BIG DATA AND ANTITRUST ENFORCEMENT

data-driven competition problems and thus contribute to

1
Giovanni Pitruzzella the economic governance of digital market. This article
focuses on the implications that Big Data have for
Keywords: Antitrust; Big Data; Mergers; antitrust enforcement, i.e. on the potential application of
Privacy; Competition policy competition law to those (pathological) situations in
which Big Data might be used by a dominant company
to foreclose competitors or to exploit consumers, might be
Abstract: The interaction between information, a relevant factor in the assessment of mergers’
innovation and market outcomes is shaping the modern anticompetitive effects or might facilitate collusive
digital industries of the 21st century: the business models behavior.
of search engines, social networks, e-commerce websites
and marketplaces are highly reliant on the ability to
gather and process large amount of data. At the same
time, it is increasingly recognized that the use of Big 1. INTRODUCTION
Data by online platforms and intermediaries has far-
reaching consequences not only on economic activity, but The way information is created, distributed and
also on social and political mechanisms: technological used has always been a defining characteristics
developments affecting information flows affect the of societies and economies. Indeed, over
organization of markets as well as the nature of centuries, technological developments affecting
individual interactions and the functioning of the information flows have shaped the
political process. A set of complementary policy tools is organization of markets, the functioning of the
needed to define a comprehensive governance of online political process, as well as the nature of
markets that effectively protects competition, consumers individual interactions. In recent times, the
as well as individuals’ privacy and media pluralism. Internet has disrupted the relationship between
Not all sensitive issues raised by Big Data are also information, society, democracy and the
competition issues. However, because of the high degree economy in many ways. It has changed the way
of concentration that characterizes online markets, individuals interact and disclose personal facts
antitrust policy finds itself at the crossroad between Big and opinions on a variety of sensitive matters,
Data and the transformative effects that the Internet is the way news are produced and distributed, as
having on the economy and on society. Antitrust well as the way companies compete and
enforcement is well equipped and sufficiently flexible to markets are organized.
adapt its analytical tools to deal effectively with potential
The novel interaction between information,
innovation and market outcomes is shaping the
modern Internet industries of the 21st century:
1 Chairman of the Italian Competition Authority new search engines, social networks, e-
commerce websites and marketplaces have

DOI: 10.12870/iar-12609 77
emerged. But also traditional industries are in the exchange of online services. Also,
affected by the availability of an unprecedented targeted online advertising may provide more
amount of information and by the increased relevant information to consumers and may be
ability to store and process that information less intrusive than non-targeted ads. Moreover,
made possible by new ICT technologies. consumers can benefit from being exposed
from offers targeted to each consumer’s
In some industries, data have become an
preferences as this reduces search costs. At the
intangible assets at least as important as
same time, however, it has been observed that
copyright and patents. Firms can certainly
individuals may bear significant costs because
benefit from the ability to analyze detailed
of the lack of data protection in the online
information about their current or potential
world, of which they may well not be aware.
customers, which can improve their marketing
The state of uncertainty associated with privacy
capabilities and lower advertising costs, and
costs has been described with the “blank
makes it possible to propose targeted offers to
cheque" metaphor: as an individual reveals
consumers. Big Data also lies at the core of
private information to other parties, she is
digital companies’ innovative services and
signing a blank cheque; the cheque may never
business models and can contribute to the
come back to her, or may come back for an
efficiency of online markets2. At the same time,
indeterminably small or large price to pay4.
however, online industries where Big Data
plays a key role often show a high degree of The behavior of online platforms and
market concentration. The correlation between intermediaries is increasingly recognized as
Big Data and market structure is not spurious; having far-reaching consequences not only on
indeed, alongside and in combination with economic activity, but also on social and
other elements, in certain circumstances Big political mechanisms. Therefore, a set of
Data might be the source of market power and complementary policy tools is needed in order
might raise barriers to entry. to define a comprehensive governance of
online markets that effectively protects
Individuals can also benefit from sharing their
different fundamental values such as
personal data. Consumers might access
competition, consumer protection, privacy and
innovative services that could not be offered
pluralism. It is important to acknowledge that
without a certain degree of information
public policy should not aim a re-distributing
sharing3. Online personal information operates
benefits and costs related to data protection
as a currency, and in some cases the only one,
between individuals and companies, but at
formulating a coherent set of rules that allow
2See OECD, Data Driven Innovation for Growth and Well-
markets to work efficiently and fairly, fostering
Being: Interim Synthesis Report, October 2014.
3 For instance, live traffic information could not be
provided without consumers disclosing their individual
4See A. Acquisti, The Economics of Personal Data and the
location and movements. Economics of Privacy, OECD Background Paper, 2010.

