United States v. Bertelsmann Se & Co. Kgaa

UNITED STATES DISTRICT COURT

FOR THE DISTRICT OF COLUMBIA

UNITED STATES OF AMERICA,

Plaintiff,

Vv. Civil Action No. 21-2886-FYP

BERTELSMANN SE & CO. KGaA,

PENGUIN RANDOM HOUSE, LLC,

VIACOMCEBS, INC., and

SIMON & SCHUSTER, INC.,

Defendants.

MEMORANDUM OPINION

John Steinbeck famously said, “I guess there are never enough books.” He apparently

meant that in the figurative sense, as a comment on the power of books to educate, to enrich, and

to explore. But today, his statement also rings true in the economic sense: The retail market for

books in the United States was over $11.5 billion in 2019 and has only continued to expand.

People want to read. And book publishers have the enormous power and responsibility to decide

which books — and therefore which ideas and stories — will be made broadly available to the

public. A publishers’ marketplace of ideas is also a marketplace of book sales, production costs,

and market share. It is this commercial market, so inextricably intertwined with the intellectual

life of our nation, that the Court examines in this case.

Penguin Random House (“PRH”) is by far the largest book publisher in the United States.

Owned by Bertelsmann SE & Co. KGaA (“Bertelsmann”), an international media and services

company, PRH annually publishes over 2,000 new books in the U.S. and generates nearly $2.5

billion in revenue. Simon & Schuster, Inc. (“S&S”), owned by the media giant Paramount

Global (formerly ViacomCBS), is the third-largest publisher in the U.S. S&S publishes about

1,000 new titles yearly and reported over $760 million in net sales in 2020.

In March 2020, ViacomCBS announced that it planned to sell S&S. Following a multi-

round bidding process, Bertelsmann and PRH signed an agreement with ViacomCBS and S&S in

November 2020 to purchase S&S for $2.175 billion. The acquisition of S&S would cement

PRH’s position as the “number one” publisher in the United States, increasing its retail market

share to almost three times that of its closest competitor.' Trial Tr. at 741:17~742:4 (Dohle).

In November 2021, the Antitrust Division of the United States Department of Justice

(“the government”) brought this action against PRH, S&S, and their parent companies (“the

defendants”), seeking to block the merger of PRH and S&S under Section 7 of the Clayton Act.

The government’s case sounds in “monopsony,” a market condition where a buyer with too

much market power can lower prices or otherwise harm sellers. Essentially, the government

alleges that the merger will increase market concentration in the publishing industry, which will

allow publishing companies to pay certain authors less money for the rights to publish their

books.

The case proceeded to trial on August 1, 2022. For twelve days, the Court heard

evidence and argument about how PRH’s acquisition of S&S would affect competition in the

“upstream” market for publishing rights. The Court heard testimony from authors, publishers,

1 In 2019, PRH had a percent share of the market for U.S. book sales and S&S had al percent share.

See ECF No. 177 (United States’ Sealed Corrected Proposed Findings of Fact and Conclusions of Law (“Govt.

PFOF-PCOL”)) § 13 (citing Defendants’ Exhibit No. DX”) 105 at 64); see also ECF No. 184 (government’s

redacted post-trial brief). Thus, the combined entity would have a market share of approximately percent. The

merged company’s next closest competitor would iii which had an percent share of

the market for book sales in 2019. See Govt. PFOF-PCOL 4 13 (citing DX 105 at 64).

literary agents, and industry executives, and admitted over 230 exhibits.? After a thorough

review of the record and careful consideration of the parties’ arguments, the Court concludes that

PRH’s acquisition of S&S is likely to substantially lessen competition to acquire “the publishing

rights to anticipated top-selling books,” which comprise the relevant market in this case. The

Court therefore will enjoin the proposed merger of PRH and S&S.

I. BACKGROUND

A. The Industry

The book industry is dominated by five major publishing houses — PRH, HarperCollins

Publishers, S&S, Hachette Book Group, and Macmillan Publishing Group, LLC — which are

known as the “Big Five.” Together, the Big Five held nearly 60 percent of the market for the

sale of trade books in 2021 (i.e., books intended for general readership, as opposed to specialized

books like textbooks or manuals). See DX 382; PX 663 at 92.

2 The Court appreciated hearing the testimony of many dedicated professionals who work in the publishing

industry. The Court heard from authors Charles Duhigg, Stephen King, and Andrew Solomon. Publisher witnesses

included: Jennifer Bergstrom, Senior Vice President and Publisher of the S&S imprint Gallery Books Group

(“Gallery”); Sally Kim, Senior Vice President and Publisher of the PRH imprint G.P. Putnam’s Sons (“Putnam”);

Liate Stehlik, President and Publisher of Morrow Group, HarperCollins Publishers; and Brian Tart, President and

Publisher of the PRH imprint Viking Penguin (“Viking”). The following literary agents testified: Elyse Cheney,

Christy Fletcher, Ayesha Pande, Gail Ross, Jennifer Rudolph Walsh (expert witness), and Andrew Wylie. The

Court also heard from top industry executives, including Markus Dohle, CEO of PRH; Dennis Eulau, Executive

Vice President and COO of S&S; John Glusman, Vice President and Editor in Chief of W.W. Norton & Company

(“Norton”); Michael Jacobs, President and CEO of Abrams Books; Jonathan Karp, President and CEO of S&S;

Madeline McIntosh, CEO of PRH in the U.S.; Brian Murray, CEO of Harper Collins Publishers; Michael Pietsch,

CEO of Hachette Book Group; Don Weisberg, CEO of Macmillan Publishers; and Steven Zacharius, CEO of

Kensington Books. The government’s economic expert was Dr. Nicholas Hill, and the defendants’ economic expert

was Professor Edward Snyder. Finally, Adriana Porro, a statistician for U.S. Department of Justice, Antitrust

Division, also testified.

The following witnesses testified by video: Christy Fletcher, John Glusman, Michael Jacobs, Andrew

Solomon, Liate Stehlik, and Steven Zacharius. See Government’s Exhibit No. (“PX”) 2008 (Fletcher Dep.); Trial

Tr. at 1740:16-17 (noting that video of Fletcher’s deposition was played at trial); DX 422 (Glusman Dep.); DX 423

(Glusman sealed); Trial Tr. at 1880:4 (noting that video of Glusman’s deposition was played at trial); PX 2005

(Jacobs Dep.); Trial Tr. at at 701:13-23 (noting that video of Jacobs’s deposition was played at trial); PX 2004

(Solomon Dep.); Trial Tr. at 689:7—8 (noting that video of Solomon’s deposition was played at trial). PX 2002

(Stehlik Dep.) at 64-65; Trial Tr. at 628:23 (noting that video of Stehlik’s deposition was played at trial); PX 2000

(Zacharius Dep.); Trial Tr. at 385:11 (noting that video of Zacharius’s deposition was played at trial),

3

The Big Five have achieved their market dominance in part by acquiring other publishers,

contributing to a trend toward consolidation in the industry. Bertelsmann entered the U.S.

publishing market by acquiring Bantam Books in 1977, which merged with Doubleday Dell in

1986 and with Random House in 1998. PRH itself was formed in 2013 when Random House

acquired Penguin Books. Since 2013, PRH has continued to acquire other publishers, including

Sasquatch Books, Rodale, Little Tiger, F& W Media, and Sourcebooks. Meanwhile, Hachette

has acquired several independent publishers in the last decade, such as Workman Publishing,

Worthy, Perseus, and Black Dog & Leventhal. See Trial Tr. at 102:13-103:4 (Pietsch), 204:3—

19 (Pietsch). In 2021, HarperCollins acquired Houghton Mifflin Harcourt, which previously was

one of the largest among the mid-size, independent publishers. See Trial Tr. at 1386:12—17

(Murray), 192:6-193:15 (Pietsch). The remaining Big Five publisher, Macmillan, has pursued

organic growth. See Trial Tr. at 1079:23—1080:18 (Weisberg).

Some smaller publishers are well respected in the industry and compete against the Big

Five — in both the upstream market for acquiring books for publication and in the downstream

market for selling books to consumers. For instance, Scholastic is one of the largest children’s

book publishers and works with some of the same authors as the Big Five, see Trial Tr. at

118:20—22 (Pietsch), 545:10—-547:2 (Karp); while Kensington, one of the largest remaining

independent publishers, is a prominent purveyor of romance novels. See PX 2000 (Zacharius) at

8; PX 2002 (Stehlik) at 64-65. In addition, Norton is a prestigious publishing house specializing

in narrative nonfiction and is favored by some best-selling authors like Michael Lewis. See Trial

Tr. at 540:24-541:24 (Karp), 544:17—25 (Karp), 550:22—551:4 (Karp); DX 422 at 7. Other

players in the industry include well capitalized, mid-sized publishers like Amazon and Disney,

which each bring in over $100 million in annual revenues from publishing. See Trial Tr. at

737:22—738:11 (Dohle).

Each publishing company is organized as an umbrella organization that houses various

“imprints.” An imprint is a trade name or brand name for an editorial group. Imprints specialize

in publishing certain types of books and thus develop reputations for success in particular genres.

See PX 530 at 2. The editors within each imprint select and acquire manuscripts for publication;

and then collaborate with authors to develop and finalize their books. See Trial Tr. at 97:1-6

(Pietsch), 1915:10-25 (Duhigg), 1919:8—-1920:3 (Duhigg). PRH has close to 100 U.S.

publishing imprints within six publishing divisions. See Trial Tr. at 812:5—11 (Dohle); Govt.

Demo. | (PRH organizational chart). Its best-known imprints include Viking, Penguin Press,

Doubleday, Riverhead, Random House, and Putnam. S&S operates three publishing groups with

around 50 imprints, including Simon and Schuster, Atria Books, Scribner, and Gallery Books.

See PX 663 (materials for prospective S&S buyers) at 91, 101.

All publishers and editors are highly motivated to secure the rights to publish new books;

indeed, identifying and acquiring books that people want to read is the essence of the business.

Yet only 35 out of 100 books turn a profit, and breakout titles drive revenues — the top 4 percent

of profitable titles generate 60 percent of profitability. See PX 151 (presentation by PRH

executives on publishing industry) at 11; Trial Tr. at 747:16-18 (Dohle), 2289:2—10 (McIntosh)

(“Where the Crawdads Sing is a great current example. Fifty Shades of Grey, Gone Girl, Girl on

the Train... . [their] sales performance so outstrips our expectation, that they deliver most of the

profit to the company.”). Publishing has therefore been described by insiders as a “portfolio

business”: The business model is to acquire a large number of high-quality books, knowing that

a substantial percentage of the titles will not be profitable. See Trial Tr. at 747:5-9 (Dohle). As

PRH CEO Markus Dohle put it, publishers are “angel investors” that “invest every year in

thousands of ideas and dreams, and only a few make it to the top.” Jd. at 747:5-9. The books

that do “make it to the top” and sell well, especially over a number of years, allow the companies

to take risks in acquiring new books and enable publishers to manage the uncertainty inherent in

“betting” on new titles. See id. at 747:5—7 (Dohle); PX 151 at 11. Books that continue to sell

after the first year of publication comprise a publisher’s “back list,” which can provide an

important source of stable revenue. Back lists allow publishers to play the “long game” because

some books take a while to become profitable. See PX 2004 at 55. By contrast, the “front list,”

which consists of books not yet released or on the market for under a year, is risky and has

poorer margins, due to the expenses of marketing and roll-out associated with the new titles. See

Trial Tr. at 118:15—119:11 (Pietsch).

B. Acquiring Books for Publication

Books begin, of course, with authors. Authors often spend years developing their ideas,

conducting research, and refining their manuscripts or proposals before submitting them for

publication. A project that is acquired may still take months or years of work before it becomes

a completed book that is ready for distribution. See, e.g., id. at 1916:14-21 (Duhigg). To

support themselves, authors often rely on “advances” from their publishers. See id. at 1925:15—

22 (Duhigg), 1941:9-1943:] (Duhigg) (advance allowed him to take time off from his job to

write, support his newborn child, buy a house, and pay living expenses). An advance is an

upfront payment against the royalties that an author may earn in the future. The advance is the

3 Royalties are payments made to the author based on a book’s sales. See Trial Tr. at 106:20—-107:25

(Pietsch). For example, authors earn 7.5 percent of sales for paperback books. See id at 255:24—256:20 (Pande),

2011:9-10 (Kim). An advance is an upfront payment of those anticipated royalties; the author is not required to pay

back the advance even if the book’s actual royalties never exceed the amount of the advance. See id. at 106:20—

107:25 (Pietsch). Advances are paid in installments, typically in quarters. Generally, the first installment is paid

upon the signing of the book contract; the second payment is made upon delivery and acceptance of the manuscript;

“single most important” term in a contract for publishing rights because in a “large number of

cases, it may be the only compensation that the author will receive for their work.” Jd. at

254:18-24 (Pande); see also PX 2002 (Stehlik) at 67. Indeed, most authors do not “earn out”

4 In

their advances, i.e., ultimately earn royalties that exceed the amount of their advances.

addition to the advance, authors care about working with editors who share their vision for the

book and who can help them to “bring the book into the world.” Trial Tr. at 97:2—6 (Pietsch);

see also id. at 1918:5-24 (Duhigg), 1943:13-17 (Duhigg) (editor is “the reason I’m there” at

PRH), 2055:10-2056:9 (Cheney), 2063:5—17 (Cheney).

Authors generally are represented by literary agents, who use their judgment and

experience to find the best home for publishing a book. They typically begin the process by

submitting the book (which might be a full or partial manuscript, or just a proposal) to multiple

imprints or editors on a preliminary basis, to gauge the level of interest in the project. See id. at

246:11—-22 (Pande), 2105:22— 25(Wylie); PX 151 at 5. Agents use their expertise to determine

which imprints and editors to target, based on factors such as the kinds of books the imprint

previously has published, how effectively they have published those kinds of books, and the

ability of the company to pay appropriate compensation. See Trial Tr. at 246:9-22 (Pande),

2117:8~-2118:25 (Ross). Agents prioritize submitting manuscripts and proposals to Big Five

imprints because of their ability to pay; an agent might send out a second round of submissions

that includes more smaller publishers if interest among the Big Five is not strong. See id. at

246:17-22 (Pande), 248:15—249:10 (Pande). Given the size of PRH and S&S, and the number of

the third installment is paid upon publication; and the fourth payment is made twelve months later. See id at 777:1—

19 (Dohle), 1829:11~15 (Walsh), 255:1-25 (Pande). It can take three to four years for an entire advance to be paid

to an author. See id. at 255:15—18 (Pande), 777:1-19 (Dohle).

4 See Trial Tr. at 108:2-12 (Pietsch) (“Roughly half the time. About half the books we [Hachette] publish

earn out their advances.”), 254:20—24 (Pande) (20% of authors represented by Pande’s agency earned out advance),

1239:25—1240:8 (Hill) (more than 85% of author contracts for anticipated top-selling books never earn out their

advance), 2101:3-5 (Wylie) (6% of books he represents earn out their advances).

7

imprints they represent, some agents “always” submit to multiple PRH imprints and to S&S. See

id, at 250:10—251:3 (Pande), 260:16—21 (Pande).

Agents try to maximize the advances paid to authors: They not only have a fiduciary

duty to achieve the best deal, see id. at 494:7—17 (Karp), 1748:21—23 (Walsh), but they also are

paid on commission (typically 15 percent of the advance), See id, at 245:18~25 (Pande); PX 151

at 5. Indeed, one prominent agent has stated that it is his job to “get an advance that an author

doesn’t earn out,” Trial Tr. at 2100:5—25 (Wylie); while a publisher explained that “typically the

most important thing for an agent representing authors [is] to get the most amount of money up

front,” see PX 2002 (Stehlik) at 67.

Other common terms in a book contract include royalty rates, audio rights, and payment

structure (i.e., the number and timing of the installments in which an advance is paid). See Trial

Tr. at 255:11-17 (Pande), 255:24-256:20 (Pande). Yet many of those other terms have become

standardized across the industry, making advance levels even more important. For instance, each

book format has a standard royalty rate that is rarely altered. See id. at 106:20—107:13

(Pietsch).° Audio rights are now always included in book contracts and may not be sold

separately. See id. at 257:14—258:6 (Pande), 622:17—25 (Karp); PX 328 (internal PRH email) at

1-2 (“We have to get those [audio] rights.”). Finally, even though authors prefer to receive

“frontloaded” advances —- payments that are made sooner, and in fewer installments — the

industry norm has shifted in the opposite direction, from payouts in thirds to payouts in quarters.

See Trial Tr. at 254:25—255:3 (Pande), 256:22—25 (Pande), 1829:11-—18 (Walsh); at 19-

20.

5 Generally, paperback books receive a 7.5-percent royalty rate; e-books and audio or digital downloads

receive 25 percent of net sales; and hardcover books get a 10-percent rate for the first 5,000 copies sold, 12.5 percent

for copies 5,000 to 10,000, and 15 percent of sales exceeding 10,000 copies. See Trial Tr. at 255:24—256:20

(Pande), 2011:4-15 (Kim).

The Court is unconvinced by the testimony of certain defense witnesses who stated that

advance levels are not the most important factor in book acquisitions, and that the author’s “fit”

and comfort level with the editor are more significant. See Trial Tr. at 1756:2—1757:8 (Walsh),

1836:5—1837:2 (Walsh), 1935:14—-20 (Duhigg), 2055:5—2056:9 (Cheney), 2063:4—17 (Cheney).

While that may be true in a small number of cases, books generally are sold to the highest bidder.

See id, at 2395:2—10 (McIntosh) (underbidder winning an auction is “rare”), 2090:20—23 (Wylie)

(“[W]e are picking one that we feel presents the strongest combination of financial terms plus

editorial engagement and context for the author.”), 2106:1—7 (Wylie) (sold books to the highest

bidder about 93% of the time). While the choice of an editor is undoubtedly significant, the

agent typically has submitted the book only to a pre-screened list of suitable editors and thus may

choose the highest bid from among those editors. See id. at 246:1—22 (Pande). The

overwhelming weight of the evidence supports the conclusion that advance levels are the

primary focus of book acquisitions.

