United States Court of Appeals
         FOR THE DISTRICT OF COLUMBIA CIRCUIT



Argued December 6, 2018            Decided February 26, 2019

                         No. 18-5214

                UNITED STATES OF AMERICA,
                       APPELLANT

                              v.

                     AT&T, INC., ET AL.,
                        APPELLEES


        Appeal from the United States District Court
                for the District of Columbia
                    (No. 1:17-cv-02511)


    Michael F. Murray, Deputy Assistant Attorney General,
U.S. Department of Justice, argued the cause for appellant.
With him on the briefs were Kristen C. Limarzi, Robert B.
Nicholson, Adam D. Chandler, Patrick M. Kuhlmann, Mary
Helen Wimberly, and Daniel E. Haar, Attorneys.

     Eric F. Citron argued the cause for amici curiae 27
Antitrust Scholars in support of neither party. With him on the
brief was Mary Jean Moltenbrey.

    Laurence M. Sandell was on the brief for amicus curiae
Cinémoi North America in support of appellant United States
of America.
                               2

    Jeffrey A. Lamken was on the brief for amici curiae
Professor William P. Rogerson, et al. in support of appellant.

     Jonathan W. Cuneo, Joel Davidow, and Gene Kimmelman
were on the brief for amici curiae American Antitrust Institute,
et al. in support of appellant.

    Jonathan E. Taylor was on the brief for amicus curiae
Open Markets Institute in support of plaintiff-appellant and
vacatur.

    Peter D. Keisler argued the cause for appellees. With him
on the brief were Joseph R. Guerra, Richard D. Klingler,
Jonathan E. Neuchterlein, C. Frederick Beckner III, Kathleen
Moriarty Mueller, William R. Drexel, Daniel M. Petrocelli, M.
Randall Oppenheimer, Jonathan D. Hacker, Katrina M.
Robson, and David L. Lawson. Aaron M. Panner entered an
appearance.

     Andrew J. Pincus argued the cause for amici curiae 37
Economists, Antitrust Scholars, and Former Government
Antitrust Officials in support of appellees. With him on the
brief were Mark W. Ryan and Michael B. Wimberly.

     Brad D. Schimel, Attorney General, Office of the Attorney
General for the State of Wisconsin, Misha Tseytlin, Solicitor
General, Jeff Landry, Attorney General, Office of the Attorney
General for the State of Louisiana, Elizabeth Baker Murrill,
Solicitor General, Hector Balderas, Attorney General, Office
of the Attorney General for the State of New Mexico, Tania
Maestas, Chief Deputy Attorney General, Mike Hunter,
Attorney General, Office of the Attorney General for the State
of Oklahoma, Mithun Mansinghani, Solicitor General, Steve
Marshall, Attorney General, Office of the Attorney General for
                               3
the State of Alabama, Robert Tambling, Assistant Attorney
General, Christopher M. Carr, Attorney General, Office of the
Attorney General for the State of Georgia, Andrew A. Pinson,
Solicitor General, Andy Beshear, Attorney General, Office of
the Attorney General for the Commonwealth of Kentucky,
Sean D. Reyes, Attorney General, Office of the Attorney
General for the State of Utah, Tyler R. Green, Solicitor
General, Peter F. Kilmartin, Attorney General, Office of the
Attorney General for the State of Rhode Island, Michael W.
Field, Assistant Attorney General, Alan Wilson, Attorney
General, Office of the Attorney General for the State of South
Carolina, and James Emory Smith, Jr., Deputy Solicitor
General, were on the bipartisan brief for amici curiae the States
of Wisconsin, et al. in support of defendants-appellees.

     Donald B. Verrilli, Jr., Justin P. Raphael, Peter C.
Tolsdorf, Steven P. Lehotsky, and Daryl Joseffer were on the
brief for amici curiae The Chamber of Commerce of the United
States of America, et al. in support of defendants-appellees.

    Thomas M. Johnson, Jr., General Counsel, Federal
Communications Commission, David M. Gossett, Deputy
General Counsel, Richard K. Welch, Deputy Associate General
Counsel, and James M. Carr, Counsel, were on the brief for
amicus curiae Federal Communications Commission in
support of neither party.

    Bruce D. Brown, Katie Townsend, and Gabriel Rottman
were on the brief for amicus curiae The Reporters Committee
For Freedom of the Press in support of neither party.

   Before: ROGERS and WILKINS, Circuit Judges, and
SENTELLE, Senior Circuit Judge.

    Opinion for the court filed by Circuit Judge ROGERS.
                               4

     ROGERS, Circuit Judge: On October 22, 2016, AT&T Inc.
announced a proposed merger with Time Warner Inc. The
government sued to enjoin this vertical merger under Section 7
of the Clayton Act, 15 U.S.C. § 18, and now appeals the denial
of its request for a permanent injunction. United States v.
AT&T Inc., 

, 254 (D.D.C. 2018). Although
it pursued three theories of antitrust violation in the district
court, the government on appeal challenges only the district
court’s findings on its increased leverage theory whereby costs
for Turner Broadcasting System’s content would increase after
the merger, principally through threats of long-term
“blackouts” during affiliate negotiations.

     At trial, the government presented expert opinion on the
likely anticompetitive effects of the proposed merger on the
video programming and distribution industry as forecast by
economic principles and a quantitative model. It also presented
statements by the defendants in administrative proceedings
about the anticompetitive effects of a proposed vertical merger
in the industry seven years earlier. The defendants responded
with an expert’s analysis of real-world data for prior vertical
mergers in the industry that showed “no statistically significant
effect on content prices.” The government offered no
comparable analysis of data and its expert opinion and
modeling predicting such increases failed to take into account
Turner Broadcasting System’s post-litigation irrevocable
offers of no-blackout arbitration agreements, which a
government expert acknowledged would require a new model.
Evidence also indicated that the industry had become dynamic
in recent years with the emergence, for example, of Netflix and
Hulu. In this evidentiary context, the government’s objections
that the district court misunderstood and misapplied economic
principles and clearly erred in rejecting the quantitative model
are unpersuasive. Accordingly, we affirm.
                                5
                                I.

      Section 7 of the Clayton Act prohibits mergers “where in
any line of commerce or in any activity affecting commerce in
any section of the country, the effect of such acquisition may
be substantially to lessen competition.” 15 U.S.C. § 18
(emphasis added). Congress acted out of concern with
“probabilities, not certainties” inasmuch as “statutes existed
[only] for dealing with clear-cut menaces to competition . . . .
Mergers with a probable anticompetitive effect were to be
proscribed by [the Clayton Act].” Brown Shoe Co. v. United
States, 

, 323 (1962). It left to the courts the
difficult task of assessing probabilities in the commercial
marketplace in the interest of “halting ‘incipient monopolies
and trade restraints outside the scope of the Sherman Act,’”
Rothery Storage & Van Co. v. Atlas Van Lines, Inc., 

, 220 (D.C. Cir. 1986) (quoting Brown 

n.32). Therefore, Section 7 “applies a much more stringent
test than does the rule-of-reason analysis under section 1 of the
Sherman Act.” 

