FEDERAL TRADE COMMISSION

Agency decision

Ask Donna

What actually matters in this document.

Text

DEPARTMENT OF JUSTICE

JOINT WORKSHOP ON MERGER ENFORCEMENT

Tuesday, February 17, 2004

9:00 a.m.

Federal Trade Commission

Conference Center

601 New Jersey Avenue, N.W.

Washington, D.C.

Reported by:

Rita M. Hemphill, CVR

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

C O N T E N T S

PAGE

Opening Remarks

Hypothetical Monopolist Test

Concentration & Market Shares

114

Monopsony

207

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

P R O C E E D I N G S

OPENING REMARKS

MR. ABBOTT:

Good morning, everyone.

Welcome

to the Joint Federal Trade Commission Department of

Justice Antitrust Division Workshop on Merger

Enforcement.

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

enjoy the facilities -- for today, tomorrow, and

Thursday.

This workshop will be held here today in

We welcome members of the audience to listen

in, although all questions during the proceedings will be

handled by members of the panels.

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

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

both the Federal Trade Commission and the Antitrust

Division, and certainly Bob Potter, head of the policy

office of the Antitrust Division, for tremendous

collaboration and support in putting on this event, which

we hope will clarify a number of issues regarding the

current application of the merger guidelines.

And we are delighted to start things out with a

bang with the leaders of the federal antitrust community.

We will lead off with brief remarks by Assistant Attorney

General Hew Pate, and follow up with remarks by FTC

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

Chairman Tim Muris.

And both General Pate and Chairman Muris have

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

away from their important time at the podium by going

into them.

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

800 telephone line to this important conference.

Suffice it to say that they are welcoming

And we hope to obtain some valuable information

which will inform our activities in the future, as a

result of the conference.

now to General Pate.

MR. PATE:

So, let me turn it over right

Thanks very much, Alden.

Thanks to

the Federal Trade Commission for providing this great

facility to host our look into the operation of the

merger guidelines, and to consider improvements -- we

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

way that the agencies carry out their responsibilities

under Section 7 of the Clayton Act and under the FTC

statute.

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

part of a very significant amount of policy activity

that's going on at the two agencies.

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

for that, in terms of bringing forward the concept of

joint work between the two agencies.

And frankly, I

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

That has been producing a great deal of

thinking, in terms of the IP and antitrust hearings,

which are now drawing toward a final close, hopefully

with the publication soon of the antitrust and IP report.

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

the context of the hearings that were held there, and

this merger conference is yet another example.

Both of the agencies, likewise, have been

active individually on the policy front.

I think of the

report on patent practices and procedures, that the Trade

Commission has issued, which is a very important

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

Division will be heading up a conference looking at

pricing practices.

And I think all of this is really an important

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

my colleagues at the Federal Trade Commission, starting

with Tim, for putting in place the circumstances that

have made for a very good relationship there.

Obviously, at the Division, merger enforcement

is a key part of our mission.

Depending on the activity

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

two-thirds of our effort goes into merger enforcement.

As I have mentioned elsewhere, merger enforcement is

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

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

antitrust world, certainly different from cartel

enforcement, where there is little or no serious

controversy about the need to aggressively intervene

where cartel activity is detected.

It is different, too, from the unilateral

conduct sphere, where it's very important for the

agencies to have a very high degree of caution about

where they intervene, because of the possibility of

chilling competitive conduct in the guise of combating

anti-competitive conduct.

Obviously, it's a gross over-simplification,

but I suggest that merger enforcement falls somewhere in

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

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

discussion about what the best use of the tool set that

we have to evaluate mergers.

Obviously, merger enforcement requires us to

make predictive judgments.

Section 7 is described often

as an incipiency statute.

That's obviously true, but

that term can carry different, very freighted meanings,

depending on who the speaker might be.

And what we're about here is to try to

determine how best we can make predictive judgments about

proposed transactions to make sure that those that would

bring efficiency and better outcomes for consumers are

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

allowed to go forward.

And likewise, that those that

have the real prospect for substantial lessening of

competition don't happen.

Obviously, the Horizontal Merger Guidelines are

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

doing that.

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

revised with some significant language changes, and some

significant changes in approach.

The merger guidelines of 1982, which have

Certainly the agencies, even apart from those

textual changes, have, throughout this period,

incorporated advances in economic thinking and

improvements in our own practices and learning.

The guidelines are flexible, they are

practical.

that both agencies are in a period of trying to emphasize

transparency.

that those in the bar who are very experienced repeat

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

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

transparent as possible to the public generally.

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

incidents of the issuance of closing statements at both

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

public, when we decide that government intervention is

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

not warranted.

And I think you can see that in terms of the

release of data that has preceded this conference.

agencies jointly, as you know, have released data on

merger challenges from 1999 to 2003.

The

The Federal Trade Commission has now released

significant data on cases in which a challenge was not

brought, and also issued some data that shed light on

other factors, customer complaints, hot documents, other

things that the agencies looked at in evaluating a

proposed merger.

set the stage for some very interesting discussion and

improved learning in this conference.

And we hope very much that this will

Transparency, as I have said, is important, not

only for the repeat players in the merger bar who are

likely to be most interested in this, but for other

enforcers, both here and internationally.

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

sound way, but explaining what sound enforcement entails.

I think that

Whether we like it or not, the two federal

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

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

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

discussion of what appropriate enforcement criteria are.

And I hope this conference will contribute to that.

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

We certainly have a wide range of topics and a

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

anyone who is involved in merger policy.

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

the next three days, it would break.

If we had an

Fortunately, we have got some very good

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

great deal of time into preparing to keep these panels

under control and hope to get some insight from them.

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

here?"

conference has been called because anyone thinks

something is broken.

conference has been put together because there is some

pre-ordained goal of producing any particular output from

the conference.

really in a position to answer until we get the benefit

of hearing what reactions and thinking, based on the data

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

have done independently over the past several years.

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

and what might come next.

But just as with the process

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

improve our procedures, this conference does demonstrate

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

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

keeping up with the best thinking on substance.

So, with that, I will conclude simply with a

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

a great deal of time into organizing these panels.

look forward to participating in some more of this,

including a round table Thursday afternoon, and will keep

things moving now by introducing my good friend, Tim

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

conference.

Thanks very much for being here.

(Applause.)

MR. MURIS:

certainly a pleasure to be here.

interesting three days.

Justice are, obviously, our most important partners and

colleagues, and I greatly appreciate their help and work

on their effort.

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

the Justice Department staff.

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

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

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

The data we released,

I particularly wanted to thank our economist

who worked on that data.

It was a lot of work.

few of them out there in the audience.

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

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

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

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

practitioners, academic and enforcement officials here

over the next few days.

I won't repeat the mistake I made about a

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

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

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

$500-an-hour economists every assembled."

long time ago when billing rates were lower.

them took great offense, and it was immediately obvious

to me and everyone else that this person billed

considerably higher than $500 an hour.

(Laughter.)

MR. MURIS:

notes.

This was a

And one of

And the others were all making

And I felt like a facilitating practice, so --

(Laughter.)

MR. MURIS:

As Hew mentioned -- you know, well,

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

the impact of the guides.

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

the states and internationally.

And as Hew mentioned, they

There are over 60 countries that have merger

control regimes, and you can see the influence on the

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

guidelines almost everywhere.

Let me tell you what my view of what we can

achieve over the next three days.

think we all know, are not a cookbook.

provide specific details on every aspect of a merger

investigation.

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

competitive.

explore state of the art application of the guidelines by

those with the most experience at using them.

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

website so that people can refer to it.

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

Justice -- listening on their computers.

people listening on their 800 number.

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

and other people can refer to it as well.

I know there are

There are other

We are actually

As Hew mentioned, we at the FTC have also

emphasized transparency over the 32 months that I have

been at the Commission.

several cases in which we did not sue, including three

cases very recently.

We have released statements in

We have released the two data sets that Hew

mentioned.

The first one contained market share and

concentration levels associated with the FTC and DOJ's

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

decisions to challenge mergers in a wide range of product

markets.

on 151 horizontal merger investigations from the

beginning of fiscal year 1996 through the end of fiscal

year 2003.

And then a few weeks ago, the FTC released data

Unlike the December data, this release included

data when no enforcement action was taken, as well as

data on additional key facts.

involved the numbers of significant competitors.

practitioners I know think and talk in these terms,

rather than concentration.

One statistic we released

Many

These data also look at enforcement decisions,

depending on whether hot documents exist or strong and

credible customer complaints are received.

should become core information in a healthy debate about

the level and direction of merger policy.

These data

I note that our recent data released largely

reflects cases in which I did not participate.

participated in cases involving only about 10 percent of

the markets analyzed.

Indeed, I

As I have said elsewhere, current merger

practice reflects a bipartisan approach.

of this and similar data in the future should help us

understand merger enforcement in practice.

And the release

The data we released highlights several

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

important issues in merger analysis.

long-standing debate about the significance of

concentration and HHI numbers.

released and the breadth of analysis we will hear this

week will finally put to rest the notion that HHI levels

have any special significance, except at very high

levels.

One involves the

I hope the data we

Instead, the agencies try to answer the

ultimate question.

We consider several variables that have an impact on the

likely level of competition in post-merger markets.

Fealty to the original guideline numerical levels was

abandoned as the agencies gained experience.

the guidelines were amended to codify the existing

practice of giving great weight to qualitative factors.

Will the merger impair competition?

In 1992,

The 1982 and 1984 guidelines had given more

emphasis to quantitative thresholds, particularly

involving HHI levels above 1800.

and Bill Baxter discussing what the appropriate levels

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

conversations that Bill retained some of his views that

he had expressed in the 1960s about the strong

presumptions one could gather from concentration.

I remember Jim Miller

In any event, the 1992 guidelines reduce the

significance of the 1800 threshold by inviting fuller

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

consideration of other conditions that help predict

whether price increases are likely, post-merger.

Thus, the pre-eminence that some would continue

to give to concentration or HHI numbers is misplaced.

State of the art merger analysis has moved well beyond a

simplistic causality of high concentration leading to

anti-competitive effects.

The number of competitors is certainly

important.

Four to three gets our attention quicker than

six to five.

elevate a single fact or number to dispositive

significance.

an increased likelihood of anti-competitive effects

before we act.

But current merger practice does not

The totality of the evidence must point to

Much of this experience with merger

investigations is captured in the guidelines themselves.

One of the salutary effects that the guidelines have is

the transparency they help bring to government.

help us clarify enforcement policy and doctrine so that

practitioners and their clients can make better

judgments.

They

Government officials should explain the basis

on which they exercise their authority.

expect clear and consistent enforcement actions.

course, application of the guidelines is not always

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

Stakeholders can

obvious.

Hence, the high billing rates.

We constantly strive to bring more transparency

to our merger process, and we hope this workshop will

result in a better understanding of current merger

policy.

transparency for consumers in the business community is

the feedback this workshop should provide for the

agencies.

Equally important with providing increased

We want to obtain important information that

will assist us in doing our jobs.

learn from you over the next three days.

from the most experienced practitioners who work with the

guidelines every day, as well as academics doing state of

the art research.

guidelines are working, what you perceive to be their

strengths and weaknesses, what are the issues in which

there is agreement and disagreement, and what areas you

consider important for further study.

Research and clarification.

guidelines must respond to new legal and economic

analysis.

at this workshop.

started.

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

HYPOTHETICAL MONOPOLIST TEST

MR. WERDEN:

Good morning.

