# FEDERAL TRADE COMMISSION

> Briefs, arguments, decisions, and more.

URL: https://www.frixlaw.com/law-library/documents/agency%3Aftc%3A4560ab493ca1cc50

## Record

- **Collection:** Agency decision
- **Document type:** Agency decision

## Text

1
1
2
3
4

FEDERAL TRADE COMMISSION

5

DEPARTMENT OF JUSTICE

6

JOINT WORKSHOP ON MERGER ENFORCEMENT

7
8
9
10
11

Tuesday, February 17, 2004

12

9:00 a.m.

13
14
15
16

Federal Trade Commission

17

Conference Center

18

601 New Jersey Avenue, N.W.

19

Washington, D.C.

20
21
22
23
24
25

Reported by:

Rita M. Hemphill, CVR

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

2
C O N T E N T S

1
2
3

PAGE
Opening Remarks

3

Hypothetical Monopolist Test

17

Concentration & Market Shares

114

Monopsony

207

4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

3
1

P R O C E E D I N G S

2

-

3

-

-

-

-

OPENING REMARKS

4

MR. ABBOTT:

Good morning, everyone.

Welcome

5

to the Joint Federal Trade Commission Department of

6

Justice Antitrust Division Workshop on Merger

7

Enforcement.

8

the new FTC conference center -- and we hope you all

9

enjoy the facilities -- for today, tomorrow, and

10

Thursday.

11

This workshop will be held here today in

We welcome members of the audience to listen

12

in, although all questions during the proceedings will be

13

handled by members of the panels.

14

We will start out today -- and I am not going

15

to go into thanks, but I owe a debt of thanks to staff at

16

both the Federal Trade Commission and the Antitrust

17

Division, and certainly Bob Potter, head of the policy

18

office of the Antitrust Division, for tremendous

19

collaboration and support in putting on this event, which

20

we hope will clarify a number of issues regarding the

21

current application of the merger guidelines.

22

And we are delighted to start things out with a

23

bang with the leaders of the federal antitrust community.

24

We will lead off with brief remarks by Assistant Attorney

25

General Hew Pate, and follow up with remarks by FTC

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

4
1

Chairman Tim Muris.

2

And both General Pate and Chairman Muris have

3

quite a few significant accomplishments, but I won't take

4

away from their important time at the podium by going

5

into them.

6

you, and also to the audience that is listening in on the

7

800 telephone line to this important conference.

8

Suffice it to say that they are welcoming

And we hope to obtain some valuable information

9

which will inform our activities in the future, as a

10

result of the conference.

11

now to General Pate.

12

MR. PATE:

So, let me turn it over right

Thanks very much, Alden.

Thanks to

13

the Federal Trade Commission for providing this great

14

facility to host our look into the operation of the

15

merger guidelines, and to consider improvements -- we

16

hope -- in thinking, or developments in thinking, in the

17

way that the agencies carry out their responsibilities

18

under Section 7 of the Clayton Act and under the FTC

19

statute.

20

I think that it's fair to say that this is one

21

part of a very significant amount of policy activity

22

that's going on at the two agencies.

23

think Tim Muris deserves the lion's share of the credit

24

for that, in terms of bringing forward the concept of

25

joint work between the two agencies.

And frankly, I

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

5
1

That has been producing a great deal of

2

thinking, in terms of the IP and antitrust hearings,

3

which are now drawing toward a final close, hopefully

4

with the publication soon of the antitrust and IP report.

5

That's happening in the health care antitrust arena, in

6

the context of the hearings that were held there, and

7

this merger conference is yet another example.

8

Both of the agencies, likewise, have been

9

active individually on the policy front.

I think of the

10

report on patent practices and procedures, that the Trade

11

Commission has issued, which is a very important

12

document, and likewise, in just a few weeks we at the

13

Division will be heading up a conference looking at

14

pricing practices.

15

And I think all of this is really an important

16

part of our mission, and I want to thank and acknowledge

17

my colleagues at the Federal Trade Commission, starting

18

with Tim, for putting in place the circumstances that

19

have made for a very good relationship there.

20

Obviously, at the Division, merger enforcement

21

is a key part of our mission.

Depending on the activity

22

level, it might be fair to say that roughly one-third to

23

two-thirds of our effort goes into merger enforcement.

24

As I have mentioned elsewhere, merger enforcement is

25

different from some of the other things that we do in the

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

6
1

antitrust world, certainly different from cartel

2

enforcement, where there is little or no serious

3

controversy about the need to aggressively intervene

4

where cartel activity is detected.

5

It is different, too, from the unilateral

6

conduct sphere, where it's very important for the

7

agencies to have a very high degree of caution about

8

where they intervene, because of the possibility of

9

chilling competitive conduct in the guise of combating

10

anti-competitive conduct.

11

Obviously, it's a gross over-simplification,

12

but I suggest that merger enforcement falls somewhere in

13

the middle, that there is a broad consensus about what we

14

are trying to achieve, but plenty of room for debate and

15

discussion about what the best use of the tool set that

16

we have to evaluate mergers.

17

Obviously, merger enforcement requires us to

18

make predictive judgments.

Section 7 is described often

19

as an incipiency statute.

That's obviously true, but

20

that term can carry different, very freighted meanings,

21

depending on who the speaker might be.

22

And what we're about here is to try to

23

determine how best we can make predictive judgments about

24

proposed transactions to make sure that those that would

25

bring efficiency and better outcomes for consumers are

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

7
1

allowed to go forward.

And likewise, that those that

2

have the real prospect for substantial lessening of

3

competition don't happen.

4

Obviously, the Horizontal Merger Guidelines are

5

the Agency's most important statement of how we go about

6

doing that.

7

been revised now several times -- in 1984, 1992, 1997 --

8

revised with some significant language changes, and some

9

significant changes in approach.

10

The merger guidelines of 1982, which have

Certainly the agencies, even apart from those

11

textual changes, have, throughout this period,

12

incorporated advances in economic thinking and

13

improvements in our own practices and learning.

14

The guidelines are flexible, they are

15

practical.

16

that both agencies are in a period of trying to emphasize

17

transparency.

18

that those in the bar who are very experienced repeat

19

customers of the agencies have a good idea of how we do

20

things, but rather that we have an obligation to be as

21

transparent as possible to the public generally.

22

At the same time, I think it's fair to say

It may not be quite enough simply to say

You can see that in terms of the increased

23

incidents of the issuance of closing statements at both

24

agencies, to try to give some amount of guidance to the

25

public, when we decide that government intervention is

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

8
1
2

not warranted.
And I think you can see that in terms of the

3

release of data that has preceded this conference.

4

agencies jointly, as you know, have released data on

5

merger challenges from 1999 to 2003.

The

6

The Federal Trade Commission has now released

7

significant data on cases in which a challenge was not

8

brought, and also issued some data that shed light on

9

other factors, customer complaints, hot documents, other

10

things that the agencies looked at in evaluating a

11

proposed merger.

12

set the stage for some very interesting discussion and

13

improved learning in this conference.

14

And we hope very much that this will

Transparency, as I have said, is important, not

15

only for the repeat players in the merger bar who are

16

likely to be most interested in this, but for other

17

enforcers, both here and internationally.

18

we need to be very mindful, not only of enforcing in a

19

sound way, but explaining what sound enforcement entails.

20

I think that

Whether we like it or not, the two federal

21

agencies are not the only game in town in terms of how

22

merger policy is going to go forward, and it's incumbent

23

on us to show our reasoning and be part of the broader

24

discussion of what appropriate enforcement criteria are.

25

And I hope this conference will contribute to that.

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

9
1

We certainly have a wide range of topics and a

2

wide range of speakers, most of whom are very familiar to

3

anyone who is involved in merger policy.

4

antitrust ego-meter to put up here on the podium during

5

the next three days, it would break.

6

If we had an

Fortunately, we have got some very good

7

moderators, mainly from the two agencies, who have put a

8

great deal of time into preparing to keep these panels

9

under control and hope to get some insight from them.

10

One thing you might ask is, "What is our goal

11

here?"

12

conference has been called because anyone thinks

13

something is broken.

14

conference has been put together because there is some

15

pre-ordained goal of producing any particular output from

16

the conference.

17

really in a position to answer until we get the benefit

18

of hearing what reactions and thinking, based on the data

19

that's released, based on work that many of the panels

20

have done independently over the past several years.

21

Well, it's certainly not the case that this

Nor is it the case that the

I think that's a question that we're not

Until we hear that, it's hard to say whether

22

and what might come next.

But just as with the process

23

side, where both agencies have done a good deal to try to

24

improve our procedures, this conference does demonstrate

25

our open-mindedness to try to make sure that we are

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

10
1
2

keeping up with the best thinking on substance.
So, with that, I will conclude simply with a

3

very big thank you to those at both agencies who have put

4

a great deal of time into organizing these panels.

5

look forward to participating in some more of this,

6

including a round table Thursday afternoon, and will keep

7

things moving now by introducing my good friend, Tim

8

Muris, for his opening remarks to help us kick off the

9

conference.

I

Thanks very much for being here.

10

(Applause.)

11

MR. MURIS:

12

certainly a pleasure to be here.

13

interesting three days.

14

Justice are, obviously, our most important partners and

15

colleagues, and I greatly appreciate their help and work

16

on their effort.

17

Thank you very much, Hew.

It is

This should be an

Hew and the Department of

And I wanted to thank our staff at the FTC and

18

the Justice Department staff.

19

which I will discuss briefly here in a second, is quite

20

interesting, and I am sure many of those who will be

21

talking over the next few days will remark on that data.

22

The data we released,

I particularly wanted to thank our economist

23

who worked on that data.

It was a lot of work.

24

few of them out there in the audience.

25

of people to read a lot of our memoranda from the past.

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

I see a

It required a lot

11
1

It is certainly a pleasure to me, and an honor,

2

that -- I noticed Hew's interesting comment about the

3

ego-meter -- that we will have so many prominent

4

practitioners, academic and enforcement officials here

5

over the next few days.

6

I won't repeat the mistake I made about a

7

decade ago, when I was at a gathering of -- a conference

8

with many of the leading economists, and I was on a

9

panel, and I said, "This is the greatest gathering of

10

$500-an-hour economists every assembled."

11

long time ago when billing rates were lower.

12

them took great offense, and it was immediately obvious

13

to me and everyone else that this person billed

14

considerably higher than $500 an hour.

15

(Laughter.)

16

MR. MURIS:

17

notes.

This was a
And one of

And the others were all making

And I felt like a facilitating practice, so --

18

(Laughter.)

19

MR. MURIS:

As Hew mentioned -- you know, well,

20

whatever one's billing rate -- we are going to discuss

21

the impact of the guides.

22

have an impact not only at the federal level, but also at

23

the states and internationally.

And as Hew mentioned, they

24

There are over 60 countries that have merger

25

control regimes, and you can see the influence on the

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

12
1

guidelines almost everywhere.

2

Let me tell you what my view of what we can

3

achieve over the next three days.

4

think we all know, are not a cookbook.

5

provide specific details on every aspect of a merger

6

investigation.

7

analyze -- to analyze whether a merger is likely anti-

8

competitive.

9

explore state of the art application of the guidelines by

10

those with the most experience at using them.

11

The guidelines, as I
They don't

They do detail a methodology, whether to

And what the workshop is going to do is

We will publish an edited transcript on our

12

website so that people can refer to it.

13

a lot of people at the FTC -- I'm sure at Department of

14

Justice -- listening on their computers.

15

people listening on their 800 number.

16

going to film this so I can watch it on my exercise bike,

17

and other people can refer to it as well.

18

I know there are

There are other

We are actually

As Hew mentioned, we at the FTC have also

19

emphasized transparency over the 32 months that I have

20

been at the Commission.

21

several cases in which we did not sue, including three

22

cases very recently.

23

We have released statements in

We have released the two data sets that Hew

24

mentioned.

The first one contained market share and

25

concentration levels associated with the FTC and DOJ's

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

13
1

decisions to challenge mergers in a wide range of product

2

markets.

