Role of Market Power in the Digital Economy

Home , AOL Mail, Bing Shopping, Engadget, FanHouse, Internet Icon (season 2), Moviefone

DRAFT 12/13/2017

New York State Bar Association Annual Meeting

Unilateral Conduct Committee: Role of Market Power in the Digital Economy

Moderator Outline by Eric Hochstadt, Weil, Gotshal & Manges LLP and Kellie Lerner, Robins Kaplan LLP1

I. Overview of Market and Monopoly Power

a. Market Power

i. The technical definition of market power under Section 2 is “the ability to raise prices above those that would be charged in a competitive market.”2

b. Monopoly Power

i. The technical definition of monopoly power is “the power to control prices or exclude competition.”3

ii. Monopoly power is analyzed in Section 2 monopolization and attempted monopolization analyses, as well as the Section 7 merger analysis.

c. Key Distinctions in Analyzing Market and Monopoly Power

i. Courts often use these terms interchangeably.4

ii. The technical definitions of market and monopoly power, which were established in NCAA and DuPont, appear to require a higher burden of proof

1 The authors would like to thank Charles N. Hurley and Shannon R. Rozell for their substantial assistance in preparing this detailed outline for the session. The views expressed herein are intended to be summary in nature and should not be interpreted to reflect any specific position taken by the authors or their respective law firms.

2 NCAA v. Bd. of Regents, 468 U.S. 85, 109 n.38 (1984).

3 United States v. E.I. duPont de Nemours & Co., 351 U.S. 377, 391 (1956).

4 See, e.g., NCAA, 468 U.S. 85, 109 n.38, 112 (defining “market power” and noting that it is linguistically different from “monopoly power,” while also citing DuPont for the definition of both market and monopoly power); Hanover Shoe v. United Shoe Mach. Corp., 392 U.S. 481, 481 n.3 (1968) (using “market” and “monopoly” power interchangeably when discussing Section 2 monopolization); Tops Mkts v. Quality Mkts., 142 F.3d 90, 97-98 (2d Cir. 1998) (“Monopoly power [is] also referred to as market power.”); Int’l Distrib. Ctrs, Inc. v. Welsh Trucking Co., Inc., 812 F.2d 786, 791 n.3 (2d Cir. 1987) (stating that “‘Market power’ is a synonym for ‘monopoly power.’”).

C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\(FINAL VERSION) DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX for “market power” than “monopoly power” since it requires proof that the prices were raised beyond a specified level.5

iii. No Supreme Court decision has explicitly distinguished between monopoly power and market power.

II. Direct Evidence and Circumstantial Evidence of Market Power

a. Direct Evidence

i. Generally, direct evidence is found where the firm profited from raising prices substantially above competitive levels, i.e. proving that defendant actually set prices above competitive levels or excluded competition.6

b. Circumstantial Evidence

i. Circumstantial evidence is found through a market structure analysis, i.e. where defendant possesses a dominant share of the relevant market that is protected by entry barriers.7

c. Screening for Market Power –Types of Circumstantial Evidence

i. Dominant Share of the Relevant Market

1. Monopolization – Courts have had difficulty drawing the lines for “dominant market share.”8 • 90% is considered to be enough9; • 60% to 64% is doubtful and is probably not enough10; • 33% is certainly not enough11; • 70% or more plus substantial barriers and lack of expanded output for existing competitors establishes a prima facie case of monopoly power;12

5 Compare NCAA, 468 U.S. 85, with E.I. duPont, 351 U.S. 377.

6 United States v. Microsoft Corp., 253 F.3d 34, 51 (D.C. Cir. 2001).

7 Id. at 56-57.

8 United States v. Aluminum Co. of America, 148 F.2d 416, 424 (2d Cir. 1945).

9 Id.

10 Id.

11 Id.

2 C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\(FINAL VERSION) DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX • 80% allows for an inference of monopoly power, but that inference is rebuttable by showing entry into the market is easy;13 • 50% to 70% range provides the greatest uncertainty, but anything below 50% is generally not enough.

2. Attempted Monopolization – A dominant market share for attempted monopolization is established by proving that there is a “dangerous probability of achieving monopoly power”14 • 50% or greater is generally considered a dangerous probability15; • 30% to 50% is rarely considered a dangerous probability16; • 30% or less is never considered a dangerous probability.17

ii. Existence of Significant Barriers to Entry

1. Entry barrier is defined as a cost that the entrant would have to bear that the incumbent would not or any condition that would inhibit entrants on a substantial scale as a response to the incumbent’s price increase.18 If entry is easy, courts will probably not find monopoly power, regardless if a firm has 100% market share.

2. Specific Types of Entry Barriers:

12 Spirit Airlines v. Nw. Airlines, 431 F.3d 917, 935-36 (6th Cir. 2005).

13 Eastman Kodak Co. v. Image Tech. Servs, 504 U.S. 451, 481 (1992).

14 Spectrum Sports v. McQuillan, 506 U.S. 447 (1993).

15 See, e.g., Am. Tobacco Co. v. United States, 328 U.S. 781, 797 (1946) (66% is sufficient); Kelco Disposal v. Browning-Ferris Indus., 845 F.2d 404, 409 (2d Cir. 1988) (55% is sufficient).

16 See, e.g., Smith Wholesale Co. v. Philip Morris USA, Inc., 219 F. App’x 398, 410-11 (6th Cir. 2007) (49-56% with low barriers to entry is insufficient); U.S. Anchor Mfg. Co. v. Rule Indus., 7 F.3d 986, 1001 (11th Cir. 1993) (less than 50% does not establish dangerous probability); M & M Med. Supplies & Serv. v. Pleasant Valley Hosp., 981 F.2d 160, 168 (4th Cir. 1992) (30% to 50% shares should usually be rejected).

17 See, e.g., Arthur S. Langenderfer v. S.E. Jonhson Co., 917 F.2d 1413, 1443 (6th Cir. 1990) (19% to 29% is insufficient).

18 Microsoft, 253 F.3d at 51; see also W. Parcel Express v. UPS, 190 F.3d 974, 977 (9th Cir. 1999).

3 C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\(FINAL VERSION) DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX • Legal license requirements19; • Regulations20; • Control of natural advantages or supplies21; • Market is too small for more firms22; • IP rights23; • Exclusivity arrangements24; and • Brand name and reputation.25

iii. Existing Competitors Have Reduced Output • A firm can exercise its anticompetitive economic power by restricting the output of its rivals (excluding competition). This is generally accomplished by raising prices or critical inputs above competitive levels, which forecloses competitors from competing in the market.26

iv. Market Structure and Market Performance

1. The Market Structure Characteristics Courts Consider:27 • Relative size and strength of competitors; • Economies of scale and scope; • Probable development of the industry; • Elasticity of consumer demand; • Homogeneity of products; • Dwindling market demand; • Potential competition; and

19 See, e.g., Rome Ambulatory Surgical Ctr. v. Rome Mem’l Hosp., 349 F. Supp. 2d 389, 418-19 (N.D.N.Y. 2004) (New York’s certificate of need is an example of a license requirement).

20 See, e.g., RSA Media v. AK Media Grp., 260 F.3d 10, 15 (1st Cir. 2001).

21 See, e.g., United States v. United Shoe Mach. Corp., 110 F. Supp. 295, 342 (D. Mass. 1953).

22 See, e.g., United States v. Griffith, 334 U.S. 100, 102 (1948).

23 See, e.g., Microsoft, 253 F.3d 34.

24 See, e.g., United States v. Dentsply Int’l, 399 F.3d 181, 189 (3d Cir. 2005).

25 See, e.g., Chicago Bridge & Iron Co. v. FTC, 534 F.3d 410, 437 (5th Cir. 2008).

26 See, e.g., Townshend v. Rockwell Int’l Corp., 200 U.S. Dist. LEXIS 5070, at *36 (N.D. Cal. 2000).

27 See ABA SECTION OF ANTITRUST LAW, ANTITRUST LAW DEVELOPMENTS 237 n.59 (7th ed. 2012) (listing market structure characteristics).

4 C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\(FINAL VERSION) DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX • Pricing trends and stability of market share over time.

2. The Market Evidence Courts Consider:28 • Pricing trends; • Pricing practices; • Stability of markets over time.

III. Digital Players and Market Entry

a. Regulatory Background for Mergers

i. Section 7 of the Clayton Act, 15 U.S.C. § 18, declares unlawful mergers that “may substantially lessen competition or tend toward monopoly”;

ii. The standard is “incipiency” and for unconsummated mergers requires a prediction of probable effects;29

iii. Since United States v. Philadelphia Nat. Bank, 374 U.S. 321, 363 (1963), a “structural presumption” has guided much of merger analysis – a merger that creates a firm with an “undue percentage” of a relevant market and results in a “significant increase in market concentration” is presumptively anticompetitive;

iv. But the presumption can be rebutted by, among other things, evidence that market shares inaccurately predict likely anticompetitive effects,30 or that the incentive to raise price will be eliminated due to likely entry or efficiencies31;

In practice, the parties look to the Merger Guidelines. Regarding entry, defendants carry the burden of showing that the entry or expansion of competitors will be “timely, likely and sufficient in its magnitude, character, and scope to deter or counteract the competitive effects of concern.32 The

28 See United States v. Eastman Kodak Co., 63 F.3d 95, 104 (2d Cir. 1995); Byars v. Bluff City News Co., 609 F.2d 843, 853 n.26 (6th Cir. 1979).

29 U.S. Dep’t of Justice & Fed. Trade Comm’n, Horizontal Merger Guidelines § 7 (2010).

30 United States v. Gen. Dynamics Corp., 415 U.S. 486, 506 (1974); see also United States v. Baker Hughes Inc., 908 F.2d 981, 983 (D.C. Cir. 1990).

31 FTC v. Staples, Inc., 190 F. Supp. 3d 100, 133 (D.D.C. 2016).

32 U.S. Dep’t of Justice & Fed. Trade Comm’n, Horizontal Merger Guidelines § 9 (2010).

5 C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\(FINAL VERSION) DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX relevant time frame for consideration in this forward looking exercise is typically two to three years.33

b. Regulatory Background for Monopolization

i. Section 2 of the Sherman Act, 15 U.S.C. § 2, covers unilateral activity such as monopolization and attempted monopolization. The statute provides: “Every person who shall monopolize, or attempt to monopolize, or combine or conspire with any other person or persons, to monopolize any part of the trade or commerce among the several States, or with foreign nations, shall be deemed guilty of a felony.”

ii. Monopolization requires a showing of two elements: defendant possesses (1) monopoly power, and (2) has acquired, enhanced, or maintained that monopoly power by the use of exclusionary conduct.34 Before determining whether monopoly power exists, it is necessary to define the relevant market (geographic and product) in which the firm has control over price or competition.

c. Select Enforcement Examples Where Entry by a Digital Player was Relevant

i. Section 2 Monopolization and Digital Players:

1. United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir. 2001) – The DOJ filed suit against Microsoft, alleging that it maintained an unlawful monopoly in the operating system market through terms in its bundling agreements.35 The DOJ and Microsoft entered into a consent decree and thereafter, the United States filed a separate complaint in the District of D.C., alleging antitrust violations.

• The D.D.C. defined the relevant market as “the licensing of all Intel-compatible PC operating systems worldwide.”36 Microsoft challenged the market definition by arguing that the court left out non-Intel compatible operating systems, operating systems for non-PC devices, and middleware products.37 On

33 Staples, 190 F. Supp. 3d at 133.

34 Verizon Commc’ns v. Law Offices of Curtis V. Trinko, 540 U.S. 398, 407 (2004); see also United States v. Grinnell Corp., 384 U.S. 563, 570-71 (1966).

35 Microsoft, 253 F.3d at 47.

36 Id. at 51.

37 Id. at 52.

6 C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\(FINAL VERSION) DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX appeal, the D.D.C. upheld the lower court because there were no other products that computer users could substitute without incurring substantial costs.38

• The D.D.C. also found that Microsoft had greater than a 95% market share.39 The D.D.C., however, did not look at market share alone,40 and found that a structural entry barrier discouraged software developers from writing applications for other less popular platforms.41

• The D.D.C. further held that Microsoft violated Sections 1 (illegal tying) and 2 (illegal monopolization) of the Sherman Act. In so holding, it found that Microsoft possessed monopoly power in the market for Intel-compatible PC operating systems because it had a dominant market share in which there was a high entry barrier; Microsoft engaged in various exclusionary efforts to protect its monopoly (i.e., paid huge sums of money to induce firms to help increase Internet Explorer’s shares of browser usage and made various threats to competitors); and Microsoft’s actions were harmful to consumers since there were no commercially viable alternatives.42

2. FTC Complaint, In re Intel Corp., No. 9341, 2009 FTC LEXIS 227 (F.T.C. Dec. 16, 2009) – The FTC brought suit under Section 5 of the FTC Act against Intel for unlawfully acquiring a monopoly in the computer chip market by refusing to deal with computer makers, Compaq, Intergraph, and DEC, to force them to surrender certain IP rights.

• The FTC alleged two relevant product markets: (1) “CPUs for use in desktop, notebook, netbook computers, servers, and narrower relevant markets contained therein”; and (2) “GPUs (including all graphics processors, or chipsets with graphics processors regardless of industry nomenclature) for use in

38 Id.

39 Id.

40 Id. at 55.

41 Id.

42 See Findings of Fact, United States v. Microsoft Corp., No. 98-1232 (TPJ) (D.D.C. Nov. 5, 1999).

7 C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\(FINAL VERSION) DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX desktop, notebook, netbook computers, servers, and narrower relevant markets contained therein.”43 .

• The FTC alleged that Intel had a 75% to 80% market share in the relevant markets and that the following barriers to entry existed: product development, cost and expertise to develop manufacturing capabilities, intellectual property rights, and product’s reputation and compatibility.44 The FTC argued that because of these entry barriers and Intel’s market share, competitors were unable to enter the relevant markets.45

• The FTC and Intel entered into a consent decree in which Intel agreed to refrain from using threats, bundled prices, or other offers to stifle competition in the computer chip markets.46 The court approved the consent decree and did not decide any dispositive motions.

3. In re 1-800 Contacts, Inc., 2016 FTC LEXIS 146 (F.T.C. Aug. 8, 2016) – The FTC brought suit under Section 5 of the FTC Act against 1-800 Contacts for orchestrating unlawful, anticompetitive bidding agreements with rival online contact lens sellers, which suppressed competition by requiring companies to refrain from advertising to consumers who had previously searched for “1-800 contacts.” 1-800 Contacts failed to provide any procompetitive benefits to justify the anticompetitive conduct.

• The FTC alleged two relevant product markets: (1) “the sale of search advertising by auction in response to user queries signaling the user’s interest in contact lenses, or smaller relevant markets therein”; and (2) “the retail sale of contact lenses, or smaller relevant markets therein, including the online retail sale of contact lenses.”47 1-800 Contacts disagreed and contended that the relevant market was “all retail sales of contact lenses in the United States, which encompasses sales

43 FTC Complaint, In re Intel Corp., 2009 FTC LEXIS 227, at *13-16.

44 Id. at *16-17.

45 Id. at *17-18.

46 See FTC Decision and Order, In re Intel Corp., No. 9341, 2010 FTC LEXIS 82 (F.T.C. Oct. 29, 2010).

47 FTC Complaint, In re 1-800 Contacts, Inc., 2016 FTC LEXIS 146, at *11. (ALJ decision is currently on appeal).

8 C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\(FINAL VERSION) DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX by online retailers and by physical, or ‘brick and mortar’ retailers, including independent ECPs, optical chains, and mass merchants.”48 The administrative judge found that online retailers are a distinct submarket of all retailers of contact lenses and thus the relevant market should be confined to online contact lens retailers.49

• The administrative judge found that the market for online contact lens retailers was about 17% of the total market for contact lenses.50 1-800 Contacts had more than a 60% market share for online contact lenses.51

• The FTC administrative judge ordered that 1-800 Contacts be barred from entering into these anticompetitive agreements and required the company to stop enforcing any and all provisions in existing agreements.52

d. Select Merger Examples Where Entry by a Digital Player was Relevant

i. § 7 Mergers and Digital Players - Permitted Mergers:

1. Comcast acquires NBCU (Approval with Conditions) – There the DOJ & FCC allowed Comcast to acquire 51% of NBCU’s stake from GE under the condition that new company must offer its content to online video distributors at "the same terms and conditions" that would be available to a cable or satellite TV operator.53

48 FTC Decision, In re 1-800 Contacts, Inc., 2017 FTC LEXIS 125, at *273 (F.T.C. Oct. 27, 2017) (emphasis added).

49 Id. at *280, 302 (stating that online retailers have peculiar characteristics and distinct customers, distinct prices, specialized vendors, and industry recognition).

50 Id. at *142.

51 Id. at *143-44 (also finding that 1-800 Contacts and the 14 parties that had formal agreements with 1-800 Contacts accounted for 79% of the market).

52 Id. at *439-47.

53 See U.S. Dept. of Justice Press Release, “Statement of the Department of Justice Antitrust Division on its Decision to Close its Investigation of Comcast Corp. and General Electric Co.’s subsidiary NBC Universal Inc. (NBCU),” Jan. 18, 2011, https://www.justice.gov/opa/pr/justice- department-allows-comcast-nbcu-joint-venture-proceed-conditions.

9 C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\(FINAL VERSION) DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX • Comcast was also required to relinquish its management rights in Hulu in order to ensure it doesn't use its Hulu ownership to wield control over the digital content space. • Lastly, the conditions prohibit Comcast from discriminating between its own content and services and other companies content and services. • DOJ enforcers found that allowing this deal under these conditions permits online content distributors to step into the shoes of cable and satellite companies to license full content.

2. Google acquires AdMob (No Challenge) – There the FTC found that that although the combination of the two leading mobile advertising networks raised serious antitrust issues, the agency’s concerns ultimately were overshadowed by recent developments in the market, Apple’s entry.54

• According to the FTC’s statement, evidence gathered by the agency raised important questions about the transaction. 1) The companies competed head-to-head intensely which created innovation and allowed mobile publishers to keep a large share of the revenue generated from the sale of their ad space; and 2) the companies had significantly more resources that gave them a major advantage over smaller rivals in the business. However, these concerns, were outweighed by recent evidence that Apple is poised to become a strong competitor in the mobile advertising market.

• “As a result of Apple’s entry, AdMob’s success to date on the iPhone platform is unlikely to be an accurate predictor of AdMob’s competitive significance going forward, whether AdMob is owned by Google or not,” the Commission’s statement explained.

• In addition, a number of firms appear to be developing or acquiring smartphone platforms to better compete against Apple’s iPhone and Google’s Android, and these firms would have a strong incentive to facilitate competition among mobile advertising networks.

ii. § 7 Mergers and Digital Players - Challenged Mergers:

1. Comcast attempt to acquire TWC - The deal was proposed to take the form of a stock swap, estimated at the time of announcement to be worth about $45.2 billion. The two companies argued that the merger would

54 See Fed. Trade Comm’n Press Release, “FTC Closes its Investigation of Google AdMob Deal,” May 10, 2010, https://www.ftc.gov/news-events/press-releases/2010/05/ ftc-closes-its- investigation-google-admob-deal.

