copyrightai
copyrightai
copyrightai
Thomas A. Lambert
Report
https://gaidigitalreport.com/2020/08/09/introduction/
Chapter
https://gaidigitalreport.com/2020/08/25/rent-seeking-and-public-choice-in-
digital-markets/
This paper can be downloaded without charge from the Social Sciences Research
Network Electronic Paper Collection at:
http://ssrn.com/abstract=3728990
Thomas A. Lambert
Abstract
The branch of economics known as “public choice” uses the tools of economics to analyze
political behavior. Rejecting idealized versions of government intervention, it assumes
that individuals participating in the political system are rational self-interest maximizers
(homo economicus). A key insight of public choice is that private firms will endeavor to
boost their profits above competitive levels by co-opting government’s unique right to
coerce. Such “rent-seeking” behavior typically involves exploiting the self-interest of gov-
ernment officials and is often successful even though it tends to reduce social welfare by
softening competition, diverting productive resources to non-productive activities, and
destroying the value of competing firms’ productive investments. This article documents
instances of rent-seeking and other adverse public choice concerns in digital markets. It
considers two broad categories of rent-seeking behavior: efforts to procure an effective
subsidy and attempts to raise rivals’ costs. Examples within the former category are
(1) news publishers’ attempts to force digital platforms to purchase news snippets and
(2) efforts by producers of digital content to free-ride off the investments of operating
systems developers. Examples of rent-seeking by raising rivals’ costs include (1) domi-
nant digital platforms’ lobbying for rules with which their rivals will have difficulty com-
plying; (2) efforts by firms competing with platforms that host user-generated content to
weaken the protections of Section 230 of the Communications Decency Act; and (3) the
creation and financing of “astroturf” (fake grass-roots) groups that instigate legal action
against group members’ competitors. As they craft policies to address purported market
failures in digital markets, policy makers should account for potential government failure
in the form of enhanced rent-seeking and other public choice concerns.
INTRODUCTION
social institutions in that it alone possesses the right to use force to achieve its objectives. 1
Members of liberal societies have generally agreed that it should exercise that extraordi-
nary authority only to protect citizens’ rights or when there are strong reasons to believe
that private ordering—individuals making their own decisions about how to use the
things at their disposal—is likely to misallocate resources and thereby reduce social wel-
fare. Common situations in which such resource misallocations are likely to occur in-
clude the classic market failures of externalities, public goods, information asymmetry,
mental, tax, securities, and antitrust laws—have been justified as means of addressing
A market failure, though, is not a sufficient condition for a governmental fix. Be-
cause government interventions can themselves create losses, policymakers should al-
ways balance the expected welfare gain from averting a market failure against any
1 Max Weber, Politics as Vocation, in FROM MAX WEBER: ESSAYS IN SOCIOLOGY 78 (H. H. Gerth & C. Wright
Mills eds.,1958) (observing that government possesses “a monopoly on the legitimate use of physical force
within a given territory”).
2 An externality occurs when an actor does not bear all the cost or capture all the benefit of its actions, as
with a polluting factory. See THOMAS A. LAMBERT, HOW TO REGULATE: A GUIDE FOR POLICYMAKERS 22–29
(2017). A public good is an amenity that is capable of being consumed without being depleted (non-rival-
rous) and cannot be withheld from individuals that did not contribute to its creation (non-excludable)—
e.g., national defense. See id. at 60-66. Information asymmetry occurs when there is a great disparity be-
tween the information available to the parties to a transaction, as with a corporation’s sale of stock to an
investor. See id. at 185–91. Market power exists when there is an absence of competition because of mo-
nopoly or collusion. See id. at 135–45.
3 See id. at 29–57, 66–76, 145–53, 193–207.
should occur “at the margin,” meaning that the likely welfare gain (market failure loss
welfare loss (from government failure) that the extra bit of restrictiveness is likely to pro-
All this suggests that policymakers should carefully account for the ways that gov-
it turns out, government interventions regularly produce two sets of welfare losses. One
set occurs when interventions misallocate resources because governmental planners lack
their highest and best ends. Losses from this sort of “knowledge problem,” which was
directives become more prescriptive and less flexible, and as the governmental planners
issuing them become further removed in time and space from the processes they are di-
right to coerce may be exploited to secure private benefits. The individuals charged with
tend to make decisions that inure to their own interests and may not seek to maximize
the aggregate welfare of all citizens. Outside the government, individuals and groups
will seek to procure governmental directives that benefit them, regardless of the direc-
tives’ effects on others. And because voters face limitations on their time, attention, and
4 See generally F. A. Hayek, The Use of Knowledge in Society, 35 AM. ECON. REV. 519 (1945).
that officials exercise state power with an eye toward maximizing overall social welfare.
The branch of economics called “public choice” has predicted and documented these
tendencies, so we may refer to this second set of welfare losses as “public choice con-
cerns.” 5
This chapter examines how public choice concerns, most prominently “rent-seek-
ing” behavior, have been manifested in initiatives to regulate digital markets. The chap-
ter first summarizes key insights from public choice and describes the phenomenon of
rent-seeking. It then documents the existence of rent-seeking activity and other public
choice concerns in the regulation of digital markets. It closes with a brief observation
about how the structure of regulatory interventions may constrain or exacerbate rent-
Public choice is “the use of economic tools to deal with traditional problems of
political science”—i.e. economic analysis of political behavior. 6 Nobel laureate James Bu-
chanan, one of the fathers of public choice theory, described it as “politics without ro-
mance.” 7 In the romantic vision of democratic politics, citizens inform themselves of po-
litical candidates’ plans for exercising governmental power and then vote for those can-
didates whose plans they believe will be most beneficial. The elected candidates then
enact legislation they believe will provide the greatest benefit to the citizenry as a whole.
Bureaucrats, who answer to an elected executive who also seeks to maximize the citi-
zenry’s welfare, enforce the laws and implement the programs the legislature has
5 See generally William F. Shughart II, Public Choice, in THE CONCISE ENCYCLOPEDIA OF ECONOMICS 427 (David
R. Henderson ed., 2008).
6 Gordon Tullock, Public Choice, in NEW PALGRAVE DICTIONARY OF ECONOMICS (1987).
7 James Buchanan, Politics Without Romance: A Sketch of Positive Public Choice Theory and Its Normative Impli-
cations, in THE THEORY OF PUBLIC CHOICE–II 11 (J. Buchanan & R. Tollison eds., 1984).
enacted, with an eye toward maximizing their effectiveness for the good of society. Gov-
ernment’s monopoly over the use of force is thus effectively harnessed to protect individ-
ual rights and prevent welfare losses that would otherwise result from market failures
The problem with this romantic view of politics, public choice scholars assert, is
that it assumes people make choices in the political arena differently than in other con-
texts. 8 When people make decisions about buying and selling things, they usually seek
typically seek personal happiness, which might involve working for the public good but
often does not. Even in the area of non-pecuniary relationships, people select friends and
mates not to benefit society as a whole but to make themselves happy. Not only do peo-
ple tend to pursue their own interests, they typically do so in a logical, internally con-
sistent fashion. Accordingly, traditional economics has started with the assumption that
Rejecting the romantic vision of politics, public choice theory assumes that people
do not shed their fundamental natures when stepping into the political arena. 9 Public
choice instead embraces the economist’s “rational choice” model of human behavior and
how (and whether) to vote. They generally “vote their pocketbooks,” and, given the low
probability that any individual vote will sway an election outcome, they invest little in
educating themselves on the candidates and the issues at stake. People running for office,
8 See James M. Buchanan, The Public Choice Perspective, in POLITICS AS PUBLIC CHOICE, VOLUME 13 OF THE
COLLECTED WORKS OF JAMES M. BUCHANAN 21-22 (2000) (discussing homo economicus element of public
choice perspective).
9 See James D. Gwartney & Richard E. Wagner, Public Choice and the Conduct of Representative Government, in
PUBLIC CHOICE AND CONSTITUTIONAL ECONOMICS 7 (James D. Gwartney & Richard E. Wagner eds., 1988)
(“Since there is no evidence that entrance into a voting booth or participation in the political process causes
a personality transformation, there is sound reason to believe that the motivation of participants in the
market and political processes is similar.”).
10 See Shughart, supra note 5 (describing public choice theory).
who apparently derive utility from holding positions of power, take reasonable steps to
secure their election. They embrace positions that will generate votes in their favor,
which means the positions are either popular with voters generally or are favored by and
salient to individuals or groups that are especially likely to provide campaign financing.
Once in office, politicians make decisions to promote, support, or oppose legislation ac-
cording to the decisions’ effects on their reelection prospects. The non-elected bureau-
crats charged with enforcing legislation or adopting and implementing regulations to fill
its gaps make decisions that will benefit them personally. On discretionary matters, they
tend to support outcomes that expand their agency’s turf and budget and enhance their
Public choice theory also makes predictions about the behavior of business organ-
izations. Because the individuals who manage those organizations typically benefit when
their companies do well, they will try to harness government power to enhance their
firms’ profits. And because firm managers are rational self-interest maximizers, they will
take the tack most likely to generate success: they will play upon government officials’
self-interest.
“rent,” first introduced by David Ricardo in the 19th century, means the payment to a
factor of production in excess of the amount required to keep the factor in its current
use. 12 For example, if an employee is paid a salary of $100,000 but would remain in her
current job for $90,000, she is capturing $10,000 in rent. 13 Of course, there is nothing in-
herently troubling about pursuing rent; any employee who would stay in her job at her
11 See generally David R. Henderson, Rent Seeking, in THE CONCISE ENCYCLOPEDIA OF ECONOMICS 427 (David
R. Henderson ed., 2008).
12 DAVID RICARDO, ON THE PRINCIPLES OF POLITICAL ECONOMY AND TAXATION (1817).
current salary but nevertheless asks for a raise is technically seeking rents. The term rent-
seeking, however, is typically used more narrowly. The concept, which originated with
Gordon Tullock in 1967 and was given its label by Anne Krueger in 1974, refers to efforts
one’s transacting partner for a greater share of the surplus created by a deal, but by har-
nessing the government’s coercive powers. 14 For example, firms may lobby legislators to
provide a subsidy for a good they produce. They may seek imposition of a tariff on com-
peting foreign goods. They may try to persuade legislators or regulators to enact a rule
that imposes extra costs on their rivals. With all these endeavors, they seek to channel
government’s coercive power in a way that benefits them, and they do so by exploiting
ing scenario are acting as rational self-interest maximizers, rent-seeking is a classic public
choice concept.
away from the creation of wealth and toward its redistribution. The money a firm spends
on lobbying, public relations, and other endeavors aimed at persuading government of-
ficials is not available for creative activities like research and product development. And
it is not just money; as firm managers devote more time and attention to procuring gov-
ernmental favors, less of each is available for productive endeavors. Second, to the extent
sources away from their highest and best ends. 15 Finally, to the extent rent-seeking drives
14 See Gordon Tullock, The Welfare Costs of Tariffs, Monopolies and Theft, 5 WESTERN ECON. J. 224 (1967); Anne
O. Krueger, The Political Economy of the Rent-Seeking Society, 64 AM. ECON. REV. 291 (1974).
15 Market competition forces sellers to lower their prices toward the level of their cost. Sellers who face less
competition can charge higher prices by producing less (so that the market-clearing price rises). When
sellers fail to produce units whose cost of production and distribution is less than the value the units would
create for the buyers who would purchase them, there is a loss in welfare. See LAMBERT, supra note 2, at
143.
rivals from the market, it squanders their non-recoverable investments. 16 For example, if
a firm has installed specialized equipment but then finds itself driven out of business by
Despite the fact that the policies and rules sought by rent-seekers routinely reduce
social welfare, they are frequently enacted. There are several reasons for this. One is that
voters, who have the power to punish legislators and bureaucrats who employ govern-
ment’s coercive power in a welfare-reducing manner, 17 are often unaware of how those
officials’ decisions have harmed overall welfare—and the officials know it. As rational
self-interest maximizers, voters will invest in information needed to exercise their voting
right effectively only to the point at which the (decreasing) marginal benefit of additional
information equals the (increasing) marginal cost of obtaining it. Because each individual
how to vote is quite low, and so voters spend little to become informed. 18 They rationally
remain ignorant of how their elected officials are exercising government power, which
frees those officials to make the decisions that benefit themselves even when those deci-
sions reduce overall social welfare. Rent-seekers, in turn, are adept at ensuring that the
decisions they prefer will somehow inure to the personal benefit of the government offi-
Voters’ rational ignorance interacts with another dynamic that makes rent-seeking
initiatives particularly likely to succeed. The policies sought by rent-seekers always con-
centrate special benefits on their proponents, who therefore have an incentive to lobby
for their adoption. Many times, however, the costs of the initiatives are distributed
16 See Herbert Hovenkamp, Antitrust’s Protected Classes, 88 MICH. L. REV. 1, 18-19 (1989) (describing “WL3
losses” from monopoly rent-seeking).