7 requires more than a
“mere possibility” of competitive harm, it does not require
proof of certain harm. Brown 

n.39.
Instead, the government must show that the proposed merger is
likely to substantially lessen competition, which encompasses
a concept of “reasonable probability.” 

government nor the defendants challenge
application of the burden-shifting framework in United States
v. Baker Hughes, 

, 982–83 (D.C. Cir. 1990), for
horizontal mergers that the district court applied to consider the
effect of the proposed vertical merger of AT&T and Time
Warner on competition.           Under this framework, the
government must first establish a prima facie case that the
merger is likely to substantially lessen competition in the
relevant market. United States v. Anthem, 

, 349
                               6
(D.C. Cir. 2017). But unlike horizontal mergers, the
government cannot use a short cut to establish a presumption
of anticompetitive effect through statistics about the change in
market concentration, because vertical mergers produce no
immediate change in the relevant market share. See Dept. of
Justice & Fed. Trade Comm’n, Non-Horizontal Merger
Guidelines § 4.0 (June 14, 1984) (“1984 Non-Horizontal
Merger Guidelines”). Instead, the government must make a
“fact-specific” showing that the proposed merger is “likely to
be anticompetitive.” Joint Statement on the Burden of Proof at
Trial at 3–4. Once the prima facie case is established, the
burden shifts to the defendant to present evidence that the
prima facie case “inaccurately predicts the relevant
transaction’s probable effect on future competition,” 

(quoting Baker 

), or to
“sufficiently discredit” the evidence underlying the prima facie
case, 

rebuttal, “the burden of producing
additional evidence of anticompetitive effects shifts to the
government, and merges with the ultimate burden of
persuasion, which remains with the government at all times.”
Baker 

.

     The relevant market definition is also undisputed by the
government and the defendants. (For ease of reference, we
refer hereinafter to defendants AT&T Inc., Direct TV Group
Holdings, LLC, and Time Warner Inc. as “AT&T.”) The
district court accepted the government’s proposal that the
product market is the market for multichannel video
distribution.    Although this market definition excludes
distributors of only on-demand content, such as Netflix and
Hulu, the district court considered the impact of the increasing
presence of these distributors on the multichannel video
programming and distribution industry. AT&T, 

at 196–97.         The district court also accepted the
government’s proposed geographic market, which included
                                7
over 1,100 local multichannel video distribution markets. 

The government did not rely on any particular market
for enjoining the proposed merger; one of its experts
aggregated the alleged harms in the local markets to derive a
total measure of nationwide economic harm. See Proposed
Findings of Fact of the United States 13 (May 8, 2018).

     As the government has presented its challenges to the
district court’s denial of a permanent injunction, the question
for this court is whether the district court’s factual findings are
clearly erroneous. FED. R. CIV. P. 52(a); see FTC v. H.J. Heinz
Co., 

, 713 (D.C. Cir. 2001). This is a deferential
standard. Zenith Radio Corp. v. Hazeltine Research, Inc., 

, 123 (1969). “A finding is ‘clearly erroneous’ when
although there is evidence to support it, the reviewing court on
the entire evidence is left with the definite and firm conviction
that a mistake has been committed.” United States v. Gypsum
Co., 

, 395 (1948). Findings that are plausible in
light of the entire record are not clearly erroneous, Anderson v.
City of Bessemer City, N.C., 

, 575, 577 (1985), so
“[w]here there are two permissible views of the evidence, the
factfinder’s choice between them cannot be clearly erroneous,”

(citing United States v. Yellow Cab Co., 

, 342 (1949)); see also Cooper v. Harris, 

,
1465 (2017). A finding may be clearly erroneous when it is
illogical or implausible, 

, rests on
internally inconsistent reasoning, 

, or
contains errors of economic logic, FTC v. Advocate Health
Care Network, 

, 464 (7th Cir. 2016).

     The government contends that it has made the requisite
showing of error because the district court’s conclusion it had
failed to meet its burden of proof “rests on two fundamental
errors: the district court discarded the economics of bargaining,
and the district court failed to apply the foundational principle
                               8
of corporate-wide profit maximization.” Appellant Br. 29, 37–
38. Further, the government contends that the district court
used internally inconsistent logic when evaluating industry
evidence and clearly erred in rejecting its expert’s quantitative
model of harm.

     In Part II, we provide an overview of the video
programming and distribution industry. Then, as relevant to
the issues on appeal, we summarize the evidence before the
district court and its findings. In Part III, we address the
government’s challenges to the district court’s findings.

                                   II.

                                A.
     The video programming and distribution industry
traditionally operates in a three-stage chain of production.
Studios or networks create content. Then, programmers
package content into networks and license those networks to
video distributors. Finally, distributors sell bundles of
networks to subscribers. For example, a studio may create a
television show and sell it to Turner Broadcasting System
(“Turner Broadcasting”), a programmer, which would package
that television show into one of its networks, such as CNN or
TNT. Turner Broadcasting would then license its networks to
distributors, such as DirecTV or Comcast.

     Programmers license their content to distributors through
affiliate agreements, and distributors pay “affiliate fees” to
programmers. Programmers and distributors engage in what
are oftentimes referred to as “affiliate negotiations,” which,
according to evidence before the district court, can be lengthy
and complicated. If a programmer and a distributor fail to
reach an agreement, then the distributor will lose the rights to
display the programmer’s content to its customers. This
                               9
situation, known as a “blackout” or “going dark,” is generally
costly for both the programmer, which loses affiliate fee
revenues, and the distributor, which risks losing subscribers.
Therefore, blackouts rarely occur, and long-term blackouts are
especially rare. The evidence indicated, however, that
programmers and distributors often threaten blackouts as a
negotiating tactic, and both may perform “go dark” analyses to
estimate the potential impact of a blackout in preparation for
negotiations.

     The evidence before the district court also showed that the
industry has been changing in recent years. Multichannel video
programming distributors (“MVPDs”) offer live television
content as well as libraries of licensed content “on demand” to
subscribers.     So-called “traditional” MVPDs distribute
channels to subscribers on cable or by satellite. Recently,
“virtual” MVPDs have also emerged. They distribute live
videos and on-demand videos to subscribers over the internet
and compete with traditional MVPDs for subscribers. Virtual
MVPDs, such as DirecTV Now and YouTube TV, have been
gaining market share, the evidence showed, because they are
easy to use and low-cost, often because they offer subscribers
smaller packages of channels, known as “skinny bundles.”

