"The Apple iPod iTunes Anti-Trust Litigation"

Filing 747

Supplemental Response re 737 Administrative Motion to Seal Daubert, filed by Apple Inc.. (Attachments: # 1 Proposed Order, # 2 Plaintiffs Daubert Motion REDACTED, # 3 Exhibits 1-4 to Daubert Motion REDACTED, # 4 Exhibit 5 to Daubert Motion REDACTED, # 5 Exhibits 6, 9-10 to Daubert Motion REDACTED)(Related document(s) 737 ) (Amiri, Amir) (Filed on 12/31/2013) Modified on 1/2/2014 (jlmS, COURT STAFF). Modified on 1/2/2014 (jlmS, COURT STAFF).

EXHIBIT 6
[Filed Under Seal]

IN THE UNITED STATES DISTRICT COURT
FOR THE NORTHERN DISTRICT OF CALIFORNIA
OAKLAND DIVISION

__________________________________________
)
)
)
THE APPLE IPOD ITUNES
)
ANTITRUST LITIGATION
)
)
)
__________________________________________)

Case No. C 05 00037 YRG
C 06 04457 YRG
HIGHLY CONFIDENTIAL

EXPERT REPORT OF ROBERT H. TOPEL
August 19, 2013
(Amended)

TABLE OF CONTENTS
I.

INTRODUCTION AND BACKGROUND................................................................. - 1 A. Summary of Qualifications ............................................................................................. - 1 B. Assignment ..................................................................................................................... - 2 C. Summary of Opinions ..................................................................................................... - 3 -

II.

BACKGROUND: ITUNES, IPODS AND THE CHALLENGED CONDUCT ...... - 6 A. The Development of the iTunes/iPod/iTMS Platform .................................................... - 6 1. iTunes ........................................................................................................................ - 7 2. iPod ........................................................................................................................... - 7 3. The iTunes Store ..................................................................................................... - 10 B. The Introduction and Use of DRM ............................................................................... - 12 C. RealNetworks RealPlayer with Harmony ..................................................................... - 13 1. RealNetworks (Real), Harmony and the Real Music Store .................................... - 13 D. Challenged Conduct ...................................................................................................... - 15 E. The End of DRM .......................................................................................................... - 17 F. Quality Adjusted iPod Prices Rapidly Declined During the Period at Issue ................ - 18 -

III.

PLAINTIFFS’ AND PROFESSOR NOLL’S THEORY OF IMPACT AND
DAMAGES .................................................................................................................. - 19 -

IV.

PROFESSOR NOLL’S HEDONIC PRICE REGRESSION AND
INFERENCE OF DAMAGES ................................................................................... - 21 A. Multiple Linear Regression........................................................................................... - 21 B. Hedonic Models of Prices for Products with Changing Quality................................... - 24 -

V.

PROFESSOR NOLL’S HEDONIC DAMAGE MODEL ....................................... - 24 A. How Professor Noll Constructs His Data ..................................................................... - 25 B. What Professor Noll Claims to Find ............................................................................. - 27 -

VI.

PROFESSOR NOLL’S CLAIMS OF PRECISION AND STATISTICAL
SIGNIFICANCE ARE INCORRECT ...................................................................... - 28 A. Professor Noll’s “Before and After” Experiment: The Wrong But-For World ........... - 37 B. Professor Noll’s Confusion about the “Log of Time” .................................................. - 43 C. Additional Features and the “Quality” of Professor Noll’s Regression ....................... - 47 D. Professor Noll’s Regression Produces Implausible Results That Are Inconsistent
With The Basic Economics Of Plaintiffs’ Theory. ....................................................... - 49 -

HIGHLY CONFIDENTIAL

i

E. Professor Noll’s Regression Improperly Estimates A Single Average Overcharge
Across All Models ........................................................................................................ - 50 F. The Impact of DRM-Free Music on the Price of iPods ................................................ - 53 VII.

PROFESSOR NOLL’S DAMAGE CALCULATIONS .......................................... - 54 -

VIII. SOME FINAL POINTS: PROFESSOR NOLL’S REGRESSION HAS NO
CONNECTION TO THE CHALLENGED CONDUCT, PLAINTIFFS’
THEORY, OR APPLE’S PRICING PRACTICES ................................................. - 55 -

HIGHLY CONFIDENTIAL

ii

LIST OF APPENDICES

APPENDIX A:

Curriculum Vitae

APPENDIX B:

List of Materials Relied Upon

APPENDIX C:

Collection of iPod Characteristics

APPENDIX D:

Additional Regression Results

HIGHLY CONFIDENTIAL

iii

I.

INTRODUCTION AND BACKGROUND

1.

I am Robert H. Topel, Isidore Brown and Gladys J. Brown Distinguished Service

Professor of Economics at The University of Chicago Booth School of Business. I am also
Director of the George J. Stigler Center for the Study of the Economy and State and Co-Director
of the Energy Policy Institute at Chicago (EPIC), both at The University of Chicago. In addition,
I am a Senior Consultant at Charles River Associates (CRA), an economics consulting firm that
specializes in the application of economic theory and statistics to legal and regulatory issues.
A. Summary of Qualifications
2.

I am an economist, and I specialize in (among other things) microeconomics, which is the

study of markets, pricing, and firm and industry behavior. I received a B.A. in economics from
the University of California, Santa Barbara in 1974, and a Ph.D. in economics from the
University of California, Los Angeles in 1981. In addition to my position at the Booth School of
Business at The University of Chicago, I have been a member of the faculties in the Department
of Economics at The University of Chicago and the Department of Economics at the University
of California, Los Angeles. At these institutions, I have taught courses on Markets and Prices,
Economic Theory, Labor Markets, Empirical Methods in Economics, Compensation and
Personnel Policies, Industrial Organization and Antitrust, Business Strategy, and Law and
Economics.
3.

From 1993 to 2003, I served as Editor of the Journal of Political Economy, and from

1991 to 1993 I was a member of the Editorial Board of the American Economic Review, two of
the leading professional publications in economics and economic theory. I am also a past
founding editor of the Journal of Labor Economics (1982-92), and I currently am a member of
the Editorial Advisory Board of the International Journal of the Economics of Business and the
Advisory Board of the Economics Research Network. I am a Research Associate of the National
Bureau of Economic Research, an elected member of the Council on Income and Wealth, an
elected Founding Member of the National Academy of Social Insurance, and a Fellow of the
Stanford University Center for the Study of Poverty and Inequality. In 2004, I was elected a
Fellow of the Society of Labor Economists. In 2005, I received the Eugene Garfield Award for

HIGHLY CONFIDENTIAL

-1-

But these observations
are not independent; they are perfectly correlated — if one knows the price of one unit,
one knows the price of all of them. By incorrectly treating millions of highly correlated
observations as independent pricing decisions, Professor Noll has claimed to estimate his
model with an absurd degree of precision. Correcting his error using standard
econometric techniques that account for non-independence, I demonstrate that his
estimates of alleged “overcharges” are indistinguishable from zero.
•

Opinion 2: Professor Noll’s regression analysis produces implausible results that are
inconsistent with the economics of Plaintiffs’ theory. His results are based on
unsupported assumptions that are inconsistent with the basic facts of this case.
•

His regression purports to show that the challenged conduct had an immediate and
uniform impact on prices of all iPods sold from September 12, 2006 onward.
This is contrary to his and Plaintiffs’ theory in which the degree of “lock-in”
would be proportional to the increase in iPod owners’ accumulated stocks of
FairPlay-protected music caused by the challenged update. This effect was zero
in September of 2006, so even under Plaintiffs’ theory the update could not have
raised iPod prices at that time. Professor Noll’s regressions are incapable of
determining when an overcharge, if any, would have occurred.

•

The regression also produces “but-for” prices that are contrary to the way in
which Apple actually sets prices.

•

Opinion 3: Notwithstanding other flaws in Professor Noll’s model, his overcharge
estimates are based on an incorrect characterization of the “but-for” world that would
have existed in the absence of the challenged update.
The DRM protection technology in versions of iTunes used prior to

HIGHLY CONFIDENTIAL

-4-

that date — specifically in iTunes 4.7, 5 and 6 — has been found legal. Thus the proper
“before-and-after” estimate of the effect of the

would compare the level of iPod

prices after the introduction of iTunes 7.0 to the “but-for” price level that existed under
the legal DRM technology of iTunes 4.7 through 6. Professor Noll incorrectly measures
the but-for price level from a three-month period before the introduction of iTunes 4.7 in
October 2004 — nearly two years before the introduction of iTunes 7. This means that
he does not measure the incremental effect on iPod prices of the

contained in

iTunes 7; rather he includes the legal “effect” of iTunes 4.7 technology and functionality
in his estimate of an alleged overcharge due to iTunes 7. This error alone doubles
Professor Noll’s estimated of “overcharge” among resellers, and it increases his estimated
“overcharge” among direct purchasers by a factor of six.
•

Opinion 4: Professor Noll’s regression model is incorrectly formulated. Among other
errors:
•

His model does not and cannot separate the impact of the challenged

feature

from other factors that impact iPod prices and occurred at the same time the
feature was introduced. Despite his claim to control for relevant product
attributes and market forces, he omits many valuable attributes of iPods, iTunes,
and iTMS that affect the prices of iPods. He also fails to account for the fact that
80 percent of the iTMS music catalogue was DRM-free on January 6, 2009,
nearly four months before he turns on his indicator variable.
•

Professor Noll also incorrectly constrains his model so that iTunes 7 must have
the same percentage impact on the prices of all iPod models, even models that
could not invoke the challenged

feature and so continued to be compatible

with Harmony. There is no economic or factual support for the assumption that
any overcharge would be exactly the same for different products purchased by
different class members, especially models that could not invoke the challenged
feature. At deposition, Professor Noll suggested that he might attempt to fix
this problem by estimating product-specific price effects of the iTunes 7
introduction. But this is not a fix. Even ignoring his model’s other flaws, this

HIGHLY CONFIDENTIAL

-5-

approach yields estimated “effects” of iTunes 7 that are inconsistent with
Plaintiffs’ theory. Moreover, this approach cannot provide an “overcharge”
estimate for the iPod Touch because it did not exist prior to the introduction of
iTunes 7.
•

Professor Noll’s attempt to account for technological progress (“Moore’s Law”)
by using the “log of time” is, by his own testimony, incorrect and inconsistent
with standard econometric practice. As in other instances, each of these errors
and omissions increased Professor Noll’s estimates of overcharges.

•

Opinion 5: Correcting Professor Noll’s errors in the context of his own model causes his
estimates of “overcharge” and “damage” to vanish.

•

Opinion 6: A necessary predicate for his theory of damages is that enough iPod owners
were purchasing RMS music, and would have continued to do so, to have an impact on
iPod prices. Professor Noll merely assumes that this was the case — i.e., that iPod
owners were in fact purchasing RMS music and would have continued to do so in
sufficient quantities to impact iPod demand. To the extent this is not the case, Professor
Noll’s damages theory fails as a threshold matter. Professor Noll has offered no evidence
that anyone, let alone a material fraction of iPod owners, would have used Harmony and
RMS, or ever did. Prior to the challenged update, RMS accounted for only about three
percent of all digital downloads, a share that would have been substantially lower among
iPod owners.

II.

BACKGROUND: ITUNES, IPODS AND THE CHALLENGED
CONDUCT

11.

This section presents a brief overview of the relevant products and technologies at issue

and a timeline of relevant events.
A. The Development of the iTunes/iPod/iTMS Platform
12.

Exhibit 1 is a timeline of key events in the evolution of the iTunes/iPod/iTMS platform.

HIGHLY CONFIDENTIAL

-6-

1. iTunes
13.

iTunes is a “jukebox” software program that allows users to import, manage and play

digital music files on the users’ computers. Apple introduced the first version of iTunes for its
Mac line of computers in 2001 and iTunes for Windows in October 2003. Since iTunes was first
introduced, there have been ten major (version) updates, or just under one per year on average.
The first of these version updates, iTunes 2.0, added integration with iPods along with
improvements in CD burning capabilities and enhanced sound capabilities.5 In general, the new
version updates enhanced the operation of the music player; added or supported new features in
iPods, such as photo and/or video; and supported new content from the iTMS (e.g., movies, TV
shows, and games).6 There have also been 78 smaller updates: some added features or
supported new iPods; others were released in order to “fix” particular technical issues. For
example, iTunes 4.7 was released to support the new photo feature added to the iPod Classic,
improved FairPlay in response to hacks, and fixed other bugs.7 All told, there have been 88
different updates, the most recent of which (iTunes 11.0.4) was released on June 5, 2013.8 (See
Exhibit 2 for a history of the various versions of iTunes and the major updates to those versions
through March 2012.)
2. iPod
14.

Apple introduced the first iPod in October 2001. The iPod could play music from a

variety of sources, including a CD collection copied or ripped to the user’s computer and then
synced or loaded to an iPod using iTunes; music obtained online without DRM protection,
including from music stores that offered music without DRM; music that originally had DRM
protection but had been copied to a CD and then copied back to a computer (burned and ripped);
5

Apple Press Release, “Apple Announces iTunes 2; Best Digital Music Software Gets Even Better,” October 23,
2001 (Apple_AIIA00974863-5).

6

Apple Press Release, “Apple Introduces iPod shuffle; First iPod Under $100,” January 11, 2005
(Apple_AIIA00974708).

7

“iTunes version history,” Wikipedia, The Free Enclyclopedia,
http://en.wikipedia.org/wiki/ITunes_version_history (accessed June 25, 2013); Declaration of Dr. John P. J.
Kelly in Support of Defendant’s Renewed Motion for Summary Judgment, April 18, 2011 (hereinafter, “Kelly
declaration in support of MSJ”), ¶22.

8

This number includes 10 versions after version 1, 26 sub-version updates, and 52 more minor updates. iTunes
4.7 is not counted as a separate version, but rather as one of the sub-version updates. See “iTunes version
history,” Wikipedia, The Free Encyclopedia, http://en.wikipedia.org/wiki/ITunes_version_history (accessed
June 25, 2013 and verified July 17, 2013).

HIGHLY CONFIDENTIAL

-7-

and later, music purchased from iTMS. From the time the iPod was first introduced, Apple
released new generations that added features to the first model, which it later called the “classic,”
and introduced new iPod models that had significantly different features. (For a description of
the evolution of iPod features over time, see Exhibit 3.) The following paragraphs offer a brief
overview of these product introductions and enhancements.
15.

iPod Classic. The first iPod (iPod Classic 1st Generation) was smaller and lighter than

other devices then on the market, had only 5GB memory, and was touted for its ability to put “a
thousand songs in your pocket.”9 Five months later, Apple introduced a second-generation iPod
with twice the memory, and four months after that Apple introduced their next update with twice
again the memory. In just nine months, the capacity of the iPod had quadrupled from 5GB to
20GB. Over the years, Apple continued to add memory to the iPod Classic: by September 2006
the iPod Classic 5th Generation was available with 80GB memory and a year later, the iPod
Classic 6th Generation was available with 160GB. (See Exhibit 4 for a summary of
enhancements to iPod memory.) At the same time, Apple continued to add features. In April
2003, Apple began to offer iPods with USB connectivity in addition to the FireWire cable that
they had offered initially.10 Over time, successive generations of the iPod Classic became
smaller and lighter, had more memory and longer battery life, and were generally sold at prices
lower than the models they replaced. Later, when photo and video were introduced, the iPod
began to feature high-resolution color screens that improved with each new generation. (For a
summary of the history of the iPod Classic, see Exhibit 5a.)
16.

iPod Mini. The iPod Mini was introduced in January 2004. It was half the size of the

original iPod. With 4GB of memory the Mini held 1000 songs – the smallest portable digital
music player to do so at the time. According to Apple’s press release, it could also transfer
music at the rate of a song per second, and it could be conveniently charged using either the

9

Apple Press Release, “Apple Presents iPod; Ultra-Portable MP3 Music Player Puts 1,000 Songs in Your
Pocket,” October 23, 2001 (Apple_AIIA00974636). See also http://www.vaughanpl.info/vortex/?p=5261.

10

Although FireWire was faster at the time the iPod was introduced, it was also limited by the fact that it was not
usually standard on personal computers. See, e.g.,
http://www.qimaging.com/support/pdfs/firewire_usb_technote.pdf.

HIGHLY CONFIDENTIAL

-8-

included FireWire or USB cable.11 The 2nd Generation Mini was introduced a year later. It was
smaller, lighter, held 50 percent more music and featured increased battery life of up to 18 hours,
and it sold for the same price as the previous model. While it was in production, the Mini was
one of the most popular electronic products on the market.12 (See Exhibit 5b.)
17.

iPod Shuffle. On January 11, 2005, Apple introduced the iPod Shuffle, a digital music

player based on iTunes’ shuffle feature, which randomly selected songs from the user’s library
for placement on the iPod. In replaying the music, the Shuffle “shuffled” the music so that users
got a new combination of music every time they listened. The Shuffle was smaller and lighter
than a pack of gum and was available for less than $100.13 (See Exhibit 5c.)
18.

iPod Nano. In September 2005, Apple introduced the iPod Nano as a replacement model

for the Mini. The Nano was a full-featured iPod that was smaller and lighter than anything then
on the market and featured a redesigned color screen.14 Industry commentators praised the Nano
for its innovative design.15 PC World named it one of five ground-breaking products in the
audio category in its 2006 World Innovations Awards.16 The 1st Generation Nano was 1.6 inches
wide, 3.5 inches long, .27 inches thick and weighed 1.5 oz.; it had a stated battery life of up to 14
hours; and the screen was a 1.5-inch (diagonal) liquid crystal display panel at 176×132
resolution. The Nano could also hold and display between 15,000 and 25,000 photos.17 The
following year, Apple introduced a completely redesigned 2nd Generation Nano with twice the
11

Apple Press Release, “Apple Introduces iPod mini; Smallest 1,000 Song Music Player Ever Comes in Five
Colors,” January 6, 2004 (Apple_AIIA00974840).

12

“iPod Mini,” Wikipedia, The Free Encyclopedia, https://en.wikipedia.org/wiki/IPod_Mini (accessed July 16,
2013).

13

Apple Press Release, “Apple Introduces iPod shuffle; First iPod Under $100,” January 11, 2005
(Apple_AIIA00974708).

14

Apple Press Release, “Apple Introduces iPod nano,” September 7, 2005 (Apple_AIIA00974603).

15

The Nano turned out to be very popular – more than a million were sold in the 17 days following its
introduction. Drew Turner, Daniel, “Apple hits $1 Billion in Profit for 2005,” eWeek, October 10, 2011,
http://www.eweek.com/c/a/Apple/Apple-Hits-1-Billion-in-Profit-for-2005/ (accessed July 16, 2013). PC World
named it one five ground-breaking products in the audio category in its 2006 World Innovations Awards.
“2006 PC World Innovations Awards,” PR Newswire, January 4, 2006, http://www.prnewswire.com/newsreleases/2006-pc-world-innovations-awards-winners-unveiled-53127872 html (accessed July 16, 2013).

16

“2006 PC World Innovations Awards,” PR Newswire, January 4, 2006, http://www.prnewswire.com/newsreleases/2006-pc-world-innovations-awards-winners-unveiled-53127872 html (accessed July 16, 2013).

17

http://support.apple.com/kb/sp42.

HIGHLY CONFIDENTIAL

-9-

storage capacity and 24 hours of battery life for the same price as the previous generation.18
Apple continued to introduce new generations of Nanos with additional innovative features. In
September 2009, for example, it introduced the first Nano with a built-in video camera. (See
Exhibit 5d.)
19.

iPod Touch. The most recent addition to the iPod product line is the iPod Touch.

Introduced in September 2007, the Touch featured a multi-touch interface, together with
widescreen display and built-in Wi-Fi wireless networking, which allowed browsing and
wireless viewing of internet videos. In addition, users could browse, preview, and buy and
download songs from iTMS.19 A year later, Apple introduced the 2nd Generation Touch.
Smaller and lighter than the original, it featured a thin metal design and a 3.5-inch widescreen
glass display. With the new Touch, users could “listen to millions of songs, watch thousands of
Hollywood movies and now, thanks to the App Store [which opened at about the same time],
download and play hundreds of great games on their iPods.”20 (For a summary of the history of
the iPod Touch, see Exhibit 5e.)
3. The iTunes Store
20.

Apple opened its online music store (“iTMS”) in April 2003, which allowed users to

purchase DRM-protected music files that could be played on an Apple personal computer or,
after syncing with iTunes, on an iPod. The only way to access the iTMS was and is through
iTunes. In October 2003, Apple introduced iTunes for Windows, which allowed owners of
personal computers using the Windows operating system to use iTunes to purchase content from
iTMS and to sync that content on iPods. When the iTMS opened in April 2003, it offered
200,000 songs.21 By March 2004, there were more than 500,000 songs available, and in August
2004 Apple announced that its catalogue topped one million songs.22 On February 23, 2006,
18

Apple Press Release, “Apple Introduces the New iPod nano; World’s Most Popular Digital Music Player
Features New Aluminum Design in Five Colors & 24 Hour Battery Life,” September 12, 2006
(Apple_AIIA00974838).

19

Apple Press Release, “Apple Unveils iPod Touch,” September 5, 2007 (Apple_AIIA00974641).

20

Apple Press Release, “Apple Introduces New iPod touch,” September 9, 2008 (Apple_AIIA00974932).

21

Apple Press Release, “Apple Launches the iTunes Music Store,” April 28, 2003 (Apple_AIIA00974776).

22

Apple Press Releases, “iTunes Music Store Downloads Top 50 Million Songs,” March 15, 2004
(Apple_AIIA00974577); “iTunes Music Store Catalog Tops One Million Songs,” August 10, 2004
(Apple_AIIA00974782).

HIGHLY CONFIDENTIAL

- 10 -

Apple announced that one of its users had purchased and downloaded the one-billionth song
from iTMS. At the time, there were two million songs available.23 By January 2009, iTMS
offered approximately ten million songs, approximately eight million without DRM; and by
September 2012, it offered 26 million songs.24 The competition lagged far behind. When the
Real Music Store (RMS) opened in January 2004, it had a catalogue of 300,000 songs, which
was about half the size of the iTMS catalogue at the time.25 By the time the Amazon.com music
store debuted in September 2007 with approximately two million songs,26 iTMS had a catalogue
with more than five million songs.27
21.

Music is not the only thing available from the iTMS.28 On October 12, 2005, the iTMS

began to offer music videos, Pixar short films, and hit TV shows for $1.99.29 At the same time,

23

Apple Press Releases, “iTunes Music Store Downloads Top One Billion Songs; Scholarship at Juilliard School
of Music to be Created February 23, 2006 (Apple_AIIA00974735); Levy, Steven, The Perfect Thing: How the
iPod Shuffles Commerce, Culture, and Coolness, hereinafter “The Perfect Thing,” pp. 135-136.

24

Apple Press Release, “Changes Coming to the iTunes Store,” January 6, 2009,
http://www.apple.com/pr/library/2009/01/06Changes-Coming-to-the-iTunes-Store html (accessed July 16,
2013); “Apple Unveils New iTunes; Featuring Dramatically Simplified Design & Seamless iCloud Integration,”
September 12, 2012, http://www.apple.com/pr/library/2012/09/12Apple-Unveils-New-iTunes.html (accessed
July 16, 2013).

25

CNET News, “Real offers new tech, song store,” January 7, 2004, http://news.cnet.com/2100-10275136275.html (accessed July 16, 2013).

26

Amazon.com began to offer digital downloads in a public beta test in September 2007. However, music from
two of the major labels, Warner and Sony, was not available until January 2008. At that time Amazon’s music
catalogue reached over 3.1 million songs. See, e.g., ArsTechnica, “Amazon’s MP3 store brings more DRMfree music at lower prices than iTunes Store,” September 25, 2007,
http://arstechnica.com/uncategorized/2007/09/amazon-launches-public-beta-of-mp3-music-store/ (accessed
May 28, 2013); NY Times, “Amazon to Sell Warner Music Minus Copy Protection,” December 28, 2007,
http://www nytimes.com/2007/12/28/technology/28music html?_r=0 (accessed May 28, 2013); NY Times,
“Sony Joins Other Labels on Amazon MP3 Store,” January 11, 2008,
http://www nytimes.com/2008/01/11/technology/11sony html (accessed May 28, 2013); Antone Gonsalves,
“Amazon Adds Fourth Major Record Label To DRM-Free Music Store,” Information Week, January 10, 2008,
http://www.informationweek.com/amazon-adds-fourth-major-record-label-to/205602334 (accessed May 28,
2013).

27

Apple Press Release, “Apple Announces iTunes U on the iTunes Store; Free Content From Top Universities
Now Available.” May 30, 2007 (Apple_AIIA00974843).

28

With the release of iTunes 4.9 on June 28, 2005, Apple began to offer users the ability to discover, manage and
subscribe and listen to podcasts right on their computers. In just the first two days, iTunes customers subscribed
to more than one million Podcasts. Apple Press Release, “Apple Takes Podcasting Mainstream; Discover,
Subscribe, Manage & Listen to Podcasts Right in iTunes 4.9,” June 28, 2005 (Apple_AIIA00974620); Apple
Press Release, “iTunes Podcast Subscriptions Top One Million in First Two Days,” June 30, 2005
(Apple_AIIA00974799).

HIGHLY CONFIDENTIAL

- 11 -

Apple introduced the iPod Classic 5th Generation (video) so that consumers could take advantage
of the additional content. Over the first three weeks video was available, iTMS customers
purchased more than one million videos.30 The following September, Apple announced that the
iTMS would start selling movies. When the service first became available, the iTMS offered 75
movies from Walt Disney Pictures, Pixar, Touchstone Pictures and Miramax Films that
customers could download to watch on their computers and iPods.31 By January 2007, the iTMS
offered 250 feature films and 350 television shows.32 By the end of May, there were more than
500 movies.33 Also in September 2006, the iTMS began to offer downloads of popular video
games for fifth generation Classic iPods, all available for $4.99.34
B. The Introduction and Use of DRM
22.

Apple complied with this contractual requirement by developing and using its own proprietary
DRM called FairPlay. Other music stores used different proprietary DRM such as
RealNetworks’ Helix and Microsoft’s WMA DRM. The result was that DRM-protected music
files purchased from a vendor using one of these methods could only be loaded and played on
devices that supported that DRM technology. This meant that digital music downloads obtained
from online music stores that used Microsoft’s WMA DRM or downloads from the RMS could

29

Apple Press Release, “Apple Announces iTunes 6 With 2,000 Music Videos, Pixar Short Films & Hit TV
Shows,” October 12, 2005 (Apple_AIIA00974906). At the time, the Chairman of record label Interscope Geffen
A&M said: “Apple is giving music fans a great way to own their favorite music videos.”

30

Apple Press Release, “iTunes Music Store Sells One Million Videos in Less Than 20 Days,” October 31, 2005
(Apple_AIIA00974894).

31

Apple Press Release, “Apple Announces iTunes 7 with Amazing New Features,” September 12, 2006
(Apple_AIIA00974982). Sometime between June and September, the name of the iTunes Music Store was
changed to the iTunes Store. Nonetheless, I continue to refer to the store as the iTMS.

32

Apple Press Release, “iTunes Store Tops Two Billion Songs; 50 Million TV Shows & Over 1.3 Million Movies
Sold.” January 9, 2007 (Apple_AIIA00974561).

33

Apple Press Release, “Apple Announces iTunes U on the iTunes Store; Free Content From Top Universities
Now Available,” May 30, 2007 (Apple_AIIA00974843).

34

Apple Press Release, “Apple Announces iTunes 7 with Amazing New Features.” September 12, 2006
(Apple_AIIA00974982).

HIGHLY CONFIDENTIAL

- 12 -

not be played on iPods. Similarly, music obtained from the iTMS could not be played on
devices that used other (non-FairPlay) DRM technology.
C. RealNetworks RealPlayer with Harmony
1. RealNetworks (Real), Harmony and the Real Music Store
23.

RealNetworks operated the RMS. Like the iTMS, the RMS was an online store that sold

downloadable digital media files that could be played on personal computers and portable
devices. Similar to iTunes, RealPlayer was Real’s proprietary program for acquiring purchased
downloads from the RMS, managing digital files, and loading those files on a portable music
player.
24.

In July 2004, RealNetworks announced that version 10.5 of its RealPlayer jukebox would

include a new feature, which it called Harmony.35

.37

35

Cohen, Peter, “RealNetworks’ Harmony Promises iPod Compatibility,” PCWorld, July 26, 2004,
http://www macworld.com/article/1035237/harmony html (accessed May 28, 2013).

36

See Expert Report of David M. Martin Jr., PH.D., April 8, 2013 (hereinafter, “Martin report”), ¶27; Robbin
declaration in support of MSJ, ¶9; Kelly declaration in support of MSJ, ¶¶32-33.

37

Apple_AIIA00093860, pp. 3-5. See also Cohen, Peter, “RealNetworks’ Harmony Promises iPod
Compatibility,” PCWorld, July 26, 2004, http://www macworld.com/article/1035237/harmony html (accessed
May 28, 2013); Declaration of David F. Martin in Support of Plaintiffs’ Opposition to Apple’s Motion for
Summary Judgment, February 28, 2011, ¶¶26-27.

HIGHLY CONFIDENTIAL

- 13 -

had already dissuaded iPod owners from purchasing RMS music and using Harmony. This is
clearly a major problem for Plaintiffs’ theory, which requires that in the absence of iTunes 7
iPod owners would avail the RMS in substantial numbers. Yet, in the absence of iTunes 7,
Professor Noll doesn’t believe that iPod owners were using Harmony or RMS, an opinion that
accords with the evidence on RMS sales.
28.

Ignoring all of this, Professor Noll refers to RealNetworks’ claim that it sold 3 million

downloads during its three-week promotional sale when it launched Harmony in August 2004.42
But the relevant question is not what Harmony was doing in 2004, but rather what its sales were
at the time it was blocked in 2006.
D. Challenged Conduct
29.

On September 12, 2006, Apple introduced iTunes 7, which included many enhancements

and upgrades to iTunes functionality (see Exhibit 6).

30.

Two new iPod models were introduced at the same time as iTunes 7: the iPod Shuffle 2nd

Generation and the iPod Nano 2nd Generation. In addition, Apple added some features to the
iPod Classic 5th Generation (a.k.a. iPod video), which had been introduced the previous year, but
did not introduce a new generation of the Classic.

42

Noll declaration, p. 53, note 93.

43

See Martin report, ¶¶98-99.

HIGHLY CONFIDENTIAL

- 15 -

31.

32.

I understand that, as of the date of his April 3, 2013 report, Professor Noll did not know

which iPod models included the

and

and which did not.

44

Supplemental Declaration of Augustin Farrugia, July 3, 2013, ¶3; Martin report, ¶84.

45

Martin report, ¶¶21-22.

46

Martin report, ¶84; Supplemental Declaration of Augustin Farrugia, July 3, 2013, ¶¶2-3.

HIGHLY CONFIDENTIAL

- 16 -

E. The End of DRM
33.

As early as 1998, there was at least one paid service (eMusic) that offered paid

downloads of music without DRM protection.49 By September 2006, there were at least 24 sites
that offered DRM-free music, all but four of which offered fully or partially paid service.50 (See
Exhibit 7.) On February 6, 2007 — less than five months after the introduction of iTunes 7 —
Apple CEO Steve Jobs published an open letter called “Thoughts on Music.” In it, he called on
the major labels to allow Apple to sell their music without DRM protection: “This is clearly the
best alternative for consumers, and Apple would embrace it in a heartbeat. If the big four music
companies would license Apple their music without the requirement that it be protected with a
DRM, we would switch to selling only DRM-free music on our iTunes store.”51
34.