3

on 151 horizontal merger investigations from the

4

beginning of fiscal year 1996 through the end of fiscal

5

year 2003.

6

And then a few weeks ago, the FTC released data

Unlike the December data, this release included

7

data when no enforcement action was taken, as well as

8

data on additional key facts.

9

involved the numbers of significant competitors.

10

practitioners I know think and talk in these terms,

11

rather than concentration.

12

One statistic we released
Many

These data also look at enforcement decisions,

13

depending on whether hot documents exist or strong and

14

credible customer complaints are received.

15

should become core information in a healthy debate about

16

the level and direction of merger policy.

17

These data

I note that our recent data released largely

18

reflects cases in which I did not participate.

19

participated in cases involving only about 10 percent of

20

the markets analyzed.

21

Indeed, I

As I have said elsewhere, current merger

22

practice reflects a bipartisan approach.

23

of this and similar data in the future should help us

24

understand merger enforcement in practice.

25

And the release

The data we released highlights several

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

14
1

important issues in merger analysis.

2

long-standing debate about the significance of

3

concentration and HHI numbers.

4

released and the breadth of analysis we will hear this

5

week will finally put to rest the notion that HHI levels

6

have any special significance, except at very high

7

levels.

8

One involves the

I hope the data we

Instead, the agencies try to answer the

9

ultimate question.

10

We consider several variables that have an impact on the

11

likely level of competition in post-merger markets.

12

Fealty to the original guideline numerical levels was

13

abandoned as the agencies gained experience.

14

the guidelines were amended to codify the existing

15

practice of giving great weight to qualitative factors.

16

Will the merger impair competition?

In 1992,

The 1982 and 1984 guidelines had given more

17

emphasis to quantitative thresholds, particularly

18

involving HHI levels above 1800.

19

and Bill Baxter discussing what the appropriate levels

20

ought to be, and it was clear to me from those

21

conversations that Bill retained some of his views that

22

he had expressed in the 1960s about the strong

23

presumptions one could gather from concentration.

I remember Jim Miller

24

In any event, the 1992 guidelines reduce the

25

significance of the 1800 threshold by inviting fuller

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

15
1

consideration of other conditions that help predict

2

whether price increases are likely, post-merger.

3

Thus, the pre-eminence that some would continue

4

to give to concentration or HHI numbers is misplaced.

5

State of the art merger analysis has moved well beyond a

6

simplistic causality of high concentration leading to

7

anti-competitive effects.

8

The number of competitors is certainly

9

important.

Four to three gets our attention quicker than

10

six to five.

11

elevate a single fact or number to dispositive

12

significance.

13

an increased likelihood of anti-competitive effects

14

before we act.

But current merger practice does not

The totality of the evidence must point to

15

Much of this experience with merger

16

investigations is captured in the guidelines themselves.

17

One of the salutary effects that the guidelines have is

18

the transparency they help bring to government.

19

help us clarify enforcement policy and doctrine so that

20

practitioners and their clients can make better

21

judgments.

22

They

Government officials should explain the basis

23

on which they exercise their authority.

24

expect clear and consistent enforcement actions.

25

course, application of the guidelines is not always

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

Stakeholders can
Of

16
1

obvious.

2

Hence, the high billing rates.
We constantly strive to bring more transparency

3

to our merger process, and we hope this workshop will

4

result in a better understanding of current merger

5

policy.

6

transparency for consumers in the business community is

7

the feedback this workshop should provide for the

8

agencies.

9

Equally important with providing increased

We want to obtain important information that

10

will assist us in doing our jobs.

11

learn from you over the next three days.

12

from the most experienced practitioners who work with the

13

guidelines every day, as well as academics doing state of

14

the art research.

15

guidelines are working, what you perceive to be their

16

strengths and weaknesses, what are the issues in which

17

there is agreement and disagreement, and what areas you

18

consider important for further study.

19

Research and clarification.

20

guidelines must respond to new legal and economic

21

analysis.

22

at this workshop.

23

started.

24

We thus expect to
We will hear

We want to know how you think the

Application of the

With your help, we will continue that process
Thank you very much, and we can get

(Applause.)

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

17
1

HYPOTHETICAL MONOPOLIST TEST

2

MR. WERDEN:

Good morning.

I'm Greg Werden,

3

and I am the moderator of this, the first session of the

4

merger enforcement workshop, which focuses on the

5

hypothetical monopolist paradigm for market delineation.

6

The hypothetical monopolist paradigm was

7

elaborated, and certainly popularized, by the 1982 merger

8

guidelines, but the basic idea was not original to the

9

1982 guidelines, as I think a lot of you realize.

10

late 1970s, some of the Antitrust Division were already

11

applying the test, and there is at least one documented

12

example in print in 1978 to prove that I am right about

13

that.

14

In the

And I have uncovered antecedents much earlier

15

than the 1970s.

16

and from which I have just edited out a few words to make

17

it shorter, was written by Morris Adelman in 1953.

18

asking, 'What is the market,' we must ask, 'What

19

substitutes exist at what price for the product or

20

service in question.

21

concern were the only occupant of the allegedly separated

22

market, would it have the power to raise price?’

23

answer is yes, then the separate market exists within

24

which competition can be lessened."

25

The following passage I'm going to read,

"In

Assuming that a single business

Well, that's the idea of a hypothetical

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

If the

18
1

monopolist test, and I believe Adelman probably was the

2

first one to have it.

3

more widely, but I just dug up this 1953 article last

4

year.

5

His 1956 article has been quoted

The 1982 merger guidelines did a lot to develop

6

this basic idea into something quite useful for market

7

delineation.

8

1982 fine tuned that approach.

9

to draw on experience of 20 years plus in applying the

10

hypothetical monopolist test, and offer insights on

11

possibilities for refining its application further.

12

And revisions of the guidelines in 1984 and
Our panel today is going

I won't waste any time on introductions of our

13

speakers, I will just turn things over to our first

14

speaker, John Harkrider.

15

MR. HARKRIDER:

Thank you, Greg.

The

16

hypothetical monopolist test is one of the organizing

17

principles of the horizontal merger guidelines.

18

is a test that is increasingly applied to define markets,

19

not just in merger cases, but throughout antitrust.

20

not just in the United States, but throughout the world.

21

And it

And

But the hypothetical monopolist test is not

22

"flawless."

Indeed, three years after the 1982

23

guidelines were published, a leading IO economist

24

commented that, "The guidelines market definition test

25

has one wholly decisive defect.

It is completely non-

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

19
1

operational."

2

(Laughter.)

3

MR. HARKRIDER:

Well, history has proved that

4

that concern is overstated, as there clearly are data and

5

quantitative methods that will operationalize the

6

hypothetical monopolist test.

7

It is important to note that using reliable

8

data is critically important.

Because without such data,

9

we quite frequently resort to the Brown Shoe factors that

10

the hypothetical monopolist test was meant to supersede,

11

or at least clarify.

12

My remarks address how to render the

13

hypothetical monopolist test operational by using sound,

14

empirical methods that answer the question posed by the

15

guidelines market definition test.

16

At the outset, it's important to realize that

17

the hypothetical monopolist test asks two subsidiary

18

questions.

19

make a SSNIP unprofitable?

20

sales will, in fact, be lost as a results of a SSNIP?

21

Much of the economic literature has focused on the first

22

question.

23

critical loss attempts to answer.

24
25

The first is, what volume of lost sales will
And second, what volume of

And that is, in fact, the question that

The second question, however, what volume of
sales will be lost as a result of SSNIP is frequently not

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

20
1

answered.

2

agency or a court, may have confidence that a

3

hypothetical monopolist could not profitably impose a

4

SSNIP if it were to lose X percent of its sales.

5

that fact-finder may have no confidence that the

6

hypothetical monopolist would lose more or less than X

7

percent of its sales.

8

And as a result, a fact-finder, whether an

But

The purpose of these remarks is to provide some

9

guidance towards answering the second question:

What

10

volume of sales would the hypothetical monopolist lose if

11

it were to impose a SSNIP.

12

four forms of evidence:

13

evidence; affidavit evidence; and survey evidence.

To that end, we will discuss

historical evidence; econometric

14

The merger guidelines expressly authorize the

15

use of historical data, although they have cautioned it

16

may produce misleading results.

17

often reflect changes in cost affecting the whole

18

industry, rather than market power.

19

reactions to such changes do not illuminate how they

20

would respond to a hypothetical monopolist.

21

Past price changes will

And consumers'

Moreover, price changes typically are of short

22

duration, and thus do not satisfy the non-transitory

23

requirement of the SSNIP test.

24

guidelines themselves caution that the picture of

25

competitive conditions that develops from the use of

For that reason, the

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

21
1

historical evidence may provide an incomplete answer to

2

the forward-looking inquiry of the guidelines.

3

When considering historical evidence, it is

4

important to distinguish between two types of evidence.

5

First, evidence that consumers switched in response to

6

past price changes, and second, evidence that consumers

7

did not.

8

If a significant number of consumers switch

9

from Product X to Product Y in response to a small but

10

significant price change in X relative to Y, this seems

11

to be evidence that the two belong in the same relevant

12

market.

13

The criticism that the price change on X may be

14

cost justified or not sufficiently long in duration would

15

simply suggest that even more consumers may switch from X

16

to Y if the price increase were not cost justified,

17

imposed by monopolists, or of longer duration.

18

Alternatively, evidence that customers did not

19

switch from product X to Y in response to a small but

20

significant price change in X relative to Y, may have

21

little bearing on whether the products belong in the same

22

relevant market, as customers may be more likely to

23

switch if the price increase were not cost justified,

24

imposed by a monopolist, or of longer duration.

25

Another form of historical evidence are natural

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

22
1

experiments.

In an extreme case, if the firm is selling

2

the products in the proposed market, and had in the past

3

engaged in cartel activity, there should be little reason

4

to doubt that the market is properly defined.

5

One court wrote that, "Every price-fixing

6

conspiracy thus identifies directly, in a real-world

7

context, a group of firms which is insulated from outside

8

competitive pressure."

9

conventional market definition evidence attempts to

10

identify artificially.

11

This is precisely what

Another example of a natural experiment is the

12

court's analysis in Staples.

13

various price studies which, on the whole, the court

14

found persuasive, tended to show that a hypothetical

15

office superstore monopolist could impose a SSNIP on the

16

sale of consumable office supplies, because when it was

17

an actual office superstore monopolist, that is precisely

18

what it did.

19

In that case, the FTC's

A form of historical evidence is econometric

20

evidence.

21

in a controlled and scientific manner to answer the

22

hypothetical monopolist test.

23

Econometric evidence uses historical evidence

In fact, Judge Easterbrook, just a few weeks

24

ago, in Menasha Corp v. News Corp criticized a plaintiff

25

for failing to introduce econometric evidence of any kind

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

23
1

in defining the relevant market.

2

There are many types of econometric analysis

3

that can be used to answer the hypothetical monopolist

4

test.

5

effect, allows for a controlled natural experiment.

6

type of analysis requires geographic variation as to the

7

number of competitors, as well as information on pricing,

8

number of competitors, and factors that may influence the

9

price of goods.

One type of the sort used in Staples, which, in

10

This

A second type of analysis directly estimates

11

market elasticities.

To be done properly, this type of

12

analysis requires time series consumer-level information

13

on price and quantity.

14

that in appropriate circumstances, econometric estimates

15

of elasticities can provide perhaps the best evidence as

16

to market definition, there are significant issues with

17

respect to its use.

While there is general agreement

18

One issue involves the ability of data.

19

quite unusual for private parties to have access to price

20

and quantity information from all market participants.

21

This is, of course, less of an issue for the government,

22

but it raises important issues about transparency, as the

23

government may be unable to share this data typically

24

collected through the CID process with the merging

25

parties.

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

It is

24
1

A second issue involves the quality of the

2

data.

For example, retail scanner data, which is one of

3

the few forms of time series data that is available,

4

frequently does not give sufficient information on

5

coupons or other discounts.

