Digital conglomerates and killer acquisitions - A discussion of the competitive effects of start-up acquisitions by digital platforms
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Concurrences REVUE DES DROITS DE LA CONCURRENCE | COMPETITION LAW REVIEW Digital conglomerates and killer acquisitions – A discussion of the competitive effects of start‑up acquisitions by digital platforms Law & Economics l Concurrences N° 1-2020 l pp. 42-50 Tristan Lécuyer tristan.lecuyer@rbbecon.com Senior Associate RBB Economics, Paris
Law & Economics
Tristan Lécuyer
Digital
constitutes a violation of the publisher's rights and may be punished by up to 3 years imprisonment and up to a € 300 000 fine (Art. L. 335-2 Code de la Propriété Intellectuelle). Personal use of this document is authorised within the limits of Art. L 122-5 Code de la Propriété Intellectuelle and DRM protection.
tristan.lecuyer@rbbecon.com
Ce document est protégé au titre du droit d'auteur par les conventions internationales en vigueur et le Code de la propriété intellectuelle du 1er juillet 1992. Toute utilisation non autorisée constitue une contrefaçon, délit pénalement sanctionné jusqu'à 3 ans d'emprisonnement et 300 000 € d'amende (art.
L. 335-2 CPI). L’utilisation personnelle est strictement autorisée dans les limites de l’article L. 122 5 CPI et des mesures techniques de protection pouvant accompagner ce document. This document is protected by copyright laws and international copyright treaties. Non-authorised use of this document
Senior Associate
conglomerates and
RBB Economics, Paris
killer acquisitions –
A discussion of
the competitive
effects of start‑up
acquisitions by
ABSTRACT
Digital platforms are growingly suspected to
acquire promising start-up companies to “kill”
future competition (the “killer acquisition”
digital platforms
concern) and/or extend their domination
by acquiring complementary services
in neighbouring markets resulting
in marginalisation of rivals and higher barriers
to entry (the “digital conglomerate” concern).
This article assesses horizontal and
conglomerate effects raised by such mergers
and discusses the extent to which a change
in competition law in Europe is needed to deal
I. Introduction
with these cases. It concludes that, given
the high uncertainty regarding future
1. Digital platforms are under growing scrutiny. Commentators have been par-
competition, changes proposed by observers ticularly vocal about acquisitions of start-ups by digital platforms, whose
to the current EU merger control risk leading growing number is interpreted by some as a sign of underenforcement. A recent
to misconceived theories of harm and
reversing the burden of proof for digital
report prepared for the UK government by a digital competition expert panel,
mergers is likely to deprive customers headed by Professor Jason Furman (“the Furman report”), notes for instance
from procompetitive benefits and jeopardise that companies such as Amazon, Apple, Facebook, Google and Microsoft have
innovation in the digital sector.
engaged in close to 250 acquisitions globally in the last five years, none of which
Les plateformes numériques sont de plus have been prohibited.1 Concerns expressed with respect to these acquisitions are
en plus suspectées d’acquérir des start-ups mainly twofold.
prometteuses afin de “tuer” la concurrence
future (inquiétude liée aux “acquisitions
meurtrières”) et/ou d’étendre leur dominance 2. First, these acquisitions would be a way for digital platforms to “kill” future
via l’acquisition de services complémentaires competition by acquiring a promising start-up which might challenge their
dans des marchés voisins conduisant
à l’affaiblissement des concurrents et à incumbent position at some point in the future. European Commissioner
des barrières à l’entrée plus élevées Margrethe Vestager in particular expressed this concern on several occasions,
(inquiétude liée aux “conglomérats pointing to the need for “revisited theories of harm, so we can intervene in mergers
numériques”). Cet article étudie les effets
horizontaux et congloméraux soulevés
when the owners of ecosystems buy start-ups before they have a chance to grow.”2
par ces fusions et discute la mesure selon We refer to this concern as the “killer acquisition concern” in the remainder of
laquelle elles nécessitent un changement the paper.
du droit de la concurrence européen.
Il montre qu’étant donné l’inhérente
incertitude liée à la concurrence future, 3. Second, start-up acquisitions by digital platforms would serve the purpose of
les changements du contrôle des expanding the domination of the platform in neighbouring markets by margi-
concentrations européen préconisés
nalising rivals and raising barriers to entry through conglomerate effects. In a
par certains commentateurs risquent
de conduire à des théories du préjudice
inappropriées tandis que renverser le standard
de la preuve pour ces fusions risque 1 Furman et al. (2019), p. 91.
de priver les consommateurs de bénéfices
proconcurrentiels et de mettre en danger 2 Commissioner M. Vestager’s contribution to the conference “Shaping competition policy in the era of digitisa-
l’innovation dans le secteur numérique. tion” held in Brussels on January 17, 2019.
42 Concurrences N° 1-2020 I Law & Economics I Tristan Lécuyer I Digital conglomerates and killer acquisitions...recent public appearance, Andreas Mundt, president of concerns in the digital sector, even in situations where si-
constitutes a violation of the publisher's rights and may be punished by up to 3 years imprisonment and up to a € 300 000 fine (Art. L. 335-2 Code de la Propriété Intellectuelle). Personal use of this document is authorised within the limits of Art. L 122-5 Code de la Propriété Intellectuelle and DRM protection.
the German Bundeskartellamt, shared this concern by gnificant network effects exist, where data have a critical
Ce document est protégé au titre du droit d'auteur par les conventions internationales en vigueur et le Code de la propriété intellectuelle du 1er juillet 1992. Toute utilisation non autorisée constitue une contrefaçon, délit pénalement sanctionné jusqu'à 3 ans d'emprisonnement et 300 000 € d'amende (art.
