United States v. Apple, Inc.

DENNIS JACOBS, Circuit Judge,

dissenting:

I respectfully dissent.

This appeal is taken by Apple Inc. from a judgment in the United States District Court for the Southern District of New York (Cote, J.), awarding an antitrust injunction in favor of the United States-, 31 states, the District of Columbia, and the Commonwealth of Puerto Rico. The plaintiffs’ claims are premised on Apple’s conduct as a prospective retailer of e-books. I vote to reverse.

* * *

I have no quarrel with the district court’s conscientious findings of fact; I. affirmatively rely on them, and cite them throughout. The 156 pages of findings track communications and interactions that happened over the 48-day course of events, detail by detail. See United States v. Apple Inc., 952 F.Supp.2d 638, 655-81 (S.D.N.Y.2013) (“Apple I”). All that is needed to decide the case, however, are the schematic facts that show the architecture of the horizontal and vertical arrangements and the dynamics of the competitive forces. They are set out in a nutshell in the following paragraphs, and at somewhat greater length in the Background section of this opinion.

As Apple was preparing the launch of its first iPad tablet in 2009, the company recognized that the device could support e-books, and gave consideration to including an e-book retail platform. However, Amazon had preceded Apple in the market, had established a 90 percent ascendency in sales of e-books, and was effectively excluding new entrants by offering bestsellers at a price ($9.99) that for many books was below the prices Amazon was paying publishers.

Although Apple was positioned to enter the retail market, it was unwilling to do so on terms that would incur a loss on e-book sales (as would happen if it met Amazon’s below-cost price), or that would impair its brand and likely fail (as would happen if it charged more than Amazon). So, as a condition to its entry as a competing buyer for the publishers’ wares, Apple insisted that the publishers agree to a distribution model that would lower that barrier to retail entry.

The new distribution model was implemented by several terms in Apple’s con*341tracts with publishers: agency pricing, tiered price caps, and a most-favored-nation clause. It is conceded that none of those terms is; standing alone, illegal. Apple also encouraged publishers to implement agency pricing in their contracts with other retailers. Although publishers were unhappy about Amazon’s below-cost price for e-books (which eroded the publishers’ hardcover sales) no one publisher alone could counter Amazon. In short order, five of the country’s six largest publishers agreed to Apple’s terms and jointly pressured Amazon to adopt agency pricing. The publishers thereby prevailed in what the district court found to be a horizontal price-fixing conspiracy. The barrier to entry thus removed, Apple entered the retail market as a formidable competitor. In the deconcentrated market, Amazon’s 90 percent market share is now 60 percent.

(I acknowledge that, in adducing facts found by the district court, this opinion unavoidably casts imputations on Amazon. Fairness requires acknowledgment that Amazon has not appeared in this litigation and has not had a full opportunity to dispute the district court’s findings or characterizations. Moreover, the fact of Amazon’s monopoly alone would not support an inference that Amazon’s behavior was in any way unlawful.)

The Department of Justice, 31 states, the District of Columbia, and the Commonwealth of Puerto Rico sued Apple and the five publishers for conspiracy in unreasonable restraint of trade, in violation of § 1 of the Sherman Antitrust Act, 15 U.S.C. § 1. The publishers settled, and Apple proceeded to a bench trial. The district court ruled that Apple’s conduct as a vertical enabler of the publishers’ horizontal price conspiracy constituted a violation per se of § 1, and that (in any event) Apple’s conduct would also violate § 1 under the rule of reason. On this appeal, a majority affirms only on the ground of liability per se. See Op. of Judge Lohier, ante, at 339-40. Since I would reverse, I consider as well the rule of reason. Judge Livingston’s opinion argues (for herself alone) that the judgment could be affirmed on that alternative ground.

The district court committed three decisive errors:

• The district court ruled (and the majority affirms) that a vertical enabler of a horizontal price-fixing conspiracy is in per se violation of the antitrust laws. However, the Supreme Court teaches that a vertical agreement designed to facilitate a horizontal cartel “would need to be held unlawful under the rule of reason. ” Leegin Creative Leather Prods., Inc. v. PSKS, Inc., 551 U.S. 877, 893, 127 S.Ct. 2705, 168 L.Ed.2d 623 (2007) (emphasis added). (POINT I)
• The district court’s alternative ruling under the rule of reason was predetermined by its (erroneous) per se ruling. Thus the district court assessed impacts on competition without recognizing that Apple’s role as a vertical player differentiated it from the publishers. The court should instead have considered Apple as a competitor on the distinct horizontal plane of retailers, where Apple competed with Amazon (and smaller players such as Barnes & Noble). (POINT II)
• Apple’s conduct, assessed under the rule of reason on the horizontal plane of retail competition, was unambiguously and overwhelmingly pro-competitive. Apple was a major potential competitor in a market dominated by a 90 percent monopoly, and was justifiably unwilling to enter a market on terms that would assure a loss on sales or exact a toll on its reputation. In that connection, the district court erro*342neously deemed the monopolist’s $9.99 price as categorically good for competition because it was lower than cost, and because e-book prices rose after the monopoly was broken. (POINT III)

A further and pervasive error (by the district court and by my colleagues on this appeal) is the implicit assumption that competition should be genteel, lawyer-designed, and fair under sporting rules, and that antitrust law is offended by gloves-off competition.

