"The Apple iPod iTunes Anti-Trust Litigation"

Filing 738

MOTION for Summary Judgment and to Exclude Expert Testimony of Roger G. Noll [REDACTED], filed by Apple Inc.. Motion Hearing set for 2/18/2014 02:00 PM in Courtroom 5, 2nd Floor, Oakland before Hon. Yvonne Gonzalez Rogers. Responses due by 1/13/2014. Replies due by 1/31/2014. (Attachments: # 1 Proposed Order)(Mittelstaedt, Robert) (Filed on 12/21/2013) Modified on 12/23/2013 (jlmS, COURT STAFF).

1
2
3
4
5
6
7
8

Robert A. Mittelstaedt #60359
[e-mail]
Craig E. Stewart #129530
[e-mail]
David C. Kiernan #215335
[e-mail]
JONES DAY
555 California Street, 26th Floor
San Francisco, CA 94104
Telephone:
[phone]
Facsimile:
[phone]
Attorneys for Defendant
APPLE INC.

9

UNITED STATES DISTRICT COURT

10

NORTHERN DISTRICT OF CALIFORNIA

11

OAKLAND DIVISION

12
13
14

THE APPLE iPOD iTUNES ANTITRUST
LITIGATION

15

___________________________________

16

This Document Relates To:

17

ALL ACTIONS

18
19
20

Lead Case No. C 05-00037 YGR
[CLASS ACTION]
DEFENDANT’S NOTICE OF MOTION
AND MOTION FOR SUMMARY
JUDGMENT AND TO EXCLUDE
EXPERT TESTIMONY OF ROGER G.
NOLL
Date:
Time:
Courtroom:

February 18, 2014
2:00 PM
5

[Public Version - Redacted]

21
22
23
24
25
26
27
28
Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

TABLE OF CONTENTS

2

Page

3

NOTICE OF MOTION AND MOTION .......................................................................................... 1 

4

RELIEF SOUGHT ............................................................................................................................ 1 

5

MEMORANDUM OF POINTS AND AUTHORITIES .................................................................. 1 

6

INTRODUCTION ............................................................................................................................ 1 

7

BACKGROUND .............................................................................................................................. 2 

8

A.  Launch of the iTunes Store. ............................................................................................. 2 

9

B.  Plaintiffs’ Original Theory. .............................................................................................. 4 

10

C.  The Amended Consolidated Complaint. .......................................................................... 4 

11

1. 

13

16

The Keybag Verification Code in iTunes 7.0 ........................................................ 5 

D.  Plaintiffs’ Theory of Impact. ............................................................................................ 6 

14
15

iTunes 4.7. .............................................................................................................. 4 

2. 

12

ARGUMENT .................................................................................................................................... 7 
I. 

LEGAL STANDARD. .............................................................................................................. 7 

17

A.  Summary Judgment. ......................................................................................................... 7 

18

B.  Exclusion of Expert Testimony. ....................................................................................... 8 

19

II.  SUMMARY JUDGMENT SHOULD BE GRANTED BECAUSE PLAINTIFFS
CANNOT SHOW ANY GENUINE FACTUAL ISSUE REGARDING IMPACT
OR DAMAGES. ...................................................................................................................... 8 

20
21
22
23
24
25
26
27
28

III.  DR. NOLL’S REGRESSION MODEL IS INADMISSIBLE AND DOES NOT
RAISE A GENUINE FACTUAL ISSUE REGARDING IMPACT OR DAMAGES. ......... 12 
A.  Dr. Noll’s Regression Is Inadmissible Because It Does Not “Fit” The Facts Of
The Case Or The Economics Of Plaintiffs’ Own Theory. ............................................ 12 
B.  The Regression Does Not Properly Account For the Relevant Factors. ........................ 14 
1. 

The regression does not properly account for the lawful effect of iTunes
4.7 and thus uses the wrong “but-for” world. ...................................................... 14 

2. 

The regression does not account for the impact of the unchallenged
aspects of the
............................................................................................. 16 
i

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1
2
3

3. 

The regression does not account for major product characteristics and
other factors relevant to price............................................................................... 18 

C.  The Regression Does Not Provide Statistically Significant Results. ............................. 20 

5

IV.  DR. NOLL’S RELEVANT MARKET OPINIONS SHOULD BE EXCLUDED
AND SUMMARY JUDGMENT GRANTED FOR FAILURE OF PROOF ON
THAT ISSUE. ........................................................................................................................ 21 

6

CONCLUSION ............................................................................................................................... 25 

4

7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
ii

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

TABLE OF AUTHORITIES

2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27

Page(s)
Cases 
A&M Records, Inc. v. Napster, Inc.,
114 F. Supp. 2d 896 (N.D. Cal. 2000), aff’d 239 F.3d 1004 (9th Cir. 2001)............................... 3
American Booksellers Ass'n, Inc. v. Barnes & Noble, Inc.,
135 F. Supp. 2d 1031 (N.D. Cal. 2001) ..................................................................................... 13
Argus Inc. v. Eastman Kodak Co.,
801 F.2d 38 (2d Cir. 1986) ........................................................................................................... 8
Bazemore v. Friday,
478 U.S. 385 (1986) ................................................................................................................... 20
Blomkest Fertilizer, Inc. v. Potash Corp. of Saskatchewan,
203 F.3d 1028 (8th Cir. 2000).................................................................................................... 18
Brown Shoe Co. v. United States,
370 U.S. 294 (1962) ................................................................................................................... 22
City of Vernon v. S. Cal. Edison Co.,
955 F.2d 1361 (9th Cir. 1992).................................................................................................... 17
Concord Boat Corp. v. Brunswick Corp.,
207 F.3d 1039 (8th Cir. 2000).......................................................................................... 7, 12, 17
Daubert v. Merrell Dow Pharms.,
43 F.3d 1311 (9th Cir. 1995)........................................................................................................ 8
Daubert v. Merrell Dow Pharms., Inc.,
509 U.S. 579 (1993) ......................................................................................................... 8, 12, 13
El Aguila Food Prods., Inc. v. Gruma Corp.,
301 F. Supp. 2d 612 (S.D. Tex. 2003) aff’d, 131 F. App’x 450 (5th Cir. 2005) ....................... 14
Freeland v. AT&T Corp.,
238 F.R.D. 130 (S.D.N.Y. 2006) ......................................................................................... 19, 20
General Elec. Co. v. Joiner,
522 U.S. 136 (1997) ..................................................................................................................... 8
In re eBay Seller Antitrust Litig.,
433 Fed. Appx. 504 (9th Cir. 2011) ............................................................................................. 7
In re Graphics Processing Units Antitrust Litig.,
253 F.R.D. 478 (N.D. Cal. 2008) ........................................................................................... 8, 18