     In addition, subscription video on demand services
(“SVODs”) have also emerged on the market. SVODs, such as
Netflix, do not offer live video content but have large libraries
of content that a viewer may access on demand. SVODs also
offer low-cost subscription plans and have been gaining market
share recently. Increasingly, cable customers are “cutting the
cord” and terminating MVPD service altogether. Often these
customers do not exit the entertainment field altogether, but
instead switch to SVODs for entertainment service.
                              10
     Leading SVODs are vertically integrated, which means
they create content and also distribute it. Traditional MVPDs
typically are not vertically integrated with programmers. In
2009, however, Comcast Corporation (“Comcast”) (a
distributor and the largest cable company in the United States)
announced a $30 billion merger with NBC Universal, Inc.
(“NBCU”) (a content creator and programmer), whereby it
would control popular video programming that included the
NBC broadcast network and the cable networks of NBC
Universal, Inc. The government sued to permanently enjoin
the merger under Section 7, alleging that Comcast’s “majority
control of highly valued video programming . . . would prevent
rival video-distribution companies from competing against the
post-merger entity.” United States v. Comcast, 

145, 147 (D.D.C. 2011). The district court, with the
defendants’ agreement and at the government’s urging,
allowed the merger to proceed subject to certain remedies for
the alleged anticompetitive conduct post-merger, including
remedies ordered in a related proceeding before the Federal
Communications Commission (“FCC”). 

in
the Comcast-NBCU merger was an agreement by the
defendants to submit, at a distributor’s option, to “baseball
style” arbitration — in which each side makes a final offer and
the arbitrator chooses between them — if parties did not reach
a renewal agreement. During the arbitration, the distributor
would retain access to NBC content, thereby mitigating
concerns that Comcast-NBCU may withhold NBC
programming during negotiations in order to benefit Comcast’s
distribution subscriptions. Comcast-NBCU currently operates
as a “vertically integrated” programmer and distributor.

     Now the government has again sued to halt a proposed
vertical merger of a programmer and a distributor in the same
industry. On October 22, 2016, AT&T Inc. announced its plan
to acquire Time Warner Inc. (“Time Warner”) as part of a $108
                               11
billion transaction. AT&T Inc. is a distribution company with
two traditional MVPD products: DirecTV and U-verse.
DirecTV transmits programming over satellite, while U-verse
transmits programming over cable. Time Warner, by contrast,
is a content creator and programmer and has three units:
Warner Bros., Turner Broadcasting, and Home Box Office
Programming (“HBO”).        Warner Bros. creates movies,
television shows, and other video programs.            Turner
Broadcasting packages content into various networks, such as
TNT, TBS, and CNN, and licenses its networks to third-party
MVPDs. HBO is a “premium” network that provides on-
demand content to subscribers either directly through HBO
Now or through licenses with third-party distributors. The
merged firm would operate both AT&T MVPDs (DirecTV and
U-verse) and Turner Broadcasting networks (which license to
other MVPDs). The government alleged that “the newly
combined firm likely would . . . use its control of Time
Warner’s popular programming as a weapon to harm
competition.” Compl. 2.

      A week after the government filed suit to stop the proposed
merger, Turner Broadcasting sent letters to approximately
1,000 distributors “irrevocably offering” to engage in “baseball
style” arbitration at any time within a seven-year period,
subject to certain conditions not relevant here. According to
President of Turner Content Distribution Richard Warren, the
offer of arbitration agreements was designed to “address the
government’s concern that as a result of being . . . commonly
owned by AT&T, [Turner Broadcasting] would have an
incentive to drive prices higher and go dark with [its]
affiliates,” Tr. 1182 (April 3, 2018). In the event of a failure
to agree on renewal terms, Turner Broadcasting agreed that the
distributor would have the right to continue carrying Turner
networks pending arbitration, subject to the same terms and
conditions in the distributor’s existing contract.
                                 12

                                   B.
     The government’s increased leverage theory is that “by
combining Time Warner’s programming and DirecTV’s
distribution, the merger would give Time Warner increased
bargaining leverage in negotiations with rival distributors,
leading to higher, supracompetitive prices for millions of
consumers.” Appellant Br. 33. Under this theory, Turner
Broadcasting’s bargaining position in affiliate negotiations will
change after the merger due to its relationship with AT&T
because the cost of a blackout will be lower. Prior to the
merger, if Turner Broadcasting failed to reach a deal with a
distributor and engaged in a long-term blackout, then it would
lose affiliate fees and advertising revenues. After the merger,
some costs of a blackout would be offset because some
customers would leave the rival distributor due to Turner
Broadcasting’s blackout and a portion of those customers
would switch to AT&T distributor services. The merged
AT&T-Turner Broadcasting entity would earn a profit margin
on these new customers. Because Turner Broadcasting would
make a profit from switched customers, the cost of a long-term
blackout would decrease after the merger and thereby give it
increased bargaining leverage during affiliate negotiations with
rival distributors sufficient to enable it to secure higher affiliate
fees from distributors, which would result in higher prices for
consumers.

     To support this theory of competitive harm, the
government presented evidence purporting to show the real-
world effect of the proposed merger. Specifically, it introduced
statements in prior FCC filings by AT&T and DirecTV that
vertical integration provides an incentive to increase prices and
poses a threat to competition. Various internal documents of
the defendants were to the same effect. Third-party
                                13
competitors, such as cable distributors, testified that the merger
would increase Turner Broadcasting’s bargaining leverage.

     The government also presented the expert opinion of
Professor Carl Shapiro on the likely anticompetitive effect of
the proposed merger. He opined, based on the economic theory
of bargaining — here, the Nash bargaining theory — that
Turner Broadcasting’s bargaining leverage would increase
after the merger because the cost of a long-term blackout would
decrease. His quantitative model predicted net price increases
to consumers. Specifically, his model predicted increases in
fees paid by rival distributors for Turner Broadcasting content
and cost savings for AT&T through elimination of double
marginalization (“EDM”).         The fee increases for rival
distributors were based on the expected benefit to AT&T of a
Turner Broadcasting blackout after the merger. Professor
Shapiro determined the extent to which rival distributors and
AT&T would pass on their respective cost increases and cost
decreases to consumers. His model predicted: (1) an annual fee
increase of $587 million for rival distributors to license Turner
Broadcasting content, and cost savings of $352 million for
AT&T; and (2) an annual net increase of $286 million in costs
passed on to consumers in 2016, with increases in future years.