10 C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\(FINAL VERSION) DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX increase their overall scale, allowing the company to become more competitive, improve customer service quality, and quicken innovation.

• The DOJ opposed the merger arguing that it would reduce competition through consolidation of the cable industry, lead to higher costs of service, and give Comcast greater leverage in how it distributes content owned by its NBCUniversal division to competitors, such as over-the-top services.55 (Comcast withdrew).

2. Staples attempt to acquire Office Depot – In 2015-16, the FTC sought an injunction to enjoin the transaction pending the results of the administrative proceeding, charging that Staples, proposed $6.3 billion acquisition of Office Depot, would significantly reduce competition nationwide in the market for “consumable” office supplies sold to large business customers for their own use.

• The complaint also alleged that, in competing for B-to-B contracts, both Staples and Office Depot can provide the low prices, nationwide distribution and combination of services and features that many large business customers require. Further by eliminating competition higher prices, reduced quality and barriers on entry were sure to follow. (Merger was withdrawn).

• “Defendants' response relies in large part on the prospect that Amazon Business will replace any competition lost because of the merger. Although Amazon Business may transform how some businesses purchase office supplies, the evidence presented during the hearing fell short of establishing that Amazon Business is likely to restore lost competition in the B-to-B space in a timely and sufficient manner.”56

II. Do the Antitrust Laws Need to Adapt to Apply to “Allegedly Dominant” Digital Firms?

i. Some argue that the current framework in antitrust i.e. specifically pegging competition to consumer welfare, defined as short term price effects is

55 See U.S. Dept. of Justice Press Release, “Statement of the Department of Justice Antitrust Division on Comcast Corporations Decision to Abandon Proposed Acquisition of Time Warner Cable,” April 24, 2015, https://www.justice.gov/opa/pr/comcast-corporation-abandons-proposed- acquisition-time-warner-cable-after-justice-department.

56 Staples, 190 F. Supp. 3d at 111.

11 C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\(FINAL VERSION) DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX unequipped to capture the architecture of market power in the modern economy.57

Others push back on these notions arguing that adopting approaches such as these disregard the consumer benefits in the pursuit of conflicting goals.58 They cite the DraftKings and FanDuel 2017 merger challenge as an example of antitrust laws being equipped with the tools to handle novel, technology- laden markets when significant risk of consumer harm are identified.59 They also compare the failed merger to the Sirius/XM merger in 2008 to show that the tools are not inflexible or rigid. 60

ii. This debate has come up before in United States. v. Microsoft Corp., 253 F.3d 34, 49 - 50 (D.C. Cir. 2001).

• “We decide this case against a backdrop of significant debate amongst academics and practitioners over the extent to which ‘old economy” § 2 monopolization doctrines should apply to firms competing in dynamic technological markets characterized by network effects. In markets characterized by network effects, one product or standard tends towards dominance, because “the utility that a user derives from consumption of the good increases with the number of other agents consuming the good.’”

• “Rapid technological change leads to markets in which ‘firms compete through innovation for temporary market dominance, from which they may be displaced by the next wave of product advancements.’ Microsoft argues that the operating system market is just such a market.”

• “Whether or not Microsoft’s characterization of the operating system market is correct does not appreciably alter our mission in assessing the alleged antitrust violations in the present case. As an

57 See, e.g., Lina M. Khan, Note, Amazon’s Antitrust Paradox, 126 Yale L.J. 710 (2017).

58 See, e.g., Maureen K. Ohlhausen, Acting Chairman, Fed. Trade Commn. Address at Georgetown University, (Sept. 12, 2017), https://www.ftc.gov/system/files/document/public_statements/1253163/ georgetown_mko_9-11-17.pdf.

59 DraftKings, Inc. & FanDuel Ltd., F.T.C. Docket No. 9375 (2017).

60 See U.S. Dept. of Justice Press Release, “Statement of the Department of Justice Antitrust Division on its Decision to Close its Investigation of XM Satellite Radio Holdings Inc.’s Merger with Sirius Satellite Radio Inc.,” Mar. 08, 2008, https://www.justice.gov/archive/opa/pr/2008/March/08_at_226.html.

12 C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\(FINAL VERSION) DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX initial matter, we note that there is no consensus among commentators on the question of whether, and to what extent, current monopolization doctrine should be amended to account for competition in technologically dynamic markets characterized by network effects.”

• “Indeed, there is some suggestion that the economic consequences of network effects and technological dynamism act to offset one another, thereby making it difficult to formulate categorical antitrust rules absent a particularized analysis of a given market.”

13 C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX C:\USERS\HURLEYCH\DESKTOP\ANTITRUST OUTLINE\(FINAL VERSION) DIGITAL PLAYERS AND ENTRY IN THE MERGER CONTEXT_WEIL_96382508_1.DOCX

WHY THE DYNAMICS OF COMPETITION FOR ONLINE PLATFORMS LEADS TO SLEEPLESS NIGHTS, BUT NOT SLEEPY MONOPOLIES

David S. Evans*

August 23, 2017 Revision of July 25, 2017 Draft

Abstract

Recent claims that online platforms have secured permanent monopolies, protected by barriers to entry from network effects and stockpiles of data, and should be the focus of intense antitrust and regulatory scrutiny, are inconsistent with the economics, technology, and history of online competition. Online platforms face dynamic competition as a result of: disruptive innovation that provides opportunities for entry; competition from online platforms that have secured a toehold in one area but compete across multiple areas; the fragility of category leadership resulting from the fact that network effects are reversible and entry costs are low; and the prevalence of ad-supported models, which result in seemingly disparate firms competing for consumer attention and advertiser dollars. The last two decades of online platform competition demonstrate that category leaders are often toppled, unexpectedly, through some combination of technological change, business model innovation, and cross-platform rivalry. The palpable threat of displacement prevents online platforms from taking their customers for granted. The history of online platform competition also provides empirical refutation of the proposition that data on users protects platform leaders from competition or puts an insurmountable obstacle before entrants. All this points to online platforms facing sleepless nights since any online platform that tries the quiet life of monopoly risks catastrophe.

______

*Evans is the Chairman, Global Economics Group, and Executive Director, Jevons Institute for Competition Law and Economics and Visiting Professor, University College London. The author gratefully acknowledges research support from Facebook.

I. Introduction

Some commentators are raising concerns that the online platforms spawned by the Internet

“look unstoppable.”1 They claim these platforms lack effective competition and pose similar concerns as giant companies did a century ago. There are calls for greater antitrust scrutiny, regulation, and even breakups. A recent theme is that these firms have built impregnable moats around themselves based on the vast amounts of data they control.

The last twenty years of history should humble anyone who claims that online platforms have secure monopolies. The record is replete with forecasts, soon proved wrong and then forgotten, that the winners that took all, or most, were unbeatable. Online platforms are impelled to innovate and compete for users because so many supposed winners have in fact been beaten. Thus, as of now, comparisons between the online platforms and old titans are inapt and overwrought. Four features distinguish online platforms from the firms that once dominated traditional industries, and lead to vigorous dynamic competition in the digital economy.2

1 The Economist, “Regulating the Internet Giants: The World’s Most Valuable Resource Is No Longer Oil, But Data,” May 6, 2017, http://www.economist.com/news/leaders/21721656-data-economy-demands-new-approach-antitrust- rules-worlds-most-valuable-resource. Jonathan Taplin, “Is It Time to Break Up Google?” New York Times, April 22, 2017, https://www.nytimes.com/2017/04/22/opinion/sunday/is-it-time-to-break-up-google.html. Farhad Manjoo, “Tech’s Frightful Five: They’ve Got Us,” New York Times, May 10, 2017, https://www.nytimes.com/2017/05/10/technology/techs-frightful-five-theyve-got-us.html Richard Straub, “Managing in an Age of Winner-Take-All,” Harvard Business Review, April 7, 2015, https://hbr.org/2015/04/managing-in-an-age-of- winner-take-all. Allen P. Grunes and Maurice E. Stucke (2015), “No Mistake About It: The Important Role of Antitrust in the Era of Big Data,” Antitrust Source, 14(4), https://www.americanbar.org/content/dam/aba/publishing/antitrust_source/apr15_full_source.authcheckdam.pdf; Maurice E. Stucke and Allen P. Grunes (2016), Big Data and Competition Policy, Oxford: Oxford University Press. 2 This paper builds on several on my earlier papers concerning dynamic competition in the information-technology era. See David S. Evans, Albert L. Nichols, and Bernard Reddy (2002), “The Rise and Fall of Leaders in Personal Computer Software,” in David S. Evans (ed), Microsoft, Antitrust, and the New Economy: Selected Essays, The Milken Institute Series on Financial Innovation and Economic Growth, Vol. 2 (New York: Kluwer Academic Publishers); David S. Evans and Richard Schmalensee (2002), “Some Economic Aspects of Antitrust Analysis in Dynamically Competitive Industries, in Innovation Policy and the Economy, Vol 2, A. Jaffe, J. Lerner, and S. Stern, eds., (Cambridge, MA: MIT Press); and David S. Evans (2016), “Multisided Platforms, Dynamic Competition, and the Assessment of Market Power for Internet-Based Firms,” University of Chicago Coase-Sanders Institute for Law and Economics Research Paper No. 753, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2746095.

First, turbulent waves of disruptive innovation have shaken the business models of platform leaders and opened new avenues of entry and competition. The Internet revolution that began in the mid-1990s led to the ultimate collapse of AOL, which had built a large online network and had a first- mover advantage. The smartphone revolution started just thirteen years ago, after the opening of the commercial Internet. Microsoft failed to leverage its leadership in desktop operating systems into mobile operating systems. Its power and relevance were diminished. Meanwhile, Blackberry’s once- vaunted device and secure e-mail network business collapsed. Now, less than ten years after the start of the smartphone revolution, voice-activated artificial intelligence platforms are taking off, and promising alternatives to current ways of doing things. The winners of this new round of competition will challenge the winners of previous ones. Smooth seas going forward are unlikely given ongoing underlying technological innovations that could prompt further disruptions. Nothing similar happened over such short time periods to the firms that dominated certain sectors of the economy from the late 19th century to the late 20th century.

Second, online platforms pegged as leaders in one area compete with each other in many other areas. They identify one another as competitors and commonly encroach on each other’s turf.

Consumers, for example, are as likely to start product searches on Amazon, the leading e-commerce firm, as on Google, the leading search-engine firm.3 A century ago dominant firms largely occupied silos. U.S. Steel didn’t compete with AT&T, which didn’t compete with the Chesapeake & Ohio

Railroad, which didn’t compete with General Electric. None of those firms competed with Standard

Oil before its 1911 breakup.

Third, online platforms are more susceptible to attack by entrants than network industries of a century ago. Network effects and sunk costs made the natural monopolies around the turn of 20th

3 Greg Sterling, “Survey: Amazon Beats Google as Starting Point for Product Search,” Search Engine Land, June 28, 2016, http://searchengineland.com/survey-amazon-beats-google-starting-point-product-search-252980; Jason Del Rey, “55 Percent of online Shoppers Start Their Product Searches on Amazon,” ReCode, September 27, 2016, https://www.recode.net/2016/9/27/13078526/amazon-online-shopping-product-search-engine.

century difficult to challenge. Rivals had to sink massive amounts of capital into duplicating physical networks such as railroad tracks and telephone lines. Using multiple networks, or switching between them, was expensive for customers, even if a second network was available. However, online platforms can leverage the Internet to provide wired and wireless connections globally. People find it generally easy, and often costless, to use multiple online platforms, and many often do. The ease and prevalence of multihoming have enabled new firms, as well as cross-platform entrants, to attract significant numbers of users and secure critical mass necessary for growth. Incumbent platforms then face serious competitive pressure from new entrants—startups or other online platforms—because their network effects are reversible. Sleepy firms risk a death spiral, as Yahoo learned.

Fourth, while many online platforms are “leaders” in particular categories, such as portals, these firms compete for the scarce attention of people and then compete to sell slices of that attention to advertisers. Even when differentiation softens this competition, the ad-supported platforms are much closer rivals than were the titans around the turn of the 20th century, or the most valuable firms towards the end of the century. Unlike traditional ad-supported media, online attention platforms, such as Pinterest, don’t have to invest in printing plants or television licenses. The technology of buying online also makes the ad-supported platforms substitutes, to some degree, for e- commerce platforms.

It is true that online platforms collect much information on their users, which then provides inputs into enhancing and improving their products and services. They invest in algorithms for processing and learning from that data and rent or build server farms to hold it. However, the rise and fall of platforms over the last twenty-five years, or even the last seven, refutes the claim that, as a general matter, data prevents entry or protects dominance. Many platforms, in possession of large amounts of data, such as Orkut, have gone into sharp decline while others, such as Tinder, have grown rapidly without possessing large amounts of data until they achieved scale.

Even the most successful online platforms don’t behave as if they’ve ever had much rest or are expecting to be getting some anytime soon. Economics, history, and technology suggest that they will face sleepless nights for the foreseeable future. None of these firms look remotely ready or able to settle into the quiet life of monopoly.4 More likely, they will be forces of competition, and disruption, throughout the economy, as the incipient revolutions in physical retail, transportation, and voice-activated platforms relying on artificial intelligence (to name just a few examples) show.

That doesn’t mean that competition authorities should take naps. Like all firms, online platforms may engage in unlawful collusion or monopolization. They may attempt mergers and acquisitions that could harm competition. Some online firms may have significant market power in particular lines of commerce and abuse that market power, through unilateral or vertical practices, to squelch competition. Evidence-based analysis, sharply focused on whether there is harm to consumers, and an understanding of the history and economics of online rivalry, however, should inform vigilance over the digital economy.

II. Economics and Technology of Online Platforms

Many successful Internet firms operate platforms. Platforms are businesses that connect two or more different types of users, such as drivers and riders in the case of Uber, facilitating a mutually beneficial exchange between users.5 Modern technologies, particularly the Internet, have made it easier to create platforms and to scale them globally. The business model isn’t new. Ancient village matchmakers, medieval bazaars, nineteenth century newspapers, telecommunications networks, and

20th century credit card systems have used the same model.

4 “The best of all monopoly profits is a quiet life.” John Hicks, “Annual Survey of Economic Theory: The Theory of Monopoly,” Econometrica, January 1935 3(1), pp. 1-20. 5 See David S. Evans and Richard Schmalensee (2016), Matchmakers: The New Economics of Multisided Platforms (Boston: Harvard Business Review Press). (Matchmakers).

Platform businesses often have what economists call indirect network effects.6 One type of user, drivers for example, value platforms that enable them to connect to more of the other type of user, riders for example. Larger platforms are more valuable to their participants to the extent they can connect more of each type of user with more of each other type of user. Telecommunications networks became more valuable to their participants as they expanded their reach because people who wanted to make calls could reach more people who could receive calls and vice versa.7

Commentators often point to indirect network effects as the propellant that makes these firms unstoppable. The dynamics of competition among these firms demonstrates that these indirect network effects, while important, are neither decisive for continued success, nor insurmountable entry barriers. In fact, Internet platforms have features that tend to make them much more susceptible to competition than the 20th century giants of American industry and companies built on physical networks.

A. Online Platforms Don’t Have to Sink Capital into Physical Networks

Online platforms don’t have to sink capital into providing the physical facilities for providing services. They rely on the companies that have built regional, national, and global networks for carrying Internet traffic, mobile carriers and local broadband providers that enable users to access the

Internet, and cloud companies that rent storage and computing capacity to companies that want to distribute their products and services over the Internet. The online platforms also rely on chip makers, phone manufacturers, computer makers, and other manufacturing businesses for the devices

6 These indirect network effects often go in just one direction for media platforms. Advertisers value access to more users but people often don’t like advertising—they come for the content and tolerate the ads. See the discussion of attention platforms below. 7 Telecommunications is a classic two-sided network even though the same people make and receive calls. So long as there is a difference in the intensity that people make or receive calls the demand to make calls, or to receive calls, depends not just on the number of participants but on the extent to which each participant tends to be on one side of the call or the other. As Rochet and Tirole observed in the seminal paper on two-sided platforms, many cases in which economists have pointed to direct network effects, including telecommunications, are in fact indirect network effects. Jean-Charles Rochet and Jean Tirole (2003), “Platform Competition in Two-Sided Markets,” Journal of the European Economic Association, 1(4): 990-1029; Jean-Charles Rochet and Jean Tirole (2006), “Two-Sided Markets: A Progress Report,” RAND Journal of Economics, 37(3): 645-667. See also Leonard Waverman (2007), “Two-Sided Telecom Markets and the Unintended Consequences of Business Strategy,” Competition Policy International, 3(1): 249-256.

that users rely on to connect to their platforms. Online platforms sometimes choose to build data centers, or content distribution networks, to realize efficiencies. They generally don’t do that until they’ve become successful and the risk of making sunk-cost investments in the business is low.8

As a result, dynamic competition is much easier for online platforms than for traditional network industries. Entrants don’t have to incur significant capital costs in building physical networks. 9 And they don’t face ruinous competition with incumbents who have already made significant sunk-cost investments. Online platforms are, therefore, much more vulnerable to entry and displacement than traditional network firms.10

Google’s experience illustrates the difference in competitive dynamics for online and physical networks. It released its Chrome desktop browser in September 2008 at a time when Microsoft’s

Internet Explorer (IE) accounted for 59 percent of worldwide desktop browser use.11 Microsoft’s IE share had already declined significantly from 87 percent in 2001, largely as a result of the growth of

Firefox, which was provided by the non-profit Mozilla Foundation. 12 By May 2017, Chrome

8 For a discussion of the factors affecting the build-versus-buy decision for a data center, see Ernest Sampera, “Data Center Building vs. Outsourcing: What’s Best for Your Business,” Data Center Knowledge, February 11, 2015, http://www.datacenterknowledge.com/archives/2015/02/11/data-center-building-vs-outsourcing-whats-best-business/. 9 That doesn’t mean they are “free-riding” on those networks which, of course, earn revenue and profits from serving end users, including data charges that are based on intensity of use. 10 The physical networks aren’t the only factor that sets these online and offline platforms apart. Modern telecommunications networks have evolved, mainly since the 1980s, from state-owned monopolies or regulated firms that were shielded from competition. Even as recently as the mid to late 2000s, companies that were trying to produce smart phones, and smart-phone apps, faced significant obstacles with telecommunications companies which wanted to control the phone experience. See Matchmakers, p. 110-112. Apple was able to break the dominance of the telecommunications companies in the late 2000s—not because it was a large platform, which it wasn’t then—but because it had a highly desirable consumer product, the iPhone. See Brian Merchant (2017), The One Device: The Secret History of the iPhone (New York: Little, Brown, and Company), at 4-6; Jacquie McNish and Sean Silcoff (2015), Losing the Signal: The Untold Story Behind the Extraordinary Rise and Spectacular Fall of Blackberry (New York: Peterson Books), at 134-135; Fred Vogelstein (2013), Dogfight: How Apple and Google Went to War and Started a Revolution (New York: Sarah Crichton Books), at 3-4. 11 W3Counter, “Browser & Platform Market Share,” September 2008, https://www.w3counter.com/globalstats.php?year=2008&month=9. 12 TheCounter, “Browser Stats,” http://web.archive.org/web/20110927080509/http://www.thecounter.com/stats/2001/August/browser.php; W3Counter, “Browser & Platform Market Share,” September 2008, https://www.w3counter.com/globalstats.php?year=2008&month=9. Note that the estimates for IE’s shares come from different sources, which may not be perfectly consistent.

accounted for 58 percent of desktop browser and IE and Microsoft’s new product Edge, together only accounted for slightly less than 10 percent.13

By contrast, in 2010 Google launched Google Fiber to deliver fast sired broadband service nationwide in competition with local cable and telco providers.14 It faced significant capital costs in laying fiber optic cable underground.15 It also encountered political barriers to entry in securing rights of way.16 Six years later, with losses mounting, it pulled back.17

B. Online Platforms Leverage Operating Systems as well as the Internet for Fast

Global Distribution

Online platforms also rely on operating systems that sit on top of these physical networks to achieve fast global distribution.18 HTML, combined with various Internet protocols, provides an operating system for developing websites. Browsers enable users and web sites to connect with each other. Search engines enable users to find websites. Operating systems, such as Windows for desktops and the iOS for mobile, provide software and hardware services that online platforms can also use.