I'm Greg Werden,

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

merger enforcement workshop, which focuses on the

hypothetical monopolist paradigm for market delineation.

The hypothetical monopolist paradigm was

elaborated, and certainly popularized, by the 1982 merger

guidelines, but the basic idea was not original to the

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

late 1970s, some of the Antitrust Division were already

applying the test, and there is at least one documented

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

that.

In the

And I have uncovered antecedents much earlier

than the 1970s.

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

it shorter, was written by Morris Adelman in 1953.

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

substitutes exist at what price for the product or

service in question.

concern were the only occupant of the allegedly separated

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

answer is yes, then the separate market exists within

which competition can be lessened."

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

monopolist test, and I believe Adelman probably was the

first one to have it.

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

year.

His 1956 article has been quoted

The 1982 merger guidelines did a lot to develop

this basic idea into something quite useful for market

delineation.

1982 fine tuned that approach.

to draw on experience of 20 years plus in applying the

hypothetical monopolist test, and offer insights on

possibilities for refining its application further.

And revisions of the guidelines in 1984 and

Our panel today is going

I won't waste any time on introductions of our

speakers, I will just turn things over to our first

speaker, John Harkrider.

MR. HARKRIDER:

Thank you, Greg.

The

hypothetical monopolist test is one of the organizing

principles of the horizontal merger guidelines.

is a test that is increasingly applied to define markets,

not just in merger cases, but throughout antitrust.

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

And it

And

But the hypothetical monopolist test is not

"flawless."

Indeed, three years after the 1982

guidelines were published, a leading IO economist

commented that, "The guidelines market definition test

has one wholly decisive defect.

It is completely non-

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

operational."

(Laughter.)

MR. HARKRIDER:

Well, history has proved that

that concern is overstated, as there clearly are data and

quantitative methods that will operationalize the

hypothetical monopolist test.

It is important to note that using reliable

data is critically important.

Because without such data,

we quite frequently resort to the Brown Shoe factors that

the hypothetical monopolist test was meant to supersede,

or at least clarify.

My remarks address how to render the

hypothetical monopolist test operational by using sound,

empirical methods that answer the question posed by the

guidelines market definition test.

At the outset, it's important to realize that

the hypothetical monopolist test asks two subsidiary

questions.

make a SSNIP unprofitable?

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

Much of the economic literature has focused on the first

question.

critical loss attempts to answer.

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

answered.

agency or a court, may have confidence that a

hypothetical monopolist could not profitably impose a

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

that fact-finder may have no confidence that the

hypothetical monopolist would lose more or less than X

percent of its sales.

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

But

The purpose of these remarks is to provide some

guidance towards answering the second question:

What

volume of sales would the hypothetical monopolist lose if

it were to impose a SSNIP.

four forms of evidence:

evidence; affidavit evidence; and survey evidence.

To that end, we will discuss

historical evidence; econometric

The merger guidelines expressly authorize the

use of historical data, although they have cautioned it

may produce misleading results.

often reflect changes in cost affecting the whole

industry, rather than market power.

reactions to such changes do not illuminate how they

would respond to a hypothetical monopolist.

Past price changes will

And consumers'

Moreover, price changes typically are of short

duration, and thus do not satisfy the non-transitory

requirement of the SSNIP test.

guidelines themselves caution that the picture of

competitive conditions that develops from the use of

For that reason, the

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

historical evidence may provide an incomplete answer to

the forward-looking inquiry of the guidelines.

When considering historical evidence, it is

important to distinguish between two types of evidence.

First, evidence that consumers switched in response to

past price changes, and second, evidence that consumers

did not.

If a significant number of consumers switch

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

significant price change in X relative to Y, this seems

to be evidence that the two belong in the same relevant

market.

The criticism that the price change on X may be

cost justified or not sufficiently long in duration would

simply suggest that even more consumers may switch from X

to Y if the price increase were not cost justified,

imposed by monopolists, or of longer duration.

Alternatively, evidence that customers did not

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

significant price change in X relative to Y, may have

little bearing on whether the products belong in the same

relevant market, as customers may be more likely to

switch if the price increase were not cost justified,

imposed by a monopolist, or of longer duration.

Another form of historical evidence are natural

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

experiments.

In an extreme case, if the firm is selling

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

engaged in cartel activity, there should be little reason

to doubt that the market is properly defined.

One court wrote that, "Every price-fixing

conspiracy thus identifies directly, in a real-world

context, a group of firms which is insulated from outside

competitive pressure."

conventional market definition evidence attempts to

identify artificially.

This is precisely what

Another example of a natural experiment is the

court's analysis in Staples.

various price studies which, on the whole, the court

found persuasive, tended to show that a hypothetical

office superstore monopolist could impose a SSNIP on the

sale of consumable office supplies, because when it was

an actual office superstore monopolist, that is precisely

what it did.

In that case, the FTC's

A form of historical evidence is econometric

evidence.

in a controlled and scientific manner to answer the

hypothetical monopolist test.

Econometric evidence uses historical evidence

In fact, Judge Easterbrook, just a few weeks

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

for failing to introduce econometric evidence of any kind

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

in defining the relevant market.

There are many types of econometric analysis

that can be used to answer the hypothetical monopolist

test.

effect, allows for a controlled natural experiment.

type of analysis requires geographic variation as to the

number of competitors, as well as information on pricing,

number of competitors, and factors that may influence the

price of goods.

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

This

A second type of analysis directly estimates

market elasticities.

To be done properly, this type of

analysis requires time series consumer-level information

on price and quantity.

that in appropriate circumstances, econometric estimates

of elasticities can provide perhaps the best evidence as

to market definition, there are significant issues with

respect to its use.

While there is general agreement

One issue involves the ability of data.

quite unusual for private parties to have access to price

and quantity information from all market participants.

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

but it raises important issues about transparency, as the

government may be unable to share this data typically

collected through the CID process with the merging

parties.

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

It is

A second issue involves the quality of the

data.

For example, retail scanner data, which is one of

the few forms of time series data that is available,

frequently does not give sufficient information on

coupons or other discounts.

to opine on whether that level of competition is

important.

And so your economist needs

A third issue involves the appropriateness of

the data.

Retail scanner data, for example, may not be

the appropriate type of data to analyze mergers between

manufacturers.

If you want to know what somebody would do in a

hypothetical situation, one alternative simply is to ask

them.

guideline’s 5 to 10 percent price increase formulation of

the SSNIP test.

concrete question to ask customers in interviews and

depositions.

customer affidavits and merger litigation.

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

respect to the use of customer affidavits.

Turner go so far to suggest that customer affidavits are

"the least reliable evidence of whether consumers would

switch in response to a SSNIP."

Areeda and

Despite this criticism, the government and

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

litigants in merger cases frequently use affidavits.

review of cases involving the use of affidavits revealed

two frequent errors.

the presence of switching costs.

frequently do not cover a representative sample of

consumers.

First, affidavits frequently ignore

And second, affidavits

As to the first question, if consumers will not

switch between the products being grouped together for

the purposes of applying the hypothetical monopolist

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

switch to a product outside of the proposed market

following a SSNIP.

argued that current GQA customers would not switch to

alternatives in response to a SSNIP in GQA was evidence

that there existed a relevant market for GQA.

For example, in Englehard, the DOJ

However, the 11th Circuit rejected the DOJ's

argument in part because Englehard's GQA customers would

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

in the price of Englehard's GQA.

The guidelines acknowledge a variation of this

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

account for situations where the relevant product

constitutes a small percentage of the overall cost of the

goods or service.

Affidavits also must cover an adequate sample

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

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

in the context of critical loss.

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

to a SSNIP, a hypothetical monopolist would find it

unprofitable if it lost 8 percent of its sales.

Consider, for example,

Then, imagine the government goes out and

interviews 100 customers, randomly selected -- which, it

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

they interview 100 random customers.

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

them say that they would.

customers purchase an equal level of the good.

case the government may very well win.

And let's say that

And let's assume that all the

In that

Now, the government frequently doesn't have a

random sample.

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

the government has 100 affidavits or 200 affidavits, or

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

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

or 200 customers out there.

And two cases make this very point.

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

only a few customers who switch to alternatives to make

the price increase unprofitable, thereby protecting a

large number of customers who would have acquiesced to

higher GQA prices."

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

No matter how many customers in each end-use

industry the government may have interviewed, those

results cannot be predictive of the entire market if

those customers are not representative of the market.

In SunGard, the court made a similar point.

The court wrote:

"The sampling of customer statements

before the court is minuscule when compared to the entire

universe of defendant’s shared hotsite customers.

Although the government has submitted approximately 50

statements from customers stating that they either would

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

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

shared hotsites.

simply cannot determine whether these 50 declarations are

representative of the shared hot site client base.”

Without more information, the court

In some instances, a few customers, however,

may be able to speak to the overall demand elasticities

of an industry.

may have avoided the sampling problem when it used

affidavits of looseleaf chewing tobacco distributors as a

proxy for the views of a broad spectrum of customers.

For example, in Swedish Match, the FTC

The last form of evidence I want to talk about

is survey evidence.

A frequent criticism of consumer

affidavits is that they are form affidavits, which simply

give yes or no responses to a series of questions,

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

including whether a customer would switch in response to

a SSNIP.

really may be is a survey.

Well, that may not be an affidavit; what it

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

obtain as many affidavits as possible without taking the

time to create detailed, customer-specific affidavits, it

may be advisable to conduct a survey instead.

conducted properly by an expert trained in acceptable

survey methodology, surveys should carry more weight than

even an allegedly representative sample of affidavits.

For example, surveys may have more

methodological controls concerning leading questions,

selection bias and randomness.

easier to organize customers by key competitive traits,

reflecting the probable elasticities of their demand.

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

the government to depose survey respondents.

Indeed, if

Surveys may also make it

For the hypothetical monopolist test to serve

as something more than an instructive theoretical

paradigm, litigants, agencies, and courts should look to

empirical evidence that can directly answer the question:

How many customers will switch in response to a SSNIP?

Well, historical evidence, affidavits,

econometric evidence, and survey evidence can all be

subject to criticism.

There can be little question that

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

they are better than conjecture on what is a critical

question on market definition.

MR. WERDEN:

Thank you, John.

cross-examination phase.

MR. HARKRIDER:

And now the

Do I sit over there for this,

or --

MR. WERDEN:

Wherever you like.

You're

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

defendant.

MR. HARKRIDER:

MR. WERDEN:

Fair enough.

I am intrigued by your suggestion

that we should care whether price increases are cost-

justified in examining historical evidence of how

customers reacted to price changes.

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

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

are cost-justified.

thinking on this subject?

MR. HARKRIDER:

So, would you elaborate your

Yes.

I'm not sure if it's

particularly important whether they are cost-justified or

not.

frequently change in an industry.

those price changes, which may very well be cost-

justified price changes, you need to look at that context

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

instructive on the hypothetical monopolist test.

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

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

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

MR. WERDEN:

Okay.

Let's turn to your

suggestion that somehow switching costs have to be taken

into account in market delineation.

very simplistic hypothetical, because any real case is

way too hard to do in the time we have.

Let me give you a

Let's suppose that all the current users of a

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

The hypothetical monopolists might be limited to

exploiting the current users, and giving up on any new

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

discriminate between the two.

In either event, my question is: Is there any

way that the presence of switching costs would cause you

to depart from the standard 5 or 10 percent price

increase as the significance threshold for market

delineation?