DOI: 10.12870/iar-12609 78
the interests of individuals, the economy and is a relevant factor in the assessment of
society as a whole ensuring both innovation mergers’ anticompetitive effects or might
and the protection of fundamental rights. facilitate collusive behavior. The “antitrust
approach” - which looks at the relationship
Because of the very high degree of
between big data and competition under the
concentration that characterizes key online
established analytical framework to assess
markets, antitrust policy finds itself at the
companies’ anticompetitive conducts - suggests
crossroad between Big Data and the
that a thorough understanding of the economic
transformative effects that the Internet is
mechanisms at work in data-driven industries is
having on markets and society. However, it
necessary to develop sound policies and needs
should be clear that not all sensitive issues
to be articulated not as a matter of theory, but
raised by Big Data are also competition issues.
with reference to specific markets and services6.
For instance, it is possible that a tradeoff arises
between privacy and competition: when there
are more opportunities to collect and analyze
relevant user data that can be used as an input
2. BIG DATA AND ABUSE OF
to provide certain services, it is less likely that
DOMINANCE
Big Data becomes a barrier to entry and a
driver of market power. Similarly, while market The recognition of the importance that Big
efficiency might increase when companies are Data plays in several online industries has
able to recommend products/services in line raised the question of whether data may be
with individual consumers’ preferences, considered an essential input in some markets
segregation and polarization in societies might and therefore whether the “essential facility”
increase when individuals are exposed only to doctrine might also apply to data.
like-minded opinions and news5.
Holding a large amount of data is not a
Notwithstanding the more general public problem as such and it is unlikely to be
policy issues, this article specifically focuses on problem when other companies can get hold of
the implications that Big Data have for the same or similar data, by collecting it from
antitrust enforcement, i.e. on the potential their own users or even buying it in. In this
application of competition law to those perspective, it is clear that Big Data should not
(pathological) situations in which Big Data be thought of as an undifferentiated mass of
might be used by a dominant company to information, but needs to be considered with
foreclose competitors or to exploit consumers, specific reference to individual services and

5C.R. Sunstein, The law of group polarization, Journal of

Political Philosophy, 2002(10); C.R. Sunstein, Divided 6 See G. Pitruzzella, Big data, competition and privacy: a
Democracy in the Age of Social Media, Princeton University look from the antitrust perspective, Concorrenza e Mercato,
Press, 2017. vol. 23, 2016.

DOI: 10.12870/iar-12609 79
markets: data relevant for targeted online conditions must be verified: it must be put in
advertising might not be relevant for e- place by a dominant company; it must relate to
commerce websites. data that is objectively necessary to be able to
compete effectively on a (downstream) market;
Moreover, it is important to recognize that
it must be likely to lead to the elimination of
data, by itself, are often of low value. In fact,
effective competition on the downstream
value derives mainly from the information
market and to consumer harm8.
obtained through the analysis of data. Along
the value chain that comprises the collection, More generally, however, it is also possible that
storage, analysis and usage of Big Data, data-driven exclusionary conduct might take
different types of barriers might limit entry at place when a vertically integrated company uses
different stages7. But it is important that data acquired at one level to distort
companies’ can benefit from their investments competition in a related market.
in data-based activities that create value for A different set of potential concerns relates to
society. possible data-driven exploitative abuses of
The importance of the trade-off between static dominance. For instance, one possibility is that
and dynamic competition lies at the heart of Big Data is used to implement advanced forms
the notion of essential facility. In fact, it is clear of price-discrimination9. The concern is that, by
that the legal notion of essentiality goes well collecting comprehensive and detailed
beyond the mere recognition of the relevance information on individual users, companies
that Big Data may have in the competitive may be able to infer with a high level of detail
process. Even in those circumstances in which individual consumers’ willingness to pay for a
Big Data is an important source of competitive certain good/service and thus charge
advantage and a barrier to entry, antitrust law individualized prices. The impact that price
does not necessarily require companies to discrimination might have on consumer’s
supply the data they collect to their welfare depends upon the degree of
competitors. In fact, it is necessary to consider competition in the market. In fact, it is not
that an “obligation to supply” is a strong obvious that price discrimination by
disincentive to invest in those activities through oligopolistic firms would necessarily decrease
which big data is collected and analyzed that
might bring benefits to consumers in the forms
of innovative services.
8 See the Communication from the Commission: Guidance on its
enforcement priorities in applying Article 82 of the EC Treaty to
Therefore, for a refusal to supply to fall under abusive exclusionary conduct by dominant undertakings, OJ C 45,
Article 102 prohibition, several cumulative 24.2.2009 and the case law in Oscar Bronner (C-7/97,
judgement of 26 November 1998), IMS Health (C-
418/01, judgement of 29 April 2004) and Microsoft (T-
201/94, judgement of 17 September 2007).
7 Daniel L. Rubinfeld and Michal S. Gal, Access barriers 9 See Executive Office of the President of the United

to big data, Arizona Law Review, forthcoming. States , Big data and differential pricing, 2015.