Spotify, for example, developed a desktop app that people could download from the web and use to stream music over the Internet. While it had to negotiate music rights to make music available in particular countries it could make its website and app available to anyone with a computer and an

13 W3Counter, “Browser & Platform Market Share,” May 2017, https://www.w3counter.com/globalstats.php?year=2017&month=5. 14 Conner Forrest, “Google Fiber: The Smart Person’s Guide,” TechRepublic, May 26, 2017, http://www.techrepublic.com/article/google-fiber-the-smart-persons-guide/. 15 Adam Levy, “Google Fiber Has Been a Huge Disappointment,” Motley Fool, August 29, 2016, https://www.fool.com/investing/2016/08/29/google-fiber-has-been-a-huge-disappointment.aspx. 16 Berin Szoka, “Don’t Blame Big Cable. It’s Local Governments That Choke Broadband Competition,” Wired, July 16, 2013, https://www.wired.com/2013/07/we-need-to-stop-focusing-on-just-cable-companies-and-blame-local- government-for-dismal-broadband-competition/. 17 Mark Bergen, “Google Fiber Is Pulling Back Its Broadband Rollout As Pressure Grows to Cut Costs,” ReCode, August 25, 2016, https://www.recode.net/2016/8/25/12644888/google-fiber-broadband-cost-cuts; Jon Brodkin, “Google Fiber Division Cuts Staff by 9%, ‘Paises’ Fiber Plans in 11 Cities,” Ars Technica, October 25, 2016, https://arstechnica.com/information-technology/2016/10/google-fiber-laying-off-9-of-staff-will-pause-plans-for-10- cities/; Alexei Oreskovic, “Google Fiber’s CEO Is Stepping Down and the Company Is Halting Plans To Offer Service in Several Cities,” Business Insider, October 25, 2016, http://www.businessinsider.com/google-fibers-ceo-is-stepping- down-and-the-company-is-halting-plans-to-offer-service-in-several-cities-2016-10. 18 Online platforms sit on top of two foundational platforms, one based on software and the other based on hardware, as discussed in Matchmakers, pp. 45-48. Operating systems, as used here, are software platforms that support other software.

Internet connection. Spotify, and other streaming music apps, rapidly eroded Apple’s market leadership in downloadable music as it launched in country after country.

Mobile app platforms—principally iOS and Android—enable online platforms to rely on mobile operating systems for obtaining hardware and software services on smart mobile devices. That reduces the cost of writing apps and expands what those apps can do. Mobile app stores enable businesses to make their apps available, often for free, to every smart mobile user on the app platform.19 Ride-sharing companies, for example, have obtained widespread use by providing apps, easily and simply, to drivers and passengers through the Apple App Store and Google Play. They didn’t need significant additional physical infrastructure to provide services in almost every major city in the world.20 That has made it possible for Uber, for example, to expand its ride-sharing business globally, but also for other firms to challenge Uber without incurring the significant sunk costs associated with establishing physical networks.

C. Online Platforms Can Easily Add Features Because They Are Based on Software

When traditional businesses want to expand into a new product category, they often have to modify assembly lines, build new factories, upgrade equipment, and secure suppliers. Online platforms are built from software code, which is highly malleable, making it possible for firms to add features, including highly complex ones. 21 Facebook, as shown below, completely rebuilt its iOS app in less than six months, providing a much more desirable product for consumers.22

19 Apple and Google charge commissions only to apps that use their app store payment mechanisms. Apple requires that apps providing paid digital content to users acquired through its app store to use the iOS app store payment mechanisms. Sellers of physical goods may use alternative payment mechanisms, and sellers of digital goods may use alternative payment mechanisms for users who sign up outside of the app store. Google allows all apps to use alternatives to its Google Play. Apple and Google charge developers a small free for getting access to their software development kits and other aids for app development. 20 See Matchmakers, Chapter 3. 21 Modifying software code to add features can be risky since software engineers need to make sure they don’t break the existing functionality when they add new code and take time for highly complex software. These difficulties, though, pale in comparison to traditional industries. 22 Facebook began an effort to redesign its mobile apps in the spring of 2012, and launched completely rebuilt versions of its apps in August 2012 (iOS) and December 2012 (Android). Robert D. Hof, “How Facebook Slew the Mobile Monster,” Technology Review, March 6, 2013, https://www.technologyreview.com/s/511781/how-facebook-slew-the-mobile-

Traditional businesses also have to secure distribution for their new products. That often requires persuading physical distributors that they should make scarce store space available. By contrast, once online platforms incorporate these new features, it is trivial for them to make those features available to users. Microsoft Windows can quickly make new versions of its software available for download to everyone who has a valid license.

Traditional businesses usually obtain customer feedback slowly.23 They may have call centers to report complaints, conduct consumer surveys, and evaluate product reviews. Most importantly they observe the trajectory of sales. Online platforms obtain data from users on the performance of the platform continuously. They typically use that data to improve performance and identify opportunities for new features. For example, Pandora uses data on the songs liked by its 76 million listeners to immediately identify other songs that each listener might also like, varying with the time of day and the device used. Thus, Pandora can continuously identify new songs that its listeners didn’t know they liked.24

Online platforms typically innovate continuously, rolling out incremental as well as disruptive innovations because it is relatively easy for them to do so. Moreover, since all online platforms can engage in this continuous innovation, at fairly low cost, any platform that fails to do so would face a loss of position and, through the reversible indirect network effects discussed below, potential demise. Not surprisingly online platform providers often compete intensively with each other by adding features, to improve their basic product offerings and for entering new categories.

monster/; Matt Brian, “Facebook Launches Native App for iPhone and iPad, Rebuilt from the Ground Up To Be Twice as Fast,” The Next Web, August 23, 2012, https://thenextweb.com/apple/2012/08/23/facebook- 2/#.tnw_OTE48pGm; Mark Gurman, “Facebook Launches Much Faster Android App with Quicker Launching, Photo Loading,” 9to5Google, December 13, 2012, https://9to5google.com/2012/12/13/facebook-launches-much-faster- android-app-with-quicker-launching-photo-loading/. 23 The Internet-of-Things will improve this situation in the coming years as many physical products, such as microwaves, have embedded internet connectivity. 24 Sherice Jacob, “How Pandora Uses Data to Improve Its Service and Music Stations,” Kissmetrics Blog, May 15, 2017, https://blog.kissmetrics.com/how-pandora-uses-data/; Glen Peoples, “Ten Things Pandora Is Doing To Help Move the Industry Forward,” Medium, February 3, 2017, https://medium.com/@glennpeoples/10-things-pandora-is-doing-to- help-move-the-industry-forward-48338d3e89cb.

The history of portal competition illustrates the process, which applies generally. Table 1 below shows the timing of selected new features introduced by the major web portals Yahoo, MSN, and AOL. Yahoo and AOL offered a web search engine in 1995, and Microsoft launched Bing in

1998. In 1996, Microsoft created Expedia and offered it through its portal. Yahoo added travel services in 1997 and AOL in 2000. Yahoo and MSN (Hotmail) offered free email starting in 1997, which AOL finally matched in 2005. Yahoo offered a shopping comparison engine in 1998, followed by AOL in 1999 and Microsoft in 2006. AOL began offering mapping services with its acquisition of

MapQuest in 2000. This was matched by Yahoo in 2002 and MSN in 2005. In the mid-2000s, the main portals responded to the success of Myspace by aggressively adding social networking services.

AOL was first with AOL Journals in 2003, followed by MSN Spaces in 2004 and Yahoo! 360 in 2005.

Yahoo was the first portal to add a Q&A service, in 2005, followed by MSN in 2006 and AOL in

2007.

D. Online Platform Users Can Multihome and Switch Platforms at Low Cost

In traditional network industries, people typically single-home and switching is costly if there is even an alternative.25 For example, people have only one wired broadband provider for their household.

They would have to pay for an additional subscription to have another. It is not simple to add or drop a wired broadband provider. Consumers often get broadband as part of a bundle from their cable provider. Switching cable providers is notoriously difficult in the United States. Consumers have to get new wires into their households and new set-top boxes when they switch cable providers. Some cable providers require consumers to return the set-top box, possibly by taking it to a physical location, before they will discontinue billing. The prevalence of single-homing and switching costs are true for electricity, gas, water, and commuter rail, as well as for durable consumer products that require investments in collateral devices

25 David S. Evans (2014), “Economic Analysis of the Impact of the Comcast/Time Warner Cable Transaction on Internet Access to Online Video Distributors,” https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2600715.

Table 1: Selected Features at Major Portals by Year of Launch or Acquisition

Yahoo! MSN AOL Auctions Yahoo! Auctions (1998-2007) - - Automobiles Yahoo! Autos (2001) CarPoint (MSN Autos) (1996) AOL Auto Channel (AOL Autos) (2000) Blogging and Yahoo! 360 (2005) MSN Spaces (2004-2010) AOL Journals (2003-2008) Social Networks Tumblr (2013) AIM Pages (2006-2007) AOL Bebo (2008-2010) Discussion and Yahoo! Answers (2005) Microsoft Communities (MSN AOL Answers (2007 - ~2015) Q&A Yahoo! Groups (2001) Groups) (1991-2009) Windows Live Groups (OneDrive Groups) (2008- 2015) Windows Live QnA (MSN QnA) (2006-2009) E-mail Yahoo! Mail (1997) Hotmail (Outlook.com) (1997) AOL Mail (1989, 2005) Entertainment Yahoo! Entertainment (1999) Bing Entertainment (MSN Moviefone (1999) and Lifestyle Yahoo! Movies (1998) Entertainment) Cambio (2010) Launch Media (Yahoo! Music) AOL Makers (2012) (2001) Style Me Pretty (2012) Maps Yahoo! Maps (2002-2015) MSN Virtual Earth (Bing MapQuest (2000) Maps) (2005) Messaging Yahoo! Messenger (1999) MSN Messenger (1999-2013) AIM (1997) Skype (2011) News Yahoo! News (1996) MSN News (1996) Engadget (2005) Slate (1996-2004) Huffington Post (2011) TechCrunch (2010) Photos Yahoo! Photos (2000-2007) Xim (2014) AOL Pictures (2005-2008) Flickr (2005) OneDrive (2014) Photobucket (2014) Search Yahoo! Search (1995) MSN Search (Bing) (1998) AOL Search (1995) Shopping Yahoo! Shopping (1998) Windows Live Shopping (MSN Shop@AOL (AOL Shop) Shopping / Bing Shopping) (1999) (2006-2013) Sports Rivals.com (2007) MSN Sports (2001) FanHouse (2006-2011) Yahoo! Sports (1997) AOL Sports (2015-2016) Travel Yahoo! Travel (1997) Expedia (1996-2003) AOL Travel (2000) Farecast (MSN Travel) (2008) Video Yahoo! Video (Yahoo! Screen) MSN Video (Bing Videos) AOL TV (2000-2003) (2006-2016) (2004) AOL On Network (AOL Yahoo! View (2016) Video) (2012) AOL Build (2014)

Increasingly, online platforms are mobile-centric or at least offer an app. It is easy to add and subtract apps from devices. The number of apps that people can use is limited mainly by their ability to keep track of icons or bookmarks. Many online platforms are available to users for free.

Consumers face no significant expense in installing an additional online platform app. Nor must they agree to use an online platform exclusively.

As a result of these features, it is easy for people to multihome on several online platforms.

They can conveniently switch back and forth depending upon which they prefer for which purpose. It is also easy for people to try a new alternative online platform and decide whether to keep using it or not. The same is true for business users on online platforms, such as advertisers and gig-economy workers. Large advertisers often use multiple online platforms, including ones in the same narrow category such as search, and switch advertising budgets between them.26 Drivers for ride-sharing apps can, and many do, use more than one app.27 Multihoming is so prevalent among online platforms that it is hard to identify exceptions.

The prevalence of multihoming, and switching, between platforms is inconsistent with the claim that data provides a substantial barrier to entry. Time and again new platforms arise, with no data at inception, and acquire consumers and obtain data over time. That doesn’t mean that data isn’t valuable. It does strongly suggest that lack of data doesn’t pose significant obstacles to online platforms that develop valuable products that consumers like.28

E. Indirect Network Effects Work in Reverse

Platforms, whether online or physical, must attract enough members of each group of users to provide enough value to the other group of users.29 Buyers won’t use a marketplace, whether online or physical, unless it has enough sellers, and sellers won’t use a marketplace unless it has enough buyers. Platforms that crack this chicken-and-egg problem can grow very rapidly. More users

26 Justin Freid, “Does Bing Deserve More of Your Money?” Search Engine Watch, February 4, 2015, https://searchenginewatch.com/sew/opinion/2393523/does-bing-deserve-more-of-your-money. 27 SherpaShare, “Map: 65% Earning from Two or More Platforms,” August 13, 2014, https://www.sherpashare.com/share/map-65-earning-from-2-or-more-platforms/. 28 Although the sweeping statements about the importance of data do not have empirical support, it is possible that there are situations in which data could prove to be a barrier to entry and enhance market power. Empirical analysis is necessary to determine whether that is true, and whether the collection and possession of data facilitates consumer harm, on net, in violation of competition laws in those specific circumstances. 29Matchmakers, Chapter 5 on the solving the chicken and egg problem and securing critical mass.

in each group attract more users in the corresponding group. As noted above, economists refer to the positive feedbacks between groups as “indirect network effects.”

Indirect network effects do not necessarily lead to monopoly. They may become exhausted at some point, as additional users don’t add much value. Several platforms can soften the impact of indirect network effects by differentiating themselves on one or both sides and thereby appealing to different users. Congestion can limit the value of additional users. In fact, multisided platforms with indirect network effects are seldom monopolies.30 The exceptions that are sometimes pointed to, such as Windows, don’t prove the rule.

Some commentators claim that these network effects result in an insurmountable barrier to entry and point to some physical networks, such as telephones and railroads, as proof. There is a fundamental difference, however, between some physical platforms and most online platforms.

Indirect network effects can be difficult to reverse for physical networks because users made specific capital investments in those networks that they would have to duplicate if they joined another network. That can make it hard for customers to try another network or to switch altogether. Indirect network effects help protect physical platforms from entry. To try a new telephone service, for example, customers of AT&T’s local exchanges at the turn of the 20th century had to get an additional wired connection to their home, use an additional telephone handset, and pay for an additional subscription.31

30 For further discussion, including other reasons why indirect network effects do not lead to monopoly typically, see David S. Evans and Richard Schmalensee (2015), “The Antitrust Analysis of Multisided Platform Businesses,” in Roger D. Blair and D. Daniel Sokol (eds.), The Oxford Handbook of International Antitrust Economics, Vol. 1 (Oxford: Oxford University Press), 404-448, at 415-419; David S. Evans and Richard Schmalensee (2016), Matchmakers: The New Economics of Multisided Platforms (Boston: Harvard Business Review Press), at 124-126. 31 Even these switching costs didn’t prevent competitive entry after AT&T’s patents expired. Independent telephone companies entered in parts of cities that AT&T didn’t serve or served poorly and, by 1907, accounted for 51 percent of telephone subscribers. AT&T was eventually allowed to consolidate competing telephone exchanges and was protected by regulatory barriers to entry. “AT&T” refers to the various entities that evolved from the original Bell Telephone Company. These entities were affiliated with each other under sometimes complex ownership structures over the years. Robert Bornholz and David S. Evans, “The Early History of Competition in the Telephone Industry,” in David S. Evans, ed., Breaking Up Bell (New York: Elsevier Science Publishing Co., 1983. See also Peter Temin with Louis Galambos (1987), The Fall of the Bell System: A Study in Prices and Politics (Cambridge, UK: Cambridge University Press).

By comparison, indirect effects are easier to reverse for online platforms because of the multihoming and lower switching costs described above. Users can add another platform to try it out which makes it easier for a challenger to crack the chicken and egg problem. Users can drop a platform, or reduce their use of it, if they find a better alternative. If users of one group drop it, the platform becomes less valuable to users of the other group, and indirect network effects, working in reverse, propel the decline of the platform.

The history of communications platforms—messaging apps and social networks—over the last two decades illustrates the importance of reverse indirect network effects as well as the data that comes along with users. People value communications platforms that have more of the people they want to interact with. A naïve view of indirect network effects implies that a successful communications platform would be secure from competition since people wouldn’t join or use a platform that didn’t include most of their personal network. The flaw in that reasoning is that people can multihome on online communications platforms. A few people in a network try a platform. If enough join, and like it, then eventually all of them could switch or drop the initial platform. This phenomenon has happened repeatedly. AOL, MSN Messenger, Friendster, Myspace, and Orkut all rose to great heights, and then rapidly declined, while Facebook, Snap, WhatsApp, Line, and others quickly rose. Nothing about the underlying economics or technology of online platforms has changed that would prevent this same cycle from repeating itself going forward.

The possibility of reverse indirect network effects doesn’t mean that online platforms necessarily die. By competing dynamically, through offering new innovative features and services and increasing value to users, online platforms can retain their momentum and forestall decline. Indirect network effects can therefore act as a competitive constraint rather than an entry barrier.

