure." That is telling, I think, because it says

11 that it is kind of unlikely that a court is going to do

12 a very good job of disentangling all of these difficult

13 concepts. Now, the optimistic view is this is just the

14 beginning of the economic exploration of this topic, and

15 come the year 2010, we will understand it well and in a

16 way that is good enough for us to brief courts on it,

17 and maybe that will happen, okay, but I take from this

18 two things.

19 One is economics is making progress, that is

20 great, I hope to participate, but the other is, it is

21 pretty subtle and it will probably stay pretty subtle,

22 if not get more subtle.

23 All right, so we are doing antitrust under

24 uncertainty. We are not in the world where we can say

25 exactly what is going on and work out the welfare


1 consequences, okay? Let's take that as an assumption

2 for now.

3 Well, traditionally at this point economists

4 plunge into Bayesian mode and talk about type one errors

5 and type two errors and so on. Underlying what I am

6 going to say, there certainly is a Bayesian framework,

7 okay, but I am not going to talk explicitly in Bayesian

8 terms. I am going to talk in jurisprudential terms,

9 because my lawyer colleagues on this panel have been

10 talking economics, so I want to get back at them.

11 So, I am going to talk about the role of

12 presumptions and burdens of proof, and I am going to

13 talk about two presumptions that should be extremely

14 important in antitrust policy and about what I

15 personally think -- although I cannot prove -- is a very

16 worrying trend that has been taking place in the

17 relative strength of these two presumptions.

18 So, what are these two presumptions? Number

19 one, in economic policy generally, in market economies,

20 we have a laissez-faire presumption. The Government

21 should not intervene in stuff unless it is reasonably

22 sure that intervention will help. I think that is a

23 pretty good idea.

24 Number two, in antitrust particularly, we should

25 protect competition unless we are reasonably sure that


1 some alternative is better, okay? So, I think at a very

2 grand, 40,000-foot level, you can view a lot of what

3 goes on in antitrust jurisprudence as being a tug of war

4 or back and forth between these two presumptions.

5 Now, I put competition in quotes on this slide

6 for a reason, and that reason is when you look at it too

7 closely, things get a little out of focus, and you do

8 not exactly know what that word means, okay? And that

9 has led us, I believe, over the course of the decades

10 towards the tempting solution of redefining the word

11 "competition" to mean what is good. So, here is a test

12 of that, okay?

13 What happens when you hear someone refer to the

14 possibility that a merger to monopoly would reduce

15 marginal costs so much that it would be good for

16 efficiency and consumers? Well, if that were true,

17 let's say you knew it was true, it would be a good

18 thing. Would it be procompetitive? I think a lot of

19 people would say yes, because it is a good thing, but

20 that is ridiculous. It is not procompetitive. It is

21 pro-consumer, it is pro-efficiency. It is not

22 procompetitive.

23 So, if we are going to use words in their real

24 meaning rather than redefining them so that the

25 definition does our policy analysis for us, we have got


1 to be a little careful about doing stuff like that.

2 Now, of course, the antitrust law protects

3 "competition," so tautologically, redefining the word

4 would be a good idea, it would lead us to do good

5 policy, if we always knew what was going on, okay? So,

6 given that the law protected competition, it would be a

7 very smart move on the part of benevolent antitrust

8 enforcers and courts and so on to redefine the word

9 "competition" so that the law then protects whatever is

10 good, okay?

11 However, there is a problem with doing this. A,

12 we do not always know what is going on exactly, and B --

13 B only applies given A -- attempting to have a

14 presumption in favor of protecting competition makes no

15 sense if you define competition to mean what is good,

16 okay, because if you knew that something was good, you

17 would want to do it, and that is not a presumption in

18 favor of protecting competition. So, for there to be

19 any meaning to the presumption in favor of competition,

20 it has to be a presumption in favor of something that

21 has not yet been proved to be good, okay?

22 So, this I think casts an interesting light on

23 the slide that I heard this morning -- and I was not

24 taking notes on who said it -- but somebody said

25 something along the following lines, or if I misheard


1 it, it has certainly been said within the last week --

2 that because there are perfectly plausible efficiency

3 justifications for exclusive dealing, plaintiffs should

4 be required to prove that there is an anticompetitive

5 effect, okay? That, of course, would be obviously right

6 if we could always prove what is true, but if we cannot

7 always prove what is true, it is not obviously right.

8 It might still be right, but it is not obviously right,

9 okay?

10 So, in order to explore this, let me, with

11 tongue in cheek, put the shoe on the other foot, okay,

12 and let's suppose that we applied the same redefinition

13 to the laissez-faire presumption, okay? So, we have

14 this presumption that says the Government should not

15 intervene unless it is pretty sure that intervention is

16 a good thing, okay? So, now let's suppose that we

17 defined laissez-faire as the good outcome, and we

18 defined intervention as the bad outcome.

19 Now, if the Government wants to come along and

20 insist that you paint your bedroom walls blue, not

21 white, you can't say that is intervention, because you

22 have not proved that it is a bad thing, okay? Well,

23 that is obviously pretty stupid.

24 So, I come out of this thinking it would be a

25 good idea for us to make sure that words go on meaning


1 what they mean, and it is very dangerous -- it has had

2 some good consequences, but it is nevertheless very

3 dangerous -- to redefine words to make them do your

4 policy analysis for you.

5 So, antitrust intellectual history, to the

6 extent that I understand it -- in less than one

7 minute -- in the bad old days, anything that could be

8 presented as a reduction of competition was illegal.

9 That was bad, because quite often, things that can be

10 presented as a reduction of competition are actually

11 good. The good new days, we have got to analyze the

12 effects of things that seem to be capable of being

13 presented as a reduction in competition, because you

14 would not want to ban those things if they are actually

15 good, okay?

16 What I am worried about is the possibility that

17 we are drifting into the not so good new days where it

18 is difficult to prevent things that are in some sense

19 reductions of competition unless you can actually prove

20 that those things are bad. Now, of course, you would

21 not want policy to prevent those things unless they are

22 bad, but that is very different from unless you can

23 prove that they are bad.

24 Now, the final bullet on this slide, which is

25 quite important, I talked about these ideas very briefly


1 with some people in Europe over the summer, and they

2 were aghast. Why were they aghast? Because they said

3 we have spent years trying to move away from a

4 descriptive basis of liability towards an effects-based,

5 economics-based concept of liability, and now, you are

6 coming from over there and trying to undo that. Well, I

7 take that seriously, so I am not going gung ho on a

8 policy proposal here, but it does seem to me that if too

9 much burden of proof is being imposed, that is a

10 problem.

11 Let me finish with this slide, dark matter, do

12 the physics, okay? It is a good idea to intervene only

13 if intervention benefits efficiency or consumers. It is

14 maybe not such a good idea to intervene only if you can

15 specifically prove that and how it would do so, okay?

16 There are multiple benefits of competition in

17 most circumstances. Often, there are concrete,

18 predictable, provable price effects, okay? Merger

19 simulation has been a very powerful tool in exploring

20 that. There is also the much vaguer and harder to pin

21 down possibility that having a bunch of different firms

22 doing different things and independent of each other can

23 lead us to benefits that are much harder to prove or

24 even define or even point to ex ante, okay? I call this

25 the dark matter of competition policy, because as in


1 astrophysics, if it exists, it is quite likely to be a

2 lot bigger and more important than the stuff that you

3 can see. So, watch out when you are imposing burdens of

4 proof.

5 (Applause.)

6 MR. O'BRIEN: Thank you, Joe.

7 Our final speaker is Ben Klein. He is Professor

8 Emeritus of Economics at UCLA and a director at LECG.

9 Ben is an internationally recognized expert on antitrust

10 economics. He was a Professor of Economics at UCLA for

11 34 years where he published numerous articles on a range

12 of topics, including antitrust, contracts and

13 intellectual property.

14 He currently serves on the board of editors of

15 five academic journals. Over the past 25 years, he has

16 consulted extensively on antitrust issues and has made

17 numerous presentations to state, federal and foreign

18 regulatory agencies and courts.

19 Ben?

20 DR. KLEIN: Thank you, Dan.

21 I am going to be talking mostly about economics,

22 and although what I am going to say is subtle, you

23 should not reach the conclusion from what Joe said that

24 because the arguments are subtle that, therefore,

25 anything goes. Just find the economist that is going to


1 make the argument you want to hear. I think there is

2 truth out there. This is moving us along on coming up

3 with what is the economic foundation for some commonly

4 used procompetitive justifications.

5 This is a paper that I am working on with Andres

6 Lerner. The paper is posted on the web site, and I

7 think it is important to go through these procompetitive

8 justifications in terms of the economics, because the

9 danger I see is the exact opposite one. I think that we

10 are moving in the direction that if you find a practice

11 that does not have efficiencies, it is becoming a

12 sufficient condition, if it is something that is being

13 used by a firm, a large firm, it is a sufficient

14 condition for antitrust liability, because the very

15 nature of an exclusive dealing contract is

16 "exclusionary," and then when you get to the balancing,

17 you have nothing on one side of the scale.