BACKGROUND

From the 2007 inception of the U.S. retail market for e-books through 2009, Amazon “dominated the e-book retail market, selling nearly 90% of all e-books.” Apple I, 952 F.Supp.2d at 649. It assured its domination by charging its retail customers $9.99 for new releases and bestsellers, below the wholesale price that Amazon was paying to publishers. Id. at 649-50, 708. The popular media reported that Amazon “takes a loss on the sale of the most popular e-books.” Id. at 652. That pricing deterred potential retail competitors from entering the relevant market— “trade e-books in the United States”1— because an entrant “would run the risk of losing money if it tried or was forced to match Amazon’s pricing to remain competitive.” Id. at 658.

Amazon’s below-cost pricing was also a threat to publishers, because at a $9.99 price point, e-books cannibalized sales of (more profitable) hardcover editions. Id. at 649. Although the major publishers believed Amazon’s below-cost pricing was “predatory,” id. at 653, each publisher understood that it was powerless to take on Amazon, id. at 650. Publishers feared that Amazon might “compete directly with publishers by negotiating directly with authors and literary agents for rights,” id. at 649, and might “retaliate” against insubordinate publishers “by removing the ‘buy buttons’ on the Amazon site that allow customers to purchase books ... or by eliminating [a publisher’s] products' from its site altogether,” id. at 679. One publisher, Macmillan, suffered such retaliation when Amazon removed the “buy buttons” for print and e-book versions of Macmillan titles. Id.

Amazon’s 90 percent market share constituted a monopoly under antitrust law. See, e.g., Am. Tobacco Co. v. United States, 328 U.S. 781, 797, 66 S.Ct. 1125, 90 L.Ed. 1575 (1946) (characterizing as “a substantial monopoly” a market share of “over 80% of the fíeld”); 3B Areeda & Hovenkamp, Antitrust Law ¶ 801 (3d ed.2008). Amazon’s below-cost pricing was a barrier to entry by Apple in 2009, when it contemplated entry into the e-book retail market via the iPad.2 Apple I, 952 *343F.Supp.2d at 654, 658. Apple nevertheless undertook to develop an e-book retail platform in time for the iPad’s launch, scheduled for January 27, 2010. Id. at 654-55. However, “Apple did not have to open an e-bookstore when it launched the iPad”; and it was willing to enter the market only on the condition that its e-book retail business would be profitable, such that Apple could “compete effectively with Amazon” without adopting a loss-leadership and below-cost pricing strategy. Id. at 656-59.

Apple opened extensive negotiations with publishers to determine how if at all it could enter the e-book retail market. Id. at 655-57. Apple met with the leaders of the six largest publishing houses in the United States: Hachette, HarperCollins, Macmillan, Penguin, Random House, and Simon & Schuster. Id. at 647, 655. At the outset, Apple understood that the publishers were unhappy with Amazon’s below-cost pricing of e-books; so Apple knew that the publishers “were willing to coordinate their efforts” to combat the $9.99 price, point. Id. at 656.

After some weeks, Apple and several publishers devised a new model for e-book distribution. Amazon had been paying a wholesale price for each e-book, and reselling (often at a loss) for a retail price of its choosing. Apple’s distribution contracts • would adopt an agency system: publishers would set the retail prices of e-books sold through Apple’s platform and Apple would take a fixed-percent commission on each sale. Id. at 659. However, the agency model would expose Apple (or any retailer) to risk, because publishers might protect hardcover sales by setting retail prices for e-books so high that Apple would appear out of touch with consumers aware of Amazon’s $9.99 price. Id. Apple’s solution was twofold. First, the proposed agency contract included a most favored nation (“MFN”) clause, under which publishers must price their new releases in Apple’s store at or below the lowest price offered by any other e-book retailer. Id. at 662. The district court found that the MFN clause “effectively forced” each publisher that signed Apple’s agency contract to move its other retailers onto the agency model. Id. at 664. That is because, once Apple’s cost was set as a percentage of the retail price, the publishers would suffer if Apple matched Amazon’s $9.99 retail price. Second, the proposed contract included maximum prices for various categories of e-books. Id. at 661-62. The district court found that these tiered price caps had the effect of setting anchor prices across the e-book industry. Id. at 670. Nonetheless, as the district court observed, these terms are not inherently illegal, and “entirely lawful contracts may include an MFN, price caps, or pacing tiers.” Id. at 698.

As Apple negotiated with publishers to sign the agency contract, it told each major publisher that all signing publishers would receive the same terms. Id. at 667. In the end, five of the six largest publishers signed Apple’s agency contract. Id. at 673. (Only Random House, the country’s largest, did not. Id.) As the district court found, the five signatories represented “over 48% of all e-books in the United States” when they signed Apple’s agency contract. Id. at 648. Apple unveiled its e-book retail platform — the “¡Bookstore” — at the first public demonstration of the iPad on January 27, 2010. Id. at 678-79.

After the publishers signed on to Apple’s agency contract, they had to focus on Amazon’s adoption of the agency model because otherwise (as explained above) the MFN clause would allow Apple to match Amazon’s price for bestsellers, and pay the publishers no more than a percentage *344commission on $9.99. However, “the [publishers feared retaliation from Amazon unless they acted in unison,” id. at 670, and “needed reassurance that they would not be alone,” id. at 674. An Apple executive liaised with each of the five signatory publishers, to encourage a “united front” in their negotiations with Amazon, and to keep the publishers “apprised about who was in and how many were on board.” Id. at 678. The publishers also communicated directly with each other. Id. at 674-77. When Amazon realized that the five publishers were acting in concert, it acceded and signed the agency contracts. Id. at 680-82.

