"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

Lead Case No. C 05-00037 YGR
[CLASS ACTION]

15

___________________________________

16

This Document Relates To:

17

ALL ACTIONS

[PROPOSED] ORDER GRANTING
DEFENDANT’S MOTION FOR
SUMMARY JUDGMENT AND
EXCLUDING THE TESTIMONY OF
ROGER G. NOLL

18
19
20
21
22
23
24
25
26
27
28
0

[Proposed] Order
C 05-00037 YGR

1

Presently before the court is Defendant Apple, Inc.’s Motion for Summary Judgment and

2

to Exclude the Expert Testimony of Roger G. Noll. Having reviewed the papers on file and

3

otherwise being fully informed of the premises of Apple’s motion, this Court hereby GRANTS

4

the motion.

5

I.

BACKGROUND

6

A.

Plaintiffs’ Original Theory.

7

Plaintiffs brought this suit in 2005—when online music stores were still using DRM—

8

alleging that Apple’s use of its proprietary FairPlay DRM was an antitrust violation. Plaintiffs

9

asserted that using a proprietary, rather than a shared, system prevented music purchased from the

10

iTS to play on music players made by other companies (iTS music could be “indirectly” played

11

on competing portable devices by burning the music to a CD and importing, or “ripping,” it to a

12

computer, but loaded “directly” only to iPods). As a consequence, plaintiffs contended once they

13

had purchased iTS music, they had become locked-in to buying iPods rather than competing

14

players in the sense that their “switching costs” would be greater.

15

This Court rejected plaintiffs’ original claim, holding that it was lawful for Apple to adopt

16

its own DRM, even if the result was to make Apple’s products incompatible with competitors’

17

products. Order Decertifying Class (ECF No. 303), pp. 2, 10; Order Granting in Part Apple’s

18

Mot. Summ. J. (ECF No. 274), p. 10.

19

B.

The Amended Consolidated Complaint.

20

After the Court’s rejection of their original claim, targeted iTunes 7.0, a software update

21

issued in September 2006. Plaintiffs challenge only one aspect of iTunes 7.0: code included in

22

the software that protected against the risk that the iPod would not play songs properly as a result

23

of third-party programs attempting to manage the iPod incorrectly. Plaintiffs argue that iTunes

24

7.0 was anticompetitive because it prevented music from other music stores from directly playing

25

on certain iPods. In particular, plaintiffs argue that the updates disabled RealNetworks’

26

Harmony, technology that enabled digital music from iTS to play on non-iPod devices.

27

Plaintiffs’ theory of antitrust impact is that, by disabling Harmony, iTunes 7.0 resulted in some

28

consumers purchasing more music from iTS than they otherwise would have, which in turn
1

[Proposed] Order
C 05-00037 YGR

1

locked them into buying an iPod when they next purchased a digital music player. According to

2

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

3

say, other customers were locked out of buying iPods after iTunes 7.0 because they had large

4

libraries of Harmony music that would no longer play on the iPods on which the KVC was

5

implemented. According to plaintiffs, this lock-out led to higher prices because Apple no longer

6

had the incentive to reduce prices to attract these customers.

7

II.

8
9

DISCUSSION
A.

Summary Judgment

Summary judgment is proper under rule 56 when “there is no genuine issue as to any

10

material fact and . . . the movement is entitled to judgment as a matter of law.” Fed. R. Civ. P.

11

56(c)(2). A party seeking summary judgment bears the initial burden of informing the court of

12

the basis for its motion, and of identifying those portions of the pleadings, depositions, discovery

13

responses, and affidavits that demonstrate the absence of a genuine issue of material fact. Celotex

14

Corp. v. Catrett, 477 U.S. 317 , 323 (1986). Material facts are those that might affect the outcome

15

of the case. Anderson v. Liberty Lobby, Inc., 477 U.S. 242 , 248 (1986). The “mere existence of

16

some alleged factual dispute between the parties will not defeat an otherwise properly supported

17

motion for summary judgment; the requirement is that there be no genuine issue of material fact.”

18

Id. at 247-48 (dispute as to a material fact is “genuine” if there is sufficient evidence for a

19

reasonable jury to return a verdict for the non-moving party).