Most books are sold in exclusive negotiations between the agent and a single editor or

imprint.° Such negotiations come about because (1) a publisher has an “option” from a prior

contract with the author, which allows the publisher to take the first look at the author’s next

project and submit an exclusive bid within a limited period; (2) a publisher is willing to pay a

premium to “preempt” the book from being offered to others; (3) an agent approaches a single

editor that is a particularly good fit for a book and enters negotiations with that editor and their

imprint; or (4) only one imprint is interested after the agent has completed a round of

8 More than half of books are sold through negotiations. See Trial Tr. at 1608:20—1609:1 (Hill) (explaining

that 60% of books with advances over $250,000 involve negotiations, with the remaining 40% being auctions); see

also id, at 478:15~-479:1 (Karp) (explaining that less than half of acquisitions are auctions because S&S has “so

many repeat authors), 771:11—14 (Dohle) (auctions are “small sliver of the overall deals we do, especially the

expensive deals we do”), 1963:20-1964:18 (Kim) (explaining that Putnam imprint buys 80% of books through one-

on-one negotiations and attributing the higher percentage to its number of “franchise” authors).

9

submissions. See id. at 475:13-476:10 (Karp), 954:18-955:7 (Tart). Alternatively, an agent

might decide to utilize an auction format, which requires imprints to bid against each other to

acquire the book. Auctions may be organized in (1) a “round robin” format, where the agent

accepts competing bids in several “rounds,” eliminates the lowest bidders before proceeding to

the next round, and continues until only one bidder remains; (2) a “best bids” format, where all

bidders submit their highest bid in a single round; or (3) a hybrid format, such as “better best,”

where the bidding starts as a “round robin” and then shifts to a final round of “best bids.” See id.

at 111:8-113:4 (Pietsch), 2049:1-19 (Cheney), 2116:10-—2117:9 (Ross).

A publisher that hopes to acquire a desirable book must offer a competitive advance to be

in the running. Editors and publishers determine how much their imprint is willing to pay for a

given book. See id. at 97:16 (Pietsch). To make that determination, they estimate the

profitability of the book by generating a profit and loss statement (“P&L”). Such a statement

suggests an appropriate advance after considering (1) the anticipated sales of the book and its

expected list price (in various formats); and (2) predicted production and marketing costs. See

DX 414 (P&L Sheet); Trial Tr. at 915:5-918:21 (Tart). The most important input in a P&L —

and the driving force behind the advance offered — is the publisher’s estimate of book sales.

See Trial Tr. at 917:13-16 (Tart) (“the higher . . . sales equate to a higher advance in the P&L”),

110:19-111:1 (Pietsch) (relationship between the level of advance and projected sales is

“extremely close”). The sales prediction is based on the demand for comparable titles (referred

to as “comp titles” or “comps”), which have similar characteristics to the proposed book in terms

of subject matter, literary merit, or author background. Publishers also often confer with their

sales, publicity, and marketing teams about expected demand for a book. See id. at 842:15—

843:10 (Dohle), 914:11-18 (Tart), 1036:13-16 (Tart). But publishers do not rely only on the

10

P&L to determine an appropriate advance. They also consider factors like the editor’s

enthusiasm, their relationship with the author, whether the book might win an award, and

whether the book is in a growing category ——- if they really want a book, they will “stretch”

beyond what a P&L suggests would be profitable. Jd. at 969:4—970:14 (Tart), 967:3—10 (Tart).

Publishers’ sales estimates, broadly speaking, are reasonably reliable. Ultimately, there

is a correlation between high advances and high book sales. Books that sell well tend to have

garnered high advances, and books that receive high advances tend to sell well. See id. at 749:4—

22 (Dohle); PX 151 at 11.

C. The Competition for Books

Regardless of the method used to acquire a book’s publishing rights, the amount that is

paid is inexorably determined by competition. In an auction, a skillful agent can capitalize on

enthusiasm for a book and play bidders off against one another, knowing that a publisher will

“bid what ., . [it] need[s] to buy that book” because “it [only] takes one passionate editor at

another imprint to win that book away.” Trial Tr. at 1965:21—25 (Kim). Although the perceived

value of a book is subjective and may vary among editors, there is often a consensus among

industry players about which books will be successful. See id. at 2108:14—24 (Wylie) (“T think

there are recognizable qualities in — in books that people who have been in the business for a

long time would easily recognize.”), 310:12—24 (Pande) (agent explaining that she treats an

anticipated strong seller differently, such as by “sending it out as widely as I possibly can”).

It is not uncommon for editors and publishers to experience a “kind of auction fever,” in

which they change their sales expectations for a book and increase what they are willing to pay

for it during a competitive round-robin auction. Jd. at 180:20—181:11 (Pietsch). “[T]he interest

of other parties validates [a publisher’s] own sense of what a book is worth.” /d. The record

1]

contains numerous examples of books that sold for unexpectedly high advances and achieved

other favorable terms for their authors due to the bidding frenzy incited by competitive auctions.

For instance, in a hybrid auction of rounds followed by best bids, SS

Pe. initially received bids of between $150,000 and $400,000 from four publishers.

See PX 944-B (Porro bidding summary); Trial Tr. at 923:16~—930:18 (Tart) (discussing book’s

acquisition); PX 320 (emails). After six rounds of bidding, PRH’s Viking imprint more than

doubled its initial bid and won the book for $775,000 “over stiff competition.” PX 39 (email

from Tart); see also PX 944-B (Porro bidding summary); Trial Tr. at 923:16—930:18 (Tart). At

the best-bids stage, Viking decided to “stretch” from its initial bid clearance of $700,000 because

“there just is literally no telling what the opponents hold in their hands.” PX 326 at 1 (emails

between Tart and Viking editor Wendy Wolf). Another example is a

a which also benefitted from a rounds auction. See PX 948-B (Porro bidding

summary). In the first round, there were three bids, ranging from $200,000 to $300,000. See id.

Yet after five rounds, im received $535,000 plus a $100,000 bonus. See id.; Trial Tr. at

433:24435:17 (Karp) (discussing bidding process); PX 632 (emails discussing auction). The

record contains at least 11 other examples that illustrate the sharp increase in prices engendered

by competitive auctions, with advances increasing at least $100,000 by the end of the auction.’

7 See PX 937-B (one-round and best-bid auction with beginning high bid of $375,000 and winning bid of

$550,000); PX 938-B (four-round and best-bid auction with beginning high bid of $250,000 and winning bid of

$750,000); PX 716 (addressing auction summarized in PX 938-B); Trial Tr. at 435:21-437:19 (Karp) (same); PX

939-B (seven-round auction with beginning high bid of $300,000 and winning bid of $1.5 million); PX 940-B (five-

round and best-bid auction with beginning high bid of $400,000 and winning bid of $1.1 million for two books); PX

941-B (eight-round auction with beginning high bid of $550,000 and winning bid of $825,000); PX 947-B (two-

round and opportunity-to-improve auction with beginning high bid of $500,000 and winning bid of $600,000 and

bonuses for two books); PX 950-B (three-round and best-bid auction with a chance to improve, with beginning high

bid of $250,000 and winning bid of $700,000); PX 951-B (best-bid and opportunity-to-improve auction with initial

preempt offer of $750,000 and winning bid of $1.1 million); PX 954-B (three-round and best-bid auction with

beginning high bid of $750,000 and winning bid of $1.05 million); PX 955-B (one-round auction with opportunity

to improve, with beginning high bid of $800,000 and winning bid of $1.5 million); PX 729 (book initially received

$750,000 preemptive offer from S&S, went to auction, and then S&S made an offer 10% higher than final auction

bid, for offer of $1.1 million); Trial Tr, at 445:5-448:11 (Karp) (addressing book in PX 729).

12

Competition is also a key factor in one-on-one negotiations, where publishers must offer

high advances because they know that the agent always has the option of breaking off

negotiations and selling the book on the market. See Trial Tr. at 955:11—20 (Tart) (explaining

that in one-on-one negotiations “you know there’s competition out there’), 1847:1—-6

(Bergstorm) (“I assume I am negotiating exclusively, but I always have my competition in my

rearview mirror. But it’s one on one. And sometimes we don’t come to terms, and sometimes

they will go to someone else.”), 1966:17—24 (Sim) (“[E]ven if it’s a one-on-one negotiation .. .

we’re constantly aware that there’s competition ... .”), 114:21-115:6 (Pietsch). Some

publishers consider individual negotiations to be the most challenging way to acquire a book,

because “you are basically bidding against the author’s expectations and the agent’s

expectations,” and there are “no other market inputs [but] you know there’s competition out

there.” Jd. at 955:11—20 (Tart). In such situations, agents have bargaining leverage because the

threat of taking the book to other publishers always lurks in the background. This is particularly

true where a publisher is attempting to preempt the auction process. See, e.g., id. at 115:21—

116:5 (Pietsch) (“When we’re calculating a preempt, we want to bring an advance that we

believe the agent will consider a good advance; that they will think, yes, there’s a chance that if I

take this to auction, I might not get this much or this is the range that it might end up at. And so

we try to offer a high advance that we think will be compelling to — to the agent... .”); 303:13-

15 (Pande) (“So preemptive offers tend to be quite high because it has to incentivize us to be

willing to take the book off the table and not offer it in a competitive situation.”). As agent Gail

Ross stated: “[I]n this business, there’s always the other competitor. Whether . . . they’re

bidding or not, they’re always there.” Jd. at 2127:11-13.

13

Agents often submit a book to more than one imprint within a publishing company, see

id. at 250:10-251:3 (Pande), and publishers sometimes allow their imprints to bid against one

another in auctions. For example, PRH allows competitive bidding between its divisions, so

long as there also is an external bidder; but for imprints within the same division, PRH requires

the division to submit a “house bid.” See id. at 769:2—20 (Dohle), 935:20-—936:1 (Tart), 943 :3—

24 (Tart); PX 332 (email from PRH staff to agent explaining imprint bidding rules). A house bid

is a single bid made on behalf of more than one imprint from a particular publisher; the house bid

allows the agent to choose which imprint to work with, and each imprint might also submit a

“pitch,” ie., a letter or memo describing its editorial and other services. See PX 2002 (Stehlik) at

75-76. Hachette also allows its imprints to bid against one another if there is an external bidder,

see Trial Tr. at 239:11—23 (Pietsch), and Macmillan appears to allow some imprint competition,

compare PX 938-B (showing separate bids from two Macmillan imprints), with PX 941-B

(showing house bid from Macmillan imprints), and PX 954-B (showing one bid from Macmillan

imprints). S&S and HarperCollins, however, do not allow competitive bidding among their own

imprints but instead require their imprints to submit house bids. See Trial Tr. at 463:11—13

(Karp), 600:8-10 (Karp), 2119:11-—24 (Ross); PX 2002 (Stehlik) at 75-76. Allowing sibling

imprints to compete against each other increases the publisher’s chances of winning a title; gives

the editors from each imprint the freedom to pursue their desired projects, and allows authors

more choice in finding the most “comfortable home” and best editorial match for their books.

Trial Tr. at 839:11—840:4 (Dohle).

Although internal competition may yield benefits to publishers and authors, it is not in a

publisher’s economic interest to allow its own imprints to drive up the price of an acquisition,

and publishers therefore take steps to limit internal competition. See PX 411 (presentation by

i4

Madeline McIntosh to PRH Board) at 4 (explaining that McIntosh had “[i]ncreased background

coordination in auctions to leverage internal demand information better and avoid internal

upbidding”); Trial Tr. at 239:11—23 (Pietsch) (“Once we have only Hachette imprints bidding,

then .. . continuing to bid each other up would hurt the company’s collective P&L.”). Ample

evidence in the record demonstrates that PRH imprints often coordinate their bids within the

same auction, artificially suppressing advances. See id. at 2341:9-2345:7 (McIntosh), 2373:9—

2382:8 (McIntosh); PX 107 (email from McIntosh) (“I feel we should coordinate — shouldn’t be

forced into bidding against each other for existing authors”); PX 121 (emails between McIntosh

and PRH executive Nina von Moltke) (coordinating imprint bids). For example, PRH imprints

sometimes agree to submit the same bid. See PX 107 (“We are coordinated. Bill [Knopf

Doubleday] and Kara [Random House] will agree to a number and both offer same.”). They also

sometimes arrange to start their bidding from a lower number. See PX 116 (emails between von

Moltke and McIntosh) (deciding that PRH imprints “go in a bit lower in round 1”). Finally, PRH |

imprints sometimes decide to collectively “move up slowly” in their bidding, particularly if PRH

is the “main driver of value.” PX 336 (email from PRH executive to Tart). PRH appears to take

pride in its successful program of bid coordination. See PX 421 (email from von Moltke to

Mclntosh) (discussing auction where three top bids at $600,000 were from PRH imprints and

noting she was “[g|lad we didn’t go higher (this one definitely benefitted from the

coordination!)’). This type of behavior from independent companies would be illegal. Cf

United States v. Apple, Inc., 791 F.3d 290, 339 (2d Cir. 2015).

In competing for the most attractive new books, the Big Five have significant advantages

over smaller publishers. Most critically, the Big Five have the capital to take chances and place

bigger bets on a book’s success; that is, they can offer higher advances for more books. Indeed,

i5

agents and authors choose the Big Five because “they are most likely to pay an appropriate

advance,” see Trial Tr. at 246:19—22 (Pande); and at least one mid-size competitor observes that

the Big Five often “overpay” for books. See DX 422 (Glusman) at 76. The Big Five can afford

to take on more risk by paying higher advances because they have the most substantial back lists,

which are highly profitable. A book that does not sufficiently earn out its advance is a “loss” for

the publisher,*® but a publisher that has a steady income stream from its back list has a higher

tolerance for absorbing such losses, See Trial Tr. at 156:5-158:10 (Pietsch), 160:7-161:7

(Pietsch), 1066:22—25 (Weisberg) (“It[] . .. obviously has an impact on the bottom line if [your]

unearned advances are too large.”). PRH has the largest back list in the industry, and its back list

is the most significant source of its revenue. See id. at 2358:2—7 (McIntosh).

The Big Five also offer significant advantages in ensuring a book’s presence in the media

and visibility to its target audience. The Big Five publishers and their individual imprints have

teams dedicated solely to selling, marketing, and publicizing books, which have built critical

relationships with booksellers and the media. See PX 2004 (Solomon) at 64 (“There’s just a

whole industry that responds better to Big Five publishers.”); Trial Tr. at 259:13—260:4 (Pande),

840:5—841:12 (Dohle), 983:12—25 (Tart). Big Five publicity teams “engage with the media to

promote the book.” Trial Tr. at 1047:16~20 (Tart). Those teams can secure author interviews

on prominent programs like the Today Show, Good Morning America, or NPR, and can

persuade senior book reviewers to closely read and review the book. See PX 2004 (Solomon) at

63-64.”

8 A book need not earn out its entire advance for a publisher to profit; publishers begin to profit “at around

70 percent of earnout for most books.” See Trial Tr. at 1240:9-12 (Hill), 2258:21—25 (McIntosh).

° See PX 2004 (Solomon) at 63-64. (“You know, your book comes out and what do you want; you want to

be on NPR, you want to be on Good Morning America or the Today Show, you want to... do a radio interview

with Terry Gross. ... The publishers at the Big Five houses have more ready access to all of that. And. .. if there is

a new book that [S&S imprint] Scribner said is a really major title, it will at least be closely read by the editors of

16

The Big Five’s sales teams can help ensure that stores not only buy books but place them

in prominent displays. See Trial Tr. at 174:19-175:5 (Pietsch), 1372:11—25 (Murray), 1378:4—

1379:9 (Murray); PX 2002 (Stehlik) at 112 (“I would say generally the Big Five seem to have

more visibility in stores .. . than other smaller publishers.”). The Big Five edge extends to the

virtual marketplace; for instance, PRH hires data scientists to study Amazon’s search algorithms

and spends money to get books better positioned in Amazon’s search results. See Trial Tr. at

893:6-16 (Dohle); PX 2002 (Stehlik) at 112.

Meanwhile, marketing teams are responsible for paid advertising and use “every device

possible to find that [book’s] audience.” See Trial Tr. at 1067:17—-1068:14 (Weisberg). In

service of that goal, they produce market research and data analytics, as well as send marketing

materials to traditional outlets or social media influencers. See id. at 983:12—25 (Tart), 1849:13—

25 (Bergstrom), 1938:5—12 (Duhigg). The Big Five can even ensure that books look better when

they reach an audience, providing multiple versions of cover art for an author to choose from.

See PX 2001 Zacher) st 57 (a

SS 1a 1 2: 1920-413 Drie)

(“[Random House] came up with something like 13 or 15 different mockup book jackets to try

and figure out like which one is going to attract the reader’s eyes when it’s sitting there on a

shelf and get them to pick it up.”).

book reviews in a way that a book coming from a more obscure press is likely to go to a junior reader”); Trial Tr. at

1047:16—20 (Tart), 167:17—168:9 (Pietsch) (“And so to get [the media] to pay attention to your emails or return your

phone calls or come out to lunch or come to your pitch event, it takes a long time in developing a lot of

credibility.”), 1372:11-25 (Murray) (“And then we have publicists who have relationships with television and radio

producers.’’).

17

By contrast, smaller publishers might have a handful of staff doing all the editing,

marketing, publicity, and sales work on a book. See id. at 259:13—260:4 (Pande). Although

some of their books do well, that success is harder won and less frequent. See PX 2004

(Solomon) at 64 (“[T]here’s some fabulous books that are published by other houses and some of

them end up being successful. But it’s harder when you have fewer resources. It’s easier when

you have more resources.”). Authors want the easy advantages offered by the Big Five’s strong

publicity, marketing, and sales teams.'? Authors know that “when a publisher really gets behind

a book, particularly a big publisher, the chances are that that book is going to probably succeed

on some level.” See Trial Tr. at 335:23-336:1 (King).