28
iii

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

In re Live Concert Antitrust Litig.,
863 F. Supp. 2d 966 (C.D. Cal. 2012) ........................................................................... 18, 22, 24
In re Methionine Antitrust Litig.,
MDL No. 00-1311, 2003 U.S. Dist. LEXIS 14828 (N.D. Cal. Aug. 22, 2003)........................ 18
In re REMEC Inc. Sec. Litig.,
702 F. Supp. 2d 1202 (S.D. Cal. 2010) ...................................................................................... 18
In re Se. Milk Antitrust Litig.,
No. 08-MD-1000, 2012 WL 1032797 (E.D. Tenn. Mar. 27, 2012)....................................... 7, 17
Kline v. Coldwell, Banker & Co.,
508 F.2d 226 (9th Cir. 1974)........................................................................................................ 8
Ky. Speedway, LLC v. Nat’l. Assoc. of Stock Car Auto Racing, Inc.,
588 F.3d 908 (6th Cir. 2009)................................................................................................ 23, 25
McGlinchy v. Shell Chem. Co.,
845 F.2d 802 (9th Cir. 1988)...................................................................................................... 14
McLaughlin Equip. Co., Inc. v. Servaas,
2004 WL 1629603 (S.D. Ind. Feb. 18, 2004) ............................................................................ 24
Menasha Corp. v. News Am. Mktg. In-Store, Inc.,
354 F.3d 661 (7th Cir. 2004)...................................................................................................... 22
Mukhtar v. Cal. State Univ., Hayward,
299 F.3d 1053 (9th Cir. 2002) amended, 319 F.3d 1073 (9th Cir. 2003) .................................... 8
Oltz v. St. Peter’s Cmty. Hosp.,
861 F.2d 1440 (9th Cir. 1988).................................................................................................... 21
Rebel Oil Co. v. Atl. Richfield Co.,
51 F.3d 1421 (9th Cir. 1995)........................................................................................................ 8
Reifert v. S. Cent. Wisc. MLS Corp.,
450 F.3d 312 (7th Cir. 2006)...................................................................................................... 22
SMS Sys. Maint. Servs., Inc. v. Digital Equip. Corp.,
188 F.3d 11 (1st Cir. 1999) ........................................................................................................ 24
So. P. Commc’ns Co. v. Am. Tel. & Tel. Co.,
556 F. Supp. 825 (D.D.C. 1982) ................................................................................................ 13
Spectrum Sports, Inc. v. McQuillan,
506 U.S. 447 (1993) ................................................................................................................... 21
United Food Mart, Inc. v. Motiva Enters., LLC,
404 F. Supp. 2d 1344 (S.D. Fla. 2005) ...................................................................................... 24
iv

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1
2
3
4
5
6
7

Va. Vermiculite Ltd. v. W.R. Grace & Co.-Conn.,
98 F. Supp. 2d 729 (W.D. Va. 2000) ................................................................................... 23, 24
Va. Vermiculite, Ltd. v. W.R. Grace & Co.-Conn.,
108 F. Supp. 2d 549 (W.D. Va 2000) ........................................................................................ 25
Rules 
Federal Rules of Civil Procedure
Rule 56 ......................................................................................................................................... 1
Rule 56(c)(2) ................................................................................................................................ 7

8

Federal Rules of Evidence
Rule 702 ....................................................................................................................................... 8

9

Treatises 

10

American Bar Association, Proving Antitrust Damages: Legal and Economic Issues (2d
ed. 2010)......................................................................................................................... 18, 20, 21

11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

v

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1
2

NOTICE OF MOTION AND MOTION
PLEASE TAKE NOTICE that on February 18, 2014 at 2:00 p.m. defendant Apple Inc.

3

will and hereby does move for summary judgment and to exclude the expert testimony of Roger

4

G. Noll.

5
6
7

RELIEF SOUGHT
Apple requests that summary judgment be entered in Apple’s favor on all claims and that
the testimony of plaintiffs’ expert, Roger G. Noll, be excluded.

8

MEMORANDUM OF POINTS AND AUTHORITIES

9

INTRODUCTION

10

This motion seeks to resolve the last remaining claim in this long-running antitrust

11

lawsuit. Plaintiffs’ principal claims have already been rejected. The record, with fact and expert

12

discovery now complete, indisputably establishes that plaintiffs cannot prove their remaining,

13

implausible assertion that an Apple software update in 2006 increased iPod prices.

14

Plaintiffs’ claim relies on an incoherent chain of purported events. Plaintiffs’ theory is

15

that the 2006 update, which prevented digital music from one insignificant source of music

16

(RealNetworks music store (RMS)) from being played directly on iPods, forced RMS customers

17

to buy music from Apple instead of RMS to play on their iPods, and they bought so much music

18

from Apple that if and when they needed to replace their iPods, they were locked in to buying an

19

iPod rather than switching to a competing device as they supposedly would have done if only

20

they had continued to buy music from RMS. Plaintiffs theorize that this chain of events happened

21

so often that demand for iPods increased to such an extent that Apple was able to, and did, charge

22

higher prices for iPods than it otherwise would have.

23

Apart from the implausibility of this theoretical chain, plaintiffs have no proof of any part

24

of it. They have no evidence that anyone, much less a sizeable group of consumers, switched

25

from RMS to Apple’s music store as a result of the update, or that if anyone did, the incremental

26

amount of music they bought from Apple forced them to buy a replacement iPod even though

27

they preferred a competing device. The named plaintiffs do not fit that description. They have

28

identified no one who does. And they have not offered any survey of consumers to support their
1

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

theory. Nor do they have any evidence that Apple’s pricing of iPods turned on this purported

2

chain of events.

3

Indeed, the record evidence refutes key parts of plaintiffs’ theory. As their expert

4

admitted, RealNetworks was a minor player, with minimal sales at the time of Apple’s software

5

update.

6
7

Thus, the non-expert evidence establishes that summary

8

judgment is warranted because plaintiffs cannot show any genuine dispute of material fact that

9

Apple raised its iPod prices due to the update.

10

Plaintiffs’ economist, Dr. Noll, and his regression model cannot salvage their claim. The

11

regression purports to tease out a supposed incremental price effect of the 2006 update from the

12

multitude of factors that determine iPod pricing. But the regression is riddled with fundamental

13

errors and inconsistencies that render it unreliable and unable to support any finding of impact.

14

As shown below, the model is based on unsupported assumptions and produces results that are

15

irreconcilable with indisputable real-world facts. And Dr. Noll miscalculates the standard errors

16

to contend that his results are statistically significant. When the standard errors are properly

17

calculated, they show that the regression results are not significantly different from zero—which

18

is the expected result given the lack of any evidence that the 2006 update had any impact on

19

demand for iPods or that Apple considered it in determining iPod prices.

20

For these reasons, summary judgment should be granted in favor of Apple, and Dr. Noll’s

21

report and testimony should be excluded. Summary judgment should also be granted because

22

Dr. Noll has not come close to conducting the required analysis to support his opinions regarding

23

the relevant antitrust markets and plaintiffs have no other evidence to support their market

24

definition. Defining a proper relevant market is required for a valid claim of monopolization.

25

But plaintiffs and Dr. Noll again rely on ipse dixit, unsupported by relevant analysis.

26

BACKGROUND

27

A.

28

Apple launched the iTunes Store (“iTS”) in April 2003. As the then-head of the

Launch of the iTunes Store.

2

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

Department of Justice Antitrust Division observed in 2006, the iTS “solved a problem that some

2

observers . . . predicted might never be solved: how to create a consumer friendly, yet legal and

3

profitable, system for downloading music and other entertainment from the Internet.” Declara-

4

tion of Amir Amiri (“Amiri Decl.”) filed herewith, Ex. 1 at p. 6. In the years preceding the iTS’s

5

opening, music piracy had been rampant, with consumers using illegal peer-to-peer file sharing

6

sites like Napster to download music for free. The scale of this piracy was enormous, with as

7

many as 10,000 files being shared per second on Napster. See generally A&M Records, Inc. v.

8

Napster, Inc., 114 F. Supp. 2d 896, 902 (N.D. Cal. 2000), aff’d 239 F.3d 1004 (9th Cir. 2001).

9

The iTS was developed as a legal alternative to these illegal file-sharing sites. To

10

succeed, Apple had to (1) make the store so attractive to consumers and so easy to use that it

11

could compete with free music services, and (2) satisfy the recording industry’s concerns that

12

licensed online distribution would lead to more piracy and theft and further erode the labels’

13

sales. ECF No. 538, ¶¶ 3-5; ECF No. 470, Ex. 3, 33:20-22.

14

Between November 2002 and April 2003, Apple and the major record labels (Sony, BMG,

15

EMI, Warner, and Universal) entered agreements that permitted Apple to distribute music

16

through the iTS. ECF No. 538, ¶ 5.

17
18
19
20
21

As plaintiffs and their expert admit, the iTS was “procompetitive” and a “huge benefit”

22

with “enormous advantages” to consumers. ECF No. 176, Ex. 21 at 105:8-20; ECF No. 322,

23

¶¶ 14-15, 40. Fortune Magazine named the iTS its 2003 “product of the year,” observing that

24

“[w]ith the success of its iTunes Music Store, Apple is almost single-handedly dragging the music

25

industry, kicking and screaming, toward a better future.” Amiri Decl, Ex. 2.