20

Courts frequently grant summary judgment in antitrust cases when the plaintiffs lack

21

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

22

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

23

inadmissible. See e.g., In re Southeastern Milk Antitrust Litigation, 2012 WL 1032797, at *4–6

24

(E.D. Tenn. Mar. 27, 2012) (finding that expert’s model of antitrust injury could not identify price

25

changes caused by alleged anti-competitive conduct and granting summary judgment for

26

defendants); In re eBay Seller Antitrust Litigation, 433 Fed. Appx. 504, 506 (9th Cir. 2011)

27

(finding expert’s econometric analysis inadequate to support claim of antitrust injury and

28

affirming summary judgment for eBay); Concord Boat Corp. v. Brunswick Corp., 207 F.3d 1039,
2

[Proposed] Order
C 05-00037 YGR

1

1054-58, 1063 (8th Cir. 2000) (finding that plaintiff’s expert’s opinions on injury should have

2

been excluded and remanding for entry of judgment).

3

Plaintiffs’ claim of impact is unsupported by any record facts. For the alleged impact to

4

have occurred, RealNetworks must have been selling Harmony music to iPod owners (or

5

potential iPod purchasers) in sufficient quantities in September 2006 that a drop in those sales

6

would have affected demand for iPods to such a degree that Apple would have changed its prices

7

in response. That, in turn, requires the existence of consumers who, because of the update,

8

became either locked in (by their iTS purchases) or locked out (by their Harmony purchases), and

9

thus purchased a player different from what they would have purchased absent iTunes 7.0. And

10

the claim requires that Apple became aware of these customers (and how many there were) and

11

took them into account in setting its prices. Plaintiffs have no record proof of any of the facts.

12

First, plaintiffs do not show that Harmony had any significant share of music sales in

13

September 2006 or the months thereafter. Second, plaintiffs have no evidence regarding what

14

portion of RealNetwork’s sales was of Harmony music to iPod owners or potential iPod

15

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

16

locked out as a result of iTunes 7.0. It is not enough merely that a Harmony purchaser owned an

17

iPod or was a potential iPod purchaser. Plaintiffs must show, whether and to what extent

18

consumers’ decision about whether to purchaser an iPod was changed as a result of iTunes 7.0.

19

Fourth, despite extensive discovery related to Apple’s pricing committee (including document

20

production and depositions), plaintiffs cannot identify any evidence that Apple ever took into

21

account sales (or the lack thereof) from RealNetworks or any other online store in setting iPod

22

prices. Fifth, plaintiffs’ hypothesized price effect is irreconcilable with the real-world manner in

23

which Apple set its prices.

24

In short, plaintiffs have no documentary or percipient witness evidence to show even the

25

most basic facts necessary for their claim to be valid. Their complaint failure of proof at this

26

stage requires that summary judgment be granted. “[P]roof of injury is an essential substantive

27

element” of an antitrust claim, on which plaintiffs bear the burden. Kline v. Coldwell, Banker &

28

Co., 508 F.2d 226, 233 (9th Cir. 1974); Rebel Oil Co. v. Atl. Richfield Co., 51 F.3d 1421, 1433
3

[Proposed] Order
C 05-00037 YGR

1

(9th Cir. 1995) (“causal antitrust injury[] is an element of all antitrust suits”). Lack of evidence of

2

injury caused by the challenged is thus “fatal to the merits of any antitrust claim.” Argus Inc. v.

3

Eastman Kodak Co., 801 F.2d 38, 41 (2d Cir. 1986). That is the situation here.

4
5

B.

Exclusion of Dr. Noll’s Testimony

Rule 702 permits a qualified expert to present testimony that “will help the trier of fact to

6

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

7

facts or data,” “is the product of reliable principles and methods,” and “the expert has reliably

8

applied the principles and methods to the facts of the case.” FED. R. EVID. 702. As “gatekeeper,”

9

the Court must ensure that expert testimony “rests on a reliable foundation and is relevant to the

10

task at hand.” Daubert v. Merrell Dow Pharms., Inc., 509 U.S. 579, 597 (1993). “Maintaining

11

Daubert's standards is particularly important considering the aura of authority experts often

12

exude, which can lead juries to give more weight to their testimony.” Mukhtar v. California State

13

Univ., Hayward, 299 F.3d 1053, 1063-64 (9th Cir. 2002) amended, 319 F.3d 1073 (9th Cir.

14

2003). Litigants should not be permitted to get to trial simply by attempting to “dazzle the

15

courtroom with graphs and tables.” In re Graphics Processing Units Antitrust Litig., 253 F.R.D.

16

478, 491-92 (N.D. Cal. 2008) (rejecting an expert regression at the class certification stage).

17

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

18

“subjective belief or unsupported speculation.” Id. at 590. While the focus is on the expert’s

19

underlying methodology, “conclusions and methodology are not entirely distinct from one

20

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

21

data and the opinion proffered.” General Elec. Co. v. Joiner, 522 U.S. 136, 146 (1997). Relevant

22

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

23

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

24

(“Daubert II”).

25
26
27
28

1.