Successful authors who first publish with smaller publishers often prefer to publish their

next book with a Big Five publisher. See fd. at 291:10—292:25 (Pande). Along with their

substantial resources, Big Five publishers have developed a valuable reputation for having strong

editorial staff with experience working with the best books and authors. See PX 530 at 2

(opining that Big Five “are known for their strong editorial . . . skills’”).!! Thus, a second book

with a Big Five publisher gives the author a better chance of an even bigger success. See Trial

Tr. at 291:10-292:25 (Pande). The Big Five view the smaller publishers as a “farm team” for

spotting writing talent, and routinely lure authors away from the non-Big Five publishers with

higher advances and the promise of superior marketing, distribution, and even cover design. See

10 See PX 2002 (Stehlik) at 101 (‘Most authors want to have their books in as many locations as possible.”),

112 (suggesting that visibility is one reason “why many agents and authors prefer to go with bigger publishers”).

W See also Trial Tr. at 353:8—25 (Eulau) (acknowledging that “reputation is important” for attracting authors),

454:1 1-22 (Karp) (agreeing that S&S’s decades of credibility and success attracts authors), 535:7-20 (Karp) (noting

that a Macmillan imprint has “a long reputation, so they can claim that when they publish a writer, that writer will be

following in the tradition of other great award-winning Nobel laureates”), 1375:24-1376:23 (Murray) (“[To acquire

top authors,] [y]ou have to have... expertise and a reputation. It helps if you have published authors that are

publishing to the similar readers, you know, so you can point to similar books that maybe you published one, two, or

three years ago that were successful.”), 2005:15-2006:18 (kim) (“[A]uthors want to be published by publishers with

good reputations, good standing, you know, with booksellers and media. They want to be a part of a list that they

can be proud to say they’re a part of.”).

18

PX 530 at 2 (explaining that small publishers “become farm teams for authors who then want to

move to a larger, more financially stable major publisher”); Trial Tr. at 291:10-—292:25 (Pande),

335:412 (King) (describing smaller publishers as the “minor leagues for writers”). The trial

record contains many examples of authors who moved from non-Big Five publishers to the Big

Five after establishing a track record of success. See Trial Tr. at 292:6-12 (Pande) (“I have had

several authors who have moved from small publishers to larger publishers. One of the authors’

name is Lad[ee] Hubbard. Her first book was published by Melville House. And for her second

book, she moved to [HarperCollins]. My author Lisa Ko moved from Algonquin to Penguin

Random House.”); DX 423 (Glusman) at 15-19

ee Of course, there are exceptions, as the defendants point out.

See, e.g., ECF No. 178 (Defendants’ Sealed Proposed Findings of Fact) 70 (citing PX 2001

(Zac)

a see also ECF No. 182 (Defendants’ Redacted Proposed Conclusions of

Law) ff] 9-10).

Self-publishing is not a significant factor in the publishing industry. Self-published

books are rarely published in print and are typically limited to online distribution. See Trial Tr.

at 173:13-23 (Pietsch), 1108:2—9 (Weisberg). The authors of self-published books cannot pay

themselves an advance. See id. at 173:8-15 (Pietsch) (remarking that for advances above

$100,000, “I do not consider [self-publishing] a threat at all because . . . [s]elf-published authors

v Defendants proposed conclusions of fact and law are contained in the same document (ECF No. 178) but

are separately enumerated. The Court will refer to the proposed conclusions of fact as “Defs. PFOF” and proposed

conclusions of law as “Defs. PCOL.”

19

can’t pay themselves an advance against royalties”). Moreover, individual authors generally do

not have relationships with media or distributors necessary to ensure that their books are visible

to a potential audience. See id. at 173:13—-23 (Pietsch) (“Self-published authors . . . don’t have

the ability to attract the attention of media. ... Imagine how hard it is . . . for one person who has

a book they published entirely on their own to say: Give me your attention. Review my book.

Promote my book. And so they simply don’t have access to the general-interest market that we

and the other Big 5 publishers address routinely. That’s our business.”). In short, self-

publishing cannot compete with the experience and resources of publishing companies. See id.

at 173:13-174:2 (Pietsch); PX 2004 (Solomon) at 52-53 (“I think a commercial publisher sells

more books, garners more reviews, gains more attention, does all kinds of things... . [ [as an

author] don’t have all of those business competencies that are involved.”); see also Trial Tr. at

2898:8—18 (Snyder) (positing that “self-publishing is not a relevant constraint”).

Il. LEGAL STANDARDS

Section 7 of the Clayton Act prohibits mergers and acquisitions “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. The “fundamental

purpose” of Section 7 is “to arrest the trend toward concentration, the tendency to [monopoly or

monopsony], before the [buyer’s or seller’s] alternatives disappear[] through merger... .”

United States v. Phila. Nat’l Bank, 374 U.S. 321, 367 (1963). Thus, Congress “sought to assure

.,. the courts the power to brake this force at its outset and before it gathered momentum.”

20

Brown Shoe Co. v. United States, 370 U.S. 294, 317-18 (1962).

To this end, “Congress used the words may be substantially to lessen competition to

indicate that its concern was with probabilities, not certainties.” FTC v. HJ. Heinz Co., 246 F.3d

708, 713 (D.C. Cir. 2001) (emphasis in original) (cleaned up) (quoting Brown Shoe, 370 U.S. at

323). The government “must prove the alleged Clayton Act violation by a preponderance of the

evidence,” i.e., that the merger would more likely than not violate the statute; but “‘[S]ection 7

does not require proof that a merger or other acquisition will cause higher prices [or

anticompetitive effects] in the affected market. All that is necessary is that the merger create an

appreciable danger of such consequences in the future.”” United States v. Anthem, Inc., 236 F.

Supp. 3d 171, 192 (D.D.C. 2017) (first quoting United States v. SunGard Data Sys., Inc., 172 F.

Supp. 2d 172, 180 (D.D.C. 2001); and then quoting Hosp. Corp. of Am. v. FTC, 807 F.2d 1381,

1389 (7th Cir. 1986)). '4 Section 7’s probabilistic standard “creates a relatively expansive

definition of antitrust liability” and “subjects mergers to searching scrutiny.” California v. Am.

Stores Co., 495 U.S. 271, 284 (1990).°

B The government’s theory is that the combined defendants would exercise market power on the buy-side of

the publishing market, i.¢., monopsony. Although most antitrust law has developed under sell-side theories of harm,

ie., monopoly, monopsony analysis relies on similar principles. Under the Horizontal Merger Guidelines, discussed

infra at n.15, “to evaluate whether a merger is likely to enhance market power on the buying side of the market, the

Agencies employ essentially the framework described ... for evaluating whether a merger is likely to enhance

market power on the selling side of the market.” U.S. Dep’t of Justice & Fed. Trade Comm/’n, Horizontal Merger

Guidelines § 12 (2010), https://www.justice.gov/sites/default/files/atr/legacy/20 10/08/1 9/hmg-2010-pdf (“Merger

Guidelines”). “The kinship between monopoly and monopsony suggests that similar legal standards should apply to

claims of monopolization and to claims of monopsonization.” Weyerhaeuser Co. v. Ross-Simmons Hardwood

Lumber Co., 549 U.S, 312, 321-22 (2007). “Monopsony and monopsony power are the equivalent on the buying

side of monopoly and monopoly power on the selling side.” United States v. Syufy Enters. 903 F.2d 659, 663 n.4

(9th Cir. 1990) (quotations omitted).

4 Although defendants quote United States vy. Marine Bancorporation, 418 U.S. 602, 623 n.22 (1974), for the

proposition that a merger’s anticompetitive effects must also be “imminent” to violate the Clayton Act, see Defs.

PCOL 4 1, the full quotation from that case is that “the loss of competition ‘which is sufficiently probable and

imminent’ is the concern of § 7,” Marine Bancorp., 418 U.S. at 623 n.22 (emphasis added) (citation omitted). The

Court understands that description of the standard to be functionally indistinguishable from the D.C. Circuit’s

formulation of the Section 7 standard, as described above.

15 In United States v. AT&T, Inc., the D.C. Circuit described the Section 7 standard of proof as follows:

“(T]he government must show that the proposed merger is likely to substantially lessen competition, which

21

Although “Congress neither adopted nor rejected specifically any particular tests for

measuring the relevant markets [where commerce is affected] .. . [nJor [defined] . . . the word

‘substantially,’” Brown Shoe, 370 U.S. at 320-21, the D.C. Circuit has taken a burden-shifting

approach to Section 7 cases. See AT&T, Inc., 916 F.3d at 1032; Heinz, 246 F.3d at 715; United

States vy. Baker Hughes Inc., 908 F.2d 981, 982-83 (D.C. 1990). The Baker Hughes test, as it

has come to be known, has a preliminary requirement and three steps. At the threshold, the

government must demonstrate the existence of a relevant market. See Marine Bancorp., 418

US. 602, 618 (1974); FTC v. RAG-Stiftung, 436 F. Supp. 3d 278, 291-92 (D.D.C. 2020). Once

it has done so, the first step of the test allows the government to establish a prima facie case and

a presumption of anticompetitive effects by demonstrating undue concentration within that

relevant market. See Anthem, 855 F.3d at 349; Heinz, 246 F.3d at 715. The second step shifts

the burden to the defendants, who must demonstrate in rebuttal that real-world conditions make

market concentration alone an unreliable predictor of the merger’s anticompetitive effects. See

Anthem, 855 F.3d at 349-50; Heinz, 246 F.3d at 715. Ifthe defendants successfully rebut the

prima facie case, the burden shifts back to the government in the third step “and merges with the

eficompasses a concept of ‘reasonable probability.’” 916 F.3d 1029, 1032 (D.C. Cir. 2019) (emphasis deleted)

(quoting Brown Shoe, 370 U.S. at 323 n.39). The parties dispute the meaning of this language. The defendants

argue that AT&T “require[s the government] to prove that a merger is ‘likely’ to cause substantial harm to

competition, [not] only that harm ‘may’ occur.” Defs. PCOL § 2. The government points to AT&T's explanation

“that this standard encompasses a concept of ‘reasonable probability,’” arguing that AT&T requires something less

than what the defendants propose. See Govt. PFOF-PCOL ff 38, 40; see also id. 7 38 (arguing for an “appreciable

danger” standard). The root of these competing formulations may be uncertainty over how the government’s

preponderance-of-the-evidence burden interacts with Section 7’s already probabilistic standard; combined, the two

standards require the government to prove “by a preponderance of the evidence” that the effect of a challenged

merger or acquisition “may be substantially to lessen competition.” Anthem, 236 F. Supp. 3d at 192. Like the

district court in AT&T, this Court “need not further toil over discerning or articulating the daylight, if any, between

‘appreciable danger,’ ‘probable,’ ‘reasonably probable,’ and ‘likely’ as used in the Section 7 context.” United States

v. AT&T, 310 F. Supp. 3d 161, 189 n.16 (D.D.C. 2018). The selection of any of the competing permutations is not

outcome-determinative in this case.

22

ultimate burden of persuasion, which remains with the government at all times.” Baker Hughes,

908 F.2d at 983; accord Anthem, 855 F.3d at 350.

IH. ANALYSIS

The government contends that the merger of PRH and S&S would harm competition to

acquire the publishing rights to “anticipated top-selling books,” resulting in lower advances for

the authors of such books and less favorable contract terms. The defendants do not dispute that

if advances are significantly decreased, some authors will not be able to write, resulting in fewer

books being published, less variety in the marketplace of ideas, and an inevitable loss of

intellectual and creative output. See Trial Tr. at 772:8—25 (Dohle). The defendants vigorously

contest, however, whether advances would decrease after the merger: They contend that

competition would not be harmed and that advances would actually rise.

A. Market Definition

The first step in merger analysis is the identification of a relevant market. See Marine

Bancorp., 418 U.S. at 618. Market definition “helps specify the line of commerce and section of

the country in which the competitive concern arises”; and allows the Court to evaluate any

anticompetitive effects by “identify[ing] market participants and measur[ing] market shares and

market concentration.” Merger Guidelines § 4.'° “Determination of the relevant market is a

necessary predicate to a finding of a violation of the Clayton Act because the threatened

monop[sony] must be one which will substantially lessen competition ‘within the area of

effective competition.’” United States v. EL. du Pont de Nemours & Co., 353 U.S. 586, 593

16 The Merger Guidelines “outline the principal analytical techniques, practices, and the enforcement policy

of the Department of Justice and the Federal Trade Commission” for merging competitors under federal antitrust

laws. Merger Guidelines § 1, They “describe the principal analytical techniques and the main types of evidence on

which the Agencies usually rely to predict whether a horizontal merger may substantially lessen competition.” /d.

Although the Merger Guidelines are not binding, courts have consistently looked to them for guidance in merger

cases. See, e.g., FTC v. Sysco Corp., 113 F. Supp. 3d 1, 38 (D.D.C. 2015).

23

(1957) (quoting Standard Oil Co. of Cal. v. United States, 337 U.S. 293, 299 n.5 (1949)). But

defining a relevant market is not an end unto itself; rather, it is an analytical tool used to

ascertain the “locus of competition.” Brown Shoe, 370 U.S. at 320-21; see also Merger

Guidelines § 4.1.1 (“[T]he purpose of defining the market and measuring market shares is to

illuminate the evaluation of competitive effects.”). Accordingly, the Supreme Court has

emphasized that market definition under the Clayton Act was intended by Congress to be “a

pragmatic, factual” analysis and “not a formal, legalistic one.” Brown Shoe, 370 U.S. at 336.

Market definition has two components: the relevant geographic market and the relevant

product market. See Brown Shoe, 370 U.S. at 324; Anthem, 236 F. Supp. 3d at 193. Here, the

parties agree that the relevant geographic market is the global market for the acquisition of U.S.

publishing rights. See Govt. PFOF-PCOL 4 125; ECF No. 1 (Complaint) 4 40; ECF No. 56

(Amended Answer) § 40 (agreeing “that Penguin Random House and Simon & Schuster compete

with each other and with many other publishers to acquire rights to publish books in the United

States and that authors who sell U.S. publishing rights can reside anywhere in the world.”). The

parties strenuously dispute, however, the boundaries of the appropriate product market.

The government defines the relevant product market as the one for publishing rights to

anticipated top-selling books. See Govt. PFOF-PCOL {ff 15, 63. Anticipated top-selling books

are those that are expected to yield significant sales, and for which authors therefore receive

higher advances. See id. | 15. The government contends that such books have distinctive

characteristics, including the need for extra marketing, publicity, and sales support to allow them

to reach broader audiences. See id. {J 15, 64-68, 87, 93-119.

The proposed market for anticipated top-selling books is a submarket of the broader

publishing market for all trade books. See id. J 124; see also Defs. PCOL {[ 9-10 (explaining

24

that the “market for the acquisition of a// U.S. trade books” is an appropriate, broader market).

Under the government’s monopsony theory, the authors of anticipated top-selling books are

“targeted sellers” against whom the merged defendants might lower the prices paid for the

authors’ wares, See Govt. PFOF-PCOL {ff 55-58, 69-76; see also Merger Guidelines § 4.1.4 (If

a monopsonist could “profitably target a subset of [sellers] for price [de]creases, the

[government] may identify relevant markets defined around those targeted [sellers].”); cf FTC v.

Wilh. Wilhelmsen Holding ASA, 341 F. Supp. 3d 27, 46-47 (D.D.C. 2018) (“[A Jntitrust markets

can be based on targeted customers”); Sysco, 113 F. Supp. 3d at 38-40 (discussing definition of

markets based on targeted customers). In the monopsony context, “[a] submarket exists when

[buyers] can profitably [cut] prices to certain targeted [sellers] but not to others, in which case

regulators may evaluate competitive effects separately by type of [seller].” Anthem, 236 F. Supp.

3d at 195 (cleaned up).

Courts evaluate relevant product markets in the monopsony context in two ways: by

considering qualitative, “practical indicia” as described by the Supreme Court in the Brown Shoe

case, 370 U.S. at 325; and by examining “supply substitution” and applying the “hypothetical

monopsonist test,” which are discussed in detail, infra. The parties in this case focus their

arguments on whether “practical indicia” support the finding of a market to publish “anticipated

top-selling books.” Because the parties choose to fight on the battlefield of “practical indicia,”

that is where the Court begins its analysis.

1, Practical Indicia

“TWlithin [a] broad market, well-defined submarkets may exist which, in themselves,

constitute product markets for antitrust purposes. The boundaries of such a submarket may be

determined by examining such practical indicia as industry or public recognition of the

25

submarket as a separate economic entity, the product’s peculiar characteristics and uses, unique

production facilities, distinct [sellers], distinct prices, sensitivity to price changes, and

specialized vendors.” Brown Shoe, 370 U.S. at 325, These indicia are “practical aids” as

opposed to “talismanic” criteria “to be rigidly applied,” FTC v. Swedish Match, 131 F. Supp. 2d

151, 159 (D.D.C. 2000) (cleaned up); thus, “submarkets can exist even if only some of these

factors are present.” TC v. Staples, Inc., 970 F. Supp. 1066, 1075 (D.D.C. 1997) (“Staples I’).