26

The labels continued to demand that all licensees of their music use DRM for several

27

years after the iTS was launched. The first major label to drop the requirement was EMI, which

28

permitted Apple to sell music without DRM beginning in early 2007. See Id., Ex. 3 at Ex. 1. By
3

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

early 2008, all of the major labels permitted Amazon to sell their music without DRM. Id.

2

Finally, all the major labels dropped the DRM requirement for Apple in early 2009. In January

3

2009, Apple announced that about 8 million songs (or 80% of the music sold on the iTS) would

4

be available DRM-free from the iTS. Id. All songs on the iTS were available DRM-free

5

beginning in March 2009. Id.

6

B.

7

Plaintiffs brought this suit in 2005—when the labels were still requiring Apple and others

Plaintiffs’ Original Theory.

8

to use DRM—alleging that Apple’s use of FairPlay was an antitrust violation. They asserted that,

9

as a consequence of Apple’s use of its own proprietary system, music purchased from the iTS was

10

directly playable on iPods but not on music players made by other companies. (iTS music could

11

be “indirectly” played on competing devices by burning the music to a CD and importing, or

12

“ripping,” it to a computer.) They argued that Apple had thereby locked iTS purchasers into

13

buying iPods rather than competing players in the sense that their “switching costs” to change to a

14

non-iPod device would be greater.

15

Rejecting that claim, this Court held that it was lawful for Apple to adopt its own DRM,

16

even if the effect was to make Apple’s products incompatible with competitors’ products. ECF

17

No. 303, pp. 2, 10; ECF No. 274, p. 10.

18

C.

19

Plaintiffs responded by filing an amended complaint (ECF No. 322) with a new theory.

20

Having failed in challenging Apple’s decision to use a proprietary DRM, they now challenged

21

later updates to Apple’s software that they claim excluded competitors. ECF No. 322, ¶ 52.

22
23

The Amended Consolidated Complaint.

1.

iTunes 4.7.

The amended complaint focused on a feature in iTunes 4.7 issued in October 2004 to

24

enhance the FairPlay encryption system. The update fixed a vulnerability in FairPlay that had

25

been exploited by hacks that stripped DRM from iTS music.

26
27
28

Apple did so by issuing iTunes 4.7, which changed the FairPlay
4

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

encryption method and disabled the hacks.

2

Plaintiffs alleged that iTunes 4.7 also unlawfully shut down RealNetworks’ “Harmony”

3

technology. As plaintiffs explained it, RealNetworks had “analyzed the firmware within the

4

iPod” and “discern[ed]” Apple’s “software code.” ECF No. 322, ¶ 53. By cracking FairPlay,

5

RealNetworks devised a way to make music purchased from its website mimic Apple’s then-

6

existing encryption scheme and make it falsely appear to the iPod that the RealNetworks-

7

protected music was actually iTS music protected by FairPlay. ECF No. 538, ¶¶ 50-54.

8

RealNetworks released this “Harmony” technology in July 2004 and promoted it by selling

9

Harmony music at half-price (49 cents a song) for three weeks. When Apple released iTunes 4.7,

10

the update disabled Harmony as to the iPods with the update, because Harmony mimicked the

11

previous encryption method, not the updated one.

12

This Court (Ware, J.,) granted summary judgment to Apple on plaintiffs’ challenge to

13

iTunes 4.7. ECF No. 627. The Court ruled that iTunes 4.7 was a valid product improvement

14

because—notwithstanding any effect it had on Harmony—it restored FairPlay’s integrity,

15

Id. at 8.

16

2.

17

The Keybag Verification Code in iTunes 7.0

Plaintiffs also challenged an aspect of the iTunes 7.0 update issued in September 2006.

18

That update included a number of enhancements. For example, it enabled the sale of movies for

19

the first time from the iTS, supported video at four times higher resolution than the previous

20

version, and introduced Apple’s distinctive “Cover Flow” method to browse through content on

21

the iPod. See Amiri Decl., Ex. 4 at Ex. 6. iTunes 7.0

22
23
24
25

1

26
27
28
5

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1
2
3

It also made the DRM more secure by eliminating one avenue of attack. Id.

4

As they did in challenging iTunes 4.7, plaintiffs alleged that the

was anticompetitive

5

because it prevented music from other music stores from directly playing on certain iPods. In

6

particular, plaintiffs argue that the updates again disabled RealNetworks’ Harmony, which

7

RealNetworks had re-launched in April 2005 after again cracking Apple’s DRM software.

8
9

Judge Ware denied Apple’s motion for summary judgment on this issue, finding that there
were issues of fact regarding whether the

was truly a product improvement. ECF No. 627,

10

pp. 11-12.2 Apple did not seek summary judgment on lack of impact or competitive injury at the

11

time because discovery was still ongoing. This Court’s scheduling orders explicitly preserve

12

Apple’s right to seek summary judgment on that issue after the completion of discovery.

13

D.

14

Plaintiffs’ theory of antitrust impact is that, by disabling Harmony, the

Plaintiffs’ Theory of Impact.
resulted in

15

some consumers purchasing more music from iTS than they otherwise would have. They assert

16

this increased the switching costs of those consumers to the point that they were locked in to

17

buying an iPod when they next purchased a digital music player. According to plaintiffs, that

18

increased demand for iPods led Apple to raise iPod prices. At the same time, they say, other

19

customers were locked out of buying iPods after the

20

Harmony music that would no longer play on the iPods on which the

21

According to plaintiffs, this lock-out led to higher prices because Apple no longer had the

22

incentive to reduce prices to attract these customers. Amiri Decl., Ex. 8 at p. 7.

23
24
25
26
27
28

because they had large libraries of
was implemented.

As shown below, plaintiffs offer no documentary evidence or testimony from percipient
witnesses to establish any of the facts necessary to support this theory or to refute the record
2

In addition to iTunes 7.0, plaintiffs’ experts also describe iTunes 7.4, an update released in
September 2007, as preventing Harmony from working on iPods. But they do not present a
separate impact or damages analysis related to that update. Instead, their position is that, because
the
had already disabled Harmony as of September 2006, iTunes 7.4 had no additional
adverse effect on iPod prices.
6

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

evidence contradicting it. Instead, they rely solely on a regression model proffered by their

2

economist, Roger Noll. The regression purports to perform a before and after analysis, compare-

3

ing iPod prices before the

4

that influence iPod prices, finding the amount of iPod prices that plaintiffs say is attributable to

5

the

with those after and, after supposedly controlling for other factors

6
7
8
9
10
11
12
13
14
15

ARGUMENT
I.

16

LEGAL STANDARD.
A.

17

Summary Judgment.

Summary judgment is proper when “there is no genuine issue as to any material fact and

18

. . . the movant is entitled to judgment as a matter of law.” Fed. R. Civ. P. 56(c)(2). Courts

19

frequently grant summary judgment in antitrust cases when the plaintiff lacks admissible

20

evidence that the challenged conduct caused antitrust injury—including where, as here, the

21

plaintiff seeks to prove injury though expert economic evidence that the court finds inadmissible.3

22
23
24
25
26
27
28

3

See e.g., In re Se. Milk Antitrust Litig., No. 08-MD-1000, 2012 WL 1032797, at *4–6
(E.D. Tenn. Mar. 27, 2012) (granting summary judgment where the expert’s model could not
identify price changes caused by alleged anti-competitive conduct); In re eBay Seller Antitrust
Litig., 433 Fed. Appx. 504, 506 (9th Cir. 2011) (affirming summary judgment where expert’s
econometric analysis was inadequate to support claim of antitrust injury); Concord Boat Corp. v.
Brunswick Corp., 207 F.3d 1039, 1054-58, 1063 (8th Cir. 2000) (finding that expert’s opinions on
injury should have been excluded and remanding for entry of judgment).
7

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

B.

Exclusion of Expert Testimony.