6

to opine on whether that level of competition is

7

important.

8

And so your economist needs

A third issue involves the appropriateness of

9

the data.

Retail scanner data, for example, may not be

10

the appropriate type of data to analyze mergers between

11

manufacturers.

12

If you want to know what somebody would do in a

13

hypothetical situation, one alternative simply is to ask

14

them.

15

guideline’s 5 to 10 percent price increase formulation of

16

the SSNIP test.

17

concrete question to ask customers in interviews and

18

depositions.

19

customer affidavits and merger litigation.

20

This may explain the practical appeal of the

It gives lawyers and economists a

It also has led to the frequent use of

There are, however, significant issues with

21

respect to the use of customer affidavits.

22

Turner go so far to suggest that customer affidavits are

23

"the least reliable evidence of whether consumers would

24

switch in response to a SSNIP."

25

Areeda and

Despite this criticism, the government and

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

25
1

litigants in merger cases frequently use affidavits.

2

review of cases involving the use of affidavits revealed

3

two frequent errors.

4

the presence of switching costs.

5

frequently do not cover a representative sample of

6

consumers.

7

A

First, affidavits frequently ignore
And second, affidavits

As to the first question, if consumers will not

8

switch between the products being grouped together for

9

the purposes of applying the hypothetical monopolist

10

test, it may not be meaningful to ask whether they would

11

switch to a product outside of the proposed market

12

following a SSNIP.

13

argued that current GQA customers would not switch to

14

alternatives in response to a SSNIP in GQA was evidence

15

that there existed a relevant market for GQA.

16

For example, in Englehard, the DOJ

However, the 11th Circuit rejected the DOJ's

17

argument in part because Englehard's GQA customers would

18

not switch to other sellers of GQA in response to a SSNIP

19

in the price of Englehard's GQA.

20

The guidelines acknowledge a variation of this

21

point when they speak of adjusting the size of a SSNIP to

22

account for situations where the relevant product

23

constitutes a small percentage of the overall cost of the

24

goods or service.

25

Affidavits also must cover an adequate sample

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

26
1

of customers, and it's important to consider this point

2

in the context of critical loss.

3

you hire Barry Harris, and he tells you that in response

4

to a SSNIP, a hypothetical monopolist would find it

5

unprofitable if it lost 8 percent of its sales.

6

Consider, for example,

Then, imagine the government goes out and

7

interviews 100 customers, randomly selected -- which, it

8

is doubtful, that that is the case, but let's say that

9

they interview 100 random customers.

10

92 of them say that they wouldn't switch and eight of

11

them say that they would.

12

customers purchase an equal level of the good.

13

case the government may very well win.

14

And let's say that

And let's assume that all the
In that

Now, the government frequently doesn't have a

15

random sample.

And so, when you think of it this way, if

16

the government has 100 affidavits or 200 affidavits, or

17

if a private party has 100 or 200 affidavits, that may

18

not be meaningful if there are, of course, more than 100

19

or 200 customers out there.

20

And two cases make this very point.

21

Englehard, the 11th Circuit said, "It is possible that

22

only a few customers who switch to alternatives to make

23

the price increase unprofitable, thereby protecting a

24

large number of customers who would have acquiesced to

25

higher GQA prices."

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

In

27
1

No matter how many customers in each end-use

2

industry the government may have interviewed, those

3

results cannot be predictive of the entire market if

4

those customers are not representative of the market.

5

In SunGard, the court made a similar point.

6

The court wrote:

"The sampling of customer statements

7

before the court is minuscule when compared to the entire

8

universe of defendant’s shared hotsite customers.

9

Although the government has submitted approximately 50

10

statements from customers stating that they either would

11

not or could not switch from shared hot sites, there were

12

more than 7,500 customers that currently used defendant’s

13

shared hotsites.

14

simply cannot determine whether these 50 declarations are

15

representative of the shared hot site client base.”

Without more information, the court

16

In some instances, a few customers, however,

17

may be able to speak to the overall demand elasticities

18

of an industry.

19

may have avoided the sampling problem when it used

20

affidavits of looseleaf chewing tobacco distributors as a

21

proxy for the views of a broad spectrum of customers.

22

For example, in Swedish Match, the FTC

The last form of evidence I want to talk about

23

is survey evidence.

A frequent criticism of consumer

24

affidavits is that they are form affidavits, which simply

25

give yes or no responses to a series of questions,

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

28
1

including whether a customer would switch in response to

2

a SSNIP.

3

really may be is a survey.

4

Well, that may not be an affidavit; what it

And so, if the goal of a client is simply to

5

obtain as many affidavits as possible without taking the

6

time to create detailed, customer-specific affidavits, it

7

may be advisable to conduct a survey instead.

8

conducted properly by an expert trained in acceptable

9

survey methodology, surveys should carry more weight than

10

even an allegedly representative sample of affidavits.

11

For example, surveys may have more

12

methodological controls concerning leading questions,

13

selection bias and randomness.

14

easier to organize customers by key competitive traits,

15

reflecting the probable elasticities of their demand.

16

And perhaps, as an added benefit, it's very difficult for

17

the government to depose survey respondents.

18

Indeed, if

Surveys may also make it

For the hypothetical monopolist test to serve

19

as something more than an instructive theoretical

20

paradigm, litigants, agencies, and courts should look to

21

empirical evidence that can directly answer the question:

22

How many customers will switch in response to a SSNIP?

23

Well, historical evidence, affidavits,

24

econometric evidence, and survey evidence can all be

25

subject to criticism.

There can be little question that

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

29
1

they are better than conjecture on what is a critical

2

question on market definition.

3

MR. WERDEN:

Thank you, John.

4

cross-examination phase.

5

MR. HARKRIDER:

6

And now the

Do I sit over there for this,

or --

7

MR. WERDEN:

Wherever you like.

You're

8

supposed to stand in the dock, I believe, as the

9

defendant.

10

MR. HARKRIDER:

11

MR. WERDEN:

Fair enough.

I am intrigued by your suggestion

12

that we should care whether price increases are cost-

13

justified in examining historical evidence of how

14

customers reacted to price changes.

15

I don't know how customers would know, and I

16

don't know why they would care, whether price increases

17

are cost-justified.

18

thinking on this subject?

19

MR. HARKRIDER:

So, would you elaborate your

Yes.

I'm not sure if it's

20

particularly important whether they are cost-justified or

21

not.

22

frequently change in an industry.

23

those price changes, which may very well be cost-

24

justified price changes, you need to look at that context

25

of a price change in order to determine whether it's

The point I'm trying to make is that prices
And in response to

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

30
1

instructive on the hypothetical monopolist test.

2

I'm not sure if I'm making a significant point

3

as to whether they're cost-justified or not, just simply

4

pointing out that they may very well be cost-justified.

5

MR. WERDEN:

Okay.

Let's turn to your

6

suggestion that somehow switching costs have to be taken

7

into account in market delineation.

8

very simplistic hypothetical, because any real case is

9

way too hard to do in the time we have.

Let me give you a

10

Let's suppose that all the current users of a

11

product are totally locked in, but there are new users.

12

The hypothetical monopolists might be limited to

13

exploiting the current users, and giving up on any new

14

customers, or might be able, in some circumstances, to

15

discriminate between the two.

16

In either event, my question is: Is there any

17

way that the presence of switching costs would cause you

18

to depart from the standard 5 or 10 percent price

19

increase as the significance threshold for market

20

delineation?

21

MR. HARKRIDER:

Yes, I think you make an

22

interesting point, and I think the court in Englehard

23

makes this very point.

24

customers that may be locked in, the dynamic of

25

competition you may want to be looking at is the

When you're dealing with

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

31
1

competition for new customers.

2

point.

3

I think that's the first

I think the second point is that in order for

4

any test or hypothesis to have any meaning, I think a yes

5

answer needs to mean something different, or lead to a

6

different conclusion than a no answer.

7

So, if you ask someone whether they would

8

switch in response to a SSNIP to another product that you

9

believe is in the relevant market, and you ask them the

10

same question with respect to a product that is not

11

within the relevant market, and the answer to both of

12

those questions is no, they would not switch, in order

13

for that test to have meaning you cannot reach different

14

conclusions with respect to whether one is in the

15

relevant market or the other one is in the relevant

16

market.

17

So, if the customer would not switch -- in

18

response to a SSNIP -- to another product in the relevant

19

market, I think you need to take into account switching

20

costs.

21

products that were within the relevant market and a no

22

answer to products that are not within the relevant

23

market, or the test -- or the question has no meaning.

24
25

You need to be able to get a yes answer to

MR. WERDEN:

Well, frankly, I just don't

understand why that is so.

This evidence may suggest

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

32
1

that there isn't any competition between the merging

2

firms, and that would be an interesting thing, if that

3

was the conclusion.

4

But that's not what we're asking when we're

5

delineating the market.

6

hypothetical monopolist would raise price significantly.

7

And I am really having trouble seeing how switching

8

between incumbent sellers of the relevant product who

9

are, by everyone's estimation, in the same relevant

10

market, how switching among them is relevant to the

11

question that market delineation poses.

12

We are asking whether a

MR. HARKRIDER:

Well, I think you're actually

13

answering the question the same way that I would answer

14

it.

15

relevant, and so the point I was making is that

16

affidavits that ask the question that you say is not

17

relevant are probably not relevant to the issue of market

18

definition.

I think the answer is no, it's not particularly

19

So, if the court -- if both the agency and the

20

court had focused on those customers that are not locked

21

in, I think that those are the questions that they should

22

have asked.

23

to customers that are locked in answers nothing, other

24

than the fact that these customers are locked in.

25

But asking the hypothetical monopolist test

MR. WERDEN:

It seems to me that the theory of

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

33
1

the case might be that the locked in customers will be

2

exploited, and that the merging firms will quit selling

3

to new customers.

4

of course, evidence has to be mustered to show that

5

that's a sensible strategy after the merger.

If that's the theory of the case, then

6

But if that is so, then clearly the focus in

7

market delineation ought not to be on the new customer,

8

should it?

9

MR. HARKRIDER:

Well, I think it's important --

10

I am an antitrust lawyer, and you're an economist.

And

11

not to make that distinction, but the court in Englehard

12

certainly gave no weight to the affidavit statements,

13

because of the presence of switching costs.

14

And I think that unless you're arguing and can

15

persuasively argue that Englehard is wrong, I think that

16

it's very difficult to argue that switching costs don't

17

matter in that context.

18

MR. WERDEN:

19

MR. HARKRIDER:

20

MR. WERDEN:

21

we don't have time to go into it.

22

(Laughter.)

23

MR. WERDEN:

24
25

Well, I have it here.
Okay.

And I don't read it that way.

So we're going to move on to our

next speaker, Barry Harris.
MR. HARRIS:

But

Thanks, Greg.

Two comments,

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

34
1

first.

Just -- you mentioned about being in the dock.

2

actually testified a few months ago, and there was a

3

federal judge in Rhode Island that actually makes the

4

witnesses stand.

5

I

And I didn't understand that, I thought the

6

chair was just missing, and I went to bring a chair over

7

when it was time to testify, and they told me, "No, we

8

don't do it that way, so you have to" -- and I guess it's

9

an incentive to give short answers.

10

But in any case, second thing, everyone is

11

talking about 20 years of experience.

12

I will be -- it was 30 years ago that I came to work for

13

the Antitrust Division, so I hope senility hasn't set in.

14

The one good thing about having done it so long is that I

15

am now able to do analysis by anecdote, so let me start

16

by telling a story about a hypothetical monopolist case.

17

I started working in the Antitrust Division in

18

October of 1974, and a year or two after that I was asked

19

to draft an affidavit on a case that I was working, a

20

merger case, for a possible challenge.

21

I am getting old.

And I looked at some old affidavits, and as I

22

went through them I learned, "Hey, I'm supposed to define

23

a market."

24

didn't have a clue how to define the market.

25

what seemed to me to be obvious, I went around and I

And being recently out of graduate school, I

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

So I did

35
1

talked to senior lawyers and senior economists, and they

2

gave me a lot of different suggestions.

