SSRN-id3599407 2

Anticompetitive Effects in EU Competition Law
Pablo Ibáñez Colomo*
Contents
1. Introduction ................................................................................................................................................................ 2
2. What are anticompetitive effects? Making sense of the variables.............................................................................. 6
2.1. The time variable: actual and potential effects .................................................................................................. 6
2.2. The dimensions of competition and of the counterfactual................................................................................. 7
2.3. The meaning of effects ...................................................................................................................................... 9
2.4. The probability of the effects: plausibility, likelihood and certainty ............................................................... 12
3. Anticompetitive effects in the case law: preliminary issues .................................................................................... 14
3.1. Mechanisms through which anticompetitive effects are manifested ............................................................... 14
3.2. The place of anticompetitive effects in the various legal tests ........................................................................ 16
4. The time dimension in the case law: actual and potential effects ............................................................................ 20
4.1. The prospective analysis of potential effects ................................................................................................... 20
4.2. The retrospective analysis of actual effects ..................................................................................................... 22
5. The meaning of competition and the counterfactual in the case law........................................................................ 24
5.1. Competition comprises both the inter-brand and intra-brand dimensions ....................................................... 24
5.2. Competition refers to such lawful competition which would otherwise have existed..................................... 25
5.3. Both the ex ante and ex post dimensions of the counterfactual are considered ............................................... 29
6. The meaning of effects in the case law .................................................................................................................... 31
6.1. Only appreciable effects are relevant under EU competition law ................................................................... 31
6.2. An effect is more than a competitive disadvantage or a limitation of a firm’s freedom of action .................. 34
6.3. As a rule, an anticompetitive effect cannot be equated with harm to consumer welfare................................. 36
6.4. Anticompetitive effects exist where competitive pressure is reduced ............................................................. 38
7. The probability of effects in the case law ................................................................................................................ 45
7.1. Semantic issues: capability and likelihood ...................................................................................................... 45
7.2. The law as applied: plausibility, likelihood and certainty ............................................................................... 47
8. Analysis and discussion ........................................................................................................................................... 52
8.1. Questions addressed in the case law ................................................................................................................ 52
8.2. Open questions ................................................................................................................................................ 57
8.3. Areas of friction in practice ............................................................................................................................. 61
9. Conclusions .............................................................................................................................................................. 67
*
London School of Economics and College of Europe. E-mail: [email protected]. I am grateful to Fernando
Castillo de la Torre, Gianni De Stefano, Andriani Kalintiri and Alfonso Lamadrid de Pablo for their comments on a
previous version. In accordance with the ASCOLA declaration of ethics, I am happy to clarify that I have nothing to
disclose.
Electronic copy available at: https://ssrn.com/abstract=3599407

1. Introduction
The ‘effects-based’ approach to EU competition law has been widely discussed for over two decades.1
One would thus be forgiven to assume that the meaning and scope of the notion of anticompetitive
effects in the EU legal order are, and have long been, clear. The opposite, however, is true. Some
elements of the notion have not been fully teased out, and a cloud of uncertainty surrounds the
evaluation of effects in concrete cases. It is true that the Court of Justice (hereinafter, the ‘Court’ or
the ‘ECJ’) has made seminal contributions over the years. For instance, it has consistently held that,
as a general rule, anticompetitive effects must not be equated with a negative impact on consumer
welfare.2 Beyond this point, however, many open issues remain. If, as the case law suggests, effects
amount to more than a mere competitive disadvantage, what are they? How are potential effects
measured? How is the analysis performed when the practice is not price-related? What is the requisite
level of probability? These are just some of the questions – of major practical significance – that play
a prominent role before courts and authorities.
A number of factors help explain the persistent uncertainty. The definition of its
administrative priorities by the European Commission (hereinafter, the ‘Commission’) is one of them.
It is well documented that a growing number of investigations focuses on clear-cut infringements (in
particular, cartel conduct).3 Such practices are prima facie unlawful irrespective of their impact on
1
It would be difficult to summarise the literature dealing with the effects-based approach. Suffice it to mention, inter alia,
Jacques Bourgeois and Denis Waelbroeck (eds), Ten years of effects-based approach in EU competition law: state of play
and perspectives (Bruylant 2012); Luc Peeperkorn and Katja Viertiö, ‘Implementing an effects-based approach to Article
82’ (2009) 1 Competition Policy Newsletter 17; Wouter Wils, ‘The Judgment of the EU General Court in Intel and the
So-Called “More Economic Approach” to Abuse of Dominance’ (2014) 37 World Competition 405; and Anne Witt, ‘The
Enforcement of Article 101 TFEU - What Has Happened to the Effects Analysis?’ (2018) 55 Common Market Law
Review 417.
2
See in particular Opinion of AG Kokott in Case C-95/04 P British Airways plc v Commission, EU:C:2006:133, para 86;
Case T-8/08, T-Mobile Netherlands BV and others v Raad van bestuur van de Nederlandse Mededingingsautoriteit,
EU:C:2009:343, para 38; Joined Cases C‑501/06 P, C-513/06 P, C-515/06 P and C‑519/06 P GlaxoSmithKline Services
Unlimited v Commission, EU:C:2009:610 (‘Glaxo Spain’), para 63; Case C‑286/13 P Dole Food Company Inc. and Dole
Fresh Fruit Europe v Commission, EU:C:2015:184 (‘Bananas’), para 125; Case T‑216/13, Telefónica, SA v Commission,
EU:T:2016:369, para 270.
3
For a systematic analysis of the enforcement priorities of the Commission, and how cartel conduct has gained in
relevance over the years, see Pablo Ibáñez Colomo and Andriani Kalintiri, ‘The Evolution of EU Antitrust Policy: 1966-
2017’ (2020) 83 Modern Law Review 321.

competition.4 As a result, the analysis of effects tends to be, if at all, a marginal aspect of such cases.
Second, only a fraction of decisions ever reaches review courts. In the field of merger control, and
given the time-sensitive nature of transactions, the proportion of challenges is small.5 In the context
of Articles 101 and 102 TFEU, the decisions in the very cases that demand a meaningful analysis of
effects are not always subject to judicial review. A substantial proportion of non-cartel cases
examined by the authority have been resolved, in the past decades, by means of commitment
decisions. These decisions do not formally declare whether competition law has been infringed and
are rarely challenged before the EU courts.6 In addition, the use of settlements has progressively
grown over the years and across practices. Firms’ incentives to bring an action are significantly
reduced following an agreement with an authority, and this even when the categorisation of the
practice by the Commission is not wholly uncontroversial.7
It is true that, since the adoption of Regulation 1/2003 – which sought to increase enforcement
at the national level – there has been a steady stream of preliminary references seeking guidance on
the interpretation and application of Articles 101 and 102 TFEU. These judgments provide valuable
indicators about the way in which the analysis of effects is to be conducted and about the specific
meaning of the fundamental underlying concepts. The very nature of preliminary rulings, however,
limits the extent to which they can clarify matters. Where the impact of a practice is assessed, the
analysis cannot be expected to go beyond providing the relevant factors that the national court would
4
Case C-67/13 P, Groupement des cartes bancaires v Commission, EU:C:2014:2204 (‘Cartes Bancaires’), para 51.
5
For a systematic overview of the number of cases reaching the EU courts, see Pablo Ibáñez Colomo, The Shaping of EU
Competition Law (Cambridge University Press, 2018) and the accompanying dataset.
6
See in this sense Article 9 of Council Regulation (EC) No 1/2003 of 16 December 2002 on the implementation of the
rules on competition laid down in Articles 81 and 82 of the Treaty [2003] OJ L1/1. For an exhaustive analysis of their
nature, see Ryan Stones, ‘Commitment Decisions in EU Competition Enforcement: Policy Effectiveness v. the Formal
Rule of Law’ (2019) 38 Yearbook of European Law 361.
7
On the rise of the use of settlements in EU competition law, see Niamh Dunne, ‘From Coercion to Cooperation:
Settlement within EU Competition Law’ (2019) LSE Legal Studies Working Paper, 14, available at
https://ssrn.com/abstract=3481419.

need to apply to the facts of the case.8 In some instances, the concepts used may not be fully fleshed
out.
As a result of this institutional reality, the aspects of the case law that shed light on the notion
of anticompetitive effects tend to be available in a fragmentary (and sometimes embryonic) manner
– and this, from a variety of disparate sources. The purpose of this paper, against this background, is
to bring together these fragments and present them under a coherent framework. The exercise
suggests that the notion of anticompetitive effects has a concrete meaning in EU competition law.
The case law provides the necessary and sufficient elements to understand how effects are to be
assessed in practice. The conclusion is true of all areas of EU competition law, in the sense that there
are no appreciable differences in the analysis of effects between EU merger control, on the one hand,
and Articles 101 and 102 TFEU, on the other. If there are variations, these relate to the applicable
legal test, not the relevant provision.
Some of the principles underpinning the case law are explicit. To begin with, it does not seem
possible to claim that anticompetitive effects can be equated with a competitive disadvantage, or with
a limitation of a firm’s freedom of action. Something more – namely an impact on efficient firms’
ability and incentive to compete – is required as a matter of law. Similarly, the notions of competition
and the counterfactual have acquired, over the years, a clear and stable meaning. In the EU system,
competition is best understood as actual or potential lawful competition which would have existed in
the absence of the practice (or transaction) under consideration. Finally, the Court has consistently
held that anticompetitive effects can be actual or potential; by the same token, the analysis can be
both prospective and retrospective.
Other principles are implicit or scattered across the relevant judgments. This article seeks to
tease them out from the case law. The requisite level of probability that triggers intervention, for
instance, has never been formulated explicitly. What is more, the Court has occasionally used as
8
Some examples in this sense include Case C-209/10 Post Danmark A/S v Konkurrencerådet, EU:C:2012:172 (‘Post
Danmark I’); and Case C-345/14 SIA ‘Maxima Latvija’ v Konkurences padome, EU:C:2015:784.

synonymous expressions that are suggestive of different thresholds of effects. If one looks at the plain
meaning of the words, there seems to be a difference between showing that a practice is capable of
restricting competition and requiring evidence that it is likely to do so. However, the Court has
sometimes used them as synonymous.9 Against this background, the relevant thresholds need to be
inferred from the analysis as actually conducted. In this sense, this article pays attention not only to
what it is declared in the judgments but also – and mainly – to what the Court does in concrete cases
(that is, how the analysis is performed and, in particular, how penetrating it is).
The abovementioned questions are examined as follows. Section 2 provides the framework
around which the various aspects of the case law can be organised. It identifies the main variables to
consider when giving a concrete meaning to the notion of anticompetitive effects. Thus, it examines
not only the ways in which the key concepts of ‘competition’ and ‘effects’ may be defined but also
the levels of probability, as well as the differences between the evaluation of actual and potential
effects. Sections 3 to 7, in turn, examine how the variables defined in Section 2 have been interpreted
by the Court. In this regard, it considers both the case law on Articles 101 and 102 TFEU and on
merger control. As far as Article 101 TFEU is concerned, it takes into account not only ‘by effect’
conduct (that is, conduct that only gives rise to intervention when effects are established) but also ‘by
object’ behaviour (that is, behaviour that is deemed prima facie unlawful irrespective of its effects).
Finally, Section 8 offers a summary of the main findings and identifies the aspects that are likely to
give rise to frictions in the medium to long term.
9
See for instance Case C-95/04 P British Airways plc v Commission, EU:C:2007:166; Case C-52/09 Konkurrensverket v
TeliaSonera Sverige AB, EU:C:2011:83; and Case C-23/14 Post Danmark A/S v Konkurrencerådet, EU:C:2015:651
(‘Post Danmark II’). For a discussion, see Opinion of AG Wahl in Case C-413/14 P Intel Corporation Inc v Commission,
EU:C:2016:788, paras 112-121.