2

Rule 702 permits expert testimony that “will help the trier of fact to understand the

3

evidence or to determine a fact in issue” if the testimony is “based on sufficient facts or data,” “is

4

the product of reliable principles and methods,” and “the expert has reliably applied the principles

5

and methods to the facts of the case.” Fed. R. Evid. 702. As “gatekeeper,” the Court must ensure

6

that expert testimony “rests on a reliable foundation and is relevant to the task at hand.” Daubert

7

v. Merrell Dow Pharms., Inc., 509 U.S. 579, 597 (1993). “Maintaining Daubert’s standards is

8

particularly important considering the aura of authority experts often exude, which can lead juries

9

to give more weight to their testimony.” Mukhtar v. Cal. State Univ., Hayward, 299 F.3d 1053,

10

1063-64 (9th Cir. 2002) amended, 319 F.3d 1073 (9th Cir. 2003). Litigants should not be

11

permitted to get to trial simply by attempting to “dazzle the courtroom with graphs and tables.”

12

In re Graphics Processing Units Antitrust Litig., 253 F.R.D. 478, 491-92 (N.D. Cal. 2008).

13

Reliable expert testimony is grounded “in the methods and procedures of science,” not

14

“subjective belief or unsupported speculation.” Daubert, 509 U.S. at 590. While the focus is on

15

the expert’s underlying methodology, “conclusions and methodology are not entirely distinct

16

from one another,” and “[a] court may conclude that there is simply too great an analytical gap

17

between the data and the opinion proffered.” GE v. Joiner, 522 U.S. 136, 146 (1997). Relevant

18

expert testimony “‘fit[s]’” the facts of the case and “logically advances a material aspect of the

19

proposing party’s case.” Daubert v. Merrell Dow Pharms., 43 F.3d 1311, 1315 (9th Cir. 1995).

20

II.

SUMMARY JUDGMENT SHOULD BE GRANTED BECAUSE PLAINTIFFS

21

CANNOT SHOW ANY GENUINE FACTUAL ISSUE REGARDING IMPACT OR

22

DAMAGES.

23

“[P]roof of injury is an essential substantive element” of an antitrust claim. Kline v.

24

Coldwell, Banker & Co., 508 F.2d 226, 233 (9th Cir. 1974); Rebel Oil Co. v. Atl. Richfield Co.,

25

51 F.3d 1421, 1433 (9th Cir. 1995) (“causal antitrust injury[] is an element of all antitrust suits”).

26

Lack of evidence of injury caused by the challenged conduct is “fatal to the merits of any antitrust

27

claim.” Argus Inc. v. Eastman Kodak Co., 801 F.2d 38, 41 (2d Cir. 1986). That is the situation

28

here. Plaintiffs’ claim of impact is unsupported by any record facts.
8

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

For the alleged impact to have occurred, RealNetworks must have been selling Harmony

2

music to iPod owners (or potential iPod purchasers) in sufficient quantities in September 2006

3

that a drop in those sales would have affected iPod demand so much that Apple would raise iPod

4

prices. That, in turn, requires that there were consumers who, because of the update, became

5

either (1) locked in by their increased iTS purchases and thus forced to buy an iPod when they

6

otherwise would have bought a competing player, or (2) locked out by their Harmony purchases

7

and thus prevented from purchasing any iPod with

8

became aware of these customers (and how many there were) and took them into account in

9

setting its prices before it sold any iPods with the

on it. And the claim requires that Apple

on it and before it could possibly know

10

whether

11

links in this tortured chain of proof, but the record is fundamentally at odds with it.

12
13

would impact demand. Not only do the undisputed facts fail to support any of the

First, Harmony was insignificant in 2006. Its share of music sales was less than 3% when
Apple released iTunes 7.0. Id., Ex. 4, ¶ 27, n.40.

14
15

But the proof

16

plaintiffs must adduce is what Harmony sales were in 2006. Earlier in the case, Dr. Noll admitted

17

that he would not expect many consumers to be using Harmony in 2005 or 2006 because it had

18

already been lawfully disabled by iTunes 4.7 and many consumers likely permanently abandoned

19

it at that point. Id., Ex. 9 at 147:20-148:5. And the best Dr. Noll can say now is that Harmony’s

20

share of the market in 2006 was less than 5%—“I don’t remember the exact number. I know that

21

it was small, yes.” Id., Ex. 10 at 80:22-81:5. The suggestion that Harmony could set off the

22

chain of events outlined by plaintiffs is not simply implausible, but unsupported by the record.

23

Second, plaintiffs have no evidence regarding what portion of RealNetworks’ small sales

24

was to iPod owners or potential iPod purchasers. As a threshold matter, to the extent RealNet-

25

works sold to consumers who did not own an iPod or had no intention of buying an iPod, those

26

sales could not create any relevant lock-in or lock-out. As Dr. Noll admitted, in determining

27

whether the

28

Harmony to play stuff on an iPod.” Id., Ex. 10 at 81:6-12. But plaintiffs have no proof that any

had any impact, the relevant sales are to the “fraction of people who used

9

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

iPod users or potential users were buying Harmony music, let alone enough of them to affect iPod

2

demand. As Dr. Noll admitted: “Q: Do you have any information or any estimate on how many

3

iPod users bought music from RealNetworks? . . . A: No.” Id. at 116:18-23.

4

Third, plaintiffs have no proof of the number of people who became locked in or locked

5

out as a result of the

6

was a potential iPod purchaser. Under Dr. Noll’s lock-in theory, if that purchaser already owned

7

a large iTS music library that he or she wanted to continue to play on a portable device, the

8

purchaser would already be locked in by that library to buying iPods regardless of any new

9

Harmony purchases. Similarly, any lock-in from increased iTS purchases would be irrelevant for

. It is not enough merely that a Harmony purchaser owned an iPod or

10

customers who would have bought an iPod regardless, simply because of its design, ease of use

11

and revolutionary features. As Dr. Noll admitted, plaintiffs’ theory requires that there have been

12

consumers whose decision about whether to purchase an iPod was changed as a result of the

13
14
15

. But Dr. Noll testified that he did not know how many people fell into that category: “We
have no way of knowing that.” Id. at 107:16-24.
Nor do plaintiffs otherwise offer any such evidence. Although the existence of such

16

customers is pivotal to their claim, plaintiffs have not conducted any consumer survey or other

17

investigation to try to identify these customers or determine how many might exist. Indeed, they

18

do not have evidence of, or testimony from, even a single customer (let alone enough customers

19

to affect prices) who says he or she (1) wanted to purchase a competing player rather than an

20

iPod, but (2) switched to buying music from iTS instead of RealNetworks as a result of the

21

update, and (3) thereafter bought enough additional iTS music that he or she was forced to buy an

22

iPod instead of a competing player. Nor can they identify any RealNetworks customers who

23

would have purchased an iPod but were locked out of doing so because of their Harmony

24

purchases. Even the named plaintiffs do not allege that they were locked in or out as a result of

25

. None of them asserts that he or she purchased or would have purchased music from

26

Harmony and either bought or did not buy an iPod as a result of not being able to play Harmony

27

music on that iPod.

28

Fourth, despite extensive discovery related to Apple’s pricing committee, plaintiffs
10

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

cannot identify any evidence that Apple ever took into account the amount of sales from

2

RealNetworks or any other online store in setting iPod prices. Plaintiffs’ theory is that Apple

3

increased its prices because the

4

See e.g., Amiri Decl, Ex. 8, p. 27.

supposedly increased the inelasticity of demand for iPods.

5
6
7
8
9
10
11
12
13
14

For prices to have been affected in the manner plaintiffs assert,

15

Apple not only would have had to take the

16

possibly known what effect it would have, if any. Given the enormous volume of iPod sales at

17

the time, and all of the factors that influence prices, it is complete speculation, at best, to say that

18

the pricing committee’s decision would have been influenced by the kind of unknown, miniscule,

19

and theoretical change in demand elasticity plaintiffs hypothesize.

into account but do so before it could have

20
21
22
23
24
25
26
27

This

28

finding defies reality. Plaintiffs have no evidence that Apple ever priced its iPods in that fashion.
11

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

Over the years, Apple has obviously experienced changes in demand for its iPods that are much

2

greater than anything that could have resulted from Harmony being disabled. And yet it has

3

consistently adhered to its practice of setting only aesthetic prices—and not making small

4

incremental price changes in the manner a service station might change its gasoline prices. No

5

basis exists for saying it would have abandoned that longstanding practice if only Harmony had

6

continued operating.