Two months after Jobs’ open letter, on April 2, 2007, EMI announced it was launching

new premium download service and that, effectively immediately, all of its digital catalog would
47

Declaration of Agustin Farrugia in Support of Defendant’s Renewed Motion for Summary Judgement, ¶¶31-32.

48

Noll deposition, p. 200:9-19; Exhibits 15.1, 15.2.

49

eMusic was a subscription service that primarily sold music from independent labels. See “eMusic,” Wikipedia,
The Free Encyclopedia, http://en.wikipedia.org/wiki/Emusic (accessed May 28, 2013).

50

Three of these – Kazaa, Morpheus and Limewire – were peer-to-peer file sharing sites, and all three were sued
by the record industry. Kazaa converted to a “legal” site in 2006; Morpheus filed for bankruptcy in 2008; and
Limewire was shut down by a judge in 2010. “Kazaa site becomes legal service,” BBC News,
http://newsvote.bbc.co.uk/mpapps/pagetools/print/news.bbc.co.uk/2/hi/technology/5220406.stm (accessed May
28, 2013); “Encyclopedia: Kazaa,” PCMag.com, http://www.pcmag.com/encyclopedia/term/45734/kazaa
(accessed May 28, 2013); “LimeWire,” Wikipedia, The Free Encyclopedia,
http://en.wikipedia.org/wiki/Limewire (accessed May 28, 2013); “Morpheus (software),” Wikipedia, The Free
Encyclopedia, http://en.wikipedia.org/wiki/Morpheus_(software) (accessed May 28, 2013); “National
Geographic unveils World Music downloads service,” PC Pro.co.uk,
http://www.pcpro.co.uk/news/90308/national-geographic-unveils-world-music-downloads-service (accessed
May 28, 2013).

51

Jobs, Steve, “Thoughts on Music,” February 6, 2007 (AIIA00093477).

HIGHLY CONFIDENTIAL

- 17 -

overcharge for all iPod models sold between September 12, 2006 (when iTunes 7 was introduced
and March 31, 2009 (when all music on the iTMS was
available without DRM).

IV.

PROFESSOR NOLL’S HEDONIC PRICE REGRESSION AND
INFERENCE OF DAMAGES

41.

Professor Noll uses a hedonic multiple linear regression in an attempt to estimate

damages by comparing iPod prices during the proposed class period to prices during a “before”
period that, he assumes, represents what would have occurred “but-for” the challenged conduct,
while at the same time attempting to control for factors such as product features and functionality
that would also affect iPod prices.60 Using his regressions, he claims to show that the
caused an anticompetitive increase in the price of all iPod models. His regression predicts that
the effect was immediate and constant, beginning on September 12, 2006 when the

was

introduced and lasting until March 31, 2009, the end of the class period. He also claims that the
percentage overcharge was exactly the same for each iPod model, even models that could not
invoke the
42.

.

I begin with a brief overview of regressions. Next I discuss the concept of “hedonic”

models of prices and how regressions have been used to estimate such models. With these
elements as background, I then turn to a description of Professor Noll’s regressions.
A. Multiple Linear Regression
43.

A multiple linear regression is a statistical technique for estimating the empirical

relationship between one measureable variable, call it P, and “multiple” other variables, X = {X1,
X2, X3, ….,XK}, which are often called “explanatory variables.” In the current context, P might
measure the price of a particular model (m) of MP3 player at a particular date (t), which we
could denote Pmt. For example, the retail price (P) of an iPod Touch with 8 GB of storage
capacity (m) in September 2007 (t) was $299, so Pmt = $299. X1 might be a continuous variable
that measures some valuable characteristic of the Touch, such as the megapixel resolution of its
screen, and so on.

60

Noll declaration, p. 71.

HIGHLY CONFIDENTIAL

- 21 -

44.

An indicator variable (or “dummy variable”) is used to indicate that a particular condition

is true or false. It takes the value 1.0 if the condition is true, and zero otherwise. For example,
let Xmt be an indicator for whether model m had a touch screen at date t. Then this indicator
would take the value Xmt = 1 if the model had a touch screen and Xmt = 0 otherwise. (The iPod
Touch in fact had a touch screen in September 2007, so Xmt = 1 for that model and date. On the
same date, the iPod Shuffle did not have a touch screen, so Xmt = 0 for the Shuffle.) Similarly, an
indicator variable for time periods where music downloads were DRM-free would take the value
1 during such periods, and zero otherwise.
45.

The adjective “linear” refers to the maintained assumption that in a given dataset that

records values of P and X, the relationship between P and the multiple X’s is linear. Then Pmt
may be expressed as a linear combination of the X’s:
(1)

Pmt = X1mtB1 + X2mtB2 + X3mtB3 + … + XKmtBK + Umt

In equation (1) the parameters Bj j=1,2,…,K are unknown constants that represent how much P
changes when a particular Xj increases by one unit, holding constant all the other Xi for i≠j. For
example, if “all else equal” the addition of a touch screen (X1mt increases from X1mt = 0 to X1mt =
1) was associated with a $50 increase in price, then B1 =$50. And so on.
46.

In addition to the linear combination of observed and measurable variables Xj, the right

side of equation (1) includes a “residual” quantity Umt, which represents the portion of Pmt that
cannot be represented by the linear combination of Xs that are included in the model. In general,
the most useful way to think about Umt is that it represents all of the “left out” factors that affect
P but were not included in the model, either because the person constructing the model chose not
to include them or because they are “unobserved” (i.e., not recorded in the available data).
47.

Given the model (1) for the relation between X and P, there are well-known regression

techniques available to estimate the unknown parameters Bj j=1,…K. The most basic of these
techniques is called “ordinary least squares” (OLS), and that is the technique Professor Noll used
in the hedonic regression model reported in Exhibits 13.1 and 13.2 of his report.61 Under very

61

In his deposition, Professor Noll appears to be confused about the technique he used to produce the estimates in
Exhibits 13.1 and 13.2. He claims that these exhibits show “generalized least squares” estimates with “robust”
standard errors that have been adjusted for heteroskedasticity. This is incorrect. His Exhibits 13.1 and 13.2

HIGHLY CONFIDENTIAL

- 22 -

specific assumptions, the OLS estimator bj of the unknown parameter Bj has certain “nice”
properties. One is that bj is “unbiased,” which means that while bj is itself a random variable that
will vary from one random sample to another, the expected value of bj is equal to the unknown
value of the parameter Bj. Put differently, if the OLS estimator is unbiased, then repeated
samples from the same population will generate a number of OLS estimates, but these estimates
will tend to cluster around the true value Bj.
When will the OLS estimator bj of Bj have this “nice” property of being unbiased? The

48.

key assumption is that the elements of U—the “excluded” variables—are not correlated with the
“included” variables X. The idea is quite simple and can be illustrated in the context of Professor
Noll’s regression. Suppose that iPod models that use the updated iTunes 7 also have higher
resolution displays compared to earlier models, but resolution is not included among the
explanatory variables, X. This means that display resolution is an omitted variable—in other
words, it’s in U. Then the estimated coefficient on iTunes 7 will “pick up” the effect of
resolution on price, because the indicator for iTunes 7 and the omitted factors in U are correlated
with one another. As a result, if the impact on price of higher resolution is positive and the
impact of the

on price is zero, the estimated coefficient on the iTunes 7 indicator variable

would be positive because it is “picking up” the effects of higher resolution. This is called
“omitted variables bias,” and it is, in fact, one of the many fatal flaws in Professor Noll’s
regression analysis.
I mentioned above that the least squares estimator bj of the unknown parameter Bj is itself

49.

a random variable, with a mean and a standard deviation, which in the case of regression
estimators is called the “standard error” of the estimator. The standard error SEj of bj is a
measure of the precision of the estimator—the confidence we can have that the true value Bj is
near the estimated value. SEj can also be estimated from the data under appropriate assumptions.
As explained below, Professor Noll’s implicit assumptions in this regard are wildly at odds with
the facts, which causes him to vastly overstate the precision and statistical significance of his
results.
report OLS results. Videotaped Deposition of Roger G. Noll, May 16, 2013, hereinafter “Noll deposition,” p.
196:4-22. The Appendix to his report contains other exhibits in which standard errors have been adjusted for
heteroskedasticity.

HIGHLY CONFIDENTIAL

- 23 -

B. Hedonic Models of Prices for Products with Changing Quality
50.

Economists have long recognized the difficulty of comparing prices over time for

products with changing attributes. Computers provide a good illustration. In 2004, a personal
computer with about 1 gigabyte (GB) of system memory, a 60 GB hard drive and a 15-inch
cathode ray monitor could be had for about $1200. By 2010, the same $1200 expenditure would
get a computer with 8 GB of memory, a 700 GB hard drive and 21-inch LCD monitor. The
nominal price of a standard computer system had not changed, but technical progress in storage,
memory and production made the 2010 model vastly superior. The “real” price of a standard
computer had fallen dramatically, even though the nominal price had not changed.
51.

Now consider iPods. In 2001, an original iPod Classic with 5 GB of storage capacity

sold for $399. By 2005 the same $399 nominal price fetched a Classic with photo and video
capabilities and 60 GB of storage, among other features. And by 2009, for a $249 nominal price
a consumer could purchase a 160 GB iPod Classic with further product enhancements. Like
computers, “real” prices of iPods were falling dramatically.
52.

As these examples illustrate, an analysis of product pricing that ignores the evolution of

product quality and features can be highly misleading. One way of dealing with this issue is to
employ “hedonic” models of product pricing, in which products are explicitly viewed as bundles
of changing attributes or characteristics.62 A hedonic regression model is in the (basic) form of
equation (1), above, where the XJ represent measures of product characteristics and the Bj
represent the “relative valuation” of such characteristics or qualities.

V.

PROFESSOR NOLL’S HEDONIC DAMAGE MODEL

53.

Professor Noll’s damages model is built on a hedonic regression model of iPod prices and

certain product attributes.63 The basic framework of this regression follows the form of equation
(1) above: a particular measure of iPod prices (the natural logarithm of price, ln(P)) is regressed
on a chosen set of product characteristics, X, and the estimated coefficients b are interpreted as

62

Griliches, Zvi, “Hedonic Price Indexes and the Measurement of Capital and Productivity: Some Historical
Perspectives.” NBER Working Paper 2634 (June 1988).

63

Noll declaration, pp. 71-72.

HIGHLY CONFIDENTIAL

- 24 -

measuring “the relative valuation of such qualities.”64 Putting it differently, Professor Noll starts
with the hedonic regression model described above and adds a set of indicators for the
occurrence of certain events that, in his view, affected the nature of competition between Apple
and other unspecified sellers of MP3 players. This set of “event” indicators is the centerpiece of
Professor Noll’s claim that he has measured the impact of Apple’s challenged conduct.
54.

Specifically, his model includes an indicator variable for the introduction of iTMS, which

is meant to measure the impact of iTMS on iPod prices (the iTMS coefficient), an indicator
variable for Harmony’s introduction to measure its impact on iPod prices (Harmony coefficient),
an indicator variable for iTunes 4.7 (Harmony blocked coefficient), an indicator variable for
iTunes 7 (iTunes 7.0 coefficient), as well as indicator variables for when competing online music
stores went DRM-free (competitors DRM free coefficient) and when iTMS went completely
DRM-free (iTMS all DRM free coefficient). Professor Noll constrains his model so that the
iTunes 7 variable estimates a single percentage overcharge on all iPod models sold after
September 12, 2006, whether the iPod model included

or not. The regression makes no

attempt to measure any impact of the reintroduction of Harmony, changes to iTunes or iTMS or
the introduction of the

.

55.
As I explain below, it is improper and
invalid to interpret these unreliable empirical “results” as evidence of overcharge or damage
from Apple’s challenged conduct. Professor Noll’s conclusions are unreliable and highly
misleading.
A. How Professor Noll Constructs His Data
56.

Professor Noll implements his pricing regression using a measure of the unit price of

individual iPod models and some of their characteristics. His price measure is constructed from
Apple’s transactional databases for resellers and direct purchasers, for which he estimates
separate damage models. For resellers, the data he uses for his regression encompass 2.14
64

Griliches, Zvi, “Hedonic Price Indexes and the Measurement of Capital and Productivity: Some Historical
Perspectives,” NBER Working Paper 2634 (June 1988), p. 1.

HIGHLY CONFIDENTIAL

- 25 -

million multi-unit transactions that occurred between November 2001 and December 2010. (By
multi-unit, I mean the sale of more than one iPod in a single purchase/sale transaction.) For
direct purchasers the data he uses for his regression include 36.9 million transactions — many of
which are multi-unit — that occurred between November 2001 and December 2010.65
57.

“Price” is not directly reported in either of these databases; instead the data report the

total number of units purchased and the amount paid in each individual transaction between
Apple and the buyer.

He does this for every such
transaction, for both resellers and direct purchasers. His reseller regression data are made up of
2.14 million actual multiunit transactions, but by counting each transaction hundreds or even
thousands of times, he effectively pretends that his regression model is using 113 million
independent price observations. This is a serious error. As I discuss below, by ignoring the fact
that his observations are not independent, Professor Noll exaggerates the precision of his
regression estimates and their statistical significance.
58.

It is important to understand how to interpret the regression estimates when this is done.

The dependent variable in his regression is the natural logarithm of the calculated price. A basic
mathematical “fact” about natural logarithms is that a change in the natural logarithm of some
quantity is approximately the percentage change in that quantity.

This fact is useful in
interpreting Professor Noll’s regression results — if a variable in his regression has a coefficient

65

Noll declaration, Exhibits 13.1 and 13.2.

HIGHLY CONFIDENTIAL

- 26 -

of .05, for example, this means that a unit (1.0) change in that variable is associated with (not
“caused”) a price that is about five percent higher.
59.

Professor Noll’s regression estimates are displayed in two of his exhibits—Exhibit 13.1

of his report shows his estimates for resellers, and Exhibit 13.2 shows his results for direct
purchasers. For completeness, I include Professor Noll’s Exhibits 13.1 and 13.2 as Exhibits 12a
and 12b. The main results from my review of his regressions will also be summarized in two
exhibits, Exhibit 13a for resellers and Exhibit 13b for direct purchasers. Unlike Professor Noll’s
tables, each of which reported results from only one regression, Exhibits 13a and 13b report
results from many regressions that correct Professor Noll’s errors. Because of this, for each
regression model shown in Exhibits 13a and 13b, I report the estimates for only the most relevant
variables in Professor Noll’s model. Complete results for all the variables in every regression in
my review are reported in Appendix D. (See Exhibits D1a and D1b].
60.

For clarity of exposition, I will focus much of my discussion on Professor Noll’s reseller

regression results. Nearly identical arguments apply to his direct purchaser regression, which I
will note along the way.
B. What Professor Noll Claims to Find
61.

Exhibit 12a (Noll Exhibit 13.1) shows Professor Noll’s reseller sales regression results.

His estimate of “overcharge” is based on the coefficient on the “itunes7_0” indicator.

This is the overcharge
percentage that Professor Noll uses in his Exhibit 14 to calculate damages on all reseller
transactions.
62.

Along with each coefficient estimate, Professor Noll also reports the standard error he

calculates for that estimate. In general, the standard error (standard deviation) of a coefficient
measures the precision with which the coefficient in question is estimated. When the standard
error is very small, the estimate is correspondingly very precise. Economists are also typically

66

Noll declaration, p. 81.

HIGHLY CONFIDENTIAL

- 27 -

interested in expressing the degree of confidence they have that the estimated value of the
coefficient did not arise by chance when the true effect of the variable in question is actually
zero. This can be measured by taking the ratio of a coefficient estimate to its standard error,
which is called the “t-ratio” or “t-statistic” for that estimate. Intuitively, the t-statistic measures
the distance between the estimated value of the parameter in question and zero, measured in
standard deviations. When the t-statistic is large, the estimated coefficient is “far” from (many
standard deviations away from) zero, which increases our confidence that the true effect of that
variable is not zero.
63.

When sample sizes are sufficiently large, as they are in this case, a common benchmark

for saying that an estimated coefficient is “statistically significant” is that it has a t-statistic that is
2.0 or larger (t ≥ 2.0) so that the estimated coefficient is at least twice its standard error. A tstatistic of 2.0 corresponds to approximately a five percent (p = .05) probability that a coefficient
estimate as large as the one obtained could have arisen by chance if the true value of that
coefficient is zero. This would sometimes be referred to as “statistical significance at the five
percent level.” Larger values of the t-statistic correspond to rapidly declining probabilities that
the true effect of the variable in question is zero. For example, a t-statistic of t = 2.58
corresponds to a 1-in-100 (p=.01) probability of arising by chance—”statistical significance at
the 1 percent level.” A t-statistic of t = 4.9 corresponds to a one-in-one million chance (p =
.000001).

VI.

PROFESSOR NOLL’S CLAIMS OF PRECISION AND
STATISTICAL SIGNIFICANCE ARE INCORRECT

64.

In his Exhibits 13.1 and 13.2, Professor Noll places three stars (***) next to each of his

estimated coefficients, which he points out “Denotes statistical significance at the 1% level.”67
This means that all of his estimated t-statistics are greater than 2.58 in absolute value. As
Professor Noll puts it: “The coefficients on indicators of market conditions all are highly
significant.”68 They are so significant, in fact, that Professor Noll should have noticed that
something is seriously wrong with his model—his claims of precision and statistical significance
67

Noll declaration, Exhibits 13.1, 13.2.

68

Noll declaration, p. 90.

HIGHLY CONFIDENTIAL

- 28 -

67.

In Exhibit 13.1 of his report (my Exhibit 12a), Professor Noll reports that his Reseller

Sales regression is based on 2.14 million “observations,” but this is highly misleading. An
accurate statement is that his data contained 2.14 million multi-unit transactions, each of which
allowed him to calculate a single price, as in the example above. When he ran his regression, he
“frequency weighted” each of these 2.14 million transactions by the number of identical iPods
represented in each transaction. This is equivalent to treating a single transaction involving N
iPods for which he calculated a single average price P as N independent observations, each of
which coincidently had the same price P.
But of course there
are not 28,050 independent observations; Professor Noll has simply counted the one piece of
information from this single transaction 28,050 times. The result is that the regression procedure
“thinks” it is dealing with far more than 2.14 million observations— in fact, Professor Noll’s
Reseller Sales regression output is calculated as if he had 113 million independent price
observations, though he does not report that number in his declaration.70
68.

In his deposition, Professor Noll was asked whether the 2.14 million observations

reported in his Exhibit 13.1 was the number of observations used for his standard error
calculations.
Q: When you do the log regression, Exhibit 13.1, you report that you had 2.1 million
observations.
A: Yes.
Q: Is that the number that you used for the denominator of your standard errors—for your
standard error calculations?
A: Probably, I mean, yes.
Q: Is that an appropriate number to use?
A: I mean, remember, it’s not quite that simple, because the quantity went into
regression, but yes, you know. These are heteroskedastic robust standard error estimates.
Q: So what formula did you use?
70

The numbers above refer to Professor Noll’s reseller regression. He took a similar approach in calculating the
standard errors in his regression of direct sales. In that case, Professor Noll had data on 36.9 million
transactions. However, after frequency weighting, he calculated his standard errors as if he had 42.4 million
independent observations. See Noll declaration, Exhibits 13.1 and 13.2; and Exhibits 13a and 13b to this report
(“Den DF” line, column 1).

HIGHLY CONFIDENTIAL

- 30 -

A: It’s a quantity weighted - it’s - the observations are quantity weighted, so it
doesn’t - it’s not 2.1 million. It’s a quantity weighting of all the observations.71

69.

This “quantity weighting” (or as it is more generally called, “frequency weighting”) is a

serious error. The user’s guide for SAS, the statistical software package that Professor Noll used
for his regression analysis, clearly explains the meaning of frequency weighting and its
consequence for the assumed number of observations on which the model is based:
When a FREQ statement appears, each observation in the input data set is assumed to
represent n observations, where n is the value of the FREQ variable. The analysis
produced when you use a FREQ statement is the same as an analysis produced by using a
data set that contains n observations in place of each observation in the input data set.
When the procedure determines degrees of freedom for significance tests, the total
number of observations is considered to be equal to the sum of the values of the FREQ
variable.72

An online Stata tutorial on “Choosing the Correct Weight Syntax” posted by the Carolina
Population Center at the University of North Carolina at Chapel Hill contains a specific warning
about the consequences of frequency weighting. 73 Referring to frequency weights, or
“fweights,” this tutorial specifically states:
Do not use fweights to specify sampling weights. Your variance of estimates, p-values
and standard errors will be computed incorrectly.74

70.

My inspection of Professor Noll’s computer code confirms that he did, indeed, use

frequency weights as sampling weights in his regression analysis. This is a major error when, as
here, the data consist of multi-unit aggregates in which each unit is an exact replica of the others.
Then each of these price “observations” from a particular transaction is (by construction)
perfectly correlated with the others. This matters because Professor Noll assumed exactly the

71

Noll deposition, p. 197:6-9. Professor Noll is also confused about what types of standard errors he actually
reported in Exhibits 13.1 and 13.2. They are not “heteroskedastic robust standard errors.” They are ordinary
least squares standard errors, calculated under the assumption that the 113 million observations are independent
and that their residuals have common variance. The footnotes to his report indicate that he also ran these
regressions allowing for robust standard errors, but those are not the results reported in Exhibit 13.1. (See Noll
declaration, notes 127 and 133.)

72

SAS/STAT(R) 9.2 User’s Guide, Second Edition, “FREQ Statement.” Available at http://support.sas.com/
documentation/cdl/en/statug/63033/HTML/default/viewer htm#statug_reg_sect011 htm.

73

Stata is another one of the most prominent statistics and econometrics software packages.

74

“Choosing the Correct Weight Syntax,” Carolina Population Center at the University of North Carolina at
Chapel Hill, http://www.cpc.unc.edu/research/tools/data_analysis/statatutorial/sample_surveys/weight_syntax
(accessed June 20, 2013).

HIGHLY CONFIDENTIAL

- 31 -

opposite in calculating “the variance of estimates, p-values and standard errors” for his model —
he assumed that literally millions of identical price “observations” are statistically independent,
when they clearly are not.
71.

The fundamental problem is that Professor Noll assumes his “113 million” observations

are statistically independent of one another, but they are not. Professor Noll appears to be
confused about the concept of “independence” that is required to run his regression in the way
that he did. He appears to believe that the independence of his price observations depends on
whether buyers of iPods are making independent purchase decisions. Asked whether two
consumers purchasing the same product from the same store at the same price on the same day
provide independent observations on price, he responded: “The two consumers made their
decisions independently whether to buy the iPod. The price they were charged were [sic] the
same, but it doesn’t mean the observation of the transaction is not independent.” (Noll deposition
pp. 211:7-10). This is the wrong concept. The question isn’t whether the individuals made
independent decisions about whether to purchase their iPods; it’s whether Apple set the prices of
those two iPods independently. This analysis is about the determination of prices.

Professor Noll’s answer reveals
that he has it backwards — he’s thinking of the wrong decision-maker.
72.

This erroneous assumption that the prices of 113 million iPods (in his reseller regression)

are statistically independent causes Professor Noll to vastly exaggerate what the data can show
and to totally misrepresent the accuracy and statistical significance of his results. In order to
understand the issue, one must first understand the strict assumptions of the “ordinary least
squares” (OLS) procedure he used to estimate his model. Estimates of standard errors in
regression models depend critically on what one assumes about the “covariance structure” of the
residuals, U, across observations—that is, what one assumes about how the residuals are
correlated with each other. The key but unstated assumption made by Professor Noll is that the
residuals in his data are independent of one another (uncorrelated), which in layman’s terms
means that if I knew that the residual (the portion of the price that cannot be represented by the
explanatory variables) for one iPod was, say, five percent, that information would not help me

HIGHLY CONFIDENTIAL

- 32 -

predict the residual for any other iPod in the data — my best guess would be zero, because that is
the expected value and Professor Noll assumes the residuals are independent.
73.

They are not independent as Professor Noll
assumed; they are perfectly correlated. As the saying goes: “If you’ve seen one, you’ve seen
them all.” The consequence is that Professor Noll is pretending to have much more information
than he actually does. The technical consequence of this exaggeration is that his standard errors
will be underestimated or, put the other way around, his claims about the precision and statistical
significance of his results will be greatly overstated.
74.

In the data analyzed by Professor Noll, the problem is deeper than the obvious (but

ignored) fact that all the iPods in a single transaction have the same price.

By construction, the residuals will be highly
correlated across transactions for a given buyer, but also across buyers. In the jargon of
econometrics, this means that the price observations for an iPod model at a given point in time
form a “cluster” within which the residuals are (highly) positively correlated both within and
across transactions.

HIGHLY CONFIDENTIAL

- 33 -

75.

An econometrician who neglects to account for this clustering will underestimate the true

standard errors of his results. That is what Professor Noll did when he ignored these facts and
falsely assumed that the price observations he created were independent of one another. As
Professors Joshua Angrist and Jorn-Steffen Pischke (2009) put it in their useful survey of modern
econometric methods:
A pillar of traditional cross section inference ….is the assumption that the data are
independent. Each observation is treated as a random draw from the population,
uncorrelated with the observation before or after. We understand today that this
sampling model is unrealistic and potentially even foolhardy. …We call this
correlation the clustering problem.75
76.
The same warning appears in the most recent edition of the American Bar Association’s
reference volume Proving Antitrust Damages: Legal and Economic Issues (2010):
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
errors 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.76
77.

In his deposition Professor Noll was adamant that clustering is not an issue for his model,

and he expressed some annoyance when he was asked about it:
Q: Did you cluster standard errors?
A: This is not something that requires a cluster analysis. Again, you’re perpetuating an
econometric mistake you’ve been making for two years. This is not a problem of
clustering.
…
Q: Did you cluster standard errors?
A: That would have been inappropriate. This is not a clustering problem.
…
75

Angrist, Joshua and Jorn-Steffan Pischke, Mostly Harmless Econometrics: An Empiricist’s Companion,
Princeton University Press, Princeton, New Jersey, 2009, pp. 293-294. Professor Noll cites Chapter 8 of this
book as one of the documents he considered in preparing his report, and Chapter 8 is the chapter from which
this quotation was taken. Noll declaration, Documents Considered #944.

76

ABA Section of Antitrust Law, Proving Antitrust Damages: Legal and Economic Issues, ABA Publishing,
Second Edition, 2010, pp. 145-6. The citation notes that standard methods of allowing for clustering “produce
consistent estimates of the standard errors even when there is no correlation among the error terms. In other
words, they work well in both situations.” (p. 147).

HIGHLY CONFIDENTIAL

- 34 -

A: I also didn’t fly to the moon because it’s not a fly-to-the-moon problem. It’s
inappropriate to do any cluster analysis on this data set, because it’s not a clustering
problem.
…
Q: So you didn’t do a cluster analysis?
A: There aren’t any clusters to do the analysis of. How can you do an analysis of
something that isn’t there?77

78.

Unfortunately, it is Professor Noll who is “perpetuating an econometric mistake.” If it is

in fact true that “there aren’t any clusters to do the analysis of,” he could easily have offered
evidence to prove his point: He could simply have shown that the residuals generated by his
model are uncorrelated and that clustering by, say, product family would not affect his standard
errors. Indeed, the American Bar Association volume cited above notes that standard
econometric methods of allowing for clustering, which are “used generally in practice” and
available in the software package used by Professor Noll, “produce consistent estimates of the
standard errors even when there is no correlation among the error terms. In other words, they
work well in both situations.”78 But he would not have been able to show this. As the examples
above illustrate, his assumption that the residuals in his model are independently distributed is
untrue in the extreme: the existence of clusters is obvious. However, one need not take my word
for it. One can demonstrate the existence of clusters directly from Professor Noll’s regression.
79.

I used the parameter estimates in Professor Noll’s reseller regression reported in Exhibit

13.1 of his report (my Exhibit 12a) to calculate the estimated residual for each of the 113 million
iPods in his reseller data. Then, within each product family and quarter, I divided the
observations randomly into two equally-sized groups and calculated the average residual within
each group. If Professor Noll’s assumption that his observations are independent is correct, then
these family-by-quarter clusters would be meaningless. In particular, the average residuals in
each family-quarter-group would satisfy the assumptions Professor Noll used in estimating his
model and standard errors. First, because the model assumes that the expected value of the
residuals is zero, the average value of the residual in each group should be very close to zero.

77

Noll deposition, pp. 186:6-187:22.

78

ABA Section of Antitrust Law, Proving Antitrust Damages: Legal and Economic Issues, ABA Publishing,
Second Edition, 2010, p. 147.

HIGHLY CONFIDENTIAL

- 35 -

Second, the Group-1 and Group-2 average residuals should be uncorrelated with one another —
in other words, knowing the Group-1 average residual for iPod Classics in the first quarter of
2005 should not help me predict the Group-2 residual for that same model sold in that same
quarter.
80.

The results of this exercise are shown in Exhibit 14a, which graphs each value of a

Group-1 average residual for a family-quarter cluster, measured on the horizontal axis, against
the corresponding Group-2 average residual measured on the vertical axis. The exhibit
demonstrates two important facts. First, the average within-cluster residuals are not closely
distributed around zero — they range from -0.45 to more than 0.30. With regard to withincluster correlation, if Professor Noll was correct we should find that the average residual for
Group 1 of each cluster does not help to predict the value for Group 2. In other words, the
relationship between them should have a slope of zero. But the exhibit shows exactly the
opposite — the within-cluster average residuals for Group 1 and Group 2 lie very close to a 45degree line. Put differently, if I know the Group-1 average residual within a family-quarter
cluster, I can almost exactly predict the Group-2 average residual within that cluster. These facts
are completely at odds with Professor Noll’s assertion that “there aren’t any clusters.”79 There
are clusters, and he ignored them. The existence of these clusters explains why Professor Noll’s
implausible claims of statistical significance, and his corresponding claim that clustering is not
appropriate in these data, are simply wrong.
81.

The consequences of Professor Noll’s mistake in calculating standard errors are shown in

column (2) of Exhibits 13a and 13b. Consider the reseller sales regression in Exhibit 13a. Using
standard and well-accepted econometric methods that are part of the regression package used by
Professor Noll, I calculated the standard errors of his model allowing for “clusters” of nonindependent price observations within product family and quarter. Because I have changed
nothing about Professor Noll’s regression or the variables in it, the coefficient estimates are
exactly the same as in his model.
The only thing that changes here is the claim one
can make about the precision of these estimates. The previous discussion indicates there is very
79

Repeating this exercise for Professor Noll’s direct sales regression leads to the same conclusions. (See Exhibit
14b.)

HIGHLY CONFIDENTIAL

- 36 -

high correlation of residuals within product family and quarter as a consequence of Apple’s
pricing practices and the construction of the data. Once I correct for this clustering, the standard
errors of his estimates are much larger than what Professor Noll claims and much more
reasonable as estimates of precision.
82.

This point is demonstrated by looking at the t-statistics associated with each of Professor

Noll’s coefficients. Recall that the t-statistic is the ratio of a coefficient to its standard error.
But the correct estimated standard error that
accounts for non-independence is 0.04216, which is 468 times greater than the standard error
Professor Noll claims.
That is,
it is not statistically distinguishable from zero, and thus there is no material or reliable evidence
of an “overcharge,” even in Professor Noll’s otherwise incorrect model.
83.

The same is true of Professor Noll’s direct sales regression.

Again, Professor Noll’s estimate of an “overcharge” is not
statistically distinguishable from zero.
84.