Those are the findings on which Apple was adjudged to have committed an antitrust violation. ' The putative violation amounted to: (a) embedding the agency model (complete with MFN clauses and price caps) in Apple’s own contracts with publishers and (b) encouraging the publishers to coordinate horizontally in their efforts to push the industry-wide adoption of the agency model. Apple and the publishers shared the motive to increase the publishers’ pricing power in order to deprive Amazon of its monopoly. They succeeded: as the district court noted earlier in this litigation, “Amazon’s market share in e-books decreased from 90 to 60 percent in the two years following the introduction of agency pricing.” United States v. Apple, Inc., 889 F.Supp.2d 623, 640 (S.D.N.Y.2012).

The foregoing Background accepts and relies upon the district court’s findings of fact. One cannot say the same of Judge Livingston’s opinion, which supports its legal conclusions and its market analysis with novel findings made now on appeal, i.e., remand by other means. A few examples:

• The notion that Amazon’s below-cost pricing was loss-leadership “designed to encourage consumers to adopt the Kindle,” Op. of Judge Livingston, ante, at 299 (for the Court), is a novelty, supported by neither the fact findings nor the record. At any rate, the effect of e-book pricing outside of the relevant market is irrelevant.
• The majority asserts that Amazon’s below-cost pricing was limited to only “a small loss” on only “a small percentage of its sales.” Id. at 327 (for the Court). These observations are apparently drawn from a submission by Amazon, downplaying the anti-competitive effects of its monopoly-protective pricing. The district court did not rely on these statistics, presumably because they are misleading and self-serving: they ignore that the minority of titles comprising new releases and bestsellers naturally have an outsize impact on the industry. Accordingly, the district court found that the below-cost pricing had consequences on the market, namely that a new entrant “would run the risk of losing money if it tried or was forced to match Amazon’s pricing to remain competitive.” Apple I, 952 F.Supp.2d at 658.
• I can find no record support for the narrative that Amazon’s market share was eroding before Apple’s entry, that the iPad “promised to introduce more competition with or without Apple’s iBookstore,” and that the publishers thereby enjoyed increased negotiating leverage. Op. of Judge Livingston, ante, at 334. Similarly, the assertion that Barnes & Noble disrupted Amazon’s dominance in the e-book market, see id. at 334, is supported neither by the district court’s findings nor by the record.

*345By contrast, my antitrust analysis relies on the findings made by the district court, and incorporates no others, in order (a) to avoid factual disputes with my colleagues, (b) to defer to the district court’s thorough fact findings in arriving at my legal conclusions, and (c) to respect the limited role of appellate courts.

DISCUSSION

I

The district court’s principal legal error, from which other errors flow, is its conclusion that Apple violated § 1 under the per se rule. Having found that the publishers’ coordinated strategy was a horizontal price-fixing conspiracy, and that Apple had facilitated that conspiracy in its vertical relationship with the publishers, see Apple I, 952 F.Supp.2d at 691, the district court drew the legal conclusion that these facts established a per se violation of the Sherman Act by Apple. This appeal turns on whether purely vertical participation in and facilitation of a horizontal price-fixing conspiracy gives rise to per se liability.

Section 1 of the Sherman Act “outlaw[s] only unreasonable restraints”; so a court weighing an alleged violation “presumptively applies rule of reason analysis, under which antitrust plaintiffs must demonstrate that a particular contract or combination is in fact unreasonable and anticompetitive before it will be found unlawful.” Texaco Inc. v. Dagher, 547 U.S. 1, 5, 126 S.Ct. 1276, 164 L.Ed.2d 1 (2006) (quoting State Oil Co. v. Khan, 522 U.S. 3, 10, 118 S.Ct. 275, 139 L.Ed.2d 199 (1997)). The exception, liability per se, is reserved for those categories of behavior so definitively and universally anti-eom-petitive that a court’s consideration of market forces and reasonableness would be pointless. Id. Traditionally, restraints that are per se unlawful take the form of horizontal agreements “raising, depressing, fixing, pegging, or stabilizing the price of a commodity.” United States v. Socony-Vacuum Oil Co., 310 U.S. 150, 223, 60 S.Ct. 811, 84 L.Ed. 1129 (1940).

Among modern cases, the per se rule takes aim exclusively at horizontal agreements, because “competition among the manufacturers of the same [product] ... is the primary concern of antitrust law.” Continental T.V., Inc. v. GTE Sylvania Inc., 433 U.S. 36, 52 n. 19, 97 S.Ct. 2549, 53 L.Ed.2d 568 (1977). Accordingly, the trend of antitrust law has been a steady constriction of the per se rule in the context of vertical relationships. See, e.g., Leegin Creative Leather Prods., Inc. v. PSKS, Inc., 551 U.S. 877, 901, 127 S.Ct. 2705, 168 L.Ed.2d 623 (2007) (holding that vertical agreements for minimum prices are not per se violations); State Oil Co., 522 U.S. at 7, 118 S.Ct. 275 (holding that vertical agreements for maximum prices are not per se violations); Continental T.V., 433 U.S. at 59, 97 S.Ct. 2549 (holding that vertical non-price restraints are not per se violations); White Motor Co. v. United States, 372 U.S. 253, 261-64, 83 S.Ct. 696, 9 L.Ed.2d 738 (1963) (holding that vertical territorial restraints are not per se violations). The cases have “continued to temper, limit, or overrule once strict prohibitions on vertical restraints.” Leegin, 551 U.S. at 901, 127 S.Ct. 2705.

A vertical relationship that facilitates a horizontal price conspiracy does not amount to a per se violation. In another age, the Supreme Court treated such a hub-and-spokes conspiracy as a per se violation. See Interstate Circuit, Inc. v. Paramount Pictures Distrib. Co., 306 U.S. 208, 226-27, 59 S.Ct. 467, 83 L.Ed. 610 (1939). But the per se rule has been in steady retreat.