18 Although the paper discusses many exclusive

19 dealing cases, we concentrate on Dentsply, and that is

20 what I am going to concentrate on today, and it is

21 because the court in that case used economics to reject

22 two very common procompetitive justifications, both free

23 riding and this undivided dealer loyalty justification,

24 and the principles that I am going to be giving you here

25 can be applied to a number of very different claimed


1 justifications, and we do it in the paper.

2 So, in terms of Dentsply, as I said, Dentsply

3 illustrates that actually the economic foundations for

4 procompetitive justifications are actually pretty

5 narrow, and the Court rejected Dentsply's claim, in

6 particular, that exclusive dealing was used to prevent

7 dealer free riding on manufacturer-supplied promotional

8 investments. This is the classic Howard Marvel

9 rationale, where the manufacturer makes investments in a

10 dealer, you know, like they build out a dealership or

11 engage in dealer training, and then the dealer uses

12 those manufacturer investments to sell a rival product,

13 and that is the classic free riding argument. The Court

14 rejected that, and the Court rejected the undivided

15 loyalty argument, that somehow you give somebody an

16 exclusive so they will more actively promote Dentsply's

17 product.

18 The Court rejected the free riding rationale

19 basically because the Court found it was contrary to the

20 facts, that number one, Dentsply did not make any

21 investments in the dealers that they could then free

22 ride on by using them to sell rival products. There was

23 no evidence, essentially no evidence in the case, that

24 the Dentsply dealers were actually switching buyers to

25 rival products. And finally, that there was testimony


1 by Dentsply executives that if there was not an

2 exclusive, they actually would have invested more -- you

3 see, the usual economic argument is the purpose of

4 exclusive dealing is to encourage the manufacturer to

5 make investments, and one way it is encouraged to make

6 investments is to prevent free riding that it knows that

7 these investments are going to be used to sell its

8 product, and the Court said, you know, the Dentsply

9 executives actually testified that if we did not have

10 exclusive dealing, we would have had to make more

11 promotional investments.

12 In terms of the other argument, the Dentsply

13 Court rejected the undivided loyalty argument, and here

14 it was not really just the facts. It was basically the

15 theory that this theory about enhancing dealer services

16 cannot be a justification for exclusive dealing, because

17 in general, competition between dealers is going to lead

18 them to supply the desired quantity of promotional

19 services, as the Court said, the dealers have the

20 incentive in competing with other dealers to make sure

21 that they supply the right kind of services.

22 See, basically the problem that Dentsply ran

23 into is although this undivided loyalty argument has

24 been accepted by a number of courts, Judge Robinson in

25 this case knew a lot of economics, and in particular,


1 she knew Howard Marvel's argument and had read the

2 article, and Howard Marvel was the expert that Dentsply

3 had hired for this, and she said, no, even in your

4 expert's article, he says that you can generally leave

5 it up to competition to put dealers to supply the right

6 services. It is only when you have this problem, this

7 inter-dealer free riding problem described in Sylvania,

8 you know, and that is a problem where the customer goes

9 to one full-service dealer and gets some kind of dealer

10 services and then goes to another dealer and buys the

11 product, you have that inter-dealer free riding problem,

12 and in that circumstance, maybe competition among

13 dealers will not give you the right quantity of dealer

14 services, but that is a problem that would not be

15 corrected with exclusive dealing, because even if you

16 had exclusive dealing and you had this kind of problem,

17 the exclusive dealer would say, no, get the services

18 from somebody else and then come and buy the

19 manufacturer's product from me.

20 So, as I said, although this rationale has been

21 accepted by a number of courts, Judge Robinson said, you

22 know, basically you can leave it up to competition, and

23 this undivided loyalty makes absolutely no economic

24 sense.

25 In contrast to basically the established


1 economics, I think the expanded economic framework that

2 I am going to present here shows that these arguments

3 make sense, that free riding is much more general than

4 you would think, and the dealer undivided loyalty makes

5 sense, and it is based upon two common sense business

6 propositions.

7 Number one, that manufacturers often want their

8 dealers, even dealers that are competing with one

9 another, to supply more promotion than the dealers would

10 independently provide on their own, and number two, that

11 exclusive dealing by creating this undivided dealer

12 loyalty actually increases the dealer's incentives to

13 supply these desired services and to more actively

14 promote the manufacturer's products.

15 So, in terms of the logic of what I am going to

16 do, first I am going to discuss the first proposition,

17 and hopefully people have been here in the morning and

18 heard Josh Wright, who has done this already, but

19 basically the first proposition is manufacturers, in

20 general, cannot leave it entirely up to the dealer

21 competition to get the quantity and the type of services

22 they want supplied, and the logic of the argument is

23 there is another step where, therefore, they have to

24 contract with their dealers, either explicitly or

25 implicitly, to solve this problem and to make sure that


1 they adequately promote their product. That leads to

2 these free riding problems, which I will discuss are

3 much broader than the classic Marvel free riding

4 problems. And then finally, that exclusive dealing is

5 commonly an element in those contractual arrangements

6 that gets the individual dealers' incentives then

7 aligned more with the manufacturer's incentives.

8 So, let me do the first proposition first, that

9 manufacturers often want their dealers to supply more

10 promotion than the dealers would independently decide to

11 provide, and the basic reason for this is that the

12 dealers do not take account of the manufacturer

13 profitability on incremental sales, that the dealer does

14 something that increases the manufacturer's sales, and

15 the dealer gets only a part of that incremental profit,

16 in many cases only a very small part of the incremental

17 profit.

18 Now, in general, this is not a problem for

19 dealer price and nonprice competition that has

20 significant inter-dealer quantity effects. So, in

21 general, when a dealer provides a desirable service like

22 free parking or lowers its price a little bit and makes

23 a little bit more sales, even though they might have a

24 small margin in terms of the total profit being earned

25 by the manufacturer and retailer together when they make


1 that extra sale, because they are getting consumers to

2 switch from other dealers, because there is large

3 inter-dealer effects, you get an equilibrium where you

4 get the desired quantity of the services provided, but

5 with promotional activity, the primary effect is not

6 really inter-dealer, but it is primarily inter-brand,

7 that you just make an extra sale for the manufacturer,

8 and there are no significant inter-dealer quantity

9 effects. Then you have this problem where the dealer,

10 by not taking account of the incremental profit, is

11 going to supply less than the desired promotional

12 services of pushing the manufacturer's product. In

13 addition, dealers cannot charge consumers directly for

14 those services, because the promotion is, in effect, a

15 price discount.

16 So -- I am going to have to go faster --

17 manufacturers solve this problem -- although I am going

18 to be talking about violating these contracts, I can

19 always violate, you know, this one --

20 MR. O'BRIEN: There is no red string we can pull

21 --

22 DR. KLEIN: No, there is no self-enforcement

23 problem here, although -- anyway, I am wasting my time.

24 Manufacturers solve this problem of insufficient

25 dealer promotion by contracting with and compensating


1 dealers for providing increased promotion, and the

2 contract may be explicit or it -- you know, in plenty of

3 the cases, like in Standard Fashion, they explicitly

4 said you have to have a certain amount of display space,

5 you have to have a "lady attendant" there full-time,

6 they used a few words like that. Most of the times it

7 is really understood that you are going to make your

8 best efforts, and they compensate dealers in these

9 things by giving them a valuable distributorship in the

10 sense that if they get terminated because they are not

11 pushing the product adequately, they are going to lose

12 this future rent stream, and the threat of termination

13 is what gets them to perform as desired.

14 However, because dealers are contracting to

15 supply more promotion than they would otherwise, you

16 know, do in their own independent interests, there is an

17 inherent problem in that they have an incentive to

18 violate the contract and free ride on the manufacturer's

19 compensation arrangement, basically because you are

20 getting a valuable dealership, like in Beltone, they

21 gave them an exclusive territory. In Standard Fashion,

22 they had minimum resale price maintenance. Whatever it

23 is, you have something valuable, but you are getting it

24 on all your sales, and you therefore have an incentive

25 just to do that pushing at the end and save the cost if


1 you are a dealer, and still you are getting most of the

2 compensation.

3 In terms of this contract, dealers may violate

4 the contract and free ride in three distinct ways, and

5 the first way is the standard case where the dealers use

6 the manufacturer-supplied investments to sell rival

7 products, and that is part of the contractual

8 arrangement. Look, we will give you these complementary

9 assets to help you push our product, and that is one

10 that you know about, but there is two other free riding

11 problems.

12 Second is the dealers may just use the

13 manufacturer paid for promotion to sell rival products,

14 that they are being compensated with this valuable

15 dealership, and on the margin, they are just going to

16 switch, and the profit incentive is really the same as

17 one, but you do not have to find these manufacturer

18 assets there.

19 And the third one is the dealers may just

20 under-supply the manufacturer's paid-for promotion, as I

21 said, because on the margin, they are getting paid on

22 all these inframarginal sales, and on the margin, it

23 really does not pay for them to spend all this money on

24 pushing the products on the margin if it was not for

25 this contract.