     AT&T responded by pointing to testimony of executives’
past experience in affiliate negotiations, and presenting
testimony by its experts critiquing Professor Shapiro’s opinion
and model. It purported to show through its own experts that
the government’s prima facie case inaccurately predicted the
proposed merger’s probable effect on competition. Professor
Dennis Carlton’s econometric analysis (also known as a
regression analysis, Tr. 2473 (April 12, 2018)), showed that
prior instances of vertical integration in the MVPD market had
not had a “statistically significant effect on content prices,” 

pointing to data on the Comcast-NBCU merger in
                              14
2011 as well as prior vertical integration between News Corp.-
DirecTV and Time Warner Cable-Time Warner Inc., which
split in 2008 and 2009, respectively. Professor Carlton and
Professor Peter Rossi critiqued the “inputs” used by Professor
Shapiro in his quantitative model, opining for instance that
values he used for subscriber loss rate and diversion rate were
not calculated through reliable methods. Professor Carlton also
opined that Professor Shapiro’s quantitative model
overestimated how quickly harm would occur because it failed
to consider existing long-term contracts.

     Professor Shapiro, in turn, critiqued Professor Carlton’s
econometric analysis as comparing different types of vertical
mergers. Regarding the “inputs” to his quantitative model,
Professor Shapiro conceded that he was unaware the subscriber
loss rate percentage he used (from a consultant report for
Charter Communications, Inc.) had been changed after the
report was presented to Charter executives.          He also
acknowledged that he had not considered the effects of the
arbitration agreements offered by Turner Broadcasting and that
to do so would require preparation of a new model. Tr. 2208,
2325 (Apr. 11, 2018).

     The district court acknowledged the uncertainty regarding
the measure of proof for the government’s burden because
Section 7 does not require proof of certain harm. AT&T, 

at 189–90 n.16. The government and AT&T had
used various phrases to describe the government’s burden,
including that it must show an “appreciable danger” of
competitive harm, or that it must show that harm is “likely” or
“reasonably probable.” 

court concluded that it
need not articulate the differences between these phrases
because “even assuming the ‘reasonable probability’ or
‘appreciable danger’ formulations govern here . . . [its]
conclusions regarding the [g]overnment’s failure of proof
                               15
would remain unchanged.” 

the lack
of precedent and the complexity in establishing the correct
approach in a Section 7 challenge to a proposed vertical
merger, the district court viewed the outcome of the litigation
to “turn[] on whether, notwithstanding the proposed merger’s
conceded procompetitive effects, the [g]overnment has met its
burden of establishing, through ‘case-specific evidence,’ that
the merger of AT&T and Time Warner, at this time and in this
remarkably dynamic industry, is likely to substantially lessen
competition in the manner it predicts.” 

(quoting
Proposed Conclusions of Law of the United States ¶ 25).

     Several amici urge this court to speak definitively on the
proper legal standard for evaluating vertical mergers. See
Amicus Curiae 27 Antitrust Scholars et al. Br. 10–16; Amicus
Curiae Chamber of Commerce et al. Br. 5–8; Amicus Curiae
Open Markets Institute Br. 16–20. There is a dearth of modern
judicial precedent on vertical mergers and a multiplicity of
contemporary viewpoints about how they might optimally be
adjudicated and enforced. See, e.g., Amicus Curiae 27
Antitrust Scholars et al. Br. 6–16; Amicus Curiae Chamber of
Commerce et al. Br. 17–24. The government’s guidelines for
non-horizontal mergers were last updated in 1984, over three
decades ago. See 1984 Non-Horizontal Merger Guidelines.
But there is no need to opine on the proper legal standards for
evaluating vertical mergers because, on appeal, neither party
challenges the legal standards the district court applied, and no
error is apparent in the district court’s choices, see Amicus
Curiae Open Markets Institute Br. 18–19 (citing cases). See
generally Narragansett Indian Tribe v. Nat. Indian Gaming
Com’n, 

, 1338 (D.C. Cir. 1998); Michel v.
Anderson, 

, 625 (D.C. Cir. 1994).

     The district court found that the government had “failed to
clear the first hurdle of showing that the proposed merger is
                               16
likely to increase Turner [Broadcasting]’s bargaining leverage
in affiliate negotiations.” AT&T, 

at 199.
Although acknowledging, as Professor Shapiro had opined,
that the Nash bargaining theory could apply in the context of
affiliate fee negotiations, the district court found more
probative the real-world evidence offered by AT&T than that
offered by the government. The econometric analysis of
AT&T’s expert had examined real-world data from prior
instances of vertical integration in the video programming and
distribution industry and concluded that “the bulk of the results
show no significant results at all, but many do show a decrease
in content prices.” 

(quoting Prof. Carlton, Tr. 2477
(April 12, 2018)); see 

218. The district court also
credited the testimony of several industry executives — e.g.,
Madison Bond, lead negotiator for NBCU, and Coleman
Breland and Richard Warren, lead negotiators for Turner
Broadcasting — that vertical integration had not affected their
affiliate negotiations in the past. By contrast, the testimony
from third-party competitors that the merger would increase
Turner Broadcasting’s bargaining leverage was, the district
court found, “speculative, based on unproven assumptions, or
unsupported.” 

Although Professor Shapiro’s
opinion was that the Nash bargaining theory predicted an
increase in Turner Broadcasting’s post-merger bargaining
leverage, leading to an increase in affiliate fees, the district
court found, in view of the industry’s dynamism in recent
years, that Professor Shapiro’s opinion (by contrast with
Professor Carlton’s) had “not been supported by sufficient real-
world evidence.” 

     Second, the district court found that Professor Shapiro’s
quantitative model, which estimated the proposed merger
would result in future increases in consumer prices, lacked
sufficient reliability and factual credibility to generate
probative predictions of future competitive harm. Relying on
                                17
critiques by Professor Carlton and Professor Rossi, the district
court found errors in the model “inputs,” for example, the value
used for subscriber loss rate was not calculated through a
reliable method. Neither the model nor Professor Shapiro’s
opinion accounted for the effect of the irrevocably-offered
arbitration agreements, which the district court stated would
have “real world effects” on negotiations and characterized “as
extra icing on a cake already frosted,” 

n.51, i.e.,
another reason the government had not met its first-level
burden of proof.

     The district court therefore concluded that the government
failed to present persuasive evidence that Turner
Broadcasting’s bargaining leverage would “materially
increase” as a result of the merger, 

or that the merger
would lead to “any raised costs” for rival distributors or
consumers, 

(emphasis in original). It therefore did
not address the balancing analysis offered by Professor
Shapiro’s quantitative model, nor the question whether any
increased costs would result in a substantial lessening of
competition.

                               III.