The most recent and explicit signal is given in Leegin, which explains that “the *346Sherman Act’s prohibition on ‘restraints of trade’ evolves to meet the dynamics of present economic conditions,” such that “the boundaries of the doctrine of per se illegality should not be immovable.” 551 U.S. at 899-900, 127 S.Ct. 2705 (alterations omitted). Leegin held that a manufacturer did not commit a per se violation of § 1 when it agreed with several retailers on a minimum price that the retailers could charge — a holding that overruled a century-old principle articulated in Dr. Miles Medical Co. v. John D. Park & Sons Co., 220 U.S. 373, 31 S.Ct. 376, 55 L.Ed. 502 (1911). See Leegin, 551 U.S. at 881, 127 S.Ct. 2705. Leegin reasoned that Dr. Miles had “treated vertical agreements a manufacturer makes with its distributors as analogous to a horizontal combination among competing distributors,” but that, “[i]n later cases, ... the Court rejected the approach of reliance on rules governing horizontal restraints when defining rules applicable to vertical ones.” Leegin, 551 U.S. at 888, 127 S.Ct. 2705. Dr. Miles was held to be inconsistent with “[o]ur recent cases[,] [which] formulate antitrust principles in accordance with the appreciated differences in economic effect between vertical and horizontal agreements, differences the Dr. Miles Court failed to consider.” Id.

Although the express holding of Leegin does not extend beyond the overruling of Dr. Miles, the Court’s analysis reinforces the doctrinal shift that subjects an ever-broader category of vertical agreements to review under the rule of reason. The Court first stated the subsisting scope of per se liability:

A horizontal cartel among competing manufacturers or competing retailers that decreases output or reduces competition in order to increase price is, and ought to be, per se unlawful.

Leegin, 551 U.S. at 893, 127 S.Ct. 2705. The Court then rejected per se liability for hub-and-spokes agreements, in wording that prescribes rule-of-reason review of vertical dealings that facilitate per se unlawful horizontal agreements (the type of agreement that the district court found Apple had undertaken):

To the extent a vertical agreement setting minimum resale prices is entered upon to facilitate either type of cartel [among manufacturers or among retailers], it, too, would need to be held unlawful under the rule of reason.

Id. (emphasis added). After Leegin, we cannot apply the per se rule to a vertical facilitator of a horizontal price-fixing conspiracy; such an actor must be held liable, if at all, “under the rule of reason.” Id.

Leegin is animated by the “appreciated differences in economic effect between vertical and horizontal agreements.” Id. at 888, 127 S.Ct. 2705. Since every challenged restraint is thus classified as either horizontal or vertical, one may draw certain reliable inferences: vertical agreements are not presumptively subject to per se liability; the vertical nature of the agreement is its salient feature; the influence of a vertical arrangement on a horizontal cartel (on another plane of competition) does not render the vertical arrangement per se unlawful.

Our only sister circuit to have considered this wording from Leegin arrived at the conclusion I draw. In Toledo Mack Sales & Service, Inc. v. Mack Trucks, Inc., 530 F.3d 204, 225 (3d Cir.2008), a manufacturer used its contracts with distributors to facilitate and enforce a horizontal conspiracy (among the distributors) that was itself illegal per se. See id. at 210. The Third Circuit held that Leegin’s instruction — that the vertical arrangement “would need to be held unlawful under the rule of reason” — prescribed the rule of reason as *347the proper analysis for whether the vertical conduct violated § 1. See id. at 225.

Taking the opposite tack, the majority opinion on this appeal insists that a vertical facilitator of a horizontal conspiracy is liable per se, even after Leegin. In support of that argument, the majority cites seven cases that pre-date Leegin,3 Op. of Judge Livingston, ante, at 322-24 (for the Court). The majority cites only one post-Leegin case that considers this question: namely, the Third Circuit’s analysis of a conspiracy that involved both vertical and horizontal relationships, concluding that the horizontal relationships violated § 1 per se and that pursuant to Leegin the vertical relationships “would have to be analyzed under the traditional rule of reason.”4 In re Ins. Brokerage Antitrust Litig., 618 F.3d 300, 318 (3d Cir.2010).

The majority’s holding in this case therefore creates a circuit split, and puts us on the wrong side of it.

“[Hjorizontal agreements as a class deserve stricter scrutiny than ... vertical agreements,” because horizontal agreements “pose the most significant dangers of competitive harm.” 11 Areeda & Hovenkamp, supra, ¶ 1902a, at 232. Horizontal price conspiracies are illegal per se because motives of horizontal players are aligned and dominant and create irresistible temptations. See, e.g., Adam Smith, The Wealth of Nations 207 (Collier 1902) (1776) (“People of the same, trade seldom meet together ..., but the conversation ends in a conspiracy against the public, or in some contrivance to raise prices.”).

Collusion among competitors does not describe Apple’s conduct or account for its motive. Apple’s conduct had no element of collusion with a horizontal rival. Its own rival in competition was (and presumably is) Amazon; and that competition takes place on a horizontal plane distinct from the plane of the horizontal conspiracy among the publishers. All Apple’s ener*348gy — all it did that has been condemned in this case — was directed to weakening its competitive rival, and pushing it aside to make room for Apple’s entry. On the only horizontal plane that matters to Apple’s e-book business, Apple was in competition and never in collusion. So it does not do to deem Apple’s conduct anti-competitive just because the publishers’ horizontal conspiracy was found to be illegal per se.