3

The lawyers gave me the Brown Shoe standards,

4

and some of the economists said, "Look at price

5

correlations."

Some others said, "Look at cross-

6

elasticities."

But none of the suggestions -- not one --

7

had anything to do with the actual exercise of market

8

power, it had to do with things that you might observe if

9

market power were being exercised, but it wasn't a kind

10

of a logical progression.

11

So then I finally did what I should have done

12

first, and I went to see my boss, who was George Hay, and

13

he suggested looking at it as a hypothetical monopolist.

14

And it was the first I had heard of it.

15

Greg, I hadn't gotten to reading books from 1952 in a

16

systematic way, and probably should have.

17

I wasn't like

But in any case, once the question was put with

18

that type of framework, it all started to make a lot of

19

sense to me.

20

logic of the hypothetical monopolist principle.

21

least as I understand it, the point of it is simply to

22

identify a group of producers that would be able to

23

exercise market power if -- and this, for me, is a key

24

second part -- if they were able to coordinate their

25

pricing output decisions.

And that ultimately gets to what is the

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

And at

36
1

So, it's basically a two-step process.

One

2

thing that I notice that I think often is not done

3

properly is that when you're considering whether or not

4

the group can coordinate, it seems to me you have to ask

5

what the nature of the coordination is going to be, and

6

it has to be consistent with the logic of what that

7

hypothetical monopolist is going to do.

8

In other words, as Greg points out, the basic

9

principle is would the hypothetical monopolist price in a

10

certain way and would it be profitable.

11

going to ask the questions about coordination, the second

12

step, it seems to me you have to go back to your

13

definition when you're doing your competitive effects

14

analysis.

15

Let me just reiterate that.

Well, if you're

The hypothetical

16

monopolist proposition tells you what the hypothetical

17

monopolists would do.

18

you do your competitive effects analysis it seems to me

19

it's incumbent upon you that you look at that in the

20

context of what that hypothetical monopolist would be

21

doing.

22

You look at that, and then when

And let me give you an example.

Let's say that

23

a hypothetical monopolist prices differentially -- so it

24

might be unilateral effects, might be something more

25

complicated than that.

Analysis of that market must

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

37
1

consider the likelihood that the coordination of this

2

type of differential pricing could actually be achieved.

3

The basic concept related to the hypothetical

4

monopolist principle brings me to the focus of the

5

outline that I provided Greg for this session.

6

pointed out, I often use critical loss both before the

7

agencies and in testimony.

8

several questions raised about the use of critical loss.

9

Is it appropriate?

10

appropriate?

11

And recently there have been

Under what conditions is it

And to understand, let me just give brief

12

background on critical loss.

13

loss basically does is it makes the hypothetical

14

monopolist principle operational.

15

As John

In my mind, what critical

And all critical loss is is a two-step process

16

that first identifies for any given price increase the

17

sales that can be lost before the price increase becomes

18

unprofitable.

19

much sales have to be lost before it becomes

20

unprofitable?

21

So, in effect, it's setting a target.

How

The second step considers whether the actual

22

sales loss associated with the hypothesized price

23

increase will exceed the calculated critical loss.

24

it's two distinct steps.

25

third step, because this form of the critical loss looks

So

And one might even argue it's a

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

38
1

at a particular price increase.

And in fact, the profit-

2

maximizing hypothetical monopolist can charge any number

3

of prices.

4

look at a variety of prices and try to answer that

5

question.

So, a full application of critical loss would

6

Now, a practical consideration is -- as the

7

size of the price increase goes up, the confidence you

8

have in any answers you get seem to me to become less and

9

less.

10

you're working with an observed interval of price

11

increases.

12

or 10, if you're talking about a 50 percent price

13

increase, you probably haven't observed changes like

14

that, and people's opinions are probably not as reliable.

15

Now, the answer to the first step, what is the

For example, you have greater confidence when

Whether people deal with 5 percent increases

16

critical loss, how much can you afford to lose, that

17

ultimately depends on the size of the price increase

18

you're postulating, and the cost structure of the

19

hypothetical monopolist.

20

tells you for every unit of sales you lose how much it

21

hurts.

22

In effect, the cost structure

The answer to the second step, how much will

23

actually be lost, can be ascertained by traditional

24

antitrust methods such as econometrics, other statistical

25

techniques, looking at business records, and going

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

39
1

through testimony.

2

talked about in the previous talk.

3

And this, in part, was what John had

Now, one of the questions that has been raised

4

about critical loss concerns the relationship between

5

cost price margins and demand elasticity.

6

been three or four articles that raise this point.

7

in particular, Michael Katz and Carl Shapiro have raised

8

the point that economic theory indicates that high

9

marginal cost margins tend to imply that actual loss

10

sales associated with the price increase will be small.

11

Well, the Carl Shapiro observation is based on

12

a theoretical inverse relationship between marginal cost

13

margins and demand elasticity that's encompassed in

14

what's called the Lerner index.

15

is a residual of looking at profit maximization under

16

certain circumstances.

There have
But

And that, effectively,

17

Now, there are several reasons why large gross

18

margins may be consistent with unit sales actually being

19

sensitive to price changes, and Katz and Shapiro identify

20

three of them.

21

already be coordinating prices, and that will sever the

22

relationship.

23

there is a kink in the demand curve, and also situations

24

in which there is a kink in the supply curve.

25

They note that firms in the market may

They also identify situations in which

I think there are a few other issues you have

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

40
1

to look at, and one of those, I think, is that the use of

2

the Lerner Index for this purpose fails to distinguish

3

between margins that are based on marginal cost --

4

meaning the margin of the last unit -- and those that are

5

based on the average variable cost of a significant

6

increment of quantity, where the answer may be quite

7

different.

8

The correct margins used in a critical loss

9

analysis should consider the actual average variable cost

10

associated with a significant loss of sales, often more

11

than 10 percent of current sales.

12

example, one of the situations where I presented critical

13

loss was the Poplar Bluff Hospital merger case.

14

v. Tenet.

15

And just to give an

It's FTC

And there, when we went back and we looked at

16

the costs, the question we asked -- and there is a little

17

bit of a chicken and egg thing there, and you have to be

18

careful -- but we knew from previous work that margins

19

for hospitals were somewhere in the range of 60 percent,

20

because there are a lot of fixed costs there, a lot of

21

fixed staffing.

22

So, what we did there was we got a hold of the

23

accounting system, we sat down, we worked with their

24

internal cost accounting system, and went through, item

25

by item, and asked the question in conjunction with

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

41
1

business people, "What would happen if you knew you were

2

to lose 10 percent of your sales for a year?

3

say you lost a big managed care contract.

4

happen to the individual cost elements?"

5

Like, let's

What would

So, in effect, asking what's variable over the

6

course of the year.

7

used, and we ended up with margins there that were

8

similar to other hospitals.

9

surprise -- that if you took greater change in demand,

10

say 20 percent, or took a longer period of time, say two

11

years, then a lot of costs that are not variable under

12

the hypothetical that we presented become variable, and

13

the calculated margins change.

14

And the answer we got was what we

But lesson learned -- not a

So, I guess the lesson from this is simply get

15

the right costs.

Make sure that it fits the situation

16

and the questions being asked for the specific case.

17

Now, second point with regard to the Lerner

18

Index, going back there, is that the relationship that it

19

describes is limited in some sense.

20

monopolists and dominant firms, and it also applies to

21

industries where Bertrand-type competitors sell

22

differentiated products.

23

It applies to

Now, and despite the fact that many, many years

24

ago I majored in math, I don't consider myself a

25

mathematician, so this is -- when I went back and looked

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

42
1

at Cournot models, you don't get the relationship in

2

Cournot models -- at least my understanding of that.

3

Now, it seems to me that if you're going to be

4

questioning things like critical loss, or anything else

5

based on particular models such as Bertrand-type models,

6

then the analysis in the case has to fit that model, and

7

has to incorporate all the implications of the specific

8

form of the Bertrand model -- in this case, that applies.

9

So, for example, Bertrand models tend to apply

10

best for differentiated products.

11

that you have -- the relationship that Katz and Shapiro

12

talk about that comes from the Bertrand model doesn't

13

tell you a whole lot about closeness of substitutes.

14

But the relationship

Again, my understanding is that it doesn't tell

15

a whole lot about cross-elasticities.

16

analyzing differentiated products, a key issue is are

17

these close substitutes, and what happens, for example,

18

if you raise the price of one of them.

19

doesn't apply there.

20

But when you're

The relationship

And ultimately, this all brings us back to the

21

hypothetical monopolist principle -- I see my time is

22

running out.

23

the guidelines is a major leap.

24

probably very specific circumstances where other

25

principles can apply, what it does is it focuses analysis

My overall view is that the application of
Although there are

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

43
1

specifically on the questions that it's supposed to be

2

asking, and that is can market power be exercised.

3

Actually, I assume my time is up, so I'm just

4

going to just say thank you, and leave myself open to

5

Greg's questions.

6

MR. WERDEN:

Thank you, Barry.

I want to ask

7

you a couple of questions about your reaction to Katz and

8

Shapiro, and I want to start at an extremely basic level.

9

Their most basic point is that anybody who is

10

trying to make something out of high margins ought to

11

have some idea of what competitive process produced those

12

high margins, and be able to construct a comprehensive

13

argument that takes into account the competitive process

14

that produced those high margins.

15

that?

16

MR. HARRIS:

Do you agree with

I mean, it seems almost

17

tautological.

18

process.

19

-- the last unit may be very small, but the question is

20

what does it look like over the range of the lost sales.

21

And that is basically a function of the cost structure in

22

the particular market is a market that has high fixed

23

costs.

24

minimum staffing that's not going to change?

25

Of course you want to understand the

The question, I think, is -- where we differ is

In the case of the hospitals, does it have high

Ultimately, any firm has to recover its costs,

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

44
1

or it goes out of business.

2

simply be reflecting high fixed costs.

3

MR. WERDEN:

And the high margins may

Well, unfortunately, Mike can't be

4

here today to say no, that isn't right, so I will say it

5

for him.

6

I think what Mike would say is, "Have you got

7

it backwards?"

8

are high.

9

process will, in equilibrium, allow you to recover those

10

fixed costs, and that may mean that there isn't much

11

competition.

12

wouldn't be able to recover those fixed costs.

13

Prices aren't high because fixed costs

If fixed costs are high, then some competitive

Because if there was competition, you

So, whatever the facts are, there is a

14

competitive process out there that's producing those

15

margins.

16

examine that process and try to make sense out of it.

17

I think that you're disagreeing with them at

18

square one, and saying they're wrong about saying that

19

the competitive process had to be responsible for

20

creating the high margins.

21

margins.

22

And what Mike and Carl say you need to do is to

You're saying costs create

Is that right?
MR. HARRIS:

Well, that wasn't exactly what I

23

intended to say.

I mean, I think competition ultimately

24

provides competitive prices equal to marginal costs.

25

Maybe included in that is going to be the return on some

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

45
1

costs, or people will not enter, and they're not going to

2

have as many competitors.

3

But I think part of the problem, too, is that

4

we're talking about a loss of sales here.

So if you are

5

having some costs, you may have entered -- you may have,

6

at the margin, a very low margin -- by at the margin, I

7

mean the last units or so.

8

But if you have a large amount of fixed costs,

9

as you eliminate, let's say, 10 percent or 15 percent of

10

your sales, you may have a cost curve that has a slope on

11

it.

12

curve can be substantial.

13

the shape of that cost curve is.

14

large margin for the very last unit, which is all the

15

Lerner-index relationship says.

16

MR. WERDEN:

And the difference between the price and the cost
It ultimately depends on what

Okay.

And you needn't have a

Let's change the topic only

17

very slightly.

You point out that their analysis is

18

based on the Bertrand model, and obviously, that isn't

19

the right model for all industries.

20

doubt in my mind that it's the right model for some

21

industries.