2. What are anticompetitive effects? Making sense of the variables
The notion of anticompetitive effects can be broken down into various components. Its scope will be
broader or narrower depending on how these components are fleshed out. For instance, it would be
relatively broad if effects were equated with a competitive disadvantage – or with a limitation of a
firm’s freedom of action. In such circumstances, anticompetitive effects would be found to exist in
the vast majority of – if not all – cases. Conversely, the scope would be narrower, and effects
relatively more difficult to establish in practice, if the latter were equated with harm to consumer
welfare. The notion is also sensitive to the requisite threshold of probability. Effects would be
relatively easy to establish if the bar were set at the level of plausibility; and they would be more
difficult to establish if it were necessary to show that an anticompetitive impact is certain, or virtually
certain, to occur. The purpose of this section is to identify these components and provide a template
against which the case law, and its evolution, can be mapped.
2.1. The time variable: actual and potential effects
The evaluation of effects can be based on the actual or potential impact of a practice (or transaction)
on competition. If intervention is based on the latter, intervention need not wait for the impact to be
manifested or to fully display its consequences on the market. The assessment of potential effects is
typically prospective in nature (and is often assumed to be prospective). For instance, action may take
place before a practice or transaction is fully realised. Merger control systems typically evaluate the
compatibility of concentrations prior to their implementation, and this on the basis of their anticipated
effects. The same assessment of potential effects can take place in relation to agreements between
undertakings or unilateral practices by dominant firms (or collectively dominant firms).10 For
10
The remainder of the article will refer to a ‘dominant firm’ as a shorthand for both single and collective dominance.

instance, an authority may intervene ex ante, before an auction is organised.11 The analysis may also
be prospective where the practice has already been implemented but the effects have not been
manifested to their full extent. This may be the case, for instance, where the practice is ongoing when
intervention takes place.12
A question that might arise in practice is whether the retrospective analysis of the impact of a
practice or transaction can be based on its potential effects alone or must consider its actual effects
instead. The question, in other words, is whether the ex post assessment of effects can focus
exclusively on the conditions that would have potentially prevailed in its absence; or whether the
assessment must take into account the actual context (that is, the subsequent developments that are
contemporaneous with, or that follow, its implementation). Under the first approach (which would
consider potential effects alone), the ex post evaluation would be based on hypotheticals about the
possible evolution of market conditions, irrespective of actual events; under the second, the
assessment would be constrained by the observable evolution of such conditions. The status of this
question in the case law is considered in Section 4 below.
2.2. The dimensions of competition and of the counterfactual
2.2.1. Inter-brand and intra-brand competition
It is commonplace to distinguish between the inter-brand and the intra-brand dimensions of
competition. Inter-brand competition refers, generally speaking, to the competition that exists among
firms at a given level of the value chain, and, in particular, to the rivalry that exists between the
different suppliers of a given product. Intra-brand competition, in turn, is typically defined as the
11
See for instance Joint selling of the media rights to the FA Premier League (Case COMP/C-2/38.173) Commission
Decision of 22 March 2006.
12
Google Search (Shopping) (Case AT.39740) Commission Decision of 27 June 2017 (‘Google Shopping’).

rivalry that exists among the distributors or retailers of a particular brand of a given product. In any
competition law system, a choice needs to be made, first, about whether both dimensions of
competition are relevant (or whether, instead, only inter-brand competition is) and, second, where the
two are deemed relevant, whether there is a hierarchy between the two dimensions. Since inter-brand
competition is universally understood to be more important, it can be expected to take precedence.
The status of inter-brand and intra-brand competition in the case law is addressed in Section 5.
2.2.2. The ex ante and ex post dimensions of competition
The impact of a practice (or transaction) on competition has to be measured against a benchmark. The
competition that is said to be affected, in other words, needs to be given a concrete meaning. The
most obvious benchmark to establish anticompetitive effects is the counterfactual, that is, the
evaluation of the conditions of competition that would have prevailed had the practice (or transaction)
not been implemented. An analysis of the relevant economic and legal context (including factors such
as the features of the relevant market or the regulatory conditions in which firms operate) sheds light
on this question. The operation of the counterfactual is exemplified by the so-called ‘failing firm
defence’ in merger control.13 A concentration cannot be said to have anticompetitive effects if one of
the parties would have left the market irrespective of it and there would have been no less restrictive
alternatives to dispose of the assets.14 In such circumstances, there would be no causal link between
the anticipated loss of competitive pressure and the completion of the transaction.
The counterfactual has an ex ante and an ex post dimension. When considering the conditions
of competition that would otherwise have prevailed, one should take into account that some practices
and transactions both create and restrict competition; in other words, pro- and anticompetitive effects
13
See Guidelines on the assessment of horizontal mergers under the Council Regulation on the control of concentrations
between undertakings [2004] OJ C31/5, paras 89-91.
14
Ibid, para 90.

may be closely intertwined. For instance, the development of a new technology may create a new
market or intensify rivalry in an existing one. In that regard, it can be said to be pro-competitive. Once
the technology has been developed, however, the firm may decide to keep it for itself to recoup its
investments. Accordingly, it may refuse to share its innovation with would-be rivals. Seen from a
purely ex post perspective, the decision to keep the technology for itself appears to restrict
competition (insofar as would-be rivals would be denied access to the market).
When defining the counterfactual against which effects are evaluated, it is therefore necessary
to decide whether both the ex ante and ex post dimensions are considered, or whether only the latter
is part of the analysis. If, in the example given above, only the ex post dimension were relevant, then
a refusal by the firm to share its technology with rivals would have anticompetitive effects. Seen ex
post, this behaviour necessarily limits the availability of the technology that would otherwise have
existed. This is not necessarily the case, however, when the ex ante dimension is considered. The
investment in the development of a new technology might never have been incurred if the firm had
not had the prospect of recouping it by keeping the innovation for itself. An ex ante approach to the
counterfactual does not merely assume that the pro-competitive gains resulting from a practice (or
transaction) would have existed. Instead, it considers whether they would have been manifested in
the absence of the observable ex post restraints. As explained in Section 5, the EU system considers
both the ex ante and ex post dimensions of the counterfactual.
2.3. The meaning of effects
2.3.1. Between freedom of action and harm to consumers
The assessment of the impact of a practice or transaction on competition is particularly sensitive to
the way the concept of effects itself is defined. As already suggested, anticompetitive effects would

be virtually ubiquitous (and straightforward to establish) if the concept were equated with a
competitive disadvantage or with a limitation of a firm’s freedom of action. The opposite would be
true if it were understood to amount to harm to consumer welfare. When considering the possible
meanings that can be attached to the very concept of effects, it is useful to think of such meanings as
discrete points along a spectrum. The two options mentioned above (limitation of a firm’s freedom
of action or competitive disadvantage, on the one hand; and harm to consumer welfare, on the other)
are at the opposite ends of the spectrum.
Virtually any practice that attracts the attention of competition authorities limits the freedom
of action of one or several firms and/or places them at a competitive disadvantage. A distribution
agreement, for instance, may limit the freedom of action of the reseller (and may thus place it at a
competitive advantage relative to rivals) in a variety of ways. The reseller may be prevented from
selling brands competing with those of the supplier,15 or may be precluded from offering the
contractual products via an online marketplace.16 The same can be said of conduct implemented by
dominant firms. For instance, tying places rivals on the market for the tied product at a disadvantage.
In addition, the practice limits the freedom of action of the firm’s customers. At the other edge of the
spectrum, equating anticompetitive effects with harm to consumers (in terms of, inter alia, prices,
output, quality or innovation) would require evidence that the latter are made worse off by the practice
or transaction.17 This is known to be a particularly demanding exercise.
Freedom of Market As efficient Consumer
action* structure competitor welfare
Fig. 1: The meaning of effects

* (or competitive disadvantage)
15
See, for instance, Case C-234/89 Stergios Delimitis v Henninger Bräu AG, EU:C:1991:91.
16
See, for instance, Case C-230/16 Coty Germany GmbH v Parfümerie Akzente GmbH, EU:C:2017:941.
17
For an articulation of what the assessment would involve, see, for instance David Spector, ‘From Harm to Competitors
to Harm to Competition: One More Effort, Please!’ (2006) 2 European Competition Journal 145.
10

As seen in Figure 1, it is possible to identify at least two points in the spectrum that are
between these two ends. First, one can define effects as harm to the market structure in which the
practice or transaction is implemented. In such a case, the evaluation would focus on its impact on
other firms on the relevant market. Second, it is possible to further refine the analysis and equate
effects with an impact on equally efficient firms. If the latter definition is embraced, the consequences
on firms that are not as efficient as those implementing the practice or transaction would not amount
to anticompetitive effects. The departure of less efficient rivals would be deemed a natural and
desirable consequence of the operation of the competitive process, not one that would trigger a prima
facie prohibition.
2.3.2. The appreciability of effects
A separate but related question concerns the appreciability of effects. If a threshold of appreciability
is introduced, then it would not be enough to show that a practice or transaction would have a negative
impact on competition (however this is defined). It would be necessary to establish, in addition, that
the impact in question is significant (or appreciable). Suppose that evidence of consumer harm is
required to establish effects. In such a case, the authority would need to show not only that prices
would go up following the practice or transaction, but also that the rise would be appreciable. Suppose
now that harm to the market structure is enough to establish effects. In the context of an exclusive
dealing agreement, for instance, it would be necessary to show not only that access to the market
would be foreclosed to new entrants, but that the agreements concluded by the supplier in question
make an appreciable contribution to rivals’ departure from the relevant market.
Competition law systems can address appreciability in a variety of ways. The approach seems
to depend, at least in part, on the meaning that is attached to the concept of effects. If, for instance, a
negative impact on competition is equated with consumer harm, an appreciable restriction is one that
11

leads to a price increase (or equivalent harm to other parameters) above a certain level (for instance
5%).18 An alternative approach to appreciability is one that revolves around the market power – or a
proxy for market power – enjoyed by the firm(s) involved in the practice or transaction. For instance,
one could safely assume that the effects of a practice implemented by a dominant firm (that is, a firm
with substantial market power19) will, if established, be appreciable. Conversely, practices and
transactions implemented by firms with a modest degree of market power, could be assumed (or at
least presumed) not to have appreciable effects. The choices made by the Court in this regard are
explained in Section 6.
2.4. The probability of the effects: plausibility, likelihood and certainty
The probability threshold has already been mentioned as a crucial factor in the definition of
anticompetitive effects. Competition law systems focus on certain practices and transactions that
typically have at least some probability of affecting competition. All the examples mentioned above
– including exclusivity obligations, tying, mergers and acquisitions – can, in certain circumstances,
have negative effects on rivals and/or consumers. Accordingly, if the requisite probability threshold
were set at a sufficiently low level, such practices would be found to have anticompetitive effects in
virtually every instance. For the same reason, the need to show anticompetitive effects would become
a formality and the distinction between practices prohibited ‘by object’ and ‘by effect’ would be
meaningless in practice. Conversely, the threshold of effects may be set at such a high level that it
becomes difficult to meet by an authority or claimant (which would make conduct de facto lawful).
This would be the case, in particular, if the authority or claimant were required to show that the
practice or transaction is certain to have a negative impact on competition.
18
For an example of the articulation of this approach, see US Department of Justice and Federal Trade Commission,
Horizontal Merger Guidelines (2010).
19
OECD, ‘Evidentiary Issues in Proving Dominance/Monopoly Power’ DAF/COMP(2006)35.
12