The bottom line is this: Professor Noll’s standard error calculations, and hence his claims

of “statistical significance,” are indisputably wrong. This conclusion alone is enough to
undermine his claims to have identified a meaningful “overcharge” associated with the
challenged conduct. But this failure to correctly represent the precision of his estimates is only
one of many mistakes that materially affect his results and claims.
A. Professor Noll’s “Before and After” Experiment: The Wrong But-For World
85.

The measured product characteristics in Professor Noll’s regression, such as the storage

capacity of a device or functionalities such as the ability to store and display photo or video files,
are used to control for “other factors” that affect the price of iPods. His real interest centers on a

HIGHLY CONFIDENTIAL

- 37 -

set of “event” variables that are meant to identify “market conditions” and the impact of Apple’s
challenged conduct on iPod prices. Specifically, Professor Noll views these variables as
representing a “before and after” comparison of iPod prices. As he describes it: “The ‘beforeafter’ method compares the prices of the reference products (here, iPods) before and/or after the
occurrence of the anticompetitive acts with prices in the damage period.”80 He goes on to
explain that “[t]he initial period before iTS was created and the period after the launch of the iTS
establish the ‘before’ period for measuring Apple’s market power prior to the events surrounding
the release of Harmony, the attempts to disable Harmony, and the movement to DRM-free
files.”81 Though this explanation is a bit confusing, the specification of his model demonstrates
that what Professor Noll is trying to say is that his “before” period — his “but-for” world
benchmark — is a three-month period between the launch of Harmony in July 2004 and the
release of iTunes 4.7 in late October of 2004.
86.

I understand the court has determined that the relevant FairPlay technology contained in

iTunes version 4.7 — which initially “blocked” Harmony for several months — is not
anticompetitive. This means that Professor Noll’s “before” but-for benchmark for measuring the
impact of the allegedly anticompetitive introduction of the

in September 2006 is incorrect.

Other flaws aside, the logic of Professor Noll’s “before and after method” implies that the
“before” period should be the period before September 2006, when the

did not exist,

FairPlay technology was legal, and Harmony was capable of interoperating with all iPod models.
87.

Professor Noll’s reseller sales regression is a good template for illustrating how he claims

to find an anticompetitive increase in iPod prices. Exhibit 15a is a timeline that shows the
periods when each of his six “event” indicators is “on” — that is, when the relevant indicator
takes a value of 1 instead of zero. On each line that indicates an event I also show Professor
Noll’s estimate of the impact that event had on iPod prices. The “omitted” time period is the
period before the opening of the iTMS in April 2003.82

80

Noll declaration, p. 71.

81

Noll declaration, p. 73.

82

Exhibit 15b contains a similar timeline for the direct sales regression.

HIGHLY CONFIDENTIAL

- 38 -

90.

The third indicator in the sequence of Professor Noll’s “events” is the introduction of

iTunes 4.7 in October 2004, which included FairPlay technology with which Harmony could not
initially interoperate. Professor Noll refers to this event as “harmony_blocked,” and in his
deposition he refers to it as “4.7.” He turns this indicator on in October 2004 and turns it off on
September 12, 2006, the day iTunes 7 was introduced. This choice is conspicuous because, as
shown in Exhibit 15a, “harmony_blocked/4.7” is the only one of his event indicators that is ever
turned off—once turned on, all the others remain on until the end of the data.

91.

Exhibit 15a shows that the next indicator to turn on is for iTunes 7, on September 12,

2006. On exactly the same day, Professor Noll turns off his indicator for iTunes 4.7.

92.

Professor Noll’s decision to turn off his “harmony_blocked/4.7” indicator when he turns

on his iTunes 7 indicator is very important. Recall the example above, where I examined what
Professor Noll’s regression would show if we hypothetically turned off the iTMS indicator on
the same day that the Harmony indicator turned on. In the example, by omitting the impact of
the iTMS during the Harmony period, the Harmony indicator was forced to “pick up” the
84

Noll declaration, p. 80.

HIGHLY CONFIDENTIAL

- 40 -

omitted effect of the iTMS. This biased the estimated effect of Harmony, and as a result, its
coefficient roughly doubled.
93.

Precisely the same thing happens here: by turning off the iTunes 4.7 indicator during the

period where the iTunes 7 indicator is on, Professor Noll forces the iTunes 7 indicator to pick up
the omitted “effect” of iTunes 4.7.

In other words, Professor Noll’s decision to turn off his
indicator for the existence of iTunes 4.7 technology has the effect of doubling his estimate of
impact and, therefore, damages, for the reseller class. Note that nothing else in his regression
changes — the coefficient on every other variable would have been exactly the same had he left
“on” the iTunes 4.7 indicator.
94.

The exercise in the previous paragraph is not just arithmetic. Professor Noll’s decision to

turn off the iTunes 4.7 indicator on September 12, 2006 precisely defines the “but-for” world
that he assumes would have existed in the absence of the challenged conduct. Specifically, by
turning off the iTunes 4.7 indicator he assumes that the FairPlay technology included in iTunes
4.7 through iTunes 6 would not exist in the but-for world, even though this was the legal
technology in use immediately prior to the introduction of iTunes 7. Instead, Professor Noll’s
but-for world is a very brief period immediately before the release of iTunes 4.7, i.e., it is the 3month period between July 26, 2004, when Harmony was first announced, and October 26, 2004,
when iTunes 4.7 was introduced. He does not explain why this is the appropriate but-for world
for measuring the incremental effect of the challenged

, or iTunes 7, or for anything else

relevant to this case. But this choice doubles his estimated impact of the

feature of iTunes

7 on resellers.
95.

In his Second Supplemental Declaration on Class Certification Professor Noll attempted

to explain his decision to turn off his iTunes 4.7 indicator upon the release of iTunes 7.85 He
85

Second Supplemental Declaration of Roger G. Noll on Class Certification, September 23, 2011, p. 10,

HIGHLY CONFIDENTIAL

- 41 -

said: “The iTunes 7.0 release replaced the prior version of iTunes software and contained new
code for disabling Harmony. If the iTunes 4.7 release was not anticompetitive, but the iTunes
7.0 release was anticompetitive, as plaintiffs allege, then the appropriate specification is for the
indicator variable for iTunes 4.7 to be reset to zero when it is replaced by iTunes 7.0.” While
this is nothing more than an assertion, the assertion itself is incorrect. Putting aside the other
problems with his model, the logically correct “before-and-after” experiment would compare the
level of iPod prices during the period of challenged conduct to the level of prices prior to that
conduct, when Apple’s practices were legal. This is the period prior to September 12, 2006,
when iTunes 4.7 through 6 were operative. Technically, in a regression this incremental effect
on prices is found by leaving on the iTunes 4.7 indicator, which then captures the level of prices
before September 2006, in which case the coefficient on the iTunes 7.0 indicator will measure
the difference between average prices before and after the introduction of iTunes 7. As I
explained above, Professor Noll’s incorrect procedure compares prices after the introduction of
iTunes 7 to those in a three-month window in 2004, which was two years earlier. But 2004 (or
any earlier period) cannot be the but-for world of Plaintiffs’ theory if technology and prices in
2005 and 2006 were legal.
96.

In other words, by turning off the “harmony_blocked/4.7” indicator in his direct sales regression,
Professor Noll overstates the claimed incremental effect of the

feature of iTunes 7 by a

factor of more than six. And of course his reported damages are overstated by this factor as well.
97.

The points just discussed are formally presented in column (3) of Exhibits 13a (for

resellers) and 13b (for direct purchasers). As indicated, the incremental effect of iTunes 7 in
Professor Noll’s model can be found by leaving “on” the indictor for “harmony_blocked/4.7,”
just as other indicators are left on.

HIGHLY CONFIDENTIAL

- 42 -

B. Professor Noll’s Confusion about the “Log of Time”
98.

As described by Professor Noll, “Moore’s Law” is an empirical regularity regarding

technical progress: “the amount of functionality that can be placed on a semiconductor of a given
size doubles every 18 months.”86 This is one of the reasons “prices for consumer electronics
generally fall through time.”87 To capture this effect in his model, Professor Noll includes a
variable he refers to as “the log of time” in his log price regressions. Professor Noll also offers
the negative estimated coefficient on “the log of time” in his regression as a test of his model’s
validity.
99.

Based on his discussions of this “log of time” variable in his report and in his deposition,

one can only conclude that Professor Noll is confused both about the use and meaning of
logarithmic variables in regression analyses and about their connection to economic theory.
While it is quite common in empirical research to control for the impact of technical progress
and other time-related factors by including a time trend, in nearly 40 years as an empirical
researcher and journal editor I have never seen an economist use the “log of time” in any
regression. Not even once. There is good reason for this: the “log of time” makes absolutely no
sense as an economic control variable. And as it turns out, had Professor Noll used more
conventional and widely accepted methods to control for technical progress, his results would
have been quite different and his “damage” estimates much smaller. These points require some
technical explanation.
100.

Consider Moore’s Law — as Professor Noll describes it in his report this “law” says the

functionality that can be placed on a given size semiconductor doubles every 18 months. Turned
into a prediction for prices of electronic goods, this means that each increment of time of a given
size — say a month or a year — should reduce costs and prices of an electronic device with
given functionality by a constant percentage amount. For example, if between 2002 and 2003
prices fall by 20 percent due to this form technical progress, then the constant pace of such
86

Noll declaration, p. 18.

87

Noll declaration, p. 80.

HIGHLY CONFIDENTIAL

- 43 -

progress implies that prices should fall by 20 percent between 2003 and 2004 or between 2007
and 2008. Each additional year reduces price by a constant percentage amount. Recalling that
the change in the logarithm of a given quantity is approximately the percentage change in that
quantity, this constant-percentage-change relationship between price and the passage of time
may be represented mathematically as ln(P) = C + At. In this equation, “t” is a measure of time
such as the number of months from a given starting point—t = 1, 2, 3, … — and “A” is the
constant percentage amount by which price changes between one month and the next. For
example, if A = -0.02 then price falls at a rate of 2 percent per month. In other words, the proper
representation of technical progress that causes prices to fall at roughly a constant rate — as
implied by Moore’s Law — is a linear relationship between the logarithm of price and time
counted in integer units, not “the log of time”.
101.

What would be the meaning of using the “log of time” — as advocated by Professor Noll

— rather than simply “time” to describe technical progress? The mathematical relationship
would be ln(P) = C + Bln(t), which would mean that each percentage increment of “time” causes
the same (B) percentage change in price. So suppose we start time at t=1 in November of 2001
as Professor Noll does in his data set. The change in ln(t) between month 1 (November) and
month 2 (December) of 2001 is ln(2) - ln(1) = .693.

In other words, Professor Noll’s
use of “the log of time” to represent unmeasured technical progress actually constrains his model
so that the effect of such progress is effectively zero during most of the period of interest, and
particularly during the period in which iTunes 7 is relevant.

HIGHLY CONFIDENTIAL

- 44 -

103.

In his deposition Professor Noll was asked about this odd choice to use “the log of time”

in his regression. His explanation was exactly backwards — there is no other way to put it.
Apparently thinking that he is offering a technical defense for why the “log of time” is an
appropriate explanatory variable, he actually explains why it is not appropriate. Here is his
answer, with emphasis added:
Q: And why did you decide to use the log of time rather than time in levels?
A: Because the — if you — there’s no reason to believe that you have an exponentially
increasing price in time. That the — if you, if you, you know if you think about if you
use level, if the dependent variable’s a log of price, and you’re measuring time by, you
know, the first - the first month is month one and the second month is month two. When
you go from one to two you’re doubling it. That’s a hundred percent increase, all right.
When you’re going from the hundredth month to the hundred-and-first month that’s a
one percent increase.
One would not expect that the – the effect of time would be such that going from 100 to
101 was one percent of going from one to two. The only – the way you get that to be a
constant rate of increase is to take the logarithm of time.88

104.

As I explained above, I completely agree that one would not expect “the effect of time

would be such that going from 100 to 101 was one percent of going from one to two.” But that
is precisely what Professor Noll forces his model to do when he uses the “log of time” as a
control for technical progress. His statement “The only way you get that to be a constant rate of
increase is to take the logarithm of time” is flatly wrong. He has it backwards. Using the
logarithm of time forces a non-constant rate of change, which, as I showed above, rapidly
approaches zero. The mathematical truth is that “the only way you get that to be a constant rate”
is to measure time in levels, not in logarithms. In response to further questions about “the log of
time” Professor Noll demonstrates complete confusion, cogently explaining why what he did is
exactly wrong.
Q: Why do you say that the choice of the starting time period is arbitrary?
A: Well, I could - I could start my – my first period could be 1842, and I could measure
time by every month since January 1842, all right. And in principle, that shouldn’t
affect anything. I mean what I pick as the starting date shouldn’t matter, because what
I’m interested in is the effect of technological change during the period of the data.89

88

Noll deposition, pp. 24:17-25:9, emphasis added.

89

Noll deposition, p. 27:4-12, emphasis added.

HIGHLY CONFIDENTIAL

- 45 -

105.

Professor Noll’s reasoning is correct—choosing January of 1842 as the starting date from

which time is measured “shouldn’t affect anything.” And it wouldn’t if he had measured time in
levels instead of logs. To be precise, had he chosen January 1842 as the starting date, then
November 2001 (the first month in his data) would be month 1919 instead of month 1, and
December 2001 would be month 1920. The level of time would change by one month between
November and December, and this would be true no matter what starting year was used. In
contrast, if time is measured from January of 1842 the passage of one calendar month between
November and December of 2001 changes “the log of time” by ln(1920) - ln(1919) = 0.0005,
which is vanishingly small. However, if time is measured from November of 2001 the exact
same passage of one calendar month changes “the log of time” by ln(2) - ln(1) = 0.693, which is
1330 times larger.
106.

Professor Noll correctly offers that “what I pick as a starting date shouldn’t matter.” If

time is measured in levels, the arbitrary choice of starting date will not matter. However, if one
uses his “log of time” variable the arbitrary choice of a starting date matters enormously. In
other words, Professor Noll’s deposition testimony is a clear acknowledgement that his reliance
on “the log of time” is indisputably wrong. At another point in this exchange Professor Noll
offers “that’s why you – you always use the log of time in a logarithmic equation.”90 As I noted
earlier, in nearly 40 years as a professional economist I have never seen a research paper that
uses “the log of time in a logarithmic equation,” because it would be wrong. Professor Noll’s
remark would have been correct had he said “never” instead of “always.”
107.

Correcting this error in Professor Noll’s reasoning and model is easy — all one needs to

do is use time measured in levels (months) instead of the “log of time” in his regression. The
effect of doing so is shown in column (4) of Exhibits 13a and 13b.

90

Noll deposition, p. 26:15-17.

HIGHLY CONFIDENTIAL

- 46 -

C. Additional Features and the “Quality” of Professor Noll’s Regression
108.

At page 80 of his report Professor Noll offers a way to “test” the quality of a regression:

“Another test of the quality of a regression is whether the estimated coefficients are consistent
with expectations derived from economic theory.” He then offers an example: “For example,
prices for consumer electronics generally fall through time due to the presence of ubiquitous
learning by doing and Moore’s Law. The positive coefficient on the logarithm of time bears out
this expectation.”

In any case, we have already established that the “logarithm of time” is a meaningless
construct for measuring Moore’s Law and technical progress, or for any other purpose.
109.

Professor Noll neglects to mention other, equally important estimates from his regression

that are not “consistent with expectations derived from economic theory.”

Further, if Professor
Noll’s standard errors are to be believed (and they are not — see discussion above), then these
estimates are highly statistically significant.
But there can be little doubt that these features are valuable to users, which is why
Apple puts them in iPods. He similarly “finds” that the number of downloads available from the
iTMS reduces iPod prices. According to Professor Noll’s view of economic theory these effects
should be positive, not negative as Professor Noll confidently estimates. If consistency with
theory is a test of a model’s quality, then these results indicate that Professor Noll’s regressions
fail multiple tests, and by a very wide margin.
110.

These “tests” reject the conceptual foundation of Professor Noll’s model, which is that it

should be consistent with theory — especially the Plaintiffs’ theory. But the conceptual
foundation of Professor Noll’s “before-and-after” experiment is incorrect in other ways as well.

HIGHLY CONFIDENTIAL

- 47 -

Most prominently, he fails to recognize that the challenged conduct is not the introduction of
iTunes 7 itself, but rather the introduction of a specific feature of iTunes 7. iTunes 7 had many
additions and enhancements that were unrelated to the challenged conduct.91 (See Exhibit 6.) In
addition, iTunes 7 was introduced at the same time as new iPod models that were upgrades to
previous versions, with enhanced functionalities.92 At best, his iTunes_7 indicator is a catch-all
for any and all things relevant to iPod pricing that are correlated with the time period in which
iTunes 7 or later versions existed, but that are not among the (very) limited product
characteristics he chose to include in his regression. Professor Noll does not explain why an
indicator for the mere existence of iTunes 7 would measure and identify the impact of the
allegedly anticompetitive

as opposed to the array of other features and functionalities of

both the iTunes 7 software itself and the hardware devices with which it operated.
111.

This raises the obvious question of why Professor Noll chose to include some product

attributes in his model and to exclude others. Beyond the short list of attributes that Professor
Noll included, Apple’s data contain information on many other measurable attributes of iPod
models that an economist would reasonably expect to affect pricing. For example, as indicated
in Exhibit 10, the data indicate whether a particular model and generation of iPod could be
attached by FireWire or USB, its weight, its battery life, the size of its display, its display
resolution, and the number of hours required to charge its battery. To the extent that any or all of
these measurable features are valuable — and why would they not be? — they should have been
included in Professor Noll’s model of iPod pricing. But they were not.93
112.

Column (5) of Exhibits 13a and 13b takes this step, adding additional controls for

measurable (and presumably valuable) features of iPods that Professor Noll chose to omit.94 In
91

See Exhibit 2 for a list of iTunes features included in each version and Exhibit 6 for a list of those in iTunes 7.

92

See Exhibit 3 for a description of the evolution of iPod features in successive models.

93

Professor Noll apparently was aware of the fact that he had failed to consider a number of characteristics – See
Expert Report of Dr. Michelle M. Burtis in Apple Inc.’s Opposition to Plaintiffs’ Motion to Exclude, May 2,
2011, ¶¶16-17; Noll January 18, 2011 declaration, pp. 39, 76-78, 81-82, and Exhibits 1-6; Noll April 7, 2011
deposition, pp. 87:18-88:1.

94

In choosing these characteristics, I used the following approach. I asked my staff to create a list of all iPod
models (classified by MPN) that Apple had introduced between 2001 and December 2010 (the latest date for
which I have usable price data). (Note: MPN stands for “Marketing Part Number” and is the most detailed
level at which information on iPod characteristics is available. “Marketing Part Number,” Wikipedia, The Free
Encyclopedia, http://en.wikipedia.org/wiki/Marketing_part_number (accessed June 14, 2013).) I then asked my

HIGHLY CONFIDENTIAL

- 48 -

both cases the data indicate these features should not have been excluded from the model – tests
of their joint significance solidly reject excluding these features from the regressions for both
resellers and direct purchasers. Just as importantly, including these features leads to results that
completely undermine any notion that iTunes 7 caused or allowed Apple to raise prices.

D. Professor Noll’s Regression Produces Implausible Results That Are Inconsistent
With The Basic Economics Of Plaintiffs’ Theory.
113.

The issue for damages is how and when the

might have affected

consumers’ willingness to pay for new iPods and ultimately the prices that Apple would have
charged for them. Exhibit 16 shows the way in which the

affected the

interoperability of each new iPod model as of September 12, 2006 when the
was introduced. Professor Noll forces his regression to find that the challenged update would
have had an immediate impact on prices of all iPods on the same date the update was introduced.
This is contrary to his and Plaintiffs’ theory, in which the update would cause the incremental
degree of “lock-in” would rise over time as iPod owners accumulated more iTMS music relative
to RMS.
114.

Any “lock in” created by the

would play out over time. For example, consider an

individual who purchased an iPod Nano 2nd Generation on September 12, 2006. The effect of
the

was zero on September 12, 2006, because the owner’s accumulated stock of iTMS

music had not changed. When this owner goes to replace that Nano at some future date and
assuming that that same owner had more iTMS music relative other music on her Nano, her

staff to compile available information on the characteristics of each of these models. In doing this search, they
reviewed information from Apple.com and Everymac.com. They also reviewed Apple press releases and Apple
price committee documents that were produced as part of the discovery in this case. When information was not
available or needed corroboration, they searched a variety of other websites. (For a list of the websites used and
a description of the process by which the search was conducted, see Appendix C.) Ultimately they collected
information on 39 different characteristics from which I selected for the regression those that logically appear to
be relevant for the explanation of prices and would not generate perfect collinearity with other variables already
included in the model.

HIGHLY CONFIDENTIAL

- 49 -

stock of DRM-protected music would be less portable than it otherwise would have been, which
might increase the likelihood of purchasing a new iPod instead of some other MP3 player. But
notice how distant in time this effect would be — the “lock-in” of some new owners of a
device affects their next purchase of an MP3 player.
115.

So even under Professor Noll’s theory, the September 2006 update would not have raised

iPod prices at that time. Indeed, Professor Noll notes that MP3 players are replaced roughly
every 18 months to two years, on average.95 It is therefore implausible that there would be a
material effect of

-induced “lock-in” on or soon after the introduction of iTunes 7 and the

. And even in a vaguely defined “long run,” if the incremental “lock-in” is to increase the
prices that Apple charges, then it must offset the reduced demand for iPods by other buyers who,
in Plaintiffs’ theory, would have preferred the option of purchasing music from RMS.
E. Professor Noll’s Regression Improperly Estimates A Single Average Overcharge
Across All Models
116.

Professor Noll improperly constrains his model by using a single variable for iTunes 7

that he has “on” for each model. Consequently, his model is forced to estimate a single average
percentage overcharge across all iPod models and generations of models, including models and
generations that never included the challenged update. His regression thus assumes that the
would not only cause overcharges on all other
models, but also that the percentage overcharge would be exactly the same regardless of their
feature sets or customer bases or ability to interoperate with Harmony. Estimating a single
average percentage overcharge will mask the fact that certain models might have been impacted
while others were not. It also hides the fact that his model does not include products that were
sold before and after iTunes 7. The iPod Touch, for example, was not sold in the “before”
period. This means that his regression cannot estimate the impact of iTunes 7 on the Touch by
comparing its prices before and after the introduction of iTunes 7, for the simple reason that the
Touch did not exist before iTunes 7. To overcome this, Professor Noll’s assumes that the iPod
Touch must have been impacted by the exact same percentage “overcharge” as models that were
sold before and after the introduction of iTunes 7, including models that were and were not

95

Noll declaration, pp. 4, 18.

HIGHLY CONFIDENTIAL

- 50 -

120.

At deposition, Professor Noll suggested that he may change his regression to fix this

problem by interacting the iTunes 7.0 variable with product-specific dummy variables, which
would allow for separate “effects” on each model.

121.

As explained above, if Plaintiffs are correct that use of Harmony and the RMS
were valuable to prospective iPod buyers, then models that maintained interoperability would
99

Noll deposition, pp. 85:12-87:24.

HIGHLY CONFIDENTIAL

- 52 -

Finally, a
model with product specific effects of iTunes 7 cannot estimate an effect on the price of the iPod
Touch, because the Touch did not exist prior to the introduction of iTunes 7.
122.

These flaws aside, and to demonstrate the extreme sensitivity of Professor Noll’s model,

column (6) of Exhibits 13a and 13b take the additional step of allowing for separate effects of
iTunes 7 on each iPod model other than the Touch, for which no such estimate is possible.

Again, this pattern is the opposite of Plaintiffs’ damage claim.
F. The Impact of DRM-Free Music on the Price of iPods
123.

Throughout this case, Professor Noll has made much of the fact that the widespread

availability of DRM-free music would dramatically limit Apple’s ability to lock in its iPod
customers.100 And, of course, once the iTMS began to offer DRM-free music, there could be no
additional lock in. Earlier, in the class certification portion of this case, he recognized that Apple
announced on January 6, 2009 that effectively immediately, eight million of the ten million
songs in the iTMS would be available without DRM protection and that “by April 1, 2009, all
digital audio files sold through iTMS could be purchased without FairPlay DRM.”101

In

deposition Professor Noll was asked: “Do you think that as of the time Apple was selling 80

100

See, e.g., Declaration of Roger Noll, July 15, 2008, p. 41; Reply Declaration of Roger Noll, October 19, 2009,
p. 44; Noll January 18, 2011 declaration, p. 59.

101

Noll January 18, 2011 declaration, p. 13; see also Reply Declaration of Roger Noll, October 19, 2009, p. 44
(citing Apple Press Release, “Changes Coming to iTunes Store,” January 6, 2009,
http://www.apple.com/pr/library/2009/01/06Changes-Coming-to-the-iTunes-Store html (accessed July 19,
2013)).

HIGHLY CONFIDENTIAL

- 53 -

percent of its music from its music store DRM-Free, that that had any impact on iPod demand?”
he answered: “Of course.”102 When asked where that impact would show up, he said:
It won’t show up. It won’t show up. What it will do is just reduce the magnitude to the
pre DRM-Free effect of 7.0, because the 7.0 effect would be less in the last few days of
the period, and so the overall effect on the coefficient would be to reduce it to reduce the
magnitude of the damages.
So – you have – because the 7.0 period is really long compared to the DRM-Free, it’s not
going to have much of an effect, but the effect will be to suppress the coefficient on – on
7.0.103

124.

Professor Noll’s conjecture is wrong—notwithstanding the many other flaws of his

regression and its interpretation, corrections to his “competitive market conditions” indicators for
the availability of DRM-free music have a large impact on his estimated “overcharges.” To show
this, I have reproduced the analyses in 13a and 13b, but instead of setting the indicator variable
for the iTMS being DRM-free to equal 1 as of April 1, 2009, I have set it to equal 1 three months
earlier, as of January 6, 2009, when 80 percent of iTMS music went DRM-free.104 The results of
these sensitivity tests are reported in Appendix D. [See Exhibits D2a through D2c and D3a
through D3c.] The key finding is that, again, these corrections uniformly reduce the estimated
“effect” of iTunes 7.

VII. PROFESSOR NOLL’S DAMAGE CALCULATIONS
125.

Each of the changes above has an effect on Professor Noll’s estimate of the coefficient on

his iTunes 7 indicator and thus on his estimate of damages. Although none of the estimated
effects from Professor Noll’s model are statistically significant, I have nonetheless used these
coefficients to re-estimate “damages” in order to show the sensitivity of his damage estimates.
The results of this exercise are shown in Exhibit 13c. As the exhibit shows, Professor Noll’s
damage estimates quickly vanish once his errors and omissions are corrected — and, in the end,
“damages” are negative because the estimated “effect” of iTunes 7 is negative. This is further

102

Noll deposition, p. 117:7-10.

103

Noll deposition, pp. 118:17-119:1.

104

To make sure that these results are robust, I tested these changes individually, and I find that each of them has
the effect of lowering Professor Noll’s coefficients across the board. See Appendix D, Exhibits D4a1 through
D4c3.

HIGHLY CONFIDENTIAL

- 54 -

evidence that Professor Noll’s approach to estimating impact and damages in this case is not
scientifically valid, accurate or reliable.

VIII. SOME FINAL POINTS: PROFESSOR NOLL’S REGRESSION HAS
NO CONNECTION TO THE CHALLENGED CONDUCT,
PLAINTIFFS’ THEORY, OR APPLE’S PRICING PRACTICES
126.

The fact remains that for all of these alternative specifications of Professor Noll’s model,

and for any others one can imagine, the entire exercise is futile. As explained above, any
connection between the

and Plaintiffs’ theory of “lock-in” depends on owners

of affected iPods purchasing less music from the RMS after the introduction of the
they would have had the

than

not been introduced, an effect that (if it exists at all) would

accumulate slowly over time and might affect future purchase decisions. But Plaintiffs have
provided no evidence that iPod owners ever purchased enough music from RMS to have a
material impact on iPod prices, and have certainly provided no evidence that the

had a

material incremental impact on whatever purchases there might have been. Prior to the
introduction of the

the RMS accounted for only about three percent of protected music

downloads overall, and this proportion would have been lower among iPod owners. With the
introduction of the

, the Plaintiffs’ theory suggests that the most intense users of Harmony

would be “locked out” from playing their music on new iPod models affected by the

and

thus would be unlikely to purchase an affected new iPod. Thus Plaintiffs’ entire case rests on a
small and unobserved population of iPod owners who had not used Harmony much in the past,
but allegedly would in the future. It is implausible that this unobserved and unmeasured group
had a material impact on the price of iPods, even in the long run.
127.

Professor Noll’s method for estimating damages is also inconsistent with the way in

which Apple set its prices.

105

Videotaped Deposition of Mark Donnelly, December 20, 2010 (“Donnelly deposition”), pp. 46:1-47:2.

HIGHLY CONFIDENTIAL

- 55 -

128.

As a result of the above analysis, my opinion is that there is no credible evidence that

class members were damaged by Apple’s challenged conduct.

___________________________________
Robert H. Topel

106

Donnelly deposition, p. 20:5-23.

107

Donnelly deposition, pp. 46:5-47:21, 49:21-50:12.

108

Donnelly deposition, pp. 72:2-74:7.

HIGHLY CONFIDENTIAL

- 56 -

Exhibit 1
Sources (cont):
http://www.apple.com/pr/library/2007/04/02Apple-Unveils-Higher-Quality-DRM-Free-Music-on-the-iTunes-Store.html
http://www.apple.com/pr/library/2007/06/28iPhone-Premieres-This-Friday-Night-at-Apple-Retail-Stores.html
http://www.apple.com/pr/library/2007/09/05Apple-Introduces-All-New-iPod-nano.html
http://www.apple.com/pr/library/2007/09/05Apple-Unveils-iPod-touch.html
http://www.apple.com/pr/library/2007/09/05Apple-Unveils-iPod-touch.html
http://www.apple.com/pr/library/2009/09/09Apple-Premieres-iTunes-9.html
http://www.apple.com/pr/library/2010/03/29iPad-Arrives-This-Saturday.html
http://www.apple.com/pr/library/2010/09/01Apple-Introduces-iTunes-10-With-Ping.html
http://www.apple.com/pr/library/2011/10/04Apple-to-Launch-iCloud-on-October-12.html
http://en.wikipedia.org/wiki/Ipod
http://en.wikipedia.org/wiki/Itunes
http://en.wikipedia.org/wiki/ITunes_Store
http://en.wikipedia.org/wiki/Apple_TV
http://www.wired.com/science/discoveries/news/2004/07/64341
http://news.cnet.com/RealNetworks-rekindles-iPod-tech-tussle/2100-1027_3-5685286.html
http://www.pcworld.com/article/115553/article.html
http://en.wikipedia.org/wiki/ITunes_version_history
http://en.wikipedia.org/wiki/IPad
http://en.wikipedia.org/wiki/Icloud

HIGHLY CONFIDENTIAL

Exhibit 2
iTunes Update History
iTunes
Version Update
1.0

Release Date

Major changes

1/9/2001

Original release based on SoundJam MP code

1.1

2/22/2001

External burners, improved visual effects, more supported CD burners

2.0

10/23/2001

Adds support for newly introduced iPod, CD burning improvements, equalizer/crossfader/sound enhancer added

3.0

7/17/2002

Smart playlists, more song list categories (including the My Rating column)

4.0

4/28/2003

Adds support for new iTunes Music Store, AAC audio codec, DVD burning, music sharing, GUI
improvements

4.1

10/16/2003

Music store/CD burning improvements, Windows support added, voice notes, on-the-go
playlists.