MR. HARKRIDER:

Yes, I think you make an

interesting point, and I think the court in Englehard

makes this very point.

customers that may be locked in, the dynamic of

competition you may want to be looking at is the

When you're dealing with

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

competition for new customers.

point.

I think that's the first

I think the second point is that in order for

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

answer needs to mean something different, or lead to a

different conclusion than a no answer.

So, if you ask someone whether they would

switch in response to a SSNIP to another product that you

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

same question with respect to a product that is not

within the relevant market, and the answer to both of

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

for that test to have meaning you cannot reach different

conclusions with respect to whether one is in the

relevant market or the other one is in the relevant

market.

So, if the customer would not switch -- in

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

market, I think you need to take into account switching

costs.

products that were within the relevant market and a no

answer to products that are not within the relevant

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

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

that there isn't any competition between the merging

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

was the conclusion.

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

delineating the market.

hypothetical monopolist would raise price significantly.

And I am really having trouble seeing how switching

between incumbent sellers of the relevant product who

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

market, how switching among them is relevant to the

question that market delineation poses.

We are asking whether a

MR. HARKRIDER:

Well, I think you're actually

answering the question the same way that I would answer

it.

relevant, and so the point I was making is that

affidavits that ask the question that you say is not

relevant are probably not relevant to the issue of market

definition.

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

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

court had focused on those customers that are not locked

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

have asked.

to customers that are locked in answers nothing, other

than the fact that these customers are locked in.

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

the case might be that the locked in customers will be

exploited, and that the merging firms will quit selling

to new customers.

of course, evidence has to be mustered to show that

that's a sensible strategy after the merger.

If that's the theory of the case, then

But if that is so, then clearly the focus in

market delineation ought not to be on the new customer,

should it?

MR. HARKRIDER:

Well, I think it's important --

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

And

not to make that distinction, but the court in Englehard

certainly gave no weight to the affidavit statements,

because of the presence of switching costs.

And I think that unless you're arguing and can

persuasively argue that Englehard is wrong, I think that

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

matter in that context.

MR. WERDEN:

MR. HARKRIDER:

MR. WERDEN:

we don't have time to go into it.

(Laughter.)

MR. WERDEN:

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

first.

Just -- you mentioned about being in the dock.

actually testified a few months ago, and there was a

federal judge in Rhode Island that actually makes the

witnesses stand.

And I didn't understand that, I thought the

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

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

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

an incentive to give short answers.

But in any case, second thing, everyone is

talking about 20 years of experience.

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

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

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

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

by telling a story about a hypothetical monopolist case.

I started working in the Antitrust Division in

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

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

merger case, for a possible challenge.

I am getting old.

And I looked at some old affidavits, and as I

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

a market."

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

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

talked to senior lawyers and senior economists, and they

gave me a lot of different suggestions.

The lawyers gave me the Brown Shoe standards,

and some of the economists said, "Look at price

correlations."

Some others said, "Look at cross-

elasticities."

But none of the suggestions -- not one --

had anything to do with the actual exercise of market

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

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

of a logical progression.

So then I finally did what I should have done

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

he suggested looking at it as a hypothetical monopolist.

And it was the first I had heard of it.

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

systematic way, and probably should have.

I wasn't like

But in any case, once the question was put with

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

sense to me.

logic of the hypothetical monopolist principle.

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

identify a group of producers that would be able to

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

second part -- if they were able to coordinate their

pricing output decisions.

And that ultimately gets to what is the

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

And at

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

One

thing that I notice that I think often is not done

properly is that when you're considering whether or not

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

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

it has to be consistent with the logic of what that

hypothetical monopolist is going to do.

In other words, as Greg points out, the basic

principle is would the hypothetical monopolist price in a

certain way and would it be profitable.

going to ask the questions about coordination, the second

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

definition when you're doing your competitive effects

analysis.

Let me just reiterate that.

Well, if you're

The hypothetical

monopolist proposition tells you what the hypothetical

monopolists would do.

you do your competitive effects analysis it seems to me

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

context of what that hypothetical monopolist would be

doing.

You look at that, and then when

And let me give you an example.

Let's say that

a hypothetical monopolist prices differentially -- so it

might be unilateral effects, might be something more

complicated than that.

Analysis of that market must

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

consider the likelihood that the coordination of this

type of differential pricing could actually be achieved.

The basic concept related to the hypothetical

monopolist principle brings me to the focus of the

outline that I provided Greg for this session.

pointed out, I often use critical loss both before the

agencies and in testimony.

several questions raised about the use of critical loss.

Is it appropriate?

appropriate?

And recently there have been

Under what conditions is it

And to understand, let me just give brief

background on critical loss.

loss basically does is it makes the hypothetical

monopolist principle operational.

As John

In my mind, what critical

And all critical loss is is a two-step process

that first identifies for any given price increase the

sales that can be lost before the price increase becomes

unprofitable.

much sales have to be lost before it becomes

unprofitable?

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

How

The second step considers whether the actual

sales loss associated with the hypothesized price

increase will exceed the calculated critical loss.

it's two distinct steps.

third step, because this form of the critical loss looks

And one might even argue it's a

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

at a particular price increase.

And in fact, the profit-

maximizing hypothetical monopolist can charge any number

of prices.

look at a variety of prices and try to answer that

question.

So, a full application of critical loss would

Now, a practical consideration is -- as the

size of the price increase goes up, the confidence you

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

less.

you're working with an observed interval of price

increases.

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

increase, you probably haven't observed changes like

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

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

For example, you have greater confidence when

Whether people deal with 5 percent increases

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

ultimately depends on the size of the price increase

you're postulating, and the cost structure of the

hypothetical monopolist.

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

hurts.

In effect, the cost structure

The answer to the second step, how much will

actually be lost, can be ascertained by traditional

antitrust methods such as econometrics, other statistical

techniques, looking at business records, and going

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

through testimony.

talked about in the previous talk.

And this, in part, was what John had

Now, one of the questions that has been raised

about critical loss concerns the relationship between

cost price margins and demand elasticity.

been three or four articles that raise this point.

in particular, Michael Katz and Carl Shapiro have raised

the point that economic theory indicates that high

marginal cost margins tend to imply that actual loss

sales associated with the price increase will be small.

Well, the Carl Shapiro observation is based on

a theoretical inverse relationship between marginal cost

margins and demand elasticity that's encompassed in

what's called the Lerner index.

is a residual of looking at profit maximization under

certain circumstances.

There have

But

And that, effectively,

Now, there are several reasons why large gross

margins may be consistent with unit sales actually being

sensitive to price changes, and Katz and Shapiro identify

three of them.

already be coordinating prices, and that will sever the

relationship.

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

in which there is a kink in the supply curve.

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

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

the Lerner Index for this purpose fails to distinguish

between margins that are based on marginal cost --

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

based on the average variable cost of a significant

increment of quantity, where the answer may be quite

different.

The correct margins used in a critical loss

analysis should consider the actual average variable cost

associated with a significant loss of sales, often more

than 10 percent of current sales.

example, one of the situations where I presented critical

loss was the Poplar Bluff Hospital merger case.

v. Tenet.

And just to give an

It's FTC

And there, when we went back and we looked at

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

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

careful -- but we knew from previous work that margins

for hospitals were somewhere in the range of 60 percent,

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

fixed staffing.

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

accounting system, we sat down, we worked with their

internal cost accounting system, and went through, item

by item, and asked the question in conjunction with

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

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

to lose 10 percent of your sales for a year?

say you lost a big managed care contract.

happen to the individual cost elements?"

Like, let's

What would

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

course of the year.

used, and we ended up with margins there that were

similar to other hospitals.

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

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

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

the hypothetical that we presented become variable, and

the calculated margins change.

And the answer we got was what we

But lesson learned -- not a

So, I guess the lesson from this is simply get

the right costs.

Make sure that it fits the situation

and the questions being asked for the specific case.

Now, second point with regard to the Lerner

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

describes is limited in some sense.

monopolists and dominant firms, and it also applies to

industries where Bertrand-type competitors sell

differentiated products.

It applies to

Now, and despite the fact that many, many years

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

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

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

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

Cournot models -- at least my understanding of that.

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

questioning things like critical loss, or anything else

based on particular models such as Bertrand-type models,

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

has to incorporate all the implications of the specific

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

So, for example, Bertrand models tend to apply

best for differentiated products.

that you have -- the relationship that Katz and Shapiro

talk about that comes from the Bertrand model doesn't

tell you a whole lot about closeness of substitutes.

But the relationship

Again, my understanding is that it doesn't tell

a whole lot about cross-elasticities.

analyzing differentiated products, a key issue is are

these close substitutes, and what happens, for example,

if you raise the price of one of them.

doesn't apply there.

But when you're

The relationship

And ultimately, this all brings us back to the

hypothetical monopolist principle -- I see my time is

running out.

the guidelines is a major leap.

probably very specific circumstances where other

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

specifically on the questions that it's supposed to be

asking, and that is can market power be exercised.

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

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

Greg's questions.

MR. WERDEN:

Thank you, Barry.

I want to ask

you a couple of questions about your reaction to Katz and

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

Their most basic point is that anybody who is

trying to make something out of high margins ought to

have some idea of what competitive process produced those

high margins, and be able to construct a comprehensive

argument that takes into account the competitive process

that produced those high margins.

that?

MR. HARRIS:

Do you agree with

I mean, it seems almost

tautological.

process.

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

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

And that is basically a function of the cost structure in

the particular market is a market that has high fixed

costs.

minimum staffing that's not going to change?

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

or it goes out of business.

simply be reflecting high fixed costs.

MR. WERDEN:

And the high margins may

Well, unfortunately, Mike can't be

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

for him.

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

it backwards?"

are high.

process will, in equilibrium, allow you to recover those

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

competition.

wouldn't be able to recover those fixed costs.

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

competitive process out there that's producing those

margins.

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

I think that you're disagreeing with them at

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

the competitive process had to be responsible for

creating the high margins.

margins.

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

intended to say.

I mean, I think competition ultimately

provides competitive prices equal to marginal costs.

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

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

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

have as many competitors.

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

we're talking about a loss of sales here.

So if you are

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

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

mean the last units or so.

But if you have a large amount of fixed costs,

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

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

it.

curve can be substantial.

the shape of that cost curve is.

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

Lerner-index relationship says.

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

very slightly.

You point out that their analysis is

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

the right model for all industries.

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

industries.

But there is no

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

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

the analysis lead to their conclusion that the high

margins are consistent with very narrow markets, not

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

broad ones?

MR. HARRIS:

I think it's going to depend --

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

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

Bertrand model, you have differentiated products and you

may very well have different margins with the different

firms.

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

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

what the cost elasticities are between different firms in

the market.

differentiated products, and I'm not sure --

That's important when you're dealing with

MR. WERDEN:

And they say it is.

MR. HARRIS:

That --

MR. WERDEN:

That the result is that if you're

in this Bertrand world with very high margins, then it

turns out that you need very little substitution, very

low cross elasticities among the firms in the industry in

order for the right conclusion to be that they form a

relevant market.

MR. HARRIS:

Well, I mean, in the simple model,

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

produce similar products but have different costs -- not

exactly the same products, but have different cost

structures.

You're going to have one firm that has much

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

larger margins than the other.

close competitors, but if that firm with the better cost

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

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

other firm.

sales.

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

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

there.

MR. WERDEN:

Well, the question is:

Aren't

they right in their case?

it true that high margins are almost always going to lead

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

competition among the firms in the market, there is

enough so that the market elasticity is enough lower than

the individual firm elasticities which, of course, have

to be low if the firms have high margins?