Certainty
Likelihood
Plausibility
Fig. 2: The probability of effects
The main thresholds of probability are depicted in Figure 2. At the low end of the spectrum,
one can identify a threshold of plausibility. The latter concept is understood to mean that a finding of
anticompetitive effects in the relevant economic and legal context would not be contrary to ‘logic and
experience’.20 Accordingly, this threshold is met as soon as one can identify a credible mechanism
through which the impact on competition can be manifested. At the higher end of the spectrum, one
can identify a threshold of certainty (that is, a 100% probability), or quasi certainty of such effects.
Between the two ends, one can identify a threshold of ‘likelihood’, which would be met where it can
be shown that the impact on competition is more likely than not to occur (that is, a level of probability
right above 50%). All three thresholds are relevant in the EU competition law system, as explained
in Section 6 below.
20
The expression is borrowed from Ioannis Lianos, ‘“Judging” Economists: Economic Expertise in Competition Law
Litigation’ in Ioannis Kokkoris and Ioannis Lianos (eds), The Reform of EC Competition Law: New Challenges (Kluwer
2010).
13

3. Anticompetitive effects in the case law: preliminary issues
3.1. Mechanisms through which anticompetitive effects are manifested
3.1.1. Collusion or absorption of a source competitive pressure
Anticompetitive effects, irrespective of how they are defined, may be manifested in a variety of ways.
The EU competition law system captures three mechanisms through which such effects may be
displayed. The first mechanism is the one that follows from collusion or the absorption of a source of
competitive pressure. The most straightforward example is that of a horizontal merger, which leads,
by its very nature, to the direct elimination of a rival.21 The negative effects resulting from this first
mechanism may be unilateral (that is, they may result from the increase in market power afforded by
the transaction to one or several market players) or coordinated (that is, they may flow from the
reduced incentive of the remaining market players to compete).22 Practices falling within the scope
of Article 101 TFEU can also display similar effects. An agreement between competitors softens the
competitive pressure faced by each of them; in addition, it may lead to collusive outcomes with other
market players.23 Some vertical arrangements can also lead to collusive outcomes.24
3.1.2. Exclusion of a source of competitive pressure
A second mechanism through which anticompetitive effects may be manifested is exclusion, that is,
the departure of one or several rivals from the market. Exclusionary outcomes may be caught by
21
Guidelines on the assessment of horizontal mergers (n 13) and Case 6/72 Europemballage Corporation and Continental
Can Company Inc. v Commission, EU:C:1973:22.
22
Guidelines on the assessment of horizontal mergers (n 13).
23
See in this sense the Guidelines on the applicability of Article 101 of the Treaty on the Functioning of the European
Union to horizontal co-operation agreements [2011] OJ C11/1.
24
See in this sense the Guidelines on vertical restraints [2010] OJ C130/1.
14

Article 101 TFEU, Article 102 TFEU and merger control. Just to mention an example, tying is a
concern insofar as it is a prima facie plausible mechanism through which competitors may be
foreclosed.25 To the extent that tying is implemented by means of an agreement between a supplier
and its customers, it can be caught by Article 101 TFEU; and, where the tie-in is imposed by a
dominant firm, by Article 102 TFEU. In the context of merger control – and, more precisely, where
the transaction gives rise to conglomerate concerns – the entity resulting from the transaction may
have the incentive to engage in tying behaviour, which may in turn have exclusionary effects.26
3.1.3. Exploitation of market power
Finally, anticompetitive effects may result from the exploitation by a dominant firm of its substantial
market power vis-a-vis customers and/or suppliers. As such, exploitative conduct can be caught by
Article 102 TFEU.27 In this case, anticompetitive effects are the consequence of the absence of
effective competitive constraints faced by the dominant firm. The anticompetitive behaviour in these
circumstances can be manifested in two main ways. First, exploitation may lead to the excessive
extraction of rents from operators elsewhere in the value chain. This extraction may be manifested,
for instance, in the dominant firm demanding unfairly high prices to its customers and, conversely,
in offering unfairly low prices to its suppliers.28 It may also result from imposing supplementary
obligations, unrelated to the main transaction, upon them.29
25
See in this sense the Guidance on the Commission’s enforcement priorities in applying Article 82 of the EC Treaty to
abusive exclusionary conduct by dominant undertakings [2009] OJ C45/7.
26
See in this sense the Guidelines on the assessment of non-horizontal mergers under the Council Regulation on the
control of concentrations between undertakings [2008] OJ C265/6.
27
This conclusion is something that directly follows from the letter of the Treaty. See in this sense René Joliet,
Monopolization and Abuse of Dominant Position: A Comparative Study of the American and European Approaches to
the Control of Economic Power (Martinus Nijhoff 1970).
28
Article 102(a) TFEU may consist in ‘directly or indirectly imposing unfair purchase or selling prices or other unfair
trading conditions’.
29
In accordance with Article 102(d) TFEU, an abuse may also consist in ‘making the conclusion of contracts subject to
acceptance by the other parties of supplementary obligations which, by their nature or according to commercial usage,
have no connection with the subject of such contracts’.
15

Second, exploitation may lead to the distortion of the conditions of competition on the markets
neighbouring the one in which the dominant firm operates. In other words, a dominant player may,
when exercising its substantial market power vis-à-vis customers and/or suppliers, favour some firms
at the expense of others (and thus inflict, at the very least, a competitive disadvantage on the latter).
For instance, a dominant supplier may adopt a policy of rebates that gives more favourable trading
conditions (in the form, for instance, of lower prices) to those customers that choose to deal
exclusively with it.30 In addition to the exclusionary effects that such a practice might display (what
have been termed primary-line effects in the literature), it may distort competition on the relevant
downstream market (secondary-line effects of an exploitative nature).31
3.2. The place of anticompetitive effects in the various legal tests
A practice or transaction that is incapable of having anticompetitive effects is not subject to EU
competition law.32 In that sense, one can claim that the notion is relevant across the board. This fact
does not mean, however, that it is necessary to establish anticompetitive effects, on a case-by-case
basis, in relation to every practice and transaction. Some practices – such as cartels – are deemed
prima facie unlawful irrespective of their impact. Anticompetitive effects are presumed to be an
inherent and inevitable consequence of the implementation of these practices; accordingly, their
impact need not be shown to exist in the economic and legal context of which they are a part.
Conversely, some practices are presumed compatible with Articles 101 and/or Article 102 TFEU;
30
Article 102(c) TFEU expressly provides that an abuse may consist in ‘applying dissimilar conditions to equivalent
transactions with other trading parties, thereby placing them at a competitive disadvantage’. This is an issue that
complemented the exclusionary concerns in several rebate cases, including Case 322/81 NV Nederlandsche Banden
Industrie Michelin v Commission, EU:C:1983:313 (‘Michelin I’); Case T-203/01 Manufacture française des
pneumatiques Michelin v Commission, EU:T:2003:250 (‘Michelin II’); and British Airways (n 9).
31
For an analysis of the question, see Opinion of AG Wahl in Case C-525/16 MEO – Serviços de Comunicações e
Multimédia SA v Autoridade da Concorrência, EU:C:2017:1020; and Damien Geradin and Nicolas Petit, ‘Price
Discrimination under EC Competition Law: Another Antitrust Doctrine in Search of Limiting Principles?’ (2006) 2
Journal of Competition Law & Economics 479.
32
See in particular T-Mobile (n 2), para 31 and Case C-307/18, Generics (UK) Ltd and others v Competition and Markets
Authority, EU:C:2020:52, para 154.
16

thus, they are deemed prima facie lawful without it being necessary to carry out an assessment of
their effects. Finally, there are practices and transactions that only give rise to intervention where
their impact on competition is established in the specific context in which they are implemented. For
instance, mergers scrutinised under Regulation 139/2004 can only be declared to be incompatible
with the internal market following a case-by-case assessment of their effects.
The distinction between practices that are prima facie unlawful irrespective of their impact –
on the one hand – and practices that are only prohibited if they are shown to have such effects – on
the other – is particularly apparent in the context of Article 101(1) TFEU. This provision distinguishes
between agreements that restrict competition by object and by effect. 33 Evidence of the impact on
competition is only required in relation to the latter. Thus, once a restriction by object is established,
it is not necessary to show, in addition, that the agreement in question has anticompetitive effects.34
However, this same divide can be identified in the context of Article 102 TFEU. 35 Some conduct is
prima facie prohibited as abusive without it being necessary to evaluate, on a case-by-case basis, its
impact on competition. This is the case, in particular, of tying,36 of pricing below average variable
costs,37 and of exclusive dealing (as well as loyalty and ‘loyalty-inducing’ target rebates).38 Other
practices, including ‘margin squeeze’ behaviour39 and standardised rebate schemes,40 are only caught
by Article 102 TFEU insofar as they have anticompetitive effects.
33
Article 101(1) TFEU refers to practices that ‘have as their object or effect the prevention, restriction or distortion of
competition within the internal market’.
34
The consistent line of case law on this point dates back to Joined Cases 56/64 and 58/64 Établissements Consten S.à.R.L.
and Grundig-Verkaufs-GmbH v Commission, EU:C:1966:41.
35
See in this sense Pablo Ibáñez Colomo, ‘Beyond the “more economics-based approach”: a legal perspective on Article
102 TFEU case law’ (2016) 53 Common Market Law Review 709.
36
Case T-30/89 Hilti AG v Commission, EU:T:1991:70; and Case T-83/91 Tetra Pak International SA v Commission,
EU:T:1994:246 (‘Tetra Pak II’).
37
Case C-62/86 AKZO Chemie BV v Commission, EU:C:1991:286, para 71.
38
Case 85/76 Hoffmann-La Roche & Co. AG v Commission, EU:C:1979:36, para 89; confirmed in Case C-413/14 P Intel
Corporation Inc v Commission, EU:C:2017:632, para 137. See also Michelin I (n 30) and British Airways (n 9).
39
Case C-280/08 P Deutsche Telekom AG v Commission, EU:C:2010:603.
40
Post Danmark II (n 9) and TeliaSonera (n 9).
17