F. Many Online Platforms Compete for Attention and Advertising

Facebook and Google derive virtually all of their revenue from selling advertising—97 percent for Facebook and 88 percent for Alphabet, Google’s parent.32 Amazon and Microsoft also operate significant online advertising businesses. Six of the top ten websites, and nine of the top ten mobile apps, based on U.S. unique visitors, also sell advertising and appear to derive most of their income from ads.33 They all operate “attention platforms.”34

Attention platforms, which include traditional ad-supported media such as television, solve a basic problem in commerce.35 Businesses would like to deliver ads to those consumers that are most likely to buy from them based on being exposed to the ad. Consumers generally don’t like many kinds of ads and wouldn’t seek them out. So long as businesses value delivering an ad message more than a consumer would pay to avoid it, there is room for a value-increasing exchange. Attention platforms generate value by providing users with valuable content and then charging the advertisers for the opportunity to present an ad to the consumer. Almost all physical attention platforms make access to consumers free, as is the case for free radio and TV, or charge consumers a fee that defrays the cost of distribution but not of providing the content, as is the case for many newspapers and magazines.

Online attention platforms have changed the traditional ad-supported media model in two critical ways. Traditional media show the same ad to many people and don’t have detailed information on who engaged with the ad. Everyone sees that same ad in this month’s Vogue Magazine or during a commercial break on CNN. Online attention platforms, which have an Internet connection with each viewer, can customize the ad in real time to the person viewing the ad. That increases the likelihood

32 Facebook, 10-K for Year Ending December 31, 2016, at p. 9; Alphabet, 10-K for Year Ending December 31, 2016, at p. 13. 33 The top 10 websites in the United States are based on visitor data from Alexa. Alexa, “Top Websites in the United States,” http://www.alexa.com/topsites/countries/US, visited June 16, 2017. The top 10 apps are based on April 2017 estimates of unique visitors from comScore (http://www.comscore.com/Insights/Rankings). The top 10 websites are Google.com, YouTube.com, Facebook.com, Reddit.com, Amazon.com, Wikipedia.com, Yahoo.com, Twitter.com, Netflix.com, and Ebay.com. The top 10 apps are Facebook, YouTube, Facebook Messenger, Google Search, Google Maps, Google Play, Instagram, Gmail, Snapchat, and Pandora Radio. 34 David S. Evans (2013), “Attention Rivalry Among Online Platforms,” Journal of Competition Law and Economics, 9(2): 313-357. 35 Matchmakers, pp. 128-129.

that the viewer will see a relevant ad and that the advertiser pays to present an ad to a relevant consumer. Online attention platforms use customer data to improve performance for advertisers who reach the most likely actual consumers of their products, and for consumers who receive targeted advertising and fewer irrelevant adds.

Online attention platforms compete with each other, in various dimensions, even though, from the perspective of a consumer, they seem to be doing different things. On the consumer side of the platform, they are all vying directly against each other to attract attention that can then be monetized through advertising. On the other side, they are also vying against each other for advertising spending. Advertisers, to maximize their investment, allocate their spending across attention platforms—online and physical—based on the return on investment they will make from presenting ads to consumers through these different channels.

Commentators who claim that categories are “winner-take-all” and that the winner is

“unstoppable” miss the fundamental feature of much of Internet competition. Even if a category is really winner-take-all, the so-called victor basically wins the opportunity to provide valuable services to consumers for free. The victor then has to compete for advertising dollars with all the other winners.

Twitter, for example, has “won” the micro-blogging category outside of China. A decade after its start, with 319 million monthly active users worldwide, the company has struggled to sell enough advertising to cover its costs and make a profit. Part of its difficulty, according to analysts, is competing for advertising campaigns with other online platforms, such as Google and Facebook, neither of which provide the same microblogging functionality as Twitter, but both of whom compete with Twitter for both user attention and advertiser dollars.36

36 Timothy Green, “Google’s Biggest Threat in 2014: Facebook and Twitter,” Motley Fool, January 1, 2014, https://www.fool.com/investing/general/2014/01/01/googles-biggest-threat-in-2014.aspx; Greg Miller, “Facebook’s Plan to Remain King of Social Media,” Wall Street Daily, August 20, 2015, https://www.wallstreetdaily.com/2015/08/20/facebook-social-media-competition/; Christopher Heine and Marty Swant,

Ad-supported Internet-based platforms that seem to dominate their categories are therefore quite unlike traditional firms with large market shares. The ad-supported platforms are competing for attention on one side and for advertising dollars on the other.

G. Online Platforms Face Frequent Disruptive Innovation

Online platform competition is dynamic and unpredictable because waves of disruptive innovation expand opportunities for entry and pose challenges to incumbents. Online services for computer users began in the mid-1980s with the launch of various online providers such as AOL.

These companies provided email, chat rooms, information, and other services. Computer users connected with a telephone modem. By 1994 AOL had one million users all of whom had aol.com e- mail addresses.37

In the mid-1990s the development of the World-Wide-Web, the opening of the commercial

Internet for connecting users, and the availability of easy-to-use web browsers for rendering web pages launched the desktop-centric online world. The steady development and deployment of ever faster wired broadband connections accelerated the online growth. Despite its large user base, hoard of data, and brand name, AOL was not able to make the transition to a web-based environment.

Yahoo was able to leapfrog AOL.

A bit more than a decade later the iPhone-led revolution, based on a pioneering device, operating system, and app distribution platform and followed closely by the Android platform, gradually eliminated the smart phone leaders, Microsoft, Blackberry, and Symbian. Each of those leaders had a significant installed base, data, applications, and network effects. The smart phone revolution also challenged online attention platforms, such as Yahoo, that had difficulty making the transition to mobile devices.

“What Ad Buyer’s Really Think About Google, Facebook, Twitter, and Everything in Between,” Adweek, September 5, 2016, http://www.adweek.com/digital/what-ad-buyers-really-think-about-google-facebook-twitter-and-everything- between-173268/. 37 Wired, “America, Online!” September 1, 1995, https://www.wired.com/1995/09/aol-2/.

In 2014, just seven years after Apple introduced the iPhone, Amazon launched a voice- activated platform called Alexa that relies on voice recognition and artificial intelligence to complete tasks. Developers can write apps, called “skills,” for the Alexa platform. As of January 2017, more than 8 million U.S. households have Amazon Echo devices running Alexa, and Alexa has more than

10,000 skills.38 Other companies, including Apple, Google, and Microsoft are developing similar platforms. Analysts forecast that the sales of voice-first devices such as Amazon Alexa and Google

Home will reach 24.5 million in 2017, raising the installed base to 33 million.39 It is too soon to tell who, if anyone, will fall prey to the voice-activated platforms.

Facebook’s experience exemplifies the difficulty that these disruptions can cause incumbents.

By 2012, Facebook was the leading social network. It had 900 million active monthly users compared to 35 million for Myspace, the previous leader in the U.S. and many other countries.40 In its filings for its impending IPO, it reported $3.7 billion of advertising revenue and operating income of $1.8 billion for the year ending December 31, 2011.41 It earned this advertising revenue primarily from serving ads to people who accessed Facebook on their desktop with a browser. Eyes, however, were rapidly shifting to mobile devices where its ads were less effective. Analysts sharply questioned its ability to make the transition to mobile. As its failure at developing a mobile app became apparent, its stock

38 Taylor Soper, “More Than 8M People Own an Amazon Echo as Customer Awareness Increases ‘Dramatically,’” GeekWire, January 25, 2017, https://www.geekwire.com/2017/8-million-people-amazon-echo-customer-awareness- increases-dramatically/; Brian Barrett, “Amazon Aleza Hits 10,000 Skills. Here Comes the Hard Part,” Wired, February 23, 2017, https://www.wired.com/2017/02/amazon-alexa-hits-10000-skills-plenty-room-grow/. 39 VoiceLabs, “The 2017 Voice Report by VoiceLabs,” January 15, 2017, http://voicelabs.co/2017/01/15/the-2017- voice-report/. 40 Facebook monthly active users are for May 2012. Facebook, Amendment 8 to Form S-1 Registration Statement, May 16, 2012. Myspace monthly active users are for May 2011. Felix Gillette, “The Rise and Inglorious Fall of Myspace,” Bloomberg Businessweek, June 22, 2011, https://www.bloomberg.com/news/articles/2011-06-22/the-rise-and- inglorious-fall-of-myspace. 41 Facebook, Amendment 8 to Form S-1 Registration Statement, May 16, 2012. Myspace monthly active users are for May 2011.

value plummeted. By November 30, 2012 its stock was trading at 26 percent lower than the IPO price.42

Facebook had to scramble to make an effective transition to mobile. It trained hundreds of programmers in iOS and Android development and assigned mobile engineers to every developer team so that all new features could be built with a “mobile-first” mindset. The company went from updating its mobile apps every three to six months to updating them every month.43 It introduced new ad formats, like mobile app installation ads, and switched from banner ads and pop-up ads to native ads integrated into the user’s feed.44 Facebook also switched from developing an HTML5 mobile app to native apps for iOS and Android for better speed and performance.45 Today, Facebook earns 83 percent of its advertising revenue from mobile.46

If Facebook hadn’t made the transition it would have been in a far different position.

Successfully making the move to mobile wasn’t inevitable, as the failure of AOL and Yahoo to bridge similar waves of disruptive innovation demonstrates. The threat of losing user attention and advertising, as consumers spent more time on mobile devices, impelled Facebook to innovate in ways that ultimately provided value to consumers and advertisers. This sort of do-or-die innovation plays out constantly for online platforms.

III. Dynamic Competition Among Online Platforms Between 1995 and 2016

The history of online competition shows that even the most successful online platforms at any point in time can’t sleep well at night because they face existential threats, and provides no

42 Facebook’s IPO price was $38.00, and it closed at $38.23 on its first trading day. It’s closing price on November 30, 2012, was $28.00, 73.7 percent of its $38.00 IPO price. 43 Kurt Wagner, “Inside Facebook’s Mobile Strategy,” Mashable, September 30, 2013, http://mashable.com/2013/09/20/facebook-mobile-strategy/#QN6i76SXmgq1. 44 Robert D. Hof, “How Facebook Slew the Mobile Monster,” MIT Technology Review, March 6, 2013, https://www.technologyreview.com/s/511781/how-facebook-slew-the-mobile-monster/. 45 Kurt Wagner, “Inside Facebook’s Mobile Strategy,” Mashable, September 30, 2013, http://mashable.com/2013/09/20/facebook-mobile-strategy/#QN6i76SXmgq1. 46 Facebook, 10-K for Year Ending December 31, 2016, at p. 41.

support for the view, now pushed by various commentators, that Internet leaders have secure positions in the categories they serve. To be clear, the point isn’t that Internet leaders will die regularly, but that the threat of demise made tangible by fallen leaders from the past, forces leaders of the present to engage in constant intense innovative competition.

A. Dynamics of Competition Among Current Platform Leaders

Discussions of the state of competition among Internet-based firms often begin with the observation that five of them are now the most valuable companies in the world. Given the ability of

Internet-based firms to disrupt traditional industries, globally, as well as to provide new products and services, it is not surprising that successful Internet firms have significant market caps. The market- cap charts illustrate, however, the vigor of online competition rather than its diminution.47 Every few years, new firms, that challenge old firms in important ways, bubble to the top.

Table 2 shows the top five U.S.-incorporated firms by market capitalization for 1990, 2000,

2010, and 2017. At the start of this 27-year period only two of the five Internet giants even existed—

Apple and Microsoft—but neither ranked in the top 10 firms. IBM, which was a significant provider of hardware and software for the emerging online economy, and AT&T, which provided portions of the physical networks for the digital economy, were in the top 10. IBM had the highest market cap.

Both firms possessed troves of data. Ten years later, IBM had tumbled to spot 14. Cisco, which provided infrastructure for the Internet was number 3. Microsoft had vaulted to 5.

By 2010, Apple, mainly on the strength of its 2007-introduction of the iPhone, was number 2 followed by Microsoft at number 3. Alphabet (Google accounts for almost all its revenue), based on its success as search-based advertising platform, was number 7. That year Facebook was still two years away from an IPO and Amazon ranked 32nd in market cap. Seven years later, Amazon and

47 This dynamic competition does not mean, of course, that these large Internet firms could not have significant market power in particular categories, or engage in exclusionary or predatory practices. It does caution, however, against taking narrow views of markets that ignore dynamic competition.

Facebook joined Alphabet, Apple, and Microsoft in the top five. IBM had fallen to 32nd largest firm based on market cap.

Unlike the largest firms at previous points in time, these large Internet firms compete with each other across a range of products and services, despite each having gotten a toehold in the digital economy doing completely different things from one another. They compete in the near term (what

Table 2: Top Five U.S.-Incorporated Firms by Market Capitalization 1990 2000 2010 2017

1 International Business General Electric Corp. Exxon Mobil Corp. Apple Inc. Machines Corp. 2 Exxon Mobil Corp. Exxon Mobil Corp. Apple Inc. Alphabet Inc.

3 General Electric Co. Cisco Systems Microsoft Corp. Microsoft Corp.

4 Bristol-Myers Squibb Co. Wal-Mart Stores Inc. Berkshire Hathaway Inc. Amazon.com Inc.

5 Merck & Co. Microsoft Corp. General Electric Co. Facebook Inc.

Source: S&P Capital IQ. Data for 1990, 2000, and 2010 are from December 31 of each year. Data for 2017 are for May 19, 2017. Outlier estimates for privately held companies have been removed from the S&P data. economists call static competition): Amazon, Facebook, Google, and Microsoft, for example, all compete for advertising and promotional dollars. The largest Internet firms also compete long-term

(what economists call dynamic competition): Amazon, Apple, Google, and Microsoft are all engaged in an intense race involving voice-activated platforms. Table 3 shows the current overlap between these firms in a number of categories.

Table 3: Selected Products and Features of Largest Online Platforms

Alphabet Inc. (Google) Facebook Inc. Apple Inc. Microsoft Corp. Amazon.com Inc. Search Google Search _ _ Bing Amazon.com (product search) Browser Google Chrome _ Safari Internet Explorer, Edge _

Mail Gmail _ iCloud email Outlook _ Messaging Google Allo, Google Facebook Messenger, iMessage MSN Messenger, _ Hangouts WhatsApp Yammer, Microsoft Team Voice/Video Google Duo, Google Facebook Messenger, FaceTime Skype, Office 365 Video, Echo Show, Alexa App, Calling Allo, Hangouts WhatsApp Microsoft Stream, Amazon Chime Maps Google Maps, Google _ Apple Maps Bing Maps, StreetSide _ Earth, Waze Payments/Wall Android Pay, Google Facebook Payments Apple Pay, Apple Wallet Windows Wallet Amazon Pay, Amazon ets Wallet Wallet Operating Android, ChromeOS _ iOS, macOS Windows Amazon Fire OS Systems Social Google+ Facebook, Instagram _ LinkedIn (2016) Twitch (narrow- gaming Networking focused), Goodreads (also narrow- books focused) Workplace G-Suite, Google+ Workplace iWork Office, Office365, Amazon Work Docs, Collaboration Microsoft Teams Amazon Chime and Enterprise Productivity Software Cloud Storage Drive, Google Cloud _ iCloud OneDrive, SkyDrive, Amazon Drive, Amazon and Cloud Platform (for enterprise) Microsoft Azure Web Services Computing Photo Storage Photos Photos iCloud Photo Library OneDrive Prime Photos

App Store Google Play _ App Store Windows Store Amazon App Store Artificial Deepmind, TensorFlow, Facebook AI Research Caffe, integration into Microsoft Artificial Amazon AI Intelligence integration into various various other products Intelligence Program other products Autonomous Waymo, Android Auto _ Apple Car (autonomous Software development _ Vehicles (voice activated system driving system, no programs underway for vehicles) physical car) Mobile Pixel, Android One _ iPhone Lumia Fire Handsets Tablets / Basic Android Tablets, _ iPad Surface Fire tablets PCs Chromebook Wearables Android Wear, Google _ Apple Watch Band _ Watch Virtual Google Daydream, Oculus, Augmented ARKit HoloLens, Mixed Reality Reality/Augme Tango, TiltBrush (3D reality tools nted Reality painting in VR)

Streaming Chromecast _ Apple TV Xbox Amazon Fire Stick, Device Amazon Fire TV Voice- Google Home, Google M (2015), Messenger Siri, HomePod, Can be Cortana, Bot Framework, Echo/Alexa, Amazon Lex Activated/Virt Assistant Bots (2016) built onto iMessage Tay, Zo, Ruuh, etc. ual Assistant, Chatbots, Smart Speakers Video YouTube, YouTube Live Live (2015), Videos, 360 Clips (video editing for Twitch Aggregation / Videos posting on other Live Video platforms) Entertainment: Google Play Games iTunes, App Store Windows Store (no Amazon Music, Video, Music, Movies, eBooks), Xbox Live, Game Studio, Kindle eBooks, Games Windows Games eBooks

Online AdWords, AdSense In NewsFeed _ Bing Ads, Amazon Advertising Advertising, Analytics, DoubleClick, Audience Network Bing Network (includes ad network)

Ad Network Tag Manager

Local Google Places/Reviews Places, Pages (similar Apple Maps (business Bing Places for Business _ Directory function) profiles)

News Google News, Google NewsFeed, Instant Apple News Linked-In _ AMP Articles

Shopping Google Shopping _ _ Bing Product Search Amazon.com

In fact, each of these firms has posed, and continues to pose, a significant threat to at least one of the other firms. In 2006, just ten years ago, for example, Microsoft appeared “unstoppable.” It licensed the operating system software for most personal computers in the world. It looked like it was going to extend that position into mobile. It had the third largest share of operating systems for smartphones (behind Symbian and Linux, and ahead of Blackberry)48 and the most sophisticated platform for app developers. But Apple’s iOS and Google’s Android platform leapfrogged Microsoft.

Together they eliminated Microsoft as a serious competitor in mobile.49 Importantly, though, mobile operating systems compete for user and developer interest with desktop operating systems. Microsoft still earns significant revenues from licensing Windows for desktop computers, as IBM did from mainframes in 1990, but its operating system platform is not as relevant to many of the innovative businesses in the digital economy today as it was in the past.

In 2012, Google accounted for 41.3 percent of U.S. online advertising revenue. The second largest provider, Yahoo, accounted for only 8.4 percent.50 Most of this revenue came from desktop ads but the company looked well positioned to extend its search advertising business into mobile.

Google Search had an exclusive agreement with Apple for the iPhone and was bundled into most

Android handsets. Four years later, Google faced a much more significant competitor in online

48 http://www.roughlydrafted.com/RD/RDM.Tech.Q1.07/BEC05CE1-D5EB-4E48-B46C-7385D5AADCFE.html. 49 Aside from limiting its ability to engage in anticompetitive practices the U.S. and EU litigation involving Microsoft did not place any apparent constraints on its ability to compete in mobile. 50 https://www.emarketer.com/Article/US-Digital-Ad-Spending-Top-37-Billion-2012-Market-Consolidates/1009362.

advertising. Facebook had become the second largest provider of online advertising; as of 2016 it had a 16.6 percent share compared to Google’s 41.0 percent share.51 But on mobile, which is accounting for an increasing share of people’s time, Google accounted for 31.5 percent of U.S. ad revenue, compared to Facebook’s 22.5 percent.52

Over the same time period, Google has faced increasing pressure from Amazon. Google earns virtually all of its revenue from serving ads to consumers who are looking for products. As

Amazon has grown in prominence many consumers are starting their product searches on Amazon, and in many cases then purchasing from Amazon. Estimates vary as to the exact magnitude of this shift. PowerReviews estimates that 38 percent of U.S. online shoppers start their search on Amazon, compared to 35 percent who start on Google.53 BloomReach estimates that 55 percent of U.S. online shoppers start their online search on Amazon, compared to 28 percent who start with a search engine.54

Just focusing on the firms with the highest market capitalizations today shows the firms that were leading at each point in time based on market capitalization faced more competition over the next decade rather than less. Most of the so-called Internet giants face direct competition from several of the other giants as well as from much smaller firms. History indicates that some of those smaller firms will likely become giants themselves in the coming decade. After all, three of the five so- called Internet giants were much smaller firms just a decade ago.