In the Bertrand world, isn't

MR. HARRIS:

I guess I just don't know the

MR. WERDEN:

All right.

MR. HARRIS:

Okay, thank you.

MR. WERDEN:

Our next speaker is Will Tom.

MR. TOM:

answer.

Thank you, Barry.

Thank you, Greg.

I'm not quite as

old as Barry, but I'm getting there.

And I think back to

my first year as a young lawyer in the Antitrust Division

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

in 1979.

thinkers like Greg were already using the hypothetical

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

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

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

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

market, and how to start thinking about this case.

And despite the fact that sophisticated

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

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

Barry was able to get good guidance from George Hay.

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

that perspective, I think the 1982 merger guidelines was

really an advance.

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

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

involved in that effort and in developing the guidelines

and propagating its use.

But

From

And if any ego-meters should be going

From that perspective, I think what I have to

offer are really just a handful of modest suggestions.

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

think the guidelines approach is a good one, and

notwithstanding the quote from George Stiegler that John

mentioned, it really is operational, by and large.

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

explicit recognition of the role of uncertainty as we

apply these guidelines.

that we ought always to bear in mind that the underlying

purpose of market definition is the assessment of

competitive effect.

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

Under the role of uncertainty, I guess my first

suggestion is that the staff be aware of the tendency of

the iterative hypothetical monopolist test to

unconsciously reverse the burden of proof.

Now, I know that the guidelines explicitly say

this is not intended to say anything about burden of

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

else.

burden of proof on those who are asserting that something

inherently uncertain will occur, as opposed to those who

say that something inherently uncertain will not occur.

But there is a natural human tendency to put the

And if you think about how the iterative

hypothetical monopolist test works, you start by positing

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

transitory price increase.

taken as a given, that's a fact.

would happen in response to that price increase by a

hypothetical monopolist if that price increase were to

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

market to defeat that price increase?

Right?

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

is, well, who the hell knows?

Right?

But since that

really is a non-operational approach to the practical

question of getting through the analysis of a particular

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

we think is going to happen?

And it is a very small step to slide from there

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

price increase by this hypothetical monopolist will, in

fact, be defeated, right?

iterations, you know, prove that the price increase would

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

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

broad as maybe it really is.

And you go through enough

And so, there may be a certain inherent bias in

favor of markets that are too narrow.

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

may be just a natural consequence of the way we approach

questions that are inherently uncertain.

And now, I can't

My second point about uncertainty is in a

similar vein.

And it's not really a market definition

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

step after market definition, and the next step after

that, and how it relates to the process of market

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

definition.

And I am reminded of a scene in the play or

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

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

notes."

the many steps in the merger guidelines process is to

lead to some errors in situations of an inherent

uncertainty.

And there may be a sense in which the effect of

And to take a very stylized example, imagine a

merger investigation in which the analyst concludes that

there is a 51 percent probability that the merging

parties A and B are actually in the same market

themselves, right?

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

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

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

you go from there.

Okay.

I mean, the -- you start with a

And then

Let's suppose that there is a serious

question about that first step, and maybe the merging

parties aren't even in the same market.

with 51 percent probability, they probably are in the

same market.

All right.

But you conclude

Next step you have got another

player, C.

He might be an uncommitted entrant, right?

He might be able to start selling this product, even

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

though he is not currently selling the product without

substantial cost in a very rapid period of time.

Well, let's suppose you can reject that

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

There is only a 49 percent likelihood that uncommitted

entrant would come in.

participant in the market.

And so you say C is not a

And let's suppose there is another potential

entry, D.

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

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

there is only a 49 percent likelihood of committed entry.

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

market.

And maybe he's not currently producing a

Okay.

Well, how do you analyze this merger as

a whole?

be anti-competitive, and assuming a lot of those

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

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

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

likelihood that this merger is going to do any harm,

right?

If the question is:

Is this merger likely to

I think there is a tendency for at least

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

and B are in the same market, okay?

Now, is there going

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

to be uncommitted entry?

the market?

What are the participants in

Well, more likely than not, there is not going

to be uncommitted entry, so we will reject the argument

that the parties are advancing about uncommitted entry.

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

guidelines define it?

is not going to be entry, right?

know, two merging parties in the same market.

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

Therefore, we should challenge the merger.

Well, more likely than not there

So you have got, you

Nobody

It is, I think, really difficult for the

merging parties counsel to know when the staff is

thinking this way.

process whereby the staff is properly reticent about

sharing information that they are learning from third

parties, and that can sometimes inhibit the dialogue.

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

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

of it.

You have a somewhat asymmetric

So, I throw that out as something for the

agencies to at least be cautious about in their internal

deliberations.

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

switch.

morning about affidavits and survey evidence, and so on

-- particularly in intermediate goods industries, where

the staff is relying on customer telephonic interviews in

ways that they probably are not in consumer goods.

And we're talking about that at some length this

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

going on there behind the curtain.

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

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

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

avoid some of the more obvious errors about how the

questions are phrased, and the sample of respondents

they're posed to, and so on.

I think John pointed

Where you're relying on the telephonic

interviews with customers, the danger of badly worded

questions eliciting misleading information about the

inherently unknowable is probably at its highest.

is also the situation in which fruitful dialogue between

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

difficult because of concerns about third-party

confidentiality.

And it

And it is very easy to slip into highly

unproductive dialogues that start and mostly end with

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

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

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

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

this:

what the customers say they would do and focus all of

your attention on the reasons that they give for what

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

objective facts that make switching more or less likely,

because that is something that the parties generally can

engage on.

The dialogue is much more productive if you ignore

And it is a lot more concrete and reliable, I

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

third parties -- what customers say they would do under

circumstances that they may not really have faced and

really may not have thought through.

All right.

Well, let me turn to the other

major heading, which is that assessment of competitive

effect is the underlying purpose of market definition.

And I guess the first question I want to

address is when should we depart from five percent?

know, the guidelines say that in appropriate

circumstances, we can use numbers other than five percent

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

circumstances are.

You

And let me take you through an example that's

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

easier to understand and then go from there to situations

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

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

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

which a five percent price increase is unprofitable, but

a 10 percent price increase is profitable.

would that happen?

price increase, you lose a certain block of customers.

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

make up for the loss of revenues from the customers you

would lose.

increase is unprofitable.

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

highly price-inelastic, though.

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

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

customers that you do keep.

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

others have, as well.

So if you raised the

And I gather Greg, in some

In that situation, if the guidelines simply

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

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

case for upward departure on the five percent, because

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

likelihood of competitive effect, I'm simplifying past

the coordinated effects step, and assume perfect

coordination, or assume only two players or something --

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

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

that protected against a 5 percent price increase but

didn't protect against a 10 percent price increase.

As Greg has pointed out, the guidelines say "at

least."

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

satisfied, even under the hypothetical that I am talking

about.

So, if a hypothetical monopolist could maintain

So, the case for an upward departure would have

to involve somewhat more exotic examples.

and C are merging.

hypothetical monopoly of A and B could not sustain a 5

percent price increase for the same reasons I described,

but could do so with a 10 percent price increase.

Imagine that A

B is a closer competitor than C.

And you know, should you call this a market and

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

or should we say that C is outside the market all

together, and kind of stop your investigation there.

There are scenarios in which it would make sense to keep

going.

The downward departure scenario is more

straightforward.

Five percent price increase is not

sustainable, but a one percent price increase would be,

because relatively few customers are diverted.

going to be the inverse of the example that I started

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

It's

with.

And there, I think the theoretical case is

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

hearkening back to the theme of uncertainty in the last

group of examples I talked about.

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

even with perfect coordination among all marketplace

participants, that puts an upper bound on how much damage

can be done if you fail to challenge the merger.

If, by hypothesis, a

Given the vagaries and the uncertainties of

life, maybe that should be enough.

somewhat facetiously that maybe we should delete the part

of the guidelines that says this five percent is not a

tolerance level.

tolerance level in the merger guidelines.

I have suggested

And yet there ought to be some

And I am not sure what the right tolerance

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

know, the whole process of merger analysis is an inexact

science, at best.

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

maybe you ought to stop there and let this merger go

through.

And if you're reasonably confident

All right.

I am running short of time, so let

me hasten quickly through my last two points.

aesthetic reasons maybe, if nothing else, can we do

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

Just for

something to reconcile the unilateral effects discussion

of competitive effects with the market definition

process?

And here, in particular, I focused on the

sentence in Section 1.11 of the guidelines that reads --

and I quote -- "In performing successive iterations of

the price increase test, the hypothetical monopolists

would be assumed to pursue maximum profits in deciding

whether to raise the prices of any or all of the

additional products under its control."

And I am not entirely sure what that means, but

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

sustainability of a five percent price increase in only

one of the hypothetical monopolist products, that's

sufficient to define a market.

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

analysis?

define a market based on the hypothetical monopolist

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

that any decent unilateral effects case is a merger to

monopoly?

This aspect of market definition that says you

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

one is I'm just confused.

I'm not sure that this anomaly

has actually done any harm to merger analysis, but it

certainly has made it a whole lot more confusing and

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

maybe has made the guidelines just a little bit less

transparent to pick up one of the themes raised by our

keynote speakers.

And given the amount of time, I am going to

drop my last supply substitute ability point, which Greg

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

myself up to the cross examination.

MR. WERDEN:

Okay.

Thanks very much, Will.

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

want to start with the issue of the upward departure.

might not be an accurate characterization, but we will

use that as a shorthand.

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

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

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

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

persuaded that a 5 percent price increase ought not to be

looked at, but rather a 10 or more percent increase

needed to be looked at.

than 10.

And it maybe needed to be more

And the FTC did persuade the 9th circuit of

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

in mind by an upward departure.

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

circuit had it right.

And what I want your

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

MR. TOM:

All right.

I am not going to comment

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

recently.

the beginning, one of the first things that goes is

memory.

And giving the advancing age I alluded to at

MR. WERDEN:

Well, neither of us actually knows

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

question is.

percent price increase these two swimming pool sanitizer

chemicals would be in the same market, should you just

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

My question is:

MR. TOM:

Is the idea if for a 15

Yes, I think it depends.

If there is

a credible theory of competitive effect that says, in

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

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

to let market definitions stand in the way.

And I alluded to the hypothetical situation in

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

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

hypothetical too deeply.

I posited three players, because you can

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

increase would not bring one of the party's closest

competitor into the market.

with not a hypothetical monopoly, but an actual monopoly.

I mean, yes, let's start

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

Even a five percent price increase will not bring the

closest competitor in.

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

its closest competitor?

intuitively, no.

what would be a fairly dramatic competitive effect under

that circumstance.

And I think most of us would say

It wouldn't make any sense to allow

Well, one might cavil about whether that

hypothetical is realistic in the first place, because if

a five percent price increase by the monopolist wouldn't

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

the monopolist pricing it at that level already?

And so, I posited another competitor, B, that

is closer than competitor C that A is acquiring.

current price is constrained by B.

haven't gone up already.

So the

And that's why prices

Now, the question is can the acquisition still

make a difference?

which the acquisition of the next closest competitor does

make the difference.

I think there are certainly models in

In such a circumstance, does it make sense to

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

first place so why worry?"

probably want to look at that model a little more closely

and see this competitive effect really realistic?

I think not.

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

I think you

Is it

likely to happen?

Are you convinced as a factual matter?

MR. WERDEN:

One last question relating to your

comment on tolerance level.

you suggest, to recognize uncertainty in our forecasts.