1 Dealer free riding need not involve manufacturer

2 investments or dealer switching. That is the

3 implication of this. So, for example, in free riding

4 one, which is the one you all know about, that one

5 involves manufacturer investments and dealer switching.

6 That is what the Court in Dentsply said, there is no

7 free rider problem here. But free riding, too, the

8 dealers are just using the paid promotion to sell the

9 rival products, and that one can occur without any

10 manufacturer investments whatsoever. They are just free

11 riding on the compensation arrangement.

12 Free riding number three, where dealers are

13 undersupplying what the manufacturers are paying for,

14 that one occurs without any manufacturer investments or

15 without any dealer switching, okay, and exclusive

16 dealing may be used to mitigate all thee forms of free

17 riding, and it prevents free riding types one and two by

18 just preventing the switching of sales to rival

19 products, and it prevents free riding number three by

20 creating this undivided dealer loyalty by promoting the

21 incentive of the dealers to promote the manufacturer's

22 product more intensively that aligns the incentives.

23 So, how does exclusive dealing, that third type,

24 how does the exclusive dealing increase the dealer's

25 incentive to promote? And remember, we are operating in


1 the context, you know, why did the Dentsply Court reject

2 this as making absolutely no economic sense? And that

3 is because there is all this competition between

4 dealers, and that is all that is necessary to get the

5 services provided unless there is a Sylvania type

6 problem, and the example that we go through in the paper

7 is this.

8 Consider this case where a customer is thinking

9 about buying a car and is leaning towards the purchase

10 of a Honda, and he goes into a Toyota dealership to

11 check out the Toyota, but really, you know, it is -- but

12 just to make sure, let me just check out the Toyota.

13 So, that is the hypothetical example.

14 Then I have this -- look at that. So -- and

15 under that -- there is a Honda and there is a Toyota,

16 and Mdh is the profit margin that the dealer earns if it

17 sells a Honda, and Mdt is the profit margin for the

18 dealer if it sells a Toyota, and the Toyota dealer is

19 deciding, what about this, even though they are leaning

20 towards the Honda?

21 Well, a nonexclusive dealer will not make its

22 best efforts to sell the Toyota if it has both cars

23 there, and basically -- now, do not get scared -- but

24 the dealer is going to choose a level of Toyota

25 promotional service as S that maximizes its


1 profitability. So, it chooses S, that maximizes the

2 profitability, which is the difference between the

3 margin on the Toyota minus the margin on the Honda,

4 times the probability that they will make the Toyota

5 sale if he starts telling them how great the Toyota is,

6 whether they will buy the Toyota, and that probability,

7 p(S) is a positive function of how much S the person

8 chooses, minus the cost of supplying S, and obviously

9 there is a positive marginal cost. The more S, the

10 higher those costs.

11 And since it costs the dealer less to sell the

12 Honda in this nonexclusive context, the dealer can earn

13 a higher net profit margin on selling the Honda, the

14 dealer goes to the one where the marginal cost of

15 providing additional services is equal to the

16 probability -- increased probability, as you supply --

17 of making the Toyota sale as you're supplying more

18 services, times the difference in the margin between the

19 Toyota and the Honda, and you can assume that they could

20 sell the Honda -- the customer comes in and sells -- and

21 wants to buy the Honda, and the salesperson can at a

22 zero cost sell the Honda and say, you know, you are

23 right, Honda is much better. Come here, I will get you

24 a good price.

25 So, under these circumstances -- for example, if


1 the dealer's margin on the two cars were the same, so

2 that Mdt and Mdh were the same number, that difference

3 would be zero, and clearly the dealer would supply

4 absolutely no services in trying to sell the Toyota. It

5 would be cheaper for the dealer to just write up the

6 sale for the Honda. But by selling the Honda rather

7 than promoting the Toyota, the dealer is free riding.

8 He is engaging in that third type of free riding that we

9 were talking about. The dealer is not switching. The

10 dealer is not actively promoting the rival Honda brand

11 as an alternative to Toyota, you know, for customers who

12 come in and want Toyota, as occurs in free riding one

13 and two. Instead, the dealer is violating the implicit

14 dealer contract for the Toyota by failing to actively

15 promote the Toyota automobiles.

16 Alternatively, if it was an exclusive, you know

17 that undivided loyalty is going to lead dealers to

18 expand their promotional efforts, and it is just going

19 to go to the point where the marginal costs of

20 additional efforts in pushing the Toyota is exactly

21 equal to how much it can make on the Toyota, times the

22 increased probability that the promotion makes it more

23 likely that they will make the sale. So, undivided

24 loyalty is clearly in that case going to lead to that,

25 and that is what you sometimes see courts saying, you


1 know, that if you do not make the -- you know, what

2 happens if you are an exclusive Toyota, basically it

3 means if you do not sell the Toyota, you do not make any

4 sale, and so it is common sense -- and, you know, this

5 is the business people who understand this -- it is

6 common sense that undivided loyalty is going to give you

7 an incentive to promote more, and in the paper, it is a

8 function of -- you still do not get to the point where

9 the dealer has the right incentive in terms of

10 maximizing the total profit of the manufacturer and

11 dealer together. That is the last thing on the

12 left-hand side. So, they still have to have these --

13 the manufacturer still has to have these implicit

14 self-enforced contracting and -- to go all the way to

15 the end, but basically the role of exclusive dealing is

16 that it aligns the incentives that are here.

17 So, I am done. The lessons, other than that I

18 put too much down here, okay? Lesson one, the Court's

19 rejection of Dentsply's procompetitive rationale is an

20 example of a common error that I think occurs in cases

21 of trying to fit the facts of a case into a preconceived

22 economic model rather than developing a model to fit the

23 facts of the case, and the preconceived theory, economic

24 theory here, that the Court adopted was basically, you

25 know, interdealer competition will lead dealers to


1 supply the type and quantity of promotional services,

2 unless you had that Sylvania type free riding problem,

3 and -- you know, because there are more likely to be

4 valid procompetitive justifications for exclusive

5 dealing, one of the implications I think is that this no

6 economic sense test is less likely to be a useful test

7 for antitrust liability, that there may be efficiency

8 justifications for exclusive -- people talk about the

9 Dentsply case as an easy case because there is nothing

10 on one side of the scale. There is obviously something

11 on one side of the scale is what I am trying to say, but

12 clearly, even though there is an efficiency

13 justification, you may have anticompetitive effects.

14 I think that the facts of that case, there were

15 significant anticompetitive effects, and Jonathan

16 Jacobson makes this point in his excellent latest

17 article in the Antitrust Law Journal. What he doesn't

18 do is he does not answer the Court's finding that there

19 was absolutely no economic basis for Dentsply's

20 undivided loyalty and free riding justification. So, in

21 that case, you would not get the wrong answer if you

22 used the no economic sense test, but the only reason you

23 do not get the wrong answer is because you do not really

24 understand the procompetitive justifications.

25 So, as I said in the beginning, I think the


1 greater danger is not that -- you know, the way some

2 people are advocating this no economic sense test as a

3 necessary condition for antitrust liability. I think

4 the danger is that the courts are going to use a no

5 economic sense test as a sufficient condition for

6 antitrust liability when a large firm uses exclusive

7 dealing, and it is not only that I am giving you that

8 there are other valid procompetitive rationales, but I

9 think as economists and as regulators we have to be more

10 humble that just because we have not figured this out

11 yet, there is lots of other procompetitive efficiency

12 justifications, and we cannot assume that the purpose of

13 a restraint is anticompetitive.

14 How much did I violate the contract by?

15 MR. O'BRIEN: Ten minutes.

16 (Applause.)

17 MR. O'BRIEN: Okay, we are going to break until

18 about five past 3:00, okay?

19 (A brief recess was taken.)

20 MR. O'BRIEN: Okay, let's get started.

21 Okay, I would like to start out our sort of

22 post-speech session here by asking folks if they would

23 like to comment on the remarks of others on the panel,

24 and I guess I will ask for a volunteer first rather than

25 being systematic about it. We will find out who has the


1 most burning comments to make about what someone else

2 said.

3 Okay, Joe.

4 DR. FARRELL: Well, I have one question for a

5 fellow panelist, which is relatively specific, I think.

6 Ben, in your model, you didn't have time to

7 present all of it, but I would like to ask, have you

8 offline, as it were, closed the loop and shown actual

9 harm to buyer, or is it just that the buyer who was

10 leaning towards buying a Honda ended up buying a Honda

11 and, of course, the Honda -- Honda likes that, the

12 dealer apparently likes that, the customer seems to like

13 that, although the welfare economics of this promotion

14 stuff, of course, are a little subtle. Toyota, of

15 course, does not like it.

16 Where does this go and how does the whole thing

17 play out with and without exclusive dealing as opposed

18 to just Toyota would like S to be higher in the short

19 run?