4.2

12/18/2003

AOL accounts with music store, GUI, and performance improvements

4.5

4/28/2004

iMix, party shuffle, CD insert printing, music store improvements, WMA to AAC conversion
(Windows only), Apple Lossless audio codec

4.6

6/9/2004

AirTunes support, minor improvements.

4.7

10/27/2004

4.8

5/9/2005

Video support, international music stores supported, security enhancements

4.9

6/28/2005

Podcasting, Motorola ROKR E1 mobile phone support added

5.0

9/7/2005

GUI refined, search bar improvements, parental controls, smart shuffle, iPod Nano support

6.0

10/12/2005

GUI/music store changes, blocks DRM remover utilities, transfer videos to 5th generation iPod
classic, included a complete redesign of FairPlay

7.0

9/12/2006

Video playback/purchasing improvements, iPod games, Major GUI changes, gapless playback
and album, sync purchased content from iPod to computer, Cover Flow added, support

7.1

3/4/2007

Apple TV support, additional 2G shuffle support, GUI improvements, fixes Windows Vista issues,
enhanced sorting options, full-screen Cover Flow

7.2

5/29/2007

Fully supports Vista, iTunes Plus introduced with 256 kbit/s DRM-free music tracks, iTunes U
introduced which offers free content from some of the top universities around the United States.
Also included GUI Update for Windows Vista

7.3

6/29/2007

Support for iPhone activation/synching, GUI changes/fixes. Changes sorting pattern

7.4

9/6/2007

7.5

11/5/2007

Support for iPod Touch, Classic (6G), Nano (3G), and adds interface art for new iPod Shuffle
colors. GUI improvements;
Allows activation of iPhones outside of the United States wherever activation is available, (e.g.
United Kingdom and Germany) as well as security and stability fixes. Also included is a GUI
update for Leopard, and the ability to add custom ringtones for free. Includes support for iPod
game Phase. Shows iPod battery level in source list (iPod Nano 3G, iPod Classic, iPod Touch, and
iPhone with 1.1.2 software)

7.6

1/15/2008

Rent movies from the iTunes Store. Transfer Apple TV purchases to your computer. Allows
manual management of music on iPhones. Added support for Windows Vista 64-bit

7.7

7/10/2008

Support for iPhone 3G, iOS 2.0 and the new App Store which features application downloads for
the iPhone and iPod Touch as well as enabling the two products to act as remotes for wireless
iTunes control

8.0

9/9/2008

Genius Sidebar and playlists, Grid View, HD TV shows, Shows capacity of Apps on iPhone/iPod
Touch on device summary tab, new default visualizer, more flexible podcast options and support
for second generation iPod Touch and 4th generation iPod Nano

Copying photos to iPod Photo, GUI/performance improvements, Windows taskbar minimizing,
updated FairPlay in effort to block hacking of music from the iTMS

HIGHLY CONFIDENTIAL

Exhibit 2
iTunes
Version Update

Release Date

8.1

3/11/2009

Support for the third generation iPod Shuffle, speed improvements for browsing large libraries
and the iTunes Store, as well as 'preparing to sync' and 'optimizing photos' for syncing to iPods
and iPhones, Party Shuffle has been replaced by iTunes DJ which now has the ability to receive
requests for songs, the ability to import/convert files and CDs to iTunes Plus format, better
performance when downloading iTunes Plus songs, accessibility improvements, Genius has been
expanded to cover TV shows and movies, refined parental controls and refined auto-fill options.
Supports Multi-touch gestures

8.2

6/1/2009

Supports iPhone 3GS and iOS 3.0 Software Update for the iPhone and iPod Touch. Includes many
accessibility improvements and bug fixes

9.0

9/9/2009

New UI and redevelopment of the iTunes Store using WebKit. Genius Mixes were added, as were
Home Sharing, iTunes LPs and iTunes Extras. Support for activation/syncing of iPod touch (late
2009). Music is automatically added to the library from a watched folder. 1-Click purchases.

9.1

3/30/2010

Adds support for iPad, adds the ability to sync and organize downloaded books between iPad and
the iTunes library, and Genius Mixes can now be renamed, rearranged, or removed.
"Applications" are renamed "Apps"

9.2

6/16/2010

Added ability to sync with iPhone 4. Also added ability to sync and read books with iPhone or
iPod touch with iOS 4 and iBooks 1.1. Added ability to organize and sync PDF documents as
books, and to read PDFs with iBooks 1.1 on iPad and any iPhone or iPod touch with iOS 4. Added
option to organize your apps on iOS 4 home screens into folders using iTunes. Speed up back-ups
while syncing an iPhone or iPod touch with iOS 4. Album artwork improvements make artwork
appear more quickly when exploring your library

10.0

9/1/2010

10.1

11/12/2010

Adds new social networking layer named "Ping". Adds support for iPod shuffle 4G, iPod nano 6G,
iPod touch 4G, and Apple TV (late 2010). Renamed AirTunes to AirPlay. Adds visual
improvements to list view. Improves performance. Adds additional support for VoiceOver Kit for
iPod. New application icon.
Bug fixes. Streaming to AirTunes speakers working again. Adds Twitter connectivity to Ping.
Adds printing support and support for devices running iOS 4.2

10.2

4/18/2011

Adds support for iPad 2, and iOS 4.3. Improves Home Sharing, allowing browsing and playback
of entire iTunes libraries on devices running iOS 4.3, and brings back the colored icons in the
Preferences window

10.3

6/6/2011

Adds support for iTunes in the Cloud (beta), allowing automatic downloading of purchased
content between iTunes and iOS devices, and downloading previously purchased music. Adds
support for iBookstore on the iTunes Store

10.4

7/20/2011

10.5
10.6

10/11/2011
3/7/2012

Adds support for Mac OS X Lion. It now allows users to take advantage of the Full-Screen App
capability. GUI slightly improved. Better integration with Windows Vista and Windows 7 (Aero
effects support).
Adds support for iPhone 4S, iCloud, iTunes in the Cloud, Wi-Fi Syncing, and iOS 5.
Adds support for iPad (3rd generation). Adds the ability to play 1080p HD movies and TV shows
from the iTunes Store. Higher bit rate songs can be converted to 128, 196, or 256 kbit/s when
syncing to iOS devices or iPods. Improvements for iTunes Match. Bug fixes

Major changes

Sources:
http://en.wikipedia.org/wiki/ITunes_version_history
http://www.apple.com/pr/library/2001/01/09Apple-Introduces-iTunes-Worlds-Best-and-Easiest-To-Use-Jukebox-Software.html
http://www.apple.com/pr/library/2001/10/23Apple-Announces-iTunes-2.html
http://www.apple.com/pr/library/2002/07/17Apple-Announces-iTunes-3.html

HIGHLY CONFIDENTIAL

Exhibit 2
Sources (cont.):
http://www.apple.com/pr/library/2003/04/28Apple-Launches-the-iTunes-Music-Store.html
http://www.apple.com/pr/library/2003/10/16Apple-Updates-iPod.html
http://www.apple.com/pr/library/2004/10/26Apple-Introduces-iPod-Photo.html
http://www.oldapps.com/itunes.php?old_itunes+4#changelog
http://gigaom.com/2005/05/09/itunes-48-released
http://www.apple.com/pr/library/2005/06/28Apple-Takes-Podcasting-Mainstream.html
http://www.apple.com/pr/library/2005/09/07Apple-Introduces-iTunes-5.html
http://www.apple.com/pr/library/2005/10/12Apple-Announces-iTunes-6-With-2-000-Music-Videos-Pixar-Short-Films-Hit-TV-Shows.html
http://www.apple.com/pr/library/2006/09/12Apple-Announces-iTunes-7-with-Amazing-New-Features.html
http://appleinsider.com/articles/07/03/05/apple_releases_itunes_71_quicktime_715_more
http://www.apple.com/pr/library/2007/05/30Apple-Announces-iTunes-U-on-the-iTunes-Store.html
http://www.apple.com/pr/library/2007/05/30Apple-Announces-iTunes-U-on-the-iTunes-Store.html
http://appleinsider.com/articles/07/06/29/itunes_7_3_supports_iphone_adds_apple_tv_photo_streaming
http://www.apple.com/pr/library/2007/09/05Apple-Unveils-the-iTunes-Wi-Fi-Music-Store.html
http://www.apple.com/pr/library/2008/01/15Apple-Premieres-iTunes-Movie-Rentals-With-All-Major-Film-Studios.html
http://www.apple.com/pr/library/2003/10/16Apple-Updates-iPod.html
http://www.apple.com/pr/library/2008/09/09Apple-Announces-iTunes-8.html
http://www.macworld.com/article/1139330/itunes.html

HIGHLY CONFIDENTIAL

Exhibit 3

Evolution of iPod Features
Model Update

Major Changes

Release Date

iPod Classic: Original iPod introduced October 23, 2001
iPod P68 Oct01

10/23/2001

First iPod introduced; portable device design; large capacity; featured Auto-Sync technology
Available capacity: 5GB

iPod P95 Mar02

3/21/2002

Added 10 GB model to the family; users now can personalize their iPods with the laser
engraving
Available capacity: 5GB, 10GB

iPod P68A/97 Jul02

7/17/2002

Introduced 20 GB model; compatible with Windows; 10GB model is physically smaller than the
previous comparable models
Available capacity: 10GB, 20GB

iPod Q14 Apr03

4/28/2003

Introduced 15 GB and 30 GB models; smaller and lighter than earlier iPods; prices are lower
than previous comparable models
Available capacity: 10GB, 15GB, 30GB

iPod Q14A Sep03

9/8/2003

Upgraded 15 GB and 30 GB models to 20 GB and 40 GB, respectively, while keeping the same
introduction prices
Available capacity: 10GB, 20GB, 40GB

iPod Q14A Sep03

1/6/2004

Added 15 GB back to the family; 15 GB model is smaller, lighter, and costs less than previous
comparable models
Available capacity: 10GB, 15GB, 20GB, 40GB

iPod Q21 Jul04

7/19/2004

New 20 GB and 40 GB models have longer battery life and are in smaller sizes than the previous
comparable models
Available capacity: 20GB, 40GB

iPod Photo P98 Oct04

10/26/2004

Allow photo browsing on the high-resolution color screeen; have more memory
Available capacity: 40GB, 60GB

iPod Photo P98A Feb05

2/23/2005

Replaced the 40 GB model with the 30 GB model; longer battery life
Available capacity: 30GB, 60GB

iPod Photo P98A Feb05

6/28/2005

Merged iPod and iPod photo lines; all iPods now equiped with color display
Available capacity: 20GB, 60GB

iPod Video M25 Oct05

10/12/2005

Enhanced color display with larger screen and higher resolution; smaller and lighter; faster
battery recharge; introduced at lower prices
Available capacity: 30GB, 60GB

iPod Video M25B Sep06

9/12/2006

Upgraded 60 GB model to 80 GB while maintaining the same size; enhanced color display with
larger screen and higher resolution; supported video playback; increased game support; longer
battery life for the 30 GB model than the previous comparable models; lower introduction price
Available capacity: 30GB, 80GB

iPod Classic N25 Sep07

9/5/2007

Introduced the 160 GB model; longer battery life for music, photo, and video playback while
keeping the same introduction prices
Available capacity: 80GB, 160GB

iPod Classic N25B Sep08

9/9/2008

Introduction at same price as 80 GB model of iPod Video M25B in Sep07
Available capacity: 120GB

iPod Classic N25C Sep09

9/9/2009

Greater capacity while keeping the introduction price the same as the 120 GB model of iPod
Classic N25B Sep08
Available capacity: 160GB

HIGHLY CONFIDENTIAL

Exhibit 3
Model Update

Major Changes

Release Date

iPod Mini: Introduced January 6, 2004
iPod Mini Q22 Jan04

1/6/2004

Introduced iPod mini, smallest portable music player available; lightweight; new design in
various colors; touch-sensitive click wheel controller; large capacity
Available capacity: 4GB

iPod Mini Q22B Feb05

2/23/2005

Added 6GB model; lower introduction prices; longer battery life
Available capacity: 4GB, 6GB

iPod Nano: Introduced September 7, 2005 to replace iPod Mini
iPod Nano M26 Sep05

9/7/2005

Available in black and white for both Mac and Windows users; held 1000 songs; thinner than a
standard #2 pencil; supported photo playback
Available capacity: 2GB, 4GB

iPod Nano M26 Sep05

2/7/2006

Introduced the 1 GB model
Available capacity: 1GB, 2GB, 4GB

iPod Nano N36 Sep06

9/12/2006

Introduced 8 GB model at same price as 4 GB model from the first generation Nano; longer
battery life; improved screen resolution; aluminum body with click wheel; smaller and lighter;
additional colors; lower prices for 2 GB and 4 GB models
Available capacity: 2GB, 4GB, 8GB

iPod Nano N46 Sep07

9/5/2007

New design; enhanced video user interface; larger screen with higher resolution; supported
video playback
Available capacity: 4GB, 8GB

iPod Nano N58 Sep08

9/9/2008

Additional colors; new design with a curved aluminum and glass enclosure; incorporate
"Genius" technology; lower introduction prices than previous comparable models
Available capacity: 8GB, 16GB

iPod Nano N33 Sep09

9/9/2009

New design; built-in video camera; larger screen with higher resolution; lower introduction
prices than previous comparable models
Available capacity: 8GB, 16GB

iPod Nano N20 Sep10

9/1/2010

New design; multi-touch interface; smaller and lighter than previous comparable models at
same introduction prices
Available capacity: 8GB, 16GB

iPod Shuffle: Introduced January 11, 2005
iPod Shuffle Q98 Jan05

1/11/2005

smaller and lighter than any other iPods; 512 MB model for under $100.
Available capacity: 512MB, 1GB

iPod Shuffle N98 Sep06

9/12/2006

Built-in clip; smaller, lighter, and more affordable than previous comparable model
Available capacity: 1GB

iPod Shuffle N98A Jan07

1/30/2007

Additional colors; built-in clip; smaller, lighter, and more affordable than the previous
comparable model
Available capacity: 1GB

iPod Shuffle N98C Sep07

9/5/2007

New colors
Available capacity: 1GB

iPod Shuffle N98E Feb08

2/19/2008

New colors; higher capacity while maintaining the same size and weight at a lower introduction
price than the 1 GB model of iPod Shuffle N98C Sep07; reduced price of 1 GB model
Available capacity: 2GB

iPod Shuffle N98F Sep08

9/9/2008

New colors; lower prices
Available capacity: 1GB, 2GB

iPod Shuffle D98 Mar09

3/11/2009

Greater capacity; smaller and lighter; faster battery recharge
Available capacity: 4GB

HIGHLY CONFIDENTIAL

Exhibit 3
Model Update

Major Changes

Release Date

iPod Shuffle D55 Sep09/
iPod Shuffle D98A Sep09

9/9/2009

Lower price; button controller with voice over; in-ear headphones with remote
Available capacity: 2GB, 4GB

iPod Shuffle N12 Sep10

9/1/2010

New design
Available capacity: 2GB

iPod Touch: Introduced September 5, 2007
iPod Touch N45 Sep07

9/5/2007

Multi-touch interface; built-in wifi wireless networking; 3.5-inch display; longer battery life for
audio and video playback
Available capacity: 8GB, 16GB

iPod Touch N45 Sep07

2/5/2008

Introduced 32 GB model
Available capacity: 8GB, 16GB, 32GB

iPod Touch N72 Sep08

9/9/2008

New design; longer battery life for audio and video playback; lighter than previous comparable
models
Available capacity: 8GB, 16GB, 32GB

iPod Touch N18 Sep09/
iPod Touch N72B Sep09

9/9/2009

Introduced 64 GB model; new design; longer battery life for audio and video playback; lighter
and lower introduction prices than the previous comparable models; greater memory and
onboard RAM capacity for 32 GB and 64 GB models
Available capacity: 8GB, 32GB, 64GB

iPod Touch N81 Sep10

9/1/2010

Higher screen resolution; included front-facing camera for FaceTime; longer battery life for
audio and video playback; lighter and smaller than previous comparable models
Available capacity: 8GB, 32GB, 64GB

Sources: iPod characteristics data in Murphy/Topel reports, Wikipedia and Apple press releases, including:
http://www.apple.com/pr/products/ipodhistory/
http://www.apple.com/pr/library/2002/07/17Apple-Unveils-New-iPods.html
http://www.apple.com/pr/library/2002/07/17Apple-Unveils-New-iPods.html
http://www.apple.com/pr/library/2003/04/28Apple-Introduces-New-iPods.html
http://www.apple.com/pr/library/2004/01/06Apple-Introduces-iPod-mini.html
http://www.apple.com/pr/library/2004/07/19Apple-Introduces-the-New-iPod.html
http://www.apple.com/pr/library/2004/10/26Apple-Introduces-iPod-Photo.html
http://www.apple.com/pr/library/2005/01/11Apple-Introduces-iPod-shuffle.html
http://www.apple.com/pr/library/2005/02/23Apple-Unveils-New-iPod-mini-Starting-at-Just-199.html
http://www.apple.com/pr/library/2005/06/28Apple-Merges-iPod-iPod-photo-Lines.html
http://www.apple.com/pr/library/2005/09/07Apple-Introduces-iPod-nano.html
http://www.apple.com/pr/library/2005/10/12Apple-Unveils-the-New-iPod.html
http://www.apple.com/pr/library/2006/05/23Nike-and-Apple-Team-Up-to-Launch-Nike-iPod.html
http://www.apple.com/pr/library/2006/09/12Apple-Introduces-the-New-iPod-nano.html
http://www.apple.com/pr/library/2006/09/12Apple-Unveils-the-New-iPod-shuffle.html
http://www.apple.com/pr/library/2006/09/12Apple-Introduces-the-New-iPod.html
http://www.apple.com/pr/library/2007/09/05Apple-Unveils-iPod-touch.html
http://www.apple.com/pr/library/2007/09/05Apple-Introduces-New-iPod-classic.html
http://www.apple.com/pr/library/2007/09/05Apple-Introduces-All-New-iPod-nano.html
http://www.apple.com/pr/library/2008/09/09Apple-Introduces-New-iPod-nano.html
http://www.apple.com/pr/library/2008/09/09Apple-Introduces-New-iPod-touch.html
http://www.apple.com/pr/library/2009/09/09Apple-Introduces-New-iPod-nano-With-Built-in-Video-Camera.html
http://www.apple.com/pr/library/2010/09/01Apple-Introduces-New-iPod-touch.html
http://www.apple.com/pr/library/2010/09/01Apple-Reinvents-iPod-nano-With-Multi-Touch-Interface.html

HIGHLY CONFIDENTIAL

Exhibit 4
iPod Timeline by Memory Size
2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

O N D J F MAM J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F M

0.5 GB

0.5 GB

1 GB

1 GB

2 GB

2 GB

4 GB

4 GB

5 GB

5 GB

6 GB

6 GB

8 GB

8 GB

10 GB

10 GB

15 GB

15 GB

16 GB

16 GB

20 GB

20 GB

30 GB

30 GB

32 GB

32 GB

40 GB

40 GB

60 GB

60 GB

64 GB

64 GB

80 GB

80 GB

120 GB

120 GB

160 GB

160 GB
iPod Classic
iPod Mini
iPod Nano
iPod Shuffle
iPod Touch

Note: Special editions of U2, Harry Potter, and Product Red are excluded from this exhibit.
Source: Price change data, Price Committee Documents, Apple Press Release, Direct Sales Data, Everymac.com.

HIGHLY CONFIDENTIAL

Exhibit 5a

iPod Timeline of iPod Classic
Family

Start Date Memory

IPOD CLASSIC N25C GOOD

9/9/2009
9/9/2008
9/5/2007
9/5/2007
9/12/2006
9/12/2006

30

IPOD VIDEO M25 BEST

10/12/2005

60

IPOD VIDEO M25 BETTER

10/12/2005

30

IPOD PHOTO P98A GOOD

6/28/2005

20

IPOD PHOTO P98A BEST

2/23/2005

60

IPOD PHOTO P98A BETTER

2/23/2005

30

IPOD PHOTO P98 BEST

10/26/2004

60

IPOD PHOTO P98 BETTER

10/26/2004

40

IPOD Q21 BEST

7/19/2004

40

IPOD Q21 BETTER

7/19/2004

20

IPOD Q14B GOOD

1/6/2004

15

IPOD Q14A BEST

9/8/2003

40

IPOD Q14A BETTER

9/8/2003

20

IPOD Q14 BEST

4/28/2003

30

IPOD Q14 BETTER

4/28/2003

15

IPOD Q14 GOOD

4/28/2003

10

IPOD P97 BEST

7/17/2002

20

IPOD P97 BETTER

7/17/2002

10

IPOD P68A GOOD

7/17/2002

5

IPOD P95 BETTER

3/21/2002

10

IPOD P68 GOOD

10/23/2001

5

2007

2008

2009

2010

2011

O N D J F MA M J J A S O N D J F MA M J J A S O N D J F MA M J J A S O N D J F MA M J J A S O N D J F MA M J J A S O N D J F MA M J J A S O N D J F MA M J J A S O N D J F MA M J J A S O N D J F MA M J J A S O N D J F M

80

IPOD VIDEO M25B BETTER

2006

80

IPOD VIDEO M25B BEST

2005

160

IPOD CLASSIC N25 GOOD

2004

120

IPOD CLASSIC N25 BETTER

2003

160

IPOD CLASSIC N25B GOOD

2002

$249
$249
$349
$249
$349
$249
$399
$299
$299
$449
$349
$599
$499
$399
$299
$299
$499
$399
$499
$399
$299
$499
$399
$299
$499
$399

Note: Special editions of U2, Harry Potter, and Product Red are excluded from this exhibit.
Source: Price change data, Price Committee Documents, Apple Press Release, Direct Sales Data, Everymac.com.

HIGHLY CONFIDENTIAL

Exhibit 5b

iPod Timeline of iPod Mini
Family

Start Date Memory

IPOD MINI Q22B BEST

2/23/2005
2/23/2005
1/6/2004

4

2004

2005

2006

2007

2008

2009

2010

2011

O N D J F MA M J J A S O N D J F MA M J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F MA M J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F M

4

IPOD MINI Q22 BEST

2003

6

IPOD MINI Q22B BETTER

2002

$249
$199
$249

Note: Special editions of U2, Harry Potter, and Product Red are excluded from this exhibit.
Source: Price change data, Price Committee Documents, Apple Press Release, Direct Sales Data, Everymac.com.

HIGHLY CONFIDENTIAL

Exhibit 5c

iPod Timeline of iPod Shuffle
Family

Start Date Memory

IPOD SHUFFLE N12 BETTER 9/1/2010

2002

2003

2004

2005

2006

2007

2008

2009

9/9/2009
9/9/2009

4

IPOD SHUFFLE D98A BETTER 9/9/2009

$49

4

IPOD SHUFFLE D98A BEST

2

IPOD SHUFFLE D98 BEST

3/11/2009
9/9/2008

$79
$59

2

IPOD SHUFFLE N98F BETTER 9/9/2008

$99

4

IPOD SHUFFLE N98F BEST

1

IPOD SHUFFLE N98E BEST

2/19/2008
9/5/2007

$69
$49

1

IPOD SHUFFLE N98A BEST 1/30/2007

$79

2

IPOD SHUFFLE N98C BEST

1

IPOD SHUFFLE N98 BEST

9/12/2006
1/11/2005

1

$69
$79
$79

1

IPOD SHUFFLE Q98 BEST

2011

O N D J F MA M J J A S O N D J F MA M J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F MA M J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F M

2

IPOD SHUFFLE D55 BEST

IPOD SHUFFLE Q98 BETTER 1/11/2005

2010

0.5

$79
$149
$99

Note: Special editions of U2, Harry Potter, and Product Red are excluded from this exhibit.
Source: Price change data, Price Committee Documents, Apple Press Release, Direct Sales Data, Everymac.com.

HIGHLY CONFIDENTIAL

Exhibit 5d

iPod Timeline of iPod Nano
Family

Start Date Memory

IPOD NANO N20 BEST

9/1/2010
9/1/2010
9/9/2009
9/9/2009

2006

2007

2008

2009

2010

16

IPOD NANO N58 BEST

9/9/2008
9/5/2007
9/5/2007
9/12/2006
9/12/2006
9/12/2006

2

IPOD NANO M26C GOOD

2/7/2006

1

IPOD NANO M26 BEST

9/7/2005

4

IPOD NANO M26 BETTER

9/7/2005

2

$199

4

IPOD NANO N36 GOOD

$149

8

IPOD NANO N36 BETTER

$179

4

IPOD NANO N36 BEST

$149

8

IPOD NANO N46 BETTER

$179

8

IPOD NANO N46 BEST

2011

O N D J F MA M J J A S O N D J F MA M J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F MA M J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F M

8

IPOD NANO N58 ULTIMATE 9/9/2008

2005

16

IPOD NANO N33 GOOD

2004

8

IPOD NANO N33 BETTER

2003

16

IPOD NANO N20 BETTER

2002

$149
$199
$149
$249
$199
$149
$149
$249
$199

Note: Special editions of U2, Harry Potter, and Product Red are excluded from this exhibit.
Source: Price change data, Price Committee Documents, Apple Press Release, Direct Sales Data, Everymac.com.

HIGHLY CONFIDENTIAL

Exhibit 5e

iPod Timeline of iPod Touch
Family

Start Date Memory

IPOD TOUCH N81 BEST

9/1/2010
9/1/2010
9/1/2010
9/9/2009
9/9/2009
9/9/2009
9/9/2008
9/9/2008

16

IPOD TOUCH N72 GOOD

9/9/2008

8

IPOD TOUCH N45A BEST

2/5/2008

32

IPOD TOUCH N45 BETTER

9/5/2007

16

IPOD TOUCH N45 GOOD

9/5/2007

2009

2010

2011

O N D J F MA M J J A S O N D J F MA M J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F MA M J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F M

32

IPOD TOUCH N72 BETTER

2008

8

IPOD TOUCH N72 BEST

2007

32

IPOD TOUCH N72B GOOD

2006

64

IPOD TOUCH N18 BETTER

2005

8

IPOD TOUCH N18 BEST

2004

32

IPOD TOUCH N81 GOOD

2003

64

IPOD TOUCH N81 BETTER

2002

8

$399
$299
$229
$399
$299
$199
$399
$299
$229
$499
$399
$299

Note: Special editions of U2, Harry Potter, and Product Red are excluded from this exhibit.
Source: Price change data, Price Committee Documents, Apple Press Release, Direct Sales Data, Everymac.com.

HIGHLY CONFIDENTIAL

Exhibit 6
iTunes 7.0 Features
· Released 9/12/2006
· First version of iTunes to sell movies (All of the first 75 movies from 4 studio of the Walt Disney Company)
· Movies will become available on the iTunes Store the same day they are released on DVD, with new releases
priced at $12.99 when pre-ordered and during their first week of availability, and $14.99 thereafter, and library
titles available for just $9.99 every day
· Delivers video near-DVD quality at a resolution of 640x480, 4 times higher than the previous version
· Redesigned layout to better organize and enjoy digital music and video
· New Cover Flow which lets you visually browse through your music and video by cover art
· iPod can now be used to transfer content to different computers
· The iTunes Store now also offers downloads of popular video games for fifth generation iPods (New iPod
Classic only) available for $4.99 each
· Existing iPods can be updated with all features listed above
·

Sources:
Apple_AIIA00974436
http://www.apple.com/pr/library/2006/09/12Apple-Announces-iTunes-7-with-Amazing-New-Features.html

HIGHLY CONFIDENTIAL

Exhibit 7

S

P d

Online Music Service Summary
Y

dd

Y

P

Y
Y

d

Y
Y

B
B
B
B
B

Y
Y
Y
Y
Y

B

Y

d

Y
Y
Y
Y

B

Y

Y

Y

d

Y
Y
Y
Y
Y

d

Y
Y

d

Y

d

Y

W
Y
Y

d

Y
Y
Y
Y
Y

R

Y
Y

W

Y

W

d
Y

P d
R d
P
P P
P
P
d
P
P
R
R
d
R

Y
Y
Y
Y
Y
Y
Y
Y

d

Y
Y
Y
Y

R
R

Y

d

Y
Y

d

Y

R

Y
Y
Y

W
Y
Y

Y

d

Y
Y

R
R

R

d

HIGHLY CONFIDENT AL

Exhibit 7
Sources:
Addictech
Amazon MP3
Amie Street
Arkade
artistxite

Batanga (eLatinMusic)
Beatport
Bleep.com
Boomkat
BuyMusic
CDBaby
Cdigix
Classical Archives
Digital-Tunes
Discogs

eBay Digital Music Center
eMusic

FYE Download Zone
GoMusicNow

Grazemusic
HDtracks
iMesh
Indieburn
iTunes Music Store
Jamendo
Juno / Juno Download

Source 1
Source 2
http://www.trademarkia.com/addictech76553281.html
http://www.gizmag.com/go/7269/
http://en.wikipedia.org/wiki/Amazon_
MP3
http://en.wikipedia.org/wiki/Amie_Stre
et
http://www.arkade.com/
http://musowiki.net/index.php/ArtistXit
http://artistxite.com/
e
http://www.batanganetwork.com/batan
http://www.bizjournals.com/triad/storie ga-inc-announces-seamless-musics/2006/11/06/story6.html?page=all
purchase-experience-linked-directly-withhttp://en.wikipedia.org/wiki/Beatport http://www.beatport.com/
http://en.wikipedia.org/wiki/Bleep.com
http://www.boomkat.com/
http://www.webwiki.com/boomkat.com
http://usatoday30.usatoday.com/tech/n
http://en.wikipedia.org/wiki/BuyMusic ews/2003-07-28-buymusic_x.htm
http://www.cdbaby.com/About
http://en.wikipedia.org/wiki/CD_Baby
http://en.wikipedia.org/wiki/Cdigix
http://www.classicalarchives.com/about. http://en.wikipedia.org/wiki/Classical_
html
Archives
http://www.digital-tunes.net/
http://www.prleap.com/pr/73840/
http://en.wikipedia.org/wiki/Discogs
http://news.cnet.com/eBay-plugs-intohttp://www.ebaychatter.com/the_chatter
digital-music-market/2100-1025_3/2008/03/digital-downloa.html
5270681.html
http://en.wikipedia.org/wiki/Emusic
http://www.emusic.com/listen/#/
http://www.prnewswire.com/newsreleases/trans-world-entertainmenthttp://www.trademarkia.com/fyeannounces-launch-of-fye-download-zonedownload-zone-78763697.html
74257247.html
http://en.wikipedia.org/wiki/GoMusicN http://hardnews1.ansci.usu.edu/archive
ow
/dec2006/121306_15cents.html
http://www.billboard.biz/bbbiz/retail/t
rans-world-to-launch-digital-servicehttp://news.minnesota.publicradio.org/f
1003571570.story
eatures/2005/10/26_horwichj_graze/
http://www.berkleegroove.com/2010/04
https://www.hdtracks.com
/29/david-chesky-hd-tracks/
http://en.wikipedia.org/wiki/Imesh
http://www.prweb.com/releases/2005/0
7/prweb258343.htm
http://en.wikipedia.org/wiki/ITunes_St
ore
http://www.jamendo.com/en
http://en.wikipedia.org/wiki/Jamendo
http://en.wikipedia.org/wiki/Juno_Reco
rds
http://www.junodownload.com/