But there is no

22

And so, my question to you is, well, if it is

23

the right model and the margins are high, then doesn't

24

the analysis lead to their conclusion that the high

25

margins are consistent with very narrow markets, not

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

46
1

broad ones?

2

MR. HARRIS:

I think it's going to depend --

3

and I'm not sure I completely understand the question --

4

but I think it's going to depend -- because in the

5

Bertrand model, you have differentiated products and you

6

may very well have different margins with the different

7

firms.

8

But that doesn't say anything is -- ultimately,

9

you know, you're getting to the merger, you're asking

10

what the cost elasticities are between different firms in

11

the market.

12

differentiated products, and I'm not sure --

That's important when you're dealing with

13

MR. WERDEN:

And they say it is.

14

MR. HARRIS:

That --

15

MR. WERDEN:

That the result is that if you're

16

in this Bertrand world with very high margins, then it

17

turns out that you need very little substitution, very

18

low cross elasticities among the firms in the industry in

19

order for the right conclusion to be that they form a

20

relevant market.

21

MR. HARRIS:

Well, I mean, in the simple model,

22

for example, let's say you just have two firms and they

23

produce similar products but have different costs -- not

24

exactly the same products, but have different cost

25

structures.

You're going to have one firm that has much

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

47
1

larger margins than the other.

2

close competitors, but if that firm with the better cost

3

structure is fully maximizing price, it's going to have

4

the large margins to move up to the cost structure of the

5

other firm.

6

sales.

7

They compete, they're

But if it raises it any further, it loses

I mean, that's basically the Cellophane issue.
So, again, I'm not sure that I am understanding

8

your question, but I don't see that there is a conflict

9

there.

10

MR. WERDEN:

Well, the question is:

Aren't

11

they right in their case?

12

it true that high margins are almost always going to lead

13

to narrow markets because even if there isn't that much

14

competition among the firms in the market, there is

15

enough so that the market elasticity is enough lower than

16

the individual firm elasticities which, of course, have

17

to be low if the firms have high margins?

18

In the Bertrand world, isn't

MR. HARRIS:

I guess I just don't know the

20

MR. WERDEN:

All right.

21

MR. HARRIS:

Okay, thank you.

22

MR. WERDEN:

Our next speaker is Will Tom.

23

MR. TOM:

19

answer.
Thank you, Barry.

Thank you, Greg.

I'm not quite as

24

old as Barry, but I'm getting there.

And I think back to

25

my first year as a young lawyer in the Antitrust Division

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

48
1

in 1979.

2

thinkers like Greg were already using the hypothetical

3

monopolist test, according to Greg at least since 1978, I

4

was there as a brand new lawyer, given my first merger

5

matter to look at, and I have to say I had a complete

6

sense of befuddlement about what to do, how to define a

7

market, and how to start thinking about this case.

8

And despite the fact that sophisticated

And had I had more sense, I would have went up

9

a flight of stairs and asked Greg what to do, just as

10

Barry was able to get good guidance from George Hay.

11

I didn't, and so I muddled along as best I could.

12

that perspective, I think the 1982 merger guidelines was

13

really an advance.

14

off the charts or breaking here, I think it should be the

15

ones in front of the people like Greg, who were actually

16

involved in that effort and in developing the guidelines

17

and propagating its use.

But

From

And if any ego-meters should be going

18

From that perspective, I think what I have to

19

offer are really just a handful of modest suggestions.

20

think, on the whole, the guidelines are working well.

21

think the guidelines approach is a good one, and

22

notwithstanding the quote from George Stiegler that John

23

mentioned, it really is operational, by and large.

24
25

I
I

And my handful of modest suggestions come under
two headings.

Heading one, we ought to have a more

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

49
1

explicit recognition of the role of uncertainty as we

2

apply these guidelines.

3

that we ought always to bear in mind that the underlying

4

purpose of market definition is the assessment of

5

competitive effect.

6

try to put a little bit of flesh on it as I go along.

And heading two, I think, is

I think that's a truism, but I will

7

Under the role of uncertainty, I guess my first

8

suggestion is that the staff be aware of the tendency of

9

the iterative hypothetical monopolist test to

10

unconsciously reverse the burden of proof.

11

Now, I know that the guidelines explicitly say

12

this is not intended to say anything about burden of

13

proof, burden of persuasion, where that lies, or anything

14

else.

15

burden of proof on those who are asserting that something

16

inherently uncertain will occur, as opposed to those who

17

say that something inherently uncertain will not occur.

18

But there is a natural human tendency to put the

And if you think about how the iterative

19

hypothetical monopolist test works, you start by positing

20

a 5 percent -- or, you know, some other significant non-

21

transitory price increase.

22

taken as a given, that's a fact.

23

would happen in response to that price increase by a

24

hypothetical monopolist if that price increase were to

25

take effect?

So, the price increase is
Now you ask, well, what

Would competition come into this tentative

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

50
1
2

market to defeat that price increase?

Right?

And so, you know, the real answer in many cases

3

is, well, who the hell knows?

Right?

But since that

4

really is a non-operational approach to the practical

5

question of getting through the analysis of a particular

6

merger, you know, the next step is to say, well, what do

7

we think is going to happen?

8

And it is a very small step to slide from there

9

to saying, well, parties, you know, prove to us that the

10

price increase by this hypothetical monopolist will, in

11

fact, be defeated, right?

12

iterations, you know, prove that the price increase would

13

be defeated, and you end up with, you know, a very high

14

threshold, I think, of establishing that the market is as

15

broad as maybe it really is.

16

And you go through enough

And so, there may be a certain inherent bias in

17

favor of markets that are too narrow.

18

prove it, it's just an intuition, but I think that that

19

may be just a natural consequence of the way we approach

20

questions that are inherently uncertain.

21

And now, I can't

My second point about uncertainty is in a

22

similar vein.

And it's not really a market definition

23

question as such, it's a question of, you know, the next

24

step after market definition, and the next step after

25

that, and how it relates to the process of market

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

51
1

definition.

2

And I am reminded of a scene in the play or

3

movie, "Amadeus," where the Emperor Joseph says of a

4

Mozart opera, "Well, the only problem here is too many

5

notes."

6

the many steps in the merger guidelines process is to

7

lead to some errors in situations of an inherent

8

uncertainty.

And there may be a sense in which the effect of

9

And to take a very stylized example, imagine a

10

merger investigation in which the analyst concludes that

11

there is a 51 percent probability that the merging

12

parties A and B are actually in the same market

13

themselves, right?

14

product of one of the merging parties, and you ask in the

15

event of a SSNIP what would happen, and would the product

16

of B come in to thwart a price increase by A.

17

you go from there.

18

Okay.

I mean, the -- you start with a

And then

Let's suppose that there is a serious

19

question about that first step, and maybe the merging

20

parties aren't even in the same market.

21

with 51 percent probability, they probably are in the

22

same market.

23

All right.

But you conclude

Next step you have got another

24

player, C.

He might be an uncommitted entrant, right?

25

He might be able to start selling this product, even

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

52
1

though he is not currently selling the product without

2

substantial cost in a very rapid period of time.

3

Well, let's suppose you can reject that

4

proposition, again, by a 51 to 49 percent margin, right?

5

There is only a 49 percent likelihood that uncommitted

6

entrant would come in.

7

participant in the market.

8

And so you say C is not a

And let's suppose there is another potential

9

entry, D.

10

related product, so he would have to expend some costs.

11

And so, after careful analysis, you conclude that, well,

12

there is only a 49 percent likelihood of committed entry.

13

And therefore, more likely than not, D will not enter the

14

market.

15

And maybe he's not currently producing a

Okay.

Well, how do you analyze this merger as

16

a whole?

17

be anti-competitive, and assuming a lot of those

18

probabilities are independent, you ought to say, "Okay,

19

well, what's 51 percent times 51 percent times 51

20

percent," and you come out with about a 13 percent

21

likelihood that this merger is going to do any harm,

22

right?

23

If the question is:

Is this merger likely to

I think there is a tendency for at least

24

beginning analysts to say, well, more likely than not A

25

and B are in the same market, okay?

Now, is there going

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

53
1

to be uncommitted entry?

2

the market?

What are the participants in

3

Well, more likely than not, there is not going

4

to be uncommitted entry, so we will reject the argument

5

that the parties are advancing about uncommitted entry.

6

Now, you know, is there going to be entry as the

7

guidelines define it?

8

is not going to be entry, right?

9

know, two merging parties in the same market.

10

else in that market, no entry likely to come in.

11

Therefore, we should challenge the merger.

Well, more likely than not there
So you have got, you
Nobody

12

It is, I think, really difficult for the

13

merging parties counsel to know when the staff is

14

thinking this way.

15

process whereby the staff is properly reticent about

16

sharing information that they are learning from third

17

parties, and that can sometimes inhibit the dialogue.

18

And if they are thinking this way, it's sometimes hard

19

for the parties even to know it, let alone talk them out

20

of it.

21

You have a somewhat asymmetric

So, I throw that out as something for the

22

agencies to at least be cautious about in their internal

23

deliberations.

24
25

Third uncertainty point, if you want to call it
that, is the whole issue of customers say they wouldn't

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

54
1

switch.

2

morning about affidavits and survey evidence, and so on

3

-- particularly in intermediate goods industries, where

4

the staff is relying on customer telephonic interviews in

5

ways that they probably are not in consumer goods.

6

And we're talking about that at some length this

Again, it can be hard to tell what's actually

7

going on there behind the curtain.

8

out that when you're doing survey evidence, your survey

9

experts will take quite a lot of pains to work with you

10

to get the biases out of the questions and to, you know,

11

avoid some of the more obvious errors about how the

12

questions are phrased, and the sample of respondents

13

they're posed to, and so on.

14

I think John pointed

Where you're relying on the telephonic

15

interviews with customers, the danger of badly worded

16

questions eliciting misleading information about the

17

inherently unknowable is probably at its highest.

18

is also the situation in which fruitful dialogue between

19

the party's counsel and the staff is at its most

20

difficult because of concerns about third-party

21

confidentiality.

22

And it

And it is very easy to slip into highly

23

unproductive dialogues that start and mostly end with

24

staff saying, "Well, customers say they wouldn't switch."

25

And I know Greg has been preaching this for a very long

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

55
1

time, and I think, by and large, the agencies are doing

2

this:

3

what the customers say they would do and focus all of

4

your attention on the reasons that they give for what

5

they say what they would do, and really dig into the

6

objective facts that make switching more or less likely,

7

because that is something that the parties generally can

8

engage on.

The dialogue is much more productive if you ignore

9

And it is a lot more concrete and reliable, I

10

think, than the speculation about what, you know, what

11

third parties -- what customers say they would do under

12

circumstances that they may not really have faced and

13

really may not have thought through.

14

All right.

Well, let me turn to the other

15

major heading, which is that assessment of competitive

16

effect is the underlying purpose of market definition.

17

And I guess the first question I want to

18

address is when should we depart from five percent?

19

know, the guidelines say that in appropriate

20

circumstances, we can use numbers other than five percent

21

as your SSNIP, but they don't say what those appropriate

22

circumstances are.

You

23

And let me take you through an example that's

24

already covered by the guidelines, because I think it's

25

easier to understand and then go from there to situations

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

56
1

I think that aren't really covered by the guidelines.

2

Well, let's suppose you have got a market in

3

which a five percent price increase is unprofitable, but

4

a 10 percent price increase is profitable.

5

would that happen?

6

price increase, you lose a certain block of customers.

7

You don't make enough on the customers that you keep to

8

make up for the loss of revenues from the customers you

9

would lose.

10

increase is unprofitable.

11

I mean, why

Well, in response to a five percent

And therefore, the five percent price

The next tranche of customers you would lose is

12

highly price-inelastic, though.

13

price by 10 percent instead of 5 percent, you don't lose

14

very many more customers, but you make a lot more on the

15

customers that you do keep.

16

of his writings, has dealt with this situation, and maybe

17

others have, as well.