20 DR. KLEIN: All we do in the paper is present

21 the procompetitive efficiency justification. We do not

22 do the other side of the scale in terms of is there any

23 anticompetitive effect. In some cases, there will and

24 in some cases there will not be an anticompetitive

25 effect, and, you know, and as I suggested in Dentsply,


1 even though there may have been a legitimate

2 procompetitive rationale, forget undivided dealer

3 loyalty in that case, that does not mean that

4 arrangement was not, on net, anticompetitive and harmful

5 to consumers ultimately by creating a barrier to entry

6 to competitors.

7 But there are so many cases out there where we

8 know -- I mean, the case I love is this Joyce Beverages

9 case that I am certain Tad knows about, where you

10 have -- you have RC Cola having their distributor only

11 have one cola, the RC Cola, because they want undivided

12 loyalty. Well, in that case, RC Cola has 5 percent of

13 the cola market and a lot smaller share if you define

14 the market more broadly to have all carbonated drinks.

15 So, in that case, we clearly know there is no

16 anticompetitive effect.

17 But basically there was this mystery in the

18 literature, why are they really having an exclusive

19 dealing arrangement there, because there does not seem

20 to be any specific investments, and there does not seem

21 to be this dealer switching, but what they want is when

22 the salesman goes into the supermarket, that they are

23 going to push RC Cola and not any other brand.

24 So, if you want to get the ultimate question,

25 that would depend upon the particular case, and you


1 would have to examine that particular case.

2 DR. FARRELL: But you say you only do the

3 procompetitive justification. What do you demand of it

4 in order to call it a procompetitive justification, just

5 that Toyota would like it?

6 DR. KLEIN: Well, look, I do not want to get

7 caught up in a language thing about --

8 DR. FARRELL: Okay, sorry, no.

9 DR. KLEIN: -- you know, we will do linguistic

10 philosophy later. My feeling -- all I mean by it is

11 that somebody doing what you would consider the right

12 thing or the good thing or something and balancing it, I

13 am just looking -- I am just presenting an economic

14 foundation for this legitimate procompetitive

15 justification.

16 I mean, the crazy thing is if you look in the

17 marketing literature, people are talking about this all

18 the time. It is just economists, you know, a little bit

19 of economics can be a very dangerous thing, and it is

20 only the economists that say competition should give you

21 the services, everything is fine. So, if you talk to

22 business people, marketing people, they all know that

23 this makes -- and it makes a lot of common sense.

24 So, in some sense, as I said, Dentsply was

25 unlucky enough to have the judge that knew economics,


1 and that is the only reason they got into problems in

2 terms of the procompetitive justification, plus they

3 were unfortunate enough to choose an expert that

4 explicitly wrote in his article that the argument makes

5 absolutely no sense. So, he could not present -- he did

6 not -- Howard did not present the argument at trial, but

7 the company did in terms of answers to interrogatories,

8 and they said, what are you talking about? Your own

9 expert says this makes no economic sense.

10 And then the other interesting thing about it,

11 and this is the connection between anticompetitive and

12 procompetitive justifications is strange, because the

13 Justice Department -- and Gail would know this -- the

14 Justice Department, in trying to demonstrate the

15 anticompetitive effect, spent all this time in their

16 findings of fact to show how important this dealer

17 channel was to promoting the Dentsply products and how

18 rivals would be at a competitive disadvantage because

19 they did not have access to that channel.

20 So, you just look at all the findings of fact,

21 and it not only demonstrates that there was a

22 significant potential anticompetitive effect, but it

23 also demonstrates that there is a significant

24 procompetitive justification for motivating the dealers

25 to do a good job. So, you have that tension, but


1 basically -- well, maybe I should not monopolize this

2 thing.

3 MR. O'BRIEN: I mean, I have a follow-up for Ben

4 on this point. I mean, in talking about Dentsply and

5 just more generally your theory, there were two types of

6 free riding beyond the traditional one that you cited.

7 Dealers can free ride, effectively, on their own

8 promotion on behalf of a manufacturer, right, which

9 maybe they are doing in conjunction with the

10 manufacturer or somehow they have arranged a contract to

11 get that done, and the other one was that dealers may

12 violate this implicit contract by just under-investing

13 rather than free riding by steering customers to a

14 rival.

15 I am wondering if you have any specific evidence

16 that you can cite from the Dentsply case, if your

17 knowledge of the cases is deep enough, that one of these

18 two types of free riding that you identify in your paper

19 was actually present.

20 DR. KLEIN: Well, I do not think -- well, I

21 think the Marvel type free riding was not present, and

22 he did try to present an argument, and I think the facts

23 made it very difficult for him, and it is too bad that

24 he is not on the panel, because he would disagree, so I

25 think the first free riding did not exist, and I think


1 the second free riding did not exist basically because

2 there was no switching to rival brands.

3 I mean, I think there was one example where

4 there was some disagreement about whether they tried to

5 switch it to someone else. So, I do not think those

6 other two were there, but in terms of the third one, all

7 the evidence you need for that is that promotion is

8 important for the manufacturer to make sales, and as I

9 said, the Justice Department went to great lengths in

10 terms of trying to demonstrate the anticompetitive

11 effect to demonstrate the existence of that, so that is

12 there, and then all you need in addition to that is that

13 incremental sales are profitable for the manufacturer,

14 and those two conditions were clearly met in Dentsply.

15 DR. CALKINS: The great thing about the Howard

16 Marvel theory is that it is one that we lay people can

17 understand, namely, that it is good for everybody for a

18 manufacturer to run ads saying you can have your hearing

19 improved by getting it tested and going to a dealer and

20 getting this new improved kind of technology to use for

21 your hearing aid, and that drives consumers to go to the

22 dealership to try it out, and that is good for consumers

23 because they are finding out information.

24 It is good for the overall industry because

25 total sales of hearing aids will go up, because all


1 these consumers are being driven into the dealership and

2 are getting their ears tested, and it is all sorts of

3 wonderful stuff. And then, if, when the consumer gets

4 there, there is the old bait and switch and they are

5 sent off to buy the el cheapo discount brand, well, the

6 bad consequence of that is there will be less of that

7 advertising about the new, improved technology and, you

8 know, science of hearing aids and such, which is

9 something that is good for the whole industry, good for

10 consumers, good for everybody, it will now be lost,

11 because the manufacturer will not spend money on that.

12 So, you can easily tell a simple layperson

13 story, if you let everybody get switched off, you will

14 no longer have those ads being run, and I look forward

15 to reading the article, but merely saying that -- I

16 mean, if you then say that if you have exclusive

17 dealing, it is good for RC Cola because they are going

18 to make more money and have more sales, well, I can have

19 a warm feeling about that just because they have got a 5

20 percent share, and God knows, if you are RC Cola, you

21 are going up against people that you need all the help

22 you can get to go up against them. So, we can see

23 there.

24 You know, if you were to tell the RC Cola story

25 where you had somebody that had an 80 percent share and


1 climbing, before I then sat back and said, boy, I am

2 really concerned about maybe intervening and causing

3 harm here, I would like to at least make sure I

4 understood what is the equivalent of the lost nice

5 advertising that is going to happen if you intervene in

6 that type of situation.

7 DR. KLEIN: Well, Steve -- can I answer it,

8 then? I mean, I agree with you on your main point,

9 Steve, that, you know, with RC Cola, that we can be

10 pretty much assured that inter-manufacturer cooperation

11 or competition is going to pass on these benefits to

12 consumers, but if you are talking about -- the analogy

13 is really identical about lost advertising, because it

14 is either lost advertising by the manufacturer or lost

15 advertising by the dealers. It is just in some cases,

16 it is efficient for the manufacturer to do the

17 promotion, and in some cases, it is efficient for the

18 dealer to do the promotion, and if you do not have the

19 exclusive in one case, you do not get the manufacturer

20 advertising, and in this case, you do not get the dealer

21 pushing the product, and if you think that there is a

22 benefit from lost advertising, then it is totally

23 analogous in the two cases.

24 You know, if you start to do a calculation --

25 and you -- even in the standard case with the


1 manufacturer, you know, when the consumers are switched

2 to the discount brand, they almost always pay a lower

3 price. It is not -- you know, they are not being

4 deceived, that they think they are getting the higher --

5 DR. CALKINS: I understand. I think my problem,

6 and I will confess, I was sitting here with my back to

7 the screen, but I understood the Marvel advertising of

8 hearing, you know, development is a good thing. When

9 you hold out as your public good having a used car

10 salesman sit there and harass you into switching from

11 this model to that model, and I being a kid from

12 Detroit, noticing that you are using entirely Japanese

13 brand models, I...

14 MR. O'BRIEN: At the risk of allowing Ben more

15 monopolization time here, I just want to push it just a

16 little bit further on the Dentsply, and if others do not

17 have a question right off the bat here, and you said

18 that this third type of free riding, which is that you

19 would under-invest as a retailer --

20 DR. KLEIN: Right.

21 MR. O'BRIEN: -- you think was present, but the

22 evidence that you cited for that was that investment was

23 important, and that does not seem to demonstrate to me

24 that they actually would have necessarily under-invested

25 but for the exclusive dealing arrangement. You know,


1 was there another way for them to ensure that

2 investment? I mean, it seems to be a bit of a leap to

3 me.

