"The Apple iPod iTunes Anti-Trust Litigation"

Filing 740

Administrative Motion to File Under Seal filed by Apple Inc.. (Attachments: # 1 Declaration Kiernan Declaration ISO Administrative Motion to Seal, # 2 Proposed Order Granting Motion to Seal, # 3 Exhibit Rdacted Version of Defendant's Notice of Motion and Motion for Summary Judgment and to Exclude Expert Testimony of Roger G. Noll, # 4 Exhibit SEALED Version of Motion for Summary Judgment and to Exclude Testimony, # 5 Exhibit REDACTED - Separate Statement ISO MSJ, # 6 Exhibit SEALED - Separate Statement ISO MSJ, # 7 Exhibit REDACTED - Exhibit 3 to Amiri Decl., # 8 Exhibit SEALED - Exhibit 3 to Amiri Dec, # 9 Exhibit REDACTED - Exhibit 4 to Amiri Dec, # 10 Exhibit SEALED - Exhibit 4 to Amiri Dec, # 11 Exhibit REDACTED - Exhibit 5 to Amiri Dec, # 12 Exhibit SEALED - Exhibit 5 to Amiri Dec, # 13 Exhibit REDACTED - Exhibit 6 to 8 to Amiri Dec, # 14 Exhibit SEALED - Exhibits 6 to 8 to Amiri Dec, # 15 Exhibit REDACTED - Exh 9 to Amiri Dec, # 16 Exhibit SEALED - Exh 9 to Amiri Dec, # 17 Exhibit REDACTED - Exh 10 to Amiri Dec, # 18 Exhibit SEALED - Exh 10 to Amiri Dec, # 19 Exhibit REDACTED - Exh 11 to Amiri Dec, # 20 Exhibit SEALED - Exh 11 to Amiri Dec, # 21 Exhibit SEALED - Exh 13 to Amiri Dec, # 22 Exhibit REDACTED - Exh 14 to Amiri Dec, # 23 Exhibit SEALED - Exh 14 to Amiri Dec, # 24 Exhibit SEALED - Exh 15 to Amiri Dec, # 25 Exhibit SEALED - Exh 16 to Amiri Dec, # 26 Exhibit SEALED - Exh 17 to Amiri Dec, # 27 Exhibit SEALED - Exh 18 to Amiri Dec, # 28 Exhibit SEALED - Exh 19 to Amiri Dec)(Kiernan, David) (Filed on 12/21/2013)

Exhibit 11
[PUBLIC VERSION - REDACTED]

Confidential - Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation

Roger Noll, Ph.D.

1

UNITED STATES DISTRICT COURT

2

NORTHERN DISTRICT OF CALIFORNIA

3

OAKLAND DIVISION

4
5

THE APPLE iPOD iTUNES
ANTI-TRUST LITIGATION

Lead Case No. C 05-00037

6
7

____________________________

8

This Document Relates To:

9

ALL ACTIONS

10

____________________________

11
12
13
14

CONFIDENTIAL - ATTORNEYS' EYES ONLY

15

VIDEOTAPED DEPOSITION OF ROGER G. NOLL, PH.D.

16

Wednesday, December 18, 2013

17

Palo Alto, California

18
19
20
21
22
23
24

Reported by:
Darcy J. Brokaw
RPR, CRR, CSR No. 12584

25

Job No. 10008944

www.aptusCR.com

Page 1

Confidential - Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation

Roger Noll, Ph.D.

1

regression are correlated within a particular group

2

and you don't do anything to correct for that, what

3

would be the impact on the reported standard errors?

4
5

MS. BERNAY:

Objection.

Vague and

ambiguous.

6

THE WITNESS:

7

the question.

8

BY MR. KIERNAN:

9

Q.

I didn't completely follow

Ask it again.

If the residual errors in the regression

10

are correlated within a particular group and you

11

don't do anything to correct for that, what would be

12

the impact on the reported standard errors?