“[V]ertieal agreements are a customary and even indispensable part of the market system” and so do not represent the same presumptive threat to competition. 11 Areeda & Hovenkamp, supra, ¶ 1902d, at 240. Even a vertical agreement designed to decrease competition among competitors does not pose the threat to market competition that is posed by a horizontal agreement, for two reasons: (1) market forces (such as countervailing measures by competitors) are categorically more effective in countering anti-competitive vertical agreements, and (2) vertical agreements are so fundamental to the operation of the market that uncertainty about the legality of vertical arrangements would impose vast costs on markets. Id. at 240-41. Such market realities are driving the evolution of antitrust law, which has “rejected the approach of reliance on rules governing horizontal restraints when defining rules applicable to vertical ones.” Leegin, 551 U.S. at 888, 127 S.Ct. 2705.

The present case illustrates why per se treatment is not given to vertical agreements that facilitate horizontal conspiracies. Assuming (as is uncontested on appeal) that the publishers violated § 1 per sé through their coordination, Apple’s promotion of that horizontal conspiracy was limited to vertical dealings. ,

The per se rule is inapplicable here for another independent reason: The per se rule does not apply to arrangements with which the courts are not already well-experienced. Leegin, 551 U.S. at 887, 127 S.Ct. 2705. As the government conceded at oral argument, no court has previously considered a restraint of this kind. Several features make it mi generis: (a) a vertical relationship (b) facilitating a horizontal conspiracy (c) to overcome barriers to entry in a market dominated by a single firm (d) in an industry created by an emergent technology.

As I undertake to show in my analysis under the rule of reason, below, the restrictive market conditions Apple faced and the pro-competitive results of Apple’s conduct make its vertical dealings categorically reasonable. Even if one tests that conclusion under the rule of reason, the analysis is sufficiently complex and yields such substantial pro-competitive results that per se liability is an abdication of the duty to distinguish reasonable restraints from those that are unreasonable."

II

Having concluded first that Apple’s conduct was anti-competitive per se, corollary errors followed when the district court turned to the rule of reason. Once a court finds that a party acted unreasonably per se in a set of transactions, an epiphany is required for the court to conclude that the same party acted reasonably doing the same acts in the samé role at the same time. The influence arising from the district court’s per se accusation of wrongdoing infected all analysis that followed. Once Apple was deemed to have joined a conspiracy that was illegal per se, its goal, motive, and conduct seemingly needed (and got) no additional scrutiny — legal or moral or economic.

Having confirmed Apple’s per se liability by conflating the horizontal plane of competition among publishers with the horizontal plane of competition among retail*349ers, the district court committed the same error in its rule of. reason analysis. Thus the district court (as explained below) overstated the anti-competitive nature of Apple’s vertical dealings and overlooked the pro-competitive effects on retail competition — the horizontal plane on which Apple does e-book business. “The district court did not analyze the state of competition between ebook retailers,” as the majority concedes. Op. of Judge Livingston, ante, at 311 (for the Court) (emphasis omitted). Exactly.

Judge Livingston’s opinion succumbs to the same fallacy by declaring the majority’s own per se analysis so overwhelming that full rule-of-reason scrutiny requires no more than a “quick look.” Quick-look analysis is an appropriate tool only when “an observer with even a rudimentary understanding of economics could conclude that the arrangements in question would have an anticompetitive effect.” Cal. Dental Ass’n v. FTC, 526 U.S. 756, 770, 119 S.Ct. 1604, 143 L.Ed.2d 935 (1999). Quick-look analysis is not a tool for cutting corners. Judge Livingston’s opinion justifies quick-look analysis by referring to e-book price increases that form the majority’s earlier argument for the application of the per se rule, see Op. of Judge Livingston, ante, at 330 — price increases that, at any rate, are the expected result when monopolistic below-cost pricing dissipates.

In form and substance, Judge Livingston’s analysis demonstrates that when one starts with a finding of unreasonableness per se, the rule of reason analysis is tainted. It is called confirmation bias. The characterization of Apple’s conduct as “vigilantism” is telling. Op. of Judge Livingston, ante, at 297-98 (for the Court), 332. Use of that word either assumes the conclusion that the conduct is illegal, or else confuses it with self-help (which used to be a virtue). -

III

On this appeal, we have reached no majority as to the rule of reason. Judge Livingston writes for herself alone that, as an alternative to the per se rule, she would also affirm under the rule of reason; without a second judge supporting this conclusion, it is dicta, because our affirmance is based on the per se theory adopted by two judges. Unlike my colleagues, I must address the rule of reason, because my vote to reverse depends on my conclusion that this alternative theory of liability is every bit as untenable as liability per se.

Analysis under the rule of reason— whether conducted in full or by an untainted quick look — compels the conclusion that Apple did not violate § 1 of the Sherman Act. The issue is decided by comparing (a) the restrictive effect of Apple’s dealings with (b) the pro-competitive result of deconcentrating a market that had been dominated by a monopolist and insulated from competition through below-cost pricing.

Under the rule of reason, the initial burden rests with the plaintiffs “to demonstrate the defendants’ challenged behavior had an actual adverse effect on competition as a whole in the relevant market.” Geneva Pharms. Tech. Corp. v. Barr Labs. Inc., 386 F.3d 485, 506-07 (2d Cir.2004) (internal quotation marks omitted). Upon plaintiffs’ showing of such an effect, “the burden shifts to the defendants to offer evidence of the pro-competitive effects of their agreement,” and then “the burden shifts back to the plaintiffs to prove that any legitimate competitive benefits offered by defendants could have been achieved through less restrictive means.” Id. The reasonableness of the restraint then boils down to whether the dominant effect of the agreement is to promote competition or restrain it. Id.