DOI: 10.12870/iar-12609 80
consumer welfare in the same way that price and market power are those that have involved
discrimination by a monopolist would. In well-known digital companies that manage
addition, this is a type of conduct where the global search engines, social networks and
difference between efficiency and equity communications services12.
considerations becomes particularly relevant: Merger review is probably the area in which we
price discrimination may well improve shall more readily witness the development of
(allocative) efficiency by allowing companies to our understanding of the role of Big Data for
expand the size of the market, but may lead to
market power and assess the potential theories
outcomes that are perceived as unfair. of harm related to the combination and holding
However, it is not clear if and how antitrust of large sets of data. Notwithstanding the fact
enforcement should be guided by fairness that turnover-based thresholds might not allow
considerations that entail placing different competition authorities to scrutinize all mergers
weights on the welfare of different groups of where Big Data plays a central role, merger
consumers. control is also the area where, in the near
future, competition authorities might more
directly affect the evolution of digital markets.
3. BIG DATA AND MERGER REVIEW In general, considerations pertaining to Big
Data might be relevant in the analysis of both
As in all other industries, mergers and horizontal and vertical/conglomerate anticompetitive
acquisitions play a big role in shaping the effects of mergers, as the combination of the
structure of digital markets. In 2016, digital parties’ datasets post-merger may have an
transformation from cloud, mobile, social and impact on the merged entity's market power
big data analytics technologies resulted in high- and/or on barriers to entry/expansion in the
value merger deals10. In fact, M&A can be an market for actual or potential competitors. Any
important tool for corporations to keep pace such data combination could only be
with the disruption coming from digital implemented by the merged entity to the extent
technologies. Many mergers affect a deep layer it is allowed by applicable data protection rules
of the value chain: the number of deals with respect to the collection, processing,
targeting big-data technology reached over one
hundred in the second quarter of 201611.
However, the mergers that have fueled 12 See, for instance: GOOGLE/DOUBLECLICK,
attention on the relationship between Big Data Commission Decision of 11 March 2008, Case No.
COMP/M.4731; MICROSOFT/YAHOO!, Commission
Decision of 18 February 2010, Case No.
COMP/M.5727; INTEL/MCAFEE, Commission
Decision of 26 January 2011, Case No. COMP/M.5984;
10See EY, Global technology M&A report: April–June FACEBOOK/WHATSAPP, Commission Decision of 3
2016. October 2014, Case No. COMP/M.7217;
MICROSOFT/SKYPE, Commission Decision of 7
11 Ibid.
October 2011, Case No. COMP/M. 6281.