18

So if you raised the

And I gather Greg, in some

In that situation, if the guidelines simply

19

said, you know, look at a SSNIP and see what would

20

happen, then you would say, "Well that would be a good

21

case for upward departure on the five percent, because

22

after all, the purpose of the guidelines is to detect the

23

likelihood of competitive effect, I'm simplifying past

24

the coordinated effects step, and assume perfect

25

coordination, or assume only two players or something --

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

57
1

it wouldn't make any sense to have a set of guidelines

2

that protected against a 5 percent price increase but

3

didn't protect against a 10 percent price increase.

4

As Greg has pointed out, the guidelines say "at

5

least."

6

a price increase of 10 percent, then the SSNIP test is

7

satisfied, even under the hypothetical that I am talking

8

about.

9

So, if a hypothetical monopolist could maintain

So, the case for an upward departure would have

10

to involve somewhat more exotic examples.

11

and C are merging.

12

hypothetical monopoly of A and B could not sustain a 5

13

percent price increase for the same reasons I described,

14

but could do so with a 10 percent price increase.

15

Imagine that A

B is a closer competitor than C.

A

And you know, should you call this a market and

16

examine the competitive effects of a merger of A and C,

17

or should we say that C is outside the market all

18

together, and kind of stop your investigation there.

19

There are scenarios in which it would make sense to keep

20

going.

21

The downward departure scenario is more

22

straightforward.

Five percent price increase is not

23

sustainable, but a one percent price increase would be,

24

because relatively few customers are diverted.

25

going to be the inverse of the example that I started

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

It's

58
1

with.

2

And there, I think the theoretical case is

3

there for a downward departure I guess I have qualms --

4

hearkening back to the theme of uncertainty in the last

5

group of examples I talked about.

6

price increase of five percent or more couldn't happen,

7

even with perfect coordination among all marketplace

8

participants, that puts an upper bound on how much damage

9

can be done if you fail to challenge the merger.

10

If, by hypothesis, a

Given the vagaries and the uncertainties of

11

life, maybe that should be enough.

12

somewhat facetiously that maybe we should delete the part

13

of the guidelines that says this five percent is not a

14

tolerance level.

15

tolerance level in the merger guidelines.

16

I have suggested

And yet there ought to be some

And I am not sure what the right tolerance

17

level is, but it should be one that recognizes that, you

18

know, the whole process of merger analysis is an inexact

19

science, at best.

20

that there is a fairly low upward bound on any damage,

21

maybe you ought to stop there and let this merger go

22

through.

23

And if you're reasonably confident

All right.

I am running short of time, so let

24

me hasten quickly through my last two points.

25

aesthetic reasons maybe, if nothing else, can we do

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

Just for

59
1

something to reconcile the unilateral effects discussion

2

of competitive effects with the market definition

3

process?

4

And here, in particular, I focused on the

5

sentence in Section 1.11 of the guidelines that reads --

6

and I quote -- "In performing successive iterations of

7

the price increase test, the hypothetical monopolists

8

would be assumed to pursue maximum profits in deciding

9

whether to raise the prices of any or all of the

10

additional products under its control."

11

And I am not entirely sure what that means, but

12

I think what that means is that if you can show the

13

sustainability of a five percent price increase in only

14

one of the hypothetical monopolist products, that's

15

sufficient to define a market.

16

You know, if that's right, then why do you need

17

analysis?

18

define a market based on the hypothetical monopolist

19

raising price on one product, you know, doesn't that mean

20

that any decent unilateral effects case is a merger to

21

monopoly?

22

This aspect of market definition that says you

And I guess, you know, the bottom line on this

23

one is I'm just confused.

I'm not sure that this anomaly

24

has actually done any harm to merger analysis, but it

25

certainly has made it a whole lot more confusing and

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

60
1

maybe has made the guidelines just a little bit less

2

transparent to pick up one of the themes raised by our

3

keynote speakers.

4

And given the amount of time, I am going to

5

drop my last supply substitute ability point, which Greg

6

tells me fits better in another session, anyway, and open

7

myself up to the cross examination.

8

MR. WERDEN:

Okay.

Thanks very much, Will.

I

9

want to explore some of these suggestions you made, and I

10

want to start with the issue of the upward departure.

11

might not be an accurate characterization, but we will

12

use that as a shorthand.

It

13

And the case -- I think the only case -- that

14

really focused on this issue is the Olin case in the 9th

15

circuit where, in order to sustain the FTC's decision

16

that the merger was unlawful, the 9th circuit had to be

17

persuaded that a 5 percent price increase ought not to be

18

looked at, but rather a 10 or more percent increase

19

needed to be looked at.

20

than 10.

21

And it maybe needed to be more

And the FTC did persuade the 9th circuit of

22

this, and I think this illustrates what I think you had

23

in mind by an upward departure.

24

comment on is whether you think the FTC and the 9th

25

circuit had it right.

And what I want your

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

61
1

MR. TOM:

All right.

I am not going to comment

2

on the specific case, only because I haven't studied it

3

recently.

4

the beginning, one of the first things that goes is

5

memory.

6

And giving the advancing age I alluded to at

MR. WERDEN:

Well, neither of us actually knows

7

the facts of the case, so that's not really what my

8

question is.

9

percent price increase these two swimming pool sanitizer

10

chemicals would be in the same market, should you just

11

say, "Well, okay, put them in the same market?"

12

My question is:

MR. TOM:

Is the idea if for a 15

Yes, I think it depends.

If there is

13

a credible theory of competitive effect that says, in

14

effect, that, unless we do something there will likely be

15

an anti-competitive effect -- I wonder whether we ought

16

to let market definitions stand in the way.

17

And I alluded to the hypothetical situation in

18

which you have got three players, and they really aren't

19

-- back up a second, because I didn't go into this

20

hypothetical too deeply.

21

I posited three players, because you can

22

imagine a two-player market in which a five percent price

23

increase would not bring one of the party's closest

24

competitor into the market.

25

with not a hypothetical monopoly, but an actual monopoly.

I mean, yes, let's start

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

62
1

Even a five percent price increase will not bring the

2

closest competitor in.

3

And you say, should you, let A merge with B,

4

its closest competitor?

5

intuitively, no.

6

what would be a fairly dramatic competitive effect under

7

that circumstance.

8

And I think most of us would say

It wouldn't make any sense to allow

Well, one might cavil about whether that

9

hypothetical is realistic in the first place, because if

10

a five percent price increase by the monopolist wouldn't

11

bring closest competitors in the market, then why isn't

12

the monopolist pricing it at that level already?

13

And so, I posited another competitor, B, that

14

is closer than competitor C that A is acquiring.

15

current price is constrained by B.

16

haven't gone up already.

17

So the

And that's why prices

Now, the question is can the acquisition still

18

make a difference?

19

which the acquisition of the next closest competitor does

20

make the difference.

21

I think there are certainly models in

In such a circumstance, does it make sense to

22

say, "We should treat C as not being in the market in the

23

first place so why worry?"

24

probably want to look at that model a little more closely

25

and see this competitive effect really realistic?

I think not.

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

I think you

Is it

63
1

likely to happen?

2

Are you convinced as a factual matter?

MR. WERDEN:

One last question relating to your

3

comment on tolerance level.

4

you suggest, to recognize uncertainty in our forecasts.

5

But I want to try to flesh out exactly what you're saying

6

here.

7

I think it is important, as

Let's suppose we had a unilateral effects case

8

in which the government believed that the result of a

9

merger would be one merging firm would increase price

10

eight percent, and then the other one would increase

11

price four percent, for a market-wide average price

12

increase of three percent, or make it two percent.

13

The three and the two percent are below the

14

magic five.

15

the merging firms is below the magic five.

16

there isn't a tremendous amount of certainty, so that we

17

really don't know anything, do you really believe that

18

the agencies should say, "Well, eight percent, four

19

percent, forget it.

20

MR. TOM:

21

In fact, even the four percent for one of

It's not that much."
The assumption was there is not a

tremendous amount of uncertainty, or --

22

MR. WERDEN:

23

don't know anything.

24

uncertainty.

25

Assuming that

MR. TOM:

Well, not so much that we really
There is always significant

Yes.

No, I'm not proposing selecting

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

64
1

a magic number of five percent, or whatever, as a

2

tolerance level.

3

where we can be pretty confident about the upper bound of

4

harm, and we can't be at all confident -- or we can be

5

somewhat confident, but only barely so -- that there is

6

any harm at all, I think we ought to very seriously

7

consider letting that one through.

8

Yes, I do think that, in circumstances

MR. WERDEN:

And not a question, but just a

9

comment.

The sentence of the guidelines that you quoted,

10

in fact, does not mean what you supposed, but the

11

following sentence does.

12

(Laughter.)

13

MR. TOM:

14

MR. BLUMENTHAL:

Okay.

I stand corrected.

Thank you.

Good morning, everybody.

I am

15

Bill Blumenthal, and as the clean-up hitter, I suppose

16

I'm going to reshape my comments a little bit to address

17

some of the things that were said by the three earlier

18

batters.

19

We begin by noting that there is not a whole

20

lot of disagreement among the speakers, at least so far,

21

save for two things that my former partner and good

22

friend, Will Tom, said in passing.

23

The first was simply the reference to too many

24

notes.

And my recollection is that that was not Emperor

25

Joseph criticizing Mozart, but Mozart criticizing

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

65
1

Salieri.

And I'm pretty sure Tim Hulce was the one who

2

said it, but we're going to have to go back and check the

3

tape.

I'm not going to dwell on that point.

4

The second thing where I think we had a little

5

bit of disagreement -- and again, it was sort of a throw-

6

away line by Will -- was the observation that, all in

7

all, the guidelines are working well.

8

more time on that one because, to me, I mean, the

9

guidelines are working well, I suppose, but only because

10

most of the people who are using them know what the code

11

words mean, and are largely ignoring them as the recipe

12

for which they were originally intended.

I'm going to spend

13

And that's really the point where I am going to

14

devote most of my remarks, but let me first begin with a

15

little bit more on the areas of agreement -- or I think

16

agreement -- among us all.

17

The hypothetical monopolist test has basically

18

won.

19

most of the commentators adopt it.

20

adopted in the courts, but it has largely prevailed in

21

the courts.

22

some tools as early as 1983 or 1984.

23

have been amended substantially.

24
25

I mean, I think pretty much all analysts adopt it,
It is not universally

We generally know which tools work.

We had

Those, obviously,

We had a pretty good set of information about
which tools do not work.

I agree completely with John

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

66
1

Harkrider and Will Tom, that affidavit battles are not an

2

especially productive or illuminating form of arms race.

3

I agree with Will on the issues of burden of

4

proof, and biased narrow markets, although whether it's

5

too narrow is sort of a judgment call we can talk about.

6

But unilateral effects, I agree with Will about his

7

supply-side point that he didn't make here, but you will

8

have to read the paper.

9

And basically, I think pretty much all of us

10

would say that if market definition is to remain a

11

discreet early step in merger analysis, that the

12

hypothetical monopolist test is the right test.

13

Now, to the disagreement, let me be a little

14

bit incendiary at least, and pose to you the question of

15

whether, in fact, we ought to be retaining market

16

definition as a discreet, early step in the analytical

17

process as it is used nominally, under the guidelines.

18

For at least five years now -- and just

19

watching the way we work, watching the way the agencies

20

work -- I have been posing to people the question of

21

whether market definition is an input into the

22

competitive analysis, or whether it isn't really an

23

output.

24

that it's neither, it's really an issue of simultaneous

25

determination.

And I think that Greg has largely convinced me

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

67
1

But the one thing that I think is fairly clear

2

-- at least the way that most of us go about our business

3

day to day -- is that it is not a pre-standing form of

4

analysis done by reference to a hypothetical monopolist

5

test, where you then plug that answer into the stuff that

6

follows.

7

On the defense side, in the U.S. I have not

8

defined a market in at least 10 years -- I think it's

9

probably 15.

10

side, because that's one of the things you have to do if

11

you don't want to have your complaint thrown out of

12

court.

13

jurisdictions, because that's one of the things you have

14

to do there if you don't want to have your filing thrown

15

out of the agency.