HIGHLY CONFIDENTIAL

Source 3

http://website.informer.com/artistxite.co
m

http://www.underconsideration.com/sp
eakup/archives/001534.html

http://www.billboard.biz/bbbiz/retail/t
rans-world-to-launch-digital-service1003571570.story

http://en.wikipedia.org/wiki/Chesky_R
ecords

http://www.junodownload.com/welcom
e_to_junodownload/

Source 4

Exhibit 7
Sources:
Kazaa

Lifeway Stores
LimeWire

Misrolas

mMode Music Store
Morpheus

mp3tunes.com
Mperia
MSN Music
Music Giants

Music Rebellion

Musica360

MusicNOW
Napster 2.0

National Geographic World Music
Optus Music Store

Pandora Radio
Pass Along
PayPlay.FM

Source 1
Source 2
Source 3
http://deadspin.com/titus-young-willtake-a-nap-in-your-at-t-store496122722?utm_campaign=socialflow_dea
dspin_twitter&utm_source=deadspin_twi
tter&utm_medium=socialflow
http://www.baptiststandard.com/index.
php?option=com_content&task=view&id
=1249&Itemid=131
http://en.wikipedia.org/wiki/LimeWire
http://www.billboard.com/features/mis
rolas-com-shuts-download-storehttp://www.billboard.com/features/mis
1003844475.story#/features/misrolas-com-http://www.theorchard.com/news/2005 rolas-mobile-links-with-at-tshuts-download-store-1003844475.story
_08C.htm
1003716969.story
http://news.cnet.com/AT38T-Wirelessopens-mobile-music-store/2100-1027_3- http://www.phonenews.com/cingular-to5396072.html?tag=item
shut-down-mmode-lbs-services-1117/
http://en.wikipedia.org/wiki/MMode
http://en.wikipedia.org/wiki/Morpheus
_(software)
http://news.cnet.com/MP3tunes.comshuns-digital-rights-management/2100http://www.mp3tunes.com/cb/about/ 1027_3-5569293.html
http://en.wikipedia.org/wiki/BitPass
http://en.wikipedia.org/wiki/MSN_Mus http://news.cnet.com/8301-10784_3ic
9926476-7.html
http://www.stereophile.com/news/more http://www.wired.com/listening_post/2
_on_musicgiants/index.html
007/09/musicgiants-a-h/
http://www.avsforum.com/t/365725/ishttp://keynet.blogs.com/networks/2004 http-www-musicrebellion-com-legal-inhttp://web.archive.org/web/2006042507
/01/psychology_schm.html
usa
1714/http://www.musicrebellion.com/
http://web.archive.org/web/2007031205
http://www.latinrapper.com/news_octo http://www.cnet.com/profile/Musica360 4839/http://musica360.com/store/about
ber25d.html
.com/
us.php
http://www.internetretailer.com/2003/1
http://www.pcworld.com/article/128520
1/10/best-buy-music-now-launch-digital- http://news.cnet.com/aols-got/aol_scraps_music_now_in_favor_of_nap
music-store
musicnow/2100-1027_3-5930749.html
ster.html
http://news.cnet.com/8301-17938_105http://en.wikipedia.org/wiki/Napster_( http://techcrunch.com/2011/10/06/jon9945987-1.html
pay_service)
irwin-on-why-rhapsody-bought-napster/
http://press.nationalgeographic.com/200
6/08/03/worldmusicchanneldeliverssoun http://worldmusic.nationalgeographic.co
dtrackoftheworldwithfreedownloads/
m/
http://en.wikipedia.org/wiki/Optus_Mu
sic_Store
http://www.techhttp://lifehacker.com/219114/downloadrecipes.com/rx/1391/pandora_how_to_ri http://lifehacker.com/232533/download- of-the-day-pandoras-jarp_save_music_mp3/
of-the-day-pandora-downloader-windows windows?tag=softwarepandora
http://en.wikipedia.org/wiki/PassAlong
_Networks
http://en.wikipedia.org/wiki/PayPlay.F
M

HIGHLY CONFIDENTIAL

Source 4

http://reviews.cnet.com/musicservices/fullaudio-music-now/45059240_7-30974743-2.html

Exhibit 7
Sources:

Peer Impact
Philadelphia Orchestra
PlayNow
Puretracks

RCN Music
Rhapsody (Started as Listen.com)

Rhino Entertainment
Songtouch

Sony Connect
Streamwaves

Tower Records Digital
Traxsource
TVT Records

URGE
Virgin Digital

Vitaminic Music Club

Voy Music

Wal-Mart
Yahoo! Music Unlimited

Source 1
Source 2
http://www.technologyreview.com/new
s/407000/p2p-from-internet-scourge-to- http://en.wikipedia.org/wiki/Peer_Impa
savior/
ct
http://www.playbillarts.com/news/artic
http://www.philorch.org/recordings/
le/5262.html
http://en.wikipedia.org/wiki/PlayNow_ http://www.javamidlet.com/2007/11/09
Arena
/playnow-music-download-service.html
http://us.puretracks.com/content/viewer
http://en.wikipedia.org/wiki/Puretracks .aspx?cid=GlobalNav_Home
http://www.businesswire.com/news/ho
me/20061117005100/en/RCN-Reacheshttp://music.rcn.net/
Enhance-Music-Game-Content-Choices
http://en.wikipedia.org/wiki/Rhapsody
_(online_music_service)
http://www.allamericanpatriots.com/487
http://www.rhino.com/global/faq#restri 35056_music_led_zeppelin_rocks_over_ai
ctions_on_digital
r_verizon_wireless
http://christianmusic.about.com/od/top
http://www.songtouch.com/
10songtouchsongs/tp/tpSTsgl41805.htm
http://downloadsquad.switched.com/20
http://en.wikipedia.org/wiki/Sony_Con 07/08/30/sony-kills-off-connect-onlinenect
music-store-atrac-format/
http://en.wikipedia.org/wiki/Streamwa
ves
http://www.ipodobserver.com/ipo/artic
le/Tower_Records_Digital_Joins_Music_ http://www.dmwmedia.com/news/tag/
Download_Market/
tower_records
http://www.traxsource.com/index.php?a
ct=page&page_id=311
http://en.wikipedia.org/wiki/TVT_Reco
rds
http://en.wikipedia.org/wiki/URGE_(di http://www.pcmag.com/article2/0,2817,
gital_music_service)
2009421,00.asp
http://en.wikipedia.org/wiki/Virgin_Di
gital
http://www.prnewswire.com/newsreleases/epo-technology-announcesbundle-agreement-with-vitaminic-to-addinstant-music-access-to-the-vitaminichttp://www.vitaminic.co.uk/pressmusic-club-76927317.html
releases.html
http://www.dmwmedia.com/news/2006 http://www.hispanicbusiness.com/2005/
/06/08/starmedia-voy-music-launch-latin- 11/21/voyr_announces_launch_of_voy_
digital-music-service
musictm.htm
http://www.macobserver.com/tmo/artic
le/Inside_Walhttp://arstechnica.com/business/2011/0
Marts_Online_Music_Store_Digital_Right 8/walmart-pulling-the-plug-on-its-mp3s_Management/
store-but-not-its-drm-servers/
http://en.wikipedia.org/wiki/Yahoo!_M http://www.informationweek.com/news
usic_Unlimited
/191000022

HIGHLY CONFIDENTIAL

Source 3
http://www.fatwallet.com/forums/offtopic/943427/

http://news.cnet.com/8301-1023_310049186-93.html

http://www.gazette.com/articles/music15226-christian-songtouch.html

http://betanews.com/2007/07/24/drmfree-mp3s-coming-to-yahoourge/#comments

http://www.dmwmedia.com/news/tag/
voy
http://tech.fortune.cnn.com/2011/08/10
/itunes-triumphant-walmart-kills-itsmusic-download-store/

Source 4

Exhibit 7
Sources:
YouTube
Zune Marketplace

Source 1
Source 2
http://labnol.blogspot.com/2007/04/ho
w-to-rip-audio-from-youtube-videos.html
http://en.wikipedia.org/wiki/Zune
http://www.zune.net/en-US/

HIGHLY CONFIDENTIAL

Source 3

Source 4

Exhibit 8
Timeline of DRM-Free Downloads
2003 - 2013
By Mar. 31st, 2009, entire iTMS is available DRM-free.
Jan. 6th, 2009: Apple announces that effectively immediately
80% of the iTMS catalogue will be available DRM-free.

DRM-Free Development

By Jan. 10th, 2008, Amazon.com MP3 is fully operational with music
from all four major labels (EMI, Universal, Warner, and Sony).
Sept. 25th, 2007: Amazon.com launches public beta test of DRM-free music
store.

Apr. 2nd, 2007: EMI begins to offer DRM-free through iTunes.

Feb. 6th, 2007: Jobs issues "Thoughts on Music," an open letter about DRM-free.

By Sept. 2006, 27 sites offer DRM-free downloads, all but 4 are fully or partially paid.
iTMS launches on Apr. 28th, 2003;
eMusic has been offering DRM-free music for monthly subscription since 1998.

1/1/2003

1/1/2004

1/1/2005

1/1/2006

1/1/2007

1/1/2008

1/1/2009

1/1/2010

1/1/2011

Sources: Apple press releases, Amazon press releases, EMI press releases, sources reflected in Exhibit 7, and Wikipedia.

HIGHLY CONFIDENTIAL

1/1/2012

1/1/2013

HIGHLY CONFIDENTIAL

Exhibit 10
Attributes Considered and Attributes Ignored
by Professor Noll
Attributes Considered by Professor Noll

Attributes Ignored by Professor Noll

Class
U2
Capacity (MB)
Photo
Video and Photo
Size (cubic inches)
Cost per Unit ($/Unit)
Reprice Sale
End of Life Sale
Quantity Sold
Seasonality
Number of Songs Available to Download
Coupon (direct sales only)

USB*
Firewire*
Weight (oz)*
Display (inches)*
Resolution (pixels)*
Music Battery Hours*
Recharge Hours
OEM sale (reseller regression only)*

* Professor Noll's datasets contain this variable.
Sources: Declaration of Roger G. Noll on Liability and Damages and iPod Characterictics data in Murphy/Topel reports.

HIGHLY CONFIDENTIAL

Exhibit 11

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

Exhibit 12a

CONFIDENTIAL - ATTORNEYS EYES ONLY

Exhibit 12a
Exhibit 13.1 Reseller Sales Log Regression Results

CONFIDENTIAL - ATTORNEYS EYES ONLY

Exhibit 12b
Exhibit 13.2 Direct Sales Log Regression Results

CONFIDENTIAL - ATTORNEYS EYES ONLY

Exhibit 12b

CONFIDENTIAL - ATTORNEYS EYES ONLY

Exhibit 13a (Amended)
Summary of Resellers Data Regressions - Dependent Variable: Log Price (RPU)

HIGHLY CONFIDENTIAL

Exhibit 13b (Amended)
Summary of Direct Data Regressions - Dependent Variable: Log Price (RPU)

HIGHLY CONFIDENTIAL

Exhibit 13c (Amended)

Effects of Correcting Professor Noll's Calculation of Damages

HIGHLY CONFIDENTIAL

01/01/02

Noll's Indicator Variables

Exhibit 15a

Notes:
2) Dates are as used in Professor Noll's latest declaration.
Sources: Declaration of Roger G. Noll on Liabilities and Damages, Apple Sales Data.

HIGHLY CONFIDENTIAL

Exhibit 15b

Sources: Declaration of Roger G. Noll on Liabilities and Damages, Apple Sales Data.

HIGHLY CONFIDENTIAL

Exhibit 16

Content Interoperability of New Digital Media Players
FairPlay
(iTMS)
Before Harmony (7/24/2004)
iPods (all models)
Other portable media players

Contents' Type of DRM
Helix
(RMS)

Yes
No

No
Yes

DRM Free
Yes
Yes

Between Harmony and iTunes 4.7 [harmony_blocked] (7/25/2004 - 10/26/2004)1
iPods (all models)
Yes
Yes
Other portable media players
No
Yes

Yes
Yes

Between iTunes 4.7 [harmony_blocked] and Harmony2 (10/27/2004 - 4/25/2005)1
iPods (all models)
Yes
No
Other portable media players
No
Yes

Yes
Yes

Between Harmony2 and iTunes 7 (4/26/2005 - 9/11/2006)1
iPods (all models)
Yes
Other portable media players
No

Yes
Yes

Yes
Yes

Notes:
(1) Upgrading to iTunes 4.7 was optional until 3/21/2005 when it became mandatory for anyone who
wished to buy from the iTMS.

HIGHLY CONFIDENTIAL

Appendix A

Robert H. Topel
CURRICULUM VITAE
May, 2013
CURRENT POSITIONS
Isidore Brown and Gladys J. Brown Distinguished Service Professor of Economics,
Booth School of Business, University of Chicago
Director, George J. Stigler Center for the Study of the Economy and the State
Co-Director, Energy Policy Institute at Chicago (EPIC)
Research Associate, National Bureau of Economic Research
EDUCATION
B.A. (with High Honors), University of California, Santa Barbara, 1974
Ph.D., University of California, Los Angeles, 1980
FIELDS OF SPECIALIZATION
Microeconomics, Labor Economics, Industrial Organization, Health Economics,
Economic Policy, Energy Economics, National Security Economics
PREVIOUS ACADEMIC POSITIONS
Professor of Economics, Graduate School of Business, University of Chicago, 1986-93
Kirby Distinguished Visiting Professor of Economics, Texas A&M University, 2006
Professor of Economics, Department of Economics, University of California, Los
Angeles, 1986
Associate Professor of Economics, Department of Economics, University of California,
Los Angeles, 1985-86
Associate Professor of Economics, Graduate School of Business, University of Chicago,
1983-85
Assistant Professor of Economics, Department of Economics, University of Chicago,
1980-83
OTHER AFFILIATIONS
Research Associate, National Bureau of Economic Research, 1984—present
Senior Fellow, the Milken Institute, 1999—present

Faculty Member, MacLean Center for Clinical Medical Ethics, University of Chicago
Member, National Petroleum Council Taskforce on Transportation Fuels Supply and
Infrastructure, 2010-2012
Fellow, Center for the Study of Poverty and Inequality, Stanford University, 2006-present
Member, Brookings Panel on Economic Activity, various years.
Visiting Scholar, Board of Governors of the Federal Reserve, 1990
Research Associate, Economics Research Center, NORC, 1980—1990
Consulting Economist, The Rand Corporation, 1982—1989
Research Associate, Center for the Study of the Economy and the State, 1980—present
Faculty Research Fellow, National Bureau of Economic Research, 1981-83
Research Economist, Unicon Corporation, 1981-88
Consultant, U.S. Department of Labor, 1985-90
Partner, Chicago Partners LLC 1994-2008
Principal & Managing Director, Navigant Economics, 2008-2013
Board of Directors, Ingalls Hospitals and Ingalls Health Service, 2000-2012
Director, WGA Evans Scholars Foundation, 2011-present
Senior Consultant, Charles River Associates, 2013-present

EDITORIAL POSITIONS
Editor, Journal of Political Economy, 1993-2003
Board of Editors, American Economic Review, 1992-94
Associate Editor, Journal of Labor Economics, 1982-92
Editorial Board, International Journal of the Economics of Business, 1993-present
Member of the Advisory Board, ERN Labor Journals, 1998-present
HONORS & AWARDS
Kenneth J. Arrow Award, International Health Economics Association, 2007
Kirby Distinguished Visiting Professor, Texas A&M University, 2006
Research America Eugene Garfield Prize for Medical and Health Research, 2005
Elected Fellow, Society of Labor Economists, 2004
Elected Member, Conference on Research in Income and Wealth
Elected Founding Member, National Academy of Social Insurance
William Ladany Research Scholar, University of Chicago, 1989-91
William Fishman Research Scholar, University of Chicago, 1986-87
Smith Richardson Dissertation Fellowship in Political Economy, 1978-79
Foundation for Research in Economics and Education Fellowships, 1975-79
Chancellor’s Intern Fellow, University of California, Los Angeles, 1975-79
University Fellow, Northwestern University, 1975
General Electric Dissertation Fellowship, 1978
TEACHING EXPERIENCE
Graduate Economic Theory I, II, III
Law, Economics and Business

Competitive Strategy
Advanced Topics in Labor Economics
Advanced Topics in Microeconomics
Managing the Workplace
Industrial Organization/Antitrust
Price Theory

OTHER PROFESSIONAL ACTIVITIES
Thompson Lecture (Keynote Address), Midwest Economic Association, 2000
Nominating Committee, American Economic Association, 1996, 1997
Program Committee, American Economic Association, 2006-2007.
Organizer, Universities-NBER Research Conference: “Labor Markets in the 1990s,”
Cambridge, December 1989.
Program Chair, Labor Economics, Econometric Society Meetings, December 1989.
National Science Foundation Review Panel in Economics, 1998, 1999
Inaugural Keynote Lecture, Missouri Economics Conference, University of Missouri,
2001
Pihl Lecturer, Wayne State University, November, 2004
Keynote Address, Federal Reserve Bank of Cleveland Conference on Education and
Economic Development, November, 2004
Kirby Lecturer, Texas A&M University, 2006
Huggins Lecturer, Department of Surgery Huggins Conference, University of Chicago,
May, 2007
Keynote Address, Conference Board of Canada Conference on Medical Research,
Montreal, January 2009
Keynote Address, Council on Competitiveness Conference on Energy Policy, Argonne
National Laboratory, May 2009
Keynote Address, University of Chicago/RFF/University of Illinois Conference on
Energy Policy and the Economy, Washington, D.C., January 2010
Keynote Address, Humana Health Economics Forum, Santa Fe Institute, 2011
Keynote Address, Toyota Sustainability Conference, La Jolla, 2011
Keynote Address, Conference on Health Policy, Arizona State University, 2013
UNIVERSITY SERVICE
Director, Undergraduate Program in Economics, 1980-83
Chairman, Graduate School of Business Curriculum Review, 1990-91
Committee on Graduate Education, 1992-94
Committee on Undergraduate Education, 1993-94
Council of the University Senate, 1992-94, 1995-97, 1999-2002, 2004-2007, 2010Committee of the Council of the University Senate, 2000-2002, 2006-2007
Graduate School of Business Policy Committee, 1995-97, 1999-2001
Member, Presidential Search Committee, 1999-2000
Board of Directors, University of Chicago Laboratory Schools, 1986-92, 1998-2007
Chairman, Director Search Committee, U of C Laboratory Schools, 2002-2003
Area Coordinator, PhD Program in Economics, 2002-2008

Director, George J. Stigler Center, 2007-present
Director, University of Chicago Energy Initiative, 2008-2010
Co-Director, Energy Policy Institute at Chicago, 2010-present

BIBLIOGRAPHY
Books:
The Welfare State in Transition, with Richard Freeman and Birgitta Swedenborg.
Chicago: University of Chicago Press for NBER, 1997.
Labor Market Data and Measurement, with John Haltiwanger and Marilyn Manser.
Chicago: University of Chicago Press for NBER, 1998.
Välfärdsstat i omvandling: Amerikanskt Perspectiv pä den Svenska Modelten, with
Richard Freeman and Birgitta Swedenborg. Författarna och SNS Förlag, 1995.
Measuring the Gains from Medical Research: An Economic Approach, with Kevin M.
Murphy. Chicago: University of Chicago Press (2003).
Reforming the Welfare State: Recovery and Beyond in Sweden, with Richard Freeman
and Birgitta Swedenborg, Chicago, University of Chicago Press for NBER, 2009
Att Reformera Välfärdsstaten, with Richard Freeman and Birgitta Swedenborg, SNS
Förlag, Stockholm, 2006
Distributional Aspects of Energy and Climate Policy, ed. with Mark Cohen and Don
Fullerton, Special Issue of the BE Journal of Economic Analysis and Policy, Spring 2011.
Also, Edward Elgar Publishing, Surrey, UK, 2013.
Articles and Monographs:
“Layoffs, Inventories, and the Demand for Labor,” Ph.D. Dissertation, University of
California, Los Angeles, December 1980.
“Unemployment Insurance: Survey and Extensions,” Economica 47 (August 1980): 35179 (with F. Welch)
“Inventory Adjustments, Industry Behavior, and the Business Cycle,” presented at the
NBER Conference on Inventories and Business Cycles, March 1980 (with A. Stockman)
“Inventories, Layoffs, and the Short-Run Demand for Labor,” American Economic
Review (September 1982): 769-87.

“Experience Rating of Unemployment Insurance and the Incidence of Unemployment,”
Journal of Law and Economics (April 1984): 61-90.
“On Layoffs and Unemployment Insurance,” American Economic Review (September
1983): 541-59.
“Equilibrium Earnings, Turnover, and Unemployment: New Evidence,” Journal of Labor
Economics (October 1984): 500-22.
“Local Labor Markets,” Presented at Hoover Institution Conference on Labor Markets,
January 1983. Journal of Political Economy 94 (June 1986, part 2): 111-43.
“Estimation and Inference in ‘Two-Step’ Econometric Models,” (with K. M. Murphy)
Journal of Business and Economic Statistics 3 (October 1985): 370-80.
“Employment Risk, Sectoral Shifts, and Unemployment,” (with G. Neumann), in Studies
in Search, ed. N. M. Kiefer and G. R. Neumann. Oxford: Oxford University Press, 1989.
“Unemployment and Unemployment Insurance,” Research in Labor Economics 7 (1985):
91-135.
“Efficient Labor Contracts with Employment Risk,” (with F. Welch), Rand Journal of
Economics 17 (Winter 1986): 490-507.
“Financing Unemployment Insurance: History, Incentives, and Reform,” in
Unemployment Insurance: The Second Half Century, ed. W. Lee Hansen and J. Byers.
University of Wisconsin Press, 1990.
“Sectoral Uncertainty and Unemployment,” UCLA Department of Economics Working
Paper No. 384, September 1985 (with L. Weiss), in Employment, Unemployment, and
Labor Utilization, ed. R. A. Hart. Boston: Allen & Unwin, 1988.
“The Housing Market in the United States” (with S. Rosen), Journal of Political
Economy (August 1988): 718-40.
“What They Say or What They Do? The Use of Survey Data in Predicting Behavior”
(with K. M. Murphy), Graduate School of Business, University of Chicago, March 1985.
“Unemployment, Risk and Earnings: Theory and Evidence from a Model of Equalizing
Wage Differentials” (with K. M. Murphy), in Unemployment and the Structure of Labor
Markets, ed. J. Leonard and K. Lang. London: Basil Blackwell, 1986, pp. 103-140.
“Job Mobility, Search, and Earnings Growth: A Reinterpretation of Human Capital
Earnings Functions,” Research in Labor Economics 8 (1986): 199-233. Reprinted in
Research in Labor Economics 35th Anniversary Retrospective, 2012, 401-435.
“The Evolution of Unemployment in the United States: 1968-1985” (with K. M.
Murphy), The NBER Macroeconomics Annual, vol. 2, 1987, pp. 7-58.

“The Impact of Immigration on the Labor Market,” in Immigration, Trade, and the Labor
Market, ed. R. Freeman. University of Chicago Press for NBER, 1988.
“Efficiency Wages Reconsidered: Theory and Evidence” (with K. M. Murphy), in
Advances in the Theory and Measurement of Unemployment, pp. 204-40. Edited by
Yoram Weiss and Gideon Fishelson. London: Macmillan, 1990.
“Labor Market Adjustments to Increased Immigration,” (with R. LaLonde), in
Immigration, Trade, and the Labor Market, ed. R. Freeman. University of Chicago Press
for NBER, 1989.
“Job Mobility and the Careers of Young Men” (with M. P. Ward), Quarterly Journal of
Economics 107 (May 1992): 441-79.
“Employment Risk, Diversification, and Unemployment,” (with George Neumann)
Quarterly Journal of Economics (November 1991): 1341-1365.
“Specific Capital, Mobility, and Wages: Wages Rise with Job Seniority,” Journal of
Political Economy 99 (February 1991): 145-76. Reprinted in Outstanding Contributions
in Labor Economics, ed. Orley Ashenfelter, Worth Publishers, 1999: 162-192.
“Specific Capital and Unemployment: Measuring the Costs and Consequences of Worker
Displacement.” Carnegie-Rochester Series on Public Policy 33 (1990): 181-214.
“The Assimilation of Immigrants in the United States: Immigrant Quality and the
Changing Price of Skills,” In Immigration and the Work Force: Economic Consequences
for the United States and Source Areas, ed. G. Borjas and R. Freeman. Chicago:
University of Chicago Press for NBER, 1992.
“Trends in the American Labor Market,” GSB Chicago, vol. 12, no. 2, Winter 1990, pp.
11-16.
“Why Has the Natural Rate of Unemployment Increased over Time?” (with K. Murphy
and C. Juhn), Brookings Papers on Economic Activity 2 (1991), pp. 75-142.
“Immigrant Quality and Assimilation in the American Labor Market” (with R. LaLonde),
American Economic Review, 81 ( May 1991): 297-302.
“Unemployment and Insurance,” in Proceedings of National Academy of Social
Insurance, 1992.
“Labor Markets and Economic Growth: Lessons from Korea’s Industrialization, 19701990” (with D. Kim), Differences and Changes in Wage Structures, ed. Richard Freeman
and Larry Katz (Chicago: University of Chicago Press for NBER, 1995.
“Wage Inequality and Regional Labor Market Performance in the United States,” Labour
Market and Economic Performance: Europe, Japan, and the USA, ed. Toshiaki
Tachibanaki (New York: St. Martin’s Press, 1994).

“Discretion and Bias in Performance Evaluation” (with C. Prendergast), European
Economic Review 36 (June 1993): 355-65.
“What Have We Learned from Empirical Studies of Unemployment and Turnover?”
American Economic Review, 83 (May 1993): 110-115.
“Regional Labor Markets and the Determinants of Wage Inequality,” American
Economic Review, 84 (May 1994): 17-22.
“Ekonomiska problem i Sveriges välfärdsstat -- inledning, sammanfattning och
slutsatser,” with Richard B. Freeman and Birgitta Swedenborg. In Välfärdsstat i
omvandling: Amerikanskt Perspectiv pä den Svenska Modelten, with Richard Freeman
and Birgitta Swedenborg. Författarna och SNS Förlag, 1995.
“Lönepolitik och strukturomvandling,” with Per-Anders Edin. In Välfärdsstat i
omvandling: Amerikanskt Perspectiv pä den Svenska Modelten, with Richard Freeman
and Birgitta Swedenborg. Författarna och SNS Förlag, 1995.
“Favoritism in Organizations” (with C. Prendergast), Journal of Political Economy 104
(October 1996): 958-78.
“In Celebration of Armen Alchian’s 80th Birthday: Living and Breathing Economics”,
(with Armen A. Alchian, James M. Buchanan, Harold Demsetz, Axel Leijonhufvud, John
R. Lott, Jr., and William F. Sharpe), Economic Inquiry, Vol. XXXIV, July 1996, 412426.
“Another Look at Look Labor Market Adjustments to Increased Immigration” (with J.
Hojvat-Gallin and R. LaLonde), 1996.
“Economic Impact of International Migration and the Economic Performance of
Migrants,” (with R. LaLonde), in Handbook of Population and Family Economics, ed.
Mark R. Rosenzweig and Oded Stark (Amsterdam: North-Holland, 1997), pp. 799-850.
“Economic Troubles in Sweden's Welfare State,” in The Welfare State in Transition, ed.
Richard Freeman, Robert Topel, and Birgitta Swedenborg. Chicago: University of
Chicago Press for NBER, 1997.
“Wage Policy and Restructuring: The Swedish Labor Market Since 1960” (with PerAnders Edin), in The Welfare State in Transition, ed. Richard Freeman, Robert Topel,
and Birgitta Swedenborg. Chicago: University of Chicago Press for NBER, 1997.
“Unemployment and Nonemployment” (with Kevin M. Murphy), American Economic
Review 87 (May 1997): 295-300.
“Factor Proportions and Relative Wages: The Supply Side Determinants of Wage
Inequality,” Journal of Economic Perspectives 11 (Spring 1997): 55-74.

“Empirical Knowledge in Labor Economics,” in Labor Market Data and Measurement,
ed. John Haltiwanger, Marilyn Manser, and Robert Topel. Chicago: University of
Chicago Press for NBER, 1998.
“Labor Markets and Economic Growth,” in Handbook of Labor Economics, ed. Orley
Ashenfelter and David Card. Amsterdam: Elsevier Science B.V., 1999, pp. 2943-2984.
“Medical Research: What's It Worth?” (with K. M. Murphy), Milken Institute Review: A
Journal of Economic Policy (First Quarter 2000): 23-30.
“Entry, Product Design, and Pricing in an Initially Monopolized Market” (with S. Davis
and K. M. Murphy), University of Chicago GSB, September 2001, revised January, 2003.
Journal of Political Economy, vol. 112 no. 1, pt. 2, February, 2004, pp188-225.
“Adverse Price Effects of Entry in Markets with Few Firms” (with Steven J. Davis and
Kevin M. Murphy), Working Paper, University of Chicago, April 2001.
“The Economic Value of Medical Research” (with Kevin M. Murphy), in Measuring the
Gains from Medical Research: An Economic Approach, edited by Kevin M. Murphy and
Robert H. Topel. Chicago: University of Chicago Press, 2003, pp 41-73.
“Current Unemployment, Historically Contemplated” (with Kevin M. Murphy and
Chinhui Juhn), Brookings Papers on Economic Activity 1. Washington, D.C.: The
Brookings Institution, 2002.
“Estimation and Inference in Two-Step Econometric Models” (with K.M. Murphy),
Journal of Business and Economic Statistics, 20, issue 1, 2002: 88-97 (reprint: 20th
Anniversary issue of the most important contributions published in JBES).
“Labor Markets in the United States and Korea: Factor Proportions, Inequality, and
Unemployment” in Macroeconomic Implications of Post-Crisis Structural Change, L.J.
Cho, D. Cho and Y.H. Kim, KDI press, 2005, pp 131-60.
“Diminishing Returns? Evidence on the Costs and Benefits of Improving Health” (with
K.M. Murphy) November 2002, Perspectives in Biology and Medicine, volume 46, no. 3,
(Summer, 2003): pp108-128.
“War vs. Containment” (with S.J. Davis and K.M. Murphy), March 2003; presented at
NBER Conference on National Security Economics, November 2003.
“Black-White Differences in the Economic Value of Improving Health” (with K. M.
Murphy). Presented at the National Institutes of Health Conference on Racial Disparities
in Health Outcomes, December 2001. Perspectives in Biology and Medicine, Summer,
2004.
“The Value of Health and Longevity” (with K.M. Murphy), Revised March, 2005, NBER
Working Paper #11405, June 2005. Journal of Political Economy, October 2006, pp
871-904. Winner of the 2005 Eugene Garfield Economic Impact of Medical Research

Award, given by Research America. Winner of the 2007 Kenneth J. Arrow Award, given
by the International Health Economics Association for the best paper in health economics
published in 2006.
“The Private and Social Benefits of Education”, (with Fabian Lange), Handbook of the
Economics of Education, North-Holland, 2006
“The Social Value of Education”, Keynote Address, Federal Reserve Bank of Cleveland
Conference on Education and Economic Development, November 2004, in Education and
Economic Development, Federal Reserve Bank of Cleveland Economic Review, 2004, pp
47-58.
“On Human Capital and Economic Growth” (with Fabian Lange), Working Paper,
University of Chicago, September, 2005.
“Äterhämtning och terstaende problem I den svenska valfardsstaten—inlendning,
sammanfattning och slutsatser” in Att Reformera Välfärdsstaten, with Richard Freeman
and Birgitta Swedenborg, SNS Förlag, Stockholm, 2006, 9-34.
”Forandrade forutsattningar for svensk lonebildning” (with Peter Fredriksson), in Att
Reformera Välfärdsstaten, with Richard Freeman and Birgitta Swedenborg, SNS Förlag,
Stockholm, 2006, 65-82.
”War in Iraq versus Containment” (with Steven J. Davis and Kevin M. Murphy), for
CESifo Conference “Guns and Butter: The Economic Causes and Consequences of
Conflict”, Munich, December 2005, February, 2006.
“Social Value and the Speed of Innovation” (with Kevin M. Murphy), American
Economic Review, May, 2007.
“Unemployment”, The New Palgrave of Economics, 2008
“Critical Loss Analysis in the Whole Foods Case”, Global Competition Policy, March,
2008, @ http://www.globalcompetitionpolicy.org/index.php?&id=949&action=907.
“Wage Determination and Employment in Sweden Since 1990”, (with Peter
Fredricksson), Working Paper, University of Chicago and Uppsala University, August,
2006, revised December, 2008, in Recovery and Beyond: Reforming the Welfare State in
Sweden, Richard Freeman, Birgitta Swedenborg and Robert Topel, eds., University of
Chicago Press for NBER, 2009
“Reforming the Welfare State: Recovery and Beyond in Sweden” (with Richard Freeman
and Birgitta Swedenborg), in Recovery and Beyond: Reforming the Welfare State in
Sweden, Richard Freeman, Birgitta Swedenborg and Robert Topel, eds., University of
Chicago Press for NBER, 2009
“On the Economics of Climate Policy”, (with Gary S. Becker and Kevin M. Murphy), BE
Journal of Economic Analysis and Policy, April, 2011

“Introduction to ‘Job Mobility Search and Earnings Growth, A Reinterpretation of
Human Capital Earnings Functions”, Research in Labor Economics 35th Anniversary
Retrospective, Emerald Group Publishing 2012, 397-400.
“Some Basic Economics of National Security” (with Kevin M. Murphy), American
Economic Review, May 2013
“Competitive Discounts and Antitrust Policy” (with Kevin. M. Murphy and Edward A.
Snyder), Oxford Handbook of Antitrust Economics, 2013 (forthcoming)

Selected Congressional Testimony and Presentations:
“Unemployment and Insurance,” Testimony before the U.S. Senate Committee on
Finance, April 23, 1991.
“The Economic Value of Medical Research,” Testimony before the U.S. Senate
Committee on Health, Education, Labor, and Pensions, May 10, 2001.
“The Value of Improvements in Health and Longevity”, Presentation for Congressional
Staff and the American Cancer Society, Washington, July, 2005.
Selected Reports:
“Unemployment Insurance Financing and Unemployment: Empirical Investigation of
Adverse Incentives,” Final Report, U.S. Department of Labor Contract No. B9M22046,
November 1982.
“Unemployment and Unemployment Insurance,” Final Report, U.S. Department of
Labor, ETA, May 1984.
“Local Labor Markets,” Final Report, U.S. Department of Labor, Office of the Assistant
Secretary for Policy, March 1984.
“The Use of Survey Data in Predicting Behavior: The Case of Enlistment Intentions,”
Final Report, U.S. Department of Defense, May 1985.
“Equalizing Wage Differences,” Final Report, U.S. Department of Labor, Office of the
Assistant Secretary for Policy, August 1985.
“Sectoral Change and Worker Displacement,” Final Report, U.S. Department of Labor,
Office of the Assistant Secretary for Policy, March 1990.
Book Reviews:
Employment Hazards by W. Kip Viscusi. In Journal of Economic Literature, March
1982.