13

MS. BERNAY:

Same objection.

14

THE WITNESS:

It could be either way.

It

15

could make them higher or it could make them lower,

16

depending on the nature of the correlation.

17

BY MR. KIERNAN:

18

Q.

19
20

And why would it impact the reported

standard errors?
A.

Well, it's all built up in the -- in the

21

nature of the assumptions one makes in doing a

22

regression analysis, which is an independence of the

23

standard errors.

24

the -- if the random shock that is --

http://www.yeslaw.net/help

25

And if the standard errors -- if

(Reporter inquires.)

www.aptusCR.com

Page 23

Confidential - Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation

Roger Noll, Ph.D.

1

THE WITNESS:

If the random shock that is

2

in the regression equation does not satisfy the

3

independence assumption, then the effect on the

4

standard errors of the coefficients could be either

5

to elevate them or to reduce them, depending on the

6

nature of the violation of the independence

7

assumption.

8

BY MR. KIERNAN:

9

Q.

Okay.

And are there standard statistical

10

tests to test whether the residual errors are

11

correlated within a particular group?

12

MS. BERNAY:

Objection.

13

THE WITNESS:

Vague.

There are many such tests

14

and many such corrections.

15

existence of even statistically significant

16

correlations is small unless those correlations are

17

high.

18

But the effect is -- the

All right.
So the corrections for autocorrelation of

residuals are not something that actually matters in

20

the vast majority of cases because the -- it's

21

almost never the case there's no correlation in

22

residual errors, but it's almost never the case that

23

making a correction for the auto- -- the correlation

24

that does exist matters in terms of the regression.

25

http://www.yeslaw.net/help

19

It's also the case here that we're not

www.aptusCR.com

Page 24

Confidential - Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation

Roger Noll, Ph.D.

1

talking about a source of bias in the coefficients.

2

We're talking about a source of bias in the

3

estimated statistical significance, the --

4

BY MR. KIERNAN:

5

Q.

The standard errors?

6

A.

Yeah, the values of the -- the expected

7

value of the regression coefficients is not

8

affected.

9
10
11

Q.

The coefficients aren't affected, but the

calculations of the standard errors are affected?
A.

Right, the calculations of the standard

12

errors are affected, but the -- but the estimated

13

effect of the independent variable is the same, the

14

expected estimated effect.

15

Q.

And if the residual errors are correlated

16

within a particular group, the standard errors could

17

either be overstated or understated?

18

A.

Yes.

19

Q.

Without a correction?

20

A.

They could be.

Although, again, the --

it's not -- it's not a dichotomous issue.

22

A, they may be affected, and B, the magnitude of the

23

effect depends on the exact conditions.

24
25

Q.

They --

And to know the magnitude of effect, you'd

have to test it, you'd have to run one of the

www.aptusCR.com

Page 25

http://www.yeslaw.net/help

21

Roger Noll, Ph.D.

Confidential - Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation

·1· ·that -- that's a good way to see if there's positive
·2· ·error correlation, but it's not a good way to see if
·3· ·there's negative error correlation.
·4· · · · · · ·And the second point is that the nature of
·5· ·the error correlation may be that it's dependent on
·6· ·particular combinations of variables; and that one,
·7· ·the standard tests wouldn't even tell you that it
·8· ·exists.
·9· · · · ·Q.· In this case, did you do anything to check
10· ·whether the residual errors in your regression set
11· ·forth in Exhibits 3A and 3B to Noll 10 are
12· ·correlated with any particular group?
13· · · · · · ·MS. BERNAY:· Objection.· Vague and
14· ·ambiguous.
15· · · · · · ·THE WITNESS:· What do you mean by "group"?
16· ·BY MR. KIERNAN:
17· · · · ·Q.· Within any group.
18· · · · ·A.· What do you mean, "a group"?· I don't
19· ·understand what you mean by a group.
20· · · · ·Q.· We've been using group for the last ten
21· ·minutes.
22· · · · · · ·MS. BERNAY:· Objection.· Argumentative.
23· ·BY MR. KIERNAN:
24· · · · ·Q.· Same group that you've -- the same group
25· ·that you've been referring to.
Page 27
www.aptusCR.com

YVer1f

Roger Noll, Ph.D.