L. 335-2 CPI). L’utilisation personnelle est strictement autorisée dans les limites de l’article L. 122 5 CPI et des mesures techniques de protection pouvant accompagner ce document. This document is protected by copyright laws and international copyright treaties. Non-authorised use of this document
stating that sometimes, even when platforms have only role and where the dynamics of platform envelopment
just entered a market, they “have already all the financial are at play.
strength to exert conglomerate market power, bringing
billions of clients and their data to that new market, rolling 8. We conclude from the above discussion that, given
it up, in quite a little time.”3 We refer to this concern as the the high uncertainty regarding future competition, the
“digital conglomerate concern” in the rest of the paper. proposed changes to the current EU merger control risk
leading to misconceived theories of harm and, impor-
4. These concerns must be taken seriously as they may tantly, reversing the burden of proof is likely to deprive
entail a profound change of merger control regimes in customers from procompetitive benefits and jeopardise
Europe. Some enforcers have recently taken positions in innovation in the digital sector.
favour of a change of competition law to better assess
the potential anticompetitive effects of such acquisitions.
Former EU’s competition chief economist, Professor
Tommaso Valletti, notably advocates a complete shift of II. A growing
suspicion of
the burden of proof and the use of “structural presump-
tions” to assess acquisitions by large digital platforms.
In essence, Valletti considers that acquisitions by large
digital platform should be presumed anticompetitive
unless the parties can prove that the merger would bring anticompetitive effects
significant efficiencies to consumers.4 These proposed 9. Acquisitions of start-up companies by digital platforms
changes would be a complete reversal of the current EU are increasingly suspected to have significant anticom-
merger control where the European Commission needs petitive effects. Several academic reports have recently
to prove anticompetitive effects to prohibit a merger. expressed concerns regarding such acquisitions which,
they argue, may significantly impede effective competi-
5. This article assesses the extent to which the current EU tion even when the two merging parties are not active in
merger control is well suited to deal with these mergers the same relevant market. These reports include studies
and whether a change in competition law is needed. It commissioned by competition authorities to formulate
contributes to this discussion by providing an economic recommendations on potential changes of competition
assessment of the concerns that start-up acquisitions by policy in the digital sector (e.g., the Crémer report for
digital platforms may raise. For each of the two most the European Commission and the Furman report for
commonly raised concerns, i.e., the “killer acquisition” the UK government) as well as some academic publi-
and “digital conglomerate” concerns, it explains the cations (e.g., the Stigler report from the Chicago Booth
economic mechanisms at stake and sets out the condi- School of Business and the article of Bourreau et al. on
tions under which these acquisitions may be deemed digital conglomerates). According to these reports, such
anticompetitive by analysing both potential horizontal mergers could harm consumers in two ways which we
and conglomerate effects. describe below.
6. With respect to horizontal effects, we show that
because tangible evidence on whether a start-up will
grow into a significant competitive force constraining
1. The “killer acquisition”
the digital platform may not be available, replacing the concern
current “balance of probability” approach by a “balance
of harm” approach as advocated by the Furman report 10. The authors of these reports argue that the acquisi-
is likely to raise significant practical difficulties and risk tion of a start-up company by a large digital platform
leading to some speculative theories of harm. may result in the “early elimination of a potential rival”
before it had a chance to grow into a significant compet-
7. With respect to conglomerate effects, we explain why itive force. The concern expressed here is that, absent the
conglomerate mergers are generally procompetitive, in transaction, a start-up with “a large and/or fast growing
particular in the digital sector where further benefits user base and a high future market potential” may have
can be passed on to consumers. Under certain circums- grown into a significant competitive constraint on the
tances, however, the expansion of a digital platform in digital platform in one or several markets where it is
a neighbouring market can raise competition issues active.5 As such, the merger would give the digital platform
through conglomerate effects undermining the ability further ability to increase prices compared to the coun-
of rivals to compete. We show that the current EU terfactual in which it would have faced direct competition
framework is nevertheless well suited to deal with these from the start-up. According to these reports, this could
happen even in situations where the start-up is not active
in the same relevant antitrust market as the acquirer at
the time of the merger. The authors consider indeed that
3 MLex, 12 July 2019, quoting A. Mundt’s contribution to the conference “Concurrenc-
es 15-Year Anniversary: Which Competition Policy for 2019–2024?” held in Brussels on
July 11, 2019.
4 Valletti (2018). 5 Crémer et al. (2019), pp. 110–122.
Concurrences N° 1-2020 I Law & Economics I Tristan Lécuyer I Digital conglomerates and killer acquisitions... 43start-up companies often innovate in the fringe of a plat- overcome the significant disadvantage of not being able
constitutes a violation of the publisher's rights and may be punished by up to 3 years imprisonment and up to a € 300 000 fine (Art. L. 335-2 Code de la Propriété Intellectuelle). Personal use of this document is authorised within the limits of Art. L 122-5 Code de la Propriété Intellectuelle and DRM protection.
form’s ecosystem by offering a different good that does to supply the two services within a single ecosystem.
Ce document est protégé au titre du droit d'auteur par les conventions internationales en vigueur et le Code de la propriété intellectuelle du 1er juillet 1992. Toute utilisation non autorisée constitue une contrefaçon, délit pénalement sanctionné jusqu'à 3 ans d'emprisonnement et 300 000 € d'amende (art.
L. 335-2 CPI). L’utilisation personnelle est strictement autorisée dans les limites de l’article L. 122 5 CPI et des mesures techniques de protection pouvant accompagner ce document. This document is protected by copyright laws and international copyright treaties. Non-authorised use of this document
not presently compete, or compete very distantly, with
the platform’s goods but that may well do so in the future 14. To deal with this concern, the Crémer report
once the start-up has fully developed its market potential. advocates a modification of existing theories of harm
to assess the market position of the merging parties in a
11. Both the Crémer and Furman reports cite the example much broader space than the “narrowly defined product
of the acquisition of Instagram by Facebook in 2014. market.” According to the authors, dominance should be
Although Instagram was only competing with Facebook instead assessed in a “‘users’ space’ that may encompass
in a small segment of the market for social network ap- a broad variety of user needs.”8 We discuss this proposal
plications (photo sharing) at the time of their merger, later in this paper and show that it appears unneces-
the authors suggest (without any sound basis to support sary and carries a substantial risk of both under- and
this conjecture) that it could potentially have developed overenforcement.
into a credible competitor of Facebook absent the trans-
action by progressively entering into adjacent segments 15. It follows from the above that, while the “killer ac-
of social network platforms. According to these reports, quisition” concern relates to the horizontal effects that
this concern is more likely to materialise in the presence may result from the elimination of a potential rival, the
of strong network effects which make it more difficult “digital conglomerate” concern relates to the non-hori-
for a new entrant to reach a critical size and therefore zontal effects that the expansion of a platform in a
result in higher barriers to entry and a greater degree of neighbouring market may generate. In the following two
concentration.6 sections, we therefore discuss the extent to which these
concerns might give rise to anticompetitive effects by suc-
12. To deal with this concern, the Furman report cessively assessing horizontal and non-horizontal aspects
considers that more weight should be given to high entailed by a merger between a digital platform and a
impact/low probability events under an approach called start-up.