Handbook of Labor Economics, ed. O. Ashenfelter and R. Layard. In Journal of
Economic Literature, 1988.

Selected Comments:
“Comment on ‘Some Recent Developments in Labor Economics and Their Implications
for Macroeconomics’,” Journal of Money, Credit and Banking 20 (August 1988, part 2).
“Comment on ‘Industry Rents: Evidence and Implications’,” (by Lawrence Summers and
Lawrence Katz) Brookings Papers on Economic Activity, Brookings Institution,
Washington, D.C., 1989.
“Comment on ‘Wage Dispersion between and within U.S. Manufacturing Plants’,
Brookings Papers on Economic Activity, 1991.
“Comment on ‘Why Is the U.S. Unemployment Rate So Much Lower?’” NBER
Macroeconomics Annual, 1998, pp. 67-72.
“Comment on ‘Does Immigration Grease the Wheels of the Labor Market?’” by George
J. Borjas. Brookings Papers on Economic Activity. edited by William C. Brainard and
George L. Perry. Washington, D.C. Brookings Institution, 2001.
“Comment on ‘Where did the Productivity Growth Go? Inflation Dynamics and the
Distribution of Income’” by Ian Dew-Becker and Robert J. Gordon. Brookings Papers on
Economic Activity. edited by William C. Brainard and George L. Perry. Washington,
D.C. Brookings Institution, 2005, 135-44.

Recent Cases, Reports and Testimony
Roderick Arnold et al. v. Cargill, Inc. in the U.S. District Court for the District
of Minnesota Third Division. No. 01-CV-2086 DWF/AJB. Expert on behalf of
Cargill in racial discrimination class action. Expert Report, June 29, 2004.
Applied Medical v. Ethicon, Inc., et. al. in the Central District of California.
No. SACV 03-1329. Expert on behalf of Johnson & Johnson (Ethicon) in antitrust
litigation. Expert Report, May 24, 2005. Deposition, November 7, 2005. Trial
Testimony, August 18, 2006.
Conmed Corp. v. Johnson & Johnson, United States District Court, Southern
District of New York 03-CV8800. Expert on behalf of Johnson and Johnson in
antitrust litigation. Expert Report, August 1, 2005.
Daryal T. Nelson, et. al. v. Wal-Mart Stores Inc. and Wal-Mart Transportation
LLC., Eastern District of Arkansas Helena Division, N). 2:05 CV-00134-WRW.
Expert on behalf of Wal-Mart in racial discrimination class action. Expert Report,
April 12, 2006. Deposition, May 2, 2006. Expert Report, July 28, 2008. Rebuttal
Expert Report, September 12, 2008. Deposition, September 23 2008. Declaration,
October 2, 2008.
Sulfuric Acid Antitrust Litigation, In the US District Court for the Northern
District of Illinois Eastern Division, No. 03-C-4576. Expert on behalf of
Marsulex and Chemtrade in antitrust class action. Expert Report, February 3,
2006. Deposition March 2006.
Amy Velez, et. al. v. Novartis Corporation, Novartis Pharmaceuticals
Corporation, and Thomas Ebeling, US District Court, Southern District of New
York, 04 Civ. 09194. Economic (compensation practices) expert on behalf of
Novartis Pharmaceuticals in class action gender discrimination case. Expert
Report, November 9, 2006.
Peter P. Jonites, et. al. v. Exelon Corporation, et. al., United States District
Court, Northern District of Illinois, No. 05CV04234. Expert on behalf of Exelon
in class action labor case. Expert Report, April 16, 2007, Amended Expert Report,
June 14, 2007.

Clifford L. Whitaker, et. al. v. 3M Company, United States District Court for
the Second Judicial District of Minnesota Civ.C4-04-012239. Expert on behalf
of 3M in class action discrimination case. Expert Report, July 25, 2007.
Deposition, August 21, 2007.

Pinkowski et al,. v. 3M Company, United States District Court, District of New
Jersey, No. 205CV05668. Expert on behalf of 3M in class action age discrimination
case. Expert Report January 29, 2007. Deposition, March 2007. Trial Testimony, May 8,
2010.
Sun Microsystems, Inc., et al. v. Hynix Semiconductor, Inc., et al. (Consolidated),
Unisys Corporation v. Hynix Semiconductor, Inc., et al., Jaco Electronics, Inc. v.
Hynix Semiconductor, Inc., et al., Edge Electronics, Inc. v. Hynix Semiconductor,
Inc., et al., All American Semiconductor, Inc., et al., v. Hynix Semiconductor, Inc. et
al., DRAM Claims Liquidation Trust, by its Trustee Wells Fargo Bank, NA v. Hynix
Semiconductor, et al., United States District Court for the Northern District of
California San Francisco Division. Expert on behalf of Hynix Semiconductor, Inc. and
Hynix Semiconductor America, Inc. in antitrust class action. Expert Report, March 7,
2008. Deposition, April 25, 2008. Declaration, April 16, 2009.
Katherine Puffer, on behalf of herself and all others similarly situated, v. Allstate
Insurance Company in the U.S. District Court for the Northern District of Illinois
Eastern Division. Expert on behalf of Allstate Insurance Company in class action gender
discrimination case. Expert Report, April 1, 2008. Declaration, October 20, 2008.
William Syverson, Patrick Boone, Lee Deshler, Robert Flowers, Barry Gerard, Tina
Gleisner, Thomas Gomez, Edwin “Dana” Goodloe, Rolf Marsh, Daniel Moczan,
James Payne, and Antonio Rivera, individually and on behalf of others similarly
situated, v. International Business Machines Corporation in the U.S. District Court
for the Northern District of California San Jose Division. Expert on behalf of IBM in
class action age discrimination case. Expert Report, August 19, 2008. Deposition,
September 9, 2008. Declaration in support of IBM, December 19, 2008.
The United States of America v. Canada, in support of the United States’ request for
arbitration regarding Canadian softwood lumber subsidies. Expert on behalf of the
United States and US Dept. of Justice in softwood lumber international trade arbitration.
Expert Report, November 21, 2008. Expert Response Report, March 23, 2009. Expert
Rejoinder Report, June 19, 2009. Trial Testimony, July 20, 2009. Expert Report for the
Tribunal, June 15, 2010.
Valassis Communications, Inc. v. News America Incorporated, a/k/a News America
Marketing Group, News America FSI, Inc. a/k/a News America Marketing FSI,
LLC and News America Marketing In-Store Services, Inc. a/a/a News American
Marketing In-Store Services, LLC in the United States District Court for the
Eastern District of Michigan Southern Division, No. 06-10240. Expert on behalf of
New America in antitrust litigation. Expert Report, November 21, 2008. Rebuttal Expert
Report, December 26, 2008. Deposition, January 22, 2009. Supplemental Expert Report,

December 21, 2009. Supplemental Rebuttal Expert Report, January 14, 2010. Deposition,
January 25, 2010.

Valassis Communications, Inc., v. News America Incorporated, No. 07-706645-CZ.
Expert on behalf of News America in antitrust litigation. Expert Report, November 21,
2008. Rebuttal Expert Report, December 26, 2008. Deposition, January 22, 2009. Trial
Testimony, July 15-16, 2009.
EBay Seller Antitrust Litigation in the U.S. District Court for the Northern District
of California San Jose Division. Expert on behalf of Ebay Inc.’s opposition to plaintiffs’
motion for class certification in antitrust case. Declaration, August 14, 2009. Deposition,
August 28, 2009 and September 18, 2009.
Payment Card Interchange Fee and Merchant Discount Antitrust Litigation, in the
U.S. District Court for the Eastern District of New York. Expert on behalf of card
issuing banks, Visa and Mastercard in class action antitrust litigation. Expert Report,
December 14, 2009. Deposition, April 20, 2010.
Applied Medical Resources Corp., and Applied Medical Distribution Corp. v.
Ethicon Endo-Surgery, Inc. in the United States District Court for the Central
Division of California Southern Division, No. CV09-3605-RSWL (VBKx). Expert
Report, March 19, 2010. Deposition, April 15, 2010.
Wallace Bolden, et al. vs. Walsh Group, et.al in the U.S. District Court Northern
District of Illinois, Eastern Division, No. 06cv104. Expert on behalf of Walsh Group in
class action race discrimination case. Expert Report, December 22, 2010. Deposition,
February 18, 2011.
Jones, et al. vs. Wells Fargo Bank N.A., Wells Fargo Home Mortgage, Inc. in the
Superior Court of California, County of Los Angeles, BC337821. Expert on behalf of
Wells Fargo Bank in class action lending discrimination case. Expert Report &
Deposition. Trial Testimony, December 21, 2010.
Peterson v. Seagate U.S. LLC, 534 F.Supp.2d 996, 2008 WL398968. Expert on behalf
of Seagate Corporation in class action age discrimination case. Expert Report, October
25, 2010. Deposition, February 15, 2011.
NRLC and Railroad Employees, National Mediation Board Case Nos. A-13569; A13570; A-13572; A-13573; A-13574; A-13575; A-13592. Expert on behalf of major US
railroads in labor mediation. Expert Report, October 10, 2011. Testimony December,
2011.
Toyota Motor Corporation Hybrid Brake Marketing Sales Practices and Products
Liability Litigation in the United States District Court Central District of California
Southern Division, No. 8:10-ML-02172-CJC-RNB. Expert on behalf of Toyota in class

action product liability (hybrid brakes) case. Expert Report, May 28, 2012. Deposition,
July 19, 2012.
Text Messaging Antitrust Litigation in the United States District Court Northern
District of Illinois Eastern Division, No. 08 C 7082. Expert on behalf of wireless
providers in antitrust litigation. Expert Report, July 3, 2013.

Appendix B: List of Materials Relied Upon
Court Documents
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.

DECLARATION OF ROGER G. NOLL ON LIABILITY AND DAMAGES – April 3, 2013 (and
sources and documents considered, cited or relied upon therein)
CORRECTIONS TO DECLARATION OF ROGER G. NOLL ON LIABILITY AND DAMAGES –
May 31, 2013 (and sources and documents considered, cited or relied upon therein)
EXPERT REPORT OF DAVID M. MARTIN – April 8, 2013 (and sources and documents
considered, cited or relied upon therein)
DEPOSITION OF JEFFREY L. ROBBIN – December 3, 2010 (and sources and documents
considered, cited or relied upon therein)
DEPOSITION OF ROGER G. NOLL – September 19, 2008 (and sources and documents
considered, cited or relied upon therein)
DEPOSITION OF ROGER G. NOLL - April 7, 2011 (and sources and documents considered, cited
or relied upon therein)
DEPOSITION OF ROGER G. NOLL - May 16, 2013 (and sources and documents considered, cited
or relied upon therein)
DECLARATION OF ROGER G. NOLL – January 18, 2011 (and sources and documents
considered, cited or relied upon therein)
SUPPLEMENTAL DECLARATION OF ROGER G. NOLL - July 18, 2011 (and sources and
documents considered, cited or relied upon therein)
SECOND SUPPLEMENTAL DECLARATION OF ROGER G. NOLL – September 23, 2011 (and
sources and documents considered, cited or relied upon therein)
SECOND SUPPLEMENTAL REPORT OF DR. MICHELLE M. BURTIS – November 14, 2011 (and
sources and documents considered, cited or relied upon therein)
DECLARATION OF DR. JOHN P. J. KELLY - April 18, 2011 (and sources and documents
considered, cited or relied upon therein)
See Expert Report of Dr. John P. J. Kelly - July 19, 2013 (and sources and documents considered,
cited or relied upon therein)
DEPOSITION OF AUGUSTIN J. FARRUGIA - December 8, 2010 (and sources and documents
considered, cited or relied upon therein)
DECLARATION OF AUGUSTIN FARRUGIA – January 18, 2011 (and sources and documents
considered, cited or relied upon therein)
SUPPLEMENTAL DECLARATION OF AUGUSTIN FARRUGIA - July 2, 2013 (and sources and
documents considered, cited or relied upon therein)
DEPOSITION OF JOHN P.J. KELLY, Ph.D. - February 22, 2011 (and sources and documents
considered, cited or relied upon therein)
DECLARATION OF DAVID F. MARTIN – February 28, 2011 (and sources and documents
considered, cited or relied upon therein)
DEPOSITION OF DAVID M. MARTIN, JR., PH.D. - May 9, 2013 (and sources and documents
considered, cited or relied upon therein)

20. DEFENDANT APPLE INC.'S RESPONSES TO PLAINTIFF'S FIRST SET OF INTERROGATORIES
August 29, 2008 (and sources and documents considered, cited or relied upon therein)
21. DECLARATION OF BONNY E. SWEENEY – January 18, 2011 (and sources and documents
considered, cited or relied upon therein)
22. DECLARATION OF DAVID KIERNAN - January 18, 2011 (and sources and documents
considered, cited or relied upon therein)
23. DECLARATION OF BONNY E. SWEENEY – February 28, 2011 (and sources and documents
considered, cited or relied upon therein)
24. DECLARATION OF DAVID KIERNAN - February 28, 2011 (and sources and documents
considered, cited or relied upon therein)
25. DECLARATION OF MICHAEL SCOTT – June 6, 2006 (and sources and documents considered,
cited or relied upon therein)
26. Deposition of Michelle Burtis – September 30, 2009 (and sources and documents considered, cited
or relied upon therein)
27. Deposition of David K. Heller – December 15, 2010 (and sources and documents considered, cited
or relied upon therein)
28. Deposition of Arthur Rangel – December 17, 2010 (and sources and documents considered, cited
or relied upon therein)
29. Deposition of Eddy Cue – December 17, 2010 (and sources and documents considered, cited or
relied upon therein)
30. APPLE’S MOTION FOR SUMMARY JUDGMENT - April 18, 2011 (and sources and documents
considered, cited or relied upon therein)
31. AMENDED CONSOLIDATED COMPLAINT FOR VIOLATIONS OF SHERMAN ANTITRUST
ACT, CLAYTON ACT, CARTWRIGHT ACT, CALIFORNIA UNFAIR COMPETITION LAW,
CONSUMERS LEGAL REMEDIES ACT, AND CALIFORNIA COMMON LAW OF
MONOPOLIZATION – January 25, 2010 (and sources and documents considered, cited or relied
upon therein)
32. ORDER GRANTING IN PART AND DENYING IN PART DEFENDANT’S MOTION FOR
SUMMARY JUDGMENT; DENYING AS PREMATURE PLAINTIFFS’ MOTION FOR CLASS
CERTIFICATION – May 19, 2011 (and sources and documents considered, cited or relied upon
therein)
33. UNREDACTED MEMORANDUM IN OPPOSITION TO MOTION FOR CLASS
CERTIFICATION (December 15, 2008)
34. Plaintiffs’ Rule 30(b)(6) Notice of Videotaped Deposition of Corporate Designees (2010-11-17)
35. SFI_658005_1_AIIA_ Responses to Data Questions 12.23.10
36. Plaintiffs’ Notice of Motion and Renewed Motion for Class Certification and Appointment of
Lead Class Counsel Under Seal (April 18, 2011)
37. SFI_658005_3_Responses to Data Questions 12.23.10
38. SFI_658269_4_Responses to Reseller Transaction Data Questions 2011-01-28
39. SFI_658573_3_Updated answers to 1.6.11 Follow Up Data Questions
40. SFI_658573_4_Responses to 1.7.2011 Data Questions
41. SVI_88493_7_Responses to iPod BOM and Gross Margin Reports 2011-02-04
42. PLAINTIFFS’ MEMORANDUM IN OPPOSITION TO APPLE’S MOTION FOR SUMMARY
JUDGMENT - April 18, 2011

43. DECLARATION OF BONNY E. SWEENEY IN SUPPORT OF PLAINTIFFS' MEMORANDUM IN
OPPOSITION TO APPLE'S MOTION FOR SUMMARY JUDGMENT - April 18, 2011
44. EXPERT REPORT OF DR. MICHELLE M. BURTIS - April 18, 2011 (and sources and documents
considered, cited or relied upon therein)
45. APPLE’S OPPOSITION TO RENEWED MOTION FOR CLASS CERTIFICATION - April 18, 2011
46. APPLE’S REPLY IN SUPPORT OF ITS MOTION FOR SUMMARY JUDGMENT - April 18, 2011
47. DECLARATION OF MICHAEL T. SCOTT IN SUPPORT OF APPLE'S REPLY IN SUPPORT OF
ITS MOTION FOR SUMMARY JUDGMENT - April 18, 2011
48. DECLARATION OF CARMEN A. MEDICI IN SUPPORT OF REPLY MEMORANDUM IN
SUPPORT OF PLAINTIFFS’ RENEWED MOTION FOR CLASS CERTIFICATION - April 18, 2011
49. REPLY MEMORANDUM IN SUPPORT OF PLAINTIFFS’ RENEWED MOTION FOR CLASS
CERTIFICATION - April 18, 2011
50. Reply Declaration of Roger G. Noll 2011-03-28
51. SFI_662507_4_AIIA_ APPLE’S OPPOSITION TO RENEWED MOTION FOR CLASS
CERTIFICATION 2011-04-11
52. ORDER GRANTING IN PART AND DENYING IN PART DEFENDANT’S MOTION FOR
SUMMARY JUDGMENT; DENYING AS PREMATURE PLAINTIFFS’ MOTION FOR CLASS
CERTIFICATION 2011-05-19
53. SUPPLEMENTAL REPORT OF DR. MICHELLE M. BURTIS 2011-07-26
54. Exhibit 1 to DECLARATION OF DAVID C. KIERNAN IN SUPPORT OF APPLE’S RESPONSE
TO PROFESSOR NOLL’S JULY 18 DECLARATION 2011-01-26
55. Exhibit 2 to DECLARATION OF DAVID C. KIERNAN IN SUPPORT OF APPLE’S RESPONSE
TO PROFESSOR NOLL’S JULY 18 DECLARATION 2011-01-26
56. DECLARATION OF DAVID C. KIERNAN IN SUPPORT OF APPLE’S RESPONSE TO
PROFESSOR NOLL’S JULY 18 DECLARATION
57. APPLE’S RESPONSE TO PROFESSOR NOLL’S JULY 18 DECLARATION
58. ORDER GRANTING PLAINTIFFS’ MOTION FOR CLASS CERTIFICATION 2011-11-22
59. ORDER REGARDING SCHEDULE 2013-01-28
60. Class Certification Report Backups for Michelle Burtis
61. DECLARATION OF JEFFREY ROBBIN IN SUPPORT OF DEFENDANT’S RENEWED MOTION
FOR SUMMARY JUDGEMENT - January 18, 2011
Bates Numbered Documents
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.

Internal Apple iTMS Surveys
Internal Apple iPod Surveys
Apple_AIIA00178491
Apple_AIIA00178737
iTunes Surveys Mac customers (Aug ‘03) Windows customers (Dec ‘03) Windows Prospects (Dec
‘03) December 2003
Apple_AIIA00323221
Apple_AIIA00477469
Final Report 2005 CYQ3 iTMS Tracker Program October 2005
Apple_AIIA00355775
Apple_AIIA00526882
Apple_AIIA00178867

73. iTunes Store Tracker Program Final Report—2006 CYQ4 January 2007
74. Apple_AIIA00193099
75. iTunes Store Tracker Survey CYQ4 2008 – Apple Market Research & Analysis March 2009
76. iTunes Store Tracker Survey CYQ1 2010 – June 2010
77. Apple_AIIA00331445
78. Apple_AIIA00389307 - Wave 4
79. Apple_AIIA00331431 - Forrester – Apple iTunes Jump-Starts Windows Digital Music
80. Apple_AIIA00455825 - iPod and iPod mini Buyer Survey January 2005
81. Apple_AIIA00493743 - iPod Buyer Survey — Wave 1 July 2005
82. iPod Owner Survey - Wave 2 January 2006
83. iPod Buyer Survey FY09 – Q2
84. iPod Owner Survey - Wave 3 June 2006
85. W5 iPod/iPhone Buyer Survey - Nov 2007 Apple Confidential Internal Use Only
86. Apple_AIIA00002914
87. Apple_AIIA00115845
88. Apple_AIIA00135089
89. Apple_AIIA00178888
90. Apple_AIIA00180903
91. Apple_AIIA00195126
92. Apple_AIIA00199264
93. Apple_AIIA00199280
94. Apple_AIIA00319774
95. Apple_AIIA00320223
96. Apple_AIIA00331431
97. Apple_AIIA00331434
98. Apple_AIIA00333439
99. Apple_AIIA03335100
100. Apple_AIIA00336153
101. Apple_AIIA00351078
102. Apple_AIIA00446624
103. Apple_AIIA00445669
104. Apple_AIIA004489705
105. Apple_AIIA00458544
106. Apple_AIIA00481550
107. Apple_AIIA00481589
108. Apple_AIIA00511000
109. Apple_AIIA00549197
110. Apple_AIIA00572228
111. Apple_AIIA00572232
112. Apple_AIIA00572600
113. Apple_AIIA00572676
114. Apple_AIIA00572699
115. Apple_AIIA00573587
116. Apple_AIIA00575626
117. Apple_AIIA00576776

118. Apple_AIIA00577214
119. Apple_AIIA00093860
120. Apple_AIIA00945275
121. Apple_AIIA00945600
122. Apple_AIIA00959717
123. Apple_AIIA00959724
124. Apple_AIIA00948352
125. Apple_AIIA00948381
126. Apple_AIIA00948413
127. Apple_AIIA00948739
128. Apple_AIIA00948960
129. Apple_AIIA00948962
130. Apple_AIIA00950304
131. Apple_AIIA00951449
132. Apple_AIIA00953603
133. Apple_AIIA00954126
134. Apple_AIIA00954141
135. Apple_AIIA00954156
136. Apple_AIIA00956496
137. Apple_AIIA00956962
138. Apple_AIIA00959170
139. Apple_AIIA00959446
140. Apple_AIIA00959620
141. Apple_AIIA00959724
142. Apple_AIIA00959757
143. Apple_AIIA00959872
144. Apple_AIIA00959648
145. Apple_AIIA00959694
146. Apple_AIIA00959713
147. Apple_AIIA00959717
148. Apple_AIIA00959740
149. Apple_AIIA00959787
150. Apple_AIIA00959792
151. Apple_AIIA00959851
152. Apple_AIIA00962282
153. Apple_AIIA00964813
154. Apple_AIIA00964990
155. Apple_AIIA00969312
156. Apple_AIIA00970146
157. Apple_AIIA00970153
158. Apple_AIIA00970162
159. Apple_AIIA00970250
160. Apple_AIIA00971555
161. Apple_AIIA00974904
162. Apple_AIIA00974636

163. Apple_AIIA00974470
164. Apple_AIIA00974713
165. Apple_AIIA00974667
166. Apple_AIIA00974749
167. Apple_AIIA00974463
168. Apple_AIIA00974546
169. Apple_AIIA00974840
170. Apple_AIIA00974588
171. Apple_AIIA00974603
172. Apple_AIIA00974838
173. Apple_AIIA00974708
174. Apple_AIIA00974519
175. Apple_AIIA00974641
176. Apple_AIIA00974932
177. Apple_AIIA00093477
178. Apple_AIIA00974824
179. Apple_AIIA00974776
180. Apple_AIIA00974577
181. Apple_AIIA00974782
182. Apple_AIIA00974735
183. Apple_AIIA00974843
184. Apple_AIIA00974620
185. Apple_AIIA00974799
186. Apple_AIIA00974906
187. Apple_AIIA00974894
188. Apple_AIIA00974982
189. Apple_AIIA00974561
190. Apple_AIIA00975107
191. Apple_AIIA00975139
192. Apple_AIIA00975167
193. Apple_AIIA00975191
194. Apple_AIIA00976658
195. Apple_AIIA00291156
196. Apple_AIIA01383364
197. Apple_AIIA01358752
198. Apple_AIIA01358805
199. Apple_AIIA00549184
200. Apple_AIIA00963459
201. Apple_AIIA00355606
202. Apple_AIIA00963492
203. Apple_AIIA00963340
204. Apple_AIIA00963340
205. Apple_AIIA00093499
206. Apple_AIIA00093499
207. Apple_AIIA00966312

208. Apple_AIIA00968156
209. Apple_AIIA00943649
210. Apple_AIIA00969283
211. Apple_AIIA01358744
212. Apple_AIIA00549183
213. Apple_AIIA00549198
214. Apple_AIIA0059199
215. Apple_AIIA00549200
216. Apple_AIIA01062282
217. Apple_AIIA01238828
218. Apple_AIIA01238820
219. Apple_AIIA00957597
220. Apple_AIIA01238839
221. Apple_AIIA00951291
222. Apple_AIIA00995575
223. Apple_AIIA01287859
224. Apple_AIIA00807000
225. Apple_AIIA01121058
226. Apple_AIIA00970285
227. Apple_AIIA00531624
228. Apple_AIIA01058774
229. Apple_AIIA01090043
230. Apple_AIIA01061533
231. Apple_AIIA01122425
232. Apple_AIIA00825977
233. Apple_AIIA00470249
234. Apple_AIIA00350891
235. Apple_AIIA00511587
236. Apple_AIIA00511463
237. Apple_AIIA00511581
238. Apple_AIIA00511483
239. Apple_AIIA00511572
240. Apple_AIIA00511549
241. Apple_AIIA00511499
242. Apple_AIIA00511544
243. Apple_AIIA00511540
244. Apple_AIIA00362516
245. Apple_AIIA00511501
246. Apple_AIIA00511476
247. Apple_AIIA00511467
248. Apple_AIIA00361301
249. Apple_AIIA00586924
250. Apple_AIIA00586940
251. Apple_AIIA00589630
252. Apple_AIIA00586761

253. Apple_AIIA00568674
254. Apple_AIIA00586947
255. Apple_AIIA005587043
256. Apple_AIIA00587053
257. Apple_AIIA00586847
258. Apple_AIIA00586841
259. Apple_AIIA00586961
260. Apple_AIIA00587018
261. Apple_AIIA00586956
262. Apple_AIIA00586928
263. Apple_AIIA00826100
264. Apple_AIIA00826134
265. Apple_AIIA00826069
266. Apple_AIIA00826063
267. Apple_AIIA00826053
268. Apple_AIIA00826010
269. Apple_AIIA00708248
270. Apple_AIIA00708205
271. Apple_AIIA00708163
272. Apple_AIIA00724232
273. Apple_AIIA00193955
274. Apple_AIIA00947706
275. Apple_AIIA00947643
276. Apple_AIIA01383219
277. Apple_AIIA01364726
278. Apple_AIIA01218566
279. Apple_AIIA01218377
280. Apple_AIIA01362655
281. Apple_AIIA01362603
282. Apple_AIIA01218253
283. Apple_AIIA00445995
284. Apple_AIIA00448231
285. Apple_AIIA00448356
286. Apple_AIIA00583911
287. Apple_AIIA00319724
288. Apple_AIIA00350916
289. Apple_AIIA00953526
290. Apple_AIIA00361480
291. Apple_AIIA00511322
292. Apple_AIIA00531724
293. Apple_AIIA00544957
294. Apple_AIIA00554643
295. Apple_AIIA00961911
296. Apple_AIIA01038556
297. Apple_AIIA01239186

298. Apple_AIIA00318980
299. Apple_AIIA00463988
300. Apple_AIIA00811549
301. Apple_AIIA01024886
302. Apple_AIIA00355620
303. Apple_AIIA01035239
304. Apple_AIIA00949979
305. Apple_AIIA00795191
306. Apple_AIIA00959232
307. Apple_AIIA00960451
308. Apple_AIIA00127056
309. Apple_AIIA00319983
310. Apple_AIIA00106919
311. Apple_AIIA00337602
312. Apple_AIIA00106989
313. Apple_AIIA00949715
314. Apple_AIIA00949772
315. Apple_AIIA00954873
316. Apple_AIIA00467172
317. Apple_AIIA00180654
318. Apple_AIIA001809902
319. Apple_AIIA00954933
320. Apple_AIIA00954724
321. Apple_AIIA00960055
322. Apple_AIIA01383038
323. Apple_AIIA00960073
324. Apple_AIIA00135124
325. Apple_AIIA00963542
326. Apple_AIIA00963282
327. Apple_AIIA00963298
328. Apple_AIIA00708319
329. Apple_AIIA01383386
330. Apple_AIIA01358855
331. Apple_AIIA00974863
332. Apple_AIIA00090447
333. Apple_AIIA00098417
334. Apple_AIIA00099100
335. Apple_AIIA00974436
336. Apple_AIIA00105851
Price Committee Documents
337. Apple_AIIA01278611 - 2001-03-11 iPod March line-up Decision
338. Apple_AIIA01276219 - 2004-06-30 iPod Decision
339. Apple_AIIA01278395 - 2004-10 iBook, Xserve RAID, iPod Decision