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

here.

I think it is important, as

Let's suppose we had a unilateral effects case

in which the government believed that the result of a

merger would be one merging firm would increase price

eight percent, and then the other one would increase

price four percent, for a market-wide average price

increase of three percent, or make it two percent.

The three and the two percent are below the

magic five.

the merging firms is below the magic five.

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

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

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

percent, forget it.

MR. TOM:

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 --

MR. WERDEN:

don't know anything.

uncertainty.

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

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

tolerance level.

where we can be pretty confident about the upper bound of

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

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

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

consider letting that one through.

Yes, I do think that, in circumstances

MR. WERDEN:

And not a question, but just a

comment.

The sentence of the guidelines that you quoted,

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

following sentence does.

(Laughter.)

MR. TOM:

MR. BLUMENTHAL:

Okay.

I stand corrected.

Thank you.

Good morning, everybody.

I am

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

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

some of the things that were said by the three earlier

batters.

We begin by noting that there is not a whole

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

save for two things that my former partner and good

friend, Will Tom, said in passing.

The first was simply the reference to too many

notes.

And my recollection is that that was not Emperor

Joseph criticizing Mozart, but Mozart criticizing

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

Salieri.

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

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

tape.

I'm not going to dwell on that point.

The second thing where I think we had a little

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

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

all, the guidelines are working well.

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

guidelines are working well, I suppose, but only because

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

words mean, and are largely ignoring them as the recipe

for which they were originally intended.

I'm going to spend

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

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

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

agreement -- among us all.

The hypothetical monopolist test has basically

won.

most of the commentators adopt it.

adopted in the courts, but it has largely prevailed in

the courts.

some tools as early as 1983 or 1984.

have been amended substantially.

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

Harkrider and Will Tom, that affidavit battles are not an

especially productive or illuminating form of arms race.

I agree with Will on the issues of burden of

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

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

But unilateral effects, I agree with Will about his

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

have to read the paper.

And basically, I think pretty much all of us

would say that if market definition is to remain a

discreet early step in merger analysis, that the

hypothetical monopolist test is the right test.

Now, to the disagreement, let me be a little

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

whether, in fact, we ought to be retaining market

definition as a discreet, early step in the analytical

process as it is used nominally, under the guidelines.

For at least five years now -- and just

watching the way we work, watching the way the agencies

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

whether market definition is an input into the

competitive analysis, or whether it isn't really an

output.

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

determination.

And I think that Greg has largely convinced me

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

But the one thing that I think is fairly clear

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

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

analysis done by reference to a hypothetical monopolist

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

follows.

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

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

probably 15.

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

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

court.

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

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

out of the agency.

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

use it.

don't do it.

typically mean by market is that grouping where, at the

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

found.

sound way to go about the analysis.

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

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

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

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

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

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

but will be.

emerge from this conference.

I'm sure sooner or later, the papers will

I cited to Jon for purposes of his paper on

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

loquitur" merger analysis.

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

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

fundamentally lawless.

way of thinking about the issue.

And we, of course, in the

But it's actually a pretty sound

And I cited to Dale for a three-step analysis

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

and a lot of others have picked up.

going through the guidelines type of test, what we

usually do when we sit down with clients is pose three

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

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

happen to price?

I mean, rather than

And for this purpose, I am using price as a

surrogate for all of the other competitive variables that

you might get into.

might get into innovation, into quantity, into just all

of the things you might worry about.

And when you poke and probe, you

The first two of those are really background

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

information.

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

or the cost decrease side.

the assets are going to be more profitable in the hands

of the buyer than in the hands of the seller.

presumably, a deal wouldn't happen.

some margin increase somewhere, and the question is

what's the source of the margin increase.

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

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

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

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

the customers going to say that?

don't they care?

Why do they care?

Why

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

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

competitive variables.

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

ways.

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

have a price increase that sticks?"

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

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

But I would suggest to you that if you really

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

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

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

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

have a lot to learn from the doctors.

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

typically what the doctor does is take a bit of the

history and formulate some hypotheses, and then you flip

into a mode of rule out a problem.

medical profession does it is they run a bunch of tests

sequentially to rule out this problem, rule out that

problem.

tests that are consistent with the problem, then you

treat the problem.

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

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

hypotheses, well, then you move on.

Now, the issue we all have is how do you

translate this into legal doctrine in a system where you

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

propositions?

First, you have to define a market.

And

second, the market definition is step one of the

analysis.

paper that will eventually be forthcoming.

are several approaches, none terribly satisfactory, that

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

the mode of analysis that's actually practiced.

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

to take the next set of steps and conform guidelines to

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

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

of court?

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

of not getting thrown out of court?

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

But let me leave you with this thought.

I'm

going to close with this.

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

against Will, in saying the guidelines are working well.

It does seem to me that the guidelines are affirmatively

misleading as a recipe for how analysis is actually

conducted, you know, 12 years after their most recent

revision.

First, it does seem to me --

Second, the business community is fundamentally

confused about how you do the analysis.

I mean, it's

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

first time and the client will have worked through the

guidelines.

analysis is.

And you can't believe how far off the

Now, the 68 guidelines went 14 years before

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

yet.

And I would not lightly call for the Agencies to

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

undertake a process of revising the guidelines, because I

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

So I'm not going to say that.

But I will say that the time probably should

come one of these years -- almost certainly not before

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

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

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

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

going to seek in the way of guidelines is a pretty

transparent set of recipes that depict the enforcement

practices to the public?"

And with that, any questions, Mr. Moderator?

MR. WERDEN:

I have a few.

First, I'm actually

quite confused about what you mean when you say that the

guidelines are misleading on how the analysis is

conducted.

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

they could be misleading.

have in mind?

I'm not sure that they tell anybody how the

MR. BLUMENTHAL:

What is it, exactly, that you

Well, I guess I'm using

misleading in the FTC Bureau of Consumer Protection

sense.

(Laughter.)

MR. BLUMENTHAL:

Which basically means that you

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

have a population of readers that are interpreting the

words in a way that is not entirely consistent with what

the words literally are intended to convey, that you have

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

seems to me that the same thing applies to certainly

agency staffers overseas who try to apply our guidelines,

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

within our borders who try to apply our guidelines.

The iterative aspect of the analysis is

something that is often fundamentally missed.

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

but rather what you calculate after going through all of

the rest of the steps preceding competitive analysis,

that's something that is quite commonly missed.

The fact

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

think, is pretty fundamental.

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

simply misconstruing how you're supposed to do the

analysis.

comes in.

So that's what I mean by

And in particular, where market definition

MR. WERDEN:

Well, in that regard, let me

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

instruction manual, and perhaps best should be described

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

the competitive analysis of a merger.

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

And if viewed in that light, is there anything

misleading about the guidelines?

problematic about the way they choose to organize these

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

articles 20 years ago, fairly consistent with the case

law that constrains how the government can litigate its

cases?

MR. BLUMENTHAL:

Yes.

Is there anything

I think with that

qualification, the guidelines remain reasonably accurate.

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

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

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

which is sort of the point.

But I'm not sure that that component is

something that withstands challenge.

minor quibble.

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

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

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

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

because they don't have any other recipes.

But that's a really

And yes, subject to the observation that

MR. WERDEN:

Okay.

The

Thanks very much, Bill.

have a large number of prepared questions.

launch into them, I want to offer the panelists an

opportunity to respond to what else they have heard thus

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

But before I

far.

MR. HARRIS:

I actually have a question of you.

I think it was you asking questions about when there were

switching costs and issues with exploiting the installed

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

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

and explain better the situation that you have in mind

where a merger might harm competition.

MR. WERDEN:

Well, sure.

I had in mind a very

simple-minded situation in which every current user is

locked in to his supplier, for whatever reason.

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

that.

Don't

But suppose it is so.

The hypothetical monopolist could exploit these

guys, because they're locked in.

would be, well, what about new customers?

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

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

And a hypothetical monopolist might be able to

discriminate between the two, and might not.

And then the issue

That's usually

And one would have a somewhat different

analysis of the merger, depending on which of those

circumstances you thought you were in.

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

hypothetical monopolist could do is exploit the guys who

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

But one clear

are already locked in.

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

there any way that you would specially account for

switching costs in the application of the 5 to 10 percent

price increase test, and my answer would be no.

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

in delineating the market.

MR. HARRIS:

You

But you would not ignore it in

looking at competitive effects, is that accurate?

MR. WERDEN:

You would not ignore anything in

looking at competitive effects.

MR. HARRIS: I understand.

MR. WERDEN:

But part of what the guidelines

are trying to tell you is what category various issues

fall into.

are market delineation and some aren't."

of them aren't.

And the guidelines say, "Well, some issues

MR. HARRIS:

Okay.

And a whole lot

And then I guess the

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

these customers are locked in, is your suggestion that

there is a way in which the merger can make that

situation worse?

And if so, what is that?

MR. WERDEN:

Oh, that's an excellent question.

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

situation worse, and that's where the competitive effects

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

analysis ought to be.

get to the competitive effects analysis you conclude that

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

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

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

customers, because that's where all the competitive

action is.

my view.

It may very well be that when you

I have certainly seen mergers where that was

But you can imagine, certainly, that the

switching costs between incumbent suppliers are much

smaller than the switching costs between an incumbent

supplier and somebody outside the candidate market, in

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

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

because they're not completely locked in.

Anybody else have any comments or questions?

No?

for almost entirely praise from our panelists today.

my question is has it completely won the intellectual and

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

and I think the answers may be different.

Okay.

The hypothetical monopolist test has come in

And

Put another way, is there an alternative that

makes any sense to use?

And I think this question has to

be posed in two different ways.

perspective, and the second one, a legal one.

One, an economic/policy

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

And with

respect to the economic perspective, I pose the question

to Barry.

Does economics offer any alternative?

MR. HARRIS:

disclaimer, "Never say never."

with a few exceptions, to think of something that

performs better, and it gets back to what question are

you asking:

they were able to coordinate, could exercise market

power?

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

least the ones I have seen -- are inappropriate.

haven't seen them all.

to the same question asked in a different way.

Maybe I

Or, alternatively, they default

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

his talk, he mentioned successful price fixes, and he

also mentioned the Staples case.

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

may tell you what the market is.

successful price fix?

all the conditions are reasonably similar, what you have

done is you have done the hypothetical monopolist test,

you have just run it out in the real world.

Well, I think it is

But what is a

At the end of the day, assuming

Same in Staples.

My understanding of that case

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

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

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

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

ignored everybody's econometrics.

an important fact of that case was that the two companies

each had pricing zones that were defined by the existence

of the other company, and these were the low price

pricing zones.

But my understanding,

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

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

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

the hypothetical monopolist question.

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

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

way of getting to the same point.

MR. WERDEN:

Okay.

So there may be

Thanks, Barry.

And the

Bill, I

will ask the legal side of the question to you.

In your

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

exact remarks -- you said the hypothetical monopolist

test has "won."

opinions that cite the Brown Shoe practical indicia.

Is it your view that the courts are really

But yet we still read district court

relying on the practical indicia, or are they ultimately

really relying on the hypothetical monopolist test, or

some interesting combination of the two?

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

analysts, and the antitrust cognoscenti.

With respect to the courts, I think,

increasingly, the test is the decisive test.

still are some courts that are laggards.

question, in large part, turns on how economically

sophisticated the judge is.

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

standard which perhaps she believes makes sense with what

she thinks she has to put down to avoid getting reversed

by the court above.