4 DR. KLEIN: Well, that's --

5 MR. O'BRIEN: You know, and that sort of

6 statement strikes me as, you know, it might not be hard

7 to argue that that efficiency is there in almost every

8 case.

9 DR. KLEIN: No, that is a problem that you have

10 with these cases. That is why I said you cannot

11 adopt --

12 MR. O'BRIEN: A no economic sense test?

13 DR. KLEIN: Yes, you know, because I think the

14 efficiency is -- I would not say universally present,

15 but it is a motivation. I forgot what your other

16 question was, but, you know, it is important -- if it is

17 important to the manufacturer, we just know from the

18 economics that if there is an exclusive, the incentives

19 are going to be aligned, and if they do not make the

20 sale of that product, they are going to not make any

21 profit. So, you do know that they are going to push it

22 more.

23 So, I mean, it just follows logically, but you

24 would need to see what they adopted -- oh, yeah, so

25 you -- there may be a less restrictive way, and then we


1 can talk about a less restrictive standard here if that

2 is the question you want to move to, but in cases where

3 it looks like the practice might have also some

4 foreclosure problems and anticompetitive effects -- I

5 hope I am using the right language -- you may impose

6 this burden on the manufacturer to come up with a less

7 restrictive way of doing it, and, you know, maybe they

8 chose this not just because of the efficiency effects

9 but also because of the -- it increased their market

10 power, so...

11 MR. O'BRIEN: Okay.

12 DR. KLEIN: I mean, it makes it very, very

13 difficult in terms of this balancing. The important

14 point is, you know, you are not going to have these easy

15 cases anymore where there is nothing on -- I mean, you

16 will still have easy cases where you do not have the

17 anticompetitive effect on one side of the scale, but you

18 are not going to have these cases, I think, if you

19 accept this where there is nothing on the other side.

20 MR. O'BRIEN: Okay. Do any other panelists have

21 any questions for any of the other panelists?

22 (No response.)

23 MR. O'BRIEN: If not, we have a series of

24 propositions we would like to walk through to see where

25 we might be able to reach consensus, where we have open


1 issues, and hopefully this will spawn some additional

2 conversation about both what was said during the session

3 and perhaps some new things.

4 So, let me start with -- where is our -- oh, you

5 have got it, okay. Let's go to the first proposition.

6 Okay, I am going to read it, because we need to read it

7 for the transcript here.

8 Exclusive-dealing arrangements are analyzed

9 under the rule of reason.

10 First, does everybody agree -- and this is

11 really more for the lawyers -- that that is the way the

12 analysis of exclusive dealing goes today?

13 DR. CALKINS: Yeah. I mean, that -- yes -- yes,

14 I'll agree to say that, and B, for a whole lot of the

15 cases, it is consistent with the general idea that under

16 the rule of reason, the defendant always wins.

17 MR. O'BRIEN: Okay. So, nobody disagrees with

18 that point. Well, perhaps the point that was just made,

19 but nobody disagrees with the proposition, correct?

20 Does anybody think that there are exclusivity

21 arrangements that should be per se illegal?

22 (No response.)

23 MR. O'BRIEN: No, I guess that is the answer.

24 DR. KLEIN: Move on.

25 MR. O'BRIEN: No.


1 Does anybody think there are exclusivity

2 arrangements that are always or nearly always

3 procompetitive and thus good candidates for a safe

4 harbor?

5 DR. CALKINS: Well, presumably a very small

6 exclusive would be -- would fit anybody's idea of a safe

7 harbor.

8 MR. O'BRIEN: And when you say "small

9 exclusive," you mean a small percentage of the market

10 or --

11 DR. CALKINS: Yeah, it is very -- it is hard to

12 imagine a court or an enforcer being concerned about an

13 exclusive below -- choose your figures. Some might

14 choose 20 percent, some might choose 30 percent, some

15 might choose 40 percent, but I think everybody would

16 agree that below some percent, no agency should worry

17 about it, and no court should find illegality unless,

18 you know, you have some reason to think that that number

19 is just, you know, totally misleading and the real

20 number will be totally different in six months when the

21 contracts kick in or something.

22 MR. O'BRIEN: Okay, fair enough.

23 Anybody else? That one was --

24 DR. KLEIN: I would like to -- I would like to

25 ask Steve a question on this one. You know, your


1 opinion about the foreclosure standard somehow being

2 lower when it comes to Section 2 rather than Section 1,

3 I mean, if somebody is a monopolist or is likely to be a

4 monopolist, I could see that it is more likely that they

5 are going to meet the critical share number, but why

6 should that critical number, whether you say it is 40 or

7 whatever, why should somehow it be lower? It sounded

8 like that is what you said from your presentation,

9 should be a tougher standard.

10 DR. CALKINS: If I did, I misspoke slightly.

11 What I meant to say is that -- well, specifically, is

12 that in the Microsoft case, the defendants argued that

13 because this practice is lawful under Section 1, it

14 must, as a necessity, be lawful under Section 2, and I

15 was just saying that I do not think that is correct,

16 that, you know, take your extreme of a dominant firm

17 that everybody would agree is a monopolist on the one

18 hand, and on the other hand, your RC Cola kind of a

19 thing. I am not saying whether or not, you know,

20 exactly where one would say there is a difference, but I

21 would think that one should be much more likely to be

22 concerned about something being done by a dominant firm

23 that is --

24 DR. KLEIN: Right, obviously, but why should

25 there be a different standard under Section 2 than under


1 Section 1? I mean, I think we are in trouble here

2 basically because Justice did not appeal the Section 1

3 no liability in Microsoft and Dentsply, and if you read

4 the Court, you know that the appeals court would have

5 overturned both of those things, but -- you know, and

6 then I think we would be in a lot better shape, but the

7 idea that somehow we should have a different standard

8 and principle when you are doing the first step of a

9 Section 2 -- I agree, if somebody is a dominant firm,

10 they are much more likely to have anticompetitive

11 foreclosure under Section 1 and under Section 2, but why

12 should there be a lower hurdle showing the

13 anticompetitive effect under Section 2?

14 DR. CALKINS: Well, part of this goes -- I mean,

15 in all of this, it is trying to make a judgment about

16 how likely a particular practice is to be harmful to

17 competition, and I was just saying that -- well,

18 specifically, is that there are a whole series of sort

19 of ways that firms with fairly modest market shares have

20 been able to persuade courts to get rid of exclusive

21 dealing cases, but where you have a dominant firm, I am

22 not saying that there is a magic difference. I am just

23 saying that, as you recognized, you would think longer

24 and harder about something being done by a dominant firm

25 that is a clear monopoly than by some firm that is a


1 trivial firm, and so just because you are told that

2 something would be lawful under -- you find some Section

3 1 case out there where some foreclosure level was a

4 motion for summary judgment for a defendant, that does

5 not mean that in every case with the most extreme

6 monopolist you would grant summary judgment without

7 thinking long and hard about it.

8 MR. O'BRIEN: Okay, let's move on to the next

9 proposition.

10 Okay, I think this one will be easy, too. The

11 proposition is from Posner's Antitrust Law.

12 I propose the following standard for judging

13 practices claimed to be exclusionary: "In every case in

14 which such a practice is alleged, the plaintiff must

15 prove first that the defendant has monopoly power. All

16 the plausible cases of exclusionary practices involve

17 defendants that have monopoly power."

18 First, does everybody agree with that?

19 MR. LIPSKY: Uh-oh.

20 MR. O'BRIEN: Can exclusive dealing involving a

21 non-monopolist result in a substantial lessening of

22 competition?

23 DR. KLEIN: Yes.

24 DR. FARRELL: All statements containing the word

25 "all" are false except for this one and perhaps a


1 handful of others. I think there is a real problem with

2 a subtle, complex and imperfectly understood topic

3 having courts, judges, make grand and sweeping

4 pronouncements. The law, as I understand it, in a

5 precedent-based system tries hard not to change over

6 time, and our understanding tends to change over time,

7 and that creates a lot of trouble. So, it is not like I

8 am out here saying, oh, and the following large category

9 of cases, firms without monopoly power or without market

10 power or something, can do a lot of harm with exclusive

11 dealing. There have been some theories developed under

12 which that can happen.

13 I think the consensus currently is that that is

14 not such a big worry, but we do not really know yet, and

15 freezing stuff in place by grand pronouncements that say

16 "all," I am not sure it is such a great idea.

17 DR. CALKINS: The larger consequence, if that is

18 the law, is that any time a -- well, any time a

19 plaintiff has failed to hire one of these fancy

20 economists and satisfactorily define a market in which

21 the defendant has a well-defined market share of more

22 than 75 or 80 percent, there is a very good chance that

23 a Court would grant a motion for summary judgment or a

24 motion to dismiss, because when you have rules like

25 that, lots of courts operationalize it by saying, okay,


1 any market share below 70 percent, I grant a motion for

2 summary judgment and do not explore anything else about

3 what is going on, and that in my judgment is too

4 sweeping a broom to use. That was a bad way to phrase

5 that, wasn't it?