“balance of harm.” We discuss this proposal later in this
paper and show that it raises significant practical diffi-
culties and introduces a substantial risk of speculative
theories of harm with consequent adverse effects. III. Assessment
2. The “digital conglomerate”
of horizontal effects
16. Two types of mergers are quite confusingly called
concern “killer acquisitions” by observers:
13. These reports also consider that the combination
of a start-up’s resources (e.g., a specific set of data the – A merger with a competitor already active in the same
start-up generates), services or products with those of relevant market, even though potentially with products
the digital platform may confer the merged entity a sig- and services in an early development phase; and
nificant competitive advantage over its rivals. If this –
A merger with a potential competitor that is not
competitive advantage is likely to materially diminish the presently active in the same relevant market.
ability of rivals to compete with the merged entity, com-
petition may then be reduced and consumers harmed. This section discusses the assessment of horizontal
The Crémer report notes for instance that “frequent- effects in each of these cases.
ly the project of the bought-up start-up is integrated into
the ‘ecosystem’ of the acquirer or into one of their existing
products.”7 According to the authors, such integration
of complementary activities, referred to as “platform en-
1. Merger with
velopment”, may allow the merged entity to attract and a competitor already active
retain customers who value the newly integrated service
at the expense of rival platforms that do not benefit from
in the same market
the combination of these two services. Network effects, 17. Mergers where one of the parties is developing
it is argued, would then strengthen the dominance of a product which does not currently overlap with the
the platform by expanding their scope in the sense that other party’s product, but may well do in a foreseea-
more users would use each of the two services post-merg- ble future, are relatively common and correspond to
er than in a situation where the two services are provided the typical “innovation concerns” raised by the EC in a
independently. The authors argue that this is likely to number of cases.9 In recent years, however, more promi-
further reinforce the dominance of the platform and raise nence has been given to the possibility that these mergers
barriers to entry as any potential entrant would need to allegedly aim at killing future competition. Indeed, in
6 Ibid. 8 Ibid.
7 Ibid. 9 In recent years, this was notably the case in GE/Alstom, Pfizer/Hospira and Dow/DuPont.
44 Concurrences N° 1-2020 I Law & Economics I Tristan Lécuyer I Digital conglomerates and killer acquisitions...a much-discussed paper called “Killer acquisitions”, former.12 Anticipating the threat of competition coming
constitutes a violation of the publisher's rights and may be punished by up to 3 years imprisonment and up to a € 300 000 fine (Art. L. 335-2 Code de la Propriété Intellectuelle). Personal use of this document is authorised within the limits of Art. L 122-5 Code de la Propriété Intellectuelle and DRM protection.
Cunningham et al. (2018) assess this issue by analysing from a surrounding market, firms in the first market
Ce document est protégé au titre du droit d'auteur par les conventions internationales en vigueur et le Code de la propriété intellectuelle du 1er juillet 1992. Toute utilisation non autorisée constitue une contrefaçon, délit pénalement sanctionné jusqu'à 3 ans d'emprisonnement et 300 000 € d'amende (art.
L. 335-2 CPI). L’utilisation personnelle est strictement autorisée dans les limites de l’article L. 122 5 CPI et des mesures techniques de protection pouvant accompagner ce document. This document is protected by copyright laws and international copyright treaties. Non-authorised use of this document
the incentives that some incumbents in the pharmaceuti- may determine price and output not only based on the
cal industry may have to acquire firms that are developing structure of the market in which they evolve, but also
a potentially competing technology in order to shut taking into account potential entry by firms active in the
down the development of that particular technology. The second market. In that sense, a loss of competition can
authors argue, on the basis of their analysis, that signifi- conceivably occur if a merger between a digital platform
cantly lower development rates are observed for products and a start-up operating in an adjacent space results in
that are acquired by overlapping incumbents compared the elimination of a potential entrant. Such concern is
to (i) similar projects that are not acquired and (ii) similar not new and is recognised in the EC Horizontal Merger
projects that are acquired by an incumbent that does not Guidelines.13
have an overlapping product.
22. On the second point, however, tangible evidence on
18. However, this definition of killer acquisitions is at whether a start-up will grow into a significant competitive
odds with what observers have in mind when expressing force constraining the digital platform may not always
concerns with respect to the digital sector. First, in the be available, raising the question of how to account for
type of mergers described by Cunningham et al., the com- hypothetical competition in the future. Traditionally,
petitor is already de facto active in the market. It might competition authorities seek to assess whether a harmful
not be public knowledge, but the target is in the process outcome is more likely to materialise post-merger than
of developing and bringing to the market an overlapping not. The EC Horizontal Merger Guidelines for instance
product. Second, unlike in the digital sector where we state that the objective of merger control is to “assess
see digital platforms materially growing the start-up they whether or not a concentration would significantly impede
acquire (Instagram’s monthly active users, for instance, effective competition.”14 With respect to potential compe-
grew from 50 million in May 2012 when Facebook tition, this means assessing whether the merging parties
acquired it to more than 1 billion today),10 the mergers are more likely to become actual competitors in the
described by Cunningham et al. result in shutting down future than not.
the acquired technology.
