Unites States of America v. Apple, Inc. et al
MOTION in Limine / APPLE INC.'S MOTION IN LIMINE TO EXCLUDE CERTAIN EXPERT TESTIMONY OF PROFESSOR RICHARD GILBERT. Document filed by Apple, Inc..(Snyder, Orin)
ATTORNEY WORK PRODUCT
UNITED STATES DISTRICT COURT
SOUTHERN DISTRICT OF NEW YORK
-------------------------------- - - - - -x
UNITED STATES OF AMERICA,
12 Civ. 2826(DLC)
APPLE INC., et al.,
------------------------------ --- - - - -x
------------------------------- - - - - --
THE STATE OF TEXAS,
THE STATE OF CONNECTICUT,et al.,
12 Civ. 03394(DLC)
PENGUIN GROUP(USA)INC., et al.,
APPLE INC.'S MOTION IN LIMINE TO EXCLUDE CERTAIN EXPERT TESTIMONY
OF PROFESSOR RICHARD GILBERT
Apple Inc. "Apple")moves to exclude from trial certain testimony of Professor Richard
J. Gilbert, fihe Department of Justice's "DOJ")economic expert. The DOJ should be precluded,
pursuant to Federal Rules of Evidence 403 and 702 and Daubert v. Merrell Dow
Pharmaceuticals, Inc., 509 U.S. 579, 589-91 (1993),from eliciting testimony from Professor
Gilbert that(1)constitutes impermissible fact finding masquerading as expert testimony and (2)
is based on a flawed and unreliable methodology.
First, the DOJ has improperly turned to its expert economist to provide a "narrative" of
the DOJ's interpretation ofthe facts in this case. That is properly the role ofthe fact-finder at
trial, Professor Gilbert has offered over 120 pages of"expert" reports, much of which merely
recount the "evidence" on which the DOJ's case rests—a carefully curated set of emails and
deposition testimony that do not reflect the entire record and have little relevance to the limited
economic opinions that are offered by Professor Gilbert. This is not the first time Professor
Gilbert has been retained to engage in impermissible fact finding. A court previously precluded
him from testifying on "questions of fact on which an economic expert's opinion is not helpful,"
and for "recit[ing] some assumed facts and argument ...[with] no analysis." Hynix
Semiconductor v. Rambus, Inc., No. 00 Civ. 20905, 2008 WL 73689, at *15(N.D. Cal. Jan. 5,
2008). Fact-finding, cloaked as an expert opinion, should not be permitted in Professor Gilbert's
Second,Professor Gilbert's conclusion that the Apple agency agreements caused a "net
reduction in volume" of e-book sales is premised on an unreliable empirical analysts that
purports to compare output in a two-week period prior to the implementation ofthe Apple
agency agreements and atwo-week period after the implementation ofthe Apple agency
agreements. Professor Gilbert acknowledges, as he must, that the result of his arbitrary and
result-oriented test (i.e., whether it shows a net increase or decrease in output)changes
depending on when you draw the two-week windows. As a result, it does not provide a reliable
basis for establishing that the agency agreements restricted output in the relevant market.
Professor Gilbert's conclusions related to a supposed post-agency reduction in the rate of growth
of e-book output are also defective and should be excluded.
Federal Rule of Evidence 702 permits witnesses qualified as experts to offer opinion
testimony only if:(1)"the testimony is based upon sufficient facts or data,"(2)"the testimony is
the product of reliable principles and methods," and(3)"the expert has reliably applied the
principles and methods to the facts ofthe case." Fed. R. Evid. 702. To be admissible under Rule
702, expert testimony must be(1) based on the special knowledge ofthe expert;(2) helpful to the
finder of fact; and (3)reliable. Daubert, 509 U.S. at 589-91.
It is not sufficient for a party to show that its witness has special expertise. Under Rule
702 and Daubert,"[t]he subject of an expert's testimony must be `scientific ...knowledge,
where "`scientific' implies a grounding in the methods and procedures of science," and
"`knowledge' connotes more than subjective belief or unsupported speculation." Daubert,509
U.S. at 589-90; United States v. Dukagjini, 326 F.3d 45, 54(2d Cir. 2002) "[E]xpert testimony
should be excluded if the witness is not actually applying expert methodology. Thus, while
expert witnesses can "guid[e] the trier offact through a complicated morass of obscure terms and
concepts," United States v. Duncan, 42 F.3d 97, 101 (2d Cir. 1994), expert testimony is properly
excludable where "[persons] ofcommon understanding are ...capable of comprehending the
primary facts and of drawing correct conclusions from them," United States v. Castillo, 924 F.2d
1227, 1232(2d Cir. 1991)(first alteration in original)(internal quotations and citations omitted).