But there

I think the

I think it turns, in large

So, as an evidentiary matter, the courts

continue to rely pretty heavily on Brown Shoe.

practical indicia, of course, are a submarket test.

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

So it all gets completely commingled.

And the

But

And I think that for a long time to come we

will continue to see courts citing to the tests that have

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

beginning to see more and more courts speaking favorably

about the guidelines, about the hypothetical monopolist

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

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

without necessarily linking it back.

MR. WERDEN:

Let's get a little more specific,

and talk about the Staples case.

There is, you know, at

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

least three or four pages in the Staples case relating to

the practical indicia.

MR. BLUMENTHAL:

MR. WERDEN:

Yes.

Is there doubt in anybody's mind

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

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

persuaded the reasonably skeptical judge about this

funny-looking market was that there was very clear

pricing evidence that the number of office superstores

really was affecting prices?

MR. HARKRIDER:

MR. WERDEN:

MR. HARKRIDER:

Greg, if I can make a point?

Sure.

I think that Staples actually

provides an illustration of the relationship between

Brown Shoe and the hypothetical monopolist test.

that the hypothetical monopolist test is frequently used

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

is quantitative evidence of the sort, or very clear

evidence of the sort that directly bears on the

hypothetical monopolist test.

I think

I think where there is very little evidence as

to whether a hypothetical monopolist could, in fact,

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

econometric evidence, there is no natural experiment,

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

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

I think in that context they frequently rely upon Brown

Shoe.

weak or not determinative, then they rely upon Brown

Shoe.

And if the quantitative evidence is relatively

I think Staples is a case where you have

essentially, a natural experiment that allowed you to get

direct evidence and get some purchase on the hypothetical

monopolist test.

instructive.

And so, Brown Shoe wasn't particularly

MR. WERDEN:

Okay.

Well, let's try a different

case.

evidence in the case, but the court was totally

unpersuaded by it on both sides.

Brown Shoe factors.

Let's try Swedish Match.

There was quantitative

And the court cited

But the way I read the decision, ultimately the

judge sat back and looked at the totality of the evidence

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

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

So it's a market.

MR. HARKRIDER:

I think that that's right, but

I think that there is uncertainty when you read the

cases, whether the Brown Shoe factors are something that

gets you an ability to analyze whether a hypothetical

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

the hypothetical monopolist test is, in fact, to some

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

extent, one of the Brown Shoe factors.

And you really see courts -- and I think this

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

figured out where the hypothetical monopolist tests

relate to Brown Shoe.

I think that once you get outside of the merger

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

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

of the relevant market is important, are not in fact

looking at the hypothetical monopolist test, although it

clearly is an increasing trend to do so.

MR. WERDEN:

All right.

Let's try another

topic.

An early criticism -- and not necessarily just an

early criticism -- of the 1982 merger guidelines approach

to market delineation was that it would systematically

bias the analysis –- yielding overly broad markets.

hence, understating market shares.

And

I didn't think anybody believed that any more

until a week ago, when the American Antitrust Institute

released a statement in conjunction with this workshop,

which they officially filed with the Agencies as their

comments, which asserts "that there are common market

definition procedures" in the guidelines that "create the

potential for systematic errors in defining markets."

They don't explain.

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

But my question for the panel -- and I will

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

articulation of the hypothetical monopolist test gives

rise to the sort of systematic errors that people were

worried about back in the mid-1980s?

MR. TOM:

Certainly to the extent that people

were worried about a systematic bias toward overly broad

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

and certainly subsequent experience -- has shown that to

be completely unfounded.

I have suggested a reason in my opening remarks

why we might have a systematic bias toward overly narrow

markets.

the guidelines, but simply because of some perhaps

unfortunate interaction between the guidelines and human

nature.

equation or one hypothetical part of the analysis as fact

and another hypothetical part of the analysis as

speculative.

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

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

where the guidelines, as opposed to some view of the

facts, led to a market that they thought was overly

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

broad?

MR. HARRIS:

I'm not sure that I can answer

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

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

Dubuque hospital case, there was an oddity there, and

that was that the lesser of the two hospitals in the

period before the merger had lowered their prices by

roughly 40 percent in an effort to get more patients and

get more managed care plans.

virtually nothing.

And they had gotten

And for some reason, they left their prices

down.

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

back where they were, and presume the elasticity

symmetric, and they didn't do that.

They had an experiment sitting out there that

So, that sort of raises the question -- and

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

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

merging parties?

So it's sort of a related question.

An attempt to answer it directly, the only

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

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

the antacid case that had very, very different looking

competitors in there.

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

may or may not have been supported.

one that comes to mind.

But it's at least

MR. WERDEN:

We referred to that one as stomach

MR. HARRIS:

Okay.

MR. WERDEN:

Because it included a lot more

remedies.

than antacids.

MR. HARRIS:

Yes.

And again, I don't know the

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

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

take two every day for the calcium.

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

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

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

rather than any antitrust analysis.

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

possible for the guidelines, if applied literally, to

yield an overly broad market.

what we mean by overly broad.

It depends a little bit on

But if the breadth from which you were either

too broad or too narrow is what you would otherwise get

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

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

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

that there is almost a chronic bias in the direction of

narrow markets.

Whether that's appropriate or not is something

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

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

markets that are a lot narrower than what business people

conventionally think of as being the market.

reason for that is that business people tend to take more

of a strategic perspective on things.

a whole lot of shifting for five percent one year.

MR. WERDEN:

Of course, it's not one year in

MR. BLUMENTHAL:

enough, fair enough.

MR. WERDEN:

And you don't see

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

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

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

it at the time.

direction at the time.

And I went in writing in the other

Recalling the political climate in 1982, when

the guidelines were released, you know, as much as

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

to a perception that this was what some would have called

Reaganism run wild.

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

And that, from my perspective, at the time was

what was motivating it.

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

That's the best I can come up with.

MR. WERDEN:

And after the fact, 20 years

Okay, thanks.

Next topic.

long-standing issue in the application of a hypothetical

monopolist test is how to account for pre-merger

elevation of price above cost.

that we generally use the prevailing price, but there are

provisions in there referring to possibility of pre-

merger coordinated interaction that say maybe not.

The AAI statement that I alluded to a few

The merger guidelines say

minutes ago issued last week argued as a general matter,

that the “use of the prevailing price should be carefully

evaluated in every merger investigation.”

they didn't say what that evaluation might look like.

So, I am wondering what the panelists think

And of course,

about this.

John.

mean something like short-run marginal costs, which is

generally the way economists think of the term

"competitive price," do you think that that's really ever

the proper benchmark in a merger case?

And I will throw the first question out to

Assuming that by the term "competitive price" we

MR. HARKRIDER:

Well, I think that you can

think of -- I don't want to skip ahead in the question --

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

but you can think of many different contexts in which you

may be uncomfortable using the prevailing price, and I

think at least one context you're talking about is where

there is some evidence of pre-merger coordination.

And I think that you would want to use the

competitive price, as you defined it, if there were

evidence that absent that coordination or collusion, that

price -- that goods were, in fact, being sold at short-

run marginal cost, which may not be the case.

I think that I can imagine many instances in

which the prevailing price, but for coordination, may

not, in fact, be short-run marginal cost.

inherently suspicious of efforts to calculate short-run

marginal cost.

sold at that price, but also because people may be

relying upon accounting data in order to calculate it.

And I would be

That's both because the goods may not be

So I would be very nervous about agencies

trying to calculate that price, and then going to

customers and saying, "Okay, well, would you switch," in

response to a SSNIP on that.

So, I think the bias should be towards using

prevailing price, and I think you should use something

other than prevailing price and, in fact, the competitive

price as you defined it -- if there is evidence but for

collusion goods would be sold at that price.

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

MR. WERDEN:

Well, that gets to the next

question, which is what is that evidence?

persuade you that but for the merger, some lower prices

have to prevail in the future?

MR. HARKRIDER:

What would

I would probably say some sort

of evidence -- I am generally a quantitative person.

I would be looking towards what goods were being sold at

before the coordination.

coordination doesn't exist.

MR. WERDEN:

Or, in a market in which that

Well, you're apt to be hard-

pressed to figure out when the coordination started, or

to observe the industry in that kind of depth.

guess my question really is, since you don't know that

much about how coordination is going on -- because if you

did you would probably just refer this case over to the

criminal people and be done with it -- but it is your

observation that prices are awfully high here.

know, something must be going on.

So I

So, you

If that's the level of your analysis, how can

you incorporate that into some sort of conclusion about

what prices are likely to be, but for the merger?

MR. HARKRIDER:

But I think that the government

has -- or the Agencies have -- a burden.

that if you have no certainty that the price, but for the

prevailing price, but for the coordination, would be

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

And I think

short-run marginal costs, and it seems unclear why you

should be using that as the relevant benchmark.

I can also imagine industries where even in the

absence of coordination, because there are differentiated

products and there is some degree of market power

exercised by or held by each individual firm, that goods

would also not be sold at short-term marginal costs.

So, I just think it's very dangerous to have a

bunch of economists calculating what the price would be

without any clear evidence of what, in fact, the price

would be.

MR. WERDEN:

Okay.

In fact, the guidelines

don’t use the competitive price as an alternative to the

prevailing price; they refer to the use of a price that

is -- I forget the phrase, but "closer to" is kind of the

concept -- the competitive price, i.e. a lower price.

And they don't say which one.

But, Will, I will put it to you now.

thought you had pre-merger coordination, and certainly it

is possible -- and that might even be the theory of the

case, that the merger may help that coordination persist

where it might not otherwise -- what price would you use?

How would you come up with some alternative?

And then if you did come up with some

alternative, how would you implement it?

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

If you

MR. TOM:

This is a hard question.

I don't

think that we have got a lot of tools at our disposal to

create prices other than the prevailing price to use as

our benchmark.

I can think of some half-measures that one

might start with in appropriate cases.

You know, one

thought experiment that is sometimes useful is to -- and

I credit Jon Baker for this -- is to say, what would

happen if prices were to fall five percent?

know, what products would exit the market?

And you

And that may give you some feel for what a more

realistic market somewhat less affected by the Cellophane

fallacy would be.

into trying to estimate what prices would be in a market

without the collusion, you're in a pretty tough spot.

But if you start turning people loose

And I guess the one other thing that occurs to

me is that the -- in some cases, you may be able to find

alternative explanations for prices seeming to be

unusually high.

was meant to exclude situations in which, for example,

the high prices are, you know, a return to the cost of

innovation, or things like that.

And I don't know if your hypothetical

If you were intending to exclude that by

positing coordination, then I think you really are in a

fairly difficult spot where, you know, you really do need

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

to find some other price benchmark, but it's not obvious

what it is.

MR. HARRIS:

May I?

MR. WERDEN:

Yes, but let me throw one thing

out before.

MR. HARRIS:

Go ahead, sure.

MR. WERDEN:

I limited my question to the

scenario of pre-merger coordination because that is the

only scenario which the guidelines endorse using a price

other than the prevailing price.

suggest that there are other circumstances in which you

should do that too, feel free.

MR. HARRIS:

But if anybody wants to

I mean it strikes me in listening

to this that let's say you do have coordination, and

somehow you have reason to believe you have coordination.

Well, in effect, whether it's legal or illegal

coordination, you have what we were talking about before,

the price fix that defines the market for you.

So, it seems to me -- taking Greg's advice from

before -- is the market definition and the competitive

effects are two separate analyses.

market definition point, if you really have coordination,

and you're in around where the monopolists would price,

well then you have kind of defined your market in

advance, and a further price increase would be going past

Well, from the pure

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

what the monopolist wants.