23. The Furman report advocates for a replacement of
19. Set aside the question of the relevant notification this approach, called “balance of probability” in the
thresholds,11 the concern raised by the acquisition of a CMA guidelines,15 by an alternative approach called
firm developing an overlapping technology does not seem “balance of harm.” Under the balance of harm approach,
substantially different from the assessment of a merger the authors urge competition authorities to estimate the
between two existing products and there is no reason to outcome of any potential competitive scenarios as well as
assume that the current EU merger control is not well the probability of each of these scenarios to materialise.
equipped to deal with such cases. This way, it is argued, agencies will be able to factor into
merger assessment “low probability, high impact” events
like the possibility for a start-up, not presently active in
2. Merger with a potential the same relevant market as the acquirer, to grow into a
significant competitive force constraining the acquirer in
competitor not presently active the future.
in the same market
24. This approach comes with a number of challenges
20. The acquisition of a potential competitor that is not for merger control. Indeed, to factor into the analysis the
presently active in the same relevant market may raise possibility that a start-up might, one day in a hypothet-
unilateral effects if: ical future, exert a significant competitive constraint on
the acquirer, a competition authority will need to:
– The parties are already constraining each other via the
threat of entry; or – identify a range of possible future outcomes;
– The start-up may grow into a competitive force in the – estimate the probability of each outcome to arise; and
future.
–
assess the impact of each of these outcomes on
customer welfare.16
21. On the first point, the “contestability of markets”
may entail that firms operating in one market might be
constrained by firms operating in a second market due
to the possibility of the latter entering the market of the
12 Baumol et al. (1982).
10 See https://techcrunch.com/2018/06/20/instagram-1-billion-users. 13 EC Horizontal Merger Guidelines, para. 59.
11 Because some digital start-ups may forego short-term profits to attract and grow a signif- 14 Ibid., para. 1.
icant user base before monetising products, it seems both appropriate and important for 15 CMA guidelines on market definition, para. 2.12.
competition authorities to make sure they review any potential horizontal overlap that
may arise in a foreseeable future. This question has been the object of intense discussions 16 For a more detailed discussion of the merits and challenges raised by the “balance of
amongst practitioners recently. harm”approach, see for instance RBB Economics (2019).
Concurrences N° 1-2020 I Law & Economics I Tristan Lécuyer I Digital conglomerates and killer acquisitions... 4525. These are likely to raise significant practical difficul- –
Economies of scope: Economies of scope arise
constitutes a violation of the publisher's rights and may be punished by up to 3 years imprisonment and up to a € 300 000 fine (Art. L. 335-2 Code de la Propriété Intellectuelle). Personal use of this document is authorised within the limits of Art. L 122-5 Code de la Propriété Intellectuelle and DRM protection.
ties for competition authorities in real-life mergers since when there are efficiencies associated to the produc-
Ce document est protégé au titre du droit d'auteur par les conventions internationales en vigueur et le Code de la propriété intellectuelle du 1er juillet 1992. Toute utilisation non autorisée constitue une contrefaçon, délit pénalement sanctionné jusqu'à 3 ans d'emprisonnement et 300 000 € d'amende (art.
L. 335-2 CPI). L’utilisation personnelle est strictement autorisée dans les limites de l’article L. 122 5 CPI et des mesures techniques de protection pouvant accompagner ce document. This document is protected by copyright laws and international copyright treaties. Non-authorised use of this document
supporting evidence will need to be gathered on each of tion of two different goods or services by a single
these points not to fall into some speculative theories of firm compared to their production by two indepen-
harm. dent firms.19 Digital firms generally exhibit significant
economies of scope when it comes to product devel-
opment. If one thinks about the algorithms developed
IV. Assessment of by Google for its search engine for instance, it is easily
conceivable that such algorithms can also be used to
conglomerate effects
provide relevant search results for customers ordering
food online or willing to listen to a particular type of
music on a streaming platform. As shown by Bourreau
26. Having explored what types of horizontal concerns et al. (2019), the strong economies of scope experienced
a merger between a digital platform and a start-up may by digital firms for product development provide more
raise, we now examine the conditions under which the incentives to diversify product lines and expand in new
expansion of a digital platform in a neighbouring market markets.
might give rise to non-horizontal competition concerns.
– Consumption synergies: Complementing an offer with
an additional product or service may increase value
27. We start by setting out the economic mechanisms
for customers. When using Google’s search engine to
explaining why conglomerate mergers are generally
find a venue for dinner for instance, customers may
considered to be procompetitive, in particular in the
find it convenient to directly obtain information from
digital sector where further benefits can be passed on to
Google’s search results on as various things as restau-
consumers. We then show that the current EU framework
rants’ ratings, menus, pictures as well as access to a
appears well suited to deal with potential exclusionary
map locating each restaurant and providing optimised
concerns in situations where significant network effects
itineraries. Because customers value access to several
exist, where access to data may constitute a unique com-
products and services in one place, and because firms in
petitive advantage and platform envelopment dynamics
the digital sector often benefit from strong economies
are at play.
of scope, digital platforms tend to create ecosystems in
which they use complementarity between goods to offer
a broader scope of products and services. A merger
1. Conglomerate mergers are between a digital platform and a start-up company
generally procompetitive, in providing a complementary service can therefore
have procompetitive effects for customers who value
particular in the digital sector the decrease in opportunity cost (e.g., less time spent
searching additional information on localisation and
28. As recognised in the EC Non-Horizontal Merger
itineraries in the restaurant example) brought by the in-
Guidelines, conglomerate mergers are generally consid-
tegration of the two services into a same ecosystem.
ered to have procompetitive effects.17 First, conglomerate
mergers do not entail the elimination of a direct compet- –
Complementarity in capabilities: In addition to the
itive constraint unlike horizontal mergers. Second, when complementarity between products and services,
products sold by the merging parties are complementa- a merger between a large digital platform and a
ry to each other, the merger can bring significant benefits promising start-up may combine complementary ca-
to consumers. This is because demand for each product pabilities. Typically, start-up companies tend to focus
are interrelated: when the price of one product falls, the on the development of one particular technology,
demand for the other product increases, and vice versa. product or service. This is because resources they can
As a result, a merger between complementary products access are generally limited and all oriented towards a
provides the merged entity with incentives to decrease unique objective for which they obtained funding.20 In
prices post-merger following the same type of economic contrast, large digital firms are more likely to possess
mechanisms that give the merging parties incentives to the required resources and skills to further develop
increase prices for substitute products. and commercialise a product at a larger scale (or in a
faster way) than a smaller firm could do. By combining
29. Additional procompetitive effects of conglomer- the innovation brought by the start-up with the capa-
ate mergers exist and have been extensively analysed bilities of a larger digital firm, a conglomerate merger
by the economic literature.18 In the digital sector, these may therefore directly benefit customers through faster
effects are likely to be reinforced by the features of digital access to innovative products or services.
platforms as described in the following paragraphs.