Furthermore,"[e]xpert evidence can be both powerful and quite misleading because of
the difficulty in evaluating it." Daubert, 509 U.S. at 595 (internal quotations and citations
omitted). Therefore, courts "exercise more control over experts than over lay witnesses" under
Federal Rule of Evidence 403 and should exclude expert testimony where its probative value is
outweighed by possible prejudice. Id.
Professor Gilbert Should Not Be Permitted To Present A Factual Narrative
Of DOJ's Case
Professor Gilbert has opined that the role of an economist is to "collect facts as they are
known, and then .. ,put those facts together with economic reasoning to reach a conclusion."
"Snyder Decl. Exhibit A (Gilbert Deposition Transcript) at 145:13-17
(emphasis added). But Professor Gilberl's written submissions fall far short of even his standard.
Instead, much ofthe content in Professor Gilbert's expert reports simply summarize the DOJ's
the narrative ofthe facts. Any testimony by Professor Gilbert based on his report is untethered
from any economic conclusions that he offers and lacks grounding in Professor Gilbert's
economic expertise. Even if Professor Gilbert's summaries ofthe evidence were accurate and
complete(which they are not), testimony by Professor Gilbert that simply marshals the evidence
and impermissibly seeks to usurp the role of the fact finder and should be excluded.
Professor Gilbert's written submissions do little more than provide an expert's
imprimatur on the DOJ's factual allegations in this case, particularly in their(1)description of
conditions in the e-books market prior to Apple's entry, and (2) discussion ofthe negotiation and
implementation of the Apple agency agreements. By way of illustrative examples, Professor
Gilbert makes the following conclusory sfiatements offact, citing to a small handful of
• that Amazon "priced its ebooks very competitively"(Ex. PX-0821 ¶ 31)',
• that Amazon's "entry dramatically increased sales oftrade ebooks"(id.), and
• that"Amazon had the largest share of ebook sales due in part to its strategy of
charging low prices for ebooks"(Ex. PX-0821 ¶ 51)(without citation).
None ofthese statements is supported by economic analysis. Rather, they are "findings" offact,
proclaimed by an economist, based on record evidence. Similarly, Professor Gilbert
acknowledges that "Amazon was selling popular just-released and bestselling books at very low
(possibly negative) margins"(Ex. PX-0821 ¶ 55), but he does not conduct the economic analysis
associated with this fact. In contrast, Apple's expert, Dr. Michelle Burtis, measures Amazon's
margins or the extent to which Amazon was pricing e-books below their wholesale cost. See
Snyder Decl., Ex. A at 86:8-18.
Professor Gilbert also offers aone-sided factual characterization of"[p]ublisher actions
before the transition to agency" based on cherry-picked documents produced in discovery and
selective excerpts of deposition testimony provided to him. He draws the broad conclusion, for
example,from a review ofsix documents, but no deposition testimony discussing those
documents, that over the course of 2009, the defendant publishers met to discuss how they
"might work together to `defend against further price erosion' of their books and end Amazon's
ability to set low ebook prices"(Ex. PX-0821 ¶ 38). This inference regarding factual evidence is
tantamount to argument of counsel made through the auspices of an expert. Professor Gilbert
applies no economic analysis; he is just advocating his view of a certain subset ofthe evidence.
As this Court has noted,"[e]xpert testimony is properly excludable ...where `persons of
common understanding arecapable of comprehending the primary facts and of drawing
~ All citations to "PX-XXX"reference Plaintiffs' exhibit list
correct conclusions from them. U.S. ex rel. Anti-Discrimination Ctr. v. Westchester Cnry., No.
06 Civ. 2860(DLC),2009 WL 1110577, at * 1 (S.D.N.Y. Apr. 22, 2009)(second alteration in
original)(quoting Castillo, 924 F.2d at 1232);see also United States v. Mejia, 545 F.3d 179, 194
(2d Cir. 2008) "Testimony is properly characterized as `expert' only if it concerns matters that
the average juror is not capable of understanding on his or her own. In this bench trial, the
Court is perfectly capable ofevaluating all ofthe evidence relating to the e-books market and
actions taken by Publisher Defendants prior to Apple's entry. Professor Gilbert's factual
narrative, divorced from any economic analysis, provides no "specialized knowledge" that "will
help the trier of fact to understand" such evidence. Fed. R, Evid. 702(a). See also Taylor v,
Evans, No. 94 Civ. 8425, 1997 WL 154010, at *2(S.D.N.Y. Apr. 1, 1997)(excluding expert
opinion "presenting a narrative ofthe case which a lay juror is equally capable of constructing.