16

I mean, we defined it on the plaintiff's

We define it sometimes in some European

But as an analytical tool in the U.S., we don't

17

use it.

18

don't do it.

19

typically mean by market is that grouping where, at the

20

end of the analysis, a product -- the problem -- is

21

found.

22

sound way to go about the analysis.

23

My sense is that, by and large, the agencies
I mean, what the agencies, I think,

And from my perspective, that is a perfectly

You know, I noted that Jon Baker and Dale

24

Collins are both on the final panel on the third day.

25

And I note that here, in particular, because I was not

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

68
1

aware of that fact when I cited both of them in the paper

2

that I guess has not yet been released or put out front,

3

but will be.

4

emerge from this conference.

I'm sure sooner or later, the papers will

5

I cited to Jon for purposes of his paper on

6

"res ipsa loquitur" market definition, and "res ipsa

7

loquitur" merger analysis.

8

private side all beat on Jon for being lawless -- for

9

reasons I will get into in a minute -- because it is

10

fundamentally lawless.

11

way of thinking about the issue.

And we, of course, in the

But it's actually a pretty sound

12

And I cited to Dale for a three-step analysis

13

that I first saw him use about 12 years ago, and that I

14

and a lot of others have picked up.

15

going through the guidelines type of test, what we

16

usually do when we sit down with clients is pose three

17

short-form questions: Where is the value of the deal;

18

what are the customers going to say; what is going to

19

happen to price?

20

I mean, rather than

And for this purpose, I am using price as a

21

surrogate for all of the other competitive variables that

22

you might get into.

23

might get into innovation, into quantity, into just all

24

of the things you might worry about.

25

And when you poke and probe, you

The first two of those are really background

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

69
1

information.

2

You want to hear whether it's on the price increase side

3

or the cost decrease side.

4

the assets are going to be more profitable in the hands

5

of the buyer than in the hands of the seller.

6

presumably, a deal wouldn't happen.

7

some margin increase somewhere, and the question is

8

what's the source of the margin increase.

9

I mean, where is the value in the deal?

We take it as a given that

Otherwise,

Presumably, there is

And in terms of what customers are going to

10

say, that's partly just to find out what's going to

11

happen if we do get into an affidavit battle, and it's

12

also to poke and probe a little bit about, well, why are

13

the customers going to say that?

14

don't they care?

Why do they care?

Why

15

And it's really that third question, the what's

16

going to happen to price, what's going to happen to other

17

competitive variables.

18

as often as not, you will pose the question 15 different

19

ways.

20

"Well, in which of your products are you most likely to

21

have a price increase that sticks?"

22

will say, "Well, it's this grouping right over here," and

23

then you poke and probe on that.

We want to hear the answer.

And

One of the ways I'm fondest of posing it is,

And sometimes they

24

But I would suggest to you that if you really

25

want to handle the merger efficiently -- if we all want

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

70
1

to handle mergers efficiently -- and I pause a little bit

2

before saying this, but I think we will say it anyway, we

3

have a lot to learn from the doctors.

4

I mean, when you go in and you have a problem,

5

typically what the doctor does is take a bit of the

6

history and formulate some hypotheses, and then you flip

7

into a mode of rule out a problem.

8

medical profession does it is they run a bunch of tests

9

sequentially to rule out this problem, rule out that

10

problem.

11

tests that are consistent with the problem, then you

12

treat the problem.

13

And the way the

And at the end of the process, if you have

And if at the end of the process you don't have

14

a problem, and you can't come up with any other

15

hypotheses, well, then you move on.

16

Now, the issue we all have is how do you

17

translate this into legal doctrine in a system where you

18

have case law that has very, very strong authority on two

19

propositions?

20

First, you have to define a market.

And

21

second, the market definition is step one of the

22

analysis.

23

paper that will eventually be forthcoming.

24

are several approaches, none terribly satisfactory, that

25

one might come up with to try to reconcile case law with

And I muse a little bit about that in the

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

I mean, there

71
1

the mode of analysis that's actually practiced.

2

And it seems to me that if we were going to try

3

to take the next set of steps and conform guidelines to

4

practice as we know it, that is the fundamental problem.

5

I mean, how do we skin the cat of not getting thrown out

6

of court?

7

Agency's problem, but how would the Agency skin the cat

8

of not getting thrown out of court?

9

Well, actually, it's not my problem, it's the

But let me leave you with this thought.

I'm

10

going to close with this.

11

and this is where, again, I join up with Will, or maybe

12

against Will, in saying the guidelines are working well.

13

It does seem to me that the guidelines are affirmatively

14

misleading as a recipe for how analysis is actually

15

conducted, you know, 12 years after their most recent

16

revision.

17

First, it does seem to me --

Second, the business community is fundamentally

18

confused about how you do the analysis.

I mean, it's

19

pretty common that we will sit down with a client for the

20

first time and the client will have worked through the

21

guidelines.

22

analysis is.

And you can't believe how far off the

23

Now, the 68 guidelines went 14 years before

24

they were revised, and we're obviously not at 14 years

25

yet.

And I would not lightly call for the Agencies to

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

72
1

undertake a process of revising the guidelines, because I

2

know how complex that is and how burdensome that can be.

3

So I'm not going to say that.

4

But I will say that the time probably should

5

come one of these years -- almost certainly not before

6

the election, but maybe 2005, 2006 -- it probably would

7

be productive for people to sort of sit back and say,

8

"What do we mean by market, and how, really, are we going

9

to try to operationalize what we're doing, if what we're

10

going to seek in the way of guidelines is a pretty

11

transparent set of recipes that depict the enforcement

12

practices to the public?"

13

And with that, any questions, Mr. Moderator?

14

MR. WERDEN:

I have a few.

First, I'm actually

15

quite confused about what you mean when you say that the

16

guidelines are misleading on how the analysis is

17

conducted.

18

analysis is supposed to be conducted, so I don't know how

19

they could be misleading.

20

have in mind?

21

I'm not sure that they tell anybody how the

MR. BLUMENTHAL:

What is it, exactly, that you

Well, I guess I'm using

22

misleading in the FTC Bureau of Consumer Protection

23

sense.

24

(Laughter.)

25

MR. BLUMENTHAL:

Which basically means that you

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

73
1

have a population of readers that are interpreting the

2

words in a way that is not entirely consistent with what

3

the words literally are intended to convey, that you have

4

-- and by the way, I mean, it's not just the public.

5

seems to me that the same thing applies to certainly

6

agency staffers overseas who try to apply our guidelines,

7

and I think I would go on to say some agency staffers

8

within our borders who try to apply our guidelines.

9

It

The iterative aspect of the analysis is

10

something that is often fundamentally missed.

11

that market shares are not what you see in the documents,

12

but rather what you calculate after going through all of

13

the rest of the steps preceding competitive analysis,

14

that's something that is quite commonly missed.

15

The fact

You know, Will's point on unilateral effects, I

16

think, is pretty fundamental.

17

misleading, that people -- large numbers of people -- are

18

simply misconstruing how you're supposed to do the

19

analysis.

20

comes in.

21

So that's what I mean by

And in particular, where market definition

MR. WERDEN:

Well, in that regard, let me

22

suggest that the guidelines disclaim being a step-by-step

23

instruction manual, and perhaps best should be described

24

as an outline for how one ought to present the results of

25

the competitive analysis of a merger.

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

74
1

And if viewed in that light, is there anything

2

misleading about the guidelines?

3

problematic about the way they choose to organize these

4

issues in a manner, as you commented in some of your

5

articles 20 years ago, fairly consistent with the case

6

law that constrains how the government can litigate its

7

cases?

8

MR. BLUMENTHAL:

Yes.

Is there anything

I think with that

9

qualification, the guidelines remain reasonably accurate.

10

I'm not sure about a separate 35 percent test for

11

unilateral effects -- which I think is still in there --

12

well, depends on how you define the market, I suppose,

13

which is sort of the point.

14

But I'm not sure that that component is

15

something that withstands challenge.

16

minor quibble.

17

the disclaimer says they are not a recipe, and that the

18

public -- if you're going to say the public shouldn't use

19

them as a recipe, then they hold up pretty well.

20

problem, of course, is people do use them as a recipe

21

because they don't have any other recipes.

22

But that's a really

And yes, subject to the observation that

MR. WERDEN:

Okay.

The

Thanks very much, Bill.

I

23

have a large number of prepared questions.

24

launch into them, I want to offer the panelists an

25

opportunity to respond to what else they have heard thus

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

But before I

75
1

far.

2

MR. HARRIS:

I actually have a question of you.

3

I think it was you asking questions about when there were

4

switching costs and issues with exploiting the installed

5

base, and at least it wasn't clear to me exactly what you

6

were asking, so I ask that you kind of elaborate on it

7

and explain better the situation that you have in mind

8

where a merger might harm competition.

9

MR. WERDEN:

Well, sure.

I had in mind a very

10

simple-minded situation in which every current user is

11

locked in to his supplier, for whatever reason.

12

ask me to explain, I think it will be hard to explain

13

that.

14

Don't

But suppose it is so.
The hypothetical monopolist could exploit these

15

guys, because they're locked in.

16

would be, well, what about new customers?

17

where the competition is, of course, to attract the new

18

customers and to sign them up and to get them locked in.

19

And a hypothetical monopolist might be able to

20

discriminate between the two, and might not.

21

And then the issue
That's usually

And one would have a somewhat different

22

analysis of the merger, depending on which of those

23

circumstances you thought you were in.

24

possibility is that you can't discriminate, so all a

25

hypothetical monopolist could do is exploit the guys who

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

But one clear

76
1

are already locked in.

2

And my question is, well okay, in that event is

3

there any way that you would specially account for

4

switching costs in the application of the 5 to 10 percent

5

price increase test, and my answer would be no.

6

don't do anything different or special in that case, not

7

in delineating the market.

8
9

MR. HARRIS:

You

But you would not ignore it in

looking at competitive effects, is that accurate?

10

MR. WERDEN:

You would not ignore anything in

11

looking at competitive effects.

12

MR. HARRIS: I understand.

13

MR. WERDEN:

But part of what the guidelines

14

are trying to tell you is what category various issues

15

fall into.

16

are market delineation and some aren't."

17

of them aren't.

18

And the guidelines say, "Well, some issues

MR. HARRIS:

Okay.

And a whole lot

And then I guess the

19

follow-up question I have in the context of a merger, if

20

these customers are locked in, is your suggestion that

21

there is a way in which the merger can make that

22

situation worse?

23

And if so, what is that?

MR. WERDEN:

Oh, that's an excellent question.

24

It isn't necessarily so that the merger can make the

25

situation worse, and that's where the competitive effects

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

77
1

analysis ought to be.

2

get to the competitive effects analysis you conclude that

3

the locked-in customers are exploited as much as locked-

4

in customers can be exploited, and the merger doesn't

5

affect that, and the focus ought to be on the new

6

customers, because that's where all the competitive

7

action is.

8

my view.

9

It may very well be that when you

I have certainly seen mergers where that was

But you can imagine, certainly, that the

10

switching costs between incumbent suppliers are much

11

smaller than the switching costs between an incumbent

12

supplier and somebody outside the candidate market, in

13

which case it may very well be that there is some effect

14

on the degree of exploitation of the locked-in customers,

15

because they're not completely locked in.

16

Anybody else have any comments or questions?

17

No?

18

for almost entirely praise from our panelists today.

19

my question is has it completely won the intellectual and

20

legal battle -- and those are two different battles --

21

and I think the answers may be different.

22

Okay.

The hypothetical monopolist test has come in
And

Put another way, is there an alternative that

23

makes any sense to use?

And I think this question has to

24

be posed in two different ways.

25

perspective, and the second one, a legal one.

One, an economic/policy

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

And with

78
1

respect to the economic perspective, I pose the question

2

to Barry.

Does economics offer any alternative?

3

MR. HARRIS:

4

disclaimer, "Never say never."