Confidential - Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation

·1· · · · ·A.· I didn't refer to a group.· I don't know
·2· ·what you're talking about.· I know I fully
·3· ·intended -·4· · · · ·Q.· You used the term "cluster" -·5· · · · · · ·(Reporter admonishes.)
·6· ·BY MR. KIERNAN:
·7· · · · ·Q· ·You used the word cluster, within a
·8· ·cluster.
·9· · · · ·A.· I don't agree that there are any clusters
10· ·here.
11· · · · · · ·MS. BERNAY:· Objection.
12· ·BY MR. KIERNAN:
13· · · · ·Q.· That's not my question, Dr. Noll.· I asked
14· ·you, did you do anything to check whether the
15· ·residual errors in your regressions set forth in
16· ·Exhibit 3A and 3B are correlated within any cluster
17· ·or group?
18· · · · · · ·MS. BERNAY:· Objection.· Asked and
19· ·answered.
20· · · · · · ·THE WITNESS:· I don't know what you mean
21· ·by a group.· And you used the word "or," and I don't
22· ·believe there are any clusters.· So how can I test
23· ·for something when I don't -- I think it either
24· ·doesn't exist or I don't understand what you're
25· ·asking?
Page 28
www.aptusCR.com

YVer1f

Roger Noll, Ph.D.

Confidential - Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation

·1· · · · · · ·What is it you're asking?· Can't you just
·2· ·give me an example of what you mean by a group, and
·3· ·then we won't have to discuss it?
·4· ·BY MR. KIERNAN:
·5· · · · ·Q.· So you don't understand the question?
·6· · · · ·A.· I don't understand what you mean by a
·7· ·group, no.· I don't know what you have in mind.
·8· · · · ·Q.· And you don't know what I mean by cluster?
·9· · · · · · ·MS. BERNAY:· Objection -10· · · · · · ·THE WITNESS:· I know what you mean by a
11· ·cluster, and there aren't any in this particular
12· ·regression.
13· ·BY MR. KIERNAN:
14· · · · ·Q.· How do you know?
15· · · · ·A.· Because I know what cluster analysis is,
16· ·and it doesn't apply to this regression because this
17· ·isn't a sample.
18· · · · ·Q.· What did you do to determine if there were
19· ·clusters?· What statistical tests did you apply?
20· · · · · · ·MS. BERNAY:· Objection.
21· · · · · · ·THE WITNESS:· I looked at the definition
22· ·of a cluster, and it doesn't apply to anything in
23· ·this regression.· I know -- I know what cluster
24· ·analysis is, and it doesn't apply to this
25· ·regression, notwithstanding what many of your
Page 29
www.aptusCR.com

YVer1f

Roger Noll, Ph.D.

Confidential - Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation

·1· ·experts have said.· They're just not right.
·2· ·BY MR. KIERNAN:
·3· · · · ·Q.· Anything else other than looking at a
·4· ·definition?
·5· · · · · · ·MS. BERNAY:· Objection.· Argumentative.
·6· · · · · · ·THE WITNESS:· I know -- the report, about
·7· ·a third of this report is about what cluster
·8· ·analysis is and what kinds of problems you apply to
·9· ·it and why this isn't a cluster sample problem.· All
10· ·right.
11· · · · · · ·So, yes, there it is.· I've cited articles
12· ·in the professional literature of which I not only
13· ·have read, but I actually know what they do.· I have
14· ·taught this stuff.· So I know what I'm talking
15· ·about.· And there's references here.· It's not that
16· ·I just read a definition and decided that something
17· ·didn't apply.
18· · · · · · ·But I know, just from knowing what cluster
19· ·analysis is, that it doesn't apply here.
20· ·BY MR. KIERNAN:
21· · · · ·Q.· You just know it when you see it?
22· · · · · · ·MS. BERNAY:· Objection.· Argumentative,
23· ·misstates his prior testimony.
24· · · · · · ·Come on, David.
25· · · · · · ·THE WITNESS:· That's complete nonsense.
Page 30
www.aptusCR.com

YVer1f

Roger Noll, Ph.D.