7

In short, plaintiffs’ claim of impact is unrooted in any real-world evidence. Plaintiffs have

8

no documentary or percipient witness testimony to show even the most basic facts necessary for

9

their claim to be valid. This case has been pending for eight years. It has been over seven years

10

since Apple issued the challenged updates. If there were any evidence supporting their

11

implausible theory, Plaintiffs have had full opportunity to gather it. Their complete failure of

12

proof requires that summary judgment be granted.

13

III.

DR. NOLL’S REGRESSION MODEL IS INADMISSIBLE AND DOES NOT

14

RAISE A GENUINE FACTUAL ISSUE REGARDING IMPACT OR DAMAGES.

15

With no direct evidence supporting their claim of impact on prices, plaintiffs rely on a

16

regression model to try to estimate both impact and the amount of damages. But that model does

17

not cure plaintiffs’ failure of proof. It rests on unsupported assumptions and reaches results that

18

conflict with real-world evidence. And it is improperly engineered through selective choice of

19

variables and manipulation of data to misleadingly create the appearance of an impact where none

20

exists. It is precisely the kind of “quite misleading” expert evidence that Daubert warns should

21

not be given to the jury. Daubert, 509 U.S. at 595.

22
23
24

A.

Dr. Noll’s Regression Is Inadmissible Because It Does Not “Fit” The Facts Of
The Case Or The Economics Of Plaintiffs’ Own Theory.

Damage models that are premised on unsupported assumptions or assumptions that are at

25

odds with the facts of the case or with plaintiffs’ own theory are inadmissible. See Concord

26

Boat, 207 F.3d at 1055-57 (expert’s economic model “should have been excluded” where it

27

“construct[ed] a hypothetical market which was not grounded in the economic reality of the stern

28

drive engine market, for it ignored inconvenient evidence”); So. P. Commc’ns Co. v. Am. Tel. &
12

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

Tel. Co., 556 F. Supp. 825, 1077, 1079 (D.D.C. 1982) (concluding that “plaintiffs’ model of the

2

‘but for’ world show[ed] no damages” where there was “obvious inconsistency between real

3

world competitive conditions and [plaintiffs’] assumptions about competition in the imaginary

4

‘but for’ world”); Daubert, 509 U.S. at 591 (testimony must “fit” the facts of the case).

5

That is the case here, and is the threshold reason Dr. Noll’s regression should be excluded.

6

As shown in the preceding section, the price effect Dr. Noll’s model purports to find is

7

unsupported by any real-world evidence regarding the amount of Harmony sales, the purchasers

8

of Harmony music, the degree of any resulting lock-in or lock-out, or any awareness or considera-

9

tion by Apple’s pricing committee of any effect from the

.

10
11
12
13
14
15
16
17

It pre-

18

dicts that Apple would have changed its prices across the board to every customer. No evidence

19

supports that implausible assertion. In American Booksellers Ass’n, Inc. v. Barnes & Noble,

20

Inc.,135 F. Supp. 2d 1031 (N.D. Cal. 2001), the court refused to admit plaintiffs’ expert’s

21

economic model because it “contain[ed] entirely too many assumptions and simplifications that

22

[were] not supported by real-world evidence.” Am. Booksellers Ass’n v. Barnes & Noble, 135 F.

23

Supp. 2d 1031, 1041-42 (N.D. Cal. 2001). Among other flaws, the model did not “take into

24

account the actual retail pricing policies of plaintiffs or defendants and … simply assume[d] that

25

prices will be set in a way that maximizes profits.” Id. at 1040. Dr. Noll’s model is flawed for

26

the same reason.

27
28

Dr. Noll’s regression is also inconsistent with his own economic theory. The model finds
a single, constant percentage overcharge on every iPod sold during the damage period with
13

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

Amiri Decl., Ex. 8 at Exs. 5-6. If his theory were correct, however, the lock-in would have

2

progressively increased during the damage period as consumers purchased more iTS music.

3

Dr. Noll offers no coherent theory how the alleged overcharge amount could possibly remain

4

constant in these circumstances. In his rebuttal report, Dr. Noll suggests that lock-out from the

5

would affect iPod prices immediately. Id. at p. 27. Even if that were true, however, Dr.

6

Noll continues to assert that other customers would be locked in to buying iPods, that the extent

7

of lock-in would increase over time, and that the effect of lock-in is to “reduce[] the intensity of

8

price competition among” music players. Id. at n.7. He does not explain how, under these

9

circumstances, the price effect he postulates would remain constant over the entire two-year

10
11

damage period. Id., Ex. 11 at 85:11-85:24.
Because it rests on unsupported assumptions that conflict with the real world, Dr. Noll’s

12

regression must be excluded. See McGlinchy v. Shell Chem. Co., 845 F.2d 802, 807 (9th Cir.

13

1988) (affirming exclusion of damages study that rested on “unsupported assumptions”); El

14

Aguila Food Prods., Inc. v. Gruma Corp., 301 F. Supp. 2d 612, 625 (S.D. Tex. 2003) aff’d, 131

15

F. App’x 450 (5th Cir. 2005) (excluding damages model that “ignore[d] market realities”).

16

B.

17

A valid regression must account for the factors that may influence the characteristic being

The Regression Does Not Properly Account For the Relevant Factors.

18

measured—here, the price of iPods. Failure to do so can bias the results, by attributing an effect

19

to one event or characteristic that actually belongs to another. Dr. Noll’s regression fails to

20

satisfy this basic requirement in at least three separate respects, each of which independently

21

requires that his testimony be excluded.

22

1.

The regression does not properly account for the lawful effect of

23

iTunes 4.7 and thus uses the wrong “but-for” world.

24

Dr. Noll’s regression includes variables for events that he contends may have affected

25

iPod prices. One of these variables is his iTunes 7.0 variable, which he claims measures the

26

effect of Harmony being disabled in 2006. Other variables include such things as the opening of

27

the iTS in April 2003, the original Harmony launch in July 2004, the release of iTunes 4.7 in

28

October 2004, and the record labels’ elimination of the DRM requirement for Apple’s compete14

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

tors in early 2008. Dr. Noll includes these other variables because, if plaintiffs’ theory were

2

valid, he believes these events would be expected to affect iPod prices in the same manner

3

plaintiffs say the

4

iTunes 4.7 (which he calls “Harmony Blocked”), Dr. Noll turns the variable “on” (i.e., sets its

5

value to 1 in the equation) on the date of the event and leaves it on for the rest of the class period.

6

did. For each of these variables except the variable for the release of

Dr. Noll treats iTunes 4.7 differently. In his initial report, Dr. Noll turned the iTunes 4.7

7

variable off as of September 12, 2006, the day the

8

Dr. Noll admits that a variable should be included if it represents a factor that might plausibly

9

affect prices. ECF No. 685, p. 9. Thus, by turning the iTunes 4.7 variable off, his regression

was released. Amiri Decl., Ex. 4 at ¶ 90.

10

treated iTunes 4.7 as having no plausible effect on prices after the

11

treats the but-for world as one in which iTunes 4.7 did not exist. Amiri Decl., Ex. 4 at ¶¶ 94-95.

12

was released—and thus

That treatment was inconsistent with plaintiffs’ theory and with Dr. Noll’s own admis-

13

sions. Plaintiffs’ theory is that, by disabling Harmony in 2004 and allegedly increasing iTS sales,

14

iTunes 4.7 caused iPod prices to increase (albeit lawfully given this Court’s ruling that iTunes 4.7

15

was lawful (ECF No. 627, p. 11)). And Dr. Noll admitted that, under plaintiffs’ theory, this

16

alleged effect could persist even after the

17

Harmony in 2004 could have caused consumers to permanently abandon Harmony. Amiri Decl.