Brown Shoe’s practical indicia also may help identify a market of targeted sellers. See

FTC v. Whole Foods Market, Inc., 548 F.3d 1028, 1038-39 (D.C. Cir. 2008) (citing Brown Shoe,

370 U.S. at 325). For example, a market of “distinct [sellers],” as posited by the government,

may find “a particular [set of buyers] ‘uniquely attractive’” and “the only realistic choice” for

their products. /d. (first citing Brown Shoe, 370 U.S. at 325; then quoting Nat’! Collegiate

Athletic Ass'n y. Bd. of Regents of the Univ. of Okla., 468 U.S. 85, 111 (1984); and then quoting

SuperTurf, Inc. v. Monsanto Co., 660 F.2d 1275, 1278 (8th Cir. 1981)).

i. The $250,000 Threshold

To identify the books that are anticipated to sell well, the government focuses on the

criterion of “distinct pricing”: For analytical purposes, it defines anticipated top-selling books as

those for which publishers pay an advance of at least $250,000. See Govt. PFOF-PCOL § 64;

Brown Shoe, 370 U.S. at 325 (explaining that “distinct prices” are probative in market

definition); see also Whole Foods, 548 F.3d at 1038-39 (explaining distinct prices paid by

targeted group of customers “indicate[] the existence of a submarket of core customers”); Syufv

Enters. v. Am. Multicinema, Inc., 793 F.2d 990, 995 (9th Cir. 1986) (considering “lucrative terms

offered for the pictures by exhibitors” to define relevant market). Books that meet the $250,000-

advance threshold comprise only 2 percent of all book acquisitions, but they account for 70

26

percent of all advance spending, amounting to $1 billion annually. See Govt. PFOF-PCOL 4]

15, 68 (citing Trial Tr. at 1239:10—24 (Hill), 2904:17—2905:3 (Snyder)). Government’s Exhibit -

963 shows that the market shares of industry participants in the proposed publishing market for

anticipated top-selling books are far more concentrated than in the market for publishing books

at lower advance levels:

@Non-Big 5 =Penguin Random House aie)

100% +-

® Simon & Schuster Re ae

ee ee

90% +------ B-----~--o- > Pre ennren ate NTT.

20% Te ll

70% 4------

60%

50% von

40% denne.

Share of contracts (count)

30% jalmnaianmmaat

20% pono

10% +----++--

5% onenennnnnene tennnenp —

Anticipated Top Sellers Non-Anticipated Top Sellers

Source: Snyder Advance Data.

In the publishing market for anticipated top-selling books, the Big Five publishers hold 91

percent of the market share, while smaller publishers collectively hold only 9 percent. PX 963.

By contrast, in the publishing market for books that earn advances below $250,000, the non-Big

Five publishers have a much more substantial market share of 45 percent. Jd.

As an initial matter, the government’s use of high advances as a proxy for anticipated

book sales is logical and supported by market realities. In publishing, advances are correlated

with expected sales because books that are expected to sell well receive higher advances. See

supra Section IB. In fact, advance levels are set by using P&L’s, and the defining feature of a

P&L is the sales estimate. See id.; Trial Tr. at 917:13—16 (Tart) (“[T]he higher .. . sales equate

to a higher advance in the P&L.”). Moreover, industry practices indicate that $250,000 is a

reasonable place to draw the line: S&S and two of the three PRH adult divisions require

approval from senior publishers or executives for advance offers of $250,000 or more: and

Publishers Marketplace, a major industry publication, categorizes deals for $250,000 or more as

“significant.”!’ See Trial Tr. at 1233:5—-135 (Hill), 459:5—8 (Karp), 1993:1—3 (Kim), 914:22—

915:2 (Tart), 2261:12-2262:5 (McIntosh): PX 989 (as

public recognition” of a distinct category of books that receive advances at or above the

$250,000 level. Brown Shoe, 370 U.S. at 325.

The defendants take aim at the $250,000 threshold that the government has chosen to

bound the market. See Defs. PCOL *¥ 23-25, 28-41. Most significantly, they argue that the

$250,000 threshold is either too high or too low to define a submarket for anticipated top selling

books. Jd. {¥ 28-41; Defs. PFOF §§ 33-43. Specifically, the defendants rely on their Exhibit

438 to argue that the advance threshold should be set at $50,000 to capture the point at which the

Big Five begin to dominate the market for acquiring books:

ms As for other Big Five publishers, Hachette does not have a company-wide approval policy, but its different

imprints require approval for offers from above $100,000 to above $250.000. See Trial Tr. at 232:21-233:5

(Pietsch) (“All our publishers have advance approval levels and they are clustered right around [$250,000].”).

Hachette also tracks the books it lost for advances of $500,000 or more. See PX 790.

. See Trial Tr. at 1438:8—11 (Murray).

See id. at 1101:7—

1102:13 (Weisberg): DX 408.

Dr. Hill’s Revised Reply Report Figure 1 (2019-2021)

($0 - $49,999, $50,000 - $999,999, $1,000,000+)

mw Non-Big 5 @ Penguin Random House #@

a Simon & Schuster = a

Share of Contracts (Count)

$0 - $49,999 $50, 000 - $999,999 $1,000,000+

See Defs. PFOF § 37 (“[T]he data establish that j/competitive conditions differ based on market

shares and author preferences, the difference begins with books acquired for advances of

$50,000 or more,” where the market share of non-Big Five publishers is reduced from 58% to

17%.); Defs. PCOL ff 31-34. Adieraadeaks. the defendants contend that the threshold should be

set at $1 million to identify the books by celebrity, franchise, or onciutly authors that are

most clearly destined for success. See Defs. PFOF *] 37-40: Defs. PCOL {§ 35-41. Ifthe

relevant market were properly defined at the lower ($50,000) or higher ($1 million) advance

level, the defendants urge, the government could not show a sufficient decrease to competition or

harm to authors. See Defs. PFOF {ff 38, 42-43.

The defendants’ excessive concern over the specific dollar threshold betrays a

misunderstanding of why the threshold was chosen. The market that the government seeks to

define is the one for anticipated top-selling books, and the $250,000 demarcation was adopted

only as an analytical tool to help it group together the books in question. The govemmment’s

economic expert, Dr. Nicholas Hill, also conducted his analyses at other numerical thresholds

(including $150,000, $250,000, $500,000, and $1 million) and observed consistent outcomes at

those various high-dollar amounts. See PX 960: Trial Tr. at 1254:7—25 (Hill), 1259:2—12 (Hill),

1233:14-20 (Hill). Thus, the $250,000 cutoff is merely useful; it is not intended to be a rigid

bright line, but rather is helpful “[flor analytical purposes” to facilitate the assessment of anti-

competitive effects. FTC v. Staples, Inc., 190 F. Supp. 3d 100, 118 & n.10 (D.D.C. 2016)

(“Staples IT’) (“[T]here is no ‘magic place that’s the right place’ to draw the line.” (quoting

government expert’s testimony)). Accordingly, the Court rejects this argument against the

government’s defined market.

The Court is unswayed by the defendants’ tactic of enumerating other markets or

submarkets in which competition would not be harmed by the merger. In addition to proposing

submarkets at the $50,000- and $1 million- advance levels, the defendants also declare that the

government could not prove anticompetitive effects from the merger in the broad market of

publishing rights for all U.S. trade books, or in the downstream market for retail book sales. See

Defs. PFOF 9{ 29-31. Those protestations are beside the point because the Clayton Act

prohibits mergers that may substantially lessen competition “in any line of commerce or in any

activity affecting commerce.” 15 U.S.C. § 18 (emphasis added). Thus, even if alternative

submarkets exist at other advance levels, or if there are broader markets that might be analyzed,

the viability of such additional markets does not render the one identified by the government

unusable. See United States v. Cont’l Can Co., 378 U.S. 441, 456-58 (1964) (validating a

relevant product market of glass and metal containers, even though “there may be a broader

product market made up of metal, glass and other competing containers”); United States v.

Aluminum Co. of Am., 377 U.S. 271, 275 (1964) (explaining that even though insulated

aluminum conductor and insulated copper conductor could both be in “a single product market,”

that “does not preclude their division for purposes of [Section] 7 into separate submarkets”); see

also Brown Shoe, 370 U.S. at 325; Anthem, 236 F. Supp. 3d at 201-02 .

30

Ample precedent supports the government’s use of a numerical cutoff to identify a

submarket. It is common for courts to use seemingly arbitrary criteria to home in on a segment

of a broader industry. See Wilhelmsen, 341 F. Supp. 3d at 51 (market of customers with fleets of

10 or more global maritime vessels); Anthem, 236 F. Supp. 3d at 195 (market of companies with

5,000 or more employees); Staples IT, 190 F. Supp. 3d at 118 (market of customers who spend

$500,000 or more annually on office supplies). In Wilhelmsen, Judge Chutkan approved a

relevant market “defined around the FTC’s preferred set of targeted customers” — “Global

Fleets.” 341 F. Supp. 3d at 48, 58. The government characterized “Global Fleets” as “fleets of

10 or more globally trading vessels.” fd. at 51. Although the defendants argued “that the Global

Fleets construct is premised on arbitrary thresholds,” the court found that such fleets “are a

distinct group with distinct needs,” even though the “choice of ten globally trading vessels was

arbitrary in the sense that the number ten is not compelled by a specific market reality.” Jd. at

51-54. Judge Chutkan explained that the government’s expert “chose ten as a starting point for

developing a series of statistical estimates, the non-statistical implications of which support the

appropriateness of regarding Global Fleets as a distinct customer group.” /d. at 55. In other

words, the cutoff of ten ships to define “Global Fleets” was an appropriate analytical tool, just as

the choice of a $250,000-minimum advance level to define “anticipated top-selling books” is

appropriate for analytical purposes. At bottom, such “construct[s]” provide a “useful way to

discuss and predict economic conditions” because their “key aspects correspond to elements of

the existing marketplace that would make it possible to profitably target a subset of customers

[or sellers] for price increases [or decreases] post-merger.” Id. at 52 (quoting Sysco, 113 F.

Supp. 3d at 38).

31

The government’s focus on anticipated top-selling books also is consistent with cases in

which courts have recognized the “high end” of other broad markets as distinct submarkets for

antitrust purposes. See, e.g., Int’l Boxing Club of N.Y., Inc. v. United States, 358 U.S. 242, 251

(1959) (affirming district court’s conclusion “that nonchampionship fights are not ‘reasonably

interchangeable for the same purpose’ as championship contests” and explaining that defining

the relevant market “involves distinction in degree as well as distinctions in kind”); Whole

Foods, 548 F.3d at 1032 (recognizing relevant submarket of “premium, natural, and organic

supermarkets” that “generally target affluent and well educated customers”); O’Bannon v. Nat'l

Collegiate Athletic Ass'n, 7 F. Supp. 3d 955, 986-88 (N.D. Cal. 2014) (recognizing relevant

submarket of “elite football and basketball recruits”), rev’d in part on other grounds, 802 F.3d

1049 (9th Cir. 2015), Thus, the relevant market defined here falls comfortably within the

parameters set by numerous applicable precedents.

The defendants nevertheless fault the government for defining its submarket by “price

alone,” contending that any correlation between advance level and expected sales shows only

that books “are valued along a continuum.” See Defs. PCOL ff 12, 24-25 (emphasis in

original). They argue that the existence of “a spectrum of price or value” is insufficient to

establish a submarket and, accordingly, that the government’s market is not appropriately

defined. Id.'* Once again, such arguments overlook the purpose of the $250,000 threshold as an

18 In support, the defendants primarily rely on Jn re Super Premium Ice Cream Distribution Antitrust Litig.

691 F. Supp. 1262 (N.D. Cal. 1988), aff'd sub nom. Haagen-Dazs Co. v. Double Rainbow Gourmet Ice Creams,

Inc., 895 F.2d 1417 (9th Cir. 1990), and United States v. Oracle Corp., 331 F. Supp. 2d 1098 (N.D. Cal. 2004). The

Court finds both In re Super Premium Ice Cream and Oracle inapposite. Neither case concerns a market of targeted

sellers or buyers, as relevant to this case. Further, in Jn re Super Premium Ice Cream, the evidence showed that

consumers who bought higher-priced “super premium” ice cream could and would buy lower-quality ice cream as a

substitute. See 691 F. Supp. at 1268. Here, authors of anticipated top-selling books have no alternative to selling

their books to a publisher because, as addressed in Section L.C, self-publishing is not a realistic alternative. Because

they have no reasonable substitute, authors can be targeted for and impacted by a decrease in prices, in a manner that

ice-cream customers could not have been targeted by a price increase. As for Oracle, there, the government

32

analytical tool that facilitates the examination of market shares and anticompetitive effects. The

threshold number need not represent an exact point at which the market begins to distinguish a

product. See Wilhelmsen, 341 F. Supp. 3d at 54-55; Anthem, Inc., 236 F. Supp. 3d at 200

(accepting a 5,000-employee threshold to define “national accounts” even though the “threshold

may exclude some products that would meet the needs of smaller employers”); Staples IT, 190 F.

Supp. 3d at 118 & n.10 (“[T]here is no ‘magic place that’s the right place’ to draw the line.”

(quoting government expert’s testimony)). Rather, a threshold will necessarily represent a

“starting point” for “statistical estimates, the non-statistical implications of which support the

appropriateness of regarding” anticipated top-selling authors as a “distinct [seller] group” that

buyers can target. Wilhelmsen, 341 F. Supp. 3d at 55.

ii. The Remaining Brown Shoe Factors

Aside from distinct pricing, the government argues that the remaining Brown Shoe

factors demonstrate that there is a relevant submarket for the publishing rights to anticipated top-

selling books. See Govt. PFOF-PCOL 4 87-114, 117-121. The government contends that such

books have “peculiar characteristics and uses,” in that they require stronger marketing, publicity,

and sales support, which allow them to reach a broader audience of readers. Id. ff] 87 (quoting

Brown Shoe, 370 U.S. at 325), 93-95. In addition, authors of anticipated top-selling books are

“distinct sellers,” in that they (1) care more about their publishers’ reputation and services, which

ensure wider distribution of their books; (2) may receive more favorable contract terms than

other authors; and (3) face different competitive conditions, as demonstrated by the dominant

attempted to define a market of “high function” software and tried to use a minimum sale price of $500,000 to

identify such software. See 331 F. Supp. 2d at 1158. Yet there were several flaws in the data analyzed and

presented by the government’s expert. See id. at 1158-59. Nor did the government offer any other qualitative or

quantitative evidence to define the market. See id. Here, Dr. Hill’s data suffers from no similar flaws and the

government has marshalled evidence beyond the advance price to show practical indicia of a submarket for the

publishing rights to anticipated top-selling books.

33

market share of the Big Five (91%) in publishing anticipated top sellers. See id. {| 66, 93-114,

117-119. For all those reasons, the government argues, anticipated top-selling books are in a

different category from books that are expected to sell relatively few copies, and publishers can

target their authors for price decreases.

The defendants, however, insist that all books are in the same market. They argue that

books at all advance levels go through an identical editing, marketing, and distribution process;

that there is no difference in the personnel who handle such books; that the contracts for all

books are negotiated in the same way; and that any special terms in the contracts for some books

simply result from an agent’s leverage. See Defs. PFOF { 47-67; Defs. PCOL 4 21. Further,

they contend that publishers cannot predict which books will be top sellers. See Defs. PFOF 4

78 (“[P]ublishers generally have no objective criteria for reaching in advance a consensus on

whether a book is likely to be a top selling book.”), 79 (arguing that publishers “cannot easily

predict top sellers,” other than books by celebrity, franchise, or prize-winning authors), 75

(asserting that every book is individual and author atypical) (citing Trial Tr. at 1068:12—13

(Weisberg), 1952:20-25 (Duhigg)); Defs. PCOL 4 21.

The Court has no trouble recognizing that anticipated top-selling books are distinct from

the vast majority of books that do not carry the same expectations for success. Obviously, the

entire publishing industry is dedicated to selling books; and all editors and publishers naturally

are very focused on discovering and acquiring the books that they believe will drive sales.

Evidence strongly supports the conclusion that, from the perspective of editors and publishers,

not all books are created equal. Beyond advances, contracts for books that are expected to sell

well are more likely to include favorable terms like higher royalty rates, higher levels of

marketing support, “glam” packages (e.g., for hair, makeup, and wardrobe services), and airfare

34

for authors,'? See Trial Tr. at 988:2-8 (Dohle) (“Very rarely, if ever, will I negotiate the other

royalty rates, but if it were to happen, it would be at that very, very top tier advance level.”),

1132:17-23 (Weisberg) (“top end” authors can negotiate terms such as payment schedules,

bonuses, and glam budgets), 1819:9—1820:2 (Walsh) (customization of contract terms is

generally correlated with higher advances), 1828:8—18 (Walsh) (higher marketing commitments

are expected for higher advance books); DX 21 at 5 (agent demanding “a publisher ready to

commit incredible energy and resources”). Publishers print more of the books they think will do

well; circulate more advance copies of such books to reviewers or influencers to create

excitement; push for interviews with more media outlets; and schedule book-tour appearances in

more locations.” See Trial Tr. at 1373:12-1374:3 (Murray); PX 986. Anticipated top-selling

books also get more attention from marketing and sales teams.”! For example, Dr. Hill

determined that S&S and PRH spend, on average, under $10,000 on marketing for books with

advances under $250,000, and between $40,000 and $90,000 on marketing for books with

advances over $250,000:

19 For example, Crown, a former publishing division of PRH that later merged with the Random House

division, produced guidelines for marketing support based on expected sales and advance levels. See PX 986; Trial

Tr, at 2275:19-2278:1 (McIntosh). Under those guidelines, books with expected sales of more than 25,000 units or

advances of more than $150,000 were to receive a dedicated publicist, book tours with stops in 5 to 15 markets,

extensive national media engagement, prominent placement on PRH and partner websites, and targeted social media

pushes, See PX 986 at 2~4. By contrast, books with advances and sales under those thresholds were to have only a

contact in the publicity department, smaller book tours (if any), and limited media engagement. See id. at 5-9.

20 Some of the defendants’ witnesses testified that all books are anticipated to sell well. See Trial Tr, at

576:17-24 (“[A] good editor worries about every book that he or she acquires, making every book profitable.”),

1810:17-1812:4 (Walsh) (“T would say I always anticipate that what I am working on is going to be a best seller.”).

That assertion is not credible. Although an agent, author, or publisher may “hope” every book will be a hit, that is

not the same as anticipating or expecting that a book will do well. See id at 1813:4—6 (The Court: “You don’t

expect every single book you work on to be a best seller or top seller?” Walsh: “Right. I hope.”); see also id. at

328:2-10 (King) (explaining that he chose a smaller publisher for a book that “wasn’t a crafted best seller”),

593:21-594:8 (Karp) (recognizing that some books are “midlist” books that publishers are “hoping,” but not

anticipating, will be hits).