– Exit strategy: Being acquired by an established firm
might give further incentives for start-ups to innovate
in the first place. These incentives have been studied at
17 EC Non-Horizontal Merger Guidelines, para. 11 and following.
19 Panzar et al. (1977).
18 For a detailed discussion of procompetitive effects of conglomerate mergers, see for in-
stance Bishop et al. (2005). 20 See for instance Churchill et al. (1983).
46 Concurrences N° 1-2020 I Law & Economics I Tristan Lécuyer I Digital conglomerates and killer acquisitions...length by the economic literature on the exit strategy 33. In Microsoft/LinkedIn for instance, a core theory of
constitutes a violation of the publisher's rights and may be punished by up to 3 years imprisonment and up to a € 300 000 fine (Art. L. 335-2 Code de la Propriété Intellectuelle). Personal use of this document is authorised within the limits of Art. L 122-5 Code de la Propriété Intellectuelle and DRM protection.
of investors financing start-ups. Cumming et al. (2008) harm considered by the Commission was that, post-merg-
Ce document est protégé au titre du droit d'auteur par les conventions internationales en vigueur et le Code de la propriété intellectuelle du 1er juillet 1992. Toute utilisation non autorisée constitue une contrefaçon, délit pénalement sanctionné jusqu'à 3 ans d'emprisonnement et 300 000 € d'amende (art.
L. 335-2 CPI). L’utilisation personnelle est strictement autorisée dans les limites de l’article L. 122 5 CPI et des mesures techniques de protection pouvant accompagner ce document. This document is protected by copyright laws and international copyright treaties. Non-authorised use of this document
notably carried out an empirical study of the exit er, Microsoft could pre-install LinkedIn on all Windows
strategies of some venture capital funds that invested PCs and integrate LinkedIn into Microsoft Office while
in European start-ups. The authors found that the limiting the interoperability of other professional social
prospects that a start-up will be acquired by a larger networks with Microsoft’s products.22 The Commission
company in the future constitute an important driver considered that these actions could significantly enhance
of the expected returns of investors which is often LinkedIn’s visibility compared to competing professional
deemed superior to the prospects that the start-up will social networks and consequently strengthen LinkedIn’s
become an independent listed company. market position to the detriment of rivals due to the
existence of strong direct network effects (the more pro-
fessionals are active on a particular network, the more
2. Conglomerate concerns valuable this network becomes to users, which in turn
attracts more users, and so on).
in the digital sector:
Network effects, data and 34. The Commission also considered that multi-homing
was limited due to the high level of time and dedication
platform envelopment required to connect with other members and regularly
update profiles. Unlike in Facebook/WhatsApp, it consid-
30. Conglomerate mergers can raise competition issues
ered that multi-homing was therefore unlikely to mitigate
through anticompetitive foreclosure when sales of one
any detrimental consequence of network effects on com-
product are bundled or tied to the sales of another
petition. As a result, Microsoft had to offer a number of
product, such that post-merger customers would switch
commitments to address these competition concerns and
a substantial amount of their purchases from rivals to
obtain the Commission’s approval.
the merged entity for one or the two products.21 If this
results in such a decrease in rivals’ profits that they are
no longer able to exert a significant competitive con- 2.2 Conglomerate concerns
straint on the merged entity in one or the two markets,
the merger must be considered harmful to consumers.
and the importance of data
Importantly, however, if the merger is not likely to bring 35. Another significant feature of digital platforms is
together products that can be sold together in a way that the importance of data. Although the use and the sig-
may hurt rivals, a conglomerate merger is considered pro- nificance of data differ depending on the market, many
competitive under the current guidelines. digital platforms collect data on their users that they sub-
sequently use as an input to improve the services and
31. In the following, we discuss the extent to which the products they sell on each side of the platform. Typically,
current EU framework is well suited to deal with these a firm like Google uses the data it collects on users’
concerns in mergers between a digital platform and searches both to improve its search algorithms on the
a start-up in light of the characteristics of the digital user side of the platform and to sell more targeted ad-
sector—namely, the existence of significant network vertisement products on the other side of the platform.
effects, the importance of data and the dynamics of A relevant question for merger control is therefore
platform envelopment. whether a merger can lead to the acquisition of a par-
ticular set of data which could confer the merged entity
2.1 Conglomerate concerns and network
a significant competitive advantage over rivals. If the
ability of rivals to exert a significant competitive con-
effects straint on the merged entity is materially reduced as a
result of this competitive disadvantage, the merger must
32. When a market is characterised by strong network
be considered as causing harm to consumers.
effects, exclusionary concerns raised by a conglomer-
ate merger are more likely to materialise. This is because
36. In Apple/Shazam for instance, the Commission
the integration of complementary activities may allow
developed a new theory of harm aiming to assess whether
the merged entity to attract and retain customers who
Apple, through the acquisition of the music recognition
value the newly integrated service at the expense of rival
app Shazam, could gain access to critical data about the
platforms that do not benefit from the combination of
preferences of music streaming customers that it could
these two services. Network effects in one of the markets
use to undermine the ability of rival music streaming
may then spread to the other market once services are
platforms to compete.23 In particular, the Commission
provided together compared to a situation where the two
services are provided independently.
22 Microsoft/LinkedIn EC decision, Case M.8124 (2016). In this case, the Commission also
assessed horizontal and vertical effects arising from the acquisition of LinkedIn data by
Microsoft.