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

Dr. Noll’s regression must be excluded because it is premised on assumptions that are at
odds with the facts of the case and with plaintiffs’ own theory. Damages models premised on the
4

[Proposed] Order
C 05-00037 YGR

1

types of unsupported assumptions on which Dr. Noll relies are inadmissible. See Concord Boat,

2

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

3

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

4

drive engine market, for it ignored inconvenient evidence”); So. Pacific Comm’ns Co. v. Am. Tel.

5

& Tel. Co., 556 F. Supp. 825, 1076-77 (D.D.C. 1982) (concluding that “plaintiffs’ model of the

6

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

7

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

8

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

9

relevant). Dr. Noll’s model does not “fit” the actual evidence in several, fundamental respects.

10

First, as discussed, the price effect the model purports to find is unsupported by any real

11

world evidence regarding the amount of Harmony sales, the purchasers of Harmony music, the

12

degree of any lock-in or lock-out of any consumer, or any awareness or consideration by Apple’s

13

pricing committee of any effect from iTunes 7.0. Second, Dr. Noll’s hypothesized price effect is

14

irreconcilable with the real world manner in which Apple set its prices. Apple has consistently

15

followed a strategy of setting prices at set intervals (e.g., $399, $249, $199). Yet, Dr. Noll’s

16

model finds that, in response to sales from a store with a miniscule market share, Apple would

17

have departed from this consistent practice and sold iPods at varying price points, such as

18

$184.17 or $276.72. This finding defies reality. Third, Dr. Noll applies a single, constant

19

percentage overcharge to every affected iPod sold during the damage period. If his theory were

20

correct, however, the lock-in would have progressively increased during the damage period as

21

consumers purchased more iTS music. Dr. Noll offers no coherent theory how the alleged

22

overcharge amount could possibly remain constant in these circumstances.

23

Because it rests on unsupported assumptions that conflict with the real world, Dr. Noll’s

24

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

25

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

26

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

27

131 F. App’x 450 (5th Cir. 2005) (excluding expert damages model where it “ignore[d] market

28

realities”).
5

[Proposed] Order
C 05-00037 YGR

2.

1
2

The Regression Does Not Properly Account For the Relevant Factors.

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

3

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

4

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

5

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

6

requires that his testimony be excluded.

7
8
9

a.

iTunes 4.7 Variable

First, Dr. Noll’s regression includes variables for events that he contends may have
affected iPod prices, including iTunes 4.7. In his initial report, Dr. Noll turned the iTunes 4.7

10

variable off as of September 12, 2006, the day iTunes 7.0 was released. By turning the iTunes 4.7

11

variable off, his initial regression treated iTunes 4.7 as having no plausible effect on iPod prices

12

after iTunes 7.0 was released—and thus treats the but-for world as one in which iTunes 4.7 did

13

not exist.

14

In his rebuttal report, Dr. Noll revised his regression to leave on the iTunes 4.7 variable as

15

to those iPods on which iTunes 7.0 was not implemented, but he continues to turn it off as to the

16

models as to which iTunes 7.0 was implemented. Thus, Dr. Noll again treats iTunes 4.7 as

17

having no effect on those models once iTunes 7.0 was released. Dr. Noll compounds this error by

18

not turning on the iTunes 7.0 variable for those iPod models which were not affected by iTunes

19

7.0. By doing so, he treats the alleged conduct as not having any possible effect on the demand

20

for unaffected models.

21
22

b.

iTunes 7.0 Variable

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

23

alleged unlawful overcharge in iPod prices. However, plaintiffs challenge only one aspect of

24

iTunes 7.0—the software code that resulted in Harmony incompatibility. They do not challenge

25

the other aspects of iTunes that are much more likely to affect demand for iPods, including new

26

album covers views of music, TV shows, movies with better browsing capability and videos with

27

“near DVD’ quality.” Dr. Noll’s regression must control for these other, unchallenged aspects of

28

iTunes 7.0 to ensure that the coefficient on the iTunes 7.0 variable captures only the price impact
6

[Proposed] Order
C 05-00037 YGR

1

of the challenged conduct. By failing to do so, the model cannot ascribe any potential price-

2

elevating impact of iTunes 7 to the challenged conduct as opposed to the other value-enhancing

3

features of iTunes 7. Expert damages calculations that fail to “segregate the losses, if any, caused

4

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

5

and must be excluded. City of Vernon v. S. California Edison Co., 955 F.2d 1361, 1372-73 (9th

6

Cir. 1992) (damages study that treated lawful acts as “contribut[ing] to the damage figure” did not

7

constitute competent evidence of damages) That is the situation here.

8
9

c.