18

Ex. 10 at 68:2-70:16 (“Q: And could that consumer expectation continue even after 7.0 is issued?

19

A: Exactly, it could.”). For such consumers, it was iTunes 4.7

20

fewer purchases from RealNetworks and whatever effect that may have had on iPod prices. By

21

nonetheless turning off the iTunes 4.7 variable when he turned on the iTunes 7.0 variable, Dr.

22

Noll caused the iTunes 7.0 variable to capture this continuing effect from iTunes 4.7. See id., Ex.

23

3 at ¶102; Ex. 4 at ¶¶ 92-93.

24

was issued because the initial shutdown of

) that resulted in

In his rebuttal report, Noll responds by revising his regression to leave on the iTunes 4.7

25
26
27
28

This, however, does not address the problem,
15

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

because the persistence effect he admits could have existed applies to all iPods, not just those

2

unaffected by 7.0. As discussed, wholly apart from the

3

permanently dissuaded from using Harmony once it was disabled by iTunes 4.7, as Dr. Noll has

4

admitted. This effect does not depend on which iPod the customer owns or whether the

5

implemented on it. It is an effect on consumers, resulting from the earlier disabling of Harmony,

6

not from the iPod they later purchase.

7
8
9

, some customers may have been

was

In his latest model, Dr. Noll then compounds this error by not turning on the iTunes 7.0
variable for those iPod models on which

was not implemented. By doing so, he treats

as not having any possible effect on demand for those models. But that treatment is

10

contrary to Dr. Noll’s own assertion that “[t]he effect on prices [from iTunes 7.0] is not

11

necessarily limited to just the products that were sold that had 7.0 in them.” Id., Ex. 10 at 46:2-4.

12

According to Dr. Noll, that is because disabling Harmony in 2006 was a “market-defining event.”

13

Id., Ex. 10 at 48:24-49:17. Thus, contrary to what he does in current model, Dr. Noll previously

14

admitted that turning on the iTunes 7.0 variable for all models is “the right way to do it.” Id., Ex.

15

10 at 49:14.

16

The effect of these errors is to find an impact where none exists. Fixing these errors—by

17

leaving on the iTunes 4.7 variable throughout the period and by turning on the iTunes 7.0 variable

18

as to all models—reduces the claimed damages to zero. Id., Ex. 14 at ¶ 7, ¶ 15 & Ex. JT-2a JT2b.

19

Damages for resellers are likewise reduced to zero even if only the latter error is corrected—i.e.,

20

the iTunes 7.0 variable is turned for all models, without changing his treatment of the iTunes 4.7

21

variable. Id.

22
23

2.

The regression does not account for the impact of the unchallenged
aspects of

.

24

Dr. Noll’s regression includes a single “iTunes 7.0” variable that he asserts measures the

25

alleged unlawful overcharge in iPod prices. Id., Ex. 3 at p. 81; see also Ex. 3 at ¶ 80. However,

26

plaintiffs challenge only one aspect of the iTunes 7.0—

27

aspects of iTunes 7.0 that are much more likely to affect demand for iPods, “including new album

28

covers views of music, TV shows, movies with better browsing capability and videos with ‘near
16

. They do not challenge the other

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

DVD’ quality.” Id., Ex. 3 at ¶ 113; see also Ex. 4 at ¶ 110. Reflecting all of these value-

2

enhancing features, Apple described iTunes 7.0 at the time as “the most significant enhancement

3

to the world’s most popular music jukebox and online music and video store since it debuted in

4

2001.” Id., Ex. 12. Dr. Noll’s regression must control for these other, unchallenged aspects of

5

iTunes 7.0 to ensure that the coefficient on the iTunes 7.0 variable captures only the price impact

6

of the challenged conduct. Id., Ex. 3 at ¶¶ 80, 108; Ex. 4 at ¶ 48.

7

Dr. Noll does not deny the need to control for the other, unchallenged aspects of iTunes

8

7.0—and he admits that failing to control for them would result in erroneously attributing to the

9

a price effect caused by something else. Id., Ex. 10 at 34:19-25, 35:2-5. He asserts,

10

however, that all of the attributes of iTunes 7.0 that might affect iPod prices are captured in his

11

other variables. When pressed to identify those other variables, he pointed to the variables for the

12

type of iPod (classic, mini, nano and shuffle) and his variables for storage capacity and photo and

13

video capability. Id. at 31:11-32:21.

14

This is insufficient. Dr. Noll cannot explain how any of these variables would capture the

15

effect of the other, unchallenged aspects of iTunes 7.0. The fact that a given iPod is a nano or a

16

classic does not shed any light on whether the ability to browse by album cover increased the

17

value of that iPod. Nor does the iPod’s storage capacity answer that question. Similarly, ability

18

to display a video is not the same as being able to play, for the first time, a full-length movie.

19

Because iTunes 7.0 included “unique attributes” that Dr. Noll’s model ignores, the model

20

cannot accurately determine whether “any price-elevating impact of iTunes 7—assuming there

21

even was one—was due to the

22

Amiri Decl., Ex. 3 at ¶ 113. Expert damages calculations that fail to “segregate the losses, if any,

23

caused by acts which were not antitrust violations from those that were” will not assist the trier of

24

fact and must be excluded.4 That is the situation here.

25
26
27
28

as opposed to other value-enhancing features of iTunes 7.”

4

City of Vernon v. S. Cal. Edison Co., 955 F.2d 1361, 1372-73 (9th Cir. 1992) (damages
study that treated lawful acts as “contribut[ing] to the damage figure” was not competent
evidence of damages); see also Concord Boat, 207 F.3d at 1056-57 (expert opinion as to the
alleged overcharge “should not have been admitted because it … did not separate lawful from
unlawful conduct”); In re Se. Milk Antitrust Litig., 2012 WL 1032797, at *5-6 (expert damages
(continued)
17

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

3.

2

The regression does not account for major product characteristics and
other factors relevant to price.

3

Dr. Noll recognizes that product features are important determinants of price, and thus his

4

regression includes variables for a few such characteristics—e.g., storage capacity, photo and

5

video capability, and size (cubic inches). Id., Ex. 10 at 30:22-32:21 & Ex. 7 at Ex. 13. But he

6

omits numerous other characteristics that are equally (if not more) likely to affect price, including

7

battery life, display size, weight, screen resolution, and type of connector (USB or FireWire). Id.,

8

Ex. 3 at ¶ 110; see also Ex. 4 at ¶ 111, Ex. 10.

9

“Omitted variable bias arises when important explanatory variables that have been

10

omitted from the regression model are correlated with included explanatory variables. . . . This

11

misspecification will bias the resulting coefficient estimates, and make these estimates unreliable

12

for damage estimation.” American Bar Association, Proving Antitrust Damages: Legal and

13

Economic Issues 148 (2d ed. 2010). Thus, courts have repeatedly recognized that failing to

14

control for major variables renders a regression unreliable and inadmissible.5 That is the

15

circumstance here. As Dr. Noll recognized by including some product characteristic variables in

16

his regression, product features are important to determining price—and there is no basis for

17

thinking that storage capacity (which he included) is relevant while battery life, screen size,

18

screen resolution, weight and type of connector (all of which he excluded) are irrelevant. As Dr.

19

Noll has admitted, a variable should be included if “prices plausibly could be affected by it.”