3.2.1. Anticompetitive effects and prima facie unlawful conduct
When a practice is deemed prima facie unlawful irrespective of its effects (both under Articles 101
and 102 TFEU), it is presumed to be at least ‘capable’ of having such effects.41 In other words, its
negative impact on competition is deemed to exist and need not be established on a case-by-case
basis. The Court has been explicit about the presumption of effects that underpins prima facie
unlawful conduct. In the context of Article 101(1) TFEU, for instance, it has explained that price-
fixing by cartels (which amounts to a ‘by object’ infringement) is known from experience to be a
source of allocative inefficiency.42 The same is true in the context of Article 102 TFEU. Underpinning
the prima facie prohibition of exclusive dealing and loyalty rebates, for instance, there is a
presumption that these practices are capable of restricting competition.43
As far as prima facie unlawful conduct is concerned, arguments about the absence of
anticompetitive effects may be advanced to escape the prohibition. In other words, the underlying
presumption is open to rebuttal by a defendant. Thus, it is possible for the parties to an agreement, or
for a dominant firm, to show that the practice is incapable of having a restrictive impact on
competition in the relevant economic and legal context. That the parties can show that an agreement
is incapable of restricting competition was expressly recognised, in the context of Article 101(1)
TFEU, in Murphy.44 In relation to Article 102 TFEU, this point was made explicit in Intel.45 The two
judgments clarify that defendants may bring arguments pertaining to the nature of the practice and
the relevant economic and legal context. These factors are examined at length below.
41
T-Mobile (n 2), para 31.
42
Cartes Bancaires (n 4) para 51; confirmed in Maxima Latvija (n 8), para 19; and Case C-228/18 Gazdasági
Versenyhivatal v Budapest Bank Nyrt. and others, EU:C:2020:265, para 36.
43
For a discussion on presumptions generally, and on this presumption in particular, see Chapter 6 in Andriani Kalintiri,
Evidence Standards in EU Competition Enforcement (Hart Publishing 2019).
44
Joined Cases C-403/08 and C-429/08 Football Association Premier League Ltd and Others v QC Leisure and Others
and Karen Murphy v Media Protection Services Ltd, EU:C:2011:631 (‘Murphy’), para 140.
45
Intel (n 38), paras 138-140.
18

3.2.2. The case-by-case assessment of anticompetitive effects: ‘standard effects’ and ‘enhanced
effects’ tests
The case-by-case assessment of the impact of a practice is the rule in the EU competition law system.
As already pointed out, the compatibility of all mergers within the meaning of Regulation 139/2004
is evaluated on the basis of a case-specific inquiry. In the context of Article 101(1) TFEU, the Court
has held that the ‘by object’ category is to be interpreted restrictively.46 The analysis of effects can
be undertaken in light of two sets of legal tests.47 The default approach is what can be termed the
‘standard effects’ test, which requires an evaluation of the impact of the practice (or transaction) on
competition in the relevant economic and legal context. There are two additional sub-tests, which
depart from the default test and can be termed ‘enhanced effects’ analysis. They apply in exceptional
circumstances. Under them, it is necessary to satisfy additional obligations. At the very least, an
authority or claimant would need to show, first, that an input or platform is indispensable for
competition on a neighbouring market and, second, that lack of access to the said input or platform
would lead to the elimination of all competition on that market.48
3.2.3. Anticompetitive effects and prima facie lawful conduct
Some practices are prima facie lawful in the sense that they are presumed to be compatible with
Article 101(1) and/or 102 TFEU without it being necessary to evaluate, case-by-case, their effects.
For instance, selective distribution agreements are deemed compatible with Article 101(1) TFEU
46
Cartes Bancaires (n 4) para 58; and Budapest Bank (n 42), para 54.
47
For an extensive analysis of the question, see Pablo Ibáñez Colomo, ‘Legal Tests in EU Competition Law: Taxonomy
and Operation’ (2019) 10 Journal of European Competition Law & Practice 424.
48
See in this sense Joined Cases C-241/91 P and C-242/91 P Radio Telefis Eireann (RTE) and Independent Television
Publications Ltd (ITP) v Commission, EU:C:1995:98 (‘Magill’); Case C-7/97 Oscar Bronner GmbH & Co. KG v
Mediaprint Zeitungs- und Zeitschriftenverlag GmbH & Co. KG and others, EU:C:1998:569; and Case C-418/01 IMS
Health GmbH & Co. OHG v NDC Health GmbH & Co. KG, EU:C:2004:257.
19

where the conditions set out by the Court in Metro I are fulfilled.49 The same is true of the restraints
found in franchising agreements and which are necessary to preserve the know-how of the franchisor
and the uniformity and reputation of its formula.50 Concerning potentially abusive conduct, above-
cost pricing is prima facie compatible with Article 102 TFEU. Quantity rebates (that is, rebates that
are incremental in nature and transaction-specific) are also prima facie lawful, and this insofar as they
are deemed to reflect the cost savings made by the supplier.51
It is possible for an authority or claimant to rebut the presumption of legality by showing that
prima facie lawful conduct has or would have anticompetitive effects in a given economic and legal
context. For instance, a selective distribution system may have a negative impact on competition
where a network of similar agreements leads to price rigidities, forecloses access to the market and/or
precludes other forms of distribution.52 Similarly, there may be instances, explored in detail below,
where quantity rebates and above-cost prices amount to an abuse of a dominant position. For instance,
an authority or claimant may be able to show that a given rebate scheme does not truly reflect the cost
savings made by the supplier and is designed to foreclose competition.53
4. The time dimension in the case law: actual and potential effects
4.1. The prospective analysis of potential effects
The Court has consistently held that EU competition law is concerned with both actual and potential
effects.54 Accordingly, the fact that an authority or claimant has failed to established the actual impact
49
Case 26/76 Metro SB-Großmärkte GmbH & Co. KG v Commission, EU:C:1977:167 (‘Metro I’).
50
Case 161/84 Pronuptia de Paris GmbH v Pronuptia de Paris Irmgard Schillgallis, EU:C:1986:41.
51
Hoffmann-La Roche (n 38), para 90; and Post Danmark II (n 9), para 28.
52
Case 75/84 Metro SB-Großmärkte GmbH & Co. KG v Commission, EU:C:1986:399 (‘Metro II’), para 40.
53
See for instance Case C-163/99 Portugal v Commission, EU:C:2001:189, paras 52-53.
54
See in particular Joined Cases 142 and 156/84 British-American Tobacco Company Ltd and R. J. Reynolds Industries
Inc. v Commission, EU:C:1987:490, para 39; Case C-7/95, John Deere Ltd v Commission, EU:C:1998:256, para 77;
Joined Cases C-215/96 and C-216/96 Carlo Bagnasco and Others v Banca Popolare di Novara soc. coop. arl. and Cassa
20

of a practice does not in itself mean that a restriction has not been established to the requisite legal
standard.55 That potential effects are sufficient to trigger intervention is self-evident in the context of
merger control, which is by definition based on the ex ante evaluation of concentrations. It is also
true in relation to Articles 101 and 102 TFEU. According to the case law, the analysis of potential
effects is relevant where the analysis is prospective in nature. Thus, it may come into play in the two
circumstances identified above: where the practice has not yet been implemented and where the
practice has already been implemented at the time of the adoption of the decision and the analysis
considers the future evolution of the relevant market. These two scenarios were considered by the
Court in British American Tobacco56 and are depicted in Figures 3 and 4.
Adoption of the
decision Court review
Analysis of potential effects
Fig. 3: The analysis of potential effects (I)
Implementation of Adoption of the

the practice Court review
decision
Analysis of actual effects Analysis of potential effects

Fig. 4: The analysis of potential effects (II)
Figures 3 and 4 show, in addition, that there is typically a lag between the adoption of a
decision by a competition authority and the review of its legality by a review court. As a result, by
the time the court reaches its decision, it is possible to assess whether the potential effects identified
by the authority are confirmed by subsequent market developments. These developments may reveal
di Risparmio di Genova e Imperia SpA, EU:C:1999:12, para 34; TeliaSonera (n 9), para 64; and Post Danmark II (n 9),
para 66.
55
John Deere (n 54), para 78.
56
British American Tobacco (n 54).
21

that the forecasts made by the authority failed to materialise. For instance, the assessment might have
been based on the premise that the barriers to entry on the relevant market and/or the practices would
have prevented the arrival or growth of rivals,57 and this premise may have been subsequently shown
to be false; alternatively, the assessment may have been based on the premise that network effects
would result in the market ‘tipping’ in favour of the firm involved in the practice.58
The question, in this regard, is whether the evaluation of the legality of the decision (and, in
particular, of whether potential effects have been established to the requisite legal standard) can be
informed by the actual developments that follow its adoption. According to the case law, the legality
of a Commission decision must be assessed on the basis of the evidence that is available when
adopted.59 Thus, the review court would not be able to consider the subsequent evolution of the market
in its assessment. The General Court (hereinafter, the ‘GC’) addressed this point in Microsoft I.60 The
fact that the theories underpinning the analysis of effects did not unfold as predicted by the
Commission was not considered.61 For the same reason, the fact that the anticompetitive effects failed
to materialise is not necessarily conclusive.62 Whether or not anticompetitive effects are established
to the requisite legal standard in such a scenario depends on the relevant probability threshold, which
is discussed below in Section 6.
4.2. The retrospective analysis of actual effects
The EU courts have addressed the question of whether the retrospective analysis can be based on
potential effects alone. The issue, in other words, is whether evidence shedding light on the evolution
of markets following the implementation of the practice (but before the adoption of the decision) can
57
See for instance Michelin II (n 30), paras 235-246.
58
See in this sense Case T-201/04 Microsoft Corp v Commission, EU:T:2007:289 (‘Microsoft I’).
59
Joined Cases 15 and 16/76 France v Commission, EU:C:1979:29..
60
Microsoft I (n 58), para 260.
61
Ibid, para 943.
62
Ibid, paras 560-564.
22

be disregarded by a claimant or authority. The issue has been addressed by the Court only indirectly,
and this when enunciating the general principles and when providing guidance to national courts on
the analysis of effects. According to a consistent line of case law, the impact of a practice on
competition is to be assessed in the ‘actual context’ in which it is implemented.63 This principle can
be interpreted as suggesting that the observable market developments at the time of the
implementation of the practice cannot be ignored.
This idea is confirmed when one examines the guidance provided to national courts. In Post
Danmark I, for instance, it explained that, absent evidence of an exclusionary purpose, prices that
cover the bulk of the costs attributable to a good or service are, ‘as a general rule’, not abusive (and
this even when they fall below average total costs).64 In such circumstances, the practice would only
be caught by Article 102 TFEU if it is shown to have anticompetitive effects.65 In its ruling, the Court
engaged in a retrospective analysis of the impact of the conduct and noted that the dominant firm’s
rival had not been driven out of the market and had, in fact, been able to maintain its distribution
network and win two customers back.66 Evidence in this sense was deemed to constitute a strong
indicator that the practice did not display exclusionary effects.
The GC faced the question directly when it considered the legality of the Commission decision
in Servier.67 In its Krka judgment, it held that the retrospective analysis of the anticompetitive effects
of a practice cannot ignore the actual developments that follow its implementation but that precede
the adoption of a decision. Thus, it concluded that the Commission had erred when claiming that,
when examining the impact of an agreement concluded between an actual and a potential competitor,
it can discharge its legal burden merely by considering hypotheticals about the evolution of markets
63
See in particular Case 56/65 Société Technique Minière, EU:C:1966:38, 249; John Deere (n 54), para 76; Case C-
238/05, Asnef-Equifax, Servicios de Información sobre Solvencia y Crédito, SL and Administración del Estado v
Asociación de Usuarios de Servicios Bancarios, EU:C:2006:734, para 49; Generics (n 32), para 116.
64
Post Danmark I (n 8), para 38.
65
Ibid, para 39.
66
Ibid.
67
Perindopril (Servier) (Case AT.39612) Commission Decision of 9 July 2014. See also Case T-691/14 Servier SAS and
others v Commission, EU:T:2018:922; and Case T-684/14 Krka Tovarna Zdravil d.d. v Commission, EU:T:2018:918.
23