340. Apple_AIIA01278345 - 2005-01 iPod, Xserve, Displays, Keyboard & Mouse, Q97, Q88, PowerBook
Decision
341. Apple_AIIA01277415 - 2005-06 iPod Decision
342. Apple_AIIA01277431 - 2005-06-09 iPod Decision
343. Apple_AIIA01384542 - 2005-08-15 PowerBook, iPod, Xserve RAID Decision
344. Apple_AIIA01277614 - 2006-01-18 iPod Decision
345. Apple_AIIA01278341 - 2006-01-18 PCM Back-up iPod Decision
346. Apple_AIIA01277439 - 2006-01-19 iPod Decision
347. Apple_AIIA01277541 - 2006-01-20 iPod Decision
348. Apple_AIIA01277329 - 2006-09 iPod Presentation
349. Apple_AIIA01276233 - 2007-04 iPod Presentation
350. Apple_AIIA01276249 - 2007-08 iPod, iPod nano Decision
351. Apple_AIIA01276307 - 2007-08 iPod, iPod nano, iPod shuffle Decision
352. Apple_AIIA01276342 - 2007-08-16 iPod, iPod nano, iPod shuffle Decision
353. Apple_AIIA01276402 - 2007-09 iPod Presentation
354. Apple_AIIA01276439 - 2007-10 iPod touch, app bundle pricing Decision
355. Apple_AIIA01276447 - 2007-10 iPod touch, software bundle Decision
356. Apple_AIIA01276570 - 2007-10 iPod touch, software bundle, iPhone Decision
357. Apple_AIIA01276584 - 2007-11 iPod touch, software bundle, iPhone Decision
358. Apple_AIIA01276598 - 2007-12-03 iPod touch, iPhone Decision
359. Apple_AIIA01276612 - 2008-01 iPod touch, iPhone Decision
360. Apple_AIIA01278046 - 2008-02-01 iPod shuffle Decision
361. Apple_AIIA01276714 - 2008-03-01 iPod shuffle, iPod touch Decision
362. Apple_AIIA01276791 - 2008-08 iPod touch, iPod classic, iPod nano Decision
363. Apple_AIIA01276935 - 2008-12 iPod shuffle Decision
364. Apple_AIIA01277156 - 2009-02 iPod shuffle Decision
Data
365. Upgrades by month (version 1).xlsb - For US Units and Revenue for Upgardes by month Since
January,2009 to 17th October,2010
366. AIIA_US iPod Sales_Family Level_Direct-Indirect-OEM.xlsx
367. US Sales_iPods_Direct_Indirect_OEM_FY02-FY10
368. NPD_DigitalOnlyJuly06_Dec11.xlsx
369. NPD_PhysicalandDigital_July06_Dec11.xlsx
370. NPDMP3Aug07Update.xls
371. NPDMP3October2005Update.xls
372. NPDUSMP3April2009.xls
373. NPDUSMP3July2010.xls
374. NPDUSMP3June08Update.xls
375. NPDUSMP3PriceBandJuly2010.xls
376. BDF percentages to apply to reseller data.csv
377. rslr_rc_data.sas7bdat
378. IpodClassic2002Q1v3.txt
379. IpodClassic2002Q2v3.txt
380. IpodClassic2002Q3v3.txt

381. IpodClassic2002Q4v3.txt
382. IpodClassic2003Q1v3.txt
383. IpodClassic2003Q2v3.txt
384. IpodClassic2003Q3v3.txt
385. IpodClassic2003Q4v3.txt
386. IpodClassic2004Q1v3.txt
387. IpodClassic2004Q2v3.txt
388. IpodClassic2004Q3v3.txt
389. IpodClassic2004Q4v3.txt
390. IpodMINI2004Q2v3.txt
391. IpodMINI2004Q3v3.txt
392. IpodMINI2004Q4v3.txt
393. RecordCheck_Ipod Sales.xlsx
394. IpodMini2005Q1.txt
395. IpodMini2005Q2.txt
396. IpodMini2005Q3.txt
397. IpodMini2005Q4.txt
398. IpodMini2006to2011Q1.txt
399. IpodShuffle2005.txt
400. IpodShuffle2006.txt
401. IpodShuffle2007Q1.txt
402. IpodShuffle2007Q2.txt
403. IpodShuffle2007Q3.txt
404. IpodShuffle2007Q4.txt
405. IpodShuffle2008Q1.txt
406. IpodShuffle2008Q2.txt
407. IpodShuffle2008Q3.txt
408. IpodShuffle2008Q4.txt
409. IpodNano2005.txt
410. IpodNano2006Q1SC2.txt
411. IpodNano2006Q1SR3.txt
412. IpodNano2006Q2.txt
413. IpodNano2006Q3.txt
414. IpodNano2006Q4.txt
415. IpodNano2007Q1SC2.txt
416. IpodNano2007Q1SR3.txt
417. IpodNano2007Q2.txt
418. IpodNano2007Q3.txt
419. IpodNano2007Q4.txt
420. IpodNano2008Q1SC2.txt
421. IpodNano2008Q1SR3.txt
422. IpodNano2008Q2.txt
423. IpodNano2008Q3.txt
424. IpodNano2008Q4.txt
425. RecordCheck2 - Copy.xls

426. RecordCheck2.xlsx
427. IpodShuffle2009.txt
428. IpodShuffle2010.txt
429. IpodShuffle2011.txt
430. IpodTouch2008Q1.txt
431. IpodTouch2008Q2.txt
432. IpodTouch2008Q3.txt
433. IpodTouch2008Q4.txt
434. IpodTouch2009Q1.txt
435. IpodTouch2009Q2.txt
436. IpodTouch2009Q3.txt
437. IpodTouch2009Q4.txt
438. IpodTouch2010Q1.txt
439. IpodTouch2010Q2.txt
440. IpodTouch2010Q3.txt
441. IpodTouch2010Q4.txt
442. IpodTouch2011Q1.txt
443. IpodTouch2011Q2.txt
444. IpodTouch2011Q3.txt
445. IpodTouch2011Q4.txt
446. IpodClassic2011Q3V2.txt
447. IpodNano2011Q3V2.txt
448. IpodShuffle2011Q3V2.txt
449. IpodTouch2011Q3V2.txt
450. price_group_MISC.csv
451. price_group_1.csv
452. price_group_7.csv
453. price_overirde_12.csv
454. price_overirde_code14.csv
455. SFI_791971_1_PriceOverrideReasonDesc.xls
456. AppleDataExtract_iPod BDF Accruals from 2001 to 2011.csv
457. RetailerCodeNames.csv
458. RetailerCodeNames_2.csv
459. 2002_Ipod_Sales
460. 2003_Ipod_Sales
461. 2004_Ipod_Sales
462. 2005_Ipod_Sales
463. 2006_Ipod_Sales
464. 2007_Ipod_Sales
465. 2008_Ipod_Sales
466. 2009_Ipod_Sales
467. 2010_Ipod_Sales
468. 2011_Ipod_Sales
469. 2002-2004_Ipod_Sales
470. 2005_Ipod_Sales

471. 2006_Ipod_Sales
472. 2007_Ipod_Sales
473. 2008_Ipod_Sales
474. 2009_Ipod_Sales
475. 2010_Ipod_Sales
476. 2011_Ipod_Sales
477. AppleDataExtract_iPod BDF Accruals from 2001 to 2011.xls
478. AIIA_iPod-OEM_HP.xls
479. RetailerCodeNames.xls
480. NPDUSMP3May2009.xls
481. NPDUSMP3July2010.xls
482. 2000-2007 (NPD data).xls
483. ~5844701.xls
484. BDF percentages to apply to reseller data.csv
485. Merged iPod features and costs_direct sales.csv
486. Merged iPod features and costs_direct sales_CKD.csv
487. 2002_Ipod_Sales to 20011_Ipod_Sales
488. Merged iPod features and costs_CKD.csv
489. AIIA - US iTS P&Ls_Q3FY03-Q4FY10
490. iTunes Music Store P&L Data
NPD Monthly Trend Reports
491. Apple_AIIA00109997
492. Apple_AIIA00485090
493. Apple_AIIA00489689
494. Apple_AIIA 00127123
495. Apple_AIIA00129531
496. Apple_AIIA00135174
497. Apple_AIIA00358751
498. Apple_AIIA00358724
499. Apple_AIIA00358776
500. Apple_AIIA00358791
501. Apple_AIIA00358816
502. Apple_AIIA00358856
503. Apple_AIIA00358883
504. Apple_AIIA00431268
505. Apple_AIIA00485090
506. Apple_AIIA00489689
507. Apple_AIIA00505446
508. Apple_AIIA00511004
509. Apple_AIIA00536766
510. Apple_AIIA00555045
511. Apple_AIIA00559475
512. Apple_AIIA00621881
513. Apple_AIIA00625999

514. Apple_AIIA00630064
515. Apple_AIIA01085453
516. Apple_AIIA 00649138
517. Apple_AIIA00733250
518. Apple_AIIA00743040
519. Apple_AIIA00972703
520. Apple_AIIA01061216
521. Apple_AIIA01083128
522. Apple_AIIA01083366
523. Apple_AIIA01110277
524. Apple_AIIA 01121207
525. Apple_AIIA01121373
526. Apple_AIIA01279027
527. Apple_AIIA01326081
528. Apple_AIIA01326098
529. Apple_AIIA01326113
530. Apple_AIIA01326130
531. Apple_AIIA01326147
532. Apple_AIIA01326164
533. Apple_AIIA01326181
534. Apple_AIIA01326198
535. Apple_AIIA01326300
536. Apple_AIIA01326266
537. Apple_AIIA01358885
538. Apple_AIIA01358902
539. Apple_AIIA01358919
540. Apple_AIIA01358970
541. Apple_AIIA01358987
542. Apple_AIIA01358953
543. Apple_AIIA01358963
544. Apple_AIIA01359004
NPD Music Retail Shares
545. Apple_AIIA00167793
546. Apple_AIIA00187687
547. Apple_AIIA00188105
548. Apple_AIIA00189930
549. Apple_AIIA00327772
550. Apple_AIIA00429468
551. Apple_AIIA00662062
552. Apple_AIIA00678009
553. Apple_AIIA00725368
554. Apple_AIIA00946216
555. Apple_AIIA00946224
556. Apple_AIIA00946391

557. Apple_AIIA00946640
558. Apple_AIIA00946649
559. Apple_AIIA00946675
560. Apple_AIIA00960247
561. Apple_AIIA00960296
562. Apple_AIIA00960300
563. Apple_AIIA00960340
564. Apple_AIIA00962442
565. Apple_AIIA00965666
566. Apple_AIIA00965893
567. Apple_AIIA00966687
568. Apple_AIIA00967103
569. Apple_AIIA00968523
570. Apple_AIIA00968532
571. Apple_AIIA00968691
572. Apple_AIIA00968757
573. Apple_AIIA01205535
574. Apple_AIIA01205543
575. Apple_AIIA01205576
576. Apple_AIIA01205585
577. Apple_AIIA01205587
578. Apple_AIIA01205590
579. Apple_AIIA01205610
580. Apple_AIIA01205623
581. Apple_AIIA01206292
582. Apple_AIIA01206330
583. Apple_AIIA01208708
584. Apple_AIIA01208710
585. Apple_AIIA01208997
586. Apple_AIIA01233725
NPD Surveys
587. Apple_AIIA01309399
588. Apple_AIIA00178974
589. Apple_AIIA01309615
590. Apple_AIIA00945746
591. Apple_AIIA00964631
592. Apple_AIIA00178967
593. Apple_AIIA01217873
594. Apple_AIIA00968565
595. Apple_AIIA01309456
596. Apple_AIIA00952642
597. Apple_AIIA00481774
598. Apple_AIIA00945746
599. Apple_AIIA00948134

600. Apple_AIIA00952622
601. Apple_AIIA00957936
602. Apple_AIIA01214028
603. Apple_AIIA01215207
604. Apple_AIIA01215317
Apple Marketing Requirements Documents (MRDs)
605. Apple_AIIA01333844
606. Apple_AIIA00989705
607. Apple_AIIA00989713
608. Apple_AIIA01052191
609. Apple_AIIA01054705
610. Apple_AIIA00982117
611. Apple_AIIA00989717
612. Apple_AIIA00114451
613. Apple_AIIA00984872
614. Apple_AIIA01055013
615. Apple_AIIA00115742
616. Apple_AIIA01337461
617. Apple_AIIA_B_001604
618. Apple_AIIA01068163
619. Apple_AIIA01064431
620. Apple_AIIA01067923
621. Apple_AIIA01337470
622. Apple_AIIA01063781
623. Apple_AIIA01068809
624. Apple_AIIA01057554
625. Apple_AIIA01066740
Public Books and Articles
626. Apple Press Releases relating to iPod, iTunes, iTunes Music Store or iTunes Store, iPad, Apple
TV, & iPhone
627. Joshua D. Angrist and Jörn-Steffen Pischke, Mostly Harmless Econometrics: An Empiricist’s
Companion. March 2008
628. Jobs, Steve, “Thoughts on Music”, February 6, 2007
629. Brad Stone, “Want to Copy iTunes Music? Go Ahead, Apple Says.” New York Times. January 7,
2009
630. Mitchell Hall, “Apple iTunes going DRM-free.” The National Business Review. January 7, 2009
631. Smith, Tony, "Apple plugs PyMusique iTunes 'hole.'" The Register. March 22, 2005,
http://www.theregister.co.uk/2005/03/22/apple_blocks_pymusique/ (accessed May 30, 2013)
632. “Apple cuts iTunes pricing, eases copy protection.” NBCNews.com. January 6, 2009
633. Brent Schlender, “How Big Can Apple Get?” Fortune Magazine. February 21, 2005
634. Tony Smith, “Downloading digital music - Majors and minors, players and platforms, lawsuits
and licenses.” The Register. December 24, 2004

635. “FairPlay – QTFairUse,”, Wikipedia, The Free Encyclopedia,
http://en.wikipedia.org/wiki/FairPlay, (accessed May 28, 2013).
636. “FireWire vs. USB 2.0.” Q Imagining Technical Note
637. “Apple iPod shuffle 2nd Gen (Silver) 1 GB Specs.” EveryiPod.com. Accessed: May 28, 2013
638. “iPod: Not Just iCandy.” PCMag.com Accessed: May 28, 2013
http://www.pcmag.com/article2/0,2817,19362,00.asp
639. “iTunes strips songs' copy protection.” Boston.com January 7, 2009 Accessed: May 28, 2013
www.boston.com/business/technology/articles/2009/01/07/itunes strips songs copy protecti
on/#
640. Levy, Steven, The Perfect Thing: How the iPod Shuffles Commerce, Culture and Coolness. Simon
& Schuster Paperbacks. New York, NY 2006
641. Katie Dean, “Listen Up: Songs Now 79 Cents.” The New Yorker. May 28, 2003
642. Richtel, Matt, “TECHNOLOGY; Apple Introduces What It Calls an Easier to Use Portable Music
Player,” NY Times, October 24, 2001,
http://www.nytimes.com/2001/10/24/business/technology-apple-introduces-what-it-calls-aneasier-to-use-portable-music-player.html (accessed 9/25/2012)
643. “New & Noteworthy: New Apple retail stores to open; LCD rackmounts; Apple slams Microsoft
proposal; more” CNET News November 28, 2001
644. Copeland, Michael V., “Unlocking the iPod: If you want to buy content for Apple's iPod, you
have to go through iTunes. Navio Systems is hoping to change that.,” Business 2.0 Magazine
(downloaded from CNN Money.com), May 22, 2006.
http://money.cnn.com/2006/05/22/technology/business2_launchpad0522/ (accessed May 28,
2013)
645. Bell, Ian , “RealNetworks Snaps Apple’s iPod Exclusivity”, Digital Trends, July 26, 2004,
http://www.digitaltrends.com/gadgets/realnetworks-snaps-apples-ipod-exclusivity/ (accessed
accessed May 28, 2013).
646. “Publication 946, How to Depreciate Property.” Internal Revenue Service. 2012
647. Tony Smith, “Real '49c a song' promo pushes downloads to 3m.” The Register. September 9, 2004
648. Peter Cohen, “RealNetworks' Harmony promises iPod compatibility.” PCWorld. July 26, 2004
http://www.macworld.com/article/1035237/harmony.html (accessed May 28, 2013)
649. Borland, John, "iTunes hack disabled by Apple." CNET News. March 21, 2005,
http://news.cnet.com/Apple-disables-iTunes-hack/2100-1027_3-5628616.html#addcomm
(accessed May 30, 2013)
650. “The SURVEYREG Procedure: Syntax.” SAS/STAT(R) 9.22 User's Guide Accessed: June 19, 2013
support.sas.com/documentation/cdl/en/statug/63347/HTML/default/viewer.htm#statug_sur
veyreg_sect006.htm
651. SAS/STAT(R) 9.2 User's Guide, Second Edition, “FREQ Statement”., Available at
http://support.sas.com/
documentation/cdl/en/statug/63033/HTML/default/viewer.htm#statug_reg_sect011.htm
652. Tim Lambert, “Why You Need to Correct for Clustering - Deltoid.” Science Blogs. September 10,
2003 scienceblogs.com/deltoid/2003/09/10/cluster/
653. Tony Smith, “Sony opens US music download store.” The Register. May 4, 2004 Accessed: May
29 2013 http://www.theregister.co.uk/2004/05/04/sony_connect_us/

654. Mariano, Gwendolyn, "Listen.com to tune in new music service," CNET News, December 2, 2001,
http://news.cnet.com/Listen.com-to-tune-in-new-music-service/2100-1023_3-276476.html
(accessed July 16, 2013)
655. Benny Evangelista, “Music firms open online services, but will fans pay?” San Francisco
Chronicle. December 3, 2001 Accessed: May 30, 2013
656. Evangelista, Benny, "Industry starting to endorse Net music / Listen.com to offer songs from all
five major labels," Listen.com, July 1, 2002, http://www.sfgate.com/business/article/Industrystarting-to-endorse-Net-music-2801248.php (accessed July 16, 2013).
657. CNET Staff, “New & Noteworthy: CNET Reviews,” November 28, 2001,
http://reviews.cnet.com/8301-13727_7-1036303-263.html (accessed 9/25/2102).
658. Kahney, Leander “The Cult of iPod,” San Francisco, CA: No Starch Press, Inc. (2005), p. 3
659. 2006 PC World Innovations Awards,” PR Newswire, January 4, 2006,
http://www.prnewswire.com/news-releases/2006-pc-world-innovations-awards-winnersunveiled-53127872.html (accessed July 16, 2013).
660. Bajarin, Ben, “Why Competing with Apple Is So Difficult,”Time Magazine. July 01, 2011Accessed:
July 18, 2013 http://techland.time.com/2011/07/01/why-competing-with-apple-is-sodifficult/#ixzz2ZN8itJ6i
661. “iPod Mini,” Wikipedia, The Free Enclyclopedia, https://en.wikipedia.org/wiki/IPod_Mini
(accessed July 16, 2013)
662. Drew Turner, Daniel, “Apple hits $1 Billion in Profit for 2005,” eWeek, October 10, 2011,
http://www.eweek.com/c/a/Apple/Apple-Hits-1-Billion-in-Profit-for-2005/ (accessed July 16,
2013)
663. “Closed Platform,” Wikipedia, The Free Enclyclopedia,
http://en.wikipedia.org/wiki/Closed_platform (accessed July 17, 2013).
664. Bloomberg Businessweek, "RealPlayer: Master Music Manager," May 25, 2005,
http://www.businessweek.com/stories/2005-05-25/realplayer-master-music-manager,
(accessed May 28, 5/28/2013)
665. )"Managing Music with Windows Media Player," PCUG,
http://www.pcug.org.au/pcug/training/upcd02dm01/dmLesson01Media%20Player.pdf,
January 17, 2004 (accessed May 28, 2013)
666. Ku, Andrew, "Dell's Digital Jukebox 20 - Music, the Dell Way," AnandTech, June 20, 2004,
http://www.anandtech.com/show/1354/9 (accessed June 4, 2013)
667. Boswell, Wendy, "Manage your music with Winamp," Lifehacker, July 8, 2007,
http://lifehacker.com/276000/manage-your-music-with-winamp (accessed May 28, 2013)
668. Sauners, Gareth J M, "Anapod - transfer files to your iPod via Explorer," November 19, 2007,
http://blog.garethjmsaunders.co.uk/archives/2007/11/19/anapod-transfer-files-to-your-ipodvia-explorer/ (accessed May 28, 2013)
669. Ulrich, Adrian, "Gnupod; Manage your iPod," GNU, last updated September 20, 2007,
http://www.gnu.org/software/gnupod/gnupod.html (accessed May 28, 2013)
670. "How to Use the EphPod IPod Manager," eHow, http://www.ehow.com/how_2145859_useephpod-ipod-manager.html (accessed May 28, 2013)
671. “Napster,” Wikipedia, The Free Enclyclopedia, http://en.wikipedia.org/wiki/Napster (accessed
July 16, 2013)
672. Kravitz, David, “Dec. 7, 1999: RIAA Sues Napster,” Wired, December 7, 2009,
http://www.wired.com/thisdayintech/2009/12/1207riaa-sues-napster/ (accessed July 16, 2013)

673. Tyson, Dan, “The 25 Worst Tech Products of all Time,” PCWorld, May 26, 2006,
http://www.pcworld.com/article/125772/worst_products_ever.html (accessed July 16, 2013)
674. “iTunes Store,” Wikipedia, The Free Enclyclopedia, http://en.wikipedia.org/wiki/ITunes_Store
(accessed July 16, 2013)
675. Lewis, Peter, “Product of The Year Apple iTunes Music Store,” Fortune Magazine (downloaded
from CNN Money), December 22, 2003,
http://money.cnn.com/magazines/fortune/fortune_archive/2003/12/22/356108/ (accessed
July 16, 2013)
676. U.S. Department of Justice, Interoperability Between Antitrust and Intellectual Property, Thomas
O. Barnett Presentation to the George Mason School of Law Symposium Managing Antitrust
Issues In a Global Marketplace, Washington, D.C., September 13, 2006,
http://www.usdoj.gov/atr/public/speeches/218316.htm (accessed July 16, 2013)
677. CNET News, “Real offers new tech, song store,.” January 7, 2004, http://news.cnet.com/21001027-5136275.html (accessed July 16, 2013)
678. ArsTechnica, “Amazon’s MP3 store brings more DRM-free music at lower prices than iTunes
Store.” September 25, 2007
679. “Amazon to Sell Warner Music Minus Copy Protection.” NY Times. December 28, 2007
680. “Sony Joins Other Labels on Amazon MP3 Store.” NY Times. January 11, 2008
681. Antone Gonsalves, “Amazon Adds Fourth Major Record Label To DRM-Free Music Store,”
Information Week, January 10, 2008
682. “Application Store,” The Free Encyclopedia, http://en.wikipedia.org/wiki/Application_store
(accessed July 16, 2013)
683. Dalrymple, Jim, “iPod DRM faces another reverse-engineering challenge,” PCWorld, November
22, 2005, http://www.macworld.com/article/1048070/ipoddrm.html (accessed June 17, 2013)
684. Healey, Jon. “Napster and Rhapsody compete for headlines,” Los Angeles Times. September 20,
2006 http://opinion.latimes.com/bitplayer/2006/09/ (accessed June 17, 2013)
685. Bernoff, Josh, “Apple iTunes Jump-Starts Windows Digital Music,” Forrester Research, Inc.,
October 17, 2003
686. Leonard, Devin, "Songs In The Key Of Steve Steve Jobs may have just created the first great legal
online music service. That's got the record biz singing his praises." Fortune Magazine,
downloaded from CNNMoney.com, May 12, 2003
http://money.cnn.com/magazines/fortune/fortune_archive/2003/05/12/342289/ (accessed
May 28, 2013).
687. John Berra, Walled Garden or Open Field?, Automation World, Feb. 2, 2012
688. Hee-Yeon Cho, Choon-Sung Nam, Dong-Ryeol Shin, 2010 International Conference on
Electronics and Information Engineering (ICEIE 2010), pg. V2-141
689. Roland M. Müller, A Comparison of Inter-Organizational Business Models of Mobile App Stores:
There is More than Open vs. Closed, Journal of Theoretical and Applied Electronic Commerce
Research, Vol. 6, Issue 2 (Aug. 2011), pg. 72-73
690. Gobry, Pascal-Emmanuel, “How Amazon Makes Money From The Kindle,” Business Insider,
October 18, 2011, http://www.businessinsider.com/kindle-economics-2011-10 (accessed July 16,
2013)
691. Akasie, Jay, “With New Kindle, Bezos Proves Ecosystems Matter More Than Hardware,”
September 7, 2012 http://www.minyanville.com/sectors/technology/articles/amazon-newkindle-amazon-kindle-family/9/7/2012/id/43794 (accessed July 16, 2013)

692. “Closed Platform,” Wikipedia, The Free Enclyclopedia,
http://en.wikipedia.org/wiki/Closed_platform (accessed July 17, 2013)
693. “MP3 Software Review”, 10TopTenReviews, http://mp3-software-review.toptenreviews.com/
(accessed June 10, 2013)
694. “Music Management software,” CNET, http://download.cnet.com/windows/musicmanagement-software/ (accessed June 10, 2013)
695. Levine, Robert, "Unlocking the iPod; Jon Johansen became a geek hero by breaking the DVD
code. Now he's liberating iTunes - whether Apple likes it or not." Fortune Magazine, downloaded
from CNNMoney.com, October 23, 2006,
http://money.cnn.com/magazines/fortune/fortune_archive/2006/10/30/8391726/ (accessed
May 28, 2013)
696. Gannes, Liz, “DVD Jon Fairplays Apple,” Gigaom.com, October 2, 2006,
http://gigaom.com/2006/10/02/dvd-jon-fairplays-apple/ (accessed May 28, 2013).
697. “PyMusique,” Wikipedia, The Free Encyclopedia, http://en.wikipedia.org/wiki/PyMusique
(accessed May 28, 2013)
698. Borland, John, “RealNetworks rekindles iPod Tech Tussle,” CNETNews, April 26, 2005,
http://news.cnet.com/2100-1027_3-5685286.html (accessed May 28, 2013)
699. Brian S. Hall, “The Apple Walled Garden is Grounded in Old Fashioned Product Superiority,”
Techpinions.com. March 4th, 2013 http://techpinions.com/the-apple-walled-garden-isgrounded-in-old-fashioned-product-superiority/14796 (accessed July 18, 2013)
700. Graham, Jefferson, “Real says digital song sale doubled market share,” USA Today, September
09, 2004, http://usatoday30.usatoday.com/tech/techinvestor/corporatenews/2004-09-08real_x.htm (accessed July 2, 2013)
701. Jemima Kiss, “How Big Is the iPod Installed Base?” Guardian, September 9, 2009 (reporting
discussion with executive at Forrester Research)
702. Larry Dignan, “Tablet Replacement Rates: More Like an MP3 Player than PC,” ZDNet January 4,
2011 (reporting a Forrester Research study)
703. “Mobile Phone Lifecycles,” GSM Association, 2006 (reporting that about half of phone sales are
replacements and that the replacement rate is about 18 months)
704. “The Life Cycle of a Cell Phone,” U.S. Environmental Protection Agency, 2005 (reporting cell
phone replacement rate of 18 months)
705. John Paczkowski, “I Got a Fever, and the Only Prescription is… More iPhone!” All Things
Digital, June 25, 1010 (reporting that the replacement cycle for iPhones is 14.7 months)
706. Victor H., “Americans Replace Their Cell Phones Every 2 Years, Finns – Every Six, a Study
Claims,” Phonearena.com, July 11, 2011
707. Griliches, Zvi, “Hedonic Price Indexes and the Measurement of Capital and Productivity: Some
Historical Perspectives,.” NBER Working Paper 2634 (June 1988)
708. “Choosing the Correct Weight Syntaxis,” Carolina Population Center at the University of North
Carolina at Chapel Hill,
http://www.cpc.unc.edu/research/tools/data_analysis/statatutorial/sample_surveys/weight_
syntax (accessed June 20, 2013)
709. William H. Greene, Econometric Analysis, 7th Edition, Prentice Hall, 2012, Chapter 11. See also
Joshua Angrist and Jorn-Steffan Pischke, Mostly Harmless Econometrics, Princeton University
Press, 2009, Chapter 8, “Clustering and Serial Correlation in Panels

710. “Proving Antitrust Damages: Legal and Economic Issues”, ABA Publishing, Second Edition,
2010
711. "Apple offers variable pricing, more DRM-free music on iTunes," Los Angeles Times, Janaury 6,
2009, http://latimesblogs.latimes.com/technology/2009/01/apple-blinks-st.html (accessed June
28, 2013)
712. Burnette, Ed, "Apple iTunes goes all DRM free, with three price tiers," ZDNet.com,
http://www.zdnet.com/blog/burnette/apple-itunes-goes-all-drm-free-with-three-pricetiers/702 (accessed June 28, 2013)
713. McLean, Prince, "iTunes Store goes DRM Free, offers over-the-air downloads," Apple Insider,
January 6, 2009, http://appleinsider.com/articles/09/01/06/itunes_store_goes_drm_free
(accessed June 28, 2013)
714. "iTunes Store and DRM-free music: What you need to know," MacWorld,
http://www.macworld.com/article/1138000/drm_faq.html (accessed January 7, 2009)
715. Kennedy, Peter E., “Estimation with Correctly Interpreted Dummy Variables in Semilogarithmic
Equations,” The American Economic Review, Vol. 71, No.4 (Sep., 1981), 801.
716. “iTunes version history,” Wikipedia, The Free Encyclopedia,
http://en.wikipedia.org/wiki/ITunes_version_history (accessed June 25th, 2013)
717. Mook, Nate, “Apple Blocks iTunes DRM Hack.” BetaNews. March 21, 2005
http://betanews.com/2005/03/21/apple-blocks-itunes-drm-hack/ (accessed July 18th, 2013)
718. “emusic,” Wikipedia, The Free Encyclopedia, http://en.wikipedia.org/wiki/Emusic (accessed
May 28, 2013)
719. “Kazaa site becomes legal service,” BBC News,
http://newsvote.bbc.co.uk/mpapps/pagetools/print/news.bbc.co.uk/2/hi/technology/522040
6.stm (accessed 5/28/2013)
720. “Encyclopedia: Kazaa,” PCMag.com, http://www.pcmag.com/encyclopedia/term/45734/kazaa
(accessed 5/28/2013)
721. “LimeWire,” Wikipedia, The Free Encyclopedia, http://en.wikipedia.org/wiki/Limewire
(accessed 5/28/2013)
722. “Morpheus (software),” Wikipedia, The Free Encyclopedia,
http://en.wikipedia.org/wiki/Morpheus_(software) (accessed 5/28/2013)
723. “National Geographic unveils World Music downloads service,” PC Pro.co.uk,
http://www.pcpro.co.uk/news/90308/national-geographic-unveils-world-music-downloadsservice (accessed 5/28/2013)
724. “iPod,” Wikipedia, The Free Encyclopedia, http://en.wikipedia.org/wiki/Ipod (accessed
5/28/2013)
725. “iTunes,” Wikipedia, The Free Encyclopedia, http://en.wikipedia.org/wiki/Itunes (accessed
5/28/2013)
726. “iTunes Store,” Wikipedia, The Free Encyclopedia, http://en.wikipedia.org/wiki/Itunes_Store
(accessed 5/28/2013)
727. “Apple TV,” Wikipedia, The Free Encyclopedia, http://en.wikipedia.org/wiki/Apple_TV
(accessed 5/28/2013)
728. Associated Press, “RealNetworks: Files Play on IPod,” Wired Magazine. July 25, 2004
http://www.wired.com/science/discoveries/news/2004/07/64341 (accessed 7/17/2013)
729. Eric Dahl, “The Playlist: Legal Digital Tunes Go P2P,” PCWorld, April 9, 2004
http://www.pcworld.com/article/115553/article.html (accessed 7/17/2013)