17 Voters’ ability to punish bureaucrats is indirect. Voters elect the executive, who exerts control over bu-
reaucrats.
18 See Shughart, supra note 5 (discussing voters’ rational ignorance).
the domestic producers of a product—likely a small group—but imposes costs in the form
of slightly higher prices on all the domestic consumers of that product—likely a large
group. The total cost of the tariff may well exceed the benefit to the domestic firms it
favors, but because each consumer bears just a tiny portion of that cost, no one is willing
to incur the cost of counter-lobbying against the tariff. After all, anyone who did so
would bear all the cost of her lobbying while capturing only a small portion of any benefit
produced by her efforts. 19 Public choice thus predicts that policies involving “concen-
trated benefits and diffused costs” will be enacted even when they are, on the whole,
welfare-reducing.
Of course, lobbying for a special benefit at the expense of the general public is
difficult. Consumers may withhold business from firms they think are abusing govern-
ment power, and, despite voters’ rational ignorance, government officials will want to
avoid any appearance that they are favoring the interests of a few over those of the public
at large. Rent-seekers may therefore seek to hide behind groups that share their policy
goals but for public-spirited, rather than self-interested, reasons. Bruce Yandle has
groups that in the early 20th century pushed hardest for liquor prohibition. 20 Baptists,
who emphasized the social evils of alcohol, made a passionate and public “pro-social”
case for prohibition. Bootleggers promoted prohibition too, but did so behind the scenes
and in the hopes of squelching competition and earning monopoly profits on their illegal
products.
Taken together, voters’ rational ignorance, the fact that protectionist rules often
create concentrated benefits but diffused costs, and the bootleggers-and-Baptists dy-
namic render rent-seeking an often successful strategy for enhancing private profits. And
19 See generally MANCUR OLSON, JR., THE LOGIC OF COLLECTIVE ACTION: PUBLIC GOODS AND THE THEORY OF
AND GROUPS (1965).
20 Bruce Yandle, Bootleggers and Baptists—The Education of a Regulatory Economist, 7 (3) REGULATION 12 (1983).
success is especially likely when the officials wielding government power have broad
discretion, regular contact with those they regulate, and little direct political accountabil-
ity. That combination frequently exists when a government agency is charged with con-
tinual oversight of some narrow sector of the economy. A key insight of public choice
theory is that such agencies tend to become “captured” by their regulatees and end up
exercising their discretionary authority in a manner that preserves or enhances the regu-
latees’ profits by protecting them from competition. 21 In exchange, the regulatees provide
benefits—from personal affection, to perks, to future job prospects—to the officials wield-
ing power. Once again, all individuals are pursuing their own interests.
Having laid the foundation, we turn now to consider some recent examples of
rent-seeking activity and other public choice concerns in the regulation of digital markets.
The regulation of digital markets presents all sorts of opportunities for rent-seek-
ing, and examples of such behavior abound. While there are many different ways to co-
opt the government’s coercion right to secure a private benefit, most instances of rent-
seeking fall into two categories: seeking some sort of subsidy or seeking to foreclose com-
petition. Parts A and B of this Section describe recent examples of those two strategies.
Part C then considers why such strategies may succeed in digital markets even when the
A. Procuring a Subsidy
Two prominent examples of firms’ seeking to exploit government power over dig-
ital platforms to procure some sort of subsidy are the campaign by traditional news pub-
lishers to extract payment from Google and Facebook and the efforts of certain software
21See George Stigler, The Theory of Economic Regulation, 2(1) BELL J. ECON. & MGT. SCI. 3, 3 (1971) (observing
that in regulated industries, “as a rule, regulation is acquired by the industry and is designed and operated
primarily for its benefits”).
and digital content providers—most notably, music streaming service Spotify and video
game producer Epic Games—to free-ride off the investments Apple and Google have
fell to its lowest level since 1940, and between 2008 and 2018, revenues at U.S. newspapers
fell from $37.8 billion to $14.3 billion, a 62% decline. 22 During the same period, U.S. news-
room employment dropped by nearly half, from 71,000 to 38,000 workers. 23 Similar
The primary reason for newspapers’ difficulties is competition created by the In-
22 Elizabeth Grieco, Fast Facts about the Newspaper Industry’s Financial Struggles as McClatchy Files for Bank-
ruptcy, PEW RESEARCH CENTER (Feb. 14, 2020), https://www.pewresearch.org/fact-tank/2020/02/14/fast-
facts-about-the-newspaper-industrys-financial-struggles/.
23 Id.
24 E.g. Esther Kezia Thorpe, How People in the UK Are Accessing News: 6 Key Findings, DIGITAL PUBLISHING
over time. In the early 1970s, the Gallup organization began polling Americans about their trust in mass
media. Early polls found that around seven-in-ten Americans trusted the media a “great deal” or “fair
amount” (68% in 1972, 69% in 1974, 72% in 1976). In 2019, the figure had fallen to 41%, up from a low of
32% in 2016. The partisan divide in media trust is stark, with 69% of Democrats, 36% of independents, and
only 15% of Republicans expressing a “great deal” or “fair amount” of trust in the media. See Megan
Brenan, Americans’ Trust in Mass Media Edges Down to 41%, GALLUP (Sept. 26, 2019), https://news.gal-
lup.com/poll/267047/americans-trust-mass-media-edges-down.aspx. These figures correlate somewhat
with the political leanings of journalists. In 1971, the breakdown of journalists belonging to a political party
was 35.5% Democrat versus 25.7% Republican. By 2002, the composition had shifted to 35.9% Democrat
versus 18% Republican. And by 2013, the balance was 28.1% Democrat to 7.1% Republican. See Chris
10
like Craigslist, which offers consumers free classified advertising for most products and
services, eroded newspapers’ ability to charge monopoly rates. 26 It did not help that
newspapers cannibalized their own classified revenues by creating sites like Cars.com
and Autotrader.com. 27 With respect to non-classified advertising, the Internet has cut
chief concern—physical newspapers must contend with the entire Internet, and even the
In light of the challenges the Internet has created for their advertising-focused
funding model, newspapers have sought to employ the government’s coercive power to
increase their revenues. Outside the U.S., media groups have successfully lobbied for
rules requiring Google and Facebook to make payments to newspapers. News organiza-
tions in Germany, France, and Spain procured copyright law amendments requiring the
platforms to pay licensing fees when they display excerpts or photographs from a pub-
lisher’s news articles. 28 In Germany, Google was able to avoid such fees by procuring
liability releases from publishers, which found that traffic to their websites would plunge
without Google’s help in directing readers to them. 29 The Spanish legislature foreclosed
Cillizza, Just 7 Percent of Journalists are Republicans. That’s Far Fewer than even a Decade ago., WASH. POST (May
6, 2014), https://www.washingtonpost.com/news/the-fix/wp/2014/05/06/just-7-percent-of-journalists-are-
republicans-thats-far-less-than-even-a-decade-ago/.
26 From 2000 to 2012, U.S. newspapers’ classified revenue dropped from $19.6 billion in 2000 to $4.6 billion
in 2012. See John Reinan, How Craigslist Killed the Newspapers’ Golden Goose, MINN. POST. (Feb. 3, 2014),
https://www.minnpost.com/business/2014/02/how-craigslist-killed-newspapers-golden-goose/.
27 See Jack Shafer, Don’t Blame Craigslist for the Decline of Newspapers, POLITICO (Dec. 13, 2016),
https://www.politico.com/magazine/story/2016/12/craigslist-newspapers-decline-classifieds-214525.
28 Eric Auchard, Google to Shut Down News Site in Spain over Copyright Fees, REUTERS (Dec. 11, 2014),
https://www.reuters.com/article/us-google-spain-news/google-to-shut-down-news-site-in-spain-over-
copyright-fees-idUSKBN0JP0QM20141211).
29 Id.
11
that possibility by declaring that Spanish publishers’ rights to license fees was “inaliena-
ble,” leading Google to remove its Google News service from its Spanish site. 30
When the French government required Facebook and Google to pay for the use of
photos and snippets, the two companies again sought to procure the sort of liability re-
leases used in Germany. 31 For French publishers that did not release Google from liability
for the display of snippets and photos from their news stories, Google began displaying
only titles and URLs linking to the stories. 32 Publishers then complained that such a dis-
play format reduces their traffic, whereupon the French competition authority entered an
interim order requiring Google to continue displaying snippets in accordance with pub-
lishers’ wishes. 33 At least temporarily, then, French publishers have succeeded in per-
suading the government to force Google and Facebook to display the snippets publishers
believe will maximize their traffic and to pay the publishers for use of those snippets.
law rather than copyright. In late 2017, after an intense lobbying campaign by news pub-
lishers including Rupert Murdoch’s News Corporation, the Australian government di-
their market power.34 The Commission released a Preliminary Report on its findings in
30 Id.
31 See AFP, Facebook Refuses to Pay French Media for Links, THE LOCAL.FR (Oct. 26, 2019), https://www.thelo-
cal.fr/20191026/facebook-refuses-to-pay-french-media-for-links.
32 Id.
33 Natasha Lomas, France’s Competition Watchdog Orders Google to Pay for News Reuse, TECHCRUNCH (Apr. 9,
2020), https://techcrunch.com/2020/04/09/frances-competition-watchdog-orders-google-to-pay-for-news-
reuse/.
34 See Anne Davies, World Watches as Australian Regulator Rules on Facebook and Google, THE GUARDIAN (Dec.
2, 2018), https://www.theguardian.com/media/2018/dec/03/world-watches-australian-regulator-facebook-
google.
12
December 2018 35 followed by a Final Report in June 2019. 36 In the Final Report, the Com-
mission recommended that Google and Facebook collaborate with news publishers to
produce a voluntary code of conduct that would include an obligation on the part of the
pets they displayed. 37 In December 2019, the Australian government accepted that rec-
ommendation and directed the ACCC to see that such a voluntary code was negotiated
by November 2020.38
In April 2020, shortly before Google was to report its progress in negotiating the
voluntary code, the Australian government changed course and directed the ACCC to
set forth mandatory rules requiring digital platforms to share their advertising revenue
with news publishers and to provide the publishers with advance notice of any algorithm
changes that could affect page rankings and displays. The government cited news pub-
lishers’ loss of advertising revenue due to the COVID-19 pandemic as a basis for its ab-
rupt shift to a mandatory code. 39 In announcing the change, Australian Treasurer Josh
Frydenberg acknowledged the Spanish and French experiences in which Google and Fa-
cebook removed news snippets in response to payment obligations, and he insisted that
Australia would avoid such evasion to “become the first country in the world to
2019), https://www.accc.gov.au/system/files/Digital%20platforms%20inquiry%20-%20final%20report.pdf.