B. Dynamic Competition Among Attention Platforms

51 http://bluemoondigital.co/wp-content/uploads/2016/12/eMarketer_US_Ad_Spending- eMarketers_Updated_Estimates_and_Forecast_for_2015%E2%80%932020.pdf. 52 http://www.adweek.com/digital/u-s-digital-advertising-will-make-83-billion-this-year-says-emarketer/. 53 http://searchengineland.com/survey-amazon-beats-google-starting-point-product-search-252980. 54 https://www.recode.net/2016/9/27/13078526/amazon-online-shopping-product-search-engine.

Many significant Internet-based firms over the last two decades have operated attention platforms that focus on developing content that attracts viewers and selling advertisers access to those viewers.55 The flux in the leading firms, based on the amount of attention they attract, shows the dynamism in this significant part of the digital economy. Table 4 shows the top 10 properties at various points in time based on published reports by comScore. Of the top 10 sites in each year, 8 dropped off from 1996-2006, 4 dropped off from 2006-2011, and 4 dropped off from 2011-2016.

Table 4: Top Ten Web Properties by Unique Visitors

Jan-1996 Jan-2006 Jan-2011 Jan-2016 1 AOL.com Yahoo! Yahoo! Google 2 WebCrawler.com MSN-Microsoft Google Facebook 3 Netscape.com Time Warner Network Microsoft Yahoo 4 Yahoo.com Google Facebook.com Microsoft 5 Infoseek.com eBay AOL, Inc. Amazon 6 Prodigy.com Amazon Ask Network AOL, Inc. Comcast 7 Compuserve.com Ask Jeeves Turner Digital NBCUniversal 8 UMich.edu MySpace.com Viacom Digital CBS Interactive New York Times 9 PrimeNet.com Glam Media Apple Inc. Digital Mode Media (formerly 10 Well.com Viacom Online CBS Interactive Glam Media) Source: comScore. Note that prior to 2013, comScore’s top web properties were based on desktop visitors only, while after 2013 they are based on combined desktop and mobile visitors.

The dynamics of competition among attention platforms is seen in the rapid rise of Snapchat.

This camera-based communication application started in September 2011. Snap reported an average of 158 million daily users in its S-1 filing in February 2017. It showed its first ad in October 2014. In

2015, it generated more than $400 million in advertising revenue. As of June 1, 2017, it had a market cap of $25.2 billion.56

Like many attention platforms, Snap started by coming up with a differentiated way of attracting attention. It focused on providing ephemeral camera-based communications between

55 These results are consistent with David S. Evans (2013), “Attention Rivalry Among Online Platforms,” Journal of Competition Law and Economics, 9(2): 313-357; Andre Boik, Shane Greenstein, and Jeffrey Prince (2016), “The Empirical Economics of Online Attention,” NBER Working Paper No. 22427. 56 S&P Capital IQ.

people.57 This attracted users who likely shifted some of their time from other online and physical attention platforms such as Twitter. Once it had attracted significant user attention it could offer advertisers the ability to reach those users with ads. Snap developed innovative advertising products that it used, together with its user base, to compete for ad spending with other attention platforms.

That gave advertisers another option and likely diverted spending from other platforms.

Competition from Snapchat for users and advertisers places pressure on other platforms.

Those platforms can compete with different features to attract users or they can mimic features that

Snap has introduced. Snap’s success, for example, induced a competitive response from Facebook on both Instagram (a camera-based property it owns) and on its main Facebook property. Facebook launched Instagram Stories, a feed of ephemeral content similar to Snapchat’s central feed.58 In turn,

Snapchat has been working to improve its audience segmentation ability in order to attract advertisers away from Facebook.59 Snap also has to worry about competition from entrants as well as other platforms. As it points out “[T]he barrier to entry for entrants is low, and the switching costs to another platform are also low.”60

C. The Durability of Platform Leadership

The rise and fall of Yahoo illustrates the forces that propel firms to great heights but then accelerate their decline. Yahoo went live in January 1995. By January 1996, Yahoo had about 5.8 million monthly unique visitors.61 Having gone public in 1996 Yahoo’s market value as of July 1, 1998 was $8.3 billion and $46.2 billion on July 1, 2004, after the dotcom bubble bust, but before its

57 Snap, Amendment No. 3 to Form S-1 Registration Statement, at p. 1. 58 Instagram, “Introducing Instagram Stories,” August 2, 2016, http://blog.instagram.com/post/148348940287/160802- stories; Casey Newton, “Instagram’s New Stories Are a Near-Perfect Copy of Snapchat Stories,” The Verge, August 2, 2016, https://www.theverge.com/2016/8/2/12348354/instagram-stories-announced-snapchat-kevin-systrom-interview. 59 Christopher Heine, “Will Snapchat’s Data Play Help Fend Off Competition From Facebook and Instagram?” Adweek, April 2, 2017, http://www.adweek.com/digital/will-snapchats-data-play-help-fend-off-competition-from-facebook-and- instagram/. 60 S-1, p. 15. 61 comScore, “Surfing Down Memory Lane to January 1996: comScore Media Metrix Revisits First-Ever Web Site Rankings,” February 25, 2004, https://www.comscore.com/Insights/Press-Releases/2004/02/1996-First-Ever-Web-Site- Rankings.

investment in Alibaba.62 The company had first mover advantages, had network effects from various user groups, and access to capital.

Yahoo faced significant competition, however, from new attention platforms. Over the 2000s, as the online advertising industry matured, many firms started introducing services, and content, to attract the attention of users. Despite a head start and a major user base, Yahoo failed to deliver a successful search engine, which became an important service for attracting users.63 In the 2010s, as consumer attention shifted from desktop to mobile and from browsers to apps, Yahoo failed to develop content, and related advertising products, that were a good fit for smartphones.64 Its share of digital advertising revenue plummeted from 12.8 percent in 2010 to 4.2 percent in 2015. Yahoo was sold to Verizon for slightly more than $4 billion in a deal that closed on June 13, 2017.65

The history of messaging apps further illustrates how disruptive innovation, resulting from technological change, can topple leaders. Early messaging programs include CTSS (1961), Talkomatic

(1973), CompuServe’s CB Simulator (1980), and Zephyr Notification Service (1980s).66 But messaging applications really took off in the mid-1990s, with the introduction of ICQ (1996), AOL’s AIM

(1997), Yahoo! Pager (1998), and MSN Messenger (1999).67

62 S&P Capital IQ. 63 Fred Vogelstein, “How Yahoo Blew It,” Wired, February 1, 2007, https://www.wired.com/2007/02/yahoo-3/; Ryan Singel, “Yahoo Gives up, Turns Search Over to Bing,” July 29, 2009, https://www.wired.com/2009/07/yahoo-gives-up/; Cara McGoogan, “Yahoo: Nine Reasons for the Internet Icon’s Decline,” The Telegraph, July 25, 2016, http://www.telegraph.co.uk/technology/2016/07/25/yahoo-9-reasons-for-the-internet-icons-decline/. 64 Vauhini Vara, “Why Yahoo Couldn’t Adapt to the Smartphone Era,” New Yorker, February 9, 2016, http://www.newyorker.com/business/currency/why-yahoo-couldnt-adapt-to-the-iphone-era; Elizabeth Dwoskin, “Yahoo’s Got Millions of Users, But It’s Still in Decline. What Went Wrong?” Washington Post, April 19, 2016, https://www.washingtonpost.com/business/economy/what-went/2016/04/19/854a6194-066f-11e6-a12f- ea5aed7958dc_story.html?utm_term=.c9f10e592209. 65 Ingrid Lunden, “Verizon Closes $4.5B Acquisition of Yahoo, Marissa Meyer Resigns,” TechCrunch, June 13, 2017, https://techcrunch.com/2017/06/13/verizon-closes-4-5b-acquisition-of-yahoo-marissa-mayer-resigns-memo/. 66 Matt Petronzio, “A Brief History of Instant Messaging,” Mashable, October 25, 2012, http://mashable.com/2012/10/25/instant-messaging-history/#CexWnxjF2Pql; Monty Munford, “The Fall… and Rise and Rise of Chat Networks,” Ars Technica, February 13, 2016, https://arstechnica.com/business/2016/02/the-fall-and- rise-and-rise-and-rise-of-chat-networks/. 67 KeriLynn Engel, “The Rise and Fall of Instant Messengers,” WhoIsHostingThis, October 22, 2014, http://www.whoishostingthis.com/blog/2014/10/22/instant-messengers/.

Of these, the most successful was AOL, which sought to preserve its advantage by making it difficult for other messenger programs to interoperate with AIM.68 The FCC was worried it would be able to leverage its monopoly into next generation services, and behavioral restraints intended to prevent this and encourage interoperability as a condition for approving the AOL merger with Time

Warner in 2000.69 Despite the FCC’s intervention, AOL didn’t become fully interoperable with third- party messaging services until 2008, after it had started rapidly losing share to competitors such as

Google Talk, MSN, and Skype.70

By mid-2011, AIM’s worldwide share had fallen to less than one percent.71 Press accounts attribute the decline to mistakes by AOL’s management and innovations by competitors.72 AIM’s user base peaked in 2007, at 63 million, and declined to only 4 million when it shut down in 2012.73 Today, the most successful messenger programs are mobile apps such as Snapchat, WeChat, Line, and

Viber.74

Apple’s music download business shows how disruptive innovation, including business models, can quickly break the dominance of online platforms. Apple introduced its downloadable music service, the iTunes Store, in 2003 after negotiating deals with the major music publishers. The distribution of music quickly shifted from physical stores, and physical media, to online downloads.

By 2010, Apple accounted for 70 percent of U.S. sales of music online and 28 percent of all U.S.

68 David Auerbach, “Chat Wars: Microsoft vs. AOL,” n+1 Magazine, Spring 2014, https://nplusonemag.com/issue- 19/essays/chat-wars/. 69 Gerald R. Faulhaber (2004), “Access and Network Effects in the ‘New Economy’: AOL-Time Warner (2000),” in John E. Kwoka and Lawrence J. White (eds.), The Antitrust Revolution, 4th Edition (Oxford University Press). 70 Christina Warren, “AIM: AOL Instant Messenger Isn’t Dead… Yet,” Mashable, March 14, 2012, http://mashable.com/2012/03/14/aim-not-dead-yet/#y7FENFPxm5qT. 71 Opswat, “Security Industry Market Share Analysis,” June 2011, https://www.opswat.com/sites/default/files/OPSWAT-Market-Share-Report-June-2011.pdf, at p. 7. 72 Jason Abbruzzese, “The Rise and Fall of AIM, the Breakthrough AOL Never Wanted,” Mashable, April 15, 2014, http://mashable.com/2014/04/15/aim-history/#xHCcwCRymPqM. 73 KeriLynn Engel, “The Rise and Fall of Instant Messengers,” WhoIsHostingThis, October 22, 2014, http://www.whoishostingthis.com/blog/2014/10/22/instant-messengers/. 74 Monty Munford, “The Fall… and Rise and Rise of Chat Networks,” Ars Technica, February 13, 2016, https://arstechnica.com/business/2016/02/the-fall-and-rise-and-rise-and-rise-of-chat-networks/.

music sales.75 Commentators suggested that its dominance was so great that entrants could only hope to complement Apple, and not unseat it:

But its dominance is so strong that the only realistic way to compete with Apple is to compete around it with offerings that complement the iTunes experience rather than attempt to unseat it, e.g., Shazam, SoundCloud, Last.FM, MXP4, Pandora, even Spotify. And more often than not, that actually means integrating into the iTunes ecosystem, thus further strengthening Apple’s role.76

In late 2008 Spotify launched its streaming music business in Sweden. Instead of downloading music, which consumers could then listen to whenever they wanted, Spotify made it possible for consumers to stream most music on demand. It negotiated licensing deals with the major music labels that gave it access to most music that people want to listen to. Spotify made money from a free ad- supported basic platform as well as a paid premium service. It expanded from Sweden to other countries including the U.S. in July 2011. Other music streaming companies, including Deezer and

Pandora, also entered using somewhat different business models.

Apple’s download business went into gradual decline 77 In the United States, music downloading by all retailers peaked in 2012 (at $2.8 billion), just after Spotify’s U.S. entry. The small decline in 2013 was followed by a much larger decline in 2014, when U.S. music downloading revenues dropped by 10 percent across all retailers.78 U.S. music downloading revenue has continued

75 NPD Group, “Amazon Ties Walmart as Second-Ranked U.S. Music Retailer, Behind Industry-Leader iTunes,” May 26, 2010, https://www.npd.com/wps/portal/npd/us/news/press-releases/pr_100526/. 76 Mark Mulligan, “Why Apple’s Dominance of the Download Market Is A Big Deal,” Forrester Blogs, August 23, 2010, http://blogs.forrester.com/mark_mulligan/10-08-23- why_apple%E2%80%99s_dominance_download_market_really_big_deal. 77 Maura McGowan, “iTunes Losing Market Share to Streaming Services,” Adweek, April 29, 2013, http://www.adweek.com/digital/itunes-losing-market-share-streaming-services-149017/; Victor Luckerson, “Spotify and YouTube Are Just Killing Digital Music Sales,” Time, January 3, 2014, http://business.time.com/2014/01/03/spotify- and-youtube-are-just-killing-digital-music-sales/; Ed Christman, “Apple Mulls Launching a Spotify Rival, Android App as Downloads Decline,” Billboard, March 21, 2014, http://www.billboard.com/articles/news/5944797/apple-mulls- launching-spotify-rival-android-app-as-downloads-decline-sources; Ed Christman and Alex Pham, “Underwhelming Start to iTunes Radio Lights Fire Under Apple,” Billboard, April 9, 2014, http://www.billboard.com/biz/articles/news/digital- and-mobile/6042224/underwhelming-start-to-itunes-radio-lights-fire-under; Hannah Karp, “Apple iTunes Sees Big Drop in Music Sales,” Wall Street Journal, October 24, 2014, https://www.wsj.com/articles/itunes-music-sales-down-more- than-13-this-year-1414166672; Chris O’Brien, “Apple Confirms Decline in Sales of iTunes Digital Music Downloads,” VentureBeat, October 28, 2014, https://venturebeat.com/2014/10/28/applele-confirms-decline-in-sales-of-itunes-digital- music-downloads/. 78 Recording Industry of America, “U.S. Sales Database,” https://www.riaa.com/u-s-sales-database/.

to decline, and by 2016 was 63 percent of its peak level in 2012 (60 percent of its 2012 peak level in inflation-adjusted terms).79 Apple’s decline was even larger than the industry average—13 to 14 percent worldwide through the first nine months of 2014.80 In subsequent years, Apple’s music download sales declined even further (16 percent in 2015 and 25 to 30 percent in 2016), with rumors in the trade press suggesting that Apple is considering abandoning its download business altogether.81

Apple, with a seemingly secure position in downloadable music, with massive amounts of data on users, including what they downloaded, could not forestall the decline of its music download business. It is now trying to catch up with the disruptors. By March 2017 Spotify had over 100 million active users, including over 50 million paid subscribers.82 In response, Apple launched its own paid streaming service in June 2015, which had 20 million subscribers as of December 2016.83

Airbnb illustrates how difficult it is to predict where competition will come from. In the early

2010s, online travel sites such as Booking.com and Expedia viewed Google as a major competitive threat. People use search engines to examine hotel options and often find their ways to online hotel booking sites in the process. As Google developed its own travel-related option the online travel sites were concerned that it would use its control over search results to tilt things in its favor.

Over the 2010s, however, Airbnb, which started in an early version in 2008, disrupted the lodging industry including the vertical search properties. It developed a marketplace for people to

79 Recording Industry of America, “U.S. Sales Database,” https://www.riaa.com/u-s-sales-database/. 80 Hannah Karp, “Apple iTunes Sees Big Drop in Music Sales,” Wall Street Journal, October 24, 2014, https://www.wsj.com/articles/itunes-music-sales-down-more-than-13-this-year-1414166672. 81 Paul Resnikoff, “Apple Terminating Music Downloads ‘Within Two Years,’” Digital Music News, May 11, 2016, https://www.digitalmusicnews.com/2016/05/11/apple-terminating-music-downloads-two-years/. Apple has denied the rumors that it is contemplating such an exit. Peter Kafka, “Apple Says It Isn’t Going To Stop Selling Music Downloads,” ReCode, May 11, 2016, https://www.recode.net/2016/5/11/11660982/apple-itunes-music-downloads-not-true. 82 Spotify, “About Us,” https://press.spotify.com/us/about/. Note that the estimate of paid users is from March 2017, while the estimate of total active users is from June 2016. 83 Shirley Halperin, “Apple Music hits 20 Million Subscribers; Execs Want ‘More, Faster – We’re Hungry!’” Billboard, December 6, 2016, http://www.billboard.com/articles/business/7604098/apple-music-20-million-subscribers-eddy-cue- zane-lowe-interview.

Figure 1: Total U.S. Music Download Revenue

Source: Recording Industry of America, “U.S. Sales Database,” https://www.riaa.com/u-s-sales-database/.

lease or rent lodging such as spare rooms, vacant apartments, and vacation homes. By 2017 it had 3 million listings in more than 190 countries. About 80 million people rented places to stay using its platform in 2016.84 As Airbnb has given people more choices, the hotel and lodging industry under competitive pressure.

Juniper’s latest research, Sharing Economy: Opportunities, Impacts, and Disruptors 2016- 2020, found that shared space platforms, such as Airbnb are allowing users to rent rooms in residential properties at rates often undercutting traditional hotel rooms, disrupting the leisure and tourism industry significantly.85

One academic study found that Airbnb reduces hotel revenue by 8-10 percent, particularly for hotels serving leisure travelers at times of peak demand.86 Morgan Stanley has estimated that Airbnb will reduce the growth rate of hotel revenue per available room by half a percentage point per year

84 Airbnb, “Airbnb’s 2016 Highlights and 2017 Trends We’re Watching,” January 3, 2017, https://www.airbnbcitizen.com/airbnbs-2016-highlights-and-2017-trends-were-watching/. 85 Hospitality Technology, “Hotel Industry Faces Increased Competition as Airbnb Revenues Hit $6B Globally,” April 27, 2016, http://hospitalitytechnology.edgl.com/news/Hotel-Industry-Faces-Increased-Competition-as-Airbnb-Revenues- Hit-$6B-Globally105231. 86 Georgios Zervas, Davide Proserpio, and John W. Byers (forthcoming), “The Rise of the Sharing Economy: Estimating the Impact of Airbnb on the Hotel Industry,” Journal of Marketing Research.

through 2020.87 It has also placed the travel sites under competitive pressure, as bookings are diverted from its platform, which serves as an intermediary between people and hotels, to Airbnb.