23 Apple/Shazam EC decision, Case M.8788 (2018). In this case, the Commission also devel-
oped a theory of harm on input foreclosure and assessed whether Apple could use Shazam’s
strong position in the market for music recognition apps to undermine the ability of Apple
21 For a more detailed discussion of the economics of bundling and tying practices in the Music’s rivals to compete by discontinuing referrals from the Shazam app towards compet-
context of conglomerate mergers, see Gore et al. (2013). ing streaming music platforms.
Concurrences N° 1-2020 I Law & Economics I Tristan Lécuyer I Digital conglomerates and killer acquisitions... 47was concerned that such data could allow Apple to better the authors call “shared user relationships.” As a result,
constitutes a violation of the publisher's rights and may be punished by up to 3 years imprisonment and up to a € 300 000 fine (Art. L. 335-2 Code de la Propriété Intellectuelle). Personal use of this document is authorised within the limits of Art. L 122-5 Code de la Propriété Intellectuelle and DRM protection.
target customers and encourage them to switch from integration of a media player into a broader operating
Ce document est protégé au titre du droit d'auteur par les conventions internationales en vigueur et le Code de la propriété intellectuelle du 1er juillet 1992. Toute utilisation non autorisée constitue une contrefaçon, délit pénalement sanctionné jusqu'à 3 ans d'emprisonnement et 300 000 € d'amende (art.
L. 335-2 CPI). L’utilisation personnelle est strictement autorisée dans les limites de l’article L. 122 5 CPI et des mesures techniques de protection pouvant accompagner ce document. This document is protected by copyright laws and international copyright treaties. Non-authorised use of this document
competitors’ platforms to Apple Music. system and associated software is likely to generate
substantial consumption synergies for customers via
37. For this strategy to significantly undermine rivals’ enhanced functionalities (e.g., easier access to videos in
ability to compete, two cumulative conditions must be met. emails). By tying its own media player with the other
features of the Windows ecosystem, Microsoft therefore
– First, similar data must not be available to rival music managed to leverage economies of scope and consump-
streaming platforms. If competitors can access similar tion synergies to make a substantial number of Windows
data without incurring significant additional costs users migrate from competing media players to Windows
compared to the merged entity, there is no reason to Media Player. This allowed Microsoft to rapidly overtake
suspect that acquisition of the data in question would Real, the largest media player at the time.
put Apple Music’s rivals at a competitive disadvantage.
41. Paradoxically, this theory is cited as an important
– Second, even if the data acquired by Apple had unique
issue arising from the acquisition of start-ups in both
characteristics allowing Apple Music to better target
Crémer and Bourreau (2019) while the characteris-
customers’ preferences, it must allow the merged entity
tics of platform envelopment actually highlight the
to divert a sufficient number of users from competi-
ability of some firms to grow into a competitive force
tors’ services to its own platform for the competitive
in a neighbouring market without resorting to external
constraint exerted by rivals to be significantly reduced
acquisitions.
post-merger.
42. First, platform envelopment suggests that firms active
It suffices to prove that one of these two conditions does
in neighbouring markets can leverage economies of scope
not hold for competitive concerns to be dismissed.
and consumption synergies to successfully enter (or
threaten to do so) the merging parties’ markets. Typically,
38. Following an in-depth investigation, the Commission
the above-mentioned expansion of Microsoft into the
concluded that none of these conditions were met. First,
market for media players has been achieved through the
it found that the data collected by Shazam was not
in-house development of the Windows Media Player
unique and that many alternative datasets were available
software and not as the result of a merger. Similarly, when
to competitors. Second, the investigation revealed that
it recently entered the market for food delivery services
Shazam’s data were unlikely to substantially increase
Uber leveraged its existing software solutions for con-
Apple’s ability to better target customers and make them
necting drivers and users, as well as optimising drivers’
switch to Apple Music.
schedules, from its transportation business. Furthermore,
Uber’s integration of the two services within a same app
2.3 Conglomerate concerns and platform environment also generated significant consumption
envelopment synergies for consumers (e.g., easy access to Uber Eats
services when using the Uber app for transportation,
39. In recent years, a new economic theory emerged to and vice versa) that may have facilitated its entry in the
describe the expansion of large platforms with a strong food delivery market. As previously discussed, these two
user base into neighbouring markets through the acqui- types of efficiencies have clear procompetitive benefits for
sition of start-ups. This theory is notably referred to as consumers.
“platform envelopment” in the Crémer report as per the
title of an article by Eisenmann et al. (2011).24 43. Second, many examples of strong market positions
in the digital sector do not result from platform envel-
40. Platform envelopment analyses entry by one platform opment. In this respect, it is particularly interesting to
into another market through the combination of “its own note that Windows Media Player has now long been
functionality with the target’s in a multi-platform bundle challenged by competing video or music players for phys-
that leverages common components and/or shared user rela- ically owned content (e.g., VLC for video) and, more
tionships.”25 One example Eisenmann et al. provide is the importantly, has been completely marginalised by the de-
launch by Microsoft of Windows Media Player in 1998. velopment of video and music streaming on the internet
Media players and PC operating systems share common and dedicated platforms (e.g., Netflix or Spotify).26 None
components. Typically, they both require developers of these companies have benefited from “platform envel-
coding software functionalities and interoperability. In opment” to build their strong market position in their
that sense, the economies of scope described above are respective digital markets.
likely to generate substantial efficiencies when develop-
ing a media player. Similarly, users of media players and 44. Because all envelopment theories seem to entail the in-
PC operating systems overlap, corresponding to what tegration of the merging parties’ products or services into
a broader ecosystem, called “multi-platform bundle” by
24 See Crémer report (2019): “The expansion of the power of established platforms with a
strong user base and a conglomerate profile into new markets pioneered by other platforms or
firms is currently debated under the heading of ‘platform envelopment’.” 26 Regarding music for instance, statistics from the Recording Industry Association of Amer-
ica (RIAA) show that in 2018 more than 80% of the US recorded music revenues have been
25 Eisenmann et al. (2011). generated by streaming. See https://www.riaa.com/u-s-sales-database.
48 Concurrences N° 1-2020 I Law & Economics I Tristan Lécuyer I Digital conglomerates and killer acquisitions...V. Conclusion
constitutes a violation of the publisher's rights and may be punished by up to 3 years imprisonment and up to a € 300 000 fine (Art. L. 335-2 Code de la Propriété Intellectuelle). Personal use of this document is authorised within the limits of Art. L 122-5 Code de la Propriété Intellectuelle and DRM protection.