     On appeal, the government contends that the district court
court (1) misapplied economic principles, (2) used internally
inconsistent logic when evaluating industry evidence, and
(3) clearly erred in rejecting Professor Shapiro’s quantitative
model. Undoubtedly the district court made some problematic
statements, which the government identifies and this court
cannot ignore. And in the probabilistic Section 7 world,
uncertainty exists about the future real-world impact of the
proposed merger on Turner Broadcasting’s post-merger
leverage. See Brown 

. At this point,
however, the issue is whether the district court clearly erred in
                               18
finding that the government failed to clear the first hurdle in
meeting its burden of showing that the proposed merger is
likely to increase Turner Broadcasting’s bargaining leverage.

     (1) Application of economic principles. The government
contends that in evaluating the evidence in support of its
increased leverage theory, the district court erroneously
discarded or otherwise misapplied two economic principles —
the Nash bargaining theory and corporate-wide profit
maximization.

          (a) Nash bargaining theory. The Nash bargaining
theory is used to analyze two-party bargaining situations,
specifically where both parties are ultimately better off by
reaching an agreement. John F. Nash, Jr., The Bargaining
Problem, 18 Econometrica 155 (1950). The theory posits that
an important factor affecting the ultimate agreement is each
party’s relative loss in the event the parties fail to agree: when
a party would have a greater loss from failing to reach an
agreement, the other party has increased bargaining leverage.
Tr. 2193–94 (Shapiro, April 11, 2018). In other words, the
relative loss for each party affects bargaining leverage and
when a party has more bargaining leverage, that party is more
likely to achieve a favorable price in the negotiation.

     The district court had to determine whether the economic
theory applied to the particular market by considering evidence
about the “structure, history, and probable future” of the video
programming and distribution industry. United States v.
General Dynamics, 

, 498 (1974) (internal
quotation marks omitted); see also Brown 

–22 & n.38. As one circuit has put it, “[t]he Nash theorem
arrives at a result that follows from a certain set of premises,”
while the theory “asserts nothing about what situations in the
real world fit those premises.” VirnetX, Inc., v. Cisco Sys. Inc.,
                               19

, 1332 (Fed. Cir. 2014). The district court
concluded that the government presented insufficient real-
world evidence to support the prediction under the Nash
bargaining theory of a material increase of Turner
Broadcasting’s post-merger bargaining leverage in affiliate
negotiations by reason of less-costly long-term blackouts. The
government’s real-world evidence consisted of statements by
AT&T Inc. and DirecTV in FCC regulatory filings that vertical
integration, such as in the proposed Comcast-NBCU merger,
can give distributors an incentive to charge higher affiliate fees
and expert opinion and a quantitative model prepared by
Professor Shapiro. The expert opinion and model were subject
to deficiencies identified by AT&T’s experts, some of which
Professor Shapiro conceded. By contrast, AT&T’s expert’s
econometric analysis of real-world data showed that content
pricing in prior vertical mergers in the industry had not
increased as the Nash bargaining theory and the model
predicted. Given evidence the industry was now “remarkably
dynamic,” the district court credited CEO testimony about the
null effect of vertical integration on affiliate negotiations.
AT&T, 

at 194.

    In other words, the record shows that the district court
accepted the Nash bargaining theory as an economic principle
generally but rejected its specific prediction in light of the
evidence that the district court credited. The district court
explained that its conclusion

         does not turn on defendants’ protestations that the
         theory is ‘preposterous,’ ‘ridiculous,’ or ‘absurd.’ . . .
         [but] instead on [its] evaluation of the shortcomings in
         the proffered third-party competitor testimony, . . . the
         testimony about the complex nature of these
         negotiations and the low likelihood of a long-term
         Turner [Broadcasting] blackout, . . . and the fact that
                                20
         real-world pricing data and experiences of individuals
         who have negotiated on behalf of vertically integrated
         entities all fail to support the [g]overnment’s
         increased-leverage theory.

(internal citations omitted).

     More concerning is the government’s contention that the
district court misapplied the Nash bargaining theory in a
manner that negated its acceptance of the economics of
bargaining by erroneously focusing on whether long-term
blackouts would actually occur after the merger, rather than on
the changes in stakes of such a blackout for Turner
Broadcasting. The government points to the district court’s
statements that Professor Shapiro’s testimony was undermined
by evidence that “a blackout would be infeasible.” AT&T, 

at 223. The district court also stated that “there has
never been, and is likely never going to be, an actual long-term
blackout of Turner [Broadcasting] content.” 

court noted that “Turner [Broadcasting] would not be willing
to accept the ‘catastrophic’ affiliate fee and advertising losses
associated with a long-term blackout.” 

n.35
(emphasis in original).

     The question posed by the Nash bargaining theory is
whether Turner Broadcasting would be more favorably
positioned after the merger to assert its leverage in affiliate
negotiations whereby the cost of its content would increase.
Considered in isolation, the district court’s statements could be
viewed as addressing the wrong question. Considered as part
of the district court’s analysis of whether the stakes for Turner
Broadcasting would change and if so by how much, the
statements address whether the threat of long-term blackouts
would be credible, as posited by the government’s increased
leverage theory. The district court found that after the merger
                                21
the stakes for Turner Broadcasting would change only slightly,
so its threat of a long-term blackout “will only be somewhat
less incredible.” 

(quoting Professor Shapiro); see
generally Brown 

–29. Recognizing
Professor Shapiro applied the Nash bargaining theory in
opining that “if a party’s alternative to striking a deal improves,
that party is more willing and able to push harder for a better
deal because it faces less downside risk if the deal implodes,”
AT&T, 

at 223 n.35, the district court rejected
the assumption underlying the government’s theory that Turner
Broadcasting would gain increased leverage from this slight
change in stakes. It relied on testimony that the small change
in bargaining position from a less costly blackout would not
cause Turner Broadcasting to take more risks, specifically
noting the Time Warner CEO’s analogy of the cost difference
between having a 1,000-pound weight fall on Turner
Broadcasting and a 950-pound weight fall on it — the
difference being unlikely to change the risk Turner
Broadcasting would be willing to take. 

n.36 (Jeff
Bewkes, Time Warner CEO, Tr. 3121 (April 18, 2018)).

     The district court’s statements identified by the
government, then, do not indicate that the district court
misunderstood or misapplied the Nash bargaining theory but
rather, upon considering whether in the context of a dynamic
market where a similar merger had not resulted in a
“statistically significant increase in content costs,” the district
court concluded that the theory inaccurately predicted the post-
merger increase in content costs during affiliate negotiations.