2 See Trial Tr. at 258:14—21 (Pande) (“I would say that there’s a pretty clear relationship between the level of

the advance and the amount of resources that the publisher invests in the marketing and publicity of the book.”),

490:13-492:4 (Karp) (“[T]he big obvious books that we spent a lot of money for, they definitely have to be

marketed and publicized aggressively.”), 1373:1—I1 (Murray), 2001:12-2002:4 (Kim); PX 989 (Putnam post-

publication P&L sheet) (showing general correlation between advance level and marketing spending),

35

$100,000

$90,000

$80,000

$70,000

$60,000

$50,000

$40,000

$30,000

$20,000

$10,000

$0 +

0 - $250K $250K - $500K $500K - $750K $750K - $1M

Advance bins

See PX 972.

The fact that the Big Five publish 91 percent of anticipated top sellers also supports a

finding that the authors of such books have unique needs and preferences. See PX 963.

Although smaller publishers can sometimes put out an anticipated top-selling book, it is the Big

Five who have the back lists and the marketing, publicity, and sales advantages necessary to

consistently provide the high advances and unique services that top-selling authors need. See

supra Section I.C (discussing Big Five’s publishing advantages). It is precisely those specialized

needs that make the authors of anticipated top-selling books vulnerable to targeting for price

reductions. Publishers of anticipated top-selling books know that such authors are not able to

find adequate substitutes for publishing their books because of their unique needs and

preferences. See id. Those publishers therefore can target authors of anticipated top-selling

books for a decrease in advances (prices) because it is not as likely that such a price decrease will

cause the publishers to lose a book. See Wilhelmsen, 341 F. Supp. 3d at 56—57 (finding targeted

buyer market where market was characterized by individual negotiations and customers had

unique needs and preferences); Staples I, 190 F. Supp. 3d at 127 (finding targeted buyer market

where industry recognized customers as a distinct group that needed specific prices and

services); see also Merger Guidelines § 4.1.4.

36

Although the defendants proclaim that no one in the industry uses the term “anticipated

top seller,” Defs. PFOF {jf 87-88, that does not mean that such books do not exist. See

Wilhelmsen, 341 F. Supp. 3d at 51-52 (rejecting defendant’s argument “that the definition of

Global Fleets does not accord with commercial reality, given that [defendants do not] use the

FTC’s definition of that term... .”); see also Le v. Zuffa, LLC, 216 F. Supp. 3d 1154, 1159,

1165-66 (D. Nev. 2016) (denying motion to dismiss that was based in part on defendant’s

argument that “Elite Professional MMA fighters” is not a term used in the industry). In fact,

market participants have other names for expected top sellers, such as “lead titles” or “priority

titles.”*? Regardless of nomenclature, clear evidence demonstrates that the practice of

identifying and giving special support to the books that will drive sales is common. The

government’s defined market thus reflects “commercial realities” in the publishing industry.

United States v. Grinnell Corp., 384 U.S. 563, 572 (1966). Le

The defendants’ position that individual publishers are unable to anticipate which books

will be top sellers is unsupportable. That contention is contradicted by the universal industry

practice of making a sales estimate for every single book before offering an advance, and

credible testimony that there is often consensus among editors and publishers about which books

will be popular with readers. See supra Section I.C; Trial Tr. at 2108:14-24 (Wylie) (“But I

think there are recognizable qualities in — in books that people who have been in the business

22 “Lead titles” or “priority titles” are expected to sell well and receive more significant marketing, publicity,

and sales support. See PX 986 (Crown intemal guidance identifying “lead titles” as books with a sales goal of

75,000 units or advances of $500,000, and advising increased marketing, sales, and publicity support for those

titles); PX 2005 at 24-27; Trial Tr. at 1071:4-1072:15 (Weisberg) (defining lead title as “top of the list”), 1988:19-

25 (Kim) (“So every season, we have two or three titles that we really designate as lead titles, titles that we feel we

really want the sales team to really love and read, books that we feel we want to put a lot of attention on and

marketing support for.”). They also generally receive high advances. See Trial Tr. 1071:4-1072:15 (Weisberg)

(“[I]f we spend a lot of money on a book and it’s a book that everybody loves, it becomes a lead title.”), 2268:9-

2269:22 (McIntosh) (“[I]f we had a really high sales expectation at time of acquisition and then by the time we’re

ready to publish the book, we still have a sales — a high sales expectation, then it would seem logical to me that —

that there could be a high advance attached to that.”).

37

for a long time would easily recognize.””). The defendants’ high share of the book-acquisition

market and their substantial profit margins strongly indicate that they are successfully choosing

books that people want to read. See PX 994; Trial Tr. at 781:3-5 (Dohle), 1492:2—3 (Hill). To

be sure, editors often offer a range of advances for any given book, and the defendants correctly

note that there are many examples of books that were unexpected best sellers, such as Stephen

King’s Carrie, or Marie Kondo’s The Life-Changing Magic of Tidying Up. See Defs. PFOF

79 79, 81. But it is commonplace for multiple editors to gravitate to the same book, as evidenced

by the routine occurrence of competitive auctions; and the defendants do not dispute that there is

a general correlation between author advances and book sales, see Trial Tr. at 749:4—22 (Dohle);

PX 151 at 11. That is strong evidence that the book-acquisition process is not random. Indeed,

whenever a publisher submits a bid of $250,000 or more for a book, that publisher has

determined that the book is likely to be a top seller and knows that the competitors for the book

are likely to be limited to the Big Five. See Trial Tr. at 153:10—-13 (Pietsch) (other Big Five

publishers are Hachette’s main competitors for books with advances over $250,000); PX 530 at 2

(Big Five publishers are S&S’s “biggest competitors . . . since they are the most likely to come

up with high advance payments required ....”), These practical indicia in the publishing

industry strongly support the existence of the identified relevant market.

One high-end submarket case that the Court finds highly relevant is Syufy Enterprises v.

American Multicinema, Inc. In Syufy, the Ninth Circuit upheld a relevant submarket “for [the]

exhibition of industry anticipated top-grossing motion pictures in the San Jose area.” 793 F.2d at

994, Anticipated blockbusters, the court explained, “are identifiable . . . on the basis of such

criteria as national advertising support, longer playtimes, guaranteed rentals, famous stars,

directors and producers, booking in first class theatres, and lucrative terms offered for the

38

pictures by exhibitors.” Jd. at 994-95. Those indicia are analogous to some of the features of

anticipated top-selling books, such as: more substantial marketing, publicity, and sales support;

authors who are prominent or have a track record of success; and higher advances. Moreover,

the appellant in Syufy challenged the existence of the market for “anticipated top-grossing

motion pictures” by making arguments similar to those pressed by the defendants here, insisting

that the market was “ex post facto and ad hoc,” that “all first run films are in substantial

competition with each other,” and that such films “possess no special characteristics that

differentiate them from less successful films from an ex ante perspective.” Jd. at 994. This

Court joins the Ninth Circuit in rejecting such arguments. As discussed, distinctive

characteristics set anticipated top-selling books apart from the rest of the pack.

In sum, this case demonstrates that “[w]hatever the market urged by the [government],

the other party can usually contend plausibly that something relevant was left out, that too much

was included, or that dividing lines between inclusion and exclusion were arbitrary.” FTC v.

Tronox Ltd., 332 F. Supp. 3d 187, 202 (D.D.C. 2018) (quoting 2B Phillip E, Areeda & Herbert

Hovenkamp, Antitrust Law § 530d (4th ed. 2014) [hereinafter Areeda, Antitrust Law]). Yet

“It]he Supreme Court has wisely recognized there is ‘some artificiality’ in any boundaries, but

that ‘such fuzziness’ is inherent in bounding any market.” Jd. (quoting Areeda, Antitrust Law

{ 530d); Anthem, 236 F. Supp. 3d at 193 (“The ‘market,’ as most concepts in law or economics,

cannot be measured by metes and bounds.” (quoting Times-Picayune Publ’g Co. v. United

States, 345 U.S. 594, 611 (1953)). Market definition is more art than science, see RAG-Stiftung,

436 F. Supp. 3d at 312-13, and it is critical to remember that the goal of the exercise is to enable

and facilitate the examination of competitive effects. See Brown Shoe, 370 U.S. at 320-22;

39

Cont’l Can , 378 U.S. at 452-55. In this Court’s view, the government has easily cleared the

bar.

2. Supply Substitution

The traditional way to define a relevant market in the monopsony context would be to

examine “the commonality and interchangeability of the buyers” of a certain good. Todd v.

Exxon Corp., 275 F.3d 191, 202 (2d Cir. 2001). Indeed, “‘the outer boundaries of a product

market are determined by the reasonable interchangeability of use or the cross-elasticity of

[supply] between the product[’s buyers, in the case of monopsony,]| and the substitutes for [such

buyers].’ Accordingly, the touchstone is [supply] substitution ....” See Wilhelmsen, 341 F.

Supp. 3d at 45 (quoting Brown Shoe, 370 USS. at 325).

To test the proposed market boundaries, courts commonly turn to the “hypothetical

[monopsonist] test.” Sysco, 113 F. Supp. 3d at 38. The hypothetical monopsonist test “ensures

that markets are not defined too narrowly,” on the theory that if the test identifies substitute

buyers for the product in question, such buyers should be included in the market. See Merger

Guidelines § 4.1.1 (describing hypothetical monopolist test). The hypothetical monopsonist test

assumes that there is only one buyer in the proposed market and asks whether that hypothetical

buyer, freed from price regulation, “could profitably target a subset of [sellers] for price

[decreases].” Sysco, 113 F. Supp. 3d at 38 (quoting Merger Guidelines § 4.1.4). If such a

hypothetical monopsonist could profitably impose what economists call a “small but significant

23 To define a market around a targeted seller, sellers must not only be identifiable by buyers for differential

pricing but also must be unable to engage in arbitrage or opportunistic re-selling. See Staples I, 190 F. Supp. 3d at

117-118; Merger Guidelines § 3. The foregoing discussion establishes that anticipated top-selling books are subject

to differential pricing. Authors of those books also cannot realistically engage in arbitrage by selling their books to

a third party who would then sell the books to publishers for a better price. See Trial Tr. at 1230:7-23 (Hill). As Dr.

Hill testified, publishers would still need to read the book or proposal and value it in the same manner as if the book

were submitted directly by the author. See id.

40

and non-transitory [decrease] in price” of at least five percent in the proposed market, that

indicates the existence of a relevant market. Jd. at 33-34 (quoting Merger Guidelines § 4.1.4).

Here, the government includes all publishing firms in the market to acquire the

publishing rights for anticipated top-selling books. See Govt. PFOF-PCOL {| 79. Applying

Judge Mehta’s explanation of the test to the instant facts, we arrive at the following analysis:

If enough [authors] are able to substitute away from [selling their books to]

the hypothetical [publisher monopsonist] to another [way of distributing

their books] and thereby make a [decrease in advances] unprofitable, then

the relevant market cannot include only the [publisher monopsonist] and

must also include the substitute [method of distribution]. On the other hand,

if the hypothetical [publisher monopsonist] could profitably [lower

advances to authors] by a small amount, even with the loss of some

[authors], then economists consider the [publishers] to constitute the

relevant market.

See Sysco, 113 F. Supp. 3d at 33; see also Merger Guidelines § 12 (“In defining relevant markets

[in buy-side cases], the Agencies focus on the alternatives available to sellers in the face of a

decrease in the price paid for by a hypothetical monopsonist. Market power on the buying side

of the market is not a significant concern if suppliers have numerous attractive outlets for their

goods or services.”).

The government’s expert, Dr. Hill, estimated what “actual diversions” would be for the

defined market, i.e., the percentage of authors who would switch to self-publishing in the face of

a “small but significant and non-transitory [de]crease” in advances paid for anticipated top-

selling books. He found that even if some small number of authors switched to self-publishing,

it would be profitable for publishers to decrease advances — that is, the defection of authors in

response to the lowered advances would be far less than what would be necessary to make the

decrease unprofitable. See Trial Tr. at 1245:14—-1246:9 (Hill).

41

The defendants do not dispute that the relevant market of “publishing rights to anticipated

top-selling books” passes the hypothetical monopsonist test. See Trial Tr. at 2897:18-2898:18

(Snyder). They instead argue that the test is inapposite here because it does not address the

alleged arbitrariness of the $250,000 threshold for bounding the market, see id.; Defs. PFOF

{ 44; indeed, submarkets at all but the lowest advance thresholds should pass the hypothetical

monopsonist test because self-publishing generally is a poor substitute for the services of an

established publisher. See Trial Tr. at 2898:8-18 (Snyder) (“[S]elf-publishing is not a relevant

constraint.”); supra Section I.C (further detailing inadequacy of self-publishing). The

incongruence of the hypothetical monopsonist test here is not surprising because it examines

substitutes for the buyers in the market, while the government’s proposed market is one of

“targeted sellers”: In this case, the test and the market-definition dispute are focusing on

different sides of the market.2* Although the Court agrees that the hypothetical monopsonist test

sheds no light on the contested issues in this case, it is sufficient to note for present purposes that

the test is a standard analytical tool in merger cases; and that it concededly supports the

government’s definition of the relevant market. Defendant’s other objections to the relevant

market have been addressed supra.

24 For this reason, the defendants’ argument that the government has not defined the “narrowest market,” as

required by some case law, lacks merit. See Defs. PCOL §] 35-41 (citing Merger Guidelines § 4.1.1; RAG-Stiftung,

436 F. Supp. 3d at 292; Sysco, 113 F. Supp. 3d at 26-27; United States v. H & R Block, Inc., 833 F. Supp. 2d 36,

58-60 (D.D.C. 2011); FTC v. Arch Coal, Inc., 329 F. Supp. 2d 109, 120 (D.D.C. 2004)). The cases relied upon by

the defendants focus on defining the market by reference to demand substitution and applying the hypothetical

monopolist test, while the instant case concerns a market defined by targeted sellers as articulated by the Merger

Guidelines. See Merger Guidelines § 4.1.4 (“If prices are negotiated individually with customers, the hypothetical

monopolist test may suggest relevant markets that are as narrow as individual customers... . Nonetheless, the

Agencies often define markets for groups of targeted customers, i.e., by type of customer, rather than by individual

customer. By so doing, the Agencies are able to rely on aggregated market shares that can be more helpful in

predicting the competitive effects of the merger.”).

42

B. Prima Facie Case

1, Market Concentration

Once the relevant market has been established, the next step is straightforward: “[T]he

government must show that the merger would produce ‘a firm controlling an undue percentage

share of the relevant market, and would result in a significant increase in the concentration of

firms in that market.’” See Heinz, 246 F.3d at 715 (quoting Phila. Nat'l Bank, 374 U.S. at 363

(alterations omitted)). Market concentration is fundamental to merger analysis. “That

competition is likely to be greatest when there are many sellers, none of which has any

significant market share, is common ground among most economists, and was undoubtedly a

premise of congressional reasoning about the antimerger statute.” Phila. Nat’ Bank, 374 U.S. at

363 (cleaned up); see also Heinz, 246 F.3d at 715 (“[W]here rivals are few, firms will be able to

coordinate their behavior, either by overt collusion or implicit understanding, in order to restrict

output and achieve profits above competitive levels.” (cleaned up)). Thus, demonstrating post-

merger “‘undue’” market concentration “establishes a ‘presumption’ that the merger will

substantially lessen competition.” Heinz, 246 F.3d at 715 (first quoting Phila. Nat’l Bank, 374

U.S. at 363; and then quoting Baker Hughes, 908 F.2d at 982). In Philadelphia National Bank,

the Supreme Court held that a significant change in concentration that results in a combined

market share of at least 30 percent is sufficient to establish the legal presumption that a merger

violates Section 7, 374 U.S. at 331, 364 (merger to a 36% market share with the top four banks

controlling a combined 78%); see also Hosp. Corp., 807 F.2d at 1384 (determining that FTC’s

finding that transaction was unlawful was supported by substantial evidence where defendant's

market share was raised from 14% to 26% and the market share of the four largest firms from

79% to 91%); of. Heinz, 246 F.3d at 711, 715~17 (holding FTC established presumption through

43

statistics about the change in market concentration where defendants would have a combined

market share of 32.8%).

The government’s expert, Dr. Hill, calculated market shares based on a comprehensive

set of data from more than sixty publishers. See Trial Tr. at 1251:12—1252:3 (Hill). According

to his calculations, the merging firms account for nearly half (49 percent) of the publishing

market for anticipated top-selling books, and the newly constituted “Big Four” that would

emerge after the deal would control approximately 91 percent. Trial Tr. at 1254:3—6 (Hill).

Government’s Exhibit 959 graphically depicts the post-merger market shares:

60%

SON ss sen ESR RSME SS ERS SRS Hi goroverepesonsencewnnncnns

40% }- c wenn eset e ence ne ent e tne eee ee ene tea ne eee teen nae en eee n een ee nena nner enn aera tener ne eeee ee eeeee eed

30% | -MMMMMM -- 222 nner ec cnn cee nec eceenennnneeenenn ae nent ene ennnnnn nn ee cece nene nn entec canst aenencneessea ey

20% }- === IID, ~~ === anne ences ence ene ce ccc ene nce nent nectar cnc e nent see neneened

10% | - ~~" "pppgggg ------222neeeeenn ose eneeene cece cnc e ee ene nen e cece cee nnnnne eee nce cennnnnneneeneecennned

0% ~~, oe

Po a

Random House

The second-largest market participant post-merger would oo A a with 24 percent of

the market, while ees and i would have 10 percent and 9 percent, respectively.

See PX 959. The non-Big Four would have the remaining 9 percent. See id. Dr. Hill also

calculated market shares using different advance thresholds to bound the relevant market and

found similar results. See PX 960. The post-merger market shares undoubtedly portray a highly

concentrated market dominated by four main players, with the leading, merged company holding

an “undue percentage share.” The 49-percent share that the post-merger PRH would hold is far

above the levels deemed too high in other cases. See, e.g., Phila. Nat'l Bank, 374 U.S. at 364

(36%); of Heinz, 246 F.3d at 711, 715—17 (32.8%). The substantial market share of the

proposed combined entity justifies a strong presumption of anticompetitive effects. See Baker

Hughes, 908 F.2d at 991 (“The more compelling the prima facie case, the more evidence the

defendant must present to rebut it successfully.”). Moreover, the high concentration must be

considered in the context of an undeniable trend in consolidation in the publishing industry. See

United States v. Pabst Brewing Co., 384 U.S, 546, 552-53 (1966) (“[A] trend toward

concentration in an industry, whatever its causes, is a highly relevant factor in deciding how

substantial the anti-competitive effect of a merger may be.”).