– that is, without considering the actual context in which the agreement in question displayed its
effects.68 In the same vein, it held that the case law discussed in the preceding section – which
concerned the prospective analysis of potential effects – is not applicable to instances in which it is
retrospective.69 The approach embraced in Krka, which is in turn consistent with the case law, reflects
the approaches captured in Figures 3 and 4.
5. The meaning of competition and the counterfactual in the case law
5.1. Competition comprises both the inter-brand and intra-brand dimensions
In Consten-Grundig, the Court unambiguously held that the EU competition law system (and the
‘principle of freedom of competition’) is not only concerned with inter-brand competition, but also
with intra-brand restrictions.70 Accordingly, one cannot rule out, from the outset, that conduct
restraining intra-brand rivalry (such as a distribution agreement providing for territorial protection)
amounts to a breach of Articles 101 and/or 102 TFEU merely because it might increase inter-brand
competition.71 Following Consten-Grundig, there has been a consistent line of enforcement against
practices with the object and/or effect of restricting intra-brand competition alone. Of these, restraints
aimed at limiting cross-border trade represent the largest fraction.72 Other practices include resale
price maintenance73 and selective distribution.74
68
Krka (n 67), paras 317-344.
69
70
Consten-Grundig (n 34), 342.
71
Ibid.
72
These cases include Glaxo Spain (n 2) and Murphy (n 44).
73
Case 243/83 SA Binon & Cie v SA Agence et messageries de la presse, EU:C:1985:284 and Pronuptia (n 50).
74
Metro I (n 49); Metro II (n 52); and Coty (n 16).
24

5.2. Competition refers to such lawful competition which would otherwise have existed
The Court has consistently held, from the early days, that competition, for the purposes of the Treaty,
must be understood as referring to such competition which would have existed in the absence of the
practice or transaction.75 Accordingly, anticompetitive effects need to be established against the
relevant counterfactual, or, as the Court has consistently held, ‘within the actual context in which it
would occur in the absence of the [practice or transaction] in dispute’.76 This is true across the board.
In this sense, the meaning of competition does not change depending on whether the assessment under
Article 101(1) TFEU relates to the object of a practice or to its effects. Similarly, it does not change
depending on whether it is a transaction examined under the Regulation 139/2004 or a practice subject
to Articles 101 and/or 102 TFEU. The need to consider the counterfactual has two main implications.
First, it is necessary to establish a causal link between the practice or transaction and any actual or
potential effects. Second, a practice that is necessary to attain a pro-competitive aim is not restrictive
of competition, whether by object or as a result of its impact on competition. These two implications
are examined in turn.
5.2.1. There must be a causal link between the practice or transaction and any effects
The Court has expressly held that there must be a causal link between a practice or transaction, on
the one hand, and any actual or potential effects, on the other. Accordingly, where the observed or
expected impact (or, more generally, the absence of competition) would have occurred in its absence,
the practice or transaction is not restrictive of competition (whether by object or effect). The need to
establish a causal link was made explicit, in the context of merger control, in Kali & Salz.77 In relation
75
Société Technique Minière (n 63), 250.
76
See above, n 63.
77
Joined cases C-68/94 and C-30/95 France and others v Commission, EU:C:1998:148 (‘Kali & Salz’).
25

to Article 102 TFEU, the Court held, in Post Danmark II, that any anticompetitive effects must not
be purely ‘hypothetical’ and must be ‘attributable’ to the dominant firm.78 These judgments are in
line with a consistent line of case law, which under Article 101 TFEU, makes it necessary to establish
the restriction against the relevant counterfactual.79
The evaluation of the causal link between the practice or transaction and its effects is to be
undertaken in light of the economic and legal context of which it is a part. The analysis may reveal
the absence of a causal link, and this for a variety of reasons – including the very nature of the relevant
market and its evolution. Two examples illustrate the idea. To begin with, the absence of competition
may not be attributable to the practice or transaction, but to the regulatory context, which may
preclude any inter-brand and/or intra-brand rivalry. In relation to Article 101(1) TFEU, for instance,
the Court has held that an agreement would not be capable of having anticompetitive effects where
there are ‘insurmountable’ barriers to entry.80 Such insurmountable barriers may exist, for instance,
where there is an intellectual property right.81 They may also exist where the regulatory framework,
taken together, amounts to a de facto monopoly precluding entry.82
A causal link would also fail to exist, in a scenario of exclusion (or exploitation), where the
deterioration of the conditions of competition can be attributed to the inefficiency of the firms that
are or may be driven out of the market. In line with what has been explained above, the Court declared,
in Post Danmark I, that the exclusion of those firms that are ‘less attractive to consumers from the
point of view of, among other things, price, choice, quality or innovation’ does not give rise to
anticompetitive effects under Article 102 TFEU.83 The departure of such firms from the market can
78
Post Danmark II (n 9), para 47 and 65.
79
See in particular Case C‑382/12 P MasterCard Inc. and Others v Commission, EU:C:2014:2201, para 161; Generics (n
32), paras 103-122; and Budapest Bank (n 42), paras 55, 82 and 83.
80
C‑373/14 P Toshiba Corporation v Commission, EU:C:2016:26, paras 31-34; and Generics (n 32), para 45.
81
See for instance Case 262/81 Coditel SA, Compagnie générale pour la diffusion de la télévision, and others v Ciné-Vog
Films SA and others, EU:C:1982:334 (‘Coditel II’); Case C-9/93 IHT Internationale Heiztechnik GmbH and Uwe
Danzinger v Ideal-Standard GmbH and Wabco Standard GmbH, EU:C:1994:261; and Generics (n 32), para 45.
82
Case T-360/09, E.ON Ruhrgas AG and E.ON AG v Commission, para 104. See also Case T-370/09, GDF Suez SA v
Commission, para 97.
83
26

be expected to occur irrespective of the implementation of the practice.84 For the same reason, above-
cost pricing and quantity rebates (as defined by the Court in Post Danmark II) are deemed prima facie
lawful. It is reasonable to presume that an equally efficient rival would not be excluded as a result of
such practices. In the context of merger control, the Court embraced the same idea by accepting, as a
matter of principle, a variation of the ‘failing firm defence’.85
5.2.2. A practice that is objectively necessary is not restrictive of competition
The evaluation of the counterfactual may reveal not only that a practice does not lead to a deterioration
of the conditions of competition, but that it actually improves them. If the assessment reveals that the
practice in question is objectively necessary to attain a legitimate or pro-competitive aim, it will not
amount to a restriction of competition, whether by object or effect. There is a long line of Article
101(1) TFEU case law expressly addressing this point, starting with Société Technique Minière.86 An
agreement may prove to be objectively necessary for a variety of reasons. It may be the case, for
instance, that a licensee would not have undertaken the necessary investments to produce and market
the contractual goods in the absence of the territorial restraints provided for by virtue of the
agreement.87 It may also be the case that the parties have complementary capabilities that allow them
to engage in a project that they would not have been able to pursue individually.88
A variation of the objective necessity test is the so-called ancillary restraints doctrine, which
has been recognised as such by the Court. Under the doctrine, the question is whether the agreement
would have been concluded in the absence of some clauses. The question, in other words, is whether
84
See in this sense Luc Peeperkorn and Vincent Verouden, ‘The Economics of Competition’, in Jonathan Faull and Ali
Nikpay (eds), The EU Law of Competition (3rd edn, Oxford University Press 2014), 1.43-1.49.
85
Kali & Salz (n 77), para 115.
86
Société Technique Minière (n 63), 250: ‘[...] [I]t may be doubted whether there is an interference with competition if
the said agreement seems really necessary for the penetration of a new area by an undertaking [...]’; Generics (n 32), paras
103-111; and Budapest Bank (n 42), paras 82-83.
87
Case 268/78 L.C. Nungesser KG and Kurt Eisele v Commission, EU:C:1982:211, paras 55-58.
88
Joined Cases T-374/94, T-375/94, T-384/94 and T-388/94 European Night Services Ltd v Commission, EU:T:1998:198,
para 145.
27

the clauses are objectively necessary for the agreement to exist in the first place. For instance, a non-
compete obligation may be necessary for the buyer to agree to the acquisition of a business.89 To the
extent that it is, it would not restrict competition, whether by object or effect. Similarly, the Court
explained in Pronuptia that a firm would not be willing to engage in a franchising agreement if it is
unable to preserve its know-how and the uniformity and reputation of its formula.90 Accordingly, the
clauses that are objectively necessary to achieve these aims fall outside the scope of Article 101(1)
TFEU altogether.
The case law addressing objective necessity revolves around the application of Article 101(1)
TFEU. A remaining question, accordingly, is whether the same principles apply in the context of
Article 102 TFEU. If they did, a dominant firm would be able to avoid a prima facie finding of abuse
on objective necessity grounds. Thus, the practice would fall outside the scope of the prohibition since
it would not have anticompetitive effects (the conditions of competition would have improved, not
deteriorated, following the implementation of the practice). The possibility to invoke objective
necessity would be distinct from, and complementary to, the possibility for the dominant firm to
provide an objective justification and/or to show that the efficiencies to which the practice gives rise
outweigh any restrictive effects. Unlike objective necessity, the latter two would only come into play
following the prima facie finding of an abuse.
It would be reasonable to assume that the objective necessity test is also relevant in the context
of Article 102 TFEU. This is so, first, because there are express references to objective necessity in
the case law.91 Second, consistency would demand that the same principles apply across the
competition law system. It would not be obvious to justify why objective necessity would apply only
in relation to Article 101(1) TFEU. Such an approach would amount to attaching different meanings
to the concept of competition depending on the applicable provision, which is not an easily tenable
89
Case 42/84 Remia BV and others v Commission, EU:C:1985:327, para 19.
90
Pronuptia (n 50), paras 16-17.
91
Case 311/84 Centre belge d'études de marché - Télémarketing v SA Compagnie luxembourgeoise de télédiffusion and
Information publicité Benelux, EU:C:1985:394, para 27.
28

position. One should consider, in this regard, that there are instances in which the same practice can
be subject to both Articles 101 and 102 TFEU.92 It is difficult to see how the same conduct could be
simultaneously found to be objectively necessary (and thus pro-competitive and incapable of having
restrictive effects) under Article 101(1) TFEU and prima facie abusive (and thus capable of having
restrictive effects) under Article 102 TFEU.
5.3. Both the ex ante and ex post dimensions of the counterfactual are considered
5.3.1. The role of ex ante considerations in the analysis of effects
An important conclusion that follows from the interpretation and application of the counterfactual by
the Court is that both its ex ante and ex post dimensions are considered. The case law shows that the
pro-competitive gains resulting from a practice cannot simply be assumed to have existed. For the
same reasons, any observable ex post restrictions cannot be examined in isolation. There are abundant
examples in the case law showing how ex ante considerations play a role in the analysis. In Nungesser,
the Court deemed justified the concerns expressed by the interveners in the case, who explained that
the relevant technology (and the resulting innovation) was the outcome of ‘years of research and
experimentation’.93 The assessment of any ex post restraints (which in the case gave territorial
protection to the licensee), accordingly, would have to pay due regard to the ‘nature of the product’.94
The cases spelling out the ancillary restraints doctrine further illustrate how the ex ante
dimension of the counterfactual plays a role in the analysis. In Pronuptia, the Court noted that
franchising agreements lead to pro-competitive gains by allowing the franchisor to benefit from its
formula without investing its own capital, on the one hand; and by giving access to franchisees to
92
Joined cases C-395/96 P and C-396/96 P Compagnie Maritime Belge Transports SA v Commission, EU:C:2000:132,
para 33; and Generics (n 32), para 146.
93
Nungesser (n 87), para 56.
94
Ibid, para 58.
29