6 MR. O'BRIEN: Okay, Tad?

7 MR. LIPSKY: I think I can agree with the last

8 sentence there, that all the plausible cases -- I am a

9 little confused, though, whether this statement in

10 context, was it limited to exclusive dealing or is it

11 meant to be applied more broadly to other types of

12 exclusionary practices? I guess that there -- you know,

13 I am trying to recall. Wasn't there a -- there were

14 some Commission consent decrees in cases involving water

15 pumps for fire trucks. It was a multiple defendant

16 situation where there was actually a fairly plausible

17 theory of cartelizing, and I do not think you could have

18 found, at least not with any logical consistency, that

19 both of the competitors were monopolists.

20 So, I guess that is a limiting case, but I would

21 be closer to agreeing with this if you were talking

22 about cases other than those in which a cartelizing

23 theory for challenging the exclusive dealing was the

24 theory of liability.

25 Am I right about this FTC decree? Does anybody


1 remember that?

2 DR. CALKINS: There was a pump case. There

3 was -- there was a case like that.

4 MR. LIPSKY: Okay, so it is actually -- it is

5 probably real, presumptively real.

6 MR. VITA: It is called Waters Hale (ph).

7 MR. LIPSKY: Excellent, okay, thank you.

8 DR. KLEIN: If Posner had restated it in terms

9 of market power instead of monopoly power --

10 MR. LIPSKY: That would be fine.

11 DR. KLEIN: -- I assume we could all agree,

12 right?

13 MR. LIPSKY: Yes, that would be fine.

14 MR. O'BRIEN: So, this statement is about

15 monopoly power or market power on the part of the

16 defendant. I am wondering if any of you think that

17 conditions relating to market power or market structure

18 in the downstream market have an effect on the extent to

19 which exclusive dealing can be anticompetitive. That

20 was not stated well, but what should we make of the

21 downstream market structure in terms of the likelihood

22 that exclusive dealing can have an anticompetitive

23 effect?

24 DR. FARRELL: Well, I mean, I talked briefly

25 earlier about the developing economics of understanding


1 the role of downstream competition in that and, you

2 know, fairly plausible seeming analyses have come out

3 with very different answers so far, so watch this space,

4 and that perhaps should be a pretty strong warning

5 against making strong statements at this point.

6 MR. O'BRIEN: Would you be willing to say that

7 some kind of barrier to entry in the downstream market

8 is necessary for anticompetitive exclusive dealing?

9 DR. FARRELL: Well, I think -- see, you are

10 talking about a lot of abstract nouns here, and I am

11 sorry, I cannot put on a southern U.S. accent, but I

12 would like to.

13 DR. WERTHER: Can you do any U.S. accent?

14 MR. LIPSKY: I thought that was a Berkeley

15 accent.

16 DR. CALKINS: You have got such a lovely accent.

17 DR. FARRELL: I think Strunken White might have

18 said if you are getting confused, try to decrease the

19 abstract nouns and increase the active verbs, and I

20 think that is a pretty good proscription for thinking

21 straight. So, let's try that.

22 Instead of talking about market power and market

23 share and dominance and exclusive dealing and so on,

24 let's ask the following question: If I come up with a

25 better way of doing things than the incumbent is doing


1 or I am less greedy than the incumbent and I am willing

2 to give consumers a better deal, am I stymied in my

3 attempt to do so by these deals that people have struck?

4 That is the core question, and a lot of the time, the

5 answer will be no, I am not stymied if there are small

6 shares of this or that. Sometimes I will be.

7 So, for example, if you look at the Microsoft

8 case, Microsoft had no need to completely keep NetScape

9 out and wasn't trying to keep NetScape out and charge a

10 lot of money for Internet Explorer. They just had to

11 make sure that NetScape did not become sufficiently

12 widely distributed that people would start writing to it

13 and say, yeah, I -- that is a rather different case from

14 the one we would generically think of. You have to be

15 careful and I think should be pretty reluctant to kind

16 of lay down these firm rules.

17 Now, having said that, I also am very aware

18 that, you know, attempts to do full-blown rule of reason

19 analysis are also dangerous, right, given the subtlety

20 of what is going on and given the capabilities and

21 noncapabilities of courts.

22 I am a big fan in theory of -- I have never been

23 up close when it has happened -- of court-appointed

24 experts. I think that could probably improve the

25 process a lot quite generally, but especially when you


1 are dealing with subtle and difficult issues.

2 DR. CALKINS: Clearly everybody would say that

3 it matters how easy our new entrant can gain access to

4 the customers to whom it is trying to sell, and if it is

5 very easy to do that, then exclusive dealing will not

6 present any problems. As you phrased the question, you

7 used the magic word "entry barriers," and as you know,

8 that has lots of different definitions, and choose your

9 right definition and defendants will almost always

10 prevail; choose different definitions, and they might

11 not.

12 It also raises the question as to whether you

13 are looking at a total exclusion standard or at simply

14 making it much more expensive, time-consuming and risky

15 in order to gain access, and so you have staked out a

16 position or the quote here has staked out a position

17 which might mean things that I would not be comfortable

18 with.

19 MR. O'BRIEN: Right. So, just one follow-up to

20 that, I guess this is directed to Joe, the Fumagalli and

21 Moto models and the Simpson and Wickelgren model, in the

22 simplest cases, you have homogeneous producers

23 downstream with no economies of scale or very small

24 economies of scale, and it strikes me in the context of

25 those models that it would be very easy for a firm to


1 enter at both levels and disentangle any anticompetitive

2 effect that is being contemplated. I am wondering if

3 you have thought about that, or maybe I am wrong.

4 DR. FARRELL: You know, it has been a while

5 since I read the models, so I do not remember

6 technically whether what you say is right. Clearly if

7 you really have homogeneous products and fixed costs and

8 sunk costs are very small, then you would think -- and

9 you would want to know why not if somebody was claiming

10 not -- that a firm could enter at both levels.

11 On the other hand, there certainly are

12 industries where at any given time the industry may

13 behave quite competitively involving the pass-through

14 dynamics that we were talking about, and yet there are

15 big sunk costs lying behind it, and that may be the more

16 relevant case for that kind of analysis.

17 MR. O'BRIEN: Anyone else? Okay, next slide.

18 Okay, I think this is an uncontroversial slide

19 as well. We will see. Maybe the questions will be more

20 interesting.

21 "Exclusive-dealing arrangements --" this is a

22 quote from Jefferson Parish. "Exclusive-dealing

23 arrangements 'may be substantially procompetitive by

24 ensuring stable markets and encouraging long-term

25 mutually advantageous business relationships.'"


1 Yes, Joe?

2 DR. FARRELL: I hate to be a curmudgeon, but

3 stable markets are not exactly what antitrust aims for.

4 Actually, maybe we should try to encourage unstable

5 markets where the status quo could be disrupted at any

6 moment by some pesky firm that maybe has not shown up

7 before, or maybe has, and is willing to take a lower

8 margin or has a better way of doing things.

9 Now, I am not saying that the basic point here,

10 that exclusive dealing arrangements "may be good" is

11 wrong, but I do not like that language.

12 MR. O'BRIEN: Okay. Well, you pick the --

13 DR. CALKINS: And while you are complaining, you

14 could complain about the mutually advantageous business

15 relationship, because that could be good for consumers,

16 and if it is just dividing up a surplus between two

17 businesses, it could be bad for consumers.

18 DR. KLEIN: Yeah, I --

19 MR. O'BRIEN: Ben Klein, do you have a view on

20 that?

21 DR. KLEIN: Well, who knows what Justice

22 O'Connor is referring to, but if she means by

23 encouraging long-term mutually advantageous business

24 that it encourages people to make specific investments

25 in the relationship, relationship-specific investments,


1 then I think she is correct and that she should not go

2 through it now, but that is one of the problems I had

3 with Joe's presentation, is that the Segal and Whinston

4 criticism of that rationale for exclusive dealing is

5 just wrong, and it is logically correct, but there is

6 assumptions being made in that that are very, very

7 unrealistic, and in particular, they are just -- well, I

8 better not go into it.

9 But, you know, so if that is what she is saying,

10 I would agree with her very much, but it is so vague,

11 right, but if she is just saying there is -- it

12 sometimes may be good...

13 MR. O'BRIEN: What efficiencies, which I assume

14 are the second object of this sentence, there are

15 numerous efficiencies that have been discussed about

16 exclusive dealing that we might classify into that

17 second phrase. What are the most significant and most

18 likely in an exclusive dealing arrangement?

19 And similarly, what efficiencies have been

20 asserted most often do you think are least likely to

21 actually exist?

22 DR. CALKINS: Oh, the best is the classic Marvel

23 free riding, manufacturers spending money, bringing in

24 the customer, then there's the old bait and switch to

25 the other product. That would be the classic and the


1 best.

2 MR. O'BRIEN: So, what efficiencies are often

3 asserted in exclusive dealing cases that you think may

4 not actually exist very often? Anybody?