Confidential - Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation

·1· ·There is -·2· ·BY MR. KIERNAN:
·3· · · · ·Q.· I'm just trying to understand what you did
·4· ·other than reading some books to determine if there
·5· ·are clusters in the case.
·6· · · · · · ·MS. BERNAY:· Objection.· Argumentative.
·7· · · · · · ·THE WITNESS:· There is no such thing as a
·8· ·test for whether you ought to use cluster analysis
·9· ·in a regression that doesn't satisfy the conditions
10· ·for clustering.
11· ·BY MR. KIERNAN:
12· · · · ·Q.· Okay.· That's what you teach your
13· ·students?
14· · · · · · ·MS. BERNAY:· Objection.· Argumentative.
15· · · · · · ·THE WITNESS:· Of course it is.
16· ·BY MR. KIERNAN:
17· · · · ·Q.· On page 34 of Noll 10 -- let me know when
18· ·you get there.
19· · · · ·A.· I'm there.
20· · · · ·Q.· The first paragraph, the last third, you
21· ·state that "Professors Murphy and Topel do not test
22· ·whether the mean residual errors from this procedure
23· ·are statistically significantly different from zero,
24· ·which would have to be the case if the errors within
25· ·a cluster are correlated."
Page 31
www.aptusCR.com

YVer1f

Roger Noll, Ph.D.

Confidential - Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation

·1· · · · ·A.· Yes.
·2· · · · ·Q.· Did you perform that analysis?
·3· · · · ·A.· No, because I don't believe there are
·4· ·clusters.· The premise of that paragraph is if you
·5· ·assume a cluster analysis is appropriate, here's
·6· ·something you do.· And they didn't do it.· But I
·7· ·don't think you should even do that because it's not
·8· ·a cluster sample problem.
·9· · · · ·Q.· If it turns out that within a group,
10· ·within a cluster -- we can use the one defined by
11· ·Professors Murphy and Topel -- the mean residual
12· ·errors are statistically significantly different
13· ·from zero, what would that tell you?
14· · · · ·A.· Nothing.
15· · · · ·Q.· Why not?
16· · · · ·A.· Because as I said before, you only get
17· ·that far if you have a cluster sampling problem, and
18· ·we don't have a cluster sampling problem.· So
19· ·there's no point in testing for cluster, the
20· ·presence of clustering effects if you don't have a
21· ·cluster to begin with.
22· · · · · · ·This is a paragraph written on if there -23· ·if it were a sample -- if the way I had done the
24· ·analysis was to sample some transactions according
25· ·to a subset of the models of iPods that were out
Page 32
www.aptusCR.com

YVer1f

Roger Noll, Ph.D.

Confidential - Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation

·1· ·could get still get a high squared with a very small
·2· ·subset getting big prediction errors.
·3· · · · · · ·(Reporter inquires.)
·4· · · · · · ·THE WITNESS:· You can have a high
·5· ·R-squared in a regression and still have a group of
·6· ·predictions that were -- where the prediction error
·7· ·is large.· And then you would -- you would still
·8· ·want to address whether that group -- you had some
·9· ·omitted variable for that group or something.
10· · · · · · ·But again, that's not really likely to
11· ·happen if you already have group identifiers.· See,
12· ·again, the -- by definition, if you have group
13· ·identifiers, the residual error within that group is
14· ·going to be zero.· The mean residual error is going
15· ·to be zero, because that's what regression analysis
16· ·does.
17· · · · · · ·So that's why, for example, the most
18· ·conventional solution to cluster problems is to use
19· ·group identifiers, indicator variables, to get the
20· ·mean of those residual errors for each group to
21· ·zero.
22· ·BY MR. KIERNAN:
23· · · · ·Q.· In this case, did you perform any
24· ·statistical test to determine or to test your
25· ·independence assumption?
Page 45
www.aptusCR.com

YVer1f

Roger Noll, Ph.D.