Airbnb's competitive threat to the OTAs [online travel agencies] is larger than that faced by hotels. First, the more hotel demand Airbnb cannibalizes – impacting occupancy and RevPAR [revenue per available room] – the harder it will be on the OTAs. Second, the fact that the OTAs over-index toward leisure (80%+ of Expedia/Priceline bookings from leisure demand) puts them in more direct competition with Airbnb. Airbnb's cannibalization of non-hotel categories – like bed and breakfasts and vacation rentals – further impacts OTA demand…. We see further long-term OTA risk if Airbnb ever decides to allow the traditional hotels to list on its site.88

Expedia, recognizing this threat, purchased an Airbnb competitor, HomeAway in 2015.

While Google was improving its online travel business and online travel platforms were complaining about Google leveraging its position in search to do so, these players ended up facing much more serious competition from a new entrant that started on a shoe string and within a few years developed a global booking business.

Between the mid-2000s and early 2010s several platforms that looked like unstoppable leaders, protected by supposedly insurmountable network effects and first-mover advantages, fell from their perches. In many cases, disruptive technological change laid the groundwork for rivals to enter. Rapid increases in wired broadband speeds made streaming music more reliable, which facilitated the entry of Pandora. The development and deployment of faster and more capacious wireless broadband resulted in the dramatic move of users and developers to smart mobile devices.

That in turn set in motion developments that accelerated the decline of desktop-centric platforms such as AOL and Yahoo and put others, such as Facebook and Google, at significant risk.

The point, as noted above, is not that online platform leaders face inevitable displacement.

Rather, it is that threat of disruptive innovation, in technologies and business models, and the

87 Morgan Stanley, “Global Insight: Who Will Airbnb Hurt More – Hotels or OTAs?” November 15, 2015. 88 Morgan Stanley, “Global Insight: Who Will Airbnb Hurt More – Hotels or OTAs?” November 15, 2015.

knowledge that seemingly secure platform leaders have been toppled, forces online platform leaders to engage in constant innovation to keep attracting users and holding on to their positions.

IV. Barriers to Entry and Expansion and the Role of Data

It is easy to start an online platform in the sense that it often takes little more than renting cheap servers, putting some effort into writing code, and running things from home, or cheap space, at first. Many successful online platforms have started this way. Credit cards are the financial instrument of choice for obtaining liquidity. Entrepreneurs can make considerable progress before they tap angel investors and venture capitalists.

Although entry in this sense is easy, success is anything but. Most online platforms have to figure out how to get a critical mass of users on board before they can create significant value and grow. Few succeed at this.89 YouTube, for example, struggled for a year to figure out how to get enough people to upload videos and enough people to view videos to make their platform useful to either. More than 40 other companies, including Google Video, tried and failed.

Successful platforms are the ones that figured out the right combination of products, prices, design, and ignition strategies.90 They then have some measure of protection from competition simply because any challenger has to solve the same difficult chicken-and-egg problem. History has shown, however, that enough entrepreneurs succeed in figuring out how to grow their platforms to keep competitive pressure on incumbents, and that major online platforms enter each other’s categories to compete as well.

89 David S. Evans and Richard Schmalensee (2016), Matchmakers: The New Economics of Multisided Platforms (Boston: Harvard Business Review Press), Chapter 5. 90 This combination may seem obvious ex-post but seldom is a priori. Facebook, for example, focused on developing a social network in which people had to be real and enforced this early on by limiting membership to people with valid .edu addresses at select schools. That turned out to be important but at the time one could make reasonable arguments for less exclusive approaches.

Recently, commentators have argued that data results in a significant barrier to entry.

According to the Economist,

Internet companies’ control of data gives them enormous power. … With data there are extra network effects. By collecting more data, a firm has more scope to improve its products, which attracts more users, generating even more data, and so on… Vast pools of data can thus act as protective moats.91

At least one book and a number of articles have raised similar concerns.92

One can make various conceptual arguments as to why data should or shouldn’t be important for online competition, and there is a thriving literature that does so.93 A fundamental point, however, is that as a general description of how online platform competition works the data-barrier to entry theory is inconsistent with the facts. The history described in the previous section doesn’t support the view that data acts either as a significant barrier to entry for online platforms or as an asset that protects incumbent platforms from competition.

AOL, Friendster, Myspace, Orkut, Yahoo, and many other attention platforms had data on their users. So did Blackberry and Microsoft in mobile. As did numerous search engines including

AltaVista, Infoseek, and Lycos. Microsoft did in browsers. Yet, in these and other categories data didn’t give the incumbents the power to prevent competition. Nor is there any evidence apparent that data increased the network effects for these firms at least in a way that gave them a substantial lead over challengers.

In fact, firms that, at their inception, had no data whatsoever displaced the leaders. When

Facebook launched its social network in India, in 2006, in competition with Orkut, it had no data on

Indian users since it didn’t have any users. That same year Orkut was the most popular social network

91 The Economist, “Regulating the Internet Giants: The World’s Most Valuable Resource Is No Longer Oil, But Data,” May 6, 2017, http://www.economist.com/news/leaders/21721656-data-economy-demands-new-approach-antitrust- rules-worlds-most-valuable-resource. 92 See, for example, Maurice E. Stucke and Allen P. Grunes (2016), Big Data and Competition Policy, Oxford: Oxford University Press. 93 See, for example, Anja Lambrecht and Catherine E. Tucker (2015), “Can Big Data Protect a Firm from Competition,” Working Paper, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2705530; Daniel L. Rubinfeld and Michal S. Gal (2017), “Access Barriers to Big Data,” Arizona Law Review, 59: 339-381.

in India with millions of users and detailed data on them.94 Four years later Facebook was the leading social network in India.95 A similar story was true for Spotify. When it entered the U.S. 2011, Apple had more than 50 million iTunes users and was selling downloaded music at a rate of one billion songs every four months.96 It had data on those people and what they downloaded. Spotify had no users, and no data, when it started. Yet it has been able to grow to become the leading source of digital music in the world. In all these cases the entrants provided a compelling product, got users, obtained data on those users, and grew.

Although it is possible that data provides some online platforms with important advantages, which could result in barriers to entry, the historical evidence refutes the proposition that data, as a general matter, provides online platforms with permanent advantages or places insurmountable obstacles before new firms.97

V. Online Platforms, Sleepy Monopolies, and Sleepless Nights

A century or ago there were sound concerns, in the United States, that some of the firms that had emerged from the rapid process of industrialization could exert significant market power over buyers and were protected from competition. In some cases, those firms had natural advantages, coming from scale economies and network effects, while in other cases, they obtained their position

94 Alexa, “Top Sites in India,” August 30, 2006, https://web.archive.org/web/20060830074546/http://www.alexa.com:80/site/ds/top_sites?cc=IN&ts_mode=country &lang=none. 95 Caitlin Fitzsimmons, “Facebook Overtakes Orkut in India,” Adweek, August 25, 2010, http://www.adweek.com/digital/facebook-overtakes-orkut/. 96 Leena Rao, “Apple: iTunes Now Has 20M Songs; Over 16B Downloads,” TechCrunch, October 4, 2011, https://techcrunch.com/2011/10/04/apple-itunes-now-has-20-million-songs-over-16-billion-downloads/. 97 The literature on the big-data barrier to entry for online platforms doesn’t provide meaningful empirical evidence that data results in significant barrier to entry or magnifies network effects. The articles typically provide an argument as to why big data might have these effects. Of course, there could be specific situations in which data could provide a barrier to entry and result in significant market power. As Gal and Rubinfeld (2017), point out, it is unlikely that there is a single explanation for how data affects the competitive process. Daniel L. Rubinfeld and Michal S. Gal (2017), “Access Barriers to Big Data,” Arizona Law Review, 59: 339-381.

through merger to monopoly or secured political or regulatory barriers to entry. Some manufacturing industries also evolved into stable oligopolies for which consumers might have benefited from more competition and innovation.

Today’s digital economy looks diametrically opposed. The underlying technology, and business models, facilitate entry and enable firms, with the right formula, to attain global scale quickly, and to challenge incumbent platforms in one or more dimensions. It also results in reversible network effects, unlike physical network industries, that can pose existential threats to the category leaders. History has borne this out over the last several decades as successive waves of online platforms rise and fall. There is no apparent reason, based on changes in economics or technology, why this time is different.

These facts, and history are well known to executives of and investors in online platforms.

They know that some largely new type of business could peel off platform users, and cause accelerating declines. That forces all online platforms, no matter how secure they may seem, to keep innovating and providing value to users. And each needs to worry about other successful platforms in addition to the proverbial inventor in the garage.

None of the firms act as if it has a moat around it, protected by data, network effects, or anything else. There don’t seem to be sleepy monopolies littering the digital economy. Few who participate in this sector can sleep well at night.

37 Role of Market Power in the Digital Economy New York State Bar Association Annual Meeting

Patrick DeGraba U.S. Federal Trade Commission* January 25, 2017

*The views expressed in this presentation are my own and do not necessarily represent those of the Federal Trade Commission or any individual Commissioner

1 Overview

• What is Market Definition? • Why Do Market Definition? • Product Market Definition: – Hypothetical Monopolist Test – Targeted Customers • Critical Loss • Geographic Market Definition • Market Definition in Some Cases

2 What is Market Definition?

• Ultimate goal of merger analysis is competitive effects. Will market power be created and exercised? • Useful to separate products likely to significantly constrain the merging firms’ prices from those that will clearly not significantly constrain price. • Market Definition is an over inclusive first approximation of those products that will likely constrain prices.

3 What is Market Definition?

• Unfortunate Use of the English Language • There is no “actual market.” • How can you define something that does not exist? • Like “defining” Washington metropolitan area • We might be better served with a term such as “substitutes triage” or “likely competing products delineation.” • Back of the envelope competitive effects

4 Why Do Market Definition?

• Need to have a set of competing products to calculate market shares. – Mergers of firms with high market shares in highly concentrated market most likely to lead to higher prices. • Quick way to eliminate obviously non-problematic mergers • High concentration establishes prima facie case • Might help organize an investigation. • Provide a starting point to be used in a more comprehensive competitive effects analysis, though competitive effects analysis often includes firms outside the formal market definition • Section 7 cases seem to require market definition. • “Line of business” equated with market definition

5 Why Market Definition is Controversial

• Economic analysis does not require market definition to obtain competitive effects results. • Merger guidelines- “[competitive effects] analysis need not start with market definition.” – Evanston- Judge found effects results compelling even though he disagreed with FTC market definition. • Market definition can wrongly supplant competitive effects. • The “zero – one” effect – All competitors in a market often treated equally. » Hershey - “…these 19 other hospitals within a 65 minute drive of Harrisburg provide a realistic alternative that patients would utilize.” (DCD p10) – Competitors not in the market are ignored completely • A market defined too broadly can short circuit competitive analysis – Whole Foods • A market defined too narrowly can short circuit competitive analysis – Lundbeck

6 How to do Market Definition

• Intuitively - Find a set of products to which “enough customers that matter” might switch if the prices of the products under investigation increase. – Market definition is driven by what consumers view as close substitutes. ͻ There are two dimensions – Product Market – based entirely on customer willingness to substitute among products with different attributes – Geographic Market – Based either on customer location or supplier location

7 The Hypothetical Monopolist Test: Product Market • The Hypothetical Monopolist Test uses an iterative process to “define” a market – Begin with a product sold by each of the merging firms as a candidate market. – Ask if a hypothetical monopolist that controlled this product could profitably impose a “small but significant and nontransitory” increase in price” (SSNIP) above the benchmark (often prevailing) level. • If yes, stop. This product is a market. • If no, go to the next step. – Expand candidate market by adding the product to which the most sales divert, and return to the SSNIP question.

8 The Hypothetical Monopolist Test: Product Market Illustrated

• Product “A” is a candidate product market. • Raising product A’s price by a SSNIP is unprofitable because of significant substitution to product B and some substitution to C. • Now consider A and B as the candidate market. • Raise A’s and/or B’s prices by a SSNIP. • Substitution to C (and others) does not make the price increase unprofitable. • Then, A & B comprise a relevant product market. C is not in the market.

1 2

C A C A B B

Consumer substitution to B defeats price increase. Insifficient consumer substitution to C to defeat price increase. A and B comprise a market.

9 SSNIP test that raises all prices is over- inclusive • When the price of product A increases the price of substitutes will typically increase as well, but not by as much. – Understates the prices of products not in the candidate market. • The typical increase in price makes the substitutes less of a competitive constraint than it would be if its price remained unchanged.

• The SSNIP test keeps the price of substitutes unchanged. • Thus, the SSNIP overstates the competitive constraint a substitute will impose if its price is allowed to adjust. – Can overstate the prices of products in the candidate market. • If the price of merging products increases 5%, close substitutes would increase < 5% • Having all products in candidate market increase price by 5% overstates customers likely willingness to “leave” the candidate market and substitute to another product compared to those who would leave the candidate market if the merging parties raised their prices and the rest of the market responded.

10 SSNIP test is never under-inclusive

• Choose a candidate market. – If candidate market is too narrow, then the test tells us SSNIP is not profitable – So candidate market expands marginally – Expansion stops when the smallest market is reached • A “significant” competitor may not be included in the market – A competitor may be noticeable but not big enough to make a SSNIP unprofitable and so is correctly excluded from the market • SSNIP won’t find an overly broad candidate market • Choose a candidate market. – If candidate market is too narrow, then the test tells us SSNIP is not profitable – So candidate market expands marginally – Expansion stops when the smallest market is reached • A “significant” competitor may not be included in the market – A competitor may be noticeable but not big enough to make a SSNIP unprofitable and so is correctly excluded from the market • SSNIP won’t find an overly broad candidate market

11 Critical Loss and the SSNIP

• SSNIP test can be done by comparing something called critical loss to something called predicted loss.

• Critical loss is a percentage that we calculate from the size of the SSNIP and the profit margin of the candidate market. – Relatively straightforward – Parties will argue over size of profit margin

• Predicted loss is estimated from data and/or documentary evidence. – Evidence used can be different for each case – More disagreement among economic experts

12 Critical Loss and the SSNIP

• A price increase (including a SSNIP) implies – sales of some units will be lost – remaining units will be sold at a higher price

• Unit sales loss reduces profits • Higher price increases profits

• The number of lost units that causes the profit reduction to just equal the profit increase from higher prices from a SSNIP is called the “Critical Loss.”

13 Critical Loss and the SSNIP

• Critical Loss (CL) is the number of lost units such that the profit reduction from lost sales just equals the profit increase from a higher price on the remaining sales.

%ǻp p-mc %CL = Where M = M + %ǻp p • Derivation • Profit Increase = ȴp(q – ȴq) Profit reduction = (p – mc)ȴq • CL = ȴq such that the Added Profit just equals the Lost Profit • ȴp(q – ȴq) = (p – mc)ȴq

qǻp CL = ǻq* = p + ǻp-mc • • Divide both sides by q and both fraction terms by p to get %CL

14 Critical Loss and the SSNIP

• Examples with a 5% SSNIP and 5%, 20% and 45% margins • Increasing margins imply decreasing critical losses

5 %CL = = 1/2 = 50% 5 + 5

5 %CL = =1/5 =20% 20 + 5

5 %CL = =1/10 =10% 45 + 5

15 Critical Loss and the SSNIP

• The “Predicted Loss” (also called the “Actual Loss”) is the number of lost unit sales that the hypothetical monopolist is predicted to lose due to the price increase. • The predicted loss is estimated from data and documentary evidence (See later slides).

• If predicted loss is greater than the critical loss, then the SSNIP is not profitable and the candidate products are not a relevant market. • If predicted loss is less than the critical loss, then the SSNIP is profitable and the candidate products are a relevant market.

• A key issue, then, is how to reliably estimate the Predicted Loss.

16 Critical Loss and Market Definition

• Predicted Loss estimates are based on data that allow us to estimate diversion among products or price effects. – Nestle/Dreyer’s (2003) – super premium ice cream • Looked at retail data and observe entry by Dreyer’s resulted in lower prices of other super premium ice cream. • Could calculate diversion between super-premium and premium ice cream base on retail data. – Sysco (2015) – broadline foodservice distribution • looked at bidding data and inferred the next best alternative to one of the merging parties would be if that party raised it prices by a SSNIP (FTC v Sysco; M&O at 36 https://www.ftc.gov/enforcement/cases-proceedings/ftc-v-sysco-usf-holding-corp-us-foods-inc )

17 How Should We Interpret High Margins? • High margins mean the critical loss will be small:

%ǻp • Remember: %CL = M + %ǻp

• But high margins also suggest that price elasticity is low, which means predicted loss will be small.

• M = 1/Eown where Eown is elasticity of demand • While not dispositive this creates an inherent tension from a claim of low critical loss but high predicted loss. • O’Brien and Wickelgren (2003) and/or Katz and Shapiro (2002)

18 Critical Loss and Market Definition

• SSNIP/CL analysis by itself is rarely dispositive. Courts will also check that such results are consistent with other evidence of market definition. • Data may not always be available to perform a SSNIP test. 1. Interviews with market participants 1. Buyers 2. Sellers 2. Depositions (investigational hearings) 3. Documents – Business and Marketing Plans – Sales reports – Internal and third-party industry studies – E-mail discussions of pricing and competition

19 Targeted Customers Product Market Definition

• Products sold to an identifiable group of customers can be a separate market if: – Seller can charge that group of customers a different price than other customers – No arbitrage between this group and other groups • Targeted customers is one of the most important ideas in the merger guidelines given court’s reliance on market definition – Targeted customers allows analysis closer to competitive effects analysis • Can rule out tangential competitors that are not significant • Can focus on specific aspects of competition for specific customer groups • Recent examples: – Sysco – Broadline foodservice distribution to national customers – Whole Foods – Core customers

20 Geographic Market

• An area outside of which products in the product market do not compete significantly enough with products inside the area. • Grocery stores in Rockville, Md don’t compete with those in DC • Typically important if customers purchase in a limited area. – Pinnacle/Ameristar • Local markets St Louis and Lake Charles • Exclude Las Vegas. – Evanston Northwestern Health • Determine which hospitals near merging hospitals are good substitutes • SSNIP test can be applied to geography just like products. • Start with merging location and raise the price. • If SSNIP not profitable add a competitor and try again. • SSNIP is replacing Elzinga-Hogarty in hospital mergers

21 Geographic Market

• Targeted customers. – If national firms can discriminate by location, each location could potentially by a separate market. – Things to consider • If price is national but promotions are local, each locality could be a market • If non-price attributes like service quality are chosen on a local basis • If two national competitors have different local competition in different cities – FCC uses local geographic markets for mobile service (AT&T-T-Mobile staff report para 32-34 https://transition.fcc.gov/transaction/att-tmobile.html ) • Local casinos might offer promotions by zip code. • In many cases geographic component adds nothing – Arch Coal – ~10 mines and all of the electricity producer that used that coal were identified. What help is geography?