20
21
22
23
24
25
26
27
28

model measuring price impact of lawful merger rather than price impact of the alleged conspiracy
could not establish antitrust injury).
5

In re Live Concert Antitrust Litig., 863 F. Supp. 2d 966, 973-83 (C.D. Cal. 2012)
(excluding expert regression that failed to account for major variables); In re REMEC Inc. Sec.
Litig., 702 F. Supp. 2d 1202, 1273-74 (S.D. Cal. 2010) (excluding regression that made “no
attempt to account for other possible causes” of changes in stock price); In re Graphics
Processing Units, 253 F.R.D. at 496-97 (finding plaintiffs’ econometric models to be “grossly
lacking” where they omitted “factors that would likely have an impact on prices); In re
Methionine Antitrust Litig., MDL No. 00-1311, 2003 U.S. Dist. LEXIS 14828, at *9-10 (N.D.
Cal. Aug. 22, 2003) (holding that failure to take into account variables that could affect price
rendered analysis unreliable); Blomkest Fertilizer, Inc. v. Potash Corp. of Saskatchewan, 203
F.3d 1028, 1038 (8th Cir. 2000) (finding econometric model not to be probative where the model
failed to take into account relevant events).
18

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

ECF No. 685, p. 9. That standard is more than satisfied as to the omitted characteristics here.

2

Not only are the attributes on their face just as likely (or more likely) to affect demand as the

3

attributes Dr. Noll included,

4
5

Moreover, under plaintiffs’ theory, each of these characteristics would be

6

expected to be correlated with the iTunes 7.0 variable, as each would be expected to make iPods

7

more desirable and thus increase demand.

8

The relevance of these characteristics is confirmed by regressions run by Apple’s experts,

9

which included variables for the excluded characteristics. When the proper but-for world is used,

10

controlling for even a few of the iPod features Dr. Noll omitted reduces the claimed damages to

11

zero. Amiri Decl., Ex. 14 at ¶ 6, ¶12-14, ¶¶16-22 & Ex. JT-1a, JT-1b. By failing to control for

12

the relevant product characteristics, Dr. Noll’s regression erroneously attributes iPod price

13

changes to

14

overcharge where none exists.” Id. ¶¶ 79-80; see also Amiri Decl. Ex. 4 at ¶ 110. The regression

15

is thus inadmissible. As the court recognized in excluding the regression model in Freeland v.

16

AT&T Corp., 238 F.R.D. 130 (S.D.N.Y. 2006), “an unsupported assumption that changes in

17

quality would never affect . . . price is insufficient to justify its exclusion from the regression

18

analysis, particularly in light of evidence to the contrary.” Id. at 148.

that actually are the result of the omitted variables, and “produce[s] an

19

In his rebuttal report, Dr. Noll does not dispute the need to control for product character-

20

istics that could affect price. Nor does he deny that the product characteristics he omitted could

21

affect price. He argues instead that his other variables capture any effect from the omitted

22

features. Amiri Decl., Ex. 8 at pp. 8, 31. But he does not explain how, for example, a variable

23

for storage capacity (or video capability) would measure the impact of introducing batteries with

24

longer battery life. Or how a variable for the size of the iPod captures the value of enhanced

25

screen resolution. He suggests that his nano variable would capture the effect of any feature “that

26

is found only in an iPod nano.” Id. at p. 31. But he does not assert that the features he has

27

excluded are unique to the nano or to any other specific iPod model. In fact, they are not. Battery

28

life, display size, weight, screen resolution, and type of connector (USB or FireWire) are not
19

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

specific to any model – any more than are the storage capacity or video capability variables Dr.

2

Noll opted to include in his model.

3

Like the expert in Freeland, Dr. Noll has “offered no empirical analysis to support his

4

conclusion” that the features he has excluded are irrelevant to price, despite their obvious

5

importance to inconsumers. 238 F.R.D. at 146. His regression is thus “so incomplete as to be

6

inadmissible as irrelevant.” Id. at 147 (quoting Bazemore v. Friday, 478 U.S. 385, 400 (1986)).

7

C.

8

Dr. Noll’s regression is inadmissible for an additional, independent reason: his damages

9

estimates are not statistically significant. Dr. Noll asserts that his regression should be accepted

The Regression Does Not Provide Statistically Significant Results.

10

because the low standard errors on the coefficients he reports supposedly show that the

11

coefficients are statistically significant. Amiri Decl., Ex. 6 at p. 80. But he has vastly

12

underestimated the standard errors. When properly calculated, the standard errors show that the

13

coefficient for 7.0 is statistically insignificant—i.e., no different from zero and thus meaningless

14

to show any impact on prices.6

15

As Dr. Noll explains, a fundamental assumption relied on in regression analysis is that the

16

errors or residuals (the portion of the price that cannot be explained by the explanatory variables)

17

are uncorrelated or independent of one another—i.e., that “the magnitude of the error in one

18

observation does not reveal any information about the magnitude of the error in another

19

observation.” Id., Ex. 8 at pp. 36-38; Ex. 11 at 23:9-24:7; Ex. 4 at pp. 32-33. If the errors within

20

a group are correlated (not independent), they are said to be clustered and must be addressed by

21

standard correction methods. Id., Ex. 4 at pp. 34-35; Proving Antitrust Damages, p. 147

22

(allowing for clustering is “used generally in practice”). If no correction is made, the standard

23

errors will be miscalculated and will under- or over-estimate the model’s precision. Amiri Decl.,

24
25
26
27
28

6

Standard errors measure the precision with which the coefficients are estimated—the
confidence that the true value is near the estimated value. The smaller the standard error, the
more precise the results. Standard errors can be used to calculate a t-statistic, which reflects the
degree of confidence that the estimate for the coefficient did not arise by chance when the true
effect of the variable in question is actually zero. Amiri Decl., Ex. 3 at pp. 46-47; Ex. 4 at pp. 2223, 27-37.
20

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

Ex. 11 at 25:15-20; Ex. 4 at pp. 34-37; Ex. 3 at pp. 50-51.7

2

Dr. Noll admits that there are standard procedures to test whether the errors in a regression

3

are correlated. Id., Ex. 14 24:9-14. But he did not employ any of those procedures, even after

4

Apple’s experts explained that the errors in his regression are in fact correlated. Id. at 27:9-32:8,

5

45:23-46:9. Had he run such a test, he would have discovered that errors within certain groups or

6

clusters are highly correlated, revealing that the independence assumption is false. Id., Ex. 14,

7

¶ 10.

8

Correcting for this fundamental error renders Dr. Noll’s estimated “overcharges”

9

statistically indistinguishable from zero. On his regression for reseller transactions, the t-statistic

10

for the iTunes 7.0 co-efficient goes from 37.599 to a corrected 0.577. Id., Ex. 14, at ¶ 9 & n.9 &

11

Exhs. JT-1a, JT-1b & App. D1a, Dib. And in the direct sales regression, it goes from 448.711 to

12

1.536 Id. T-statistics of less than 2 signify the result is not statistically significant. Id., Ex. 4 at

13

¶ 63. Because its results are not statistically significant, Dr. Noll’s “model provides no reliable

14

evidence that the iTunes 7 update had any material effect on the prices of iPods.” Id., Ex. 3 at ¶

15

99; see also Ex. 4 at ¶¶ 79-84.

16

IV.

DR. NOLL’S RELEVANT MARKET OPINIONS SHOULD BE EXCLUDED AND

17

SUMMARY JUDGMENT GRANTED FOR FAILURE OF PROOF ON THAT

18

ISSUE.

19

To prove a monopolization claim, the plaintiff must define the relevant product market.

20

Spectrum Sports, Inc. v. McQuillan, 506 U.S. 447, 455-59 (1993). Doing so requires identifying

21

“the pool of goods or services that enjoy reasonable interchangeability of use and cross-elasticity

22

of demand.” Oltz v. St. Peter's Cmty. Hosp., 861 F.2d 1440, 1446 (9th Cir. 1988). Dr. Noll

23

purports to define two relevant product markets: the device market, which he opines includes

24
25
26
27
28

7

Proving Antitrust Damages, pp. 145-46 (“There can be substantial consequences from
estimating the standard errors for the coefficient estimates as if the errors were uncorrelated when
they are in fact correlated. With positive correlation between the error terms, the incorrectly
estimated standard error generally will be biased downward, making the regression coefficients
seem to be more precisely estimated than they really are. As a result, a statistical test on the
coefficients may yield what appears to be a statistically significant result but is not.”).
21

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

only portable digital media players; and the music market, which he opines consists solely of

2

digital audio files. Amiri Decl., Ex. 6 at pp. 23-42. This testimony should be excluded, because

3

it is not the product of reliable economic principles and methods.