Confidential - Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation

·1· · · · · · ·MS. BERNAY:· Objection.· Asked and
·2· ·answered.
·3· · · · · · ·THE WITNESS:· I have -- I have not
·4· ·performed a test of the independence assumption as
·5· ·you've put it in that way, no.· It would be
·6· ·unnecessary, because there are no groups with
·7· ·outlying residual errors in the R-squared spot.· And
·8· ·by definition, the mean residual errors by group are
·9· ·going to be zero.
10· ·BY MR. KIERNAN:
11· · · · ·Q.· And if statistical tests show that mean
12· ·residual errors within groups are correlated, that
13· ·does not affect your analysis or any of your
14· ·opinions in any way?
15· · · · · · ·MS. BERNAY:· Objection.· Calls for
16· ·speculation.
17· · · · · · ·THE WITNESS:· It might or it might not,
18· ·depending on what the reason for finding that
19· ·correlation was, that statistically significant
20· ·correlation was.· It would purely depend on the way
21· ·the test was performed and the way the groups were
22· ·created and the way the residual errors were
23· ·calculated.· All right.· That's what it would depend
24· ·on.
25· ·///
Page 46
www.aptusCR.com

YVer1f

Roger Noll, Ph.D.

Confidential - Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation

·1· ·understand it.
·2· · · · ·Q.· Well, yeah, okay.
·3· · · · · · ·MR. KIERNAN:· Move to strike the last
·4· ·part.
·5· ·BY MR. KIERNAN:
·6· · · · ·Q.· So the new iPod owners includes both
·7· ·purchasers of iPods with 7.0 and iPods without 7.0?
·8· · · · ·A.· To differing degrees, yes.
·9· · · · ·Q.· Okay.· And in the rest of the paragraph,
10· ·you state that the "lock-in would not have affected
11· ·the demand for subsequent iPods for a long period
12· ·because these purchasers would not soon make
13· ·repurchase decisions."
14· · · · · · ·What's the basis for that?
15· · · · ·A.· That it's information we have about how
16· ·long people own electronic devices.· They don't buy
17· ·a new electronic device with the same frequency they
18· ·buy music.
19· · · · ·Q.· And that's the 18-month to two-year period
20· ·that you referred to in your initial merits report?
21· · · · ·A.· I think that's those -- I don't remember
22· ·from memory, but that sounds about right as to
23· ·frequency of repurchase of iPods.
24· · · · ·Q.· Okay.
25· · · · ·A.· The mean frequency.· Some are more, some
Page 49
www.aptusCR.com

YVer1f

Roger Noll, Ph.D.

Confidential - Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation

·1· ·are buying replacement iPods are the -- what this is
·2· ·about is the fact that it's mainly people who are
·3· ·buying replacement iPods that experience the
·4· ·immediate effect of lock-in.· They are, to some
·5· ·degree, locked in.· And the point of Harmony was to
·6· ·reduce the degree to which existing users of iPods
·7· ·were locked in.
·8· ·BY MR. KIERNAN:
·9· · · · ·Q.· Okay.· And I think I'm following you now,
10· ·Dr. Noll.· When you're referring to "new iPod
11· ·purchasers" in the last paragraph on page 27, you're
12· ·referring to customers who did not own an iPod
13· ·before that time?
14· · · · ·A.· Yes.· This is new purchasers.· This is not
15· ·replacement purchasers.
16· · · · ·Q.· Okay.· And for a new purchaser in late
17· ·2006, your opinion is for most of the damages
18· ·period, they would not be an important factor
19· ·affecting iPod prices because they wouldn't purchase
20· ·a replacement for the 18- to 24-month period?
21· · · · ·A.· The -- yeah.· Of course, it's not
22· ·dichotomous, it's continuous.· Their importance
23· ·grows through time.· But initially, it would not be
24· ·important because you don't replace your iPod every
25· ·month.· All right.· So it would take a while before
Page 51
www.aptusCR.com

YVer1f

Roger Noll, Ph.D.