*350Analysis begins with an accounting of anti-competitive effects. Apple’s vertical conduct consisted of negotiating the terms of its own contracts. Of course, every Contract is a restraint of trade to some extent, see Nat’l Collegiate Athletic Ass’n v. Bd. of Regents of Univ. of Okla., 468 U.S. 85, 98, 104 S.Ct. 2948, 82 L.Ed.2d 70 (1984); so this fact alone is neither here nor there.

The agency agreement that Apple signed with each publisher was innocuous: as the parties agree, each term — including the agency structure, MFN clause, and price caps — is absolutely legal. The district court so found expressly:

The Plaintiffs do not argue, and this Court has not found, that the agency model for distribution of content, or any one of the clauses included in the Agreements, or any of the identified negotiation tactics is inherently illegal. Indeed,. entirely lawful contracts may include an MFN, price caps, or pricing tiers.

Apple I, 952 F.Supp.2d at 698. The main restraint resulting from Apple’s vertical conduct was the shifting of pricing power from e-book retailers to e-book publishers. And this effect operated as a restraint only in the sense that Amazon faced pressure to adopt an agency model and to charge 'prices set by the five publishers, which of course remained in competition with each other, and with the publishers who account for the remaining 52 percent of the industry.

The district court opinion and the plaintiffs’ briefs fixate on the idea that Apple ended Amazon’s $9.99 price for most new releases and bestsellers, and that consumers would have preferred a lower price. But the consumer’s near-term preference for low prices is not an object of antitrust law. See Brooke Grp. Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209, 237, 113 S.Ct. 2578, 125 L.Ed.2d 168 (1993). The district court charts the short-term price developments, treating the end of below-cost pricing as anti-competitive and observing with disapproval the natural tendency for prices to rise to competitive levels. The rule of reason promotes competition; it can be safely assumed that if competition sharpens, prices will take care of themselves.

As to the pro-competitive effects, the rule of reason must take account primarily of the deconcentrating of the e-book retail market. The benefit of increasing the number of firms in a market derives from the “inverse correlation between concentration and competition.” Eleanor M. Fox, Economic Concentration, Efficiencies and Competition: Social Goals and Political Choices, in Industrial Concentration and the Market System 137, 149 (Eleanor M. Fox & James T. Halverson eds., 1979). As the district court found, Apple was weighing its entry into the retail e-book market, and the agency structure was the only way Apple would enter the market. Nobody has proposed — before or since Apple’s entry — any “less restrictive means” by which Apple could have achieved the same competitive benefits. See Geneva Pharms., 386 F.3d at 507 (plaintiffs’ burden to prove viable and less restrictive alternative). Apple’s challenged conduct broke Amazon’s monopoly, immediately deconcentrat-ed the e-book retail market, added a platform for reading e-books, and removed barriers to entry by others. And removal of a barrier to entry reduces for the long term a market’s vulnerability to monopolization.5 These effects sound in the basic *351goals of antitrust law. Even if only quick-look analysis were appropriate in this case, these effects would vindicate Apple’s conduct. (Judge Livingston’s opinion discounts this pro-competitive effect by noting the open question whether “below-cost pricing is unlawfully anti-competitive,” thereby suggesting that Apple’s dismantling of the entry barrier could be pro-competitive only if the barrier was itself a Sherman Act violation. Op. of Judge Livingston, ante, at 331-32. But it is no matter whether the insuperable barrier that Apple tore down had been raised lawfully or not.)

Another pro-competitive effect is the encouragement of innovation, a hallmark and benefit of competition. Apple began retailing e-books in conjunction with its release of the iPád, a device that integrated cutting-edge functions and applications, just one of which was the capacity for users to buy and read e-books. It is impossible to know the likely course of innovation, and pro-competitive effects of innovation cannot be measured; nevertheless, the encouragement of innovation must be afforded considerable weight under the rule of reason. See generally 2B Areeda & Hovenkamp, supra, ¶ 407. Apple’s business is not the technology of the clothespin.

The restraint of Apple’s vertical conduct was no more than a slight offset to the competitive benefits that now pervade the relevant market.6

How else could the competitive benefits have been realized in this market? In the course of this litigation, three theories have been offered for how Apple could have entered the e-book market on less restrictive terms. Each theory misapprehends the market or the law, or both. The absence of alternative means bespeaks the reasonableness of the measures Apple took.

Theory 1: Apple could have competed with Amazon on Amazon’s terms, using wholesale contracts and below-cost pricing. This was never an option. The district court found as fact that: a new entrant into the e-book retail market “would run the risk of losing money if it tried or was forced to match Amazon’s pricing to remain competitive,” Apple I, 952 F.Supp.2d at 658; Apple was “not willing” to engage in below-cost pricing, id. at 657; and Apple could have avoided this money-losing price structure simply by forgoing entry to the market, see id. at 659. Even if Apple had been willing to adopt below-cost pricing, the result at best would have been duopoly, and the hardening of the existing barrier to entry. Antitrust law disfavors a durable duopoly nearly as much as monopoly itself. See 6 Areeda & Hovenkamp, supra, ¶ 1429.