4

Dr. Noll did not employ the economic methods courts have recognized as valid for

5

defining a relevant market. He admits that economists typically determine relevant markets by

6

estimating cross-elasticity of demand. Id. at pp. 23-24. But he did not do that. He similarly

7

alludes to the “hypothetical monopolist” test endorsed by the Department of Justice. But he did

8

not employ that method either. Nor did he conduct any other thorough quantitative analysis of

9

prices, costs or output of potential competing products. Like the expert excluded in Menasha

10

Corp. v. News Am. Mktg. In-Store, Inc., 354 F.3d 661, 664 (7th Cir. 2004), he “introduced no

11

econometric analysis of any kind” related to market definition. He asserts that economists rely on

12

surveys of buyers, or statements of executives in the industry. But he does not present any such

13

evidence.

14

Rather than relying on any accepted economic methods, Dr. Noll offers his own truncated

15

review of what he perceives to be the similarities or dissimilarities between certain products, with

16

an occasional citation to a news article or other report from the internet. He does not purport to

17

be applying any economic expertise in doing so. Nor does he profess to have any industry or

18

other experience that would make him an expert on what consumers consider to be substitutes for

19

digital music and music players. This portion of his report could have just as easily been written

20

by a non-expert lawyer or anyone else with the ability to run internet searches. It is not valid

21

expert opinion. See Live Concert, 863 F. Supp. 2d at 994 (excluding expert report purporting to

22

define relevant product market where expert “purported to rely on ‘industry information’ … [but]

23

did not utilize a reliable methodology for interpreting and applying this information”).8

24
25
26
27
28

8

Brown Shoe Co. v. United States, 370 U.S. 294 (1962), held that “practical indicia” may
determine “[t]he boundaries of … a submarket.” Id. at 325. But the Ninth Circuit has never held
that the relevant product market can be defined “exclusively by reference to these ‘practical
indicia.’” In re Live Concert, 863 F. Supp. 2d at 985-86; accord Reifert v. S. Cent. Wisc. MLS
Corp., 450 F.3d 312, 320 (7th Cir. 2006). And Dr. Noll did not in any event employ any
economic expertise related to any practical indicia.
22

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

In addition to not applying any relevant expertise, Dr. Noll’s non-expert product review is

2

flawed. As to the player market, he refers only to CD players and cell phones, without discussing

3

any other potentially competing devices for playing digital music, such as notebook computers

4

and home stereo systems—products that Apple’s witnesses testified they considered to be

5

competitive products. Amiri Decl., Ex. 13 at 48:4-10. Likewise, Dr. Noll does not address

6

whether customers might substitute free downloads from peer-to-peer file sharing sites for iTS

7

music, disregarding evidence that the availability of free downloads was a major challenge to the

8

success of paid music stores like iTS. Id., Ex. 3 at ¶ 125. Instead, Dr. Noll “focused on finding

9

the perfect substitute[s] for [iPods and iTMS music],” ignoring that some “flexible elastic

10

buyer[s]” will switch to “substitutes, even if the substitutes are imperfect.” Va. Vermiculite Ltd.

11

v. W.R. Grace & Co.-Conn., 98 F. Supp. 2d 729, 737 (W.D. Va. 2000) (excluding expert

12

testimony for failure to “thoroughly examine substitutes when defining the market”); Ky.

13

Speedway, LLC v. Nat’l. Assoc. of Stock Car Auto Racing, Inc., 588 F.3d 908, 916-17 (6th Cir.

14

2009) (affirming order striking expert’s “report and deposition testimony because he had included

15

only Busch series and open-wheeled races as possible substitutes for attending live NASCAR

16

stock-car racing events or watching them on television” without considering “other sports—and

17

possibly other forms of entertainment—as substitutes”).

18
19
20

To the extent Dr. Noll does address possible substitutes, he excludes them based on
improper considerations. For example,


Dr. Noll assumes that cell phones must be able to download music over wireless carriers’

21

networks to be substitutes for iPods. Amiri Decl., Ex. 6 at pp. 29-31. But iPods never had

22

that capability during the class period. Id., Ex. 3 at ¶ 130. Only the iPod Touch could

23

download music at all, and only over a Wi-Fi connection, not a wireless network. Id.

24



Dr. Noll excludes on-demand and non-interactive streaming music from the market

25

because those services were not fully available on mobile devices during the class period.

26

Id., Ex. 6 at pp. 33-39. But support for mobile devices is not relevant to consumers who

27

purchased music to play on other devices, and Dr. Noll presents no analysis as to the

28

volume of iTS sales for use on such other devices. Id., Ex. 3 at ¶ 126.
23

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR



1

Dr. Noll opines that a streaming music service (like Pandora) is “most likely” not a

2

substitute for iTS music because it is not fully customizable. Id., Ex. 6 at p. 37. But he

3

offers no data or reasoned analysis to explain why a service must be fully customizable

4

before consumers will choose it over permanent downloads.


5
6

Dr. Noll offers only untested, subjective opinions for his conclusion that CDs are not close
substitutes for downloads, (see id., Ex. 6 at pp. 39-41),

7
8
9

Expert opinions based on untested assumptions and subjective opinions, as opposed to

10

market data, are routinely excluded as inadmissible under Daubert.9 Even “reasonable” or

11

“intuitively obvious” presumptions cannot form the basis for expert opinions. See United Food

12

Mart, Inc. v. Motiva Enters., LLC, 404 F. Supp. 2d 1344, 1349-50 (S.D. Fla. 2005) (expert

13

opinion based on “reasonable” assumption that gas stations located within two miles on one

14

another on the same road compete excluded because expert did not “test [it] by reference to

15

specific facts of this case”); SMS Sys. Maint. Servs., Inc. v. Digital Equip. Corp., 188 F.3d 11, 25

16

(1st Cir. 1999) (refusing to credit expert opinion based on an “intuitively obvious proposition”

17

and sources the expert “neither attaches nor discusses”).

18

In sum, Dr. Noll’s proposed relevant product definitions must be excluded because they

19

are the product of untested, subjective opinions regarding customer preferences, as opposed to

20

any economic analysis of the markets, the interchangeability of products, or actual consumer

21

behavior. And, because—absent Dr. Noll’s opinion—plaintiffs have no basis upon which to prove

22

a relevant market, summary judgment should be granted on this ground as well. In re Live

23

Concert, 843 F. Supp. 2d at 1000 (granting summary judgment for lack of proof of relevant

24

market after excluding expert witness opinion); Va. Vermiculite, Ltd. v. W.R. Grace & Co.-Conn.,

25
26
27
28

9

See In re Live Concert, 863 F. Supp. 2d at 994 (excluding expert’s relevant market
definition where expert “relied on his own subjective opinion in order to determine which
performers qualify as ‘rock’ artists”); McLaughlin Equip. Co., Inc. v. Servaas, No. 98-127-C-T/K,
2004 WL 1629603, at *6 (S.D. Ind. Feb. 18, 2004) (excluding expert’s relevant market opinion
where it was “based on insufficient data,” including only one piece of market data).
24

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR

1

108 F. Supp. 2d 549, 576 (W.D. Va 2000) (same); Ky. Speedway, 588 F.3d at 919 (affirming

2

summary judgment after concluding that expert testimony on relevant market had been properly

3

excluded).

4
5

CONCLUSION
Dr. Noll’s testimony regarding impact, damages and relevant market should be excluded

6

and summary judgment granted to Apple.

7

Dated: December 20, 2013

8

Respectfully submitted,
JONES DAY

9
10

By: /s/ Robert A. Mittelstaedt
Robert A. Mittelstaedt

11

Counsel for Defendant APPLE INC.

12
13

SFI-847506v2

14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
25

Motion for Summary Judgment and
To Exclude -- C 05-00037 YGR