5 DR. KLEIN: I hope nobody says this focused

6 dealer effort, but I guess one of the things I should

7 say is the justification that Microsoft offered, the

8 procompetitive justification for the exclusive dealing

9 arrangement with the Internet access providers, sounded

10 like a focus -- the way you presented it, it sounded

11 like a focused dealer incentive, but what they wanted

12 was -- the argument they presented was something to the

13 effect that they wanted the developers to focus on the

14 Windows APIs, which meant they wanted to have a monopoly

15 in Windows so that when developers were developing their

16 programs, they would only develop Windows programs,

17 which is a very different argument than, you know, you

18 want -- they did not want the Internet access providers

19 to promote their product. That is not what they were

20 doing.

21 They were talking about a different type of

22 focus there, but that argument I think the Court

23 correctly rejected as making no sense other than you

24 want a monopoly. You want to maintain your monopoly.

25 MR. O'BRIEN: What significance, if any, should


1 be given to observing an exclusive dealing arrangement

2 in a similar competitive market when you are analyzing a

3 case where there is exclusive dealing, maybe in a market

4 that exhibits some more market power in some ways than

5 the other, but otherwise has similarities?

6 DR. FARRELL: Well, at a technical level, there

7 certainly have been analyses that show that in some

8 circumstances, exclusive dealing engaged in by, let's

9 say, all members of an oligopolistic manufacturing

10 sector, whether downstream industry, can soften

11 competition and be in that sense anticompetitive, even

12 conditional on, you know, a flourishing oligopoly

13 structure, and let's face facts, we are never dealing

14 with perfectly competitive industries when we are

15 talking about these cases, so oligopoly is what you mean

16 by the word "competitive" here.

17 There are other analyses that suggest that

18 exclusive dealing can actually sharpen competition. I

19 think it is fair to say that that literature is both

20 unsettled and in a state of nonferment, the nonferment

21 because nobody seems very excited about it. People are

22 really more interested in the monopoly-preserving

23 possibilities I think than the oligopoly-softening

24 possibilities, and that may be a legitimate choice of

25 emphasis, where to put our intellectual resources, or it


1 may just be, you know, what happens to be fun for

2 assistant professors to do these days.

3 DR. KLEIN: I think we have to be very careful

4 when we start talking about oligopoly-softening, and I

5 guess Joe would say I have this bias, this laissez-faire

6 bias, but I can imagine unilateral behavior -- you know,

7 a gasoline company decides they are going to locate

8 their station not next to another station but a couple

9 of blocks away, because if they locate it next to the

10 station, it is going to be more intensive competition.

11 People are going to be able to compare the prices.

12 We do not want to go in and micro-regulate the

13 competitive process. You know, you hire an economist,

14 and let's assume they draw the welfare triangles, and

15 they say consumers are better off if that person puts

16 the station next to the other station, and even though

17 it has -- let's assume it has the effect of sharpening

18 competition if we do that, we do not want to regulate

19 that behavior, at least I do not want to, even though

20 the calculation would come out that way.

21 So, I think it is dangerous to start talking

22 about oligopoly-softening of competition in general, and

23 basically I guess I have a prior that we are just going

24 to mess things up and we should just leave it up to the

25 competitive process, unless there is a -- you know, you


1 have this first step where you need some major

2 anticompetitive effect in terms of foreclosure.

3 So, I guess my comment was not totally

4 irrelevant, because we are talking about Section 2

5 unilateral behavior, even though it has nothing to do

6 with exclusive dealing.

7 DR. CALKINS: Trying to psycho-analyze your

8 question, I think you were -- I am guessing that you

9 were referring to the argument you sometimes see made

10 that, look, over here in this market, which we all

11 stipulate is competitive, this practice is occurring,

12 and so, therefore, it must follow as the night follows

13 the day that when that same practice is being engaged in

14 by this complete and total monopolist, it deserves

15 summary judgment very promptly on that ground alone,

16 and --

17 MR. O'BRIEN: That is a good psycho-analysis.

18 Yes, that is what I was hoping somebody would address.

19 DR. CALKINS: And I myself do not buy into that

20 theory in the little that I have done thinking about it,

21 but my thinking is still at a preliminary stage.

22 MR. VITA: Well, it is not so much -- maybe,

23 Dan, a competitive market versus a noncompetitive

24 market, but the individual -- the size of the firm or

25 the mark -- the firm's specific market power. Like the


1 RC Cola example somebody alluded to before, RC has some

2 exclusive relationship with its bottlers or something, I

3 think it was, and you look at RC Cola, and they are a

4 small fry. I mean, they do not matter anywhere. So,

5 you look at that and you say, well, obviously they are

6 doing that. They cannot possibly have any kind of

7 foreclosure mode or some monopolization motive. It has

8 to be some sort of value creation that induces them to

9 do that.

10 Is it fair to say that when you do -- then you

11 look at Coke, for example, maybe doing the same kind of

12 thing, some other firm with substantial market share or

13 market power possibly? At least it says you have got to

14 consider the efficiency story. You can't rule it out.

15 There is a possibility that there is value creation,

16 that there is something inefficient about it, but not

17 necessarily -- the fact that RC does it doesn't

18 vindicate Coke's usage, that debate is not over, but

19 that does say to you -- you know, we have got to take

20 that seriously.

21 DR. FARRELL: Yeah, I think you said it right,

22 you know, unless there is something about that industry

23 or market that I do not know, you can presumably infer

24 from RC's use of these exclusives that there is

25 something other than monopoly preservation going on, but


1 that does not mean that there is not monopoly

2 preservation going on.

3 DR. KLEIN: Exactly.

4 DR. FARRELL: It does not mean there is either.

5 MR. LIPSKY: Thanks.

6 MR. O'BRIEN: Okay, next slide.

7 Okay, so anticompetitive effects, this is a --

8 this is actually a quote from Dennis Carlton's paper on

9 the Aspen and Kodak case.

10 "In the presence of scale economies, exclusive

11 dealing can be a way of depriving Firm 2 (or its

12 distributors) of the necessary scale to achieve

13 efficiencies, even though, absent the exclusivity, Firm

14 1 and Firm 2 would both be large enough to achieve

15 efficiency."

16 So, this is the standard scale economy argument

17 about excluding your rivals so that it cannot reach

18 efficient scale, and I guess my question is, does the

19 panel see that as the primary anticompetitive theory of

20 exclusive dealing that we ought to be focused on?

21 MR. LIPSKY: Well, I will take a stab at that.

22 Certainly, you know, in a static sense, it is hard to

23 argue with this proposition, and I think this is

24 consistent with the notion that there are stories

25 associated with exclusive dealing where you are trying


1 to compel two-stage entry basically, and I think some of

2 those are good stories. Probably we would not agree on

3 which ones were good stories.

4 I heard John Jacobson the other day talking

5 about Pullman, and I disagreed with him on that one, and

6 then Motion Picture, and I disagreed with him on that

7 one, but it does not -- and also United Shoe Machinery,

8 and I disagree with him on that one, but I think we

9 could find -- I think we could find a two-stage entry

10 story that held together, and so I would say I agree

11 with this.

12 But I would also interject -- and I have said in

13 other contexts -- there is kind of an endemic temptation

14 or tendency in the system, in the investigation and the

15 litigation system, to underestimate supply flexibility.

16 I mean, you know, supply flexibility is not -- or new

17 entry is not always an answer, and so I would hate for

18 my remarks to be misconstrued. There are industries in

19 which the barriers to entry are such that if you have a

20 two-stage story, it is a serious problem, but I think

21 there is a tendency to look at what is right in front of

22 you, to, you know, fail to predict the rise of the

23 Internet or the mobile phone, you know, falling in price

24 by 75 percent over five years or, you know, some other

25 alarming and unpredicted new technology or new


1 development, and because the dynamic aspect is so

2 important, I think this is a theme that needs to be

3 hammered again and again.

4 So, what I guess I am saying, yes, I agree with

5 this, but it is narrow -- I would like to make my

6 agreement as narrow as humanly possible.

7 MR. O'BRIEN: Anybody else?

8 DR. KLEIN: Tad, you sounded like an expert

9 witness there.

10 DR. CALKINS: I was hoping that Tad could tell

11 me how to get a mobile phone bill that is 75 percent

12 lower.

13 MR. O'BRIEN: So, Joe, based on your remarks, I

14 guess I would ask, do you think this is the primary

15 story of competitive harm that we should be focused on

16 in analyzing exclusive dealing, or should some of the

17 other theories that you mentioned, I guess in particular

18 Simpson/Wickelgren, maybe some of these two-stage models

19 of oligopoly where exclusive dealing can play a role,

20 are those things we should be concerned about, or is

21 this number one and number two?

22 DR. FARRELL: Well, I disagree with the

23 question. I think the primary focus should be based on

24 what is going on in the market at hand, and we should

25 adjust the tools to fit the facts and not prejudge what


1 the theory is going to be.