     Of course, it was not enough for the government merely to
prove that after the merger the costs of a long-term blackout
would change for Turner Broadcasting. Its theory is that
Turner Broadcasting’s bargaining leverage would increase
sufficiently to enable it to charge higher prices for its content.
                               22
The district court’s focus on the slight change in the cost of a
long-term blackout in finding Turner Broadcasting’s
bargaining leverage would not meaningfully change aligns
with determining whether the government’s evidence
established that a change in the post-merger stakes for Turner
Broadcasting would likely allow it to extract higher prices
during affiliate negotiations. The district court reasoned that
because long-term blackouts are very costly and would
therefore be infeasible for Turner Broadcasting even after the
merger, there was insufficient evidence that “a post-merger
Turner [Broadcasting] would, or even could, drive up prices by
threatening distributors with long-term blackouts.” 

(emphasis in original). In finding the government failed to
“prov[e] that Turner [Broadcasting]’s post-merger negotiating
position would materially increase based on its ownership by
AT&T,” 

the district court reached a fact-specific
conclusion based on real-world evidence that, contrary to the
Nash bargaining theory and government expert opinion on
increased content costs, the post-merger cost of a long-term
blackout would not sufficiently change to enable Turner
Broadcasting to secure higher affiliate fees. Witnesses such as
a Turner Broadcasting president Coleman Breland, AT&T
executive John Stankey, and Time Warner CEO Jeff Bewkes,
whom the district court credited, testified that after the merger
blackouts would remain too costly to risk and that any change
in that cost would not affect negotiations as the government’s
theory predicted.

     Not to be overlooked, the district court also credited the
efficacy of Turner Broadcasting’s “irrevocable” offer of
arbitration agreements with a no-blackout guarantee. It
characterized the no-blackout agreements as “extra icing on a
cake already frosted.” AT&T, 

at 241 n.51. In
crediting Professor Carlton’s econometric analysis, the district
court explained that it was appropriate to consider the analysis
                              23
of the Comcast-NBCU merger because the Comcast-NBCU
merger was similar to the proposed merger — a vertical merger
in the video programming and distribution industry. There the
government had recognized, “‘especially in vertical mergers,
that conduct remedies,’ such as the ones proposed [in the
Comcast case], ‘can be a very useful tool to address the
competitive problems while preserving competition and
allowing efficiencies’ that ‘may result from the transaction.’”
AT&T, 

at 169 n.3 (quoting Hr’g Tr. 15:16–21
(July 27, 2011), Comcast, 

145). Like there,
the district court concluded the Turner arbitration agreements
would have “real-world effect.” 

n.30.

     The post-merger arbitration agreements would prevent the
blackout of Turner Broadcasting content while arbitration is
pending. AT&T, 

at 217. As mentioned, Turner
Broadcasting “irrevocably offer[ed]” approximately 1,000
distributors agreements to engage in baseball style arbitration
in the event the parties fail to reach a renewal agreement, and
the offered agreement guarantees no blackout of Turner
Broadcasting content once arbitration is invoked. AT&T’s
counsel represented the no-blackout commitment is “legally
enforceable,” Oral Arg. Tr. 53:7–8, and AT&T “will honor”
the arbitration agreement offers, Oral Arg. Tr. 55:2, 61:22–25,
62:1–10. Consequently, the government’s challenges to the
district court’s treatment of its economic theories becomes
largely irrelevant, at least during the seven-year period.
Counsel for Amici Curiae 27 Antitrust Scholars explained that
arbitration agreements make the Nash bargaining model
premised on two-party negotiations “substantially more
complicated,” Oral Arg. Tr. 50:10, and Professor Shapiro
acknowledged that taking the arbitration agreements into
account would require “a completely different model.” Tr.
2325 (April 11, 2018).
                               24
     Further, the government’s contention that the district court
failed to properly weigh the probative force of the defendants’
statements in FCC filings is unavailing. During licensing and
rulemaking proceedings before the FCC, DirecTV stated “a
standard economic model” (i.e., the Nash bargaining theory)
predicts that the proposed Comcast-NBCU merger “would
significantly increase the prices other MVPDs pay for NBCU
programming,” and two years later stated, similar to AT&T
Inc. comments, that “vertically integrated MVPDs have an
incentive to charge higher license fees for programming that is
particularly effective in gaining MVPD subscribers than do
non-vertically integrated MVPDs.” 1 The district court took
judicial notice of these statements pursuant to Federal Rule of
Evidence 201, explaining it was “hesitant” to assign significant
evidentiary value to the prior regulatory filings because AT&T
and DirecTV made the statements acting as competitors whose
positions would be affected by FCC review. AT&T, 310 F.
Supp. 3d at 206. FCC rules require all regulated parties —
whether applicants seeking to transfer licenses in connection
with a proposed merger or competitors who oppose the merger
— to provide only “[t]ruthful and accurate statements to the
Commission” in adjudicatory proceedings. 47 C.F.R. § 1.17;
see FCC Amicus Curiae Br. 3. The statements were admissible
as party admissions pursuant to Federal Rule of Evidence

1
 In re Matter of Applications of Comcast Corp., General Electric
Co. and NBC Universal, Inc., for Consent to Assign Licenses or
Transfer Control of Licensees, MB Docket No. 10-56, Reply
Comments of DirecTV, Inc., 4 (Aug. 19, 2010); In re Matter of
Revision of the Commission’s Program Access Rules et al., MB
Docket No. 12-68 et al., Comments of DirecTV, LLC 19 (June 22,
2012); In re Matter of Revision of the Commission’s Program Access
Rules et al., MB Docket No. 12-68 et al., Comments of AT&T Inc.
22 (June 22, 2012); In re Matter of Revision of the Commission’s
Program Access Rules et al., MB Docket No. 12-68 et al., Reply
Comments of AT&T Inc. 2 (July 23, 2012).
                               25
801(d)(2), see Talavera v. Shah, 

, 309 (D.C. Cir.
2011), yet even as admissions, the district court had to evaluate
their persuasive force in the circumstances before it, and the
district court did. See 

; cf. General

; Owens v. Republic of Sudan, 

, 790 (D.C. Cir. 2017).

     The district court accepted the FCC statements as
probative of the proposition that the Nash bargaining theory
could apply in the context of affiliate fee negotiations. But it
concluded generic statements that vertical integration “can”
allow an entity to gain an unfair advantage over rivals were
“informed by the state of the market at the time . . . and the
particular inputs to the models presented to the FCC.” AT&T,

at 206. As such the FCC statements were “not
particularly probative of whether [the proposed merger] could
do the same with its programming in today’s more competitive
marketplace,” with the rising presence of virtual MVPDs and
SVODs, like Netflix and Hulu. 

see also 

The district court also noted that many of the statements
in the FCC regulatory filings related to whether a vertically
integrated programmer would withhold content, which
Professor Shapiro opined would not occur here because it
would be profitable for the merged firm to continue to license
Turner programming. 

see also 

Tr. 2218,
2293. Once the district court credited AT&T’s expert’s
opinion based on an econometric analysis that the similar
Comcast-NBCU merger had not had a “statistically significant
effect on content costs,” 

the district court could
understand that the defendants’ admissions at the time of the
Comcast-NBCU merger offered little probative support for the
government’s increased leverage theory.