37 Id. at 32. The code was to include a commitment that “where the digital platform obtains value, directly
or indirectly, from content produced by news media businesses, that the digital platform will fairly nego-
tiate with news media businesses as to how that revenue should be shared, or how the news media busi-
nesses should be compensated.” Id.
38 See Press Release, Josh Frydenberg, Response to Digital Platforms Inquiry, Treasurer of the Common-
lationship Between Digital Platforms and Media Companies, Treasurer of the Commonwealth of Australia
(Apr. 20, 2020), https://ministers.treasury.gov.au/ministers/josh-frydenberg-2018/media-releases/accc-
mandatory-code-conduct-govern-commercial.
13
successfully require payment for content.” 40 The ACCC unveiled a draft of its mandatory
Throughout their lobbying campaigns, news organizations have insisted that dig-
ital platforms are unfairly taking and profiting from their content. 42 But news publishers
can easily prevent digital platforms from displaying their published material. They can
use the Robots Exclusion Standard (Robots.txt) to prevent content scraping, 43 and Google
provides tools publishers may use to block snippets or control their length.44 Despite
these capabilities, newspapers typically refrain from restricting the display of snippets
for an obvious reason: they benefit from the traffic digital platforms provide. Tradition-
ally, newspapers paid to have headlines and short excerpts of content publicized in order
to increase readership and thereby enhance sales, subscriptions, and advertising reve-
nue. 45 Google and Facebook provide such publicity for free, and newspapers obviously
value it, as evidenced by the ease with which Google was able to procure liability releases
40 Josh Frydenberg, Here’s News — We’ll Hold Facebook and Google to Account, THE AUSTRALIAN (Apr. 20,
2020), https://www.theaustralian.com.au/commentary/heres-news-well-hold-facebook-and-google-to-ac-
count/news-story/92f84d6e4442512249c3ac5dd8a0516d.
41 See Exposure Draft of Treasury Laws Amendment (News Media and Digital Platforms Mandatory Bar-
and other web robots about which areas of the website should not be processed or scanned. See About /
robots.txt, ROBOTS.TXT, https://www.robotstxt.org/robotstxt.html (last visited Oct. 01, 2020).
44 See GOOGLE SEARCH, ROBOTS META TAG, DATA-NOSNIPPET, AND X-ROBOTS-TAG SPECIFICATIONS, https://de-
3, 2020), https://australia.googleblog.com/2020/05/responding-to-revised-publisher-code.html.
46 See Auchard, supra note 28. Germany’s largest publisher, Axel Springer, abandoned a plan to block
Google for refusing to pay for content after a consortium of around 200 German publishers saw their online
14
Far from simply protecting their content, which they could do themselves, power-
ful legacy media companies have succeeded in convincing the Australian government to
force Google and Facebook to continue providing news publishers with free publicity, to
pay the publishers when they do so, and to give the publishers information needed to
secure a preferred place in search rankings. The media companies insist that these man-
dates are needed because the platforms’ ad tech services extract an excessive portion of
revenues from advertisements on the publisher’s websites. 47 But the publishers need not
utilize those services in selling display ads. They could sell their advertising space di-
rectly and pocket all the advertising revenue, or they could utilize competing intermedi-
aries; as the ACCC acknowledged, competitors exist at every stage of the digital adver-
tising sales chain. 48 News publishers that continue to use the digital platforms’ ad tech
ing to pay the most for it. Google, of course, demands compensation for that valuable
ad-matching service, but it charges similar fees to all web publishers that use its ser-
vices—blogs, affinity group webpages, medical informational portals, and digital news
sites alike. When news organizations lobby for special rules for their websites, they are
traffic plunge after they blocked the company. Id. Traffic from Facebook also appears to be quite important
to publishers. When Facebook experimented in several countries with moving professional news stories
into a separate feed called Explore, publishers in the countries complained that traffic to their news sites
plummeted. See Sheera Frenkel, Nicholas Casey & Paul Mozur, In Some Countries, Facebook’s Fiddling Has
Magnified Fake News, N.Y. TIMES (Jan. 14, 2018), https://www.nytimes.com/2018/01/14/technology/facebook-
news-feed-changes.html?smid=fb-nytimes&smtyp=cur).
47 See, e.g., NEWS CORP AUSTRALIA, SUBMISSION TO THE AUSTRALIAN COMPETITION AND CONSUMER
COMMISSION: RESPONSE TO THE DIGITAL PLATFORMS INQUIRY PRELIMINARY REPORT 13, 30 (Mar. 1, 2019),
https://www.accc.gov.au/system/files/News%20Corp%20Australia%20%28March%202019%29.pdf.
48 See DIGITAL PLATFORMS INQUIRY FINAL REPORT, supra note 36, at 128. See also DANIEL S. BITTON & STEPHEN
LEWIS, CLEARING UP MISCONCEPTIONS ABOUT GOOGLE’S AD TECH BUSINESS, Appendix A (May 5, 2020),
https://www.accc.gov.au/system/files/Google%20-%20Report%20from%20Daniel%20Bit-
ton%20and%20Stephen%20Lewis%20%285%20May%202020%29.pdf (cataloguing competitors at each
level of ad tech stack).
15
prising in light of public choice theory. It stands to reason that elected officials pursuing
their own interests would more likely support news outlets than digital platforms. The
former control press coverage and therefore have great sway over voters who, again, are
platforms may have money to spend on campaigns and can publicize or attempt to hide
news stories, but they do not actually create the stories that may affect votes.
paign to force platforms to share advertising revenues and favor the publishers with ad-
vance notice of algorithm changes, various public interest groups have offered their sup-
Government officials who might normally be reluctant to take revenue from one set of
private businesses and give it to another with greater political sway could therefore reas-
sure themselves—and any skeptical voters—that they were simply taking the actions nec-
In the United States, news publishers have sought to extract rents from digital
platforms by lobbying for an exemption from the antitrust laws. 50 Their efforts
49 See, e.g., Victor Pickard, Journalism’s Market Failure Is a Crisis for Democracy, HARV. BUS. REV. (Mar. 12,
2020), https://hbr.org/2020/03/journalisms-market-failure-is-a-crisis-for-democracy; Matt Stoller, Tech Com-
panies Are Destroying Democracy and the Free Press, N.Y. TIMES (Oct. 17, 2019), https://www.ny-
times.com/2019/10/17/opinion/tech-monopoly-democracy-journalism.html; Ofcom: Newspapers Play ‘Vital
Role’ In Democratic Society, NEWS MEDIA ASS’N (June 21, 2018), http://www.newsmediauk.org/News/ofcom-
newspapers-play-vital-role-in-democratic-society; Barry Lynn, Google and Facebook are Strangling the Free
Press. Democracy is the Loser, THE GUARDIAN (July 26, 2018), https://www.theguardian.com/commentis-
free/2018/jul/26/google-and-facebook-are-strangling-the-free-press-to-death-democracy-is-the-loser.
50 Nitasha Tiku, Publishers Could Get a New Weapon Against Facebook and Google, WIRED (Mar. 7, 2018),
https://www.wired.com/story/bill-would-let-publishers-gang-up-versus-facebook-and-google/ (reporting
that the “prime driver of the bill [creating an antitrust exemption for newspapers] is the News Media Alli-
ance,” a trade association made up of 2,000 American and Canadian newspapers, which “lobb[ied] for such
an exemption for a year”).
16
2018. 51 According to a press release announcing the bill, it would allow “small publishers
to band together to negotiate with dominant online platforms to improve the access to
and the quality of news online.” 52 In reality, the bill would create a four-year safe harbor
for “any print or digital news organization” to jointly negotiate terms of trade with
Google and Facebook. 53 It would not apply merely to “small publishers” but would in-
stead immunize collusive conduct by such major conglomerates as Murdoch’s News Cor-
poration, the Walt Disney Corporation, the New York Times, Gannet Company, Bloom-
berg, Viacom, AT&T, and the Fox Corporation. 54 The bill would permit news organiza-
tions to fix prices charged to digital platforms as long as negotiations with the platforms
were not limited to price, were not discriminatory toward similarly situated news organ-
izations, and somehow related to “the quality, accuracy, attribution or branding, and in-
teroperability of news.” 55 Given the ease of meeting that test—since news organizations
could always claim that higher payments were necessary to ensure journalistic quality—
the bill would enable news publishers in the United States to extract rents via collusion
51 Journalism and Competition Preservation Act of 2018, H.R. 5190, 115th Cong. (2018), https://www.con-
gress.gov/bill/115th-congress/house-bill/5190/text.
52 See Press Release from Office of U.S. Representative David Cicilline , CICILLINE, COLLINS INTRODUCE BILL
To qualify for the safe harbor, a news organization must simply (1) have a dedicated professional staff that
creates original news, (2) be commercially marketed, and (3) produce content that is at least 25% current
news or related material. Id.
54 See John M. Yun, News Media Cartels Are Bad News for Consumers, COMPETITION POL’Y INT’L 2 (Apr. 2019).
55 Journalism and Competition Preservation Act of 2019, supra note 51, at § 3(b).
17
2. App Developers’ Efforts to Free-Ride off the Investments of Mobile Platform Provid-
ers
A second example of firms’ seeking rents in the form of an effective subsidy in-
volves attempts by Spotify, Epic Games, and other developers of digital applications to
use antitrust law to take a free ride on investments made by the producers of mobile
operating systems.
In January 2007, Apple released the first generation of its revolutionary iPhone—
tasks or providing entertainment. 56 The iPhone’s operating system, iOS, was originally
quickly realized, however, that the availability of additional, high-quality apps would
enhance the iPhone’s attractiveness. On July 10, 2008, the day before the second-genera-
tion “iPhone 3G” was released, Apple debuted the App Store, a platform for distributing
iPhone apps created by third-party software developers. 58 Originally launched with just
552 apps, 59 the App Store now hosts nearly 1.85 million, the vast majority of which are
produced by third-parties. 60 The iPhone ecosystem, however, remains closed; Apple does
not license its iOS operating system to other hardware producers, and the App Store is
the only outlet through which an iPhone user may license an app without violating the
terms of use on the iOS and thereby voiding the warranties on the iPhone and disabling
56 See John Markoff, Apple Introduces Innovative Cellphone, N.Y. TIMES (Jan. 10, 2007), https://www.ny-
times.com/2007/01/10/technology/10apple.html.
57 Id.
58 See Jason Snell & Peter Cohen, Apple opens iTunes App Store, MACWORLD (July 10, 2008),
https://www.macworld.com/article/1134380/app_store.html).
59 Id.
60 See STATISTA, NUMBER OF APPS AVAILABLE IN LEADING APP STORES AS OF 1ST QUARTER 2020,
18
a number of its security features. 61 Apple’s revenue from the iPhone primarily comes
from hardware sales, 62 so it takes pains to ensure that all software operated on the iPh-
one—its iOS as well as any third-party apps—works flawlessly. Any software glitches
Originally quite unique, the iPhone soon encountered close competition with the
system sponsored by Google and licensed for free to hardware producers throughout the
revenue, as Android products typically include Google Search, from which Google gen-
erates search ad revenue, and Google’s Chrome browser, which enables Google to collect
user data for targeted search ads. 65 Google also earns revenue from its own app store,
Google Play, which is usually included in Android devices. 66 Unlike Apple’s iOS, An-
droid permits users to install apps acquired from sources other than the app store that is
Because producers of Android-based products need not pay for the operating system (as
61 See Tim Worstall, The Problem with Apple’s Closed Apps Universe, FORBES (Aug. 31, 2012),
https://www.forbes.com/sites/timworstall/2012/08/31/the-problem-with-apples-closed-apps-uni-
verse/#5894e2f9794b.