methods they would only have been able to obtain after ‘considerable efforts’ and thus benefit from
the franchisee’s goodwill, on the other.95 Accordingly, some ex post restraints resulting from the
franchising agreement (such as non-compete obligations or limitations relating to the outlets from
which the contractual goods may be sold) are inextricably linked to these ex ante pro-competitive
gains. To the extent that they are, they cannot be dissociated from them.
Beyond the ancillary restraints doctrine, the case law on refusals to deal captures effectively
the role that ex ante considerations play in the analysis. Seen ex post, any refusal by a vertically-
integrated company to deal with a would-be rival has anticompetitive effects. However, the Court has
consistently held that such refusals are only abusive in exceptional circumstances.96 The stringent,
‘enhanced effects’ test laid down in Magill, Bronner and IMS Health can only be understood if the
ex ante dimension of the counterfactual is taken into account. The ex ante factors that confine to
exceptional circumstances the instances in which a refusal amounts to an abuse were discussed by
AG Jacobs in Bronner. The Opinion emphasises that imposing an access obligation too readily would
harm firms’ ex ante incentives to invest and innovate.97 Thus, an exclusive focus on the observable
ex post restraints would disregard the pro-competitive gains resulting from the very same restraints
and lead to less, not more, competition.
5.3.2. How ex ante considerations are incorporated in the analysis
The examples above are useful not only to show how ex ante considerations may lead to the
conclusion that a practice is incapable of having anticompetitive effects, but also the different ways
in which they may be incorporated in the analysis. In some circumstances, ex ante considerations are
evaluated on a case-by-case basis in light of the relevant economic and legal context. Such was the
95
Pronuptia (n 50), para 15.
96
See in this sense Magill (n 48), Bronner (n 48) and IMS Health (n 48).
97
Opinion of AG Jacobs in Case C-7/97 Oscar Bronner GmbH & Co. KG v Mediaprint Zeitungs- und Zeitschriftenverlag
GmbH & Co. KG and others, EU:C:1998:264.
30

case, for instance, in Nungesser. In these instances, the firm(s) involved in the practice may provide
evidence showing that the effects of any ex post restraints cannot be dissociated from the ex ante
gains resulting from it – and, by the same token, why, by failing to do so, an authority may have erred
in law. For instance, the parties may provide ‘serious indicia’98 showing why, in the specific
circumstances of a case, the benefits resulting from an exclusive distribution agreement would not
have been achieved without the restriction of active and passive sales.99
There are other circumstances where ex ante considerations are incorporated in the legal test
itself. In fact, the ‘enhanced effects’ and prima facie lawfulness tests, described above, are precisely
designed to take into account the ex ante dimension of the counterfactual. These legal tests seek to
capture the presumption that any ex post restraints are inextricably linked to the pro-competitive gains
resulting from the practice. This is obvious, as explained above, in relation to ancillary restraints. By
making them prima facie lawful, the Court acknowledges that the agreements of which they are a part
(and the resulting pro-competitive gains) would not take place without some ex post restraints.
Similarly, the ‘enhanced effects’ test (and, in particular, the indispensability condition) is crafted to
preserve firms’ ex ante incentives to invest and innovate (and, by doing so, the pro-competitive gains
resulting from them).
6. The meaning of effects in the case law
6.1. Only appreciable effects are relevant under EU competition law
Only appreciable effects are relevant in the EU competition law system. As the Court declared in
Völk, practices with an ‘insignificant effect’ on competition (de minimis) fall outside the scope of
98
Budapest Bank (n 42), paras 82-83.
99
Guidelines on vertical restraints (n 24), para 61.
31

Article 101 TFEU.100 Appreciability is also enshrined in the merger control regime. In accordance
with Article 2 of Regulation 139/2004, only transactions that would lead to a significant impediment
to effective competition can be declared to be incompatible with the internal market. 101 It seems clear
from Völk that the appreciability of the restrictive effects is assessed by reference to the market power
enjoyed by the firm(s) involved – or, to use the Court’s expression, their weak or strong position on
the relevant market(s) affected by the practice or transaction.
In principle, the assessment of the market power of a firm requires a case-by-case evaluation.
However, both the Court and the Commission have resorted to proxies that dispense from a context-
specific inquiry. In Expedia, the ECJ held that, where an agreement is found to restrict competition
by object, the fact that it is capable of affecting trade between Member States is sufficient to conclude
that its effect on competition is appreciable within the meaning of Völk.102 For practices that are not
prima facie unlawful irrespective of their effects, market shares tend to be used as a proxy. In
accordance with the Preamble to Regulation 139/2004, the fact that the market share of the parties
does not exceed 25%, is an indicator that the transaction is not liable to have significant effects.103 In
the context of Article 101 TFEU, the Commission has published, over the years, several versions of
its De Minimis Notice, which also revolve around market shares.104 This is also the technique used
in the Guidelines on vertical restraints,105 on horizontal co-operation agreements106 and on technology
licensing107 provide, which shed additional light in this regard.
A logical corollary to the above is that the de minimis doctrine has no role to play in abuse of
dominance cases. Given that appreciability depends on the market power enjoyed by the firm(s)
100
Case 5/69 Franz Völk v S.P.R.L. Ets J. Vervaecke, EU:C:1969:35.
101
See Article 2 of Council Regulation (EC) No 139/2004 of 20 January 2004 on the control of concentrations between
undertakings [2004] OJ L24/1. See also Case T-399/16 CK Telecoms UK Investments Ltd v Commission, EU:T:2020:217.
102
Case C-226/11 Expedia Inc. v Autorité de la concurrence and others, EU:C:2012:795.
103
Regulation 139/2004 (n 101), Recital 32.
104
Notice on agreements of minor importance which do not appreciably restrict competition under Article 101(1) of the
Treaty on the Functioning of the European Union [2014] OJ C291/1.
105
Guidelines on vertical restraints (n 24).
106
Guidelines on horizontal co-operation agreements (n 23).
107
Guidelines on the application of Article 101 of the Treaty on the Functioning of the European Union to technology
transfer agreements [2014] OJ C89/3.
32

involved in a practice or transaction, and given that a precondition for the application of Article 102
TFEU is a finding of a position of substantial market power, any anticompetitive effects – provided
that they are established to the requisite legal standard – are appreciable by definition. Against this
background, it is only logical that the Court, in Post Danmark II, ruled that it is not appropriate to set
a de minimis or appreciability threshold in abuse cases.108 When implemented by a firm with
substantial market power, ‘any further weakening of the structure of competition may constitute an
abuse of a dominant position’.109 Accordingly, it is not necessary for a claimant or authority to show
that the effects, once shown, are appreciable.
A key question that has not yet been addressed by the Court relates to the requisite level of
appreciability. As the law stands, it is possible to gain an idea of the practices or transactions that are
unlikely to yield appreciable effects. There is no dispute that the effects of a practice implemented by
a dominant firm will, if established, be appreciable. However, the Court has not identified (directly
or by proxy) the degree of market power above which the impact of a practice or transaction on
competition becomes appreciable. It would be reasonable to assume that the threshold of
appreciability is below the level of dominance.110 However, it is unclear where it lies and, by the same
token, how far below dominance it is. There is, in other words, a grey area where effects may or may
not be appreciable depending on the circumstances of the case. Crucially, the EU courts have clarified,
in this regard, that the issue of appreciability is to be assessed on a case-by-case basis. Thus, it would
not be sufficient for an authority or claimant to show that the parties exceed the market share
thresholds as defined in the De Minimis Notice.111
108
Post Danmark II (n 9), para 73.
109
Ibid, para 72.
110
See in this sense the Guidelines on the application of Article 81(3) of the Treaty [2004] OJ C101/97, para 26.
111
Case T-9/93 Schöller Lebensmittel GmbH & Co. KG v Commission, EU:T:1995:99, para 75. See also Expedia (n 102).
33

6.2. An effect is more than a competitive disadvantage or a limitation of a firm’s freedom of
action
From the early days, the Court has held, directly and indirectly, that a limitation of a firm’s freedom
of action does not lead, in and of itself, to anticompetitive effects. Similarly, the fact that one or
several firms are placed at a disadvantage cannot be equated with a negative impact on competition.
These questions were first addressed in relation to the interpretation of Article 101(1) TFEU. 112 The
essence of this consistent line of case law is captured in the rulings dealing with the status of exclusive
dealing. Single-branding restrains, by definition, the freedom of action of the buyer (which is
precluded from selling products competing with the supplier’s); in addition, it would inflict a
competitive disadvantage on the buyer where other outlets are entitled to sell several brands of a
product. In spite of these facts, the Court ruled that exclusive dealing is not restrictive by its very
nature, and that its effects need to be assessed by looking at the impact of the agreement on the
relevant market.113 This position has been held across a broad range of practices that limit a firm’s
freedom of action and/or inflict a competitive disadvantage; including, just to mention some
examples, non-compete obligations,114 non-challenge clauses in a licensing agreement115 or bans on
the use of online marketplaces.116
In the context of merger control, it has never been seriously disputed that limiting a firm’s
freedom of action and/or placing firms at a competitive disadvantage is as such insufficient to
establish anticompetitive effects. This point is exemplified by conglomerate mergers. A conglomerate
transaction involving firms in neighbouring markets typically gives the merged entity a competitive
112
See in particular Case 23/67 SA Brasserie de Haecht v Consorts Wilkin-Janssen, EU:C:1967:54; Delimitis (n 15); Case
C-250/92 Gøttrup-Klim e.a. Grovvareforeninger v Dansk Landbrugs Grovvareselskab AmbA, EU:C:1994:413, paras 28-
34; Case C-309/99 JCJ Wouters, JW Savelbergh and Price Waterhouse Belastingadviseurs BV v Algemene Raad van de
Nederlandse Orde van Advocaten, EU:C:2002:98, para. 97.
113
Delimitis (n 15), paras 10-13.
114
Remia (n 89).
115
Case 65/86 Bayer AG and Maschinenfabrik Hennecke GmbH v Heinz Süllhöfer, EU:C:1988:448.
116
Coty (n 16).
34