DOI: 10.12870/iar-12609 81
storage and usage of personal data. Moreover, conglomerate effects, the theory of harm
an analysis of the merger party’s ability (and normally revolves around possible foreclosure
incentives) to combine the different datasets effects that can arise when the merged parties
would have to be undertaken. Recently, on this have the ability and incentives to exclude rivals
issue, the European Commission has fined from a market to the detriment of consumers.
Facebook 110 million euros for providing Big Data might be relevant both directly, when
incorrect and misleading information on the the foreclosure strategy rests on access to data,
technical possibility of automatically matching and indirectly, when Big Data, and most
Facebook and WhatsApp users' identities13. In notably data-driven network effects, might act
any case, any theory of harm would have to be as a source of barrier to entry15.
considered in light of a range of factors such One final aspect refers to the analysis of the
as: (i) the scope of the relevant product market; impact of the merger on competition also in
(ii) market structure; (iii) the importance of data terms of privacy. It has been observed that if
for the supply of the specific service; (iv) the consumers value privacy as a desirable
different sources (including third-party) characteristic of a service, a reduction in
through which companies relevant privacy is analogue to a reduction in the quality
data/information can be collected. of the service provided16. In the
For instance, the analysis of Big Data is likely Facebook/WhatsApp merger, the Commission
to be particularly relevant in the assessment of had stated that any privacy-related concerns
horizontal mergers among companies active in flowing from the increased concentration of
online advertising, where Big Data is a key data as a result of the transaction do not fall
input in the provision of targeted ads. Insofar, within the scope of the EU competition law
in the cases that it has examined, the European rules but within the scope of the EU data
Commission has found that there will continue protection rules17. More recently, in its review
to be a large amount of internet user data that
are valuable for advertising purposes and that
are not within the merged entity’s exclusive 15 For instance, in the recent Microsoft/LinkedIn
control14. When mergers raise vertical and/or merger, the European Commission identified potential
competition problems deriving from the: (i) tying
between Windows PC operating system and LinkedIn;
and (ii) integration of LinkedIn features into
Office/Outlook and the degradation of interoperability
13 The Commission at the time of its decision had also
between the latter and third-party PSN services. The
carried out an 'even if' assessment that assumed user
theory of harm was also based on the analysis of the role
matching as a possibility, concluding that, even in this
that (data-driven) network effects have on barriers to
scenario, its conclusions as to the lack of anti-
entry.
competitive effects of the proposed transaction would
stand. 16 See OECD, Big Data: Bringing competition policy to the
digital era, October 2016.
14 See, for instance, FACEBOOK/WHATSAPP,
Commission Decision of 3 October 2014, Case No. 17FACEBOOK/WHATSAPP, Commission Decision of
COMP/M.7217. 3 October 2014, Case No. COMP/M.7217.

DOI: 10.12870/iar-12609 82
of the Microsoft/LinkedIn merger the rational companies would naturally exploit the
Commission has concluded that foreclosure same degree of price transparency to identify
effects in the market for professional social the best positioning of their offers in the
networks might negatively affect consumers market. Indeed, sellers can closely monitor –
because of the marginalization of competitors also through ad-hoc software that can be bought
that offer a greater degree of privacy protection off the shelves – (the evolution of) prices of
to users than LinkedIn (or make the entry of their competitors and easily define and
any such competitor more difficult)18. implement dynamic pricing strategies.
Therefore, privacy considerations appear to Empirical evidence shows that dynamic pricing
become increasingly relevant also in merger is extremely common20.
review. It has been observed that these features of
online market environment might foster the
ability of companies to reach, monitor and
enforce collusion21. For instance, reaching an
4. BIG DATA AND COLLUSION
agreement may be easier when companies use
The high degree of transparency that the same algorithm to define prices. Also, it is
characterizes online markets and the well known that sustainability of collusion
widespread use of pricing algorithms, which requires effective enforcement, i.e. detection
exploit the increasing ability to gather and and retaliation of deviant firms. The use of
process large amounts of online data, may have pricing algorithms might increase the likelihood
an impact on collusive behavior19. that cheating on the collusive “agreement” will
be detected as well as the speed of punishment.
Online price transparency increases both In fact, companies can monitor their rivals’
consumers’ and companies’ ability to compare prices and detect changes in real time. Also,
price information online. On the demand side, (strategic) dynamic interaction occurs more
price comparison tools and electronic quickly than in traditional markets, thereby
marketplaces allow consumers to quickly easing companies’ ability to implement harsh
compare prices of products across a large punishment strategies, making punishment
number of online sellers and my thus facilitate stronger, swifter and more certain.
consumer choice, leading firms to compete
more aggressively on prices. At the same time, The impact that online pricing algorithms
might have as a tool to sustain “traditional”
explicit collusion – by increasing transparency
18 MICROSOFT / LINKEDIN, Commission Decision
of 6 December 2014, Case No. COMP/M.8124.
19 See OECD, Algorithms and Collusion, May 2017; 20 See European Commission, Final report on the E-
Ezrachi, A. and M. E. Stucke, Virtual Competition: The
commerce Sector Inquiry, COM(2017) 229 final.
Promise and Perils of the Algorithm-Driven Economy, Harvard
University Press, 2016. 21 See OECD, Algorithms and Collusion, May 2017.