62 See STATISTA, SHARE OF APPLE’S REVENUE BY PRODUCT CATEGORY FROM THE 1ST QUARTER OF 2012 TO THE
https://www.androidauthority.com/how-does-google-make-money-from-android-669008/.
66 Id.
19
cheaper than iPhones. 68 And as the operating system is open-source, Android permits
ness, though, also means that Android products are more susceptible to viruses, mal-
ware, and other security risks. 70 When it comes to apps, Android offers more variety,71
as app review in Google’s Play Store is less stringent than in Apple’s App Store 72 and
Android app developers may also distribute outside the Play Store. Again, however, this
openness comes at a cost: Android apps tend to be of lower quality than iOS apps, largely
because app developers must support many more versions of Android than of iOS.73 In
terms of users, Android barely edges out iOS in the United States, with 51.8% of
smartphone subscriptions versus 47.4% for iOS, but it dominates globally, with a 74.14%
68 See Simon Hill, Android vs. iOS: Which Smartphone Platform is the Best?, DIGITAL TRENDS (May 10, 2020)
(observing that Android phones tend to be more affordable than iPhones); Sean Keach, Android Phones
Nearly Three Times Cheaper than iPhone, TRUSTED REVIEWS (Feb. 2, 2015), https://www.trustedre-
views.com/news/android-phones-nearly-three-times-cheaper-than-iphone-2924886.
69 See Hoffman, supra note 64.
70 See Max Eddy, SecurityWatch: Android vs. iOS, Which Is More Secure?, PC MAGAZINE (Apr. 24, 2019); Lucas
Mearian, Android vs iOS security: Which is Better?, COMPUTERWORLD (Aug. 7, 2017), https://www.comput-
erworld.com/article/3213388/android-vs-ios-security-which-is-better.html (observing that “[w]hile all mo-
bile devices have inherent security risks, Android has more vulnerabilities because of its inherent open-
source nature, the slow pace with which users update the OS and a lack of proper app vetting”).
71 See STATISTA, NUMBER OF APPS AVAILABLE IN LEADING APP STORES AS OF 1ST QUARTER 2020,
Play Store When You Launch Your First App, YALANTIS, https://yalantis.com/blog/apple-app-store-and-
google-play-store/ (last visited Oct. 01, 2020); Mary Aleksandrova, How to Publish Your App on App Store and
Google Play? A Comprehensive Go-to-Market Guide, EASTERN PEAK (Jan. 3, 2018), https://eastern-
peak.com/blog/how-to-publish-your-app-on-app-store-and-google-play-comprehensive-go-to-market-
guide/.
73 See Jerry Hildenbrand, After 10 years, Android Apps Are Still Worse than their iOS Counterparts, ANDROID
20
Both Apple and Google earn revenues through their app stores. Apple retains 30%
of revenues from app licenses and in-app purchases of digital goods. 75 It earns no reve-
nue, however, from distributing free apps, which are typically ad-supported; from the
use of App Store apps to purchase physical goods and services such as household items,
delivered food, and ride-sharing services; or from out-of-app purchases of digital goods
that purchasers then enjoy on App Store apps, such as when a user who has downloaded
a free streaming video app then subscribes to the relevant streaming service outside the
app. 76 For in-app digital subscriptions, Apple’s revenue share drops to 15% after the first
year.77 Google’s revenue share from sales of Android apps through Google Play closely
mirrors Apple’s take through the App Store: 30% for paid apps and in-app purchases of
digital goods, with in-app subscription takes falling to 15% after the first year. 78 Other
Android app stores determine their own revenue shares. Some collect a lower percentage
of revenues on app sales, 79 and some developers of paid Android apps avoid fees alto-
There are, of course, significant costs in running a third-party app store. Apart
from developing and maintaining the technology required to produce apps, catalogue
them and bring them to the attention of interested users, and process distribution and
payment, platforms must ensure that the apps they distribute will work as described and
77 See id.
21
will not create technological glitches for users. App stores create value for the developers
whose apps they distribute by implicitly certifying that those apps are “safe and effec-
tive.” And that requires close examination of each app submitted for distribution, both
free and paid. Apple’s vetting process is particularly rigorous, 81 and consequently App
Store apps are generally considered to be of higher quality than those distributed through
Google Play and other Android app stores. 82 Not surprisingly, then, distributing through
the App Store is more lucrative for app developers. 83 In the third quarter of 2019, for
example, the App Store generated revenue of $14.2 billion on 8 billion first-time down-
loads, while Google Play earned just over half as much ($7.7 billion) on nearly three times
from the services app stores provide. As of March 2020, Spotify had 286 million monthly
active users, the vast majority of whom downloaded its app through Apple’s App Store
or Google Play. 85 Through the App Store alone, users have downloaded the Spotify app
or an update to it more than 300 million times. 86 Around 55% of Spotify’s active users
use its free service. 87 As they pay nothing to Spotify, distributing to them generates no
revenue for the App Store or Google Play. It does, however, make money for Spotify,
81 See Kif Leswing, Inside Apple’s Team that Greenlights iPhone Apps for the App Store, CNBC (June 22, 2019),
https://www.cnbc.com/2019/06/21/how-apples-app-review-process-for-the-app-store-works.html.
82 See supra notes 72–73 and accompanying text.
83 See C. Scott Brown, Top App Developers Make Gobs of Cash from Apple, Much Less from Google, ANDROID
are subscribers).
22
which sells ads on its free service and can earn more as the number of free users grows. 88
Any revenues the app stores receive for certifying, marketing, and distributing Spotify’s
app would come from in-app subscriptions to Spotify’s ad-free service, Spotify Premium.
Spotify, however, has made it difficult for the app stores that certify and distribute
its app (and its many upgrades) to receive any compensation for their efforts. In 2016, it
removed iOS users’ ability to upgrade to its Spotify Premium service using Apple’s in-
app payment system (“In-App Purchase”), and it thereby circumvented the requirement
to share revenue with Apple. 89 Users of the iOS app must leave the app and go to
Spotify’s app, users may upgrade to Premium by remaining in the app and submitting a
credit card number directly to Spotify, but they may not use Google’s more convenient
in-app payment system, Google Play Billing, which would retain a share of revenue for
Google.
Both Google and Apple have continued to support Spotify’s app and upgrades
despite the lack of payment for doing so. Apple, however, has refused to allow Spotify
to include an option for in-app payments that evade Apple’s revenue share by foregoing
Apple’s In-App Purchase system. It has also refused to approve versions of Spotify’s app
that would explicitly direct users on how to sign up for Spotify Premium outside the app,
and it has insisted that Spotify not directly notify app users about how to do so. Apple
download app with the ability to upgrade to enhanced service for a fee—while using
other in-app payment systems or directing users to pay for upgrades outside the app
https://www.cnet.com/news/apple-fires-back-spotify-pays-fees-on-less-than-1-percent-of-members/.
Some other digital service providers, including Netflix for video streaming, have taken the same tack. See
Nicole Nguyen, How App Makers Break Their Apps to Avoid Paying Apple, WALL ST. J. (June 28, 2020),
https://www.wsj.com/articles/how-app-makers-break-their-apps-to-avoid-paying-apple-11593349200.
23
could quickly dry up its App Store revenues, which are used to cover the considerable
cost of vetting, marketing, and distributing apps. Under the policies Spotify is demand-
ing, any developer of a paid app could avoid contributing to the App Store by charging
nothing for the app itself, locking all its functionality, and directing users outside the app
Since 2015, Spotify has engaged in an intensive effort to persuade government of-
ficials to force Apple to certify and distribute the Spotify app and updates without shar-
ing in Spotify’s revenues. The company first launched a lobbying campaign to procure
legislation or regulation that would achieve its goals. 90 It arranged meetings with con-
gressional leaders and hired a team of lobbyists to press the case that Apple’s insistence
on compensation for app distribution was a means to promote Apple’s own streaming
music service, Apple Music, over Spotify’s.91 Spotify also pressed its case to officials from
the U.S. Department of Justice and to the Federal Trade Commission’s Technology Task
90 See Tony Romm, Spotify Makes Case Against Apple in Congress, POLITICO (July 12, 2015), https://www.po-
litico.com/story/2015/07/spotify-makes-case-against-apple-in-congress-119976.
91 According to a Politico report, lobbyists for Spotify held a number of “secretive congressional meetings”
in 2015 to press their case against Apple. They met with, among others, staff members from the offices of
Rep. Bob Goodlatte (R-VA), chair of the House Judiciary Committee; Rep. Tom Marino (R-PA), chair of
House Judiciary’s antitrust subcommittee; and Sen. Al Franken (D-MN), a member of the antitrust subcom-
mittee of the Senate Judiciary Committee. Id. Spotify hired four outside lobbying shops to assist with its
efforts, id., spending $740,000 on federal lobbying in 2015. See David McCabe, Spotify Picks up Lobbyist Amid
Fight with Apple, THE HILL (August 12, 2016), https://thehill.com/policy/technology/291315-spotify-picks-
up-lobbyist-amid-fight-with-apple.
24
Force.92 And in March 2019, with much fanfare, 93 it filed an antitrust complaint against
Spotify’s efforts are paying off. On June 16, 2020, the European Commission an-
nounced that it had opened an investigation of Apple’s App Store policies. 95 It will ex-
amine whether Apple has violated European antitrust law by (1) refusing to allow pay-
ments within the app using non-Apple payment methods and (2) restricting app devel-
opers’ ability to instruct users to make payments outside the app.96 The Commission will
determine whether these policies, which are aimed at preventing free-riding on Apple’s
efforts to certify, market, and distribute apps, nevertheless “breach EU competition rules
position.” 97 The impetus for its investigation, the Commission explained, was Spotify’s
2019 complaint along with a similar complaint by “an e-book and audio book distribu-
tor.” 98
for no compensation—to certify, promote, and distribute Spotify’s app to the attractive
customer base Apple has assembled. The fact that Spotify has pursued action against
Apple but not Google, which shares Apple’s position as both the effective gateway to a
92 See Diane Bartz & Stephen Nellis, Exclusive: Antitrust Probers in Congress Ask Spotify to Detail Alleged Apple
Abuses – Sources, REUTERS (Oct. 4, 2019), https://www.reuters.com/article/us-tech-antitrust-apple-exclu-
sive/exclusive-antitrust-probers-in-congress-ask-spotify-to-detail-alleged-apple-abuses-sources-
idUSKBN1WJ1Y3.
93 When it filed its claim with the EU, Spotify launched a website entitled “Time to Play Fair,”
https://www.timetoplayfair.com, that is aimed at making its case to the public. The website includes a
summary of Spotify’s allegations and a “media kit” for reporters.
94 See Mark Sweney, Apple’s 30% App Store Commission Unfair, Spotify Claims, THE GUARDIAN (Mar. 13, 2019),
https://www.theguardian.com/technology/2019/mar/13/spotify-claim-apple-30-percent-app-store-com-
mission-unfair-european-commission-complaint.
95 See EUROPEAN COMMISSION PRESS RELEASE, ANTITRUST: COMMISSION OPENS INVESTIGATIONS INTO APPLE’S
97 Id.
98 Id.
25
base of mobile phone users 99 and a Spotify rival in music streaming, 100 suggests that
Spotify aims to require Apple to follow Google in permitting Spotify to evade revenue-
evade payment, but not while maintaining a seamless app experience for its users, who
Google and Apple, however, are not in the same position when it comes to Spotify.