Confidential - Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation

·1· ·you start to see an effect.
·2· · · · · · ·There would be other things happening that
·3· ·would cause it to have some effect on demand, like
·4· ·multiple iPods within the same family unit that want
·5· ·access to the same music, things like that.
·6· · · · · · ·But the main immediate effect of lock-in
·7· ·is the existing or established base.· It's not the
·8· ·new people.· And the new people would just gradually
·9· ·through time get added to the people who are
10· ·affected by lock-in in terms of their effect on the
11· ·demand for iPods.
12· · · · · · ·MR. KIERNAN:· Let's just take a short
13· ·break.
14· · · · · · ·THE VIDEOGRAPHER:· Off the record at
15· ·10:05.
16· · · · · · ·(A brief recess was taken.)
17· · · · · · ·THE VIDEOGRAPHER:· On the record at 10:18.
18· ·BY MR. KIERNAN:
19· · · · ·Q.· In the regression that is in Exhibits 3A
20· ·and 3B to Noll 10, you turn on the dummy variable
21· ·for 7.0 at different times for different models; is
22· ·that right?
23· · · · ·A.· Well, the indicator is on for a model that
24· ·has 7.0 on it.· So anything that was released after
25· ·the first date would have it, quote, turned on,
Page 52
www.aptusCR.com

YVer1f

·1· ·BY MR. KIERNAN:
·2· · · · ·Q.· So the existing Harmony sales -- the sales
·3· ·from Harmony that already have completed, they will
·4· ·continue to impact the price of models that did not
·5· ·have 7.0?
·6· · · · · · ·MS. BERNAY:· Objection.· Vague and
·7· ·ambiguous.
·8· · · · · · ·THE WITNESS:· Yes.· I mean, the -- first
·9· ·of all, Harmony isn't sales.· All right.· But
10· ·Harmony is out there on people's computers, and it
11· ·can still be used to load songs onto devices
12· ·regardless of whether RealNetworks continues to
13· ·distribute it.· It's still out there, so it would
14· ·still continue to affect people, people's behavior.
15· ·BY MR. KIERNAN:
16· · · ·

25· ·///

22· · · · ·Q.· Did you do anything to determine whether
23· ·or not 7.0 impacted the demand for iPods that did
24· ·not contain 7.0?
25· · · · · · ·MS. BERNAY:· Objection.

Confidential - Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation

Roger Noll, Ph.D.

1

would be another anticompetitive effect of 7.0

2

because the consumers who bought those things were

3

harmed.

4

because I think the supply conditions in the MP3

5

market would not make it possible for anybody to

6

sustain a significant price increase in that market

7

just because of 7.0.

8

BY MR. KIERNAN:

9

Q.

But I have basically assumed that's zero

Okay.

And is your opinion, Dr. Noll, that

10

the impact on prices caused by 7.0 would be

11

immediate because of the lock-out effect?

12

A.

Yes.

13

Q.

Okay.

14
15

And why is that?

Why would it be

immediate?
A.

Because there's this group of people,

16

there's just a continuous flow of people who want to

17

replace an MP3 player that were available to Apple

18

before 7.0 was introduced and for at least some

19

models are not available to them afterwards.

20
21
22
23

25

MS. BERNAY:

Objection.

www.aptusCR.com

http://www.yeslaw.net/help

24
Vague and
Page 78

Confidential - Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation

Roger Noll, Ph.D.