730. “iPad,” Wikipedia, The Free Encyclopedia, http://en.wikipedia.org/wiki/iPad (accessed
5/28/2013)
731. “iCloud,” Wikipedia, The Free Encyclopedia, http://en.wikipedia.org/wiki/iCloud (accessed
5/28/2013)
732. Haire, Meaghan, “A Brief History of the Walkman,” Time Magazine. July 01, 2009
http://www.time.com/time/nation/article/0,8599,1907884,00.html (accessed 5/28/2013)
733. Simcoe, Robert J., “The Revolution in Your Pocket: FIFTY YEARS AGO A TRANSISTOR RADIO
CALLED THE TR1 STARTED THE REMAKING OF THE WORLD BY SEMICONDUCTORS,”
Invention & Technology Magazine Fall 2004 Volume 20, Issue 2.
http://web.archive.org/web/20081203130622/http://www.americanheritage.com/articles/ma
gazine/it/2004/2/2004_2_12_print.shtml
734. Smith, Tony, “Ten years old: the world's first MP3 player. Who remembers the MPMan F10
now?” The Register. March 10, 2008
http://www.theregister.co.uk/2008/03/10/ft_first_mp3_player/ (accessed 7/17/2013)
735. Cohen, Peter, “PyMusique Author Hacks Apple's ITunes Fix,” PCWorld. March 23, 2005
http://www.pcworld.com/article/120146/article.html (accessed 7/17/2013)
736. “Pandora Radio,” Wikipedia, The Free Encyclopedia,
http://en.wikipedia.org/wiki/Pandora_Radio (accessed 5/28/2013)
737. Perez, Juan Carlos, “Sony Spins New Online Music Store,” PCWorld. May 5, 2004
http://www.pcworld.com/article/115995/article.html (accessed 5/28/2013)
738. “Amazon.com Launches Public Beta of Amazon MP3, a Digital Music Store Offering Customers
Earth’s Biggest Selection of a la Carte DRM-Free MP3 Music Downloads,” Business Wire.
September 25, 2007
http://www.businesswire.com/portal/site/google/index.jsp?ndmViewId=news_view&newsId
=20070925005710&newsLang=en (accessed 7/17/2013)
739. Stone, Brad, “Sony Joins Other Labels on Amazon MP3 Store,” New York Times. January 11, 2008
http://www.nytimes.com/2008/01/11/technology/11sony.html?_r=1&adxnnl=1&ref=technolo
gy&adxnnlx=1366826676-ymmK7p9HEqdahXynzAiDzA& (accessed 7/17/2013)
740. “Old Version of iTunes,” OldApps.com
http://www.oldapps.com/itunes.php?old_itunes+4#changelog (accessed 7/17/2013)
741. Appleyard, David, “iTunes 4.8 Released,” Gigaom. May. 9, 2005
http://gigaom.com/2005/05/09/itunes-48-released/ (accessed 7/17/2013)
742. AppleInsider Staff, “Apple releases iTunes 7.1, QuickTime 7.1.5, more,” Apple Insider. March 5,
2007
http://appleinsider.com/articles/07/03/05/apple_releases_itunes_71_quicktime_715_more
(accessed 7/17/2013)
743. AppleInsider Staff, “iTunes 7.3 supports iPhone, adds Apple TV photo streaming,” Apple
Insider. June 29, 2007
http://appleinsider.com/articles/07/06/29/itunes 7 3 supports iphone adds apple tv photo
_streaming (accessed 7/17/2013)
744. Dalrymple, Jim, “Apple releases iTunes 8.1 update,” MacWorld. March 11, 2009
http://www.macworld.com/article/1139330/itunes.html (accessed 7/17/2013)
745. Hugh Hurt, “EMI's Last-Ditch Effort: DRM-Free Music.” Wired. April 3, 2007 Accessed: May 28,
2013 http://www. wired.com/entertainment/music/news/2007 /04/emi_business0403

746. Richard MacManus, “EMI Music DRM-free: What It Means For The Online Music Industry.”
ReadWrite.com April 2, 2007 Accessed: May 28, 2013 http://readwrite.com/2007
/04/02/emi_music_drm_free
747. Candace Lombardi, “iTunes goes DRM-free with EMI.” CNET News. May 30, 2007 Accessed:
May 28, 2013 http://news.cnet.com/iTunes-goes-DRM-free-with-EMI/2100-1027_3-6187457.html
748. “EMI Music Launches DRM-Free Superior Sound Quality Downloads Across Its Entire Digital
Repertoire.” April 2, 2007 Accessed: May 28, 2013 http:/ /shaungroves.com/2007 /04/emi -drmpress-release/
749. “Apple to End Music Restrictions.” BBC News. January 7 2009 Accessed: May 28, 2013
http://news.bbc.co.uk/2/hi/technology/7813527.stm
750. “SharpMusique.” Wikipedia. Accessed: May 31, 2013 http://en.wikipedia.org/wiki/PyMusique
751. Eric Bleeker, “1 Acquisition Apple Doesn't Want to See: Samsung Buying Pandora.” The
MoHeyFool. March 3, 2013 Acessed: March 5, 2013
http://www.fool.com/investing/general/2013/03/03/1-acquisition-apple-doesnt-want-to-seesamsung-buy.aspx
752. “The History & Future of Real Networks,” Internet Video Magazine. Accessed: May 11, 2013
http://internetvideomag.com/Articles-2004/022324Real.htm
753. Adrian Covert, “A decade of iTunes singles killed the music industry.” Fortune Magazine. April
25, 2013 http://money.cnn.com/2013/04/25/technology/itunes-music-decline/index.html
(accessed 7/17/2013)
754. John Borland, “Cell phones heading into iPod territory” CNET News. July 1, 2004 Accessed:
April 23, 2013 news.cnet.com/Cell-phones-heading-into-iPod-territory/2100-1041_35254146.html
755. John Borland, “Cell phones take iPod challenge,” CNET News. November 12, 2004 Accessed:
April 23, 2013 news.cnet.com/Cell-phones-take-iPod-challenge/2100-1039_3-5449095.html
756. John Borland, “iTunes ushers in a year of change,” CNET News. April 26, 2004 Accessed: April
23, 2013 news.cnet.com/iTunes-ushers-in-a-year-of-change/2100-1027_3-5199227.html
757. John Borland, “Music moguls trumped by Steve Jobs?” CNET News. April 15, 2005 Accessed:
April 23, 2013 news.cnet.com/Music-moguls-trumped-by-Steve-Jobs---page-3/2100-1027_35671705-3.html
758. Ryan Naraine, “Music Store Embedded in New RealPlayer 10,” InternetNews.com. January 07,
2004 Accessed: May 30, 2013 http://www.internetnews.com/ecnews/article.php/3296171/Music+Store+Embedded+in+New+RealPlayer+10.htm
759. Marc Ferranti, “RealNetworks Puts Harmony on any Player.” IDG News Service. July 26, 2004
http://www.techhive.com/article/117062/article.html (accessed 7/17/2013)
760. Colin Cameron and Douglas L. Miller, “robust Inference with clustered Data,” in Handook of
Empirical Economics and Finance, edited by Aman Ullah and David E. A. Giles, CRC Press 2011
761. Kennedy, Peter, “A guide to Econometrics,” Second Ed.
762. Daniel L. Rubinfeld, Quantitative Methods in Antitrust, in 1 ISSUES IN COMPETITION LAW
AND POLICY 723 (ABA Section of Antitrust Law 2008)
763. “7 Digital About Us,” http://about.7digital.com/ (accessed 7/17/2013)
764. “7digital,” Wikipedia, The Free Encyclopedia, http://en.wikipedia.org/wiki/7digital (accessed
7/17/2013)
765. Clegg, Dean, “Addictech,” http://www.trademarkia.com/addictech-76553281.html (accessed
7/17/2013)

766. Hanlon, Tim, “digiwax = DRM-free MP3 + Vinyl,” Gizmag. May 15, 2007
http://www.gizmag.com/go/7269/ (accessed 7/17/2013)
767. “Amazon MP3,” Wikipedia, The Free Encyclopedia,
http://en.wikipedia.org/wiki/Amazon_MP3 (accessed 7/19/2013)
768. “Amie Street,” Wikipedia, The Free Encyclopedia, http://en.wikipedia.org/wiki/Amie_Street
(accessed 7/19/2013)
769. “Google Play,” Wikipedia, The Free Encyclopedia, http://en.wikipedia.org/wiki/Google_Play
(accessed 7/19/2013)
770. “Arkade Overview,” http://www.arkade.com/content/Bottom+Banner/Arkade+Overview
(accessed 7/19/2013)
771. “Artistxite About Us,” http://artistxite.com/us/magazine/staticabout/8280/About-Us.html
(accessed 7/19/2013)
772. “ArtistXite,” http://musowiki.net/index.php/ArtistXite (accessed 7/19/2013)
773. “WWW.ARTISTXITE.COM,” http://website.informer.com/artistxite.com (accessed 7/19/2013)
774. “Bandcamp,” Wikipedia, The Free Encyclopedia, http://en.wikipedia.org/wiki/Bandcamp
(accessed 7/19/2013)
775. “Bandcamp Terms of Use” http://bandcamp.com/terms_of_use (accessed 7/19/2013)
776. “Beatport,” Wikipedia, The Free Encyclopedia, http://en.wikipedia.org/wiki/Beatport (accessed
7/19/2013)
777. “Bleep.com” Wikipedia, The Free Encyclopedia, http://en.wikipedia.org/wiki/Bleep.com
(accessed 7/19/2013)
778. “About Boomkat” http://www.boomkat.com/about/ (accessed 7/19/2013)
779. “Boomkat - Your independent music specialist,” http://www.webwiki.com/boomkat.com
(accessed 7/19/2013)
780. “BuyMusic,” Wikipedia, The Free Encyclopedia, http://en.wikipedia.org/wiki/BuyMusic
(accessed 7/19/2013)
781. Graham, Jefferson, “BuyMusic's downloads strike a sour note,”USAToday. July 28, 2003
http://usatoday30.usatoday.com/tech/news/2003-07-28-buymusic_x.htm (accessed 7/19/2013)
782. “iTunes Music Store vs. BuyMusic.com,”
http://www.underconsideration.com/speakup/archives/001534.html (accessed 7/19/2013)
783. “CD Baby – About Us,” http://www.cdbaby.com/About (accessed 7/19/2013)
784. “CD Baby,” Wikipedia, The Free Encyclopedia http://en.wikipedia.org/wiki/CD Baby
(accessed 7/19/2013)
785. “Cdigix,” Wikipedia, The Free Encyclopedia http://en.wikipedia.org/wiki/Cdigix (accessed
7/19/2013)
786. Young, Jeffrey R., ” Cdigix Ceases Operations, Citing Poor Economy,” The Chronicle of Higher
Education. February 12, 2009 http://chronicle.com/blogs/wiredcampus/cdigix-ceasesoperations-citing-poor-economy/4528
787. “About Classical Archives,” http://www.classicalarchives.com/about.html (accessed
7/19/2013)
788. “Classical Archives ,” Wikipedia, The Free Encyclopedia
http://en.wikipedia.org/wiki/Classical_Archives (accessed 7/19/2013)
789. “Digital-tunes Homepage” http://www.digital-tunes.net/ (accessed 7/19/2013)
790. “Underground music service Digital Tunes launches! “ PRLEAP.COM April 27, 2007
http://www.prleap.com/pr/73840/ (accessed 7/19/2013)

791. “Discogs,” Wikipedia, The Free Encyclopedia http://en.wikipedia.org/wiki/Discogs (accessed
7/19/2013)
792. All documents and other information, including information from various websites, as listed or
cited in the report or the footnotes, exhibits, and appendices thereto

Appendix C: Collection of iPod Characteristics
1.

The iPod characteristics research was initiated based on the list of order numbers (also

called “MPN” or “Product Identifier”) provided in the direct sales and reseller sales transaction
data from Professor Noll’s latest declaration backup materials. The order numbers were then
individually searched online, and the available characteristics from all sources consulted were
collected. For example, by searching order number “M8737LL/A” online, we collected
information from sources such as“support.apple.com”, “everymac.com”, and “amazon.com”.
The list of characteristic variables is a result of the union of the information collected from the
different sources.
2.

The full list of sources consulted includes Apple.com, Apple Price Documents (provided

by client), Everymac.com (also known as “everyipod.com”), abt.com, amazon.com,
bestbuy.com, brokerbin.com, buy.com, cdw.com, cnet.com, collegestoreonline.com,
cowboom.com, ebay.com, facebook.com, falabella.com, flash-memory-store.com,
ipodused.com, lowerpricestoday.com, milo.com, model spec, mp3-players.toptenreview.com,
nexttag.com, onyougo.com, outlet.amazonwebstore.com, overstock.com, pacificgeek.com,
partnumber.org, pcsuperstore.com, personafile.com, pricecheck.co.za, pricegrabber.com,
reviewindex.net, shop.com, shop.neobits.com, techforless.com, Todoclon.com, toolowtogo.com,
topperise.ch, warrantylife.com, wikipedia, youfindit.ca.
3.

After the search of characteristics was completed for all order numbers, we compared and

consolidated the information from various sources for each order number to obtain the order
number level data. We first ranked the sources for a given order number depending on its
reliability and the completeness of its information (referred to as “the ranking process” below).
Reliability is determined by degree of consistency between the information reported by the
source and Apple.com. Apple.com is considered the most reliable source, and all characteristics
available on the Apple site were used for in constructing our characteristics dataset. However
when information for certain characteristics or order numbers was not available on Apple.com
(e.g., “discontinued dates”), we would refer to everymac.com as our second best source for
characteristics since it has the most comprehensive technical specifications for most iPods, and
its information is consistent with that on Apple.com when the latter is available. With these two

sources (apple.com and everymac.com), we were able to collect information on the
characteristics for most models across all iPod families.
4.

We further compared characteristics of the order numbers within the same iPod family to

collapse the information from the order number level to the family level. Order numbers within
the same family should have the same characteristics. Through comparison, we resolved
differences in characteristics by cross-checking the information against the comparable iPod
description in the corresponding Apple press release. For example, suppose “order No. 123” and
“order No. 321” belong to the same iPod family. We checked their characteristics for internal
consistency and also for consistency with the description on the Apple’s product launch press for
the respective family. If all sources agree, we use the characteristics for the family. If not, we
adopt the information from the Apple press release when available. There are a few models for
which some characteristics are not reported by Apple. In that case, we repeat the ranking process
and use the information from the most reliable source. When using information from sources
other than Apple.com, we verified the data by crosschecking multiple other sources. For
example, if Apple.com does not report the battery life for iPod X and everymac.com reports 8
hours, we verified the data against a third reliable source such as “Amazon.com” or “CNET
review,” if possible.
5.

Final results are compared with the characteristics provided in Professor Noll’s latest

declaration backup materials. The non-trivial differences between the two datasets are
summarized in the table below. (In making this table, we exclude, for example, a difference in
product width due to rounding – 2.43 inches vs. 2.4 inches.)

No.

VARIABLE

Class

MODEL
family

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
29
30
31
32
33
34
35
36
37
38
39
40

1st Reprice - Month
End of Life - Price
End of Life - Price
End of Life - Price
End of Life - Price
End of Life - Month
End of Life - Month
End of Life - Year
End of Life - Year
End of Life - Year
End of Life - Year
Music Battery Hours
Music Battery Hours
Music Battery Hours
Size (cubic inches)
Size (cubic inches)
Size (cubic inches)
Size (cubic inches)
Size (cubic inches)
Size (cubic inches)
Weight (oz )
Onboard Ram (MB)
Onboard Ram (MB)
Playback Support Formats
Playback Support Formats
Playback Support Formats
Playback Support Formats
Playback Support Formats
Playback Support Formats
Playback Support Formats
Playback Support Formats
Playback Support Formats
Playback Support Formats
Playback Support Formats
Connectivity
Connectivity
Connectivity
Connectivity
Connectivity
Connectivity

IPOD CLASSIC
IPOD CLASSIC
IPOD CLASSIC
IPOD SHUFFLE
IPOD SHUFFLE
IPOD SHUFFLE
IPOD SHUFFLE
IPOD SHUFFLE
IPOD SHUFFLE
IPOD SHUFFLE
IPOD SHUFFLE
IPOD TOUCH
IPOD TOUCH
IPOD TOUCH
IPOD MINI
IPOD CLASSIC
IPOD CLASSIC
IPOD CLASSIC
IPOD CLASSIC
IPOD CLASSIC
IPOD SHUFFLE
IPOD TOUCH
IPOD TOUCH
IPOD MINI
IPOD NANO
IPOD NANO
IPOD CLASSIC
IPOD SHUFFLE
IPOD SHUFFLE
IPOD SHUFFLE
IPOD SHUFFLE
IPOD SHUFFLE
IPOD SHUFFLE
IPOD SHUFFLE
IPOD CLASSIC
IPOD CLASSIC
IPOD CLASSIC
IPOD CLASSIC
IPOD CLASSIC
IPOD CLASSIC

IPOD P68 GOOD
IPOD PHOTO P98A BEST
IPOD PHOTO P98A GOOD
IPOD SHUFFLE D98 BEST
IPOD SHUFFLE N98A BEST
IPOD SHUFFLE N98A BEST
IPOD SHUFFLE N98E BEST
IPOD SHUFFLE D98 BEST
IPOD SHUFFLE N98A BEST
IPOD SHUFFLE N98C BEST
IPOD SHUFFLE N98E BEST
IPOD TOUCH N72 BEST
IPOD TOUCH N72 BETTER
IPOD TOUCH N72 GOOD
IPOD MINI Q22 BEST
IPOD P68A GOOD
IPOD P97 BEST MAC
IPOD P97 BEST WIN
IPOD P97 BETTER MAC
IPOD P97 BETTER WIN
IPOD SHUFFLE D55 BEST
IPOD TOUCH N18 BEST
IPOD TOUCH N18 BETTER
IPOD MINI Q22 BEST
IPOD NANO N20 BEST
IPOD NANO N20 BETTER
IPOD PHOTO P98A BEST
IPOD SHUFFLE D98 BEST
IPOD SHUFFLE N98 BEST
IPOD SHUFFLE N98A BEST
IPOD SHUFFLE N98C BEST
IPOD SHUFFLE N98E BEST
IPOD SHUFFLE N98F BEST
IPOD SHUFFLE N98F BETTER
IPOD PHOTO P98 BEST
IPOD PHOTO P98 BETTER
IPOD PHOTO P98A BEST
IPOD PHOTO P98A BETTER
IPOD PHOTO P98A GOOD
IPOD PHOTO P98A GOOD

COMPARISON
u2 harrypotter
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

Murphy/Topel

Noll

Jul 2002
349
349
79
79
Sep 2007
Sep 2008
2009
2007
2008
2008
36
36
36
3 89
7 62
8 21
8 21
7 03
7 03
0 60
256
256
doesn't include Apple lossless
includes Apple lossless
includes Apple lossless
includes Apple lossless
doesn't include HE-AAC
includes audible
includes audible
includes audible
includes audible
includes audible
includes audible
doesn't include firewire
doesn't include firewire
doesn't include firewire
doesn't include firewire
doesn't include firewire
doesn't include firewire

Aug 2002
249
249
69
49
Feb 2008
Mar 2009
2010
2008
2009
2009
24
24
24
3 60
7 49
8 06
8 06
6 91
6 91
0 61
128
128
includes Apple lossless
doesn't include Apple lossless
doesn't include Apple lossless
doesn't include Apple lossless
includes HE-AAC
doesn't include audible
doesn't include audible
doesn't include audible
doesn't include audible
doesn't include audible
doesn't include audible
includes firewire
includes firewire
includes firewire
includes firewire
includes firewire
includes firewire

Appendix D
Additional Regression Results

Exhibit D1a (Amended)

HIGHLY CONFIDENTIAL

Exhibit D1a (Amended)

HIGHLY CONFIDENTIAL

Exhibit D1a (Amended)

HIGHLY CONFIDENTIAL

Exhibit D1a (Amended)

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

Exhibit D2a (Amended)

HIGHLY CONFIDENTIAL

Exhibit D2a (Amended)

HIGHLY CONFIDENTIAL

Exhibit D2a (Amended)

HIGHLY CONFIDENTIAL

Exhibit D2a (Amended)

HIGHLY CONFIDENTIAL

Exhibit D2b (Amended)

HIGHLY CONFIDENTIAL

Exhibit D2b (Amended)

HIGHLY CONFIDENTIAL

Exhibit D2b (Amended)

HIGHLY CONFIDENTIAL

Exhibit D2b (Amended)

HIGHLY CONFIDENTIAL

Exhibit D2c (Amended)

Note: Professor Noll's calculation of damages contains the following errors:
- Coefficients on iTunes 7 were calculated relative to the wrong but-for world.
- Standard errors were estimated incorrectly due to ignored clustering of residuals.
- Percentage Overchages were calculated using an incorrect formula.

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

Exhibit D3a (Amended)

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

Exhibit D3b (Amended)

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

Exhibit D4a1

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

Exhibit D4a2 (Amended)

HIGHLY CONFIDENTIAL

Exhibit D4a3 (Amended)

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

Exhibit D4b1

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

Exhibit D4b2 (Amended)

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

HIGHLY CONFIDENTIAL

Exhibit D4b3 (Amended)

HIGHLY CONFIDENTIAL

Exhibit D4c3 (Amended)
Effects of Correcting Professor Noll's Calculation of Damages

HIGHLY CONFIDENTIAL

1
2
3
4
5
6
7
8
9

Robert A. Mittelstaedt #60359
[e-mail]
Craig E. Stewart #129530
[e-mail]
David C. Kiernan #215335
[e-mail]
Amir Q. Amiri #271224
[e-mail]
JONES DAY
555 California Street, 26th Floor
San Francisco, CA 94104
Telephone:
[phone]
Facsimile:
[phone]
Attorneys for Defendant
APPLE INC.

10

UNITED STATES DISTRICT COURT

11

NORTHERN DISTRICT OF CALIFORNIA

12

OAKLAND DIVISION

13
14
15

THE APPLE iPOD iTUNES ANTI-TRUST
LITIGATION.

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

16

DECLARATION OF DR. ROBERT
TOPEL

17
18
19
20
21

I, Robert H. Topel, Ph.D submitted my Amended Expert Report in the above-captioned
matter on August 19, 2013, with supplemental exhibits served on November 8, 2013. The report
and corrected exhibits are true and correct and based on my own personal knowledge.

22

I declare the forgoing is true and correct to the best of my knowledge and belief.

23

Executed this 19 day of December, 2013 in Payson, Arizona.

24
25
26
Robert H. Topel, Ph.D.

27
28
-1-

EXHIBIT 9
[Filed Under Seal]

Robert Topel, Ph.D.

Confidential -Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation
Page 46

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

Q. All right. Is it your understanding that
the dataset that Professor Noll used in his
regression constituted a sample of Apple's
transaction data?
A. My understanding is that although there
could be missing transactions in the translation
that was given to Professor Noll, I don't believe
there are. So it would be all the transactions.
Q. So it's a population consisting of the
entire transactional database except for possibly a
few data points that had to be discarded? Is that
your view?
A. I don't know about "had to be discarded."
It is what he got.
Q. Okay. Let me rephrase then.
So is it your understanding that
Professor Noll relied on the complete Apple
transaction database?
A. As complete as he got it.
Q. So it's not a sample of the transaction
database?
A. You mean it's not a random draw from the
transaction database?
Q. Well, let's start with that. Is it a
random draw from the transaction database --

Page 47

1
A. No. As I understand it -2
(Reporter admonishes. )
3
THE WITNESS: I'm sorry.
4
I think it -- I think it's the universe of
5 what was handed over to me. He used every
6 observation.
7 BY MS. SWEENEY:
8
Q. Now, how do you interpret the standard
9 error when the data on which a regression is run are
10 not a sample from the underlying population but,
11 rather, constitute the whole population itself?
12
A. Your question is whether in the data as we
13 get it the experiment could be replicated.
14
So like if the transaction database -- if
15 all the transactions were running in, the data would
16 be what they are in that circumstance.
17
Q. What do you mean when you're referring to
18 "that circumstance"?
19
A. You might get a different set of outcomes
20 from running the -- running the process again. So
21 we've got all the data that came from one process of
22 selling Apples -- selling iPods.
23
Q. So is there a difference when you're using
24 a population versus a sample and you're determining
25 the statistical significance of standard errors?

Page 48

1

2
3
4

5

6
7
8
9

Page 49

A. I don't understand your question, but for
the purposes of this analysis, I don't think there
is.
Q. And why don' t you think there is a
difference?
A. Because we're asking whether people are
making independent pricing decisions, and we have
to -- we have to think about the statistical
properties of those decisions.

Page 46 .. 49
www.aptusCR.com

Robert Topel, Ph.D.

Confidential -Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation
Page 50

Page 51

Page 52

Page 53

Page 50 .. 53
www.aptusCR.com

Robert Topel, Ph.D.

Confidential - Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation
Page 54

Page 55

25
Page 56

Page 57

4
Q. So you say that Professor Noll
5 underestimated the true standard errors of -MR. MITTELSTAEDT: Where are you reading
6
7 from, Bonny?
8
MS. SWEENEY: I was looking for that. I'm
9 just going to paraphrase. Then we won't have to
10 find it. I'm not going to ask about that particular
11 sentence.
12 BY MS. SWEENEY:
13
Q. You say that Professor Noll underestimated
14 the true standard errors of his results by not
15 accounting for clustering. And you rely on a couple
16 sources. One of them is Angrist and Pischke? Is
17 that how you say that?
18
A. Pischke.
19
Q. Pischke.
20
And then you also rely on a monograph
21 published by the ABA antitrust section called
22 "Proving Antitrust Damages."
23
Are there any other works in economics
24 that you can identify that support your assertion?
25
A. Yes.

Page 54 .. 57
www.aptusCR.com

Robert Topel, Ph.D.

Confidential -Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation
Page 94

Page 95

1 apply a different clustering adjustment method to
2 serially correlate residuals versus cross-sectional
3 residuals?
4
A. It depends on whether one wants to assume
5 some parametric form for serial correlation or not.
6
Q. I'm not sure I can understand that answer,
7 so I'll just -8
A. It's kind of a statistical question, so
9 it's a statistical answer.
10
Q. So is a gross layperson way of describing
11 your answer to say maybe?
12
A. There are different ways of dealing with
13 serial correlation. And the clustering method,
14 where you cluster within some group, is a robust way
15 of doing that.
16
Others might decide, well, I have a
17 particular form in which I think the serial
18 correlation occurs. So they might parameterize the
19 covariance matrix of the residuals in that way.
20
(Reporter inquires.)
21
THE WITNESS: The covariance matrix of the
22 residuals in that way.
23 BY MS. SWEENEY:
24
Q. Okay. So it is possible to use different
25 methodologies for serially correlated residuals

1 versus cross-sectional residuals, correct?
2
A. Sure.
3
Q. You didn't do that here, right? You just
4 used the same methodology for both kinds of
5 residuals?
6
A. Ours are robust, yes. They're robust
7 estimates, given the definition of the groups.
Q. So that's why you didn't use an
B
9 alternative method; is that right?
10
A. Yes. As I said, this was sufficient to
11 demonstrate my point.
12
Q. Are you familiar with the shortcomings
13 that Professor Noll identified in the dataset that
14 he used?
15
A. I read them.
16
Q. Did you do anything when you conducted
17 your regression analyses to correct for those
18 shortcomings?
19
A. The specific ones that he referred to?
20
Q. Yes.
21
A. No. We took the data, as he had analyzed,
22 as a given, and we showed the flaws in his analysis.
23
Q. Did you ask Apple for any additional data?
24
A. Not that I recall. We got this data, I
25 think -- as I indicated, we went out and looked for
Page 97

Page 96

1
2
3
4
5
6
7
8
9
10
11

1
other attributes. That was collected from Apple's
2
websites and the websites of others.
Q. Did you ask Apple for any further
3
explanation of the observations of different aspects 4
5
of the dataset that you received?
6
A. Well, if you define "you" there as my
7
staff and me -- and my staff might have been trying
8
to understand something that they didn't
understand -- I can't tell you that, say, Ricardo
9
didn't have a conversation with somebody. But it
10
11
wasn't on my radar screen.
12
13
14

15
16

17
18
19
20
21
22
23

24
Q. So you just took the data as you received
25 it from Professor Noll and applied your clustering

24
25

adjustment, correct?
A. I believe that's a simplified way of
putting it. I think that my staff went to the
trouble of reconstructing the data in the way they
thought it should be reconstructed and comparing it
to Professor Noll, because I told them to go out and
reproduce what Professor Noll had done. And they
said the data came out pretty much the way
Professor Noll had it, so we're going to go with
that.
Q. Well, what do you mean by "reconstructed"?
A. We got the same files, as I recall, that
Professor Noll got. And so he ran his regressions
on a particular dataset. I think we were able to
reproduce what Professor Noll did.
Q. But going to this point of what
Professor Noll described as shortcomings in the
data, you didn't do any further adjustments to
account for the discounts or coupons, et cetera,
correct?
A. I don't think I had those information.
Q. Did you ask for that information?
A. I physically didn't ask for it.
Q. Did you -- I'm sorry.
A. I think I asked somebody at some point, is

Page 94 .. 97
www.aptusCR.com

Robert Topel, Ph.D.

Confidential -Attorneys' Eyes Only
The Apple iPod iTunes Anti-Trust Litigation
Page 106

1

2

3
4
5

6
7
8

9

10
11

12
13

14
15

Page 107

those things that's statistically independent, then
you're okay, statistically, assuming that the
observations are independent.
BY MS. SWEENEY:
Q. So is it your view that Roger's regression
analysis suffers from an omitted variable problem?
A. You're on another problem now. That's not
the reason for the clustering analysis that we did.
Q. What is the reason for the clustering
analysis that you did?
A. I think all the reasons I've given earlier
today: that Professor Noll is pretending to have
1
the data generate much more independent variation,
vastly more independent variation than they actually
do.

------------------------------------------,P~a_g_e_1~0~8~----------------------------------------,P~a-g_e_1~0~9~ 1

16

17
18
19

20
21
22
23
24
25

BY MS. SWEENEY:
Q. When you say that a missing predicate is
that people were in fact purchasing from the Real
Music Store for usage on iPods in significant
nUillbers, what do you mean when you say "significant
nUillbers"?
A. Well, the question is how many for lockin
is the incremental number -(Reporter inquires.)
THE WITNESS: The question for lockin is

Page 106 .. 109
www.aptusCR.com

EXHIBIT 10
[Filed Under Seal]

Kevin Murphy, Ph.D.

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

Page 134
1 for any particular reason, that much closer to the
2 regression line.

Page 135

Page 136

Page 137

1

10

.

23
Q. Now, we talked earlier about your
24 criticism that -- strike that.
25
We talked earlier about whether iPod

Page 134.. 137
www.aptusCR.com