Theory 2: Apple could have entered the e-book retail market using the wholesale model and charged higher prices than Amazon’s. The district court foreclosed this theory as well; it found that Apple refused to impair its brand by charging “what it considered unrealistically high prices.” Apple I, 952 F.Supp.2d at 659. Even if Apple had been willing to tarnish its brand by offering bad value for money, the notion that customers would actually have bought e-books from Apple at the higher price defies the law of demand. *352Higher prices may stimulate sales of certain -wines and perfumes — not e-books.7

Nor could Apple justify higher prices for the e-books by competing on the basis of its new hardware, the iPad, because there is inter-operability among platforms. And if Apple had attempted to pursue this hardware-based competition by programming its iPad to run the iBookstore but to reject Amazon’s Kindle application, Apple might have been exposed to an entirely different antitrust peril. See United States v. Microsoft Corp., 253 F.3d 34, 50-80 (D.C.Cir.2001) (en banc); Google Android, No. 40099 (Eur. Comm’n Apr. 15, 2015) (antitrust proceedings brought by European Commissioner for Competition against Google for favoring Google’s own applications on mobile devices that use Google’s operating system).

Theory 3: Apple could have asked the Department of Justice to act against Amazon’s monopoly. Counsel for the United States actually proposed this at oral argument. At the. same time, however, he conceded that the Department of Justice had already “noticed” Amazon’s e-book pricing and had chosen not to challenge it because the government “regarded it as good for consumers.” Any request from Apple would therefore have been futile. True, Apple could not have known that the Antitrust Division would have adopted the position that below-cost pricing is not a concern of antitrust policy: who could have guessed that the government would adopt a policy that is- primitive as a matter of antitrust doctrine and illiterate as a matter of economics? Nevertheless, hindsight reveals that government antitrust enforcement against Amazon was not an option.

More fundamentally, litigation is not a market alternative. This observation has especial force in markets that are undergoing rapid technological advance, where the competitive half-life of a product is considerably more brief than the span of antitrust litigation. A requirement that potential market entrants litigate instead of enter the market on restrictive (but legal and reasonable) terms, would license monopoly for the duration.

Apple took steps to compete with a monopolist and open the market to more entrants, generating only minor competitive restraints in the process. Its conduct was eminently reasonable; no one has suggested a viable alternative. “What could be more perverse than an antitrust doctrine that discouraged new entry into highly concentrated markets?” In re Text Messaging Antitrust Litig., 782 F.3d 867, 874 (7th Cir.2015).

Application of the rule of reason easily absolves Apple of antitrust liability. That is why at oral argument the government analogized this case to a drug conspiracy, in which every player is a criminal — at every level, on every axis, whether big or small, whether new entrant or recidivist. The government found the analogy useful — and necessary — because in an all-criminal industry there is no justification or harbor under a rule of reason.

IV

Because T see no antitrust violation, I need not consider Apple’s separate challenge to the injunction itself. My eol-*353leagues, for their own good reasons, do not reach that challenge either. Yet the injunction and its shortcomings bear upon the institutional interest of the courts; and Apple’s challenge deserves some response. In my view, the injunction warps the role of a neutral, court-appointed referee into that of an adversary party, with predictable consequences.

The monitor is an arm of the district court, and owes loyalty in that direction only. See Fed.R.Civ.P. 53(a). But the injunction redirects the loyalty of the monitor to Apple’s chief adversary in the litigation, the Department of Justice. Under the injunction, the DOJ recommends the monitor (Injunction ¶ VI(A)), approves the monitor’s fees (id. ¶ VI(I)), and mediates disputes between the monitor and Apple (id. HITVI(E), (H)). Thus the injunction first creates a neutral fact-finding office, and then gives an adversary the ability to decide who holds the office, how much he gets paid (out of the other side’s pocket), and how broadly he may reach and inquire. Reciprocally, the monitor is directed to inform the government if he “discovers or receives evidence that suggests” further antitrust violations, whether or not related to this litigation. (Id. HVI(F).) This is a device that must misfire.

As events have happened (and were seemingly fore-ordained) the monitor has reason to look to the DOJ with gratitude and loyalty. The DOJ recommended Michael Bromwich as monitor, and the district court appointed him. United States v. Apple Inc., 787 F.3d 131, 134-35, 2015 WL 3405534, at *2 (2d Cir. May 28, 2015). Without a meaningful cap on his fee, Bromwich proposed that defendant Apple compensate him at $1,265 per horn* — an eye-popping rate for service as an agent of a court. Id. at 136, 2015 WL 3405534 at *3. (Because Bromwich lacks antitrust expertise, he proposed to add an actual antitrust lawyer to the team at $1,025 an hour. Id.) When Apple challenged that tariff as unreasonable, Bromwich explained that the injunction gave Apple no standing to object: “the fees and expenses to be paid to the monitor and his team are not set by Apple; they are set by the monitor, with approval reserved for the DOJ and the Plaintiff States.” Id. (quoting Bromwich). Bromwich was right, which is telling: the injunction contemplated no role for the judge.

Once the Department of Justice selected him and approved his hourly fee, Bromwich drew up his own mandate. Although the injunction' contemplated that the monitor would check sufficiency of an antitrust policy that Apple was to prepare in 90 days (and Apple’s compliance with it), Bromwich started his inquiry immediately on his appointment; he multiplied interviews, document inspections, and discontents; he demanded to interview Apple executives without the presence of Apple’s chosen counsel; and he took aim at the competitive culture of the corporation generally — a culture that is obviously aggressive, but just as obviously no business of the courts. See id. at 134-36, 140-41, 2015 WL 3405534 at *2-3, *7.