1

ambiguous.

2
3
4
5
6
7
8
9
10
11
12
13

So I'm not aware of any documents that
explicitly analyze lock-in and lock-out.
Q.

Okay.

So what are you relying on for the

14

basis of your opinion that the impact on price would

15

be immediate due to lock-in or lock-out as a result

16

of 7.0?

17

A.

It would -- why a demand curve becomes

18

more inelastic is not really relevant to the

19

question of what is an increase in the degree of

20

inelasticity of the demand curve on price.

There's

21

no -- it doesn't matter what the cause is.

If the

22

demand curve becomes more inelastic, the profit

23

maximizing price goes up.
(Reporter inquires.)

25

THE WITNESS:

http://www.yeslaw.net/help

24

The profit maximizing price

www.aptusCR.com

Page 79

Confidential - Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation

Roger Noll, Ph.D.

1
2

goes up.
And so the basis for it is the economic

3

theory of lock-in, the optimal pricing of a firm

4

that has a product with lock-in and other switching

5

costs, and then the empirical test of whether that

6

theory predicts what actually happened.

7

BY MR. KIERNAN:

8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
http://www.yeslaw.net/help

24
25

www.aptusCR.com

Page 80

Roger Noll, Ph.D.

Confidential - Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
http://www.yeslaw.net/help

24
25

www.aptusCR.com

Page 81

Roger Noll, Ph.D.

Confidential - Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
http://www.yeslaw.net/help

24
25

www.aptusCR.com

Page 82

Confidential - Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation

Roger Noll, Ph.D.

1

in general, creating incompatibilities caused prices

2

to go up.

3

that the creation of incompatibility causes prices

4

to go up, which is only true if it makes demand

5

curves more inelastic.

6

And there's evidence in all those reports

So anything in any report that discusses

7

the relationship between lock-in and prices is

8

evidence in support that 7.0 did, in fact, cause the

9

demand to become more inelastic.

10

BY MR. KIERNAN:

11

Q.

With respect to the impact on prices on

12

iPods as a result of the lock-out effect, would you

13

expect the impact to be the same throughout the --

14

well, would you expect the impact on prices to

15

remain the same, to remain constant?

16
17

MS. BERNAY:

Objection.

Calls for

speculation.

18

THE WITNESS:

I would expect it to be,

19

yeah, mostly constant through time, although toward

20

the end of the period, it might become worse.

21

But on the other hand, there are other
things in the model that would offset that, such as

23

the move to DRM-free files by competitors of the

24

iTunes Store.

25

///

http://www.yeslaw.net/help

22

www.aptusCR.com

Page 85

Roger Noll, Ph.D.

Confidential - Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation

·1· ·guys.
·2· · · · ·Q.· Right.· Okay.
·3· · · · · · ·(Reporter inquires.)
·4· · · · · · ·THE WITNESS:· The installed base of the
·5· ·other guys.
·6· ·BY MR. KIERNAN:
·7· · · · ·Q.· The installed base for the non-iPods?
·8· · · · ·A.· The non-iPods.
·9· · · · ·Q.· All right.
10· · · · · · ·So focusing on the installed base for
11· ·iPods, would the impact of 7.0 on locking them into
12· ·iPods or increasing the lock-in of those people on
13· ·iPods, that would occur over time depending upon
14· ·their future purchases, correct?
15· · · · ·A.· Yeah.· And it's not just that.· As well,
16· ·it's also -- it doesn't really happen until they buy
17· ·their next device.
18· · · · · · ·So there's not going to be an
19· ·instantaneous effect.· It's going to be a longer
20· ·term effect.· It's the lack of competition for the
21· ·other guys that is the principal short-term effect.
22· · · · ·Q.· I've got to get one update.
23· · · · · · ·Since your last deposition, have you
24· ·purchased any iPods?
25· · · · ·A.· Probably not since my last one.· I bought
Page 109
www.aptusCR.com

YVer1f