Eisenmann et al., it is unclear why the current framework
Ce document est protégé au titre du droit d'auteur par les conventions internationales en vigueur et le Code de la propriété intellectuelle du 1er juillet 1992. Toute utilisation non autorisée constitue une contrefaçon, délit pénalement sanctionné jusqu'à 3 ans d'emprisonnement et 300 000 € d'amende (art.
L. 335-2 CPI). L’utilisation personnelle est strictement autorisée dans les limites de l’article L. 122 5 CPI et des mesures techniques de protection pouvant accompagner ce document. This document is protected by copyright laws and international copyright treaties. Non-authorised use of this document
used by the Commission to assess mergers involving
potential bundling and tying practices would need ad- 48. As discussed throughout this article, the source of
justment. Indeed, the above discussion of the Microsoft/ competitive concerns arising from a start-up acquisi-
LinkedIn case indicates that the Commission is well able tion by a digital platform active in a different market
to address such theories of harm and impose the appro- sometimes closely relates to the efficiencies that the
priate remedies. merger generates. Assuming that these effects are anti-
competitive risks prohibiting benign mergers but also,
45. In particular, the Crémer report’s proposal to assess more importantly, mergers likely to entail substantial pro-
market position of the merging parties in a “users’ competitive effects to consumers. It is therefore critical
space” appears both arbitrary and unnecessary and, in to properly assess the procompetitive effects a merger is
consequence, carries a substantial risk of both under- likely to generate and to compare them to the potential
and overenforcement. To see why, it seems helpful to anticompetitive effects that have been identified.
conduct the following “thought experiment” to identify
potential outcomes that would result from this approach 49. Departing from this approach to reverse or harden
in practice. the standard of proof for digital mergers, as some com-
mentators suggest, constitutes a risky path to follow for
46. Amazon and Uber have many similarities: they are the fairness and the predictability of merger control in
both large digital platforms that connect users on the Europe. Given the high uncertainty regarding future
one side and service providers on the other (merchants competition, the changes proposed by the Furman and
of physical goods for the former, drivers and restaurants Crémer reports risk leading to misconceived theories
for the latter). There are probably many customers who of harm while reversing the burden of proof for digital
use both platforms such that substantial consumption mergers is likely to deprive customers from procompet-
synergies may be generated if both services were to be itive benefits and jeopardise innovation in the digital
integrated into a single ecosystem. Similarly, given the sector.
nature of the resources used by both companies, there
may exist significant economies of scope associated to 50. Increasing challenges and debates are likely to
the supply of the two services altogether (in terms of continue on how to make sure competition authorities
product development and underlying platform algorithm properly assess the potential impact of start-up acqui-
for instance). However, this does not mean that Amazon sitions by digital platforms. However, it is critical to
is constrained by Uber, and vice versa. If prices of, say, continue assessing mergers based on a careful assessment
goods sold on Amazon were to increase, no customers of the economic mechanisms through which a merger
will switch to services provided by Uber such that may harm consumers and not on abstract theories of
the existence of the Uber platform is irrelevant to the harms or presumptions of anticompetitive effects. n
ability of Amazon to profitably raise prices. Assessing
dominance on a users’ space that would include Amazon
and Uber would therefore materially understate the
market power of these digital platforms on each of the
markets they operate.
47. Conversely, considering now, as part of this “thought
experiment”, a hypothetical merger between Amazon
and Uber, the same logic can lead competition author-
ities to overstate the competitive constraints the two
platforms are imposing on each other. Indeed, assessing
the impact of this hypothetical merger on a “users’ space”
would give unjustified importance to the other platform
in the competitive assessment while, as discussed above,
the two platforms do not exert any significant competi-
tive constrain on each other pre-merger.
Concurrences N° 1-2020 I Law & Economics I Tristan Lécuyer I Digital conglomerates and killer acquisitions... 49References
constitutes a violation of the publisher's rights and may be punished by up to 3 years imprisonment and up to a € 300 000 fine (Art. L. 335-2 Code de la Propriété Intellectuelle). Personal use of this document is authorised within the limits of Art. L 122-5 Code de la Propriété Intellectuelle and DRM protection.
Eisenmann, T., Parker, G. & Van Alstyne, M. (2011),
Ce document est protégé au titre du droit d'auteur par les conventions internationales en vigueur et le Code de la propriété intellectuelle du 1er juillet 1992. Toute utilisation non autorisée constitue une contrefaçon, délit pénalement sanctionné jusqu'à 3 ans d'emprisonnement et 300 000 € d'amende (art.
L. 335-2 CPI). L’utilisation personnelle est strictement autorisée dans les limites de l’article L. 122 5 CPI et des mesures techniques de protection pouvant accompagner ce document. This document is protected by copyright laws and international copyright treaties. Non-authorised use of this document
Platform envelopment, Strategic Management Journal,
Baumol, W. J., Panzar, J. C. & Willig, R. D. (1982), 32(12): 1270–1285.
Contestable Markets and the Theory of Industry
Structure, Harcourt Brace Jovanovich. Furman, J., Coyle, D., Fletcher, A., McAuley, D. &
Marsden, P. (2019), Unlocking digital competition,
Bishop, S., Lofaro, A., Rosati, F. & Young, J. (2005), Report of the Digital Competition Expert Panel.
The Efficiency-Enhancing Effects of Non-Horizontal
Mergers, Report for DG Enterprise and Industry, Gore, D., Lewis, S., Lofaro, A. & Dethmers, F. (2013),
European Commission. The Economic Assessment of Mergers under European
Competition Law, Cambridge University Press, ch. 5.