advantage over rivals operating only on one of the markets. If there is an overlap between the
customers of both products, this competitive advantage would allow the entity to offer a wider
portfolio of products to its customers. The entity may also have the ability and the incentive to engage
in tying and bundling. As confirmed in Tetra Laval, such consequences are, in and of themselves,
insufficient to show that the transaction would have anticompetitive effects.117 These principles are
now encapsulated in the Non-Horizontal Merger Guidelines issued by the Commission.118
The evolution in the interpretation of Article 102 TFEU also leads to the conclusion that the
anticompetitive effects of the practice cannot merely be inferred from a competitive disadvantage
and/or a limitation of a firm’s freedom of action. The Court has consistently held that, in order to
establish anticompetitive effects (and, more precisely, exclusionary effects), it is necessary to
ascertain whether a practice makes entry ‘more difficult, or impossible’.119 The precise meaning of
this expression, which can be interpreted in more ways than one (including in support of the idea that
a competitive disadvantage is enough to trigger intervention) needs to be teased out from an analysis
of individual judgments. The trajectory of the case law since at least Deutsche Telekom120 sheds light
on the nature of the assessment, both in relation to exclusionary and exploitative strategies.
Suffice it to mention some examples. A ‘margin squeeze’, which necessarily places rivals at
a competitive disadvantage by forcing them to sell below cost, is not in itself abusive.121 The
anticompetitive effects of the practice need to be established in light of its impact on the relevant
market.122 The same can be said, more generally, of conduct that amounts to below-cost pricing but
is not predatory within the meaning of AKZO,123 and of standardised rebates.124 In relation to the
117
Case C-12/03 P Commission v Tetra Laval BV, EU:C:2005:87; and Case T-5/02 Tetra Laval BV v Commission,
EU:T:2002:264.
118
Guidelines on the assessment of non-horizontal mergers (n 26), paras 39, 47-57, 67, 72-77, 103 and 111-118.
119
British Airways (n 9), para 68; Case C-280/08 P Deutsche Telekom (n 39), para 177; TeliaSonera (n 9), para 63 Post
Danmark II (n 9), para 31. This expression has occasionally been used in the context of Article 101 TFEU. See in
particular Maxima Latvija (n 8), para 29.
120
Deutsche Telekom (n 39).
121
Ibid, para 250.
122
Ibid, para 254.
123
Post Danmark I (n 8).
124
Post Danmark II (n 9).
35

latter, the criteria identified by the Court in cases like Post Danmark II and Intel are the same as those
deemed relevant under Article 101 TFEU in cases like Delimitis.125 Similarly, in Generics, the Court
held that, for the effects of a pay-for-delay agreement to trigger intervention under Article 102 TFEU,
the said effects need to go beyond the mere impact it has on the freedom of action of the generic
producer receiving the payment.126 Concerning, finally, exploitative conduct, the Court declared in
MEO that a competitive disadvantage does not suffice, in and of itself, to show that discriminatory
pricing amounts to a distortion of competition on the relevant market.127
6.3. As a rule, an anticompetitive effect cannot be equated with harm to consumer welfare
In the same way that an analysis of the case law shows that a limitation of a firm’s freedom of action
and/or a competitive disadvantage are insufficient to establish anticompetitive effects, it is clear that,
as a matter of principle, evidence of direct harm to consumers is not necessary to show such effects.
The question of whether anticompetitive effects can be equated with consumer harm has emerged
where the mechanisms through which these effects are manifested are collusion (and more precisely
the absorption of a source of competitive pressure) and exclusion. It has been contended that reduced
competitive pressure is, in and of itself, insufficient to justify intervention. In other words, the
elimination of a source of rivalry (whether through exclusion or collusion) would not amount, from
this perspective, to anticompetitive effects. The argument, according to this interpretation of the
notion, is that it would be necessary to show, in addition, that reduced rivalry makes or would make
consumers worse off.128
125
Delimitis (n 15), paras 15-36. Compare with Post Danmark II (n 9), paras 29-46 and Intel (n 38), paras 139-140.
126
Generics (n 32), paras 161 and 172.
127
Case C-525/16 MEO – Serviços de Comunicações e Multimédia SA v Autoridade da Concorrência, paras 25-26.
128
See for instance Spector (n 17) and Pinar Akman, ‘“Consumer” versus “Customer”: The Devil in the Detail’ (2010)
37 Journal of Law and Society 315.
36

The Court has consistently held, in contradiction with the abovementioned view, that the EU
competition law system is concerned not only with the protection of consumers, but also with the
‘structure of the market’ and therefore with ‘competition as such’.129 Accordingly, and as clarified in
British Airways, it would not be necessary to show that an exclusionary strategy would make
consumers worse off for it to be caught by Article 102 TFEU (or at least not in principle).130 Similarly,
the Commission would not need to show that a merger would lead to a price increase for end-users.
For instance, the mere fact that a transaction would eliminate the main source of competitive pressure
faced by the acquiring firm (and that there would be no perspective of new entry replicating such
constraints) would be enough to declare its incompatibility with the internal market.131
An analysis of the case law suggests that there is only one instance in which direct evidence
of consumer harm is required to establish an abuse of a dominant position. In Magill (as confirmed
in IMS Health), the Court held that a refusal to license an intellectual property right amounts to an
abuse of a dominant position where, inter alia, the behaviour prevents the emergence of a new product
for which there is potential consumer demand.132 This condition adds to indispensability and the
elimination of ‘all competition’ conditions, mentioned above. The implication of the ‘new product’
condition is that, for a refusal to license to amount to a violation of Article 102 TFEU, the
anticompetitive effects would need to go beyond those that follow inevitably from the operation of
the intellectual property rights; and that the harm to consumers should be both direct and
substantial.133
129
Glaxo Spain, para 63.
130
British Airways (n 9), paras 103-108.
131
Case T-342/07 Ryanair Holdings plc v Commission, EU:T:2010:280, paras 224, 225 and 445.
132
Magill (n 48), para 54; and IMS Health (n 48), para 38.
133
For a discussion, see Robert O’Donoghue and Jorge Padilla, The Law and Economics of Article 102 TFEU (2nd edn,
Hart Publishing 2013) 559-562.
37

6.4. Anticompetitive effects exist where competitive pressure is reduced
In EU competition law, anticompetitive effects exist where a practice (or transaction) harms the
ability and/or incentive to compete of firms that are as efficient as the firm(s) involved in (or
benefitting from) it. More precisely, an analysis of the case law suggests that the relevant question in
this regard is whether the ability and/or incentive to compete are harmed to such an extent that
competitive pressure is reduced. Thus, a disadvantage and/or a limitation of a firm’s freedom of action
would not be problematic in themselves, but only insofar as they can be expected to lead to such an
outcome. In the same vein, no anticompetitive effects would exist where the competitive pressure
faced by the firm(s) involved in (or benefitting from) the practice or transaction is not altered by it.
This principle is expressed in different ways depending on the nature or practice and the mechanism
through which the impact on competition is manifested.
6.4.1. Collusion or absorption of a source of competitive pressure
Turning to the first mechanism identified above – collusion or absorption of a source of competitive
pressure – anticompetitive effects may be manifested in two ways. First, a horizontal agreement or
merger may lead, as already mentioned, to unilateral effects. Unilateral effects arise where, without
coordination, a practice or transaction has a negative impact on one or more firms’ incentive to
compete. In other words, they arise where the reduced competitive pressure faced by one or more
firms leads to an increase in the degree of market power they enjoy. The most straightforward case is
one where a practice or transaction creates or strengthens a position of single dominance, which is
38

understood to mean a position of substantial market power.134 As is true of appreciability, market
shares are the most obvious proxy for dominance.135
Unilateral effects may also arise even where a practice or transaction does not create or
strengthen a dominant position. As explained above, the threshold of appreciability seems to be below
that of dominance. According to the GC in CK Telecoms, there are two conditions to establish
unilateral effects, absent dominance, in the context of EU merger control.136 First, the transaction
must lead to an appreciable reduction of the competitive constraints faced by the merging parties prior
to the transaction. Second, it must lead to an appreciable reduction of the competitive constraints
placed upon competitors. In other words, the GC held that the Commission would need to show, to
the requisite legal standard, that the transaction reduces the incentives to compete of all firms on the
relevant market. In addition to market shares,137 an increase in market power can be established by
proxy in light of factors such as firms’ closeness of competition or rivals’ ability to expand capacity
in response to a deterioration of the conditions of competition.138
Second, and as explained above, a practice (or transaction) can also lead to anticompetitive
effects if it eliminates the incentives to compete by means of inter-firm coordination. The
circumstances in which tacit collusion (and thus the emergence of a collective dominant position) is
likely to occur, and become sustainable, were defined by the GC in Airtours.139 Horizontal and non-
horizontal mergers may thus be declared to be incompatible with the internal market if the criteria
defined in that judgment are fulfilled.140 In addition, agreements within the meaning of Article 101
134
For a definition of the notion, see Hoffmann-La Roche (n 38), para 38. See also Guidelines on the assessment of
horizontal mergers (n 13), paras 2 and 17.
135
Hoffmann-La Roche (n 38), para 41 and AKZO (n 37), para 60.
136
CK Telecoms (n 101), para 96.
137
Guidelines on the assessment of horizontal mergers (n 13), paras 14-18 and 27. See, in the same vein, Guidelines on
horizontal co-operation agreements (n 23), paras 44-46.
138
Guidelines on the assessment of horizontal mergers (n 13), paras 28-38 and Guidelines on horizontal co-operation
agreements (n 23), para 34. See also CK Telecoms (n 101), paras 227-250.
139
Case T-342/99 Airtours plc v Commission, EU:T:2002:146, para 62. See also Case C-413/06 Bertelsmann AG and
Sony Corporation of America v Independent Music Publishers and Labels Association, EU:C:2008:392, para 123.
140
Guidelines on the assessment of horizontal mergers (n 13), paras 39-57; and Guidelines on the assessment of non-
horizontal mergers (n 26), paras 79-81.
39

TFEU may also be found to have restrictive effects if they create the conditions in which coordination
with other rivals on the market becomes feasible and/or easier to sustain. Some of these scenarios are
explored by the Commission in its Guidelines on horizontal co-operation agreements.141
6.4.2. Exclusion of a source of competitive pressure
Exclusionary concerns arise where actual or potential rivals’ ability and incentive to compete is
affected to such an extent that competitive pressure is reduced. In Delimitis, the Court devised a test
to determine whether access to the market would be foreclosed as a result of a practice (together or
in combination with other practices having the same object and/or effect).142 The question, as
confirmed in Maxima Latvija, would be whether there would be ‘real, concrete possibilities’ for a
new rival to establish itself and exercise a competitive constraint on existing players.143 The same
criteria seem appropriate, mutatis mutandis, to inquire whether existing market players would retain
their ability and incentive to exercise an effective constraint in spite of the practice. These two
judgments are also compatible with the way in which the analysis is actually conducted under Article
102 TFEU and merger control. In abuse of dominance cases, a practice can be said to lead to
‘foreclosure effects’144 where it ‘hinder[s] the ability of competitors’145 to operate on the relevant
market.
As a matter of principle, and in line with what has been explained above, only the exclusion
of equally efficient rivals is relevant in EU competition law.146 This principle is the corollary to the
need to establish a causal link between the practice and any anticompetitive effects. Such a causal
link would be missing where a firm’s departure would be the consequence of its inability to provide
141
Guidelines on horizontal co-operation agreements (n 23), para 77-85; and paras 175-182.
142
Delimitis (n 15), paras 15-36.
143
Maxima Latvija (n 8), para 27.
144
Intel (n 38), paras 138, 142 and 143; and Generics (n 32), para 157.
145
TeliaSonera (n 9), para 67.
146
See above, Section 5.2.1.
40