DOI: 10.12870/iar-12609 83
and thus the ability to police the agreement – 5. CONCLUSIONS
does not appear to raise entirely new issues for
competition enforcers. However, innovative The close interaction between information,
questions, both in terms of policy and actual innovation and market outcomes is shaping the
enforcement, might arise because of the impact modern digital industries of the 21st century:
that algorithms may have on tacit collusion. the business models of search engines, social
From a policy perspective, one question is networks, e-commerce websites and marketplaces
whether or not the scope of antitrust law are founded on operators’ ability to gather and
should be revised to take into consideration the process large amount of data. At the same time,
impact that these pricing algorithms have on it is increasingly recognized that the use of Big
the diffusion of tacit collusion. However, it is Data by online platforms and intermediaries
not clear yet if this phenomenon is common or has far-reaching consequences not only on
not, and/or if there are spillover effects on economic activity and consumer welfare, but
offline competition. It is also necessary to also on social and political mechanisms:
identify the scenarios in which the use of technological developments affecting information
pricing algorithms may raise the most serious flows shape the organization of markets as well
concerns. as the functioning of the political process and
In terms of enforcement, a key issue is how the the nature of individual interactions.
traditional notion of liability extends to the Antitrust finds itself at the crossroad between
development and use of pricing algorithms. For Big Data, market power and the transformative
instance, companies might unilaterally create effects that the Internet is having on markets
and use computer algorithms to achieve a given and society. Antitrust policy is well equipped
target, such as profit maximization and the and sufficiently flexible to adapt its analytical
machines, through self-learning and tools to deal effectively with data-related
experiment, determine independently the competition problems and to contribute to the
strategy to maximise profits, which might also economic governance of online markets. Well-
result in collusive behavior22. functioning markets are necessary to foster
innovation and progress (benefiting both
companies and consumers) and to benefit
consumers: to this end, merger review and
enforcement against abuses of dominance and
cartels are certainly essential.
At the same time, digital markets need a broad
set of rules to foster the overall interests of
society safeguarding individuals’ fundamental
22See Ezrachi, A. and M. E. Stucke, Virtual Competition: rights. Antitrust is an important component of
The Promise and Perils of the Algorithm-Driven Economy,
Harvard University Press, 2016. the wider set of policy instruments that can be
used to govern digital markets, which also

DOI: 10.12870/iar-12609 84
comprise consumer protection, data protection,
media pluralism. As the EU approach shows, it
is important that these tools are used in a
complementary way, pursuing common goals
that encompass the promotion of growth and
innovation as well as the welfare of individual
consumers and citizens’ fundamental rights.

DOI: 10.12870/iar-12609 85
 6. BIBLIOGRAPHY EY, Global technology M&A report: April–June
2016,
A. Acquisti, The Economics of Personal Data and http://www.ey.com/Publication/vwLUAssets
the Economics of Privacy, OECD Background /EY-global-technology-m-and-a-report-2q16-
Paper, 2010. final-look/%24FILE/EY-global-technology-
m-and-a-report-2q16-final-look.pdf, 2016 .
Court of Justice, C-7/97 Bronner, 26
November 1998. General Court, T-201/04 Microsoft, 17th
September 2007.
Court of Justice, C-418/01 IMS Health, 29
April 2004. A. Ezrachi, A. and M. E. Stucke, Virtual
Competition: The Promise and Perils of the Algorithm-
European Commission, Communication from the
Driven Economy, Harvard University Press, 2016.
Commission: Guidance on its enforcement priorities in
applying Article 82 of the EC Treaty to abusive OECD, Data Driven Innovation for Growth and
exclusionary conduct by dominant undertakings, OJ C Well-Being: Interim Synthesis Report, October 2014.
45, 24.2.2009. OECD, Big Data: Bringing competition policy to the
European Commission, Final report on the E- digital era, October 2016.
commerce Sector Inquiry, COM(2017) 229 final, OECD, Algorithms and Collusion, May 2017.
2017.
G. Pitruzzella, Big data, competition and
European Commission, privacy: a look from the antitrust perspective,
GOOGLE/DOUBLECLICK, Case No. Concorrenza e Mercato, vol. 23, 2016.
COMP/M.4731, 11 March 2008.
D. L. Rubinfeld and M. S. Gal, 2017, Access
European Commission, barriers to big data, Arizona Law Review,
MICROSOFT/YAHOO!, Case No. forthcoming.
COMP/M.5727, 18 February 2010.
C.R. Sunstein, The law of group polarization,
European Commission, INTEL/MCAFEE, Journal of Political Philosophy, 10, 2002.
Case No. COMP/M.5984, 26 January 2011.
C.R. Sunstein, Divided Democracy in the Age of
European Commission, Social Media, Princeton University Press, 2017.
FACEBOOK/WHATSAPP, Case No.
COMP/M.7217, 3 October 2014.
European Commission,
MICROSOFT/SKYPE, Case No. COMP/M.
6281, 7 October 2011.
Executive Office of the President of the United
States , Big data and differential pricing, 2015.

DOI: 10.12870/iar-12609 86