The post-merger market also would be unduly concentrated under the Herfindahl-

Hirschman Index (“HHT”), a measure commonly used to evaluate market concentration. See

RAG-Stiftung, 436 F. Supp. 3d at 310 n.26 (explaining calculation of HHI). The HHI is a

formula “used to estimate the competitiveness of the market on the basis of the number and size

of the firms.” Areeda, Antitrust Law J 930a. It provides a “short cut to establish a presumption

of anticompetitive effect through statistics about the change in market concentration.” AT&T,

Inc., 916 F.3d at 1032. “The HHI estimates market concentration by summing the squares of the

market share of every firm in the market. ... When one assesses the competitive impact ofa

merger, the important numbers are (1) the post-merger HHI; and (2) the amount the merger

increases the HHI.” Areeda, Antitrust Law § 930a (emphasis in original). An increase in the

45

index above certain levels “establish[es] the [government’s] prima facie case that a merger is

anti-competitive.” Heinz, 246 F.3d at 716. Under the Merger Guidelines, if an acquisition (1)

increases the HHI of a relevant market by more than 200 points and (2) results in a post-

acquisition HHI exceeding 2500, it is presumptively anticompetitive. See Merger Guidelines

§ 5.3; see also Staples H, 190 F. Supp. 3d at 128; H& R Block, 833 F. Supp. 2d at 71-72

(enjoining transaction that would have given the combined firm only a 28.4 percent market share

because the transaction would have resulted in an increase in the HHI of more than 200 and a

post-acquisition HHI that would have exceeded 2500), Here, the post-merger HHI would be

3,111, with an increase of 891, well above the thresholds required to trigger the presumption

under the Guidelines. Trial Tr. at 1256:24-1258:11 (Hill), 1259:4-12 (Hill).”°

Based on the market-share analysis and the HHI analysis, the government has met its

burden to establish that the proposed merger between PRH and S&S would produce “a firm

controlling an undue percentage share of the relevant market, and [would] result[] in a significant

increase in the concentration of firms in that market.” Phila. Nat’! Bank, 374 U.S. at 363. That

showing alone is sufficient to establish a prima facie case. Heinz, 246 F.3d at 716 (“Sufficiently

large HHI figures establish the FTC’s prima facie case that a merger is anti-competitive.” (citing

Baker Hughes, 908 F.2d at 982-83 & n.3)). Notably, the defendants do not question the

accuracy of Dr. Hill’s market-share calculations, nor his application of the HHI.

The government further notes that the market shares reflect the actual competitive

dynamics in the market. See Govt. PFOF-PCOL ff 135-97. Dr. Hill compiled several different

data sets to evaluate how frequently the merging parties compete against each other and against

35 Moreover, post-merger HHIs (and the post-merger increase) also are above the presumption thresholds if

the relevant market is defined using a variety of other advance cutoffs (e.g., $150,000, $350,000, $500,000, or $1

million). Trial Tr. at 1254:7-25 (Hill); see also PX 960.

46

other publishers. As discussed in more detail below, see infra, Section ILB.2.i, the data tracks

the instances where the merging parties lost books to one another, and where they were “runners-

up” to each other in book acquisitions. Dr. Hill’s analysis of the data reveals that, as market

shares would predict, the Big Five in fact dominate book acquisitions in the relevant market.

Consistent with their market shares, when S&S loses a book, it most often loses to PRH; and

when S&S wins a book, its most likely runner-up is PRH. See Trial Tr. at 1282:15-24 (Hill)

(indicating that for books acquired by S&S, PRH was the runner-up approximately 60% of the

time); PX 970 (showing diversion ratios); see also Trial Tr. at 2927:17-2928:4 (Snyder) (PRH is

also the most frequent runner up to S&S according to Professor Snyder’s data set). Moreover, an

independent deal tracker maintained by Hachette for acquisitions above $500,000 also depicts

results consistent with market shares: Roughly 90 to 95 percent of Hachette’s losses in that

advance range were to other Big Five publishers. See PX 790 (tracking Hachette’s losses to

these publishers); Trial Tr. at 191:16-194:24 (Pietsch).

Also significant is the stability of the market shares held by the primary market

participants over time. Based on Dr. Hill’s comprehensive data set, which included information

from approximately 1,200 book contracts per year, the market shares of the Big Five in acquiring

anticipated top-selling books has remained stable for the past three years. See PX 967.

Furthermore, the Big Five’s market shares versus the non-Big Five have also been consistent:

The data demonstrate that the aggregate market share of non-Big Five publishers has been

“essentially flat.” Trial Tr. at 1482:15-25. This stability suggests that more weight should be

assigned to market shares, see Merger Guidelines § 5.3 (“The Agencies give more weight to

market concentration when market shares have been stable over time . . . .”), and thus reinforces

the presumption of anticompetitive effects based on market concentration.

AT

2. Other Evidence

The government does not rely solely on the high degree of market concentration that

would result from the merger, and the attendant presumption of anti-competitive harm; instead,

the government also “bolster[s] its prima facie case by offering additional evidence.”

Wilhelmsen, 341 F. Supp. 3d at 59. The government presents evidence that (1) the merger will

cause anticompetitive effects from the elimination of competition between PRH and S&S, and

(2) the higher concentration in the post-merger market will increase the risk of coordinated

anticompetitive conduct by the largest publishers. See Govt. PFOF-PCOL {J 135-97.

i. Unilateral Effects

Mergers necessarily eliminate the competition between the merging companies,

See Heinz, 246 F.3d at 717. The government contends that PRH and S&S currently compete

“fiercely” to publish anticipated top-selling books, and that eliminating direct competition

between them is likely to harm authors. Govt. PFOF-PCOL § 244. Indeed, “[c]ourts have

recognized that a merger that eliminates head-to-head competition between close competitors

can result in a substantial lessening of competition.” Sysco, 113 F. Supp. 3d at 61; see also

Wilhelmsen, 341 F. Supp. 3d at 59. “Unilateral effects” are those that result directly from the

elimination of competition between the merging parties. Anthem, 236 F. Supp. 3d at 216. As

explained by the Merger Guidelines, “[a] merger can enhance market power simply by

eliminating competition between the merging parties. This effect can arise even if the merger

causes no changes in the way other firms behave.” Merger Guidelines § 1. Unilateral effects

may be especially acute in a “highly concentrated market.” Staples I, 970 F. Supp. at 1083.

48

a. Head-to-Head Competition

The analysis of unilateral effects focuses on how closely the merging firms currently

compete, in order to extrapolate the effects of eliminating that competition. See Merger

Guidelines § 6.2. Evidence in the record demonstrates that PRH and S&S are close competitors

for anticipated top-selling books. Specifically, PRH is the publisher against which S&S

competes the most frequently and to which S&S loses the most. See Trial Tr. at 595:23-25

(Karp) (agreeing that PRH is the “publisher [S&S] bid[s] against the most”), 1280:17—1281:17

(Hill) (reviewing win/loss data showing that S&S loses to PRH about 60% of the time).

Meanwhile, S&S is a significant competitor to PRH, see id. at 2360:20-23 (McIntosh), 1275:25—

1276:6 (Hill), and makes a particularly strong showing in biographies, memoirs, political

nonfiction, and books about current events, see id.. at 454:23-455:3 (Karp), 455:8-11 (Karp);

PX 326 at 2 (“S&S has political bestseller chops like no other right now.”).

The government’s expert, Dr. Hill, conducted a variety of economic analyses that assess

how closely PRH and S&S compete. Dr. Hill used four different methods to calculate “diversion

ratios,” which measure head-to head competition between the merging parties by asking the

following question: If one merging party lowered advance levels, what percentage of its authors

would “divert” their business to the other merging party, as opposed to diverting to other firms in

the industry? A higher diversion ratio indicates that the merging parties are close competitors

and that the merger is more likely to lead to harm. See Trial Tr. at 1274:2-12 (Hill); see also

Merger Guidelines § 6.1.

Dr. Hill calculated diversion ratios based on: (1) diversion proportional to market shares,

which is the largest data set; (2) win/loss data, which examines which publishers the merging

parties lose to the most often; (3) runner-up data, which shows how often the other party was the

A9

“runner-up” when one of the merging parties won an acquisition; and (4) minutes from the

parties’ editorial meetings, which provide a window into how frequently one merging party bid

on a book and lost to the other party. Recognizing that each methodology has limitations, Dr.

Hill performed multiple tests “to get a holistic understanding of what diversion might look like.”

Trial Tr. at 1294:20-1295:4. All the methodologies employed by Dr. Hill pointed to the same

conclusion: that PRH is S&S’s closest competitor, and that S&S is a significant competitor to

PRH. Specifically, Dr. Hill’s diversion ratios indicate that if PRH lowered advances, between 19

and 27 percent of its authors would divert to S&S; and that if S&S lowered advances, between

42 and 59 percent of its authors would divert to PRH. The government summarized the results

of the four studies as follows:

Figure 7. Summary of Dr. Hill’s Diversion Estimates (PX-97 0)

Type of Analysis Diversion from Diversion from

PRH to S&S S&S to PRH

Diversion according to share 19% 42%

Win/loss data 19% 59%

Runner-up study 27% 59%

Editorial minutes 21% 54%

See Govt. PFOF-PCOL 4 274 (citing PX 970).

The defendants’ expert, Professor Snyder, calculated his own diversion ratios, using a

less reliable data set assembled from the records of eighteen agents who responded to subpoenas

(“agency data”).*° Although Professor Snyder’s ratios were lower, he also found that PRH is

26 Professor Snyder’s agency data is less comprehensive than Dr. Hill’s data set. The eighteen literary

agencies that provided information in response to subpoenas are only a subset of the agents in the industry. See

Trial Tr. at 2657:7—25 (Snyder). Of the 973 contracts examined by Professor Snyder, from 2018 to 2021, only 360

earned advances per title of $250,000 or more. See id. Of those 360 contracts, Professor Snyder could not identify a

clear runner-up in 61 instances. See Trial Tr. at 2658:12-15 (Snyder). The remaining pool of data, relatively small

and unrepresentative as it was, see Trial Tr. at 1289:22—1292:12 (Hill), indicated that PRH and S&S were the winner

50

S&S’s closest competitor. Professor Snyder determined that the diversion ratio from PRH to

S&S is 20 percent, and the diversion ratio from S&S to PRH is 27 percent. Trial Tr. at 2927:4-

25 (Snyder).

The competition between PRH and S&S benefits authors by increasing advances paid for

their books, and industry participants predict that the loss of that competition would be harmful

to authors. Kensington’s CEO, Steven Zacharius, testified, “I personally would expect that

[advances] would go down since there will be less competition for those authors.” PX 2000 at 3.

Macmillan’s CEO Don Weisberg testified, “My guess is less competition will . . . long-term

probably bring the advance levels down.” See Trial Tr. at 1085:3-24. Agent Ayesha Pande

testified, “I think overall [the merger] will limit the choice, the number of editors and imprints

and publishing houses that would . . . be a good home for my clients... . And I believe overall

advances for my clients would be suppressed.” See id. at 295:3-16,7”

The merger would cause an inarguable loss of competition from the elimination of

situations where PRH and S&S would have been the top two or the only two bidders for an

anticipated top seller. Dr. Hill calculates that this should happen in approximately 12 percent of

and runner-up in only 7 percent of the cases, while market shares would have predicted that they would be winner

and runner-up in 12 percent of the cases. See Trial Tr. at 1588:7-19 (Hill), 2797:20-2798:6 (Snyder). Notably,

Professor Snyder’s estimate of diversion from S&S to PRH is based on a sample of only 22 books over four years,

the smallest sample of all the data sets used to estimate diversion. Trial Tr. at 1291:15-1292:12 (Hill), 1707:20-—

1708:1 (Hil); PX 996 at 1.

27 Defendants presented testimony to the contrary, suggesting that the lost competition between PRH and

S&S would not harm authors or their advances; and that it instead might lead to an increase in advance levels. See

Defs. PFOF ff 25-26 (suggesting that savings from the merger would allow the combined company to spend more

money to acquire books, which in turn would force competitors to offer higher advances), 115-116 (stressing that

PRH and competitors have no plans to lower advances). For example, S&S CEO Jonathan Karp testified that the

company has no plans to decrease author advances or reduce title count post-merger, see Trial Tr. at 583:13-19; and

PRH Head of Global Mergers & Acquisitions Manuel Sansigre did not consider the potential for reduced author

compensation when projecting the merger’s efficiencies for PRH, see id. at 2532:25-2533:12. In addition,

HarperCollins CEO Brian Murray stated that his company had not discussed author advances decreasing due to the

merger, see id. at 1407:24-1408:12; and that

, see id, at 1447:3-19, 1452:9-14. The Court finds testimony that the merger

will have either no effect or positive effects on advances incredible. The Court instead credits the much stronger

evidence that advances will decrease after the merger, based on the market-share data, economic analyses, and the

more credible testimony regarding market dynamics discussed supra.

51

book transactions based on market share, while Professor Snyder calculates that it happened only

6 to 7 percent of the time in his data set. See id. at 1588:7-19 (Hill), 2797:20-2798:6 (Snyder).

The government’s evidence included 27 summaries of competitive episodes, over three and a

half years, in which PRH and S&S drove up advances through direct, head-to-head competition.

See id. at 660:16—25 (Porro), 664:8-10 (Porro), 686:9-22 (Porro). For example, as the only two

bidders for one book, PRH and S&S drove the advance offered from $6 million to $8 million.

See PX 958-B, As the last two bidders for another book, PRH and S&S drove the advance

offered from $685,000 to $825,000. See PX 941-B. The loss of such head-to-head match-ups

undoubtedly would harm the authors whose advances would have been bid up by the direct

competition. See generally Merger Guidelines § 6.2. The defendants argue, however, that the

incidence of harm would be too infrequent to be considered substantial. See Defs. PCOL {J 58-

65.

Even when the merging parties were not the top two bidders, S&S’s participation

strengthened competition across all auction formats — round-robin, best-bid, and hybrid.

Hachette CEO Michael Pietsch testified that a larger number of bidders leads to “more upward

pressure” so that “in general . . . the price paid at auction can increase because of the number of

participants.” See Trial Tr. at 181:7—11 (Pietsch). Dr. Hill confirmed that when a large number

imprints participate in an auction, all of them understand that they need to be more aggressive in

their bidding to prevail. See id, at 1268:2-8 (Hill) (“So this is a correlation between when you

have a large number of bidders, you may need to be more aggressive in your bidding.”);*8 see

28 Dr. Hill gave two examples that demonstrate how the number of bidders influences a publisher’s bidding

strategy in best-bids auctions. In a 2019 best-bids auction for a young adult novel, an S&S editor wrote that,

because there were only three bidders, “I think we can be more guarded in our bidding.” See Trial Tr. at 1267:15—

22 (Hill). In a 2020 best-bids auction for a book by a musician, a PRH editor wrote, “Another editor and I discussed

bringing our offer significantly down yesterday based on the sense J got from the agent that she doesn’t have many

interested bidders.” See id. at 1267:23-1268:1 (Hill).

52

also Anthem, 236 F. Supp. 3d at 220-21 (“reducing the number of national carriers from four to

three is significant” because of its likely effect on bidding behavior). A higher number of

bidders also increases the chances that an author will receive an “outlier” high bid. A book’s

perceived value may vary significantly among different editors and publishers, and an unusually

high bid for a book is likelier when there are more bidders. See Trial Tr. at 601:20-25 (Karp),

1305:18-1306:3 (Hill), 2109:3-21 (Wylie). In one notable example, one bidder offered an

advance four times higher than the next closest bidder, reflecting the winner’s unique view of the

book’s potential. See id. at 2931:16-2933:19 (Snyder). The loss of S&S as an independent

bidder would weaken bidding incentives and reduce the frequency of events like these.

As previously noted, competition among publishers influences advances even in

individual negotiations between an agent and one publisher. See supra Section I.C. That is

because publishers know that agents can shop the book to other publishers if the publisher’s offer

is not high enough. See id. Therefore, the loss of PRH as an outside competitor would weaken

authors’ leverage in one-on-one negotiations with S&S, and the loss of S&S as an outside

competitor would weaken authors’ leverage in one-on-one negotiations with PRH. This

conclusion is consistent with Dr. Hill’s expert testimony, see Trial Tr. at 1270:13-1271:18 (Hill),

as well as the Merger Guidelines. See Merger Guidelines § 6.2 (“A merger between two

competing [buyers] prevents [sellers] from playing those [buyers] off against each other in

negotiations.”).

Finally, the evidence suggests that the acquisition of S&S would reduce PRH’s

motivation to compete for publishing rights. PRH CEO Markus Dohle testified that there are

two ways to increase market share in the industry: publish more successful books or acquire

other companies that publish successful books. See Trial Tr. at 801:18-23. PRH has most

53

recently pursued a strategy of bidding more aggressively and acquiring more “big books” to

organically increase its market share. See id. at 800:15-801:3 (Dohle), 2259:5—-20 (McIntosh).

The acquisition of S&S would give PRH an alternative means of increasing its market share that

would remove the pressure on PRH to acquire more books. See id. at 802:11-18 (Dohle). Thus,

accomplishing its goal of increasing market share through the merger would cause PRH to bid

less aggressively for books than it otherwise would. See id. (Q: “After this merger, Penguin

Random House will not have as strong a need to grow its share?” Dohle: “Yes.”).

b. Economic Models

Dr. Hill used economic models to attempt to quantify the expected harm to authors from

the merger. He conceded that the models are imprecise and do not perfectly reflect the way

books are acquired in the publishing industry; but he performed the analyses to glean additional

information about the likelihood of anticompetitive harm. See id. at 1653:25—1654:9 (Hill). Dr.