Other district courts have excluded Professor Gilbert's economic opinion for the same
impermissible fact finding. In Hynix Semiconductor v. Rambus, Inc., the court excluded
Professor Gilbert from testifying on "questions offact on which an economic expert's opinion is
not helpful," finding that his report "recites some assumed facts and argument but contains no
analysis." 2008 WL 73689, at *15.
To the extent Professor Gilbert engages in permissible fact finding, his testimony should
be similarly excluded in this District as outside the realm of proper expert testimony. See In re
Rezulin Prods. Liab, Litig., 309 F. Supp. 2d 531, 546-47(S.D.N.Y. 2004)(excluding expert
testimony as improper "because it describes `lay matters which a jury is capable of
understanding and deciding without the expert's help"' where the expert "merely repeated facts
or opinions stated by other potential witnesses or in documents produced in discovery" and
"drew simple inferences from documents produced in discovery."); The Pension Comm. ofthe
Univ. ofMontreal Pension Plan v. Banc ofAmerica Securities LLC,691 F. Supp. 2d 448,469
(S.D.N.Y. 2010)(excluding expert from testifying regarding nature of alleged fraud because
"Plaintiffs' fact witnesses and counsel will undoubtedly describe the alleged fraud at trial and
allowing Weiser to testify as an expert on that subject is likely to improperly bolster those
descriptions" and would be "unduly prejudicial"). Furthermore, his factual narrative constitutes
"a needless presentation of cumulative evidence" and a "waste of time" under Federal Rule of
Evidence 403. Dukagjini, 326 F.3d at 54.
Professor Gilbert's Conclusions Related to e-book Output In The Alleged
Relevant Market Should Be Excluded
In a market where output is rising, even when some prices rise, proof ofan output
restriction is necessary to sustain a claim of anticompetitive effects. Brooke Grp. Ltd. v. Brown
& Williamson Tobacco Corp,, 509 U.S. 209, 233(1993) "[s]upracompetitive pricing entails a
restriction in output"). In such a market, output "can only have been restricted in the sense that it
expanded at a slower rate than it would have" absent the alleged anticompetitive
conduct. td. Professor Gilbert purports to show this is true ofthe ebook market when he opines
that(1)the Apple agency agreements caused a "net reduction in volume" of e-book sales, based
on an analysis oftwo-week periods before and after agency, and (2)"ebook output grew more
slowly after the switch to agency than before." As the Supreme Court has emphasized, however,
"[s]uch a counterfactual proposition is difficult to prove in the best of circumstances." Brooke
Grp., 509 U.S. at 233. Dr. Gilbert's unreliable conclusions do not provide the required proof.
Professor Gilbert's Empirical Analysis Related to Output is
Professor Gilbert's opinion that the Apple agency agreements caused a "net reduction in
volume"(Ex. PX-0821 ¶ 75)of e-book sales is premised on a patently unreliable empirical
analysis and should be excluded. Professor Gilbert reaches this conclusion by comparing sales
of certain e-book titles over an artificial two-week period in March 2010(before agency) with
sales of those same titles in a second two-week period in April 2010 (after agency), Based on
this arbitrary time-frame, he purports to find that sales for four ofthe five Publisher Defendants
decreased, non-Defendant publishers' sales increased, and overall there was a net decrease in
sales. Professor Gilbert also relies on this analysis to support other sweeping conclusions,
including that "the higher prices from agency reduced total ebook sales"(Ex. PX-0821 ¶ 185).
His empirical analysis, however, has fatal empirical and conceptual flaws.
First, Professor Gilbert offers no reliable, generally-accepted methodology to support his
arbitrary assumptions and calculations. The two-week before-and-after snapshot of the market
he "studies" is an irrational, unfounded, and litigation-driven device that cannot be presented to
the fact-finder as a reliable analysis. The unreliability ofProfessor Gilbert's approach can be
demonstrated by making small adjustments to the placement ofthe 2-week "pre" and "post"
agency windows that he compared in his analysis. Comparing the same "before" period selected
by Professor Gilbert, to an "after" period that is two weeks later than the one selected by
Professor Gilbert establishes that net e-book sales were higher in the post-agency period. This
result —which applies the same methodology used by Professor Gilbert during a slightly different
time period — is contrary to Professor Gilbert's conclusion that output declined as a result of
agency. Professor Gilbert acknowledged this shortcoming in his methodology. He does not
dispute that it is possible to find an increase in total output depending on where you draw the
arbitrary two-week windows. Snyder Decl., Ex. A at 347:15-17. Professor Gilbert's irrational
two-week before-and-after opinion should be excluded as unreliable on this basis alone.