Google provides extensive paid services to Spotify—e.g. Google Analytics 360, Google
Optimize 360, Google Ads—and therefore benefits as Spotify’s user base grows. 101 Apple,
by contrast, earns no significant revenue from Spotify apart from commissions on in-app
purchases. Google also obtains user information from Spotify, 102 and because Google’s
revenue comes primarily from targeted advertising, such information is of special value
to it. Apple, whose revenue comes primarily from hardware sales, gets little benefit from
the information it acquires in distributing apps. Apple also offers a value proposition
different from Google’s. Because of its more closed operating system and the extra vet-
ting it applies—at considerable cost—to app submissions, the apps in its store are
99 Although there are other distributors of Android apps, Google Play is by far the largest outlet for such
apps in the western world. As a practical matter, Spotify likely has to distribute through Google Play in
order to reach most American and European Android users. See Elad Natanson, The “Other” Android App
Stores—A New Frontier for App Discovery, FORBES (Sept. 3, 2019), https://www.forbes.com/sites/eladnatan-
son/2019/09/03/the-other-android-app-stores-a-new-frontier-for-app-discovery/#bce21bb6774c (observing
that App Store and Google Play “dominate app distribution in the west”).
100 Google’s YouTube Music service competes with Spotify in digital music streaming. See STATISTA, SHARE
https://www.thinkwithgoogle.com/intl/en-gb/success-stories/how-spotify-increased-premium-subscrip-
tions-using-google-optimize-360/ detailing Spotify’s use of various Google services).
102 See SPOTIFY PRIVACY POLICY, https://www.spotify.com/us/legal/privacy-policy/?version=1.0.0-GB (last
visited Oct. 01, 2020) (disclosing that Spotify “may share information with advertising partners” for pur-
poses of tailoring ads); GOOGLE, HOW GOOGLE USES INFORMATION FROM SITES OR APPS THAT USE OUR
SERVICES, https://policies.google.com/technologies/partner-sites?hl=en-US (last visited Oct. 01, 2020) (de-
tailing that apps using Google services provide information to Google that may then be used to personalize
ads).
26
perceived to be of higher quality, and its certification is therefore more valuable for de-
velopers.103 At the same time, Apple’s relentless efforts to upgrade the user experience
have attracted devoted, high-end users who are more likely to spend on apps. 104 The App
Store thus offers a particularly attractive distribution outlet to Spotify and other app de-
velopers.
In the end, Spotify is attempting to use antitrust to force Apple to act like Google
even though (1) Spotify confers benefits on Google that are not available to Apple; (2) the
certification service Apple provides is more valuable, and costlier to produce, than
Google’s; and (3) the user base Apple has cultivated is more desirable than that offered
by Google. In seeking to use antitrust law to extract greater value than it could obtain in
Whereas Spotify has directed its campaign against Apple alone, another promi-
nent app developer is attempting to use antitrust to extract greater surplus from transac-
tions on Google’s platform as well. In August 2020, Epic Games, producer of the popular
103 See Hildebrand, supra note 73; Poluliakh & Osadchiy, supra note 72.
104 See Felix Richter, Apple Users More Willing to Pay for Apps, STATISTA (July 6, 2018), https://www.sta-
tista.com/chart/14590/app-downloads-and-consumer-spend-by-platform/; Brown, supra note 83;
SENSORTOWER, supra note 84.
105 To be fair, Spotify is not alone in seeking to free-ride off Apple’s efforts to create a high-quality app
ecosystem. In addition to Epic Games, discussed next in the text, smaller app developers have also com-
plained of Apple’s insistence on compensation. See, e.g., Kif Leswing, Why Apple’s App Store is Under Fire,
CNBC (June 18, 2020), https://www.cnbc.com/2020/06/18/apple-app-store-faces-complaints-from-base-
camp-others-eu-probe.html (discussing smaller app developers’ complaints about App Store policies). The
president of Microsoft has also called for antitrust authorities to investigate Apple’s policies. See Dina Bass
& Mark Gurman, Microsoft Says Antitrust Bodies Need to Review Apple App Store, BLOOMBERG (June 18, 2020),
https://www.bloomberg.com/news/articles/2020-06-18/microsoft-says-antitrust-regulators-need-to-re-
view-app-stores. Microsoft, which operates its own app store and takes a similar—albeit lower—revenue
share from third-party app developers, see Jonny Caldwell, Microsoft Quietly Removes Pledge to Share 95% of
App Revenue on the Microsoft Store, ONMSFT (Jan. 16, 2020), https://www.onmsft.com/news/microsoft-qui-
etly-removes-pledge-to-share-95-of-app-revenue-on-the-microsoft-store (observing that Microsoft’s stand-
ard revenue share is 15%), distributes numerous apps through Apple’s App Store. See APPLE, APP STORE
PREVIEW, MICROSOFT CORPORATION, https://apps.apple.com/us/developer/microsoft-corpora-
tion/id298856275 (last visited Oct. 01, 2020) (cataloguing Microsoft apps available in the App Store).
27
Fortnite video game, filed antitrust suits challenging the app store policies of both com-
panies. 106 Epic complains that the companies restrict app distribution to their own
stores—Apple, by expressly disallowing iOS app distribution outside the App Store 107
and Google, by creating practical impediments to other avenues for Android app distri-
bution. 108 Epic also challenges each platform’s requirement that in-app purchases be pro-
cessed through the platform’s proprietary payment system (In-App Purchase for iOS;
Google Play Billing for Android).109 These restraints, Epic asserts, enable Apple and
Google to collect a supracompetitive revenue share on most in-app digital purchases, in-
cluding those that Fortnite players make when they buy access to such game features as
digital “skins” for their on-screen avatars and “battle passes” that allow them to partici-
While the Epic lawsuits are in their early days and may look different after discov-
ery, the company’s claims currently appear to be weak under United States antitrust law.
As an initial matter, the policies Epic complains of are not the product of market power;
they were put in place when Apple and Google had minuscule shares of the mobile op-
erating systems market. 110 The fact that the two companies now earn substantial revenues
from their app stores because their mobile operating systems have enjoyed success is not
illegal under U.S. law. As Judge Learned Hand famously explained, “The successful
106 Complaint for Injunctive Relief, Epic Games, Inc. v. Apple Inc., No. 3:20-cv-05640-YGR (N.D. Cal. Aug.
13, 2020) [hereinafter Epic v. Apple Complaint]; Complaint for Injunctive Relief, Epic Games, Inc. v. Google
LLC, et al., No. 5:20-cv-05671-NC (N.D. Cal. Aug. 13, 2020) [hereinafter Epic v. Google Complaint].
107 Epic v. Apple Complaint, supra note 106, at ¶¶ 64-81.
108 For example, Google forbids the distribution of other app stores through Google Play, and it allegedly
discourages app distribution through Internet websites (so-called “sideloading”) by requiring Android us-
ers to click through “dire warnings” about the security risks posed by apps that have not been vetted by
Google Play. Epic v. Google Complaint, supra note 106, at ¶¶ 26-105.
109 See Epic v. Apple Complaint, supra note 106, at ¶¶ 128-38; Epic v. Google Complaint, supra note 106, at
¶¶ 125-30.
110 See Erick Schonfeld, Smartphone Sales Up 24 Percent, iPhone's Share Nearly Doubled Last Year (Gartner),
28
competitor, having been urged to compete, must not be turned upon when he wins.” 111
There are also obvious business justifications for the complained of policies. The
first policy—placing restrictions (or, in Google’s case, warnings) on outside app distribu-
tion—helps prevent harm to users’ mobile software and hardware and builds user trust
in the mobile ecosystem. Because Apple’s primary revenue from mobile systems comes
from sales of hardware with preinstalled software (iPhones with iOS), it has a particularly
strong interest in ensuring perfect performance of the units it sells, and it therefore exerts
more control over app distribution than does Google, which earns substantial revenue
off mobile search. But Google, too, has an interest in protecting the Android brand by
quiring use of the platform’s own payment system for in-app purchases—enables each
platform to receive compensation for the tremendous value it provides to app developers
by assembling an installed base of users and certifying and marketing the developers’
apps. Producing such value is costly, which helps explain why nearly all app stores de-
mand a cut of developers’ revenues. 113 Indeed, the fees charged by Apple and Google
The most significant deficiency in Epic’s antitrust claims is that the complained of
policies in no way enhance the market power Apple and Google possess. Regardless of
111 United States v. Aluminum Co. of America, 148 F.2d 416, 430 (2d Cir. 1945).
112 Verizon Communications Inc. v. Law Offices of Curtis V. Trinko LLP, 540 U.S. 398, 407 (2004).
113 See ANALYSIS GROUP, APPLE’S APP STORE AND OTHER DIGITAL MARKETPLACES: A COMPARISON OF
29
their app store policies, Apple and Google control their operating systems and may thus
determine which, if any, third-party apps may operate on their platforms. They may
license their operating systems on the condition that users not install apps that have not
been approved, and create adverse consequences—both technological and legal—for do-
ing so. Their control over access to their operating systems—a power they will possess
whatever their app store policies are—enables them to extract a sizable share of the sur-
plus created when an app developer transacts with a licensee of the operating system.
They currently extract such surplus by taking a cut of the app developer’s revenue, and
the app developer often responds by charging higher prices to users. If, however, Apple
and Google were not able to capture surplus from app developers and users in that fash-
ion, they could easily do so in another manner. They could, for example, charge app
developers for access to critical application programming interfaces (APIs) or for the right
to be included on some list of approved apps that could bypass any technological barrier
to operability and any restrictions in the licenses held by users of the operating systems. 115
Because the policies Epic complains of do not create market power that would not other-
wise exist, a court is unlikely to conclude that they violate United States antitrust law.
Victory in a court of law, though, is not likely Epic’s goal. The circumstances sur-
rounding the filing of Epic’s lawsuits suggest that the company is primarily pursuing
victory in the court of public opinion. Epic’s lawsuits were part of a tightly orchestrated
publicity campaign. On August 13, 2020, the company breached its contracts with Apple
and Google by submitting app updates that contained obscure code allowing users to
bypass the platforms’ in-app purchasing systems and thereby avoid revenue sharing. 116
115 See Dirk Auer, The Epic Flaws of Epic’s Antitrust Gambit, TRUTH ON THE MARKET (Aug. 27, 2020),
https://truthonthemarket.com/2020/08/27/the-epic-flaws-of-epics-antitrust-gambit/. As Auer has observed,
Epic has conceded that Apple could cut off its access to crucial development tools.
116 See Nick Statt, Apple Just Kicked Fortnite off the App Store, THE VERGE (Aug. 13, 2020), https://www.thev-
erge.com/2020/8/13/21366438/apple-fortnite-ios-app-store-violations-epic-payments.
30
Apple and Google responded by exercising their contractual rights to remove the non-
conforming apps from their stores.117 Within hours of the removal, Epic filed pre-drafted
complaints—one 62 pages in length, the other 60—against the two companies. Epic then
peppered social media with a sleek video mirroring the iconic television commercial Ap-
ple released in connection with its 1984 debut of the Macintosh home computer but re-
Apple itself. 118 Epic’s clever commercial directed viewers to a Twitter hashtag,
#FreeFortnite, and the company soon hosted a “#FreeFortniteCup” to draw further atten-
tion to its cause. 119 A few days after filing its complaints and commencing its publicity
blitz, Epic generated further publicity by filing 197 pages in a pre-packaged motion for
temporary injunctive relief. 120 Of course, no such “emergency” relief would have been
needed had Epic filed its antitrust claims without first breaching its agreements. Epic’s
dramatic removal from the app stores, however, drew attention to its cause and upped
the public pressure on Apple and Google to make changes to their app store policies.