Bourreau, M. & de Streel, A. (2019), Digital
Conglomerates and EU Competition Policy, Panzar, J. C. & Willig, R. D. (1977), Economies of Scale
available at SSRN: 3350512. in Multi-Output Production, The Quarterly Journal
of Economics, 91(3): 481–493.
Churchill, N. C. & Lewis, V. L. (1983), The Five Stages
of Small Business Growth, Harvard Business Review, RBB Economics (2019), A question of balance:
61(3): 30–50. Comments on a proposed new test for UK merger
control, RBB brief, No. 59.
Crémer, J., de Montjoye, Y.-A. & Schweitzer,
H. (2019), Competition policy for the digital era, Scott Morton, F., Nierenberg, T., Bouvier, P., Ezrachi,
Report for the European Commission. A., Jullien, B., Katz, R., Kimmelman, G., Melamed
A. D. & Morgenstern, J. (2019), Report: Committee
Cumming, D. & binti Johan, S. A. (2008), Preplanned for the Study of Digital Platforms – Market Structure
Exit Strategies in Venture Capital, European Economic and Antitrust Subcommittee, George J. Stigler
Review, 52(7): 1209–1241. Center for the Study of the Economy and the State,
The University of Chicago Booth School of Business.
Cunningham, C., Ederer, F. & Ma, S. (2018),
Killer acquisitions, available at SSRN: 3241707. Valletti, T. (2018), Après moi, le déluge ! Tech giants
in the digital age, available at: https://ecp.crai.com/
wp-content/uploads/2018/12/Tommaso-Valletti-2018.
pdf.
50 Concurrences N° 1-2020 I Law & Economics I Tristan Lécuyer I Digital conglomerates and killer acquisitions...Concurrences
Editoriaux Droit & économie
Jacques Attali, Elie Cohen, Claus‑Dieter Emmanuel Combe, Philippe Choné,
Ehlermann, Jean Pisani Ferry, Ian Forrester, Laurent Flochel, Frédéric Jenny,
Eleanor Fox, Douglas H. Ginsburg, Gildas de Muizon, Jorge Padilla,
Laurence Idot, Frédéric Jenny, Arnaud Penelope Papandropoulos, Anne Perrot,
Montebourg, Mario Monti, Gilbert Parleani, Nicolas Petit, Etienne Pfister, Francesco Rosati,
Jacques Steenbergen, Margrethe Vestager, David Sevy, David Spector...
Bo Vesterdorf, Denis Waelbroeck,
Marc van der Woude...
Chroniques
Concurrences est une revue Interviews Ententes
Ludovic Bernardeau, Anne-Sophie Choné
trimestrielle couvrant l’ensemble Sir Christopher Bellamy, Lord David Currie,
des questions de droits de Thierry Dahan, Jean-Louis Debré, Isabelle Grimaldi, Michel Debroux, Etienne Thomas
l’Union européenne et interne de Silva, François Fillon, John Fingleton, Pratiques unilatérales
de la concurrence. Les analyses Renata B. Hesse, François Hollande,
William Kovacic, Neelie Kroes, Laurent Binet, Frédéric Marty,
de fond sont effectuées sous Anne Wachsmann
forme d’articles doctrinaux, Christine Lagarde, Johannes Laitenberger,
de notes de synthèse ou Emmanuel Macron, Robert Mahnke, Pratiques commerciales
Ségolène Royal, Nicolas Sarkozy, déloyales
de tableaux jurisprudentiels.
Marie‑Laure Sauty de Chalon,
L’actualité jurisprudentielle Tommaso Valletti, Christine Varney... Frédéric Buy, Valérie Durand,
et législative est couverte par Jean‑Louis Fourgoux, Rodolphe Mesa,
onze chroniques thématiques. Marie‑Claude Mitchell
Dossiers Distribution
Nicolas Ereseo, Dominique Ferré,
Jacques Barrot, Jean-François Bellis, Didier Ferrier, Anne-Cécile Martin
David Bosco, Murielle Chagny, John Connor,
Damien Géradin, Assimakis Komninos, Concentrations
Christophe Lemaire, Ioannis Lianos, Jean-François Bellis, Olivier Billard,
Pierre Moscovici, Jorge Padilla, Emil Paulis, Jean‑Mathieu Cot, Ianis Girgenson,
Robert Saint-Esteben, Jacques Steenbergen, Sergio Sorinas, David Tayar
Florian Wagner-von Papp, Richard Whish...
Aides d’État
Jacques Derenne, Bruno Stromsky,
Articles
Raphaël Vuitton
Procédures
Guy Canivet, Emmanuelle Claudel, Pascal Cardonnel, Alexandre Lacresse,
Emmanuel Combe, Thierry Dahan, Luc Gyselen, Christophe Lemaire
Daniel Fasquelle, Barry Hawk, Nathalie
Homobono, Laurence Idot, Frédéric Jenny, Régulations
Bruno Lasserre, Luc Peeperkorn, Anne Perrot, Orion Berg, Hubert Delzangles,
Nicolas Petit, Catherine Prieto, Patrick Rey, Emmanuel Guillaume
Joseph Vogel, Wouter Wils...
Mise en concurrence
Bertrand du Marais, Arnaud Sée
Pratiques Actions publiques
Jean-Philippe Kovar, Francesco Martucci,
Tableaux jurisprudentiels : Actualité
des enquêtes de concurrence, Stéphane Rodrigues
Contentieux indemnitaire des pratiques Droits européens et
anticoncurrencielles, Bilan de la pratique
étrangers
des engagements, Droit pénal et concurrence,
Legal privilege, Cartel Profiles in the EU... Walid Chaiehloudj, Sophie‑Anne Descoubes,
Marianne Faessel, Pierre Kobel, Silvia Pietrini,
Jean‑Christophe Roda, François Souty,
International
Stéphanie Yon-Courtin
Livres
Belgium, Brésil, Canada, China, Germany,
Hong‑Kong, India, Japan, Luxembourg,
Switzerland, Sweden, USA...
Sous la direction de Stéphane Rodrigues
Revues
Christelle Adjémian, Mathilde Brabant,
Emmanuel Frot, Alain Ronzano, Bastien ThomasYou can also read