Having thus been selected by an adversary party, paid at a rate approved by the adversary party, and directed to look to the adversary party for the mediation of disputes, Bromwich was (in every respect important to a lawyer) retained and run by the adversary. Apple had an unenviable choice: it could accept scrutiny by a lawyer whose incentives were corrupted by the injunction that created his office, or attack the fee and the widening scope of inquiry, thereby sharpening the confrontations created by the mechanics of the injunction

A magistrate judge has cut Bromwich’s hourly fee. Id. at 139 n. 4, 2015 WL *3543405534 at *6 n. 4. And a panel of this Court has construed narrowly the scope of the monitor’s inquiries. Id. at 136-38, 2015 WL 3405534 at *4. But the structural defect of the injunction remains: allowing an arm of the court to serve as agent of an adversary party. It would take strong stuff for a lawyer to transcend the worldly incentives of this injunction: unlimited work at the (now cut) rate of $1,000 an hour, paid by a solvent party that may expect retaliation for protesting, in order to perform a monitorship subject to extension by the court for reasons that will be influenced by input from the monitor himself.

An injunction that thus blurs the lines of the adversary system does no good for the reputation of the- courts.

. The parties did not dispute this market definition. Apple I, 952 F.Supp.2d at 694 n. 60.

. While the district court did not use the label "barrier to entry,” its findings of fact made the point clearly. In finding that a new entrant to e-book retail in 2009 "would run the risk of losing money if it tried or was forced to match Amazon’s pricing to remain competitive,” Apple I, 952 F.Supp.2d at 658, the district court left no doubt that the effect of Amazon’s below-cost pricing regime was to "impede entry and protect existing market power” — the basic operation of a barrier to entry, 2B Areeda & Hovenkamp, supra, ¶ 420c, at 78.

The majority disputes whether there was any barrier to entry under Amazon's below-cost pricing regime, because at least one competitor attempted to join the market. See Op. of Judge Livingston, ante, at 299 (for the Court), 333. Even if that entrant had any chance of success (nobody contends that it sold a meaningful number of e-books, or made any money, or reduced Amazon’s mammoth market share to less than 90 percent), that fact need not imply ease of entry because "a barrier may protect a market incumbent *343without completely excluding entry.” 2B Areeda & Hovenkamp, supra, ¶ 420a, at 73.

.The cases are cited by the majority in this order: Klor’s, Inc. v. Broadway-Hale Stores, Inc., 359 U.S. 207, 79 S.Ct. 705, 3 L.Ed.2d 741 (1959); United States v. General Motors Corp., 384 U.S. 127, 86 S.Ct. 1321, 16 L.Ed.2d 415 (1966); Toys “R" Us, Inc. v. FTC, 221 F.3d 928 (7th Cir.2000); Denny's Marina, Inc. v. Renfro Productions, Inc., 8 F.3d 1217 (7th Cir.1993); United States v. MMR Corp., 907 F.2d 489 (5th Cir.1990); Business Electronics Corp. v. Sharp Electronics Corp., 485 U.S. 717, 108 S.Ct. 1515, 99 L.Ed.2d 808 (1988); NYNEX Corp. v. Discon, Inc., 525 U.S. 128, 119 S.Ct. 493, 142 L.Ed.2d 510 (1998).

Just as unhelpfully, the majority cites dicta from a Sixth Circuit case affirming the dismissal of a lawsuit that alleged a hub-and-spokes conspiracy. See Total Benefits Planning Agency, Inc. v. Anthem Blue Cross & Blue Shield, 552 F.3d 430 (6th Cir.2008). The majority cites the case as if its holding supports the continued legitimacy of the hub-and-spokes theory after Leegin, a flawed interpretation given the Sixth Circuit's disposition on the hub-and-spokes claim. Id. at 435 (holding that plaintiffs inadequately alleged a horizontal conspiracy and that, after Leegin, "all vertical price restraints are to be judged under the rule-of-reason standard” (emphasis added)).

.The Third Circuit analyzed a network of restraints, including a conspiracy among insurance brokers, a conspiracy among insurers, and agreements that connected the brokers and insurers. The court explained Leegin s impact this way:

Under the Supreme Court's jurisprudence, virtually all vertical agreements now receive a traditional rule-of-reason analysis. See Leegin, 551 U.S. 877, 127 S.Ct. 2705. In the factual context of this case, a horizontal agreement means ... an agreement among either the brokers or the insurers in the global conspiracy. Agreements between brokers and insurers, on the other hand, are vertical and would have to be analyzed under the traditional rule of reason.

In re Ins. Brokerage Antitrust Litig., 618 F.3d 300, 318-19 (3d Cir.2010) (internal citation and footnote omitted).

. Generally speaking, entry barriers permit monopolization and monopoly power allows a firm to erect entry barriers. See, e.g., Port Dock & Stone Corp. v. Oldcastle Ne., Inc., 507 F.3d 117, 125 (2d Cir.2007); United States v. Microsoft Corp., 253 F.3d 34, 82 (D.C.Cir.*3512001) (en banc); see also Mobil Oil Corp. v. Fed. Power Comm'n, 417 U.S. 283, 302 n. 23, 94 S.Ct. 2328, 41 L.Ed.2d 72 (1974). Each is less likely to arise when the other is absent from a market.

. Amazons's below-cost prices also threatened the market for hard-copy books, see Apple I, 952 F.Supp.2d at 649, and thus the royalties of authors, who may well consider that they have some role in this industry.

. In economic terms, e-books are subject to the law of demand and therefore have negative price elasticity of demand. See generally N. Gregory Mankiw, Principles of Economics 67 (6th ed.2012). E-books are neither Veblen goods nor Giffen goods, nor do they have perfectly inelastic demand. See id. at 92-93, 453-54, 835; Laurie Simon Bagwell & B. Douglas Bernheim, Veblen Effects in a Theory of Conspicuous Consumption, 86 Am. Econ. Rev. 349 (1996).