2 Having said that, I think I said in my earlier

3 remarks that I believe this Rasmussen, Ramseyer and

4 Wiley or Segal/Whinston theory, which is being referred

5 to here, is the one that people talk about most. I tend

6 to suspect that it is the main one. I would add -- I

7 mean, you have to interpret efficiencies carefully, so,

8 for example, scale to fully reward innovation, is that

9 achieving efficiencies?

10 But broadly speaking, I think this is what most

11 economists think of most of the time when they think

12 about anticompetitive exclusive dealing, and I think

13 that may well be right, but I think we should be open to

14 whatever the facts of a particular case say.

15 MR. O'BRIEN: All right. Anybody else?

16 (No response.)

17 MR. O'BRIEN: Let's go to 7.

18 Okay, this is from the Microsoft case, and the

19 quotation is:

20 "If the monopolist's procompetitive

21 justification stands unrebutted, then the plaintiff must

22 demonstrate that the anticompetitive harm of the conduct

23 outweighs the procompetitive benefit."

24 I guess my question is -- well, first, does that

25 make sense to you, and secondly -- this is maybe more


1 for the economists, although equally for the lawyers --

2 does economics supply tools to do this?

3 DR. KLEIN: Try Joe.

4 DR. FARRELL: Well, let's see. I mean, clearly

5 in order to plunge into enforcement, we would not want

6 to go ahead if the anticompetitive harm of the conduct

7 is outweighed by the procompetitive benefit. Using the

8 term "procompetitive benefit" in -- I am not sure

9 whether it is the same way or not as Ben uses it, but I

10 am using it to mean actual benefits to efficiency and

11 consumers, not just kind of non-anticompetitive

12 rationales.

13 This, of course, is part of a bigger decision

14 tree that the Microsoft Court laid out. In thinking

15 through a burden-shifting process like that, you have to

16 think about a number of things, and I do not know how

17 much the Court thought through these things. I am

18 pretty sure I know how much they knew the necessary data

19 required to do it exactly right, which is not a

20 criticism, because nobody has that data either.

21 You have to think both about whether in most

22 cases this is true or that is true, but also about if

23 this is true, is it going to be easy to prove, or is it

24 quite likely to be true but be hard to prove? And that

25 really gets back to what I hope was the main theme that


1 came out of my talk earlier, that in my opinion, there

2 are often benefits of open, free-wheeling competition

3 that are very difficult to pin down and almost

4 impossible to prove, and I think that needs to be kept

5 in mind when we lay down these decision trees.

6 Did the Microsoft Court keep that in mind? To

7 some extent. Did it do it the right amount? I have no

8 idea, and I doubt that they really know either.

9 DR. CALKINS: If the question is should one

10 think about the competitive harm that is likely, should

11 one think about the procompetitive benefit, the answer

12 to that is entirely yes.

13 On the other hand, can you read this statement

14 to say that if there is any tiny procompetitive benefit,

15 perhaps using anybody's definition of "procompetitive,"

16 does that mean that the defendant always wins unless the

17 plaintiff is able, with great specificity, to precisely

18 quantify the anticompetitive harm, precisely quantify

19 the anticompetitive benefit, and then precisely

20 calculate that one is more than the other?

21 Well, it may well be that if that is what one

22 means, then what one is saying is that any time there is

23 any benefit that can be characterized as procompetitive,

24 the defendant will always win, and so if that is where

25 you ended up, that might not be a good place, but that


1 does not mean that you should not think about the

2 procompetitive benefit.

3 DR. KLEIN: Go ahead, Tad.

4 MR. LIPSKY: No, go ahead.

5 DR. KLEIN: No --

6 MR. O'BRIEN: Go ahead, Tad.

7 MR. LIPSKY: Well, I was just going to say that

8 we always have to consider the fact, you know, there was

9 a day not so long ago when you could expect a follow-on

10 litigation from cartel cases that were litigated and won

11 by the Department of Justice. You would get a guilty

12 plea in a price-fixing case, and then we transitioned --

13 I am not sure exactly what the history is or how we got

14 here, but then we got to the point where there was a

15 story in a newspaper saying that there was a

16 price-fixing investigation, boom, 80 private class

17 action -- purported class action treble damage suits

18 against everybody in the industry, and then we got to

19 the state where there -- you get the same thing even in

20 these conduct type cases, which are not cartel cases,

21 and there are follow-on class actions for Dentsply, and

22 there were follow-on class actions for this, that and

23 the other outside of the price-fixing area, and that

24 combined with, you know, indirect purchaser statutes and

25 all kinds of things that happen in antitrust litigation


1 generally I think creates the fear that there are some

2 legitimate procompetitive practices that the perpetrator

3 cannot afford a defense, and I think that is a very

4 troublesome phenomenon.

5 I guess the thought is provoked by Joe's comment

6 that there are -- you know, there is sort of a -- maybe

7 we should indulge a presumption that when things are

8 loosened up a little, and there are fewer strong ties,

9 you know, partial vertical relationships, maybe that is

10 the way we want markets to function, but I think the

11 system in general works pretty well if we require -- you

12 know, we always have the ultimate burden of proof on the

13 plaintiff, so that if the defendant can come up with a

14 sensible justification, a justification that can be

15 persuasive with the fact-finder, then yes, the right

16 standard is, if the defendant has something good to say

17 for his practice, let's adopt a rule that the plaintiff

18 does not win unless the plaintiff persuades that the

19 negative effect on competition outweighs the

20 procompetitive effect.

21 And true enough, part of what I was saying

22 earlier is, yes, it is the wiles of economic theory. It

23 is the unadministerable, you know, battle between the

24 economic experts and all the other facts in the case,

25 but what is the alternative? The alternative is


1 Standard Stations, or worse, and we know that is wrong,

2 so that is why I would like -- I keep trying to bring

3 into the conversation this institutional element.

4 Let's not -- once we decide it is a balance,

5 let's not just throw confetti in the air. Let's try to

6 focus on what the applied micro tells us about what

7 rationales deserve to be explored and what facts could

8 rule various theories of efficiency or theories of

9 restraint in or out. Let's organize that process so we

10 do not just have a U.S. versus IBM every time there is,

11 you know, a 13-year slog or a 14-year slog like Harmar

12 versus Coca-Cola every time we have a difficult

13 exclusive dealing issue.

14 DR. CALKINS: I really misunderstood you, Tad.

15 I thought when you said you wanted to go to the 18-month

16 model, you wanted to go back to the days of Standard

17 Stations, and I just --

18 DR. KLEIN: Per se. But to answer your question

19 about whether we have the tools to do this, I guess

20 economists have the tools -- I was on a panel with Steve

21 Salop where I said I -- even if I were the judge, I

22 wouldn't know exactly how to do it, and he said, you

23 know, that is all economists know how to do, you know,

24 want to take away your doctorate or something, but when

25 you -- obviously you have to go to balancing.


1 I mean, I am pretty cynical about this, because

2 I do not know -- I do not think the courts have done

3 this, and I do not know what to tell them to do. I

4 mean, I think they go backwards, and they figure out --

5 you know, they do some kind of implicit balancing, and

6 then they say -- they make it easy and they say it was

7 not an anticompetitive effect or there is no

8 procompetitive efficiency rationale, and I do not know

9 what exactly we should have them do, other than we know

10 we want them to hire more economists, right?

11 But it is a -- I think that is the ultimate

12 question, because you do have to do the balancing, and I

13 do -- I mean, it is a legal question, but I do think the

14 burden should be placed on the plaintiff at that point,

15 because I have this prior bias about the competitive

16 process. So, I agree with the legal rule, but then what

17 exactly are you doing -- and it should -- it should not

18 be a close thing, because that is my -- and I think that

19 is the way the law is or it should be, that it should

20 not be a very close thing that we are balancing, and it

21 should not be something -- you know, there should be

22 this first step that you have to show a very clear

23 anticompetitive effect before you go forward in any way,

24 and that is going to get rid of most of the cases.

25 Steve will say that is why the defendants win


1 all the time, but they do not always win, because you

2 have the Dentsplies and you have the Microsoft, and I

3 think that is enough to get efficiency in the economy.

4 DR. FARRELL: There is this article by Priest

5 and Klein -- I do not know if that is you --

6 DR. KLEIN: Yes, that is me.

7 DR. FARRELL: -- saying that whatever the rules

8 are, the litigated cases are going to be close ones.

9 So, I do not think we can have a rule that litigated

10 cases are not allowed to be close.

11 MR. O'BRIEN: Okay, well, we have run past our

12 time, and I think it is Ben's fault, by about four

13 minutes. So, thank you very much everybody.

14 (Applause.)

15 (Whereupon, at 4:04 p.m., the hearing was

16 concluded.)



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



4 DATE: NOVEMBER 15, 2006

6 I HEREBY CERTIFY that the transcript contained

7 herein is a full and accurate transcript of the notes

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

9 FEDERAL TRADE COMMISSION to the best of my knowledge and

10 belief.

  DATED: 12/4/2006
18 C E R T I F I C A T I O N O F P R O O F R E A D E R

20 I HEREBY CERTIFY that I proofread the transcript

21 for accuracy in spelling, hyphenation, punctuation and

22 format.


Updated June 25, 2015