5

with a few exceptions, to think of something that

6

performs better, and it gets back to what question are

7

you asking:

8

they were able to coordinate, could exercise market

9

power?

10

Well, the first thing would be my
But I'm hard pressed,

What is the group of competitors, that if

Now, it seems to me –- alternatives that -- at

11

least the ones I have seen -- are inappropriate.

12

haven't seen them all.

13

to the same question asked in a different way.

14

Maybe I

Or, alternatively, they default

And just as an example, I think it was John, in

15

his talk, he mentioned successful price fixes, and he

16

also mentioned the Staples case.

17

fair that if you can look at real-world information, it

18

may tell you what the market is.

19

successful price fix?

20

all the conditions are reasonably similar, what you have

21

done is you have done the hypothetical monopolist test,

22

you have just run it out in the real world.

23

Well, I think it is

But what is a

At the end of the day, assuming

Same in Staples.

My understanding of that case

24

-- which is imperfect; I didn't work on it -- but despite

25

all the talk about econometrics, at the end of the day,

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

79
1

if you read the decision, you will see the judge kind of

2

ignored everybody's econometrics.

3

an important fact of that case was that the two companies

4

each had pricing zones that were defined by the existence

5

of the other company, and these were the low price

6

pricing zones.

7

But my understanding,

Well, again, you can say that's real-world

8

information, and it's a different way of going about it.

9

But at the end of the day, it's telling you an answer to

10

the hypothetical monopolist question.

11

different paradigms, but I'm not aware of them.

12

ones I am aware of seem to me to be just a round-about

13

way of getting to the same point.

14

MR. WERDEN:

Okay.

So there may be

Thanks, Barry.

And the

Bill, I

15

will ask the legal side of the question to you.

In your

16

remarks -- I paraphrase -- you said -- I didn't get the

17

exact remarks -- you said the hypothetical monopolist

18

test has "won."

19

opinions that cite the Brown Shoe practical indicia.

20

Is it your view that the courts are really

But yet we still read district court

21

relying on the practical indicia, or are they ultimately

22

really relying on the hypothetical monopolist test, or

23

some interesting combination of the two?

24
25

MR. BLUMENTHAL:

Well, when I say "won," I mean

primarily in the hearts and minds of commentators,

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

80
1
2

analysts, and the antitrust cognoscenti.
With respect to the courts, I think,

3

increasingly, the test is the decisive test.

4

still are some courts that are laggards.

5

question, in large part, turns on how economically

6

sophisticated the judge is.

7

part, on how facile the judge is on reconciling the legal

8

standard which perhaps she believes makes sense with what

9

she thinks she has to put down to avoid getting reversed

10

by the court above.

11

But there

I think the

I think it turns, in large

So, as an evidentiary matter, the courts

12

continue to rely pretty heavily on Brown Shoe.

13

practical indicia, of course, are a submarket test.

14

as Brown Shoe says, a submarket can, itself, be a market.

15

So it all gets completely commingled.

16

And the
But

And I think that for a long time to come we

17

will continue to see courts citing to the tests that have

18

been articulated by other courts, it's just that we're

19

beginning to see more and more courts speaking favorably

20

about the guidelines, about the hypothetical monopolist

21

test, and that then creates a body of precedent where,

22

sooner or later, courts will be able to rely just on that

23

without necessarily linking it back.

24

MR. WERDEN:

Let's get a little more specific,

25

and talk about the Staples case.

There is, you know, at

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

81
1

least three or four pages in the Staples case relating to

2

the practical indicia.

3

MR. BLUMENTHAL:

4

MR. WERDEN:

Yes.

Is there doubt in anybody's mind

5

here that the only fact -- and it may be more than one

6

fact, but I will call it "a fact" -- that ultimately

7

persuaded the reasonably skeptical judge about this

8

funny-looking market was that there was very clear

9

pricing evidence that the number of office superstores

10

really was affecting prices?

11

MR. HARKRIDER:

12

MR. WERDEN:

13

MR. HARKRIDER:

Greg, if I can make a point?

Sure.
I think that Staples actually

14

provides an illustration of the relationship between

15

Brown Shoe and the hypothetical monopolist test.

16

that the hypothetical monopolist test is frequently used

17

as an economic test as opposed to Brown Shoe, where there

18

is quantitative evidence of the sort, or very clear

19

evidence of the sort that directly bears on the

20

hypothetical monopolist test.

21

I think

I think where there is very little evidence as

22

to whether a hypothetical monopolist could, in fact,

23

impose a SSNIP – which -- and by that I mean there is no

24

econometric evidence, there is no natural experiment,

25

there are no surveys, there are no affidavit evidence --

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

82
1

I think in that context they frequently rely upon Brown

2

Shoe.

3

weak or not determinative, then they rely upon Brown

4

Shoe.

5

And if the quantitative evidence is relatively

I think Staples is a case where you have

6

essentially, a natural experiment that allowed you to get

7

direct evidence and get some purchase on the hypothetical

8

monopolist test.

9

instructive.

10

And so, Brown Shoe wasn't particularly

MR. WERDEN:

Okay.

Well, let's try a different

11

case.

12

evidence in the case, but the court was totally

13

unpersuaded by it on both sides.

14

Brown Shoe factors.

Let's try Swedish Match.

There was quantitative

And the court cited

15

But the way I read the decision, ultimately the

16

judge sat back and looked at the totality of the evidence

17

and said, "Well, if I think of a hypothetical monopolist

18

raised the price five percent, he would make more money."

19

So it's a market.

20

MR. HARKRIDER:

I think that that's right, but

21

I think that there is uncertainty when you read the

22

cases, whether the Brown Shoe factors are something that

23

gets you an ability to analyze whether a hypothetical

24

monopolist could, in fact, increase the price, or whether

25

the hypothetical monopolist test is, in fact, to some

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

83
1

extent, one of the Brown Shoe factors.

2

And you really see courts -- and I think this

3

is one of the problems, is that the courts really haven't

4

figured out where the hypothetical monopolist tests

5

relate to Brown Shoe.

6

I think that once you get outside of the merger

7

context, you see that a great number of courts, whether

8

in a Section 1 or a Section 2 context, where a definition

9

of the relevant market is important, are not in fact

10

looking at the hypothetical monopolist test, although it

11

clearly is an increasing trend to do so.

12

MR. WERDEN:

All right.

Let's try another

13

topic.

An early criticism -- and not necessarily just an

14

early criticism -- of the 1982 merger guidelines approach

15

to market delineation was that it would systematically

16

bias the analysis –- yielding overly broad markets.

17

hence, understating market shares.

And

18

I didn't think anybody believed that any more

19

until a week ago, when the American Antitrust Institute

20

released a statement in conjunction with this workshop,

21

which they officially filed with the Agencies as their

22

comments, which asserts "that there are common market

23

definition procedures" in the guidelines that "create the

24

potential for systematic errors in defining markets."

25

They don't explain.

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

84
1

But my question for the panel -- and I will

2

start with Will -- do you believe that the guidelines'

3

articulation of the hypothetical monopolist test gives

4

rise to the sort of systematic errors that people were

5

worried about back in the mid-1980s?

6

MR. TOM:

Certainly to the extent that people

7

were worried about a systematic bias toward overly broad

8

markets, I think probably most perceptions of the time --

9

and certainly subsequent experience -- has shown that to

10

be completely unfounded.

11

I have suggested a reason in my opening remarks

12

why we might have a systematic bias toward overly narrow

13

markets.

14

the guidelines, but simply because of some perhaps

15

unfortunate interaction between the guidelines and human

16

nature.

17

equation or one hypothetical part of the analysis as fact

18

and another hypothetical part of the analysis as

19

speculative.

20
21
22

Not because of anything analytically wrong in

That is the tendency to treat one side of the

But other than that I don't see any particular
systematic biases at all.
MR. WERDEN:

Can any of the panelists think of

23

any matter they worked on, or one they didn't work on,

24

where the guidelines, as opposed to some view of the

25

facts, led to a market that they thought was overly

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

85
1

broad?

2

MR. HARRIS:

I'm not sure that I can answer

3

that, but a curious fact that I have raised with Greg in

4

a different context was -- must be 10 years now -- in the

5

Dubuque hospital case, there was an oddity there, and

6

that was that the lesser of the two hospitals in the

7

period before the merger had lowered their prices by

8

roughly 40 percent in an effort to get more patients and

9

get more managed care plans.

10

virtually nothing.

11

And they had gotten

And for some reason, they left their prices

12

down.

13

said, oh, they could raise their prices 40 percent, get

14

back where they were, and presume the elasticity

15

symmetric, and they didn't do that.

16

They had an experiment sitting out there that

So, that sort of raises the question -- and

17

it's not exactly your point -- but how low does the

18

market have to be in a situation like that to include the

19

merging parties?

20

So it's sort of a related question.

An attempt to answer it directly, the only

21

thing that comes to mind -- and I didn't look into this,

22

and I may have the facts wrong -- but I have a memory of

23

the antacid case that had very, very different looking

24

competitors in there.

25

a very broad market in which a case was brought on that

And that might be a candidate for

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

86
1

may or may not have been supported.

2

one that comes to mind.

3

But it's at least

MR. WERDEN:

We referred to that one as stomach

5

MR. HARRIS:

Okay.

6

MR. WERDEN:

Because it included a lot more

4

7

remedies.

than antacids.

8

MR. HARRIS:

Yes.

And again, I don't know the

9

process that got to the market, and I don't know very

10

much about the market, except that my doctor tells me to

11

take two every day for the calcium.

12

a lot of people at the time, and it certainly struck me

13

at the time as being a market that was way too broad.

14

And that was based on just kind of being a consumer,

15

rather than any antitrust analysis.

16
17
18

I think that struck

But other than that, I am hard-pressed to think
of one.
MR. BLUMENTHAL:

Well, I don't know that it's

19

possible for the guidelines, if applied literally, to

20

yield an overly broad market.

21

what we mean by overly broad.

It depends a little bit on

22

But if the breadth from which you were either

23

too broad or too narrow is what you would otherwise get

24

by reference to the standards in the courts, the non-

25

guideline standards, it seems to me, as it does to Will,

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

87
1

that there is almost a chronic bias in the direction of

2

narrow markets.

3

Whether that's appropriate or not is something

4

we can talk about, but for 20 years I have been saying

5

that a 5 percent one-year test is going to lead to

6

markets that are a lot narrower than what business people

7

conventionally think of as being the market.

8

reason for that is that business people tend to take more

9

of a strategic perspective on things.

10

a whole lot of shifting for five percent one year.

11
12
13

MR. WERDEN:

Of course, it's not one year in

MR. BLUMENTHAL:
enough, fair enough.

15

MR. WERDEN:

17

And you don't see

the 1992 guidelines, it's for the foreseeable future.

14

16

And the

Yes, foreseeable future.

Okay.

Fair

Do you have any theory for

why people like Bob Pitofsky thought what they thought?
MR. BLUMENTHAL:

You know, I was not among

18

those who thought it at the time, and I didn't understand

19

it at the time.

20

direction at the time.

21

And I went in writing in the other

Recalling the political climate in 1982, when

22

the guidelines were released, you know, as much as

23

anything else it seems to me sort of a knee-jerk reaction

24

to a perception that this was what some would have called

25

Reaganism run wild.

For The Record, Inc.
Waldorf, Maryland
(301)870-8025

88
1

And that, from my perspective, at the time was

2

what was motivating it.

3

after the fact, that wasn't what was motivating it.

4

That's the best I can come up with.

5

MR. WERDEN:

And after the fact, 20 years

Okay, thanks.

Next topic.

A

6

long-standing issue in the application of a hypothetical

7

monopolist test is how to account for pre-merger

8

elevation of price above cost.

9

that we generally use the prevailing price, but there are

10

provisions in there referring to possibility of pre-

11

merger coord

[Text truncated at 120,000 characters. The full text is on the page linked above.]

---

Source: Frix Law Library, https://www.frixlaw.com/law-library/documents/agency%3Aftc%3A4560ab493ca1cc50. Public record. Not legal advice.