    Thus, even viewing the statements to the FCC as
supportive of the government, the district court’s finding of the
                               26
efficacy of Turner Broadcasting’s irrevocable offers of no-
blackout arbitration agreements means the merger is unlikely
to afford Turner Broadcasting increased bargaining leverage.

         (b) Corporate-wide profit maximization. Still, the
government maintains that the reliance on past negotiation
experience indicates that the district court misunderstood, and
failed to apply, the principle of corporate-wide profit
maximization by treating the principle as a question of fact,
when “[t]he assumption of profit maximization is ‘crucial’ in
predicting business behavior.” Appellant Br. 50 (citation
omitted). This principle posits that a business with multiple
divisions will seek to maximize its total profits. It was adopted
as a principle of antitrust law in Copperweld Corp. v.
Independent Tube Co., 

, 771 (1984), holding that
a parent and a wholly-owned subsidiary are not capable of
conspiracy against each other under Section 1 of the Sherman
Antitrust Act. Companies with multiple divisions must be
viewed as a single actor, and each division will act to pursue
the common interests of the whole corporation. See 

     The district court never cited Copperweld in its opinion,
which is troubling given the government’s competitive harm
theories and expert evidence based on economic principles.
But the government’s position that the district court never
accepted this economic principle overlooks that it did “accept
Professor Shapiro’s (and the Government’s) argument that
generally, ‘a firm with multiple divisions will act to maximize
profits across them.’” AT&T, 

at 222 (citations
omitted). And it ignores that if the merged firm was unable to
exert the leverage required by the government’s increased
leverage theory, then inquiring (as the district court did of
Professor Shapiro) about an independent basis to conclude that
the firm did have such leverage is not a rejection of the
corporate-wide profit maximization principle.
                               27

     The government maintains that the district court’s
misapplication of the principle of corporate-wide profit
maximization is evident from its statement the evidence
suggests “vertically integrated corporations have previously
determined that the best way to increase company wide profits
is for the programming and distribution components to
separately maximize their respective revenues.” 

Stating that the programming and distribution divisions would
“separately maximize their respective revenues” is contrary to
the maximization principle to the extent separate units would
act against the merged entity’s common interest. See

. At this point in its opinion,
however, the district court was explaining why “that profit-
maximization principle is not inconsistent with testimony that
the identity of a programmer’s owner has not affected affiliate
negotiations in real-world instances of vertical integration.” 

The district court can be viewed as conveying its
understanding that Turner Broadcasting’s interest in spreading
its content among distributors, not imposing long-term
blackouts, would redound to the merged firm’s financial
benefit, not that Turner Broadcasting would act in a manner
contrary to the merged firm’s financial benefit. Industry
executives testified that “the identity of a programmer’s owner
has not affected affiliate negotiations in real-world instances of
vertical integration,” AT&T, 

at 222. For
instance, the Chair of Content Distribution at NBC Universal
testified that at Comcast-NBCU, he “never once took into
account the interest of Comcast cable in trying to negotiate a
carriage agreement” for NBC Universal. Tr. 2014 (Madison
Bond, NBC Universal, Chairman of Content Distribution
(April 10, 2018)); see also Tr. 1129 (Coleman Breland, Turner
Broadcasting President (April 2, 2018)).
                               28
     To the extent the government maintains this testimony is
irreconcilable with the legal principle of corporate-wide profit
maximization, it gives no credence to the district court’s focus
on “the best way to increase company wide profits,” referring
to the merged firm. AT&T, 

at 222. In other
words, the district court was explaining that real-world
evidence reflected the profit-maximization principle. Even if
the district court could have made clearer that it understood the
principle was not a question of fact, the government does not
explain how considering how that is done in a particular
industry is contrary to the principle of corporate-wide profit
maximization.

     Nor is the conclusion that the merged firm would not be
able to maximize its profits by raising prices during
negotiations inconsistent with the principle of corporate-wide
profit maximization. Based on the record evidence, the district
court could plausibly understand that the proposed merger
would not enable the merged entity to exert increased
bargaining leverage by means of long-term blackouts and,
therefore, would not affect affiliate fee negotiations to raise
content costs. See 

, 577. Finding the
distributor division’s interest would not affect Turner
Broadcasting’s negotiations with other distributors is
consistent with the evidence that when a programmer and
distributor merge, it is still in the best interests of the merged
entity as a profit maximizer to license programming broadly to
other distributors. Tr. 1129 (Breland (April 2, 2018)). That is,
instead of withholding content in an attempt to benefit the
merged entity, programmers will seek to license their content
to other distributors. In this instance, the district court
concluded the principle and the real world “fit.” Moreover,
AT&T’s view that the government’s claims of fundamental
economic errors are ultimately irrelevant in light of Turner
                               29
Broadcasting’s    irrevocable        arbitration/no     blackout
commitment is not implausible.

      Similarly, contrary to the government’s position, the
district court’s findings about post-merger negotiating are not
internally inconsistent with its finding on the cost savings of
the merger. The district court found, and the government
agreed, that the merger would result in cost savings as a result
of EDM. Pre-merger, both Turner Broadcasting and AT&T
earned margins over cost before their products reached
consumers: Turner Broadcasting earned a profit margin when
it licensed content to AT&T, and AT&T earned a profit margin
when it sold content to consumers. Post-merger, Turner
Broadcasting would not earn a profit margin when licensing
content to AT&T because the merged entity would eliminate
that cost and, according to Professor Shapiro, pass on some of
those cost savings to consumers in order to attract additional
subscribers. For there to be EDM savings, Professor Shapiro
opined, the merged firm must act on its unified interest across
divisions. Thus, Turner Broadcasting, instead of maximizing
its own revenue, would license its programming to AT&T for
a lower price. The government did not contest AT&T’s
position that a merged entity can maximize its own profits by
eliminating cost even if it has no ability to secure higher prices
from other companies during negotiations. At most, the
government challenged the sufficiency of the evidence for
finding the merged entity would not be able to increase prices
for Turner Broadcasting content. Reply Br. 13–14. But there
is record evidence to support finding that AT&T would be able
to eliminate its own costs without gaining the ability to raise
Turner Broadcasting content prices.