The question remains as to why Epic would launch a lawsuit-based publicity cam-
paign to induce Apple and Google to alter their app store policies when any changes
would not reduce the platforms’ market power but would simply induce them to extract
117 Id.
118 Id.
119 See The Fortnite Team, Join the Battle and Play in the #FreeFortnite Cup on August 23 (Aug. 20, 2020),
https://www.epicgames.com/fortnite/en-US/news/freefortnite-cup-on-august-23-2020.
120 Epic Games, Inc.’s Notice of Motion and Motion for Temporary Restraining Order and Order to Show
Cause Why a Preliminary Injunction Should Not Issue and Memorandum of Points and Authorities in
Support Thereof, No. 3:20-CV-05640-EMC (N.D. Cal. Aug. 17, 2020), https://cdn2.unrealengine.com/epic-
v-apple-8-17-20-768927327.pdf.
31
developers like Epic who no longer require any curation to garner the necessary trust from
consumers and who earn a large share of their revenue from in-app purchases. 121
Under the current system, paid apps that achieve great success effectively subsi-
dize upstarts, niche apps, and apps that are advertiser-supported. Epic may not like that
outcome, but the system has the advantage of getting developers of new and small apps
on board, expanding the offerings in each platform’s app store, and thereby building the
installed base of users from which Epic and other app developers benefit. As Auer fur-
ther observes, the current system, by facilitating a 30% revenue share from paid apps,
also reduces the incentive for Apple and Google to imitate successful apps and may
thereby operate as a “soft commitment not to expropriate developers, thus leaving them
Apple and Google have many ways to monetize control over their mobile operat-
ing systems. Given the intense competition between the two platforms, each has an in-
third-party apps so as to grow their user bases. Epic’s legally deficient lawsuits represent
an effort to put public pressure on Apple and Google to revamp their app store policies,
not in a manner that would reduce their market power—any such power would persist
even absent the complained of policies—but in a way that would advantage Epic at the
expense of other app developers and the mobile app ecosystem itself. In short, Epic is
engaged in rent-seeking.
firm may extract rents by inducing the government to raise its rivals’ costs and thereby
give it a competitive advantage. Rivals with higher costs must charge higher prices,
which permits a cost-advantaged firm either to follow their lead and earn higher margins
32
or to grow its market share by underpricing them. In an extreme case, higher-cost firms
may lose so much business that they are driven from the market altogether. In all these
scenarios, competition is weakened, and consumers suffer. We consider here several re-
cent examples of firms’ seeking to exploit government power to hobble their digital ri-
vals.
1. Google, Facebook, and the Regulation of Privacy, Artificial Intelligence, and Content
Moderation
adopted in 2016 and effective as of May 2018, endeavors to give individuals control over
the information about them that is collected or processed by digital firms. 123 GDPR re-
quires digital firms that process personal data to implement technical and organizational
measures to comply with certain data protection principles. 124 Business and data collec-
tion processes must be designed in light of those principles, and information systems
must use the highest possible privacy settings by default. 125 A user’s personal data may
not be processed unless the user has expressly consented or the processing occurs under
five other lawful bases: contract, public task, vital interest, legitimate interest, or legal
requirement. 126 Where information processing is based on consent, the data subject may
revoke it at any time. 127 Firms must clearly disclose any data collection, declare the legal
basis and purpose for processing the data, and state the length of time the data will be
retained and whether it will be shared with third parties. 128 Data subjects may demand a
portable copy of their data presented in a common format, 129 and they have the right to
33
have their data erased under certain circumstances. 130 Any business whose core activity
consists of regular or systematic processing of personal data must employ a data protec-
tion officer (DPO) to ensure compliance with GDPR, 131 and any data breach must be re-
Given the high cost of complying with these extensive requirements, one might
expect that GDPR would have impaired the European business of Google and Facebook,
both of which process vast troves of personal data. In actuality, GDPR has been a boon
to the two companies. An early study compared the tracking reach of digital advertising
firms (“ad tech vendors”) from one month before GDPR’s implementation to one month
after. 133 It found, unsurprisingly, that web tracking had decreased in the EU during the
period: trackers per page fell by 3.4% in the EU even as they grew 8.9% in the United
States. 134 But the bulk of the loss in web-tracking, which is crucial for targeted advertis-
ing, was suffered by smaller ad tech firms. Whereas the website tracking reach of the top
fifty ad tech firms besides Google and Facebook fell by 20%, and the tracking opportuni-
ties of “especially small” ad tech firms—those ranked 100 to 150—fell by around 32%,
Facebook’s website reach fell by only 6.66%, and Google’s actually grew slightly (by
0.93%). 135 The study authors thus concluded that “smaller advertisers lose” and that
In June 2019, the Wall Street Journal reported on the first full year of GDPR and
confirmed that the law appears to have benefited Google and Facebook, with both com-
133 See Bjorn Greif, Study: Google Is the Biggest Beneficiary of the GDPR, CLIQZ (Oct. 10, 2018),
https://cliqz.com/en/magazine/study-google-is-the-biggest-beneficiary-of-the-gdpr.
134 Id.
135 Id.
136 Id.
34
implementation. 137 Industry experts interviewed by the Journal suggested two reasons
for the relative good fortune of Google and Facebook under GDPR. First, the companies
have much greater resources for compliance, and firms prefer to concentrate their ad
budgets with companies whom they trust not to violate the rules. 138 In addition, because
GDPR makes it harder for third-parties to collect the personal information that is so val-
uable for targeting ads, it benefits digital firms that have direct relationships with users
and can more easily procure consent to use their data. 139 With their many user-facing
services that connect them directly to data subjects, Google and Facebook are far less re-
liant on third-party data. In fact, Google recently announced its intention to phase-out
support for third-party cookies—tiny data files that enable advertisers to track website
visitors across the Internet—in its market-leading Chrome browser. 140 That move, made
in the name of furthering user privacy, will greatly benefit Google and Facebook vis-à-vis
their ad tech rivals that lack direct user contact and therefore depend on third-party
data. 141
In its recent report Online Platforms and Digital Advertising, 142 the United Kingdom’s
Competition and Markets Authority (CMA) suggested that GDPR may have provoked
137 Nick Kostov & Sam Schechner, GDPR Has Been a Boon for Google and Facebook, WALL ST. J. (Jun 17, 2019),
https://www.wsj.com/articles/gdpr-has-been-a-boon-for-google-and-facebook-11560789219). While Eu-
rope’s digital advertising market grew by 14% in 2018, Facebook’s revenue from ads shown in Europe
increased by 40%, and “Google’s revenue in Europe, the Middle East and Africa—the vast majority of
which comes from advertising—rose 20%.” Id.
138 Id.
139 Id.
140 See Building a More Private Web: A Path Towards Making Third Party Cookies Obsolete, CHROMIUM BLOG (Jan.
2020), https://www.bloombergquint.com/businessweek/google-s-cookie-overhaul-could-reshape-the-digi-
tal-ad-industry); Ariel Bogle, Google Wants to Kill Third-party Cookies. Here's Why That Could Be Messy, ABC
SCIENCE (Jan. 20, 2020), https://www.abc.net.au/news/science/2020-01-21/google-to-kill-third-party-cook-
ies-privacy-and-competition/11882718.
142 U.K. COMPETITION AND MARKETS AUTHORITY, ONLINE PLATFORMS AND DIGITAL ADVERTISING: MARKET
35
Google’s decision. The CMA acknowledged “hear[ing] concerns that large platforms use
data protection regulations such as [GDPR] as a justification for restricting access to val-
uable data for third parties, while retaining it for use within their ecosystems, thereby
consolidating their data advantage and entrenching their market power.” 143 It further
observed that Google and Facebook “have an incentive to interpret data protection regu-
lation in a way that entrenches their own competitive advantage, including by denying
third parties access to data that is necessary for targeting, attribution, verification and fee
or price assessment while preserving their right to use this data within their walled gar-
dens.” 144 It explicitly highlighted “Google’s recent announcement that it was phasing out
support for third-party cookies on the Chrome browser, restricting publishers’ ability to
Given the competitive benefit GDPR has conferred on them, Google and Facebook
have warmed to the sorts of rules it imposes. The companies are now actively promoting
similar regulatory regimes that could entrench their dominance. For example, in a Janu-
ary 2020 op-ed in the Financial Times, the CEO of Google and its parent company Alpha-
bet argued that governments should impose broad artificial intelligence (AI) regula-
tions. 146 Asserting that “[e]xisting rules such as Europe’s General Data Protection Regu-
lation can serve as a strong foundation,” he highlighted Google’s own AI principles and
actions as a model for government mandates. 147 Facebook has similarly called for gov-
ernments to mandate actions it already takes. In meetings with EU regulators about dig-
ital platforms’ content moderation, Facebook recently proposed what tech writer Josh
146 Sundar Pichai, Why Google Thinks We Need to Regulate AI, FINANCIAL TIMES (Jan. 20, 2020),
https://www.ft.com/content/3467659a-386d-11ea-ac3c-f68c10993b04.
147 Id.
36
served, the sort of initiatives Google and Facebook are now proposing offer them a com-
petitive benefit they did not obtain from GDPR: “[I]n the case of GDPR, everyone had to
add new transparency and opt out features,” but implementation of these new proposals
would allow their proponents to “sail forward largely unperturbed while rivals and up-
148 Josh Constine, Facebook Asks for a Moat of Regulations it Already Meets, TECHCRUNCH (Feb. 17, 2020),
https://techcrunch.com/2020/02/17/regulate-facebook/. Constine catalogued Facebook’s proposed rules
and explained how the company already complies with each:
• User-friendly channels for reporting content – Every post and entity on Facebook can already
be flagged by users with an explanation of why. [See FACEBOOK, REPORTING ABUSE,
https://www.facebook.com/help/1753719584844061?helpref=page_content (last visited Oct.
01, 2020).]
• External oversight of policies or enforcement – Facebook is finalizing its independent Oversight
Board right now. [See Josh Constine, Toothless: Facebook proposes a weak Oversight Board,
TECHCRUNCH (Jan. 28, 2020), https://techcrunch.com/2020/01/28/under-consideration/]
• Periodic public reporting of enforcement data – Facebook publishes a twice-yearly report about
enforcement of its Community Standards. [See FACEBOOK, COMMUNITY STANDARDS
ENFORCEMENT REPORT, https://transparency.facebook.com/community-standards-enforce-
ment (last visited Oct. 01, 2020).]
• Publishing their content standards – Facebook publishes its standards and notes updates to
them. [See FACEBOOK, COMMUNITY STANDARDS: RECENT UPDATES, https://www.face-
book.com/communitystandards/recentupdates/ (last visited Oct. 01, 2020).]
• Consulting with stakeholders when making significant changes – Facebook consults a Safety
Advisory Board and will have its new Oversight Board. [See FACEBOOK, WHAT IS THE FACEBOOK
SAFETY ADVISORY BOARD AND WHAT DOES THIS BOARD DO?, https://www.face-
book.com/help/222332597793306 (last visited Oct. 01, 2020).]
• Creating a channel for users to appeal a company’s content removal decisions – Facebook’s
Oversight Board will review content removal appeals. [See Josh Constine, Facebook Will Pass Off
Content Policy Appeals to a New Independent Oversight Body, TECHCRUNCH (Nov. 15, 2018)
https://techcrunch.com/2018/11/15/facebook-oversight-body/.]
• Incentives to meet specific targets such as keeping the prevalence of violating content below
some agreed threshold – Facebook already touts how 99% of child nudity content and 80% of
hate speech removed was detected proactively, and that it deletes 99% of ISIS and Al Qaeda
content. [See Guy Rosen, Community Standards Enforcement Report, November 2019 Edition,
FACEBOOK NEWSROOM (Nov. 13, 2019), https://about.fb.com/news/2019/11/community-stand-
ards-enforcement-report-nov-2019/.]