Other Characteristics and Factors Affecting Price

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

10

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

11

video capability, and size (cubic inches). But he omits numerous other characteristics that are

12

equally likely to affect price, including battery life, display size, weight, screen resolution, and

13

type of connector (USB or FireWire).

14

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

15

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

16

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

17

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

18

Economic Issues 148 (2d ed. 2010). As the court recognized in excluding the regression model in

19

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

20

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

21

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

22

in Freeland, Dr. Noll has “offered no empirical analysis to support his conclusion” that the

23

features he has excluded are irrelevant to price. 238 F.R.D. at 146. His regression is thus “so

24

incomplete as to be inadmissible as irrelevant.” Id. at 147 (quotations omitted).

25
26

3.

The Regression Does Not Provide Statistically Significant Results

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

27

estimates are not statistically significant. When properly calculated, the standard errors show that

28

the coefficient for 7.0 is statistically insignificant—i.e., no different from zero and thus
7

[Proposed] Order
C 05-00037 YGR

1

meaningless to show any impact on prices.

2

A fundamental assumption relied on in regression analysis is that the errors or residuals

3

(the portion of the price that cannot be explained by the explanatory variables) are uncorrelated or

4

independent of one another. If the errors within a group are correlated (not independent), they are

5

said to be clustered and must be addressed by standard correction methods. If no correction is

6

made, the standard errors will be miscalculated and will under- or over-estimate the model’s

7

precision. .

8
9

Dr. Noll admits that there are standard procedures to test whether the errors in a regression
are correlated. But he did not employ any of those procedures, even after Had he run such a test,

10

he would have discovered that errors within certain groups or clusters are highly correlated,

11

revealing that the independence assumption is false..

12

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

13

statistically indistinguishable from zero. Because its results are not statistically significant, Dr.

14

Noll’s “model provides no reliable evidence that the iTunes 7 update had any material effect on

15

the prices of iPods.”

16
17

C.

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

18

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

19

identifying “the pool of goods or services that enjoy reasonable interchangeability of use and

20

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

21

Dr. Noll purports to define two relevant product markets: the device market, which he opines

22

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

23

solely of digital audio files.

24

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

25

defining a relevant product market. He did not estimate cross-elasticity of demand or apply the

26

“hypothetical monopolist” test endorsed by the Department of Justice. Further, Dr. Noll did not

27

conduct any other thorough quantitative analysis of prices, costs or output of potential competing

28

products. Like the expert excluded in Menasha Corp. v. News Am. Mktg. In Store Servs., Inc.,
8

[Proposed] Order
C 05-00037 YGR

1

354 F.3d 661, 664 (7th Cir. 2004), he “introduced no econometric of any kind” related to market

2

definition. He asserts that economists rely on surveys of buyers, or statements of executives in

3

the industry. But does not present any such evidence.

4

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

5

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

6

an occasional citation to a news article or other report from the internet. Such an exercise is not

7

valid expert opinion. See In re Live Concert, 863 F. Supp. 2d at 994 (excluding expert report

8

purporting to define relevant product market where expert “purported to rely on ‘industry

9

information’ … [but] did not utilize a reliable methodology for interpreting and applying this

10

information”).

11

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

12

flawed. Dr. Noll “focused on finding the perfect substitute[s] for [iPods and iTMS music],”

13

ignoring the fact that some “flexible elastic buyer[s]” will switch to “substitutes, even if the

14

substitutes are imperfect.” Virginia Vermiculite Ltd. v. W.R. Grace & Co.-Conn., 98 F. Supp. 2d

15

729, 737 (W.D. Va. 2000) (excluding expert testimony for failure “to thoroughly examine

16

substitutes when defining the market”). Kentucky Speedway, LLC v. N.A. of Stock Car Auto

17

Racing, Inc., 588 F.3d 908, 916-17 (6th Cir. 2009) (affirming order striking expert’s “report and

18

deposition testimony because he had included only Busch series and open-wheeled races as

19

possible substitutes for attending live NASCAR stock-car racing events or watching them on

20

television” without considering “other sports—and possibly other forms of entertainment—as

21

substitutes”). His opinions are therefore inadmissible.

22

III.

23
24

CONCLUSION
Accordingly, Apple’s Motion for Summary Judgment and to Exclude the Expert

Testimony of Dr. Roger G. Noll is hereby GRANTED.

25
26
27
28
9

[Proposed] Order
C 05-00037 YGR

1

Dated: ____________________________

2
3

The Honorable Yvonne Gonzalez Rogers
United States District Court Judge

4
5
6

SFI-847570v1

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

[Proposed] Order
C 05-00037 YGR