But when you're asking the competitive effects

question, it seems to me the question you want to ask is

is there something about this merger that makes it more

likely that you will be able to continue to coordinate.

But as far as the pure market definition point goes, I

think if you really understand that there is coordination

there, that, in effect, has defined the market for you.

MR. WERDEN:

Thank you, Barry.

Next question.

The hypothetical monopolist test, as articulated in the

merger guidelines, includes what's referred to in the

guidelines and elsewhere as the smallest market

principle.

Interestingly, in last week's AAI statement, it

is asserted without any elaboration:

"The smallest

market principle should be deleted from the guidelines

entirely."

I want to throw out to the panelists whether

they think that makes any sense, and whether they are

inclined to make offensive use of the smallest market

principle.

Do you think the existence of the smallest market

principle in the guidelines has caused the agencies to

miss significant competition between the merging firms?

But I will start with a question to Barry.

MR. HARRIS:

I think the answer is no, and I

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

think the reason for that goes back to what Bill

Blumenthal said, and that is in the real world what the

agencies seem to do is they basically do a hunt to see

where the problem may be.

to an answer from that.

And from -- and then they work

And I don't know that the narrowest market

principle really has much of an impact on how they

address the issues.

question.

How they present it is a different

Another thing is -- maybe this is sort of a

question for Greg -- but I always thought of the

narrowest market principle as basically having two

different possibilities, one that I think is appropriate

and one that I think may not be.

The one that's appropriate would be, well,

let's say we have a well-defined market, the hypothetical

monopolists could raise price profitably.

it's for cars, automobiles.

goes out and buys General Mills, as opposed to General

Motors, you know, making cookies and things.

Let's just say

And then the car monopolist

Well, clearly, the analysis should be in the

car market, not in the car-plus-cookie market.

But the

hypothetical car-plus-cookie monopolist certainly could

raise the price of cars.

narrowest market principle was intended to exclude those

So, it struck me that the

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

kinds of circumstances.

What I think it doesn't exclude, and probably

should not exclude, is let's say there are a variety of

strategies that a hypothetical monopolist could take.

And one involves, let's say, four firms and a different

equally plausible strategy involves seven firms.

It would seem to me that there is a separate

market that would encompass, depending on what

competitive issue you're trying to address, would address

the four firms, and a separate one that would address the

seven firms.

definition says you can't look at the issues involving

the seven firms.

And I don't see that the narrowest market

And again, it may have to do with differential

prices, or what's the next best alternative to the

hypothetical monopolist, rather than just the merging

parties, or things like that.

principle -- cutting it off at just the fourth firm

situation and saying you can't go look at the broader

market that includes seven.

MR. WERDEN:

But I don't see the

The 1982 and 1984 guidelines had a

lot of qualifying language in there about independent

competitive concerns in the larger market.

taken out in 1992.

can't look at that seven-firm market.

That was

So the suggestion may be that you

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

But I think,

clearly, that hasn't been the practice.

John, a question for you.

If you come in to

lobby the Agencies, are you prepared to try to use the

smallest market principle offensively and say that under

our own guidelines, the merging firms don't compete in

the same market?

MR. HARKRIDER:

no hesitation doing that.

examples where we have done that.

Yes, absolutely.

I would have

And let me give you two

I have been involved in three scientific

journal mergers, all of which got similar requests, and I

know of a fourth.

very clear that somebody who is buying a journal on brain

surgery is probably not going to switch to one on foot

surgery in response to a SSNIP.

you're a patient, you're in a lot of trouble.

And in each one of those cases, it was

And if they do -- if

So, I think an argument can be made -- and a

very strong argument can be made -- that each one of

these scientific journals do not belong in the same

relevant market.

the Agency, in each case they were explaining a different

theory, which was a portfolio theory, where they weren't

necessarily looking at bundled goods, but they were

trying to figure out, okay, well, if you increase the

price of a very expensive journal, does this cause some

That being said, and in due respect to

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

sort of externality that is, to some extent, outside of

that narrow market where the library, who is buying 1,000

journals, drops something else.

compete.

It doesn't technically

I think that's an instance where the Agencies

didn't necessarily follow their guidelines but were

looking at competitive effects where they could find

them.

cases the government chose not to challenge the merger.

I think another instance where that occurs is

And you know, thankfully, in each one of those

in software cases.

and then gone down to the Agencies -- where my client may

have a sort of software, let's say, that does sort of

back-office sorts of things.

industry, the company that we're acquiring has a software

that does front office sorts of things.

I have definitely been involved --

And, within a given

So, if you do sort of an analogy between, like,

Excel and Word, you know, both Excel and Word do some

sort of spell check, both Excel and Word allow you to add

and subtract and do little charts.

hopefully -- would say that they are in the same relevant

market -- at least if you're not looking at these things

as suites.

But no one --

So, I think that you can definitely -- and at

least in the case that I'm talking about, a second

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

request was issued.

And I think, for the purpose of

trying to figure out, well, do people buy these as

suites, are they within some broader relevant market.

But certainly, I think in highly differentiated goods,

especially with software goods, it frequently is the case

that goods may be either compliments -- but certainly not

substitutes, although it kind of looks to somebody who

doesn't know much about the products, that they compete.

MR. WERDEN:

Were you successful in your

software argument to the Agency?

MR. HARKRIDER:

Not successful in avoiding the

second request, at great expense to the client,

unfortunately.

of the day.

You know, we were successful at the end

A case wasn't brought.

MR. WERDEN:

Well, that's what really matters,

isn't it?

MR. HARKRIDER:

MR. WERDEN:

MR. HARKRIDER:

Yes, it is what really matters.

You get to bill your hours, and -Well, no, that's not what

really matters.

us that care a lot about the cost imposed on business

with respect to second requests, I think that, you know,

the scientific journal case is a very good example.

But I think that, you know, for those of

I mean, in the last 10 years there have been

four different second requests, all looking at the exact

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

100

same theory, where the guidelines, in fact, would seem to

suggest to you that the products don't compete.

MR. WERDEN:

Well, I won't comment on what the

theory of any of those investigations was, because I

don't know.

So, let's move on to another topic.

Section 1.11 of the guidelines states that what

constitutes small but significant non-transitory increase

in price will depend on the nature of the industry and

the Agency, at times, may use price increases larger or

smaller than 5 percent.

Will has already addressed this subject of

using price increases greater or less than five percent,

but there is one particular question I wanted to ask him

about on this subject that he didn't already address.

And that is the case of a low profit margin industry.

There is some commentary, including from Bob

Pitofsky, and some precedent -- I think all the FTC

supermarket cases, but I don't have any inside

information on that -- have used a price increase smaller

than five percent on the theory that it was appropriate

because profit margins were low.

Do you think that's right?

about the import, if any, of low profit margins?

MR. TOM:

What do you think

I wouldn't think that one should

systematically adopt a percentage increase less than five

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

101

percent in all low-profit margin industries.

a little worried about the implications of that.

I would be

I in part, for the reasons that I articulated

earlier, which is you are getting down to finely granular

judgments about what the competitive effect is likely to

be in situations where the unknowability of it all may

simply swamp what you're trying to achieve.

And given the fact that we're dealing in an

area where there are very likely pro-competitive effects

from mergers, you know, I would be concerned about it.

recognize the intuition on the other side is with profit

margins very low in an industry, you know, a relatively

small price increase is more significant to the players

in the industry.

But it's not obvious to me that catching the

additional mergers that might have anti-competitive

consequences in that circumstance is going to be worth

the risk of erring in the other direction.

MR. HARRIS:

Think what's going on with

groceries.

A very successful grocery store is going to

run maybe two percent return.

they have to pay for all the products.

And the reason for that is

So, if you limit it to questions where you only

look at five-percent price increases, you have basically

almost created situations where entry is going to solve

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

102

all your problems.

Because a five-percent price increase

is going to be increasing profits like three or four or

five-fold.

So, what that means is if you don't have some

flexibility in situations like this, you have basically

defined away those industries because basically, all the

problems are solved by entry in those kinds of markets.

MR. WERDEN:

But there is no necessary reason

why the entry analysis has to perfectly parallel the

market delineation analysis.

MR. HARRIS:

I --

MR. WERDEN:

And the question I put was for

market delineation.

MR. HARRIS:

I agree.

But again, you have to

be careful how you use the market.

against what Tim Muris said and start talking about

shares and things in that market, those shares don't mean

a whole lot if you're looking at a five-percent price

increase, because entry or uncommitted entry, I mean,

depending on the circumstances, is very important.

If you're going to go

That's going to drive the competitive analysis.

So I think almost as a way of screening within the

Agency, it seems to me more prudent to be looking at

smaller price increases and asking questions there.

maybe not.

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

But

103

But it strikes me as a danger in setting five

percent as a threshold and not being able to do this type

of analysis in these kinds of industries -- not because

it's going through a cyclical margin, but rather because

it's inherent in the make-up of the particular industry.

MR. WERDEN:

My suspicion is if one seriously

considered only a one or two percent price increase for

supermarkets, you would find that supermarkets a couple

of miles apart aren't in the same market, and that --

although I have never worked on a supermarket merger, I

have worked on a lot of local market retailing-type

mergers, and I can't imagine that you actually get the

right answer, in some sense, from using such a tiny price

increase, because you're going to keep the merging firms

from competing with each other.

MR. HARRIS:

Actually, I didn't mean that you

wouldn't necessarily use one or two, I meant that you

shouldn't limit it to the use of five percent.

wasn't suggesting that should be a standard to only use

one or two.

applies, then.

So I

So I don't know that your comment actually

Maybe I was just not clear.

MR. WERDEN:

Okay.

Well, thanks.

Final topic

-- and Bill already touched on this quite a bit -- Brown

Shoe, I think perhaps more than any other case, holds

that you cannot challenge a horizontal merger without

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

104

pleading a relevant market, and coming up with some

shares, and et cetera, doing the structural analysis.

But that's a long time ago, and antitrust law

has evolved considerably.

courts now pretty uniformly hold that you can do a rule

of reason analysis without defining a market.

And in the Section 1 area, the

So, my question is even though no court yet

hinted at the possibility of accepting a merger challenge

that doesn't involve market delineation, is sooner or

later that where the law is going?

And I will put the first question to John.

you believe that a court should, and separately would,

accept a merger challenge?

will use a fairly straight-forward unilateral effects

analysis supported by some quantitative analysis, but

with no allegation of a relevant market.

MR. HARKRIDER:

And I will make it simple.

Well, I think the question is

when.

would think probably not.

important in merger policy to be both transparent and

predictable.

need.

If you do this tomorrow, will they accept it?

I believe that it is very

I believe the courts are also aware of that

I think the fact that the current guidelines at

least speak of market definition, and aren't entirely

clear on the issue of whether market definition is

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

105

necessary with respect to unilateral effects, and given

the line of commerce provision in the statute and the

Supreme Court's interpretation of that as requiring a

relevant market, I think that going into court in a

merger case without alleging a relevant market is

something you do at your peril.

So, I think that if the question is right now

is that going to be successful, I think absolutely not.

I think that if 10 years from now you have a period of

time where the merger guidelines have made it clear that

you do not need market definition, at least in the

context of unilateral effects so that there is at least

guideline precedent for that, I believe that there

probably is some probability, just like the hypothetical

monopolist test was accepted by the courts, that that

formulation would be accepted by the courts.