See Constine, Facebook Asks for a Moat of Regulations it Already Meets, supra.
149 Constine, supra note 148.
37
Section 230 of the Communications Decency Act (CDA) provides significant legal
protection for digital platforms that allow users to post public content. Paragraph (c)(1)
of the provision states that “[n]o provider or user of an interactive computer service shall
mation content provider.” 151 The Act includes some exceptions to that rule, such as for
copyrighted material, 152 but it generally exempts digital platforms from liability based on
user-posted content on their sites. Because it encourages digital platforms to allow user
interaction—a key feature of what most people think of as “the Internet” as distinct from
dubbed Section 230(c)(1) “The Twenty-Six Words that Created the Internet.”153
A second key provision of Section 230 is paragraph (c)(2), which provides that
“[n]o provider or user of an interactive computer service shall be held liable on account
of … any action voluntarily taken in good faith to restrict access to or availability of ma-
terial that the provider or user considers to be obscene, lewd, lascivious, filthy, exces-
sively violent, harassing, or otherwise objectionable, whether or not such material is con-
stitutionally protected….” 154 This provision allows digital platforms to moderate content
posted by users without fear of liability to either the content generator or third-parties,
who might assert that the platform’s moderation of content renders it a publisher of the
content it allows. 155 As Kosseff has explained, these two provisions of Section 230, which
150 For further discussion of Section 230, see Berin Szoka & Ashkhen Kazaryan, Section 230: An Introduction
for Antitrust & Consumer Protection Practitioners, in THE GAI REPORT ON THE DIGITAL ECONOMY (2020).
151 47 U.S.C. § 230(c)(1).
153 JEFF KOSSEFF, THE TWENTY-SIX WORDS THAT CREATED THE INTERNET (2018).
155 Section 230(c)(2) was enacted to override a court decision holding that a digital platform operator’s mod-
eration of user-generated content could render it the publisher of the content it allowed. See KOSSEFF, supra
note 153, at 73 (discussing Congress’s intention to reverse the effect of Stratton Oakmont, Inc. v. Prodigy
Servs. Co., 1995 WL 323710 (May 24, 1995)).
38
were enacted before any of the currently prominent Internet platforms existed, allowed
the open Internet we now have to flourish: By freeing digital platforms from liability for
lating platforms from liability arising from good faith content moderation, paragraph
(c)(2) enabled platforms to prevent their sites from becoming flooded with objectionable
legislators to cut back on Section 230’s protections. 157 On first glance, these varied com-
panies appear to have little in common. They are alike, however, in that each (1) faces
little to no potential liability for user-generated content but (2) competes with firms that
do.
ware, artificial intelligence products, and business and cloud computing services—nor
vides a significant platform for user-generated content. But both companies compete in
cloud computing with Google and Amazon, each of which hosts a vast amount of user-
generated content (e.g. through Google’s YouTube and Amazon’s customer reviews and
user-streaming service Twitch). In 2017, IBM and Oracle teamed up to lobby for passage
of the “Stop Enabling Sex Traffickers Act” (SESTA) and the “Fight Online Sex Trafficking
Act” (FOSTA), which together amended Section 230 to allow digital platforms to be liable
under certain circumstances for user-generated content that facilitates sex-trafficking. 158
39
Sex-trafficking is a terrible thing, of course, and IBM and Oracle may have been seeking
both to reduce its incidence and to win public favor by taking a stance against it. But
thousands of other companies had those same motivations to lobby for SESTA/FOSTA,
yet did not do so. The extraordinary efforts of IBM and Oracle perhaps stemmed from
the fact that any weakening of Section 230’s protections increases the liability risk for
Google and Amazon and makes each a less formidable competitor. IBM and Oracle may
also have reasoned that once the door was opened to allowing platform liability for some
user-generated content, it would be easier to push for additional exceptions to the liability
shield. 159
IBM has recently made such a push. In a June 2019 publication entitled “A Preci-
sion Regulation Approach to Stopping Illegal Activities Online,” IBM heaped praise on
the legislators who have called for a reduction in Section 230’s protections and suggested
that the law be changed further.160 Under IBM’s proposal, a digital platform would be
protected from liability for user-generated content only if it could show that it took “rea-
sonable care” to prevent its platform from being used to further liability-creating con-
duct. 161 A result of this amendment would be that no legal claim against a platform based
on user-generated content could be dismissed at the complaint stage. The defendant plat-
form would always have to prove that it had taken reasonable steps to prevent the liabil-
ity-causing conduct (or that the conduct did not, in fact, create liability) before the claim
40
against it could be dismissed. Such a requirement would greatly increase platforms’ liti-
gation costs and would likely generate “strike suits”—meritless lawsuits filed for the pur-
platforms to install AI-based filtering technology such as IBM’s Watson Tone Analyzer,
an AI solution that assesses what content intends, not just what it says, and has been
lauded as “an important tool for sites trying to balance freedom of speech with protection
IBM and Oracle were not the only companies to lobby for SESTA/FOSTA. They
were joined by entertainment giants Walt Disney Company and 21st Century Fox, both of
which wrote to key U.S. Senators in support of the legislation.163 Those companies have
also lobbied to prevent the exportation of Section 230’s protections via trade agreements
and thereby preserve the possibility of liability for user-generated content outside the
U.S. 164
162 Terri Coles, How AI Can Help Filter the Worst of the Web, IT PRO TODAY (June 30, 2019), https://www.itpro-
today.com/data-analytics-and-data-management/how-ai-can-help-filter-worst-web. Oracle, too, offers
these sorts of solutions, which may explain why it stated in its letter in support of SESTA that “[a]ny start-
up has access to low cost and virtually unlimited computing power and to advanced analytics, artificial
intelligence and filtering software. That capability is also offered as a service in the cloud.” See Oracle Letter
to Sens. Portman and Blumenthal, supra note 158.
163 See Letter from Chip Smith, Executive Vice-President of Global Affairs, 21st Century Fox, to U.S. Senators
dios (Paramount Pictures Corp., Sony Pictures Entertainment Inc., Twentieth Century Fox Film Corp., Uni-
versal City Studios L.L.C., and Warner Bros. Entertainment Inc.) submitted extensive comments urging the
National Telecommunications and Information Administration to oppose inclusion of Section 230 protec-
tions in trade
agreemhttps://www.ntia.doc.gov/files/ntia/publications/180717_mpaa_response_to_ntia_internet_prioriti
es_inquiry_final.pdf. Disney lobbyists reportedly distributed a handout warning Congress that including
the provision in trade deals would make it difficult for Congress to change the law in a way that improved
the Internet. See McCabe, supra note 157.ents. See COMMENTS OF THE MOTION PICTURE ASSOCIATION OF
AMERICAN BEFORE THE NATIONAL TELECOMMUNICATIONS AND INFORMATION ADMINISTRATION (July 17,
2018),
41
such as Facebook, Google/YouTube, and Amazon/Twitch. Given that such content com-
petes for consumers’ attention against the entertainment offerings of major film studios,
anything that impairs the platforms hosting user-generated content tends to benefit tra-
platforms to do more to protect the film studios’ copyrights. On that front, Section 230 is
not directly relevant; the provision expressly has no effect on “any law pertaining to in-
tellectual property” and would therefore provide no protection to a digital platform ac-
cused of hosting copyrighted material. 165 When it comes to users’ posting of copyrighted
materials, the legal provision that protects platforms from liability is Section 512 of the
Digital Millennium Copyright Act (DMCA), which provides a safe harbor for digital plat-
forms that engage in good faith anti-piracy efforts and honor takedown notices from the
music and film industries. 166 Ginning up opposition to Section 230, however, is likely to
be an effective strategy for cutting back on Section 512 of the DMCA. 167 Because the two
provisions provide similar protections for digital platforms, policymakers tend to view
them—and would likely deal with them—as a package. 168 And since Section 230 insulates
platforms from liability for a broader scope of user-generated content, it is easier to find
short, there are more “Baptists” to assist with a challenge to Section 230, which can then
165 47 U.S.C. § 230(e)(2) (“Nothing in this section shall be construed to limit or expand any law pertaining
to intellectual property.”).
166 17 U.S.C. § 512.
167 See Brendan Bordelon, Copyright Liability Emerges as Latest Threat to Big Tech’s Legal Shield, NATIONAL
42
Hotel chains have also joined the opposition to Section 230. In 2016, the American
Hotel and Lodging Association (AHLA), a trade group that includes Marriott Interna-
tional, Hilton Worldwide, and Hyatt Hotels, reported to its members on a detailed plan
to impair the business of Internet-based short-term home rental platforms like VRBO (Va-
cation Rentals By Owner), HomeAway, and, most prominently, Airbnb. 169 By increasing
the lodging options for travelers, these companies reduce hotel chains’ ability to charge
high rates, particularly in times of peak demand. In AHLA’s private report, a copy of
which was obtained by the New York Times, the Association touted its successes in lobby-
ing three U.S. Senators to request an FTC investigation into the short-term rental industry
and in procuring a number of state and local ordinances restricting short-term rentals by
property owners. 170 It also announced its plans to seek a weakening of Section 230’s pro-
tections by, among other things, “[e]ngaging the copyright holder community which has
The hotel group’s attack on Section 230 is an effort to saddle Airbnb and similar
sites with liability for property owners’ violations of local ordinances regulating short-
term rentals. As AHLA members know, it would be extremely costly for home-sharing
sites to assure that posting property owners are complying with thousands of local ordi-
nances. Eliminating Section 230’s protections would therefore increase hotel competitors’
compliance costs as well as their likely liability. Notably, the hotel chains’ lobbying ef-
forts appear to be paying off: In September 2019, U.S. Representative Ed Case (D-HI) in-
troduced the Protecting Local Authority and Neighborhoods (PLAN) Act, which would
amend Section 230 to permit civil actions against Airbnb and other rental sites based on
169 See Katie Benner, Inside the Hotel Industry’s Plan to Combat Airbnb, N.Y. TIMES (Apr. 16, 2017),
https://www.nytimes.com/2017/04/16/technology/inside-the-hotel-industrys-plan-to-combat-
airbnb.html?smid=tw-share&_r=0).
170 See id.
171 See Report, American Hotel and Lodging Association, The Hotel Industry’s Plans to Combat Airbnb
(2016), https://www.nytimes.com/interactive/2017/04/16/technology/document-hotel-industry-plans-to-
combat-airbnb-excerpt.html.
43
AHLA. 173
3. Lawsuit Instigation
the legal protections on which they may rely, firms may hobble their competitors by in-
citing legal action against them. In recent years, firms competing in digital markets have
taken this tack by creating and funding groups that purport to represent the public inter-
est but are really focused on agitating for lawsuits against group members’ competitors.
Google’s acquisition of travel software firm ITA, FairSearch’s original members were
crosoft, whose Bing search engine is Google’s most prominent competitor in general
search. 176 FairSearch then began a relentless campaign to encourage antitrust enforce-
ment against Google. In the colorful words of one commentator, “FairSearch can be most
where they’ve started organically, originate them where they haven’t, and generally
172 See Protecting Local Authority and Neighborhoods Act, H.R. 4232, 116th Cong. (2019-2020)
https://www.congress.gov/bill/116th-congress/house-bill/4232/text.
173 See James Prichard, Ex-Congressman Joins Board of Lodging Industry Lobbying Group, PACIFIC BUSINESS
175 See Vlad Savov, What is FairSearch and Why Does It Hate Google So Much?, THE VERGE (Apr. 12, 2013),
https://www.theverge.com/2013/4/12/4216026/who-is-fairsearch.