Second, another defect in Professor Gilbert's analysis is that it is taken over so short a
time period that it cannot be used to argue credibly that the agency agreements had an
anticompetitive effect, especially in light ofthe overwhelming evidence that the volume of ebook sales continued to grow at very high rates following the implementation ofthe agency
agreements. Professor Gilbert's analysis and conclusions regarding reduced output in the
alleged relevant market is based on 0.4% ofthe sales of trade e-books during the period April
2010 through March 2012.[See Burtis Trial Affidavit]
To prove "harm to the whole market," Capital Imaging Assocs., P.C. v. Mohawk Valley
Medical Assocs., 996 F.2d 537, 543(2d Cir. 1993), expert testimony in antitrust cases must
"incorporate all aspects ofthe economic reality of the ...market" or else such "deficiencies in
the foundation of the opinion [and] the expert's resulting conclusions" are inadmissible as "mere
speculation." Concord Boat Corp. v. Brunswick Corp., 207 F.3d 1039, l OS7(8th Cir. 2000); see
also Blue Dane Simmental Corp. v. American Simmental Assn, 178 F.3d 1035, 1040 (8th Cir.
1999)(affirming district court's exclusion of expert testimony where the expert "attributed the
entire difference in market price" of a particular cattle breed to the introduction of a certain
subset of cattle "despite the fact that these animals made up a tiny fraction ofthe market").
Professor Gilbert's cropped snapshot of less than 1% ofthe entire alleged relevant market does
not encompass "all aspects ofthe economic reality" ofthe e-books market. It cannot serve as a
helpful basis to draw conclusions about the alleged anticompetitive effects ofthe alleged
conspiracy on the alleged relevant market.
Third, Professor Gilbert fails to account for the fact that sales of a particular title will
typically decline over time, a phenomenon that Professor Ashenfelter acknowledges. Snyder
Decl., Exhibit B (Ashenfelter Deposition Transcript) at 162:19-164:7. Performing Professor
Gilbert's same comparison for different two week periods prior to the agency agreement shows
that sales of a title typically decline over time.2 Similarly, Penguin's sales, which were not
included in Professor Gilbert's analysis because Penguin had not implemented an agency
agreement with Amazon during the time period studied, also experienced an overall decline in
sales as well. It is therefore clear that the age ofthe titles affect sales volume over time and that
Professor Gilbert's analysis, which ignores this fact, cannot be relied upon to draw conclusions
about the general impact of agency on e-book sales.
Professor Gilbert's Opinion Regarding Rate of Growth in Output is Not
Professor Gilbert acknowledges the rapid growth of e-book sales in the relevant market
before and after agency, but opines that "ebook output grew more slowly after the switch to
agency than before." The pre-agency growth rate was characteristic of a young, dynamic
market, based on a starting point of relatively few e-book sales. It is entirely unremarkable that
such agrowth-rate would slow and, as Professor Gilbert concedes, a 358% growth rate (as
calculated by him in the year prior to agency) was not sustainable. Snyder Decl., Ex. A at 361:524. Professor Gilbert admits that he cannot conclude that the slowing growth rate is due to the
adoption ofthe agency contracts. Id. at 358:21-24.
Given the rapid growth of e-book sales before and after agency, output "can only have
been restricted in the sense that it expanded at a slower rate than it would have" absent the
agency model. Brooke Grp., 509 U.S. at 233 . Professor Gilbert offers no opinion that the
alleged relevant market would have expanded at a greater rate but-for the Apple agency
agreements. Therefore, Professor Gilbert's testimony regarding the slowing rate of growth in
For example if you compare sales over the two weeks ending January 30, 2010, and February 6,2010, with
sales over the two weeks ending March 6,2010, and March 13, 2010,there is a net decrease in output.
output is unhelpful and, at best, will mislead the fact finder. See Fed. R. Evid. 403. It should be
excluded from presentation at trial.
Apple respectfully requests that this Court grant its motion to preclude Professor Gilbert
from offering the opinions discussed herein.
Dated: New York, New York
Lisa H. Rubin
Gibson, Dunn & Crutcher, LLP
200 Park Avenue, 47th Floor
New York, NY 10166
Daniel S, Floyd (Pro Hac Vice)
Daniel G. Swanson(Pro Hac Vice)
Gibson, Dunn & Crutcher, LLP
333 South Grand Avenue
Los Angeles, CA 90071
Cynthia Richman(Pro Hac Vice Admission
Gibson, Dunn & Crutcher, LLP
1050 Connecticut Avenue, N.W.
Washington, DC 20036
Howard E. Heiss
Edward N. Moss
O'Melveny &Meyers LLP
Times Square Tower
7 Times Square
New York, NY 10036
On behalfofDefendant Apple, Inc.