Hill’s primary model predicts that the merger would cause advances for PRH authors to decrease

by about 4 percent (or $44,000); and would cause advances for S&S authors to decrease by 11.5

percent (or $105,000). See id. at 1312:10-20 (Hill); PX 964 at 1-2. Although the defendants

challenge the applicability of the models and some of the inputs used by Dr. Hill, they fail to

convince the Court that the models are worthless. The economic models generally corroborate

the other evidence in the record that author advances would decrease in the wake of the merger.

Dr. Hill primarily relied on a “second-score auction” model to quantify the merger’s

potential harm to authors. See id. at 1295:21-22 (Hill). The model assumed that all book rights

are allocated using auctions, see id. at 1298:13-1302:19 (Hill); and it used a market-share input

to estimate how often the merging parties would be the top two bidders in an auction, see id. at

1305:9-17 (Hill). The model used an input of variable-profit margins to estimate the variation

54

among the bids: Using higher variable-profit margins generally would cause the model to

predict greater harm, i.e., a bigger decrease in advance level? See id. at 1305:18-1306:3 (Hill).

To measure the effect of the merger, the model looked at the instances when the merging parties

would be the top two bidders and then eliminated the second-highest bid, thereby making the

third-place bid the one that would set the amount of the advance (on the theory that the winning

bid would now only need to beat the third-place bid). See id. at 1303:7-22 (Hill).

The defendants argue that Dr. Hill’s second-score auction model is flawed because (1) it

inaccurately assumes that all book transactions involve auctions, and (2) Dr. Hill used the wrong

input for variable-profit margins. See Defs. PFOF 215; Defs. PCOL {ff 106-121; ECF No. 183

(Defs. Objections to Govt. PFOF-PCOL) at 29-36. Although the defendants are correct that the

model does not precisely reflect how book contracts are allocated among publishers, its exclusive

reliance on auctions is a reasonable simplification. See Trial Tr. at 1296:16-21 (Hill) (explaining

why he interprets the model more broadly). The market-share data captures the rate at which the

parties are winning book contracts — through negotiations, auctions or otherwise, and market

shares also reasonably predict how often the merging parties would be the winner and runner-up.

Compare PX 970, with Trial Tr. at 2927:8-25 (Snyder) (diversion according to share produces

diversion ratios broadly consistent with Dr. Hill’s runner-up study and Professor Snyder’s

agency data). Moreover, as previously discussed, competition affects advance levels even in

one-on-one negotiations, so the model’s use of auctions to simulate the result of negotiations has

some basis in market reality. Finally, similar or identical models have been used in other

antitrust cases involving industries that feature negotiations. See Wilhelmsen, 341 F. Supp. 3d at

64-65; Anthem, 236 F. Supp. 3d at 217-20; Sysco, 113 F. Supp. 3d at 24, 66-67 (noting that

29 Variable-profit margin is equal to revenue minus variable costs. The metric does not account for, ie.,

subtract, fixed costs. See Trial Tr. at 1310:2—25 (Hill).

55

customers were awarded contracts through “a request for proposal or bilateral negotiations”).

The Court understands that the second-score auction model provides only a rough approximation

of expected harm but nevertheless finds it useful. As for Dr. Hill’s allegedly mistaken inclusion

of fixed costs in some of his variable-profit margins, that was the more conservative approach:

Including the extra costs lowered the margins and reduced the model’s prediction of harm. See

Trial Tr. at 1311:18-1312:6 (Hill). Although Professor Snyder also suggested that Dr. Hill

should have included fixed costs in all the variable-profit margins, see id. at 3027:5-3030:4

(Snyder), Dr. Hill explained that the model explicitly calls for the use of variable, not fixed,

costs,°

In response to Professor Snyder’s criticism that the second-score auction model was a

poor fit for the publishing industry, Dr. Hill also ran a series of models based on the “gross

upward pricing pressure index” (GUPPI). See id. at 1315:16-1316:10 (Hill). The GUPPI

models use diversion ratios and margins as inputs, with higher diversion ratios and higher

margins leading to a higher prediction of harm. See id. at 1318:2—7 (Hill); see also supra

Section III.B.2.i.a (explaining diversion ratios). But the GUPPI models used by Dr. Hill are

more difficult for the defendants to challenge because those models were originally adopted by

the defendants’ own economists during the pre-complaint investigation of the instant merger.

See Trial Tr. at 1633:15—23 (Hill). The GUPPI models also predict a reduction in author

advances due to the merger, across different auction formats and using various diversion ratios,

including those calculated by Professor Snyder. The government summarized Dr. Hill’s findings

as follows:

30 Both the second-score auction model and the GUPPI models, discussed infra, are “explicit” that one should

use firms’ variable, not fixed, costs to implement the models. Trial Tr. at 3092:23-3093:15 (Hill). This is because

the models assume that publishers ask whether the marginal profits of acquiring one more book exceed the marginal

costs. See id.

56

Figure 8. Dr. Hill’s Estimates From Second-Score Auction and GUPPI Models (PX-964)

% Reduction in Author

Model Diversion Assumption Compensation

PRH S&S

Second-score auction model Diversion according to share 4.3% 11.6%

Multi-round auction GUPPI Diversion according to share 7.3% 19.2%

Multi-round auction GUPPI Snyder diversion ratios TA% 12.8%

Single-round and hybrid GUPPI| Diversion according to share 3.7% | 5.6%

Single-round and hybrid GUPPI| Snyder diversion ratios 3.7% 6.4%

See Govt. PFOF-PCOL 4 288 (citing PX 964). There is ample precedent for using GUPPI and

similar models to predict harm in antitrust cases. See Wilhelmsen, 341 F. Supp. 3d at 64;

Anthem, 236 F. Supp. 3d at 212; see also FTC v. Sanford Health, No. 1:17-cv-133, 2017 WL

10810016, at *12—13 (D.N.D. Dec. 15, 2017), aff'd., 926 F.3d 959 (8th Cir. 2019).

ili. Coordinated Effects

Another avenue for the government to prove competitive harm is by showing a likelihood

of “coordinated effects,” which occur when market participants mutually decrease competition in

the relevant market. AT&T, 310 F. Supp. 3d at 246 (“A proposed merger may violate Section 7

by enabling or encouraging post-merger coordinated interaction among firms in the relevant

market that harms customers.” (cleaned up)); see also Merger Guidelines § 7 (“Coordinated

interaction involves conduct by multiple firms that is profitable for each of them only as a result

of the accommodating reactions of the others.”). Coordinated effects can arise from an express

or implied agreement among competitors, see CCC Holdings, 605 F. Supp. 2d at 60; or from

“parallel accommodating conduct” among competitors without a prior agreement, Merger

Guidelines § 7. Parallel accommodating conduct involves “situations in which each rival’s

response to competitive moves made by others is individually rational, and not motivated by

57

retaliation or deterrence nor intended to sustain an agreed-upon market outcome, but

nevertheless emboldens price [decreases] and weakens competitive incentives to [raise advances]

or offer [authors] better terms.” Jd.

Coordinated effects are likelier in concentrated markets; indeed, the idea that

concentration tends to produce anticompetitive coordination is central to merger law. See Heinz,

246 F.3d at 716 (“Merger law ‘rests upon the theory that, where rivals are few, firms will be able

to coordinate their behavior, either by overt collusion or implicit understanding, in order to

restrict output and achieve profits above competitive levels.’” (quoting FTC v. PPG Indus., 798

F.2d 1500, 1503 (D.C. Cir. 1986)). Therefore, when the government has shown that a merger

will substantially increase concentration in an already concentrated market — as it has done

here, see supra Section III.B.1 — “the burden is on the defendants to produce evidence of

‘structural market barriers to collusion’ specific to this industry that would defeat the ‘ordinary

presumption of collusion’ that attaches to a merger in a highly concentrated market.” H & R

Block, 833 F. Supp. 2d at 77 (quoting Heinz, 246 F.3d at 725).

As an initial matter, a history of collusion or attempted collusion is highly probative of

likely harm from a merger. See Hosp. Corp., 807 F.2d at 1388; see also FTC y. Elders Grain,

Inc., 868 F.2d 901, 906 (7th Cir. 1989) (“[A]n acquisition which reduces the number of

significant sellers in a market already highly concentrated and prone to collusion by reason of its

history and circumstances is unlawful in the absence of special circumstances.”); H & R Block,

833 F. Supp. 2d at 78; Tronox, 332 F. Supp. 3d at 208-210; Merger Guidelines § 7.2. Thus, it is

significant that in United States v. Apple, Inc., the Second Circuit upheld a finding that between

2009 and 2019, all the “Big Six”?! publishers, except for Random House, participated in a

33 This was before Penguin Books and Random House merged, so there was a “Big Six” instead of a “Big

Five.”

58

“horizontal conspiracy . . . to raise e[-]book prices.” See 791 F.3d at 339, This coordination

9 66

involved “numerous exchanges between executives at different Big Six publishers,” “constant

communication” among the publishers “regarding their negotiations with both Apple and

Amazon,” and “frequent telephone calls among the Publisher Defendants.” Jd. at 302, 318.

“(T]he Big Six operated in a close-knit industry and had no qualms communicating about the

need to act together.” Jd. at 300. The Second Circuit concluded that the publishers engaged in

“express collusion” that was a per se violation of antitrust law. Jd. at 316, 321-29. Although

Random House did not participate in the conspiracy, Penguin Books and S&S both did, see id. at

308, and this “history of successful cooperation establishes a precondition to effective collusion

— mutual trust and forbearance.” See Hosp. Corp., 807 F.2d at 1388. The case portrays an

industry already “prone to collusion,” which may become “even more prone to collusion” after

the proposed merger of its largest and third-largest competitors. See Elders Grain, 868 F.2d at

905-06.

The Apple case provides the backdrop for trends in the industry that appear to

demonstrate that the Big Five are already engaging in tacit collusion or parallel accommodating

conduct when acquiring books. Recent years have seen the industry-wide standardization of

certain contract terms — involving payment structure, audio rights, and e-book royalties — in

ways that favor publishers over authors, suggesting that the top publishers have engaged in

coordinated conduct. Advances used to be paid to authors in two installments, but publishers

uniformly moved to paying them in three installments and then four installments, thereby

delaying authors’ compensation. See supra Section I.B. After audiobooks became a significant

source of revenue in the industry, publishers uniformly refused to acquire books without audio

rights included, thereby limiting authors’ ability to maximize their compensation and preventing

59

authors from diversifying their sources of income. See id. In addition, during the early years of

e-books, publishers uniformly shifted e-book royalty rates from 50 percent to 25 percent, thereby

reducing authors’ compensation. See Trial Tr. at 774:6-776:21 (Dohle). Thus, in an industry

where the competition to acquire anticipated top sellers is intense, the competing publishers

nevertheless choose, almost always, not to gain advantage by offering more favorable contract

terms. This phenomenon bespeaks a tacit agreement among the publishers to compete only on

the basis of advance level because it collectively benefits them not to yield on other contract

terms. Accord H & R Block, 833 F. Supp. 2d at 77-78 (“[A] highly persuasive historical act of

cooperation between [competitors]” supports the theory that “coordination would likely take the

form of mutual recognition that neither firm has an interest in an overall ‘race to free’... .”).

One example involving audio rights is illustrative. When selling the publishing rights to

a highly sought-after book, her agent attempted to hold an auction that excluded audio

rights. S&S wanted the book but refused to bid because “[t]he only way to prevent agents from

breaking off audio rights like this is to hold firm to our policy of no deals without audio rights.”

PX 652 at 2. An S&S editor ruminated, “It will be very interesting to see whether PRH,

Hachette, Harper or Macmillan participate. M[y] understanding is that they too have the ‘no

audio, no deal’ rule.” Jd. The agent was forced to restart the auction with audio rights included,

see PX 568 at 3, presumably because the book received insufficient offers or only received offers

that included audio. See PX 320 at 1 (in the first round, PRH bid for bundled audio rights in

violation of the auction’s initial rules). In the renewed auction that included audio rights, the

bidding was fervid and reflected vigorous competition.*” This episode starkly demonstrates that

32 As previously discussed, there were six rounds of bidding between four bidders, with a high bid of

$400,000 in the first round and a winning bid of $775,000 from PRH’s Viking imprint, which was $75,000 more

than Viking’s initial bid clearance. See supra Section 1.C.

60

the publishers, despite their great enthusiasm for the book, initially engaged in parallel conduct

to deny the author the ability to exclude audio rights from the auction. The parallel conduct was

effective and mutually beneficial, as the publishers all retained the opportunity to acquire the

book, with their preferred contract term concerning audio rights. Based on this evidence, the

Court finds that the Big Five publishers have engaged in tacit coordination that is profitable for

those involved.

Finally, it is significant that in a market already prone to collusion, where coordinated

conduct already appears to be rampant, PRH’s acquisition of S&S would reinforce the market’s

oligopsonistic structure and create a behemoth industry leader that other market participants

could easily follow. See PX 80E (translation of PX 80, materials for Bertelsmann board

presentation) at 13 (describing publishing industry as an oligopoly). The Big Five publishers

already control 91 percent of the relevant market. See PX 963. The merger would distill the Big

Five to a Big Four, with an overwhelmingly dominant top firm, PRH-S&S, controlling 49

percent of the market and dwarfing its nearest competitors. In the newly reconfigured market,

the top two firms, the merged entity and a would have a 74-percent market share.

See id. Under such circumstances, coordinated effects are likely through “sheer market power”

because the “post-merger market would feature two firms that control roughly three quarters” of

the market. Tronox, 332 F. Supp. 3d at 209; see also Heinz, 246 F.3d at 724 n.23 (recognizing

that “price leadership” is “a danger” in a “duopoly” market), The merger would thus increase

the market’s already high susceptibility to coordination. See Trial Tr. at 1329:18-21 (Hill).

33 Other factors that courts have found relevant to an evaluation of the likelihood of coordinated effects

include: differentiated products, transparent competitive outcomes, punishment mechanisms, and frequent purchases

for small amounts. See H & R Block, 833 F. Supp. 2d at 77-79; CCC Holdings, 605 F. Supp. 2d at 62; Arch Coal,

329 F, Supp. 2d at 144-45 . The Court sees no need to march through a discussion of those factors. Merger analysis

is industry-specific and fact-intensive. See Brown Shoe, 370 U.S, at 321-22 (“Congress indicated plainly that a

merger had to be functionally viewed, in the context of its particular industry.” (footnote omitted)). Where, as here,

61

C. Rebuttal

The government is entitled to a presumption of anticompetitive effects and has also met

its burden to establish a prima facie case. The defendants, therefore, now have the burden to

rebut the government’s case by “show[ing] that the prima facie case inaccurately predicts the

relevant transaction’s probable effect on future competition.” See Baker Hughes, 908 F.2d at

981. It is the defendants’ task to demonstrate that the market shares and the associated

presumption of illegality inaccurately reflect competitive reality. See id.; see also Heinz, 246

F.3d at 715 (“To rebut the presumption, the defendants must produce evidence that shows that

the market-share statistics give an inaccurate account of the merger’s probable effects on

competition in the relevant market.” (cleaned up)).

“There is no science to weighing the factors at play in an antitrust analysis,” and the

defendants may rebut the government’s prima facie case with any relevant “real-world

evidence.” RAG-Stiftung, 436 F. Supp. 3d at 312. For example, the defendants may meet their

burden of rebuttal by demonstrating low barriers to entry in the relevant market, see, e.g.,

Wilhelmsen, 341 F. Supp. 3d at 68; Staples I, 190 F. Supp. 3d at 133; or sophisticated

counterparts (here, authors and agents) who can blunt the impact of consolidation, see, e.g.,

RAG-Stifiung, 436 F. Supp. 3d at 315.

“Because the burden of persuasion ultimately lies with the plaintiff, the burden to rebut

must not be ‘unduly onerous.’” United States v. Anthem, Inc. (“Anthem IP’), 855 F.3d 345, 350

(D.C. Cir. 2017) (quoting Baker Hughes, 908 F.2d at 981, 991). However, “[t]he more

compelling the prima facie case, the more evidence the defendant must present to rebut it

there is a strong risk of collusion based on history, current practices, and extreme market concentration, the Court

finds it unnecessary to explore peripheral issues.

62

successfully.” See Baker Hughes, 908 F.2d at 991. Here, the government has “made out a

strong prima facie case” based on the highly concentrated market and affirmative evidence of

likely anticompetitive effects. See Wilhelmsen, 341 F. Supp. 3d at 66. The defendants,

therefore, “must make out a correspondingly strong rebuttal showing.” See id.

I. Existing Competition

The defendants assert that existing competition can and will constrain the merged

company more than market shares or the government’s evidence would suggest. See, e.g., Sysco,

113 F. Supp. 3d at 78. The defendants point to competition from other publishers, competition

from self-publishing, and internal competition within publishing houses. See Defs. PFOF

q§ 115-23, 164-85, 198-213; Defs. PCOL J] 74-79.

i. Other Publishers

The defendants argue that a combined PRH and S&S would be constrained by other

publishers, who do not plan to lower their advance offers or change their bidding strategies. See

Defs. PFOF Jf 115-23; Defs. PCOL J 74—-76. For example, HarperCollins’s CEO Brian

Murray testified that his company would not “hold back” in competing with the merged entity.

See Trial Tr. at 1385:9-15 (Murray). Consistent with that testimony, HarperCollins did not

project a decrease in its title count or its advance spending after the PRH-S&S merger was

announced. See DX 279 (HarperCollins strategy presentation for 2022) at 25. The CEOs of

other competitors, including Hachette al also stated that they would not change

their bidding strategies in response to the merger. See Trial Tr. at 211:9-13 (Hachette CEO

Michael Pietsch); Pe ee at 31. Therefore, the defendants

argue, other existing publishers stand ready to prevent any unilateral anticompetitive effects from

the merger.