176 See Greg Sterling, Microsoft Joins FairSearch Group Opposing Google-ITA Acquisition, SEARCH ENGINE LAND
44
disseminate negativity toward the Google brand. Think of it as a PR firm working to de-
complaint against Google for tying its mobile search and browser technologies to its pop-
ular Android app store, Google Play, and for the purportedly predatory act of licensing
Android at below-cost rates. 178 That effort paid off. In July 2018, the European Commis-
sion fined Google €4.34 billion for, among other things, “requir[ing] manufacturers to
pre-install the Google Search app and browser app (Chrome), as a condition for licensing
Google’s app store (the Play Store).” 179 By operating under the guise of FairSearch, Mi-
crosoft was able to obscure its role in instigating an action against its rival for behavior
Microsoft withdrew from FairSearch in December 2015, 180 but FairSearch contin-
ues to agitate for legal action against Google. 181 In doing so, FairSearch claims to repre-
sent “a group of businesses and organizations united to promote economic growth, in-
novation and choice across the Internet ecosystem.” 182 It lists at least nine companies as
members. 183 According to its official filings with Belgian authorities, however, FairSearch
is now controlled entirely by executives from two companies: Oracle and Napsers. 184
tices Regarding Android Mobile Devices to Strengthen Dominance of Google’s Search Engine (July 18, 2018),
https://ec.europa.eu/commission/presscorner/detail/en/IP_18_4581.
180 See Mark Bergen, Microsoft Quietly Retreats From FairSearch, Watchdog Behind Google Antitrust Cases, VOX
183 Id.
45
None of its other members has the right to vote on the group’s actions. 185 Neither Oracle
large stakes in the Chinese firm Tencent and a number of foreign technology companies—
has a direct stake in the mobile markets at issue in the EU’s Android investigation. The
two companies do, however, compete with Google in other markets and benefit when
panies, through FairSearch, are allowed access to otherwise confidential information re-
creating the false appearance of a grassroots campaign. 187 Another organization engaged
in astroturfing is the Free and Fair Markets Initiative (FFMI), which describes itself as “a
ing a fair, modern marketplace that works for all Americans.” 188 In addition to lobbying
for legislation restricting Amazon and investigations into its practices, FFMI has sent doz-
ens of letters and reports to members of Congress and their staffers, published numerous
opinion pieces in newspapers and online media outlets, and tweeted hundreds of social
media posts criticizing Amazon. 189 As FFMI has taken these actions, Amazon has faced
increasing scrutiny from the U.S. Department of Justice, the FTC, the EU, and numerous
185 Id.
186 Id. (“By becoming a complainant in an EU antitrust case, companies enjoy privileged access to confiden-
tial information linked to ongoing inquiries.”).
187 Id.
188 See FREE AND FAIR MARKETS INITIATIVE, ABOUT US, https://freeandfairmarketsinitiative.org/about-us/
46
payers,” 191 FFMI has publicly listed among its supporters a labor union, a Boston man-
agement professor and a California businessman. 192 The Wall Street Journal, however, re-
ported that FFMI’s asserted grassroots support is “not what it appear[s] to be.” 193 Ac-
FFMI’s true principals, according to the Journal, are several giant firms that stand to ben-
efit if Amazon falters: Simon Property Group, the largest shopping mall operator in the
U.S.; 195 Walmart Inc., the largest retailer in the U.S. 196 and the world’s second largest re-
tailer, after Amazon; 197 and Oracle, which competes with Amazon in cloud computing
and fiercely battled with it over a $10 billion Pentagon contract.198 Each of the three com-
panies was reportedly asked to pony up $250,000 to strategic communications firm Mar-
athon Strategies to support FFMI’s work. 199 Oracle admitted to doing so, and a source
confirmed to the Journal that Walmart had done so through an intermediary; Simon Prop-
191 See FREE AND FAIR MARKETS INITIATIVE, supra note 188.
192 See Grimaldi, supra note 189.
193 Id.
194 Id.
195 See NATIONAL REAL ESTATE INVESTOR, TOP 25 SHOPPING CENTER OWNERS (July 1, 2008),
https://www.nreionline.com/research-amp-data/top-25-shopping-center-owners.
196 See NATIONAL RETAIL FOUNDATION, TOP 100 RETAILERS 2019, https://nrf.com/resources/top-retailers/top-
https://www.forbes.com/sites/laurendebter/2019/05/15/worlds-largest-retailers-2019-amazon-walmart-
alibaba/#174ea6334171.
198 See Grimaldi, supra note 189.
199 Id.
200 Id.
47
As the name indicates, rent-seeking is “demand side” activity: one party is de-
manding (seeking) that the government’s right to coerce be exercised in a way that ena-
bles that party to extract rents. Given the Internet’s ability to disrupt markets and chal-
lenge traditional enterprises by, among other things, reducing transaction costs and in-
legacy firms deem it the easiest way to retain or enhance their profits. 202 But if the policies
sought are anticompetitive or socially harmful in some other way, would rent-seeking
succeed?
Public choice theory suggests why, at this moment in history, rent-seeking efforts
may be especially successful in digital markets. Recall that the fundamental premise of
public choice is that actors in the political arena are rational self-interest maximizers. At
least at the current time, political actors’ pursuit of self-interest seems likely to result in
implementation of many of the policies rent-seekers are demanding. To see why this is
so, consider the personal interests of government officials—and of those who hold them
Government officials’ personal interests may lead them to support even poorly
designed restrictions on digital platforms. Elected officials are likely to favor such re-
strictions because it allows them to take credit for “cracking down on Big Tech,” a cause
201 Ride-sharing services like Uber and Lyft and short-term rental businesses like Airbnb exemplify how
the Internet disrupts traditional businesses by allowing new providers to compete. Both services drasti-
cally reduce customers’ and providers’ costs of transacting and, through their reciprocal rating systems,
prevent the sorts of “lemons problems” that may arise when buyers and sellers possess different levels of
information about each other’s offerings. See LAMBERT, supra note 2, at 215–16.
202 See, e.g., Steve Blank, Strangling Innovation: Tesla Versus “Rent Seekers,” VENTUREBEAT (June 25, 2013),
48
that is popular with both progressives and conservatives, albeit for different reasons. 203
Google, Facebook, and Amazon because they believe the firms exploit workers and sup-
exercise excessive influence over government. 204 Conservatives support the same policies
but for a different reason: they believe the Big Tech companies are biased against them
and disdainful of their values. 205 It is hardly surprising, then, that the flurry of proposals
to rein in Big Tech includes entries from such ideologically diverse legislators as progres-
sive Senator Elizabeth Warren (D-MA) 206 and conservative Senator Josh Hawley (R-
MO). 207 Politically, Big Tech restrictions are a winner, even when—as in the case of
Enforcement officials’ self-interest may similarly lead them to favor both addi-
tional restrictions on and more enforcement action against the major technology plat-
forms. The greater the number and complexity of the restrictions it enforces, the greater
the prestige—and often the budget—of an enforcement agency. And given the political
203 Polling shows majorities of both Democrats and Republicans—68% and 67%, respectively—in favor
breaking up Big Tech companies to level the playing field for all content. Katharina Buchholz, Majority of
Americans in Favor of Breaking up Big Tech, STATISTA (Sept. 23, 2019), https://www.sta-
tista.com/chart/19440/survey-responses-breaking-up-big-tech/. See also Emily Stewart, Poll: Two-thirds of
Americans Want to Break up Companies like Amazon and Google, VOX (Sept. 18, 2019),
https://www.vox.com/policy-and-politics/2019/9/18/20870938/break-up-big-tech-google-facebook-ama-
zon-poll.
204 See, e.g., Kiran Stacey & Kadhim Shubber, Democratic Calls to Break up Big Tech Raise Fears in Silicon Valley,
FINANCIAL TIMES (Feb. 17, 2020); TIM WU, THE CURSE OF BIGNESS: ANTITRUST IN THE NEW GILDED AGE (2018).
205 See Sam Sabin, 3 in 5 GOP Voters Believe There’s Social Media Bias Against Conservatives, THE MORNING
dium.com/@teamwarren/heres-how-we-can-break-up-big-tech-9ad9e0da324c.
207 See, e.g., Ending Support for Internet Censorship Act, S. 1914, 116 Cong. (2019-2020), https://www.haw-
ley.senate.gov/sites/default/files/2019-06/Ending-Support-Internet-Censorship-Act-Bill-Text.pdf; Mary
Catherine Wellons, GOP Senator Introduces a Bill That Would Blow up Business Models for Facebook, YouTube
and Other Tech Giants, CNBC (June 19, 2019) https://www.cnbc.com/2019/06/18/sen-hawley-bill-would-re-
voke-cda-section-230-for-large-tech-companies.html.
49
salience of Big Tech, enforcement action against the leading technology platforms is es-
pecially likely to attract the attention of legislative appropriators. Enforcers with political
ambitions seem particularly likely to take on Big Tech, as doing so may boost their elec-
toral prospects. Empirical evidence shows that state attorneys general who actively par-
ticipate in multi-state litigation like the currently pending 48-state investigation of Google
are more likely to seek their state’s governorship or a seat in the U.S. Senate. 208 And across
against American technology firms have raised her profile and likely aided her quest to
When it comes to regulation of and enforcement against the major digital plat-
forms, the groups that normally rein in improvident decisions by government officials
may be ineffective. Members of the news media may harbor their own biases against big
technology platforms, which they perceive as having damaged the news business by
usurping consumer attention and advertising revenue. 210 Academics stand to gain favor-
by recent fawning press reports on entrepreneurial scholars pushing for action against
the largest technology platforms. 211 And most voters, as usual, remain rationally
208 See Colin Provost, When is AG Short for Aspiring Governor? Ambition and Policy Making Dynamics in the
Office of State Attorney General, 40 PUBLIUS 597 (2010), https://pdfs.seman-
ticscholar.org/38e1/e04f5d0357b408e1f892562ef5fd72dba776.pdf?_ga=2.125407625.249508186.1594661615-
1982199029.1594661615.
209 Mehreen Khan & Rochelle Toplensky, Vestager Discloses Ambition to Become Next EU Commission Chief,
2020) (observing that “most news executives in this country share a viewpoint on the platforms, having
seen them pull advertising dollars from the news business and spread misinformation at the expense of
professional journalism…”).
211 See, e.g., David Stretfeld, Amazon’s Antitrust Antagonist Has a Breakthrough Idea, N.Y. TIMES (Sept. 7, 2018),
50
ignorant; they are unlikely to learn how seemingly small and benign-sounding legal
requirements, various privacy mandates, and so forth, could have serious adverse conse-
quences. In short, public choice theory suggests that the current political environment is
This chapter has described public choice theory and the phenomenon of rent-seek-
ing and has documented instances of rent-seeking activity in digital markets. Under-
standing what rent-seeking is, what forms it takes, and why it so often succeeds is im-
portant for at least two reasons. First, such an understanding helps policymakers assess
whether a particular market failure warrants a regulatory fix. Because public choice con-
cerns are inevitable when government exercises its right to coerce, they should always be
and rent-seeking can assist in crafting regulatory interventions to secure as much social
welfare as possible. If individuals indeed act in the public sphere as they do in other
vention will be less likely than others to achieve their public-spirited aims. For example,
a sectoral regulator that has continual contact with its regulatees and extensive discretion
goal like “the public interest” is more susceptible to capture than a general, non-sector-
specific regulator with minimal regulatee contact, less discretion, and a narrower man-
date. Of course, the former sort of regulator will also have some advantages over the
51
latter, such as more industry expertise and a greater ability to address unforeseen circum-
stances. Subsequent chapters of this report consider the tradeoffs between different sorts
the losses likely to occur when rational self-interest maximizers exercise government’s
52