22 Geographic Market

• Elzinga-Hogarty test is not valid for hospitals. – E-H test developed in the 1970s to analyze commodity flows https://www.justice.gov/atr/chapter-4-competition-law-hospitals#2a1 • Predates SSNIP test development – Intuitively - Draw a circle on a map • If most people living inside the circle buy inside the circle… • …and very few people form outside the circle buy inside the circle, • then the circle is a geographic market – THIS TEST CAN PRODUCE MARKETS THAT ARE MUCH TOO BIG • Used for years to defeat challenges to hospital mergers. • Elzinga testified in In re Evanston Nw. Healthcare Corp., 2007, that his own test gave overbroad markets in hospital settings. – Elzinga and Swisher 2011 Inter J of Econ of Bus. – Capps, Dranove, Greenstein, and Satterthwaite, NBER April 2001 • One major problem is that it was designed for homogeneous goods but hospitals tend to be heterogeneous • Not designed for customer flows

23 Zillow Trulia Merger (2015)

• Two largest consumer-facing web portals for home buying that sell advertising to agents • FTC allowed the merger and issued public statement regarding analysis • https://www.ftc.gov/public-statements/2015/02/statement-commissioner-ohlhausen-commissioner-wright-commissioner- mcsweeny • Two markets considered – Advertising market for real estate agents – Listing information for home buyers

24 Zillow Trulia Merger (2015)

• Advertising for real estate agents – Considered “high performing” agents and “all” agents • No ability to price discriminate against high performing agents – Portals were a small portion of agents’ advertising spend – No evidence portals had higher ROI than other advertising methods – High volume agents leave these portals as adverting venues regularly – Not a significant relationship between Zillow’s advertising prices and Trulia’s presence in local geographic markets. • Competition for home buyers – Ad revenue is an incentive to innovate and grow consumer traffic – Competition from other consumer-facing portals such as Realtor.com and brokers like Redfin

25 FTC v Staples/Office Depot (2016)

• Merger between largest sellers of consumable office supplies • Defense: Amazon Business will restore lost competition • Decision: Amazon Business faced challenges District Court Decision (DCD) pg 64 • Lack of RFP experience • Can’t control third party price and delivery • No customer specific pricing • No dedicated agents for B to B space • No desktop delivery • No utilization and invoice reports • Lacked product variety and depth – No evidence that online ordering would replace RFP’s for large buyers • “If Amazon Business was more developed…the outcome of this case very well may have been different.” DCD pg 70

26 FTC v Sysco (2015)

• Merger between two largest and only national broadline foodservice distributors in the U.S. • FTC asserted a market for targeted broadline foodservice distribution to national customers. (FTC v Sysco M&O p 18) – Customers with large geographic presence would not switch their purchases to a patchwork of smaller service in response to a SSNIP • Economic expert performed a SSNIP test – Based on margin calculations determined critical loss to be 50% (M&O 35) – Calculated predicted loss to be much lower than 50% (M&O 35-37) • Looked at bidding and RFP data to determine how customers would divert to another broadliner if the price of one broadliner were to increase by 5%. • Looked at Linc database maintained by U.S. Foods services and found much less than 50% of customers would switch to non-broadline services in response to a SSNIP. (M&O 37)

27 FTC v Sysco (2015)

• Two conditions for targeted customers – Ability of sellers to price discriminate – No arbitrage • Judge found FTC expert analysis “persuasive” – Had reservations because data was not from ordinary course documents but rather assembled for the proceeding and were not necessarily complete – Found it important that results of the data analysis were consistent with ordinary course documents and business practices • “the determination of the relevant market in the end is 'a matter of business reality-[ ] of how the market is perceived by those who strive for profit in it.” (District Court Decision pg 37)

28 FTC v Whole Foods (2007)

• Merger between Whole Foods and Wild Oats • Two largest organic grocery chains • Market Definition – FTC - Premium Natural Organic Stores (PNOS) – Defendants Conventional Supermarkets, were also in the market. (Circuit Court Decision (CCD) 14) • District Judge - Because there are marginal customers that would switch to other stores there was not PNOS market. The larger market unlikely to have antitrust harm (CCD 2, 19) • Circuit Court – Viewed this case as a targeted customer case. (CCD 20) Merger between Whole Foods and Wild Oats • Two largest organic grocery chains • Market Definition – FTC - Premium Natural Organic Stores (PNOS) – Defendants Conventional Supermarkets, were also in the market. (Circuit Court Decision (CCD) 14) • District Judge - Because there are marginal customers that would switch to other stores there was not PNOS market. The larger market unlikely to have antitrust harm (CCD 2, 19) • Circuit Court – Viewed this case as a targeted customer case. (CCD 20)

29 FTC v Whole Foods (2007)

• Circuit Court found – competition between PNOS stores and other supermarkets for dry goods – no such competition for perishable goods between PNOS and conventional supermarkets (ACD 17-18) – Competition between Whole Foods and Wild Oats (and other PNOS) did affect prices of perishable goods – “Core” PNOS customers bought primarily perishable goods – Thus they could be price discriminated against – Thus the cluster of goods they purchased constituted a submarket.

30 FTC v Whole Foods (2007)

• “Diversions Ratios”: Based on analysis of Wild Oats Revenues • FTC presentation A challenged merger of retailers: FTC v Whole Foods\ Sophia Bulgaria Sept 2009

Othe r PNOS

PNOS

Whole Foods

PNOS Trader Joe's Vitamin Cottage Conventional Sup. Gourmet Sup. Mass Merchants

31 FTC v Whole Foods (2007)

• Wild Oat’s Response to Entry

Conventional Mass Whole Foods Other PNOS Trader Joe's Gourmet Sup. Sup. Merchants

Margins Implied Prices

32 Evanston Northwestern Healthcare (2005)

– FTC brought suit against a consummated merger of hospitals in Evanston Illinois. – FTC alleged a geographic market consisting of 3 hospitals two of which had merged. Defendants claimed a larger market including more hospitals – FTC presented direct evidence of price increases as a result of the merger. – 3 Interesting findings by Judge McGuire • Elzinga-Hogarty not appropriate for hospital geographic market because it returned overly large geographic markets. • The geographic market contained 7 hospitals, even though the plaintiffs proposed a three hospital market. • The direct evidence of anticompetitive effects were enough to overcome the larger market.

33 Penn State Hershey (2016)

– Hershey, the leading teaching hospital in the Harrisburg area bought nearby Pinnacle Health System

– FTC (and PA) sued, and alleged a four county market in the area (Circuit Court decision (CCD) 15) – District court found • Market is bigger - 43% of Hershey patients are outside the 4 counties (District Court Decision (DCD) p 9); up to 65 minutes away (DCD p 10) • Hershey signed long term deals with two largest insurers, locking in prices – Circuit Court found • Reliance on the 43% figure followed E-H and ignored SSNIP (CCD 16) • District Judge ignored 91% of Hershey customers were local (CCD 20) – Inflows don’t predict outflows, ignores product differentiation

• Contracts are not used to determine market definition (CCD 16) – Also consider that competition occurs on non-price dimensions

34 FTC v Lundbeck (2009)

• “…perplexing…” market definition precluded any competitive

effects analysis https://www.ftc.gov/sites/default/files/documents/cases/2010/08/110919lundbeckfindings.pdf – Lundbeck purchased Indocin IV with retail price of $78. Treats PDA in preterm babies. Purchased the only other drug to treat PDA, NeoProfen, about to receive FDA approval. Two days later raised the price of Indocin IV to ~$1500. – After FDA approval, Lundbeck priced NeoProfen similarly. – Eight doctors testified, would not switch for small price difference. – Defendant’s expert said a hospital that would prefer lower priced drug could not “drive” doctors towards cheaper drug. (CCD 6)

35 FTC v Lundbeck (2009)

• District Court Judge concluded – Doctors would not switch, implies low cross price elasticity – Low cross price elasticity implies not in same market – Not the same market implies no antitrust harm • Low cross elasticity by itself DOES NOT imply not in same market. Must be compared to own price elasticity. – Competitive Effects (symmetric linear case)

MD p - mc εji %ǻP = Where M = and D = ε 2(1-D) p ii • Casual analysis of a 5% price change may not help analyze a 1300% price change.

36 Takeaways

• Market definition is best viewed as an over inclusive first approximation of the important competing products. • Based on customer substitution to other goods. • SSNIP test results should be consistent with other evidence. • Can be useful – When structural analysis is appropriate. – As a way to organize an investigation. • Pitfalls of use – Not a prerequisite for doing competitive effects and can divert attention from the competitive analysis. – Potential for all sellers in the market to be treated as equals when they are not. • “Targeted customers” useful in focusing on competitive effects.

2018 NYSBA ANTITRUST SECTION ANNUAL MEETING : Unilateral Conduct Committee: Role of Market Power in the Digital Economy JANUARY 25, 2018

Speaker Written Materials for Nick Gaglio1

This panel explores the overarching question of whether the existing analytical framework for market power is well-suited to assessing conduct in the digital economy. These materials provide background information intended to illustrate evidence relied on by courts and antitrust agencies in evaluating tech-sector conduct and transactions. The materials begin with background on market power assessment under both Section 2 of the Sherman Act and Section 7 of the Clayton Act. They end with in-depth discussions of two topics that have received considerable attention with respect to their roles in the digital economy, and the way they interact with traditional market power analyses, specifically, free services in multi-sided markets, and the competitive constraints generated by the threat of entry, with a particular focus on network effects as a barrier to entry.2

I. WHERE WE’VE BEEN: EVIDENCE OF MARKET OR MONOPOLY POWER

Section 7

In evaluating the likelihood of a merger to create an entity with exercisable market power, courts have looked to indirect evidence, with a focus on measures of concentration. Historically, courts have commenced the analysis of a horizontal merger’s competitive effect by evaluating indirect evidence of market power, by defining an antitrust relevant market “composed of products that have reasonable interchangeability for the purposes for which they are produced—price, use and qualities considered.” United States v. E.I. DuPont de Nemours & Co., 351 U.S. 377, 404 (1956). In defining relevant markets, courts assess both product and geographic contours. Brown Shoe Co. v. United States, 370 U.S. 294, 324-25 (1962) (“The ‘area of effective competition’ must be determined by reference to a product market (the ‘line of commerce’) and a geographic market (the ‘section of the country’).”) Courts then look at the merger’s impact on concentration in those defined relevant markets. See, e.g., id. at 322 n.38. For decades, concentration at specified levels has supported a presumption of enhanced market power, e.g., from United States v. Philadelphia Nat. Bank, 374 U.S. 321, 364 (1963) through FTC v. Sysco Corp., 113 F. Supp. 3d 1, 52-53 (D.D.C. 2015).

For a similarly long time, courts have assessed whether entry was likely to mitigate that presumed enhanced market power. See, e.g., Brown Shoe, 370 U.S. at 321-22; United States v. Baker Hughes Inc., 908 F.2d 981, 987 (D.C. Cir. 1990).

1 Nick Gaglio is a partner in the New York office of Axinn. These materials could not have been prepared without the valuable help of Tal Elmatad and Maxime Fischer-Zernin. 2 The materials presented are intended to provide attendees with illustrative background material for the discussion. As such, they do not reflect any specific positions of the author, the Axinn firm or any Axinn client.

Page 2

Courts’ adoption of the framework emphasizing indirect, concentration-oriented measures reflects the enduring influence of the antitrust agencies’ Horizontal Merger Guidelines, particularly through the 1997 revisions to the 1992 Guidelines. See, e.g., FTC v. PPG Indus. Inc., 798 F.2d 1500, 1503 (D.C. Cir. 1986) (adopting HHI per 1982 Merger Guidelines); Baker Hughes, 908 F.2d at 985-86 (criticizing United States’ failure to faithfully apply the 1982 Guidelines’ framework); FTC v. Staples, Inc., 970 F. Supp. 1066 (D.D.C. 1997) (adopting efficiencies framework from 1997 version of the 1992 Guidelines). Prior to the 2010 revision, the Guidelines analysis began with market definition, the “hypothetical monopolist test” and the presumptions created by concentration levels in a properly defined relevant market. 1992 Guidelines (1997 rev.) at § 1.

In merger cases, including those involving technology markets, courts will consider more direct evidence where available. For example, in United States v. Oracle, 331 F. Supp. 2d 1098, 1117-18 (N.D. Cal. 2004), the Court rejected the government’s structural evidence as ill-suited to a unilateral effects theory in a merger between differentiated product manufacturers. Id. at 1123- 65. As a result, it required other evidence of exercisable post-merger market power. The Court considered the parties’ documents and data as well as testimony from competitors, customers, consultants and economic experts. Specifically, the Court examined win-loss and bidding data, testimony from SAP (a competitor to the parties), testimony from customers who rejected SAP in various bids and testimony from consultants that helped companies select enterprise software. Id. at 1166-68. The Court also considered the government’s expert analyses of (1) salesforce discount requests, (2) a regression analysis of reasons for discounting, and (3) a merger simulation based on an auction market model. Id. at 1168-70. The Court noted the absence of evidence that would have been relevant to the post-merger market power assessment, specifically econometric evidence of diversion. Id. at 1172-73.

The Agencies have routinely considered other evidence of market power as well. As the Court pointed out in Baker Hughes, the Guidelines also made clear that evidence of entry and other market characteristics, besides concentration, and merger-specific efficiencies, were relevant to the assessment of whether a merger would create or enhance exercisable market power. 1992 Guidelines (1997 rev.) §1.51.

Sherman Act Section 2

The Section 2 requirement of monopoly power is “the power to control prices or exclude competition.” E.I. DuPont 351 U.S. at 391. Control over prices and exclusion of competition are two sides of the same monopoly power. See E.I. DuPont 351 U.S. at 392 (“It is inconceivable that price could be controlled without power over competition or vice versa.”).

As discussed above, much like in Section 7 market power assessments, courts have most often established Section 2 monopoly power through the proxy of market share. Monopoly power is generally established indirectly, using “the defendant's relevant market share in light of other market characteristics, including barriers to entry.” Tops Markets, Inc. v. Quality Markets, Inc., 142 F.3d 90, 100 (2d Cir. 1998). Market share is often used as a proxy for monopoly power “[b]ecause a true determination of whether a firm possesses monopoly power hinges upon an

Page 3

analysis of complicated economic factors.” Am. Council of Certified Podiatric Physicians & Surgeons v. Am. Bd. of Podiatric Surgery, Inc., 185 F.3d 606, 622 (6th Cir. 1999). There is no precise level of market share at which monopoly power is inferred. United States v. Aluminum Co. of Am., 148 F.2d 416, 424 (2d Cir. 1945) (holding that a 90 percent market share was sufficient, but “it is doubtful whether sixty or sixty-four percent would be enough; and certainly thirty-three percent is not.”). Later commentary suggests 50 percent is considered a floor for finding monopoly power. 2B Phillip E. Areeda et al., ANTITRUST LAW ¶ 532c, 250 (3d ed. 2007).

Courts have long required evidence of both a dominant market share and barriers to entry to find monopoly power. See, e.g., Matsushita Elec. Indus. Co. v. Zenith Radio Corp., 475 U.S. 574, 591 n.15 (1986) (“Respondents offer no reason to suppose that entry into the relevant market is especially difficult, yet without barriers to entry it would presumably be impossible to maintain supracompetitive prices for an extended time.”); Image Tech. Servs. v. Eastman Kodak Co., 125 F.3d 1195 at 1208 (9th Cir. 1997) (“Even a 100% monopolist may not exploit its monopoly power in a market without entry barriers.”). There are a variety of definitions of barriers to entry, including “factors (such as certain regulatory requirements) that prevent new rivals from timely responding to an increase in price above the competitive level,” United States v. Microsoft Corp., 253 F.3d 34, 51 (D.C. Cir. 2001), and “additional long-run costs that were not incurred by incumbent firms, but must be incurred by new entrants’ or ‘factors in the market that deter entry while permitting incumbent firms to earn monopoly returns,” W. Parcel Express v. UPS, 190 F.3d 974, 975 (9th Cir. 1999) (quoting Los Angeles Land Co. v. Brunswick Co., 6 F.3d 1422, 1427-28 (9th Cir. 1993); Stearns Airport Equip. Co. v. FMC Corp., 170 F.3d 518, 530 (5th Cir. 1999). Examples of barriers to entry include “legal requirements, control of natural advantages or supplies, markets too small for more firms, intellectual property rights, exclusivity arrangements, large capital outlays, economies of scale, and brand name or reputation.” ABA SEC. ANTITRUST L., ANTITRUST LAW DEVELOPMENTS 234-35 (8th ed. 2017) (collecting citations to cases illustrating various barriers to entry).

While courts frequently rely on the proxies discussed above, courts also assess monopoly power “by direct evidence of a defendant's price control or exclusion of competitors from a particular market in a manner indicative of its possession of monopoly power.” Shak v. JPMorgan Chase & Co., 156 F. Supp. 3d 462, 482 (S.D.N.Y. 2016). For example, in Microsoft, the Court considered both indirect and direct forms of evidence. 253 F.3d 34, 51-58 (D.C. Cir. 2001). While the Court relied on indirect evidence in finding that Microsoft had monopoly power, it also examined and rejected Microsoft’s proffer of direct evidence that it lacked monopoly power. Id at 56-57. First, the Court considered Microsoft’s evidence of research and development spending, which Microsoft asserted a monopolist had no incentive to spend. In addition to finding Microsoft’s evidence insufficiently specific to the relevant product, operating systems, the Court also credited the lower court’s finding that exclusionary motives could also incent R&D spending. Id. at 57. Second, the Court rejected Microsoft’s evidence that its pricing of the Windows operating system was below the short-term profit-maximizing level, in other words, that it lacked “the power to control prices” as required by the Supreme Court in du Pont. Microsoft, 253 F.3d at 57. The Court placed great weight on Microsoft’s setting of its prices without regard to rivals’ prices. Id. at 57-58 (citing the lower court’s finding that “[o]ne would expect a firm in a competitive market to pay much closer attention to the prices charged by other firms in the market.”). Id.