63

The defendants’ reliance on such assurances from their competitors is insufficient. It is

not necessary for other publishers to change their maximum advances or bidding strategies for

anticompetitive unilateral effects to occur. First, and most obviously, with respect to book

acquisitions where PRH and S&S would have been the winner and runner-up, the merged entity

will acquire such books for lower advances regardless of the other publishers’ bids. See supra

Section ITI.B.1, Section II.B.2.1.

Second, in situations where PRH or S&S would have won a book, regardless of the

runner-up, the merged entity might submit a lower bid due to its decreased motivation to achieve

organic growth. See supra Section IIL.B.2. In such a case, another publisher could win the book

instead, for a lower advance than what PRH or S&S would have offered as standalone entities.

Third, publishers do not immediately offer their maximum advance when attempting to

acquire books; moreover, they initially offer higher advances when they think there is more

competition, and lower advances if they think there is less competition. See supra Section

IIL.B.2.i.a: see also Trial Tr. at 499:6-500:12 (Karp), 1267:13—-1268:8 (Hill). The general

softening of competition with the elimination of S&S as a standalone bidder, leading to the

perception of less competition in book acquisitions, would likely lead publishers to make lower

initial advance offers. See supra Section III.B.2. If subsequent bids that would have come from

PRH or S&S as separate entities are not forthcoming, or are lower than they otherwise would

have been, the other publishers could acquire books for lower advances simply by following

their existing bidding strategy.

Fourth and finally, it is not necessary for advances to decline in absolute terms for

authors to be harmed. The relevant market has been growing rapidly in recent years in response

to strong consumer demand, and advances have been rising, consistent with that growth. See PX

64

2002 (Stehlik) at 16-17; DX 422 (Glusman) at 37-38; Trial Tr. at 991:5-19 (Tart), 1990:4—-9

(Kim). If the merger goes through, the rate of growth might be offset by competitive harm,

allowing publishers to acquire books for more than they do now but for less than they would

have absent the merger. That would result in harm to authors even if there were no decline in

advances, or even if there was some (slowed) growth in the total advances paid.

The defendants also argue that non-Big Five publishers would be a significant

competitive constraint on a combined PRH and S&S. See Defs. PFOF J 164-80; Defs. PCOL

{| 74-76. The evidence shows, however, that the smaller publishers lack the resources to

compete regularly in the market for anticipated top-selling books. See supra Section I.C; Trial

Tr. at 2047:13-18 (Cheney). Individual publishers outside the Big Five rarely acquire books for

advances at or above $250,000. See PX 963.

The defendants take the novel approach of aggregating all the non-Big Five publishers

and characterizing them as a single force with a 9-percent market share — which allegedly

makes their collective power to constrain the merged company comparable to that of a Big Five

publisher. See Defs. PFOF § 164 (“When aggregated, the non-Big-Five publishers are as likely

to win as a ....”); Defs. PCOL ¥ 64 (arguing that an “effective 6-5 merger

(based on market shares) [is] at issue here”); see also Trial Tr. at 2906:10-2907:14 (Snyder 4

The defendants offer no precedent to support this economic sleight of hand, and the methodology

appears dubious. If market shares can be so readily manipulated by aggregating unaffiliated

companies, why not aggregate all the publishers that are not PRH and S&S into a single, massive

counterweight with a 51 percent market share? The defendants’ approach appears incompatible

with the way antitrust law approaches market concentration and its presumed effects on

34 Although Dr. Hill combined the non-Big Five’s market shares in his economic models as a reasonable

simplification, see Trial Tr. at 3081:12-23 (Hill), he did not treat them as one competitor in his overall analysis.

65

competition. See supra Section III.B.1. Generally, a firm with lower market share is assumed to

wield less market power, and market concentration would be considered low in an industry with

many individual firms with small market shares. See Merger Guidelines § 5.3 (“The Agencies

may measure market concentration using the number of significant competitors in the market.

This measure is most useful when there is a gap in market share between significant competitors

and smaller rivals ....”). For example, applying a HHI analysis shows that 100 firms that each

have a l-percent market share (which would produce an HHI of 100) do not represent the same

competitive landscape as two firms that each have a 50-percent market share (which would

produce an HHI of 5000).*° Indeed, the government points out that aggregating the non-Big

Five publishers does not help the defendants’ case because it yields a higher HHI and depicts a

more concentrated market. See Govt. PFOF-PCOL 134.

Professor Snyder gave counterintuitive and apparently erroneous testimony about the

significance of non-Big Five bidders in competitive auctions. He claimed that the non-Big Five

publishers, with a combined market share of 9 percent, are nevertheless the winner or runner-up

in 23 percent of auctions for anticipated top-selling books; while PRH and S&S, with a

combined market share of 49 percent, are the winner and runner-up only seven percent of the

time. See Trial Tr. at 2689:22—2690:5 (Snyder), 2827:13—23 (Snyder). The Court finds the 23-

percent figure unreliable because it was the subject of much contradictory testimony at trial,

including the credible assertion by Dr. Hill that the number should be halved. See id. at

3051:16—3053:17 (Hill). And overall, Professor Snyder’s reliance on the limited and

unrepresentative “agency data” weakened the credibility of his analyses. See supra Section

35 Summing the squares of each firm’s market share, the first HHI is calculated as 100 x 17= 100, and the

second HHI is calculated as 2 x 50? = 5000. See supra Section IM.A.2.

66

III.B.2 (note 26). Thus, the defendants’ expert fails to cast doubt on the reliability of the market-

share statistics presented by the government. See Heinz, 246 F.3d at 715.

ii. Internal Competition

The defendants argue that internal imprint competition increases competition in the

market beyond that represented in market shares. See Defs. PFOF 9 198-213; Defs. PCOL

{{ 77-79. That argument is undermined by the presumption that “[c]ompanies with multiple

divisions must be viewed as a single actor, and each division will act to pursue the common

interests of the whole corporation.” AT&T, 916 F.3d at 1043. This presumption “was adopted as

a principle of antitrust law,” id., in Copperweld Corp. v. Independence Tube Corp., 467 U.S.

752, 771 (1984) (“[T]he coordinated activity of a parent and its wholly owned subsidiary must be

viewed as that of a single enterprise .... A parent and its wholly owned subsidiary have a

complete unity of interest.”). Consistent with economic principles and common sense, internal

imprint competition should be considered only to the extent that it maximizes the profits of the

publishing house. See Areeda, Antitrust Law J 964b (“Antitrust law generally presumes that a

firm maximizes its profits in the environment in which it finds itself... .”).

Although internal competition among imprints is currently permitted by some publishers

in round-robin auctions, such competition is far from unrestrained. To the extent imprints

compete internally within the confines of a house bid, they can provide more editorial choices to

authors, but there is no price competition that allows authors to achieve the highest possible

advance level. See supra Section I.C, Even this non-price competition is discouraged. See DX

71 at 2 (email from S&S CEO Karp regarding editor guidelines) (“Duplicating work is not a

productive expenditure of one of our most valuable resources — time. You can acquire more

books if you aren’t all chasing the same ones... . Try to bail out on submission in which there

67

are more than two imprints in the building pursuing the project.”). In cases where internal

imprints do compete financially, such competition is confined to situations where there is an

external bidder. Moreover, as previously discussed, there are numerous examples in the record

of PRH using “background coordination in auctions to .. . avoid internal up-bidding.” PX 411 at

4: see also Trial Tr. 2372:17-2373:8 (McIntosh); supra Section LLC. Thus, internal competition

would have a very limited impact in mitigating anticompetitive effects in the industry. To

achieve the highest advances for authors, internal imprint competition is no substitute for

competition among independent publishers.

The defendants assert that the merged company would go even farther in permitting

internal competition than current policies allow. See Defs. PFOF § 210; DX 236. They note that

PRH CEO Dohle has promised, in a letter to agents, that S&S legacy imprints will bid against

PRH imprints even when there are no external bidders. See Defs. PFOF {] 210; DX 236. The

Court gives no weight to this unenforceable promise. Indeed, the promise calls to mind the

criminal-law concept of “consciousness of guilt”: Mr. Dohle’s extraordinary pledge appears to

reflect his awareness of how threatening the combined entity would be to authors and agents.

The promise lacks credibility for three reasons:

First, the proposed policy would not be profit-maximizing and is thus unreliable evidence

of future conduct. See Areeda, Antitrust Law § 964b. It is unclear how the new feature of the

policy — allowing internal competition even without an external bidder — would financially

benefit the combined entity. Instead, it appears that the promise was made just to get the deal

done, and once the merger is executed, there will be no economic incentive to maintain the

policy.

68

Second, the promise can be broken at will. Mr. Dohle, his successor, or his superiors

could legally change or rescind the new policy at any time. The defendants argue that such

behavior would harm the merged entity because of backlash from agents, see Defs. PFOF 210,

but evidence shows that agents have limited power over the large publishing houses, see infra

Section III.C.3.i. A unilateral promise by PRH that it will not use its market power if it acquires

S&S cannot rebut the government’s prima facie case. See H & R Block, 833 F. Supp. 2d at 82

(Even if “the Court has no reason to doubt that defendants would honor their promise [to

maintain the acquired firm’s current prices for three years post-merger], this type of guarantee

cannot rebut a likelihood of anticompetitive effects in this case.”).

Third, the promise would not prevent the merged entity from reducing internal imprint

competition. Despite the promise, PRH could coordinate bids with legacy S&S imprints in the

same way that PRH currently does among its own imprints. See supra Section I.C. The promise

also would not stop the merged company from consolidating PRH and legacy S&S imprints, as

PRH did when it reorganized the divisions within Random House in 2019, so that “there [would]

be less internal competition with[in] the focused editorial profiles in [its] divisions.” See

PX 241. Finally, the merged entity could direct its imprints to focus on non-competing genres,

thereby preventing the imprints from pursuing the same books. See id. For all the foregoing

reasons, Mr. Dohle’s promise does little to rebut the government’s prima facie case or the

presumption of anticompetitive effects.

69

iii. Self-Publishing

The defendants argue that self-publishing is a competitive constraint on the market,

particularly for celebrity and romance authors.*° See Defs. PFOF §§ 181-85; see also Trial Tr. at

566:1-11 (Karp). But, as previously discussed, self-publishing is not a reasonable substitute for

traditional publishers in the market for anticipated top-selling books. See supra Section LC,

Section I.A.2. Anecdotes about author Brandon Sanderson raising $40 million on Kickstarter,

or author Colleen Hoover having success with self-publishing, do not change the overall picture

29:66

of the industry. Sanderson’s success with self-publishing was “rare,” “a feat,” and “so

incredible.” Trial Tr. at 1076:4-10 (Weisberg), 1077:4-7 (Weisberg). Similarly, Hoover is “a

cultural phenomenon” and “the hottest author in the country.” Jd, at 524:18-25 (Karp), 560:5—

10 (Karp). Sanderson and Hoover are exceptions that prove the rule: For the overwhelming

majority of authors in the relevant market, self-publishing is no real substitute for using a

publishing house, and self-publishing therefore does little to constrain anticompetitive effects.

2. Barriers to Entry and Expansion

The defendants argue that there are few barriers to entry that would prevent new or

existing publishers from competing effectively with the merged company. See Defs. PFOF

44] 69, 124-48; Defs. PCOL §§ 80-83. New entrants to the market would presumably give

authors alternative outlets to publish their books, thereby preventing the merged entity from

lowering advances. “The existence and significance of barriers to entry are frequently . .. crucial

considerations in a rebuttal analysis. In the absence of significant barriers, a company probably

cannot maintain [sub]competitive pricing for any length of time.” Baker Hughes, 908 F.2d at

36 Defendants admitted in their closing argument that self-publishing was not a true option for authors. See

Trial Tr. at 3272:17~25 (arguing that hypothetical monopsonist test is “utterly meaningless other than addressing

whether there’s an outside option in the form of self-publishing, which nobody was arguing in this case”). In their

post-trial briefing, however, they surprisingly assert that self-publishing is a competitive threat.

70

987. To constrain the new entity, “entry [by new competitors] must be timely, likely, and

sufficient in its magnitude, character, and scope to deter or counteract the competitive effects of

concern.” Wilhelmsen, 341 F. Supp. 3d at 66-67 (quotations omitted). “The expansion of

current competitors is regarded as essentially equivalent to new entry, and is therefore evaluated

according to the same criteria.” /d. at 66 (quotations omitted).

Contrary to the defendants’ contentions, the evidence demonstrates that there are

substantial barriers to entry and expansion in the publishing market for anticipated top-selling

books. Established publishers have many advantages that are not easily replicated, including: (1)

back lists that generate substantial and consistent revenue, which in turn supports risky

acquisitions of high-advance books, see Trial Tr. at 156:5—157:14 (Pietsch); (2) large and

effective marketing, sales, and distribution teams that have relationships with media and

retailers, ___ at 6-7; (3) excellent reputations and track records of success

that attract authors, see Trial Tr. at 454:11—22 (Karp); and (4) lower variable costs due to

economies of scale, see id. at 2047:16—-18 (Cheney); see also Section IC. In addition, numerous

publisher witnesses expressed concern about a lack of access to sufficient printing capacity,

which limits the number of books that publishers can physically produce and thus limits

opportunity for expansion. See id, at 758:3-5 (Dohle), 121:8-19 (Pietsch), 364:1—9 (Eulau); a

eee at 3. Industry insiders, including PRH executives,

candidly acknowledged in trial testimony and ordinary-course documents that barriers to entry

are high in the publishing business. See Trial Tr. at 168:9-169:1 (Pietsch), 755:4-15 (Dohle),

1380:12-1381:8 (Murray); PX 80-E at 13; PX 79 (presentation by PRH executives on publishing

industry) at 8.

71

The best proof that would-be new competitors face formidable barriers to entry is the

stability of market shares in the industry: No publisher has entered the market and become a

strong competitor against the Big Five in the past thirty years. See Trial Tr. at 163:2-6 (Pietsch).

Moreover, the Big Five’s market share in acquiring anticipated top-selling books has remained

stable for the past three years. See PX 967. Thus, there is little evidence that new or existing

publishers will grow at a pace and magnitude that would allow them to discipline a merged PRH

and S&S. See Wilhelmsen, 341 F. Supp. 3d at 68 (“The fact that the merging parties have been

able to maintain high margins and market shares without witnessing notable entry and expansion

suggests that... the market... is characterized by significant barriers to entry.” (cleaned up)),

Merger Guidelines § 5.3 (“The Agencies give more weight to market concentration when market

shares have been stable over time ....”). The Big Five’s market shares are built on “decades of

credibility and success,” Trial Tr. at 454:11-22 (Karp), and they cannot be easily challenged by

less-established publishers.

Although the defendants argue that social media like “BookTok”?? and Amazon’s online

bookstore level the playing field for smaller publishers, those platforms are not new and are far

from “game-changing.” See id. at 1069:1-4 (Weisberg) (“[T]his is the business of one-to-one

word of mouth. It’s never been anything else and it’s still not, just the devices have changed.”).

Despite the current availability of “BookTok” and virtual storefronts, the Big Five still

consistently acquire the publishing rights for 91 percent of anticipated top-selling books,

demonstrating that the playing field has not been leveled in any meaningful way. See PX 994.

For example, PRH utilizes its superior resources to maximize sales even on Amazon. See supra

37 BookTok refers to activity on the social media platform TikTok where users review and promote books to

one another. See Trial Tr. at 1414:12-16 (Weisberg).

72

Section LC (PRH hires data scientists to study Amazon’s search algorithms and spends money to

get books better positioned in Amazon’s search results.).

The defendants nevertheless point to new entrants like Zando, Spiegel & Grau, and Astra

House, which have had some success in acquiring publishing rights to anticipated top-selling

books. See Defs. PFOF 9 129-32, 144, 147. Although those publishing houses are associated

with successful and well-respected editors, they lack many of the other advantages enjoyed by

the Big Five: big back lists; extensive marketing, sales, and distribution teams; and scale. As a

result, those new entrants have barely made a dent in the relevant market — their collective share

is less than one percent, see PX 968, and no one in the industry views them as substantial

competitors to the Big Five. See Trial Tr. at 249:20-250:90 (Pande), 552:21-557:3 (Karp).

Moreover, the growth of those new competitors was accompanied by a countervailing shrinkage

in the market shares of other non-Big Five publishers: The stability of the overall market share

of non-Big Five publishers indicates that the new entrants have done little to change the

competitive landscape, and that barriers to entry and expansion remain high. See id. at 1482:15—

1483:17 (Hill); PX 977.

The defendants contend that Big Five rivals like HarperCollins and well-funded

companies like Disney are poised to expand in the relevant market. See Defs. PFOF {| 124-48;

Defs. PCOL ff 81-83. To be sure, Big Five rivals face low barriers to expansion because they

have many of the same advantages that PRH and S&S have. But there is no evidence that

HarperCollins, Hachette, or Macmillan could or would compete more aggressively with the

merged company. See Trial Tr. at 177:1-15 (Pietsch), 1088:8-14 (Weisberg), 1385:9-1386:11

(Murray). The distribution of market shares among PRH, S&S, and the other Big Five

publishers, has remained relatively constant in recent years. See PX 994. The Court has every

73

reason to believe that all the industry players are already doing their best to compete; it is

therefore unlikely that the non-merging Big Five publishers could suddenly expand sufficiently

to prevent the anticipated competitive harm.

Two well-funded companies outside the Big Five highlighted by the defendants are

Amazon and Disney. Amazon acquired several high-priced books when it first started its

publishing business about a decade ago, but it has failed to make significant headway in the

industry. See Trial Tr. at 172:17-25 (Pietsch). From 2019 to 2021, Amazon’s share in acquiring

the publishing rights to anticipated top-selling books declined from under i to underff

ii. See PX 968 at 2. Amazon also struggles with selling its books outside of its own

platform. See Trial Tr. at 171:17-23 (Pietsch), The Court therefore is not convinced that

Amazon is a significant competitive constraint in the relevant market. The defendants ar

This text is long and has been trimmed here. Open the source document for the complete record.