attractive goods and/or services. However, there may be instances in which it the exclusion of less
efficient rivals could give rise to a finding of infringement. This may be the case, the Court explained
in Post Danmark II, where a practice is implemented by a dominant firm that is protected by
regulatory barriers in a partially liberalised industry.147 It is reasonable to assume that, as the law
stands, it would be for the authority or claimant to show, to the requisite legal standard, why the
departure of less efficient rivals is justified in a given case.
The case law provides several illustrations of the principle. It follows from AKZO, Deutsche
Telekom and Post Danmark I that, if a practice does not force rivals to sell below cost, it is deemed
prima facie lawful. An equally efficient rival can be expected to withstand competition that does not
involve below-cost pricing.148 A variation on this filter is the so-called ‘as efficient competitor’ test,
which is relevant when evaluating the lawfulness of conditional rebates. This test is designed to
establish whether a rebate scheme, in the circumstances in which it is implemented, would require an
equally efficient rival to sell at a loss when competing for the contestable part of customers’
demand.149 In Intel, the Court clarified that a dominant firm may rely on the ‘as efficient competitor’
test to show that a loyalty rebate scheme is not capable of having anticompetitive effects.150
Considerations pertaining to the nature of the product, the practice and the actual context in
which the latter is implemented are relevant in the (prospective or retrospective) analysis of
exclusionary effects. First, the Court has had the occasion to explain how the probability of
anticompetitive effects depends, at least in part, on the nature of the practice and its relative potential
to cause harm. This is true of price-based and non-price-based conduct. For instance, the level of a
‘margin squeeze’ has an impact on its exclusionary potential. In particular, the Court held in
TeliaSonera that a negative ‘margin squeeze’ (that is, an instance where the wholesale price charged
by the dominant firm to its downstream rivals is higher than the retail price it charges to its end-users)
147
148
AKZO (n 37), para 72.
149
Guidance (n 25), paras 39-44.
150
Intel (n 38), paras 142-143.
41

is a probable source of anticompetitive effects.151 Conversely, a pricing practice that allows a rival to
cover the bulk of its costs is unlikely to have such effects.152 The case law on rebates hints at a similar
idea. The exclusionary potential of a scheme depends on the ‘criteria and rules’ for the award of the
rebate.153 Thus, rebates conditional upon exclusivity are more likely to cause harm than standardised
schemes based on the volume supplied (in the same way that retroactive rebates are more likely to
cause harm than incremental schemes). By the same token, an outright refusal to deal has more
exclusionary potential than dealing on less favourable terms and conditions.
Second, factors pertaining to the context in which the practice (or transaction) is implemented
shed light on its impact on rivals’ ability and incentive to compete. One of these factors is the degree
of market power enjoyed by the firm(s) involved. In this sense, the extent of the dominant position
has been frequently mentioned by the Court in Article 102 TFEU cases.154 A second factor is the
coverage of the practice, that is, the fraction of the market subject to it.155 Third, the features of the
relevant market and the nature of the product may also play a prominent role. For instance, the
existence of high barriers to entry characterised by economies of scale may have an impact on rivals’
ability and incentive to compete.156 The same can be said of network effects, which may exacerbate
the foreclosure effects of a practice.157 Where the dominant firm is vertically-integrated, the
indispensable nature of the product may be a factor.158 The regulatory context is another one.159
Where the analysis is retrospective in nature, actual evidence relating to rivals’ ability and
incentive to enter and/or remain on the market is relevant, as already explained above.
Anticompetitive effects would not exist, for instance, where the contemporary evolution of the market
151
TeliaSonera (n 9), para 73.
152
153
Post Danmark II (n 9), para 29 and 32.
154
TeliaSonera (n 9), para 81; Post Danmark II (n 9), para 30; and Intel (n 38), para 139.
155
Delimitis (n 15), para 19; Case C-549/10 P Tomra Systems ASA and Others v Commission, EU:C:2012:221, paras 37-
49; Post Danmark II (n 9), para 46; Maxima Latvija (n 8), para 29; and Intel (n 38), para 139.
156
157
158
Deutsche Telekom (n 39), para 255; and TeliaSonera (n 9), paras 69-70.
159
Post Danmark II (n 9), para 39, which makes a reference to the ‘statutory monopoly’ enjoyed by the dominant firm.
42

reveals that the practice has not precluded entry.160 The experience acquired over the year yields some
valuable conclusions about the impact of practices in the actual context of which they are a part. It
appears, to begin with, that a competitive advantage (even an unparalleled one) does not necessarily
limit firms’ ability and/or incentive to compete. In Microsoft I, for instance, the Commission noted
that the tying of Windows and Windows Media Player gave the latter product an unparalleled
advantage over rivals.161 In spite of this fact, and even though the remedy failed to work as expected,
the practice did not limit rivals’ ability and incentive to compete.162
Experience also shows that a competitive disadvantage and/or a decrease in terms of market
share (as evidenced, for instance, by the loss of some customers) do not necessarily impact negatively
on rivals’ ability and incentive to compete.163 In and of themselves, the case law shows, these factors
are insufficient to establish exclusionary effects. Decades of enforcement reveal that a competitive
disadvantage may in fact have a positive impact on rivals’ incentives to compete. This point is
acknowledged by the Commission in its Non-Horizontal Merger Guidelines.164 It may spur rivals to
develop counterstrategies or to improve the quality of their products to make up for the disadvantage.
Post Danmark I, in turn, provides a concrete illustration of how a firm may retain its ability and
incentive to compete even after experiencing a decrease in its market share: as observed above, the
dominant firm’s competitor managed to retain its distribution network and win back the two major
customers that were lost following the implementation of the practice. The analysis in Post Danmark
I was retrospective in nature. Where the analysis is prospective, evidence about the features and past
evolution of the market would also be a factor to consider in this regard.165
160
Case C-457/10 P AstraZeneca AB and AstraZeneca plc v Commission, EU:C:2012:770, paras 194-203.
161
162
163
See, by analogy, Post Danmark I (n 8), para 39 and Deutsche Telekom (n 39), para 250.
164
Guidelines on the assessment of non-horizontal mergers (n 26), paras 39, 67 and 103.
165
See in particular British American Tobacco (n 54), John Deere (n 54) and Guidance (n 25), para 20.
43

Principle It is for the authority or
Case law
claimant to show why
Only the exclusion of equally departing from the principle
Intel, Post Danmark II
efficient firms is relevant is justified in casu
Relevant factors
Nature of the practice See the contrast between, e.g. Case law
incremental vs retroactive
The exclusionary potential rebates; and negative vs AKZO, TeliaSonera, Post
depends on the practice positive ‘margin squeezes’ Danmark II
Degree of market power Case law
The greater the market power, TeliaSonera, Post Danmark II,

the more likely the effects Intel
Coverage
Case law
The greater the coverage, the
Post Danmark II, Intel
more likely the effects
For instance, effects may be Case law

Other factors include the
more likely in presence of
features of the product and
network effects or large Microsoft I, TeliaSonera, Post
market, regulatory context
economies of scale Danmark II, Intel
Additional considerations
• Experience shows that a competitive advantage (even an unparalleled one) does not necessarily have a
negative impact on firms’ ability and/or incentive to compete (see e.g. Microsoft I, Microsoft/Skype)
• A decrease in market share does not necessarily reduce rivals’ ability and/or incentive to compete (see
e.g. TeliaSonera, Post Danmark I)
Fig. 5: The exclusion of a source of competitive pressure
6.4.3. Exploitation
Cases dealing with exploitative behaviour are relatively scarce. A question that these cases raise is
whether the analysis of exploitative effects differs from that undertaken under the two mechanisms
44

discussed above. This question is particularly relevant where the analysis concerns practices that have
as their object or effect the distortion of the conditions of competition on an upstream or downstream
market in which a dominant firm does not operate. The MEO judgment suggests that there are no
fundamental differences in the approach to the analysis of such distortions. As noted above, the Court
held in MEO that a competitive disadvantage does not amount, in and of itself, to a distortion of
competition within the meaning of Article 102(c) TFEU. Accordingly, the impact of the exploitative
practice on a customer’s ability and incentive to compete will have to be established in accordance
with the same criteria defined in cases involving exclusion.166
7. The probability of effects in the case law
7.1. Semantic issues: capability and likelihood
The applicable threshold of effects is not immediately obvious to infer from the case law. It requires
a careful reading, and comparison, of the relevant judgments. Part of the reason behind the absence
of clear and straightforward answers has to do with the vocabulary used when addressing the requisite
probability of anticompetitive effects. The Court has held that the actual or potential anticompetitive
effects must be ‘capable’ of occurring and/or ‘likely’ to occur.167 Other terms, such as ‘liable’ have
also been used.168 It would be reasonable to conclude from the case law, first, that ‘capability’ and
‘likelihood’ have, at least in some judgments, been used as synonymous; and, second, that these two
terms are suggestive of a single relevant threshold of effects. This is the point of view expressed by
AG Wahl in his Opinion in Intel.169
166
MEO (n 125), para 31.
167
See above n 9.
168
Murphy (n 44), para 140. See also Microsoft I (n 58), paras 560-564, where the GC addresses the point and, in
particular, whether the reference to the ‘risk’ of the elimination of all competition sets a threshold identical or similar to
that of likelihood.
169
Opinion of AG Wahl in Intel (n 9).
45

Even if this point of view were to be accepted, a number of difficulties remain. The single
most important one is that, while the two terms have been used as synonymous, they convey different
meanings. As a result, the requisite threshold of effects would vary depending on the way they are
interpreted. On the one hand, the literal meaning of the words ‘capable’ and ‘capability’ is indicative
of a low threshold of effects,170 which can be equated with plausibility. Practices and transactions
(including the examples mentioned above, such as tying, exclusive dealing or horizontal and non-
horizontal mergers) attract the attention of competition authorities and give rise to litigation precisely
because they are capable of having anticompetitive effects (or, if one prefers, because it is typically
plausible that they will have a negative impact on competition). If this interpretation were accepted,
the threshold of effects would be presumptively met as soon as it is established that the practice or
transaction has been implemented. Absent other factors pertaining to the economic and legal context,
anticompetitive effects would be deemed to follow, logically and inevitably, from the very
implementation of the practice.
The words ‘likely’ and ‘likelihood’, in turn, are indicative of a higher threshold. The literal
meaning of the words is suggestive of an event that will probably happen or is expected to happen.171
In her Opinion in Post Danmark II, AG Kokott suggested that the applicable threshold is one of
likelihood. In the Advocate General’s view, anticompetitive effects would be established when it is
‘more likely than not’ that they will be manifested.172 AG Kokott’s operational definition of the
concept is not only in line with its plain meaning, but also with the meaning attached to it in Section
2.4. above. This interpretation would, as already mentioned, place the threshold of probability right
170
The Oxford English Dictionary defines ‘capable’ – in its fifth entry – as ‘[h]aving the needful capacity, power, or
fitness for (some specified purpose or activity)’. The Cambridge Dictionary defines that a person is ‘capable’ of something
when she has ‘the ability or qualities to be able to do something’.
171
The Oxford English Dictionary defines ‘likely’ as ‘having a high chance of occurring; probable’. The Cambridge
Dictionary, along the same lines, defines ‘likely’ as ‘expected’ and ‘probably true’.
172
Case C-23/14 Post Danmark A/S v Konkurrencerådet, EU:C:2015:343, para 82: ‘According to settled case-law, it is
necessary but also sufficient that the rebates in question can produce an exclusionary effect. This is the case where, on
the basis of an overall assessment of all the relevant circumstances of the individual case, the presence of the exclusionary
effect appears more likely than its absence’. In support of her interpretation of the case law, AG Kokott refers to Post
Danmark I (n 8), paras 42 and 44.
46

above 50%. At this level, effects would no longer follow presumptively from the implementation of
the practice. For the same reason, it would be easier for the firm(s) involved in the practice or
transaction to rebut a finding of anticompetitive effects.
7.2. The law as applied: plausibility, likelihood and certainty
If one pays attention to the thresholds actually applied by the Court, it becomes possible to discern
the relevant threshold that is relevant for each of the