Even though, for example, Bill says that he

doesn't define relevant markets, at least at the lower

end, you still have safe harbors.

harbors are still things that people look at, especially

in the context of commodity goods, where there sort of is

an established precedent on what the relevant market is.

So I think it would send a very scary signal to the

business community that that market definition wasn't

necessary, and I think that you need to do so in very

I think that the safe

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

106

slow steps.

MR. WERDEN:

Go ahead.

MR. BLUMENTHAL:

Can I just add one thing?

Just to be clear, what I said is I don't define markets

on the defense side.

defensive here, but also to go on to say that on the

plaintiff's side it seems to me -- at least if you're a

private practitioner -- it's virtually malpractice per se

not to define a market, given where the courts are today.

And I would think that while the Agencies, as a

matter of not so much prosecutorial discretion, but as a

matter of wanting to move an enforcement program in a

particular direction, might try to do otherwise.

But I raise that not just to be

MR. HARKRIDER:

I think that that's fair.

And

while I do agree -- and I'm sure -- you know, obviously,

I can't speak for you, so I won't try to -- but while you

certainly say that you don't define relative markets, I

am sure you're aware of sort of the precedent.

For example, if you're doing a deal in

aluminum, or you're doing a deal in polypropylene, and

you may say to your client, "Okay, gee, there have been,

you know, 10 other cases either litigated or consent

decrees.

defined the market, and looking at our concentration

numbers, you know, we have a post-merger HHI of 1,000," I

The DOJ has defined the market, the FTC has

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

107

would be very concerned that if, at the end of the day,

the Department of Justice said, "Okay, yes, I know.

have defined markets this way and I know the guidelines

say that, you know, we have a post-merger HHI of, you

know, 35, but we're going to go and look to see if there

is a competitive effect," I just think, given the level,

there are, what, 2,000 HSR's that were filed last year,

there are 30 or 40 second requests.

open up the possibility of second requests on the

balance.

MR. TOM:

MR. WERDEN:

MR. TOM:

I would not want to

Can I just add one thing?

Sure.

While I agree with all that, I think

that the legal barrier to bringing a merger case without

defining a market is the -- “in any line of commerce, in

any section of the country,” language in Section 7 of the

Clayton Act.

the Clayton Act it was established that mergers could

also be challenged under Section 1 of the Sherman Act.

And the Supreme Court has now made clear that in

challenging conduct under Section 1 in the Sherman Act

you don't have to define a relevant market.

Of course before there was a Section 7 of

So, you know, if one were pressed to it, and if

one, for whatever reason -- which I'm having trouble

imagining -- I wanted to try to do a run around the case

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

108

law that says define a market in a merger case, there is

your precedent.

1 merger cases, and you have got a legal theory.

Indiana Federation with the old Section

MR. WERDEN:

And I could add that in the

Rockford hospital case, Posner held that the substantive

Section 1 standards for merger are identical to the

substantive Section 7 standards.

burden for the government if they went under Section 1,

at least if they went into the 7th circuit.

MR. TOM:

MR. WERDEN:

So there is no higher

Right.

Let me, in the closing minutes,

pose a fall-back to going into court without a market.

And the fall-back is that instead of leading with the

market, you trail with it.

competitive effects story -- and this is the way you

write your briefs, this is the way you try your case,

this is the way you do your oral argument.

You lead with your

Say, "Look, we've got this straight-forward

competitive effect story.

Here is evidence that supports

it.

will, but we're also going to explain to you that it's

not really going to help you understand the story.

fact, it's going to be affirmatively misleading, because

you could say the market is this, but that ignores this

factor; you could say the market is that, but that

And if you demand that we talk about markets, we

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

109

suggests that somebody is more important than he really

is, et cetera."

What would you think about litigating a case

that way?

MR. HARKRIDER:

Well, I would think that the

district court would be very concerned that if they

didn't start off with relevant market they probably would

run a significant chance of being overturned on appeal.

So I think that if you're end game is to get

away from relevant market, that's probably the first way

to start, so you could have a district court, say

something in what may effectively be -- you had to find a

relevant market, but it really wasn't necessary.

And that could be the first part of the step of

getting away from relevant market.

said before, if that's your goal, probably the way you're

suggesting it is the safe way to go, and I'm sure the

court, as the safe way to go, would probably say, "We

don't need to, but we're going to define a relevant

market in case you do think we need to."

MR. WERDEN:

I think that, as I

Anybody have any view on that?

Yes.

MR. BLUMENTHAL:

Well, I think for the

foreseeable future, whatever the mode of analysis is

internally, by the time it gets to court it has to be

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

110

translated into the framework with which the court is

familiar.

And the way the jurisprudence on something like

this tends to evolve, the cutting edge stuff happens at

the Agency.

long time, the courts hold to the old framework.

a certain point, some wise judge looks back and says,

"Look what we have been doing for the last 10 or 15

years, let me try to translate it in a way that is

consistent with what the Agencies are doing."

It slowly seeps into the courts, but for a

And at

But for purposes of pleading, for purposes of

the theater of the courtroom, I would think that that

approach is still a little bit perilous.

MR. WERDEN:

Okay.

Well, I think we will end

it on that note.

interesting discussion, and that will conclude the first

session of the workshop.

I thank our four panelists for an

(Applause.)

(Whereupon, at 12:00 p.m., a

luncheon recess was taken.)

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

111

A F T E R N O O N

MR. SIBLEY:

S E S S I O N

I'm David Sibley, from the

Antitrust Division, and this particular panel is

concerned with the role, defined as broadly and vaguely

if you like, of concentration and market shares in

antitrust analysis as influenced by the data recently

released by the Agencies.

Now, we're going to have most of the

discussions structured around three questions.

one is:

a new version of the merger guidelines.

is, what should future guidelines say on the subject of

concentration and market shares defined in any way you

want?

The first

At some point in the future there is bound to be

Whenever that

And John Kwoka, from Northeastern University,

and Steve Newborn are going to talk about that.

have encouraged them to use less than their usual amount

of gravitas, and try to seem a little bit extreme.

(Laughter.)

MR. SIBLEY:

And I

The second main question is do the

Agencies place too much or too little emphasis on market

shares and concentration when it comes to analyzing

mergers?

about that.

And Rick Rule and Jan McDavid will be talking

Then Craig Newmark, with some kibitzing from

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

112

me, is going to talk about whether the empirical evidence

in economics relating prices to concentration and a bunch

of other things, or profits to concentration and a bunch

of other things, does that say anything very specific

about the extent to which concentration ought to underlie

the guidelines or underline merger policy?

And then finally, we're going to have a talk

from Vincent Verouden, from the EC, about the EC

guidelines and the way that those guidelines treat market

shares and concentration.

Now, with the exception of Vincent, who is

going to get 10 or 15 minutes to say what he wants to

say, the rest of them, in extremely impressive display of

ego on the conference call last week, suggested that each

speaker take maybe five or 10 minutes, and everybody else

pile on.

So, I am expecting great things from them.

And

to sort of up the ante and make them feel as embarrassed

as possible if they don't come through for me here, I

will mention that at the beginning of this conference

call last week, when I had my sort of reasonable -- not

very inspired way of allocating speakers and time and

questions, someone -- in fact, everyone here has denied

that they are that someone -- said, "David, have you any

idea of the amount of ego on this conference call?"

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

113

So, that argues to me that these folks -- well,

if it was you, you must have had a cold, or tonsilitis.

MS. MCDAVID:

No, it was Steve.

MR. SIBLEY:

Okay.

So, anyway, it's going to

be a little like an economic seminar at the University of

Chicago, where the speaker gets maybe 10 minutes, and

then the audience piles on.

So, with that, we will start with the first

question.

What should future guidelines say about how

concentration and market shares are defined, analyzed,

used, whatever?

And we will start off with John Kwoka.

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

114

CONCENTRATION & MARKET SHARES

MR. KWOKA:

Out at the University of Chicago,

of course, David, you hardly get your name out before

people jump.

and a great pleasure for me to be here to discuss some

aspects of the question of where we stand with regard to

one of the most prominent and perhaps most controversial

aspects of the merger guidelines.

concentration and market shares.

But thank you very much.

It's a privilege

Namely, the role of

I want to devote these few minutes to a couple

of issues that I hope go beyond some of the old debates.

There are really two disconnects that I see with regard

to economics versus the merger guidelines versus

enforcement policy.

The first of these is the growing disconnect, I

believe, between the way the guidelines are written and

what modern economics teaches about the anti-competitive

effects of mergers.

And the second is the disconnect between the

structural criteria in the guidelines and actual agency

action, which is underscored by the data recently

released by the Justice Department and the FTC.

So, in these few minutes I will try to

summarize these thoughts.

I have more extensive written

remarks that are available in limited numbers in hard

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

115

copy here, but also will be posted on my website at

Northeastern University.

First, with regard to the guidelines and the

economics of mergers, I believe there are really three

major categories of competitive theories of the effects

of mergers:

notion of tacit or explicit agreement among parties;

unilateral effects, which has been in the guidelines now

for 10 years, but certainly has been an issue longer than

that; and strategic behavior.

cooperative effects, the long-standing

Concentration is an important predictor of the

price effects of mergers when the concern is cooperation.

I think this is the implication of a wide range of

economic theories and has been confirmed by a large body

of empirical work.

without its limitations.

work implies precisely what level or change in

concentration matters.

Neither theory nor empirical work is

Neither theory nor empirical

Certainly none of it implies that any

concentration matters, and there is very little guidance

as to what the trade-offs may be with other causal

factors.

believe, and lies at the heart of traditional merger

enforcement practices.

This is a well-established relationship, I

But concentration is much less important in the

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

116

case of unilateral effects.

In the case of unilateral

effects, what matters is demand substitutability between

the merging firm's own products.

on elasticities.

ratios.

certain conditions, on diversion ratios and elasticities,

by market shares.

Substitutability hinges

Those may be informed by diversion

And in turn, some light may be cast, under

So, that's an informational content to market

shares in the case of unilateral effects, but I believe

that their concentration plays no analogously important

role in theories of anti-competitive harm that derive

from unilateral effects.

In fact, there is a conceptual matter.

The

guideline's exercises involving market definition,

product heterogeneity and entry conditions do not matter,

either.

incorporate essentially all of the information that one

would otherwise seek on those questions.

Careful measures of the relevant elasticities

Thirdly, firms may engage in any competitive

conduct against their rivals.

And merger may make this

more profitable, more feasible, and therefore, more

likely.

behavior, and so forth, are different insofar as they do

not involve efforts directly to raise product price.

rather, are intended to handicap rivals, and thereby

Raising rival's cost, foreclosure, disciplining

For The Record, Inc.

Waldorf, Maryland

(301)870-8025

But

117

diminish their competitive effectiveness.

There is, of course, no unified theory of such

behavior, and so there is no simple enumeration of causal

factors.

other factors may matter more in analyzing and predicting

the effects of mergers where strategic behavior is a

central concern.

Concentration may matter, share may matter,

These remarks themselves probably strike no one

as very novel, but the implication is that there really

are ways in which mergers may raise competitive concern.

Each, however, has its own distinctive set of factors

that we look to for analysis and prediction of those

effects.

That's not the way the guidelines read.

The

guidelines set out a common methodology involving first

product and geographic

This text is long and has been trimmed here. Open the source document for the complete record.

This is a copy of a public record, reproduced as it was published. It is not legal advice, and it may not be the version a court would rely on. Check the official source before you cite it.