Page 4

II. WHERE WE ARE: EVIDENCE OF MARKET OR MONOPOLY POWER IN DIGITAL MARKETS

The traditional analyses of market power described above continue to play a role when applied to the digital economy; however, the digital economy presents unique challenges for these analytical tools. For instance, the consumer welfare standard, a fundamental underpinning of American antitrust theory, often characterizes increased benefits to consumers – frequently in the form of lower prices – as a sign of effective competition. But how do we evaluate what impact a merger or conduct would have when the products are free? Moreover, many digital economy products provide services to multiple sets of customers simultaneously, which introduces additional complexity when attempting to evaluate the competitive impact of a merger or certain conduct. Finally, many of these multi-sided products are characterized by network effects where the product becomes more valuable to all sides as adoption grows (i.e., market share increases); this can drive incredible value to consumers but some commentators believe it also raises questions as to whether and how these network effects present a barrier to entry of antitrust concern. Commentary abounds suggesting that these aspects of the digital economy call for a departure from the established antitrust analytical framework. For example, compare Ronald A. Cass, Antitrust for High-Tech and Low: Regulation, Innovation, and Risk, 9 J.L. ECON. & POL'Y 169, 197-99 (2013) (calling for more cautious enforcement) with The World’s Most Valuable Resource is No Longer Oil, but Data, ECONOMIST, May 6, 2017 (calling for more aggressive enforcement).

Enforcers in the United States from both sides of the aisle, meanwhile, have expressed belief that the current framework is up to the task. See, e.g., “‘Big Data’ and Competition for the Market,” Remarks of Bernard A. Nigro, Jr. Deputy Assistant Attorney General, U.S. Dept. of Justice Antitrust Div. at the Fourth Annual Tech, Media & Telecom Competition Conference (Dec. 13, 2017) (hereinafter, “Nigro M&TCC speech”) (“Existing antitrust tools have been adequate to address these issues in the past, and they are adequate now too.”); “Antitrust Enforcement in the Digital Age,” Remarks by Maureen K. Ohlhausen, Acting Chairman, U.S. Federal Trade Comm’n, Before the Global Antitrust Enforcement Symposium (Sept. 12, 2017) (hereinafter, “Ohlhausen GAES speech”) (stating “the current framework is sufficiently flexible to address these important issues”); “A U.S. Enforcer’s Perspective: Protecting Competition and Promoting Innovation,” Remarks of Federal Trade Commissioner Terrell McSweeny at the 2016 Taiwan International Conference on Competition Policy (June 29, 2016) (emphasizing that “our antitrust tools are flexible . . . we do not need a different set of rules”).

The following discussion provides selected illustrations from the digital economy of evidence relevant to market power in more recent investigations and transactions. Specifically, the illustrations relate to (a) multi-sided markets and free goods, and (b) entry, with particular focus on network effects.

a. Multi-sided markets and free goods

A multi-sided market exists when an entity serves multiple groups of customers whose demands are interdependent. Two-sided markets exist where a platform “can affect the volume of transactions by charging more to one side of the market and reducing the price paid by the

Page 5 other side by an equal amount; in other words, the price structure matters, and platforms must design it so as to bring both sides on board.” United States v. Am. Express Co., 838 F.3d 179, 184 n.3 (2d Cir. 2016), cert. granted sub nom. Ohio v. Am. Exp. Co., No. 16-1454, 2017 WL 2444673 (U.S. Oct. 16, 2017) (quoting Jean–Charles Rochet & Jean Tirole, Two–Sided Markets: A Progress Report, 37 RAND J. Econ. 645, 664–65 (2006)). Under these conditions, the interdependence of the two customer groups requires a pricing strategy conducted with consideration of “two demand curves, each of which depends on the quality-adjusted quantity purchased on the other side. The platform incurs a fixed cost for operating the platform and variable costs for servicing each side.” David S. Evans, The Antitrust Economics of Multi-Sided Platform Markets, 20 YALE J. ON REG. 325, 339–40 (2003). Specifically:

“The optimal price for side A depends on the responsiveness of demand to changes in price on side A, the responsiveness of demand on side B to changes in quality- adjusted sales on side A, and changes in variable costs on both sides. To see this, suppose we have found the optimal prices for sides A and B. An increase from the optimal price on side A, holding the optimal price on side B constant, will have the following effects: Demand on side A will fall, demand on side B will fall since side B's product is less valuable, variable costs will fall on side A and variable costs will fall on side B.” Id.

Unlike a single-sided business, which profit-maximizes by producing until marginal revenue equals marginal cost, in two-sided businesses the optimal price will depend on the price sensitivities of demand for both sides and the degree of their interdependence. Id.

In American Express, the Second Circuit recently addressed appropriate market definition when faced with two-sided platforms. 838 F.3d at 184 n.3. The Second Circuit found the District Court made a “fatal” mistake of defining a “network services market,” which failed to include the market for cardholders. Id. at 197. The District Court defined the “network services market” as the market for “core enabling functions provided by networks, which allow merchants to capture, authorize, and settle transactions for customers who elect to pay with their credit or charge card.” United States v. Am. Exp. Co., 88 F. Supp. 3d 143, 171 (E.D.N.Y. 2015), rev'd and remanded sub nom. United States v. Am. Express Co., 838 F.3d 179 (2d Cir. 2016), cert. granted sub nom. Ohio v. Am. Exp. Co., No. 16-1454, 2017 WL 2444673 (U.S. Oct. 16, 2017). The Circuit Court explained that a proper application of the hypothetical monopolist test (HMT) would have led the District Court to include both sides of the market:

“A proper application of the HMT in this case would not have merely assumed that a decrease in quantity of network services demanded by merchants facing a SSNIP would be too small to render the accompanying price increase unprofitable. The District Court instead should have considered the extent to which even a low level of merchant attrition might cause some cardholders to switch to alternative forms of payment. Application of the HMT to a two-sided market must consider the feedback effects inherent on the platform by accounting for the reduction in cardholders' demand for cards (or card transactions) that would accompany any degree of merchant attrition. Although the District Court claimed that it accounted for the two-sided features of the credit-card industry in its market definition inquiry,

Page 6

it expressly declined to define the relevant product market to encompass the entire multi-sided platform. This was error because the price charged to merchants necessarily affects cardholder demand, which in turn has a feedback effect on merchant demand (and thus influences the price charged to merchants).” 838 F.3d at 199–200 (2d Cir. 2016) (citations omitted).

The Supreme Court will review the case, leaving the antitrust bar and business community to wait to see whether courts must consider evidence from more than one side of a multi-sided market in evaluating the existence or exercise of market power.

The importance of two-sided markets will only grow as businesses continue to develop platform based solutions enabled by the Internet. For instance, social media platforms are a prime example of a two-sided platforms, and frequently present the additional feature that only one side of the market pays a cost. Unlike American Express, which charges both cardholders and merchants a fee, a social media platform is a multi-sided platform that charges only one side, the advertisers. To the extent that market power analyses depend on price theory and therefore require positive prices to define antitrust markets under the hypothetical monopolist test, the inability to identify a price charged to users presents challenges: for example, does a court or enforcer assign some measure of information or attention cost, or ignore the free side of a platform? At least one Federal Trade Commission (FTC) Commissioner has suggested an answer to this question. See “Competition Law: Keeping Pace in a Digital Age,” Remarks of Commissioner Terrell McSweeny at the 16th Annual Loyola Antitrust Colloquium (Apr. 15, 2016) (“The issue is whether to look just at price effects on the paying side of these platforms, or whether to consider harms – such as to quality and innovation – on the free side. The Guidelines’ section on innovation makes clear that we look at both sides in the merger enforcement context.”) Decisions confirm that antitrust agencies consider such evidence.

Facebook’s 2012 acquisition of Instagram involved consideration of both sides of an online, two-sided market with a free service. Consider the 2012 Facebook/Instagram acquisition. In August 2012 the FTC concluded a non-public investigation into the proposed acquisition and took no action. FED. TRADE COMM’N, FTC Closes Its Investigation Into Facebook’s Proposed Acquisition of Instagram Photo Sharing Program (Aug. 22, 2012), available at https://www.ftc.gov/news-events/press-releases/2012/08/ftc-closes-its-investigation- facebooks-proposed-acquisition. Although the FTC did not publish the reasoning behind its decision to let the acquisition proceed, the UK Office of Fair Trading’s (OFT) decision on the same deal is informative. In permitting the merger to proceed, OFT stated that social networks are two-sided markets, while defining three narrower product markets: social networking for users, a camera application for users, and an advertising space for advertisers. UK OFF. FAIR TRADING, Anticipated acquisition by Facebook Inc of Instagram Inc ¶7, 8 (Aug. 22, 2012), available at http://webarchive.nationalarchives.gov.uk/20140402232639/http://www.oft.gov.uk/shared_oft/m ergers_ea02/2012/facebook.pdf. In evaluating whether the acquisition created a risk of vertical foreclosure by Facebook, the OFT considered evidence of consumer preferences with respect to functionality and concluded that Facebook was unlikely to limit the posting of Instagram photos to other social networks because it would likely diminish the value of the app, which could cause users to switch away from Instagram. Id. at ¶ 37. This reflects an awareness of the two-

Page 7

sidedness of the market and a consideration of evidence of the non-price effects of the acquisition, i.e. the effect on quality of user experience.

The FTC considered evidence regarding both sides of a two-sided market with a free service in its review and approval of Zillow’s merger with Trulia. See FEDERAL TRADE COMMISSION, Statement of Commissioners Ohlhausen, Wright, and McSweeny Concerning Zillow, In. / Trulia, Inc. (Feb. 19, 2015), https://www.ftc.gov/system/files/documents/public_statements/625671/150219zillowmko-jdw- tmstmt.pdf. Staff assessed evidence of competition between real estate portals and other advertising outlets, the potential for post-merger diversion, and direct price competition between the merging parties on the advertising side, while also looking at the parties’ incentives to develop new features on the customer-facing side “in order to grow its consumer audience and thereby increase its advertising revenue.” Id. Finding it unlikely that the merger would enhance a combined Zillow/Trulia’s ability to exercise market power against either advertiser real estate brokers or consumer house shoppers, the FTC closed its investigation without enforcement. Id.

Various commentators have suggested alterations to the traditional assessment of market power with respect to products offered for free. For example, Michal Gal and Daniel Rubinfeld make two proposals for the analysis of market power in two-sided markets with a free service: the Agencies should consider the impact of free goods on paid goods despite the fact that traditional SSNIP tests would ignore their impact, and market power analyses of markets for free goods should look beyond price to effect on quality, consumer choice and information costs. Michal S. Gal & Daniel L Rubinfeld, The Hidden Costs of Free Goods: Implications for Antitrust Enforcement, 80 ANTITRUST L.J. 522, 553 (2016). They suggest that “in markets in which all goods are provided for free,” the hypothetical monopolist test should set “market boundaries by measuring the effects of small but significant and non-transitory changes in quality (SSNIQ), in line with the Microsoft/Skype analysis. The SSNIQ test examines switching once quality is reduced (rather than when price is increased).” Id. at 551. Similarly, David S. Evans notes that “[a] price of zero provides a red flag that the textbook models of competition and standard antitrust analysis do not apply to the product in question.” David S. Evans, The Antitrust Economics of Free, COMPETITION POL’Y INT’L, Spring 2011, at 71, 81. Evans states that courts have no choice but to consider the interdependence of each side of a two-sided market: “[w]hen an antitrust or merger analysis involves a product that is made available for free—or where the paid product in question has a twin product whose price is zero—there is no substitute for carefully considering the economic interrelationships between these products and the overall competition between providers of the paired products or one or the other product.” Id. at 84-85. By contrast, John M. Newman has suggested that the free goods can be analyzed if one acknowledged that “consumers generally pay with their attention, information, or both,” and “[i]n tying arrangements, consumers also pay with money.” John M. Newman, Antitrust in Zero- Price Markets: Foundations, 164 U. PA. L. REV. 149, 202 (2015). These proposals suggest a variety of evidence that advocates can productively bring to the agencies’ attention during an investigation involving multi-sided markets.

b. Entry barriers, including network effects

As Commissioner Ohlhausen recently noted, “during any antitrust investigation, we routinely look at entry barriers including such network effects.” Ohlhausen GAES speech at 9.

Page 8

Agency actions in digital and technology markets, as well as court decisions, confirm that evidence illuminating the same question courts and the agencies have been asking for decades – what is the likelihood and nature of entry – remains central to the market power analysis. For example, an agency may focus on evidence of future entry, as in the Department of Justice’s approval of the Sirius/XM Radio merger. See DEPARTMENT OF JUSTICE, Statement of the Department of Justice Antitrust Division on its Decision to Close its Investigation of XM Satellite Radio Holdings Inc.’s Merger with Sirius Satellite Radio Inc., (Mar. 24, 2008), https://www.justice.gov/archive/opa/pr/2008/March/08_at_226.html. There, the Antitrust Division closely examined evidence regarding new technologies under development that the Division determined would likely offer new or improved alternatives to satellite radio, including streaming internet radio to mobile devices. Id. The parties provided additional evidence that competition between Sirius and XM was limited, in part, by long-term contracts with automobile manufacturers. Id. That provided important context with which the Division could judge the sufficiency and timeliness of the anticipated entry of streaming media. Id.

The agency may also focus on evidence of recent entry, such as in the Antitrust Division’s review of Expedia’s merger with Orbitz. See DEPARTMENT OF JUSTICE, Justice Department Will Not Challenge Expedia's Acquisition of Orbitz, (Sept. 16, 2015), https://www.justice.gov/opa/pr/justice-department-will-not-challenge-expedias-acquisition- orbitz. There, the Division looked at evidence regarding post-merger pricing on both sides of the travel booking platforms, including evidence of competitive constraints the platforms placed on one another, as well as the constraints generated by rivals. Id. Significantly, the Division found that “the online travel business is rapidly evolving. In the past 18 months, for example, the industry has seen the introduction of TripAdvisor’s Instant Booking service and Google’s Hotel and Flight Finder with related booking functionality.” Id. Through review of evidence including the parties’ documents, transactional data from the parties and other industry players, and industry interviews, the Division was focused on the competitive check the emerging and innovative entrants placed on incumbents in the digital economy.

Rapidity of developments in a digital economy sector, however, will not always translate to low barriers to entry. In its review of Draftkings’ proposed combination with FanDuel, the FTC concluded that entry would be unlikely. See In re DraftKings, Inc./FanDuel Limited, F.T.C. Docket No. 9375, Complaint ¶78 (July 19, 2017). After defining a market for daily fantasy sports competitions, in contrast to season-long fantasy competitions, the administrative complaint asserted that there was substantial head-to-head competition between the parties with respect to price, competition size, competition format, and product features. See id. at ¶¶49-77. The complaint then set forth evidence the FTC considered regarding entry. First, the complaint alleged a network effects barrier, in that “[t]he largest obstacle, among many, is the difficulty and cost of acquiring a critical mass of [daily fantasy sports] users on a provider’s platform.” Id. at ¶79. As Commissioner Ohlhausen subsequently noted, “These platforms were subject to various benefits of scale, including the fact that having more players allowed the operator to run larger contests with bigger prizes.” Ohlhausen GAES speech at 10. She observed that network effects “had a significant impact on our analysis and decision to challenge the merger.” Id. at 11. The FTC did not, however, condemn the merger simply because the market was subject to network effects. As the Complaint makes clear, additional evidence was presented of failed entry: “several large, sophisticated, well-capitalized technology or sports media companies have either

Page 9 considered and rejected plans to enter the [relevant] market, or attempted to enter with little or no success.” Complaint at ¶80. Moreover, regulatory compliance obstacles facing online gambling companies created an additional barrier to entry. Id. at ¶78.

The Northern District of California similarly credited evidence that network effects in the syndication of ratings and reviews were an important barrier preventing entry in the Division’s challenge to the Bazaarvoice and Power Reviews merger. See United States v. Bazaarvoice, Inc., No. 13-CV-00133, 2014 WL 203966, at ¶¶239-252 (N.D. Cal. Jan. 8, 2014). That evidence included the testimony of economic experts, competitor documents, and parties’ own internal and public documents expressly describing syndication as an entry barrier. Id. The Court cited evidence not only that there was a network effect across the manufacturer/brand and retailer platforms, id. at ¶243, but that it created a “significant and durable” barrier, in large part based on the parties’ internal documents suggesting that competitors would have a hard time signing up either brands or retailers and thus allowing Bazaarvoice to “lock competitors out of a meaningful data set.” Id. at ¶¶243-247. But the Court did not stop there, and additionally considered evidence of switching costs, development costs and reputational barriers. Id.at ¶¶253-265. Finally, the Court also observed that there was no evidence of “even preliminary analysis of the viability of [entry into] the market.” Id. at *71.

Court decisions and agency materials make clear that evidence of disruption – actual or fear thereof – which is driving either price or non-price improvements, as well as the source of that disruption, will continue to be critical to an analysis of market power in digital markets. See e.g., Section 6.4 of the 2010 Horizontal Merger Guidelines (capturing how the agencies will consider a merger’s effect on innovation, both actual and incented); In the matter of Verisk/EagleView, F.T.C. Docket No. 9363, Complaint (Dec. 16, 2014) (challenging deal, which parties ultimately abandoned, citing evidence that acquirer recently entered market with product alleged to be uniquely positioned to disrupt target’s industry-leading technology); Fed. Trade Comm'n v. Steris Corp., 133 F. Supp. 3d 962 (N.D. Ohio 2015) (denying preliminary injunction because entry was not imminent, based on evidence that alleged disruptive entrant lacked customer commitments, lacked available capital, had not deployed capital to build a competing facility, and was not likely for business reasons to do so).

III. CONCLUSION

Calls to alter the framework for assessing market power in the digital economy will certainly continue, but the evidence utilized by agencies and courts will remain the same, at least in the near-term. And the focus of that evidence will remain on the ultimate question of competitive effects. As Deputy AAG Nigro recently noted:

Some have said that we need new tools to address these new data issues. Advocates for new tools tend to cite network effects, and argue that the winner-take-all nature of digital markets and the existence of tipping points mean that the typical means of assessing market power are ineffective. The presence of network effects may indeed change the competitive landscape, but markets subject to network effects are not inherently less competitive: economic theory shows that—although network effects may provide firms with significant competitive advantages within the market—markets with network effects can produce intense competition among firms competing for the market. In other words,

Page 10

firms compete intensely to become the solution most people choose. And, it is the promise of the rewards that come with that position that creates an incentive for firms to take outsized risks to invest in the development of innovative products and services.

Nigro M&TCC speech, at 6. At the same time, the flexibility of the antitrust toolkit, which Commissioners Ohlhausen and McSweeny have recently stressed, will continue to allow practitioners to creatively gather and present evidence that illuminates the market realities.

It is also helpful to remember, while considering commentary calling for changes, that the United States antitrust regime is a norms-based process, which is simultaneously resilient and constantly evolving. As Professor William Kovacic observed in The Modern Evolution of U.S. Competition Policy Enforcement Norms:

The evolutionary character of competition policy means that views about what constitutes a good public enforcement program will change over time as understanding about the operation of the economy grows . . . . Good enforcement practice requires continuing reassessment of the economic foundations of specific interpretations of the antitrust statutes . . . . The intellectual status quo at any moment usually reflects a synthesis of older and newer thinking rather than a wholesale displacement of earlier perspectives.

71 ANTITRUST L.J. 377 (2003). Again, this presents an opportunity for practitioners to marshal evidence that promotes a new understanding of the operation of the digital economy, while working within the existing analytical framework.