    (2) Inconsistent reasoning in evaluating trial testimony.
The government further maintains that the district court used
                                30
internally inconsistent reasoning when evaluating testimony
from witnesses in the industry.

     At trial, third-party distributors and executives from
Comcast-NBCU and Time Warner testified about negotiations
in the video programming and distribution industry. Third-
party distributors testified about their concerns, and their
reasons, that Turner Broadcasting would gain increased
bargaining leverage as a result of the proposed merger.
Comcast-NBCU and Time Warner executives testified that the
interests of an affiliated distributor did not affect negotiations
in their prior experiences negotiating on behalf of vertically
integrated companies. The district court concluded that the
third-party distributor testimony “fail[ed] to provide
meaningful, reliable support for the [g]overnment’s increased
leverage theory,” AT&T, 

at 211, while the
executives’ testimony “undermine[d] the persuasiveness of the
[g]overnment’s proof,” 

The district court declined
to credit the third-party distributors’ testimony because “there
is a threat that [third-party distributor] testimony reflects self-
interest,” 

yet dismissed the suggestion that testimony
from the Time Warner executives should be discounted as
potentially biased due to self-interest, 

     The government contends this reasoning was inconsistent
because self-interest existed on both sides of the issue of
whether the proposed merger would have anticompetitive
effects. Even so, the potential for self-interest was not the only
reason the district court found third-party distributor testimony
of little probative value. Much of the third-party competitor
testimony, the district court found, “consisted of speculative
concerns,” 

and did not contain any analysis or factual
basis to support key assumptions, such as how Turner
Broadcasting’s bargaining leverage would change and how
many subscribers distributors would lose in a blackout. By
                                31
contrast, the Time Warner executives’ testimony did “not
involve promises or speculations about the employees’ future,
post-merger behavior” and instead recounted “what these
executives previously experienced when working within a
vertically integrated company.” 

Their testimony
was uniform among all testifying witnesses and corroborated
by that of a Comcast-NBCU executive — a competitor of
AT&T. To the extent the government also maintains the
district court improperly discounted the third-party distributor
testimony because it contradicted Professor Shapiro’s opinion
that Turner Broadcasting would not actually withhold content
from other distributors, any error in that regard does not
demonstrate the district court clearly erred in discounting their
testimony for the independent reasons that it rested on
speculative, future predictions and lacked adequate factual
support.

     (3) Rejection of Professor Shapiro’s quantitative model.
Finally, the government contends that the district court clearly
erred in rejecting Professor Shapiro’s quantitative bargaining
model. Specifically, that the district court erred in finding
insufficient evidence to support Professor Shapiro’s
calculations of fee increases for rival distributors and in finding
no proof of any price increase to consumers.

     Preliminarily, the court does not hold that quantitative
evidence of price increase is required in order to prevail on a
Section 7 challenge. Vertical mergers can create harms beyond
higher prices for consumers, including decreased product
quality and reduced innovation. See Amicus Curiae Open
Markets Institute Br. 4–12. Indeed, the Supreme Court upheld
the Federal Trade Commission’s Section 7 challenge to Ford
Motor Company’s proposed vertical merger with a major spark
plug manufacturer without quantitative evidence about price
increases. Ford Motor Co. v. United States, 

, 567–
                               32
69, 578 (1972). Here, however, the government did not present
its challenge to the AT&T-Time Warner merger in terms of
creating non-price related harms in the video programming and
distribution industry, and we turn to the government’s
challenges to the district court’s handling of the quantitative
evidence regarding the proposed merger’s predicted effect on
consumer price.

     Professor Shapiro presented a quantitative model that
predicted an annual net increase of $286 million being passed
on to consumers in 2016, with increasing costs in future years.
This figure was based on the model’s predictions of an annual
fee increase of $587 million for rival distributors to license
Turner Broadcasting content and cost savings of $352 million
for AT&T. The district court accepted Professor Shapiro’s
testimony about the $352 million cost savings from the merger.
But it found that insufficient evidence supported the inputs and
assumptions used to estimate the annual costs increases for
rival distributors, crediting criticisms by Professor Carlton and
Professor Rossi. Indeed, the district court found that the
quantitative model as presented through Professor Shapiro’s
opinion testimony did not provide an adequate basis to
conclude that the merger will lead to “any” raised costs for
distributors or consumers, “much less consumer harms that
outweigh the conceded $350 million in annual cost savings to
AT&T’s customers.” AT&T, 

at 241 (emphasis
in original).

     Whatever errors the district court may have made in
evaluating the inputs for Professor Shapiro’s quantitative
model, the model did not take into account long-term contracts,
which would constrain Turner Broadcasting’s ability to raise
content prices for distributors. The district court found that the
real-world effects of Turner Broadcasting’s existing contracts
would be “significant” until 2021 and that it would be difficult
                               33
to predict price increases farther into the future, particularly
given that the industry is continually changing and
experiencing increasing competition. This failure, the district
court found, resulted in overestimation of how quickly the
harms would occur. Professor Shapiro acknowledged that
predictions farther into the future, after the long-term contracts
expire, are more difficult. Tr. 2317 (April 11, 2018). Neither
Professor Shapiro’s opinion testimony nor his quantitative
model considered the effect of the post-litigation offer of
arbitration agreements, something he acknowledged would
require a new model. And the video programming and
distribution industry had experienced “ever-increasing
competitiveness” in recent years. AT&T, 

at
241. Taken together, the government’s clear-error contention
therefore fails.

     It is true that the district court misstated that the
government had not proven that any price increases would
“outweigh the conceded $350 million in annual cost saving to
AT&T’s customers.” 

testified that the
merger would result in $352 million cost savings to AT&T and
that not all those savings would be passed on to consumers.
The $352 million, therefore, was not cost savings to consumers
but to AT&T. But the district court did not weigh increased
prices for consumers against cost savings for consumers, and
instead found that the government had not shown at the first
level that the merger was likely to lead to any price increases
for consumers because of the failure to show that costs for rival
MVPDs would increase as a result of Turner Broadcasting’s
increased leverage in affiliate negotiations after the merger.
Counsel for the government and AT&T agree the error
regarding the consumer savings value alone would not require
remand because the district court’s opinion was not based on
balancing any price increases against cost savings to
consumers. Oral Arg. Tr. 36–37, 57:1–13. Consequently,
                               34
because the government failed to meet its burden of proof
under its increased leverage theory at the first level, the error
regarding cost savings was harmless error, see Czekalski v.
LaHood, 

, 453 (D.C. Cir. 2009); FED. R. CIV. P.
61.

     Accordingly, because the district court did not abuse its
discretion in denying injunctive relief, see 

–53, we affirm the district court’s order denying a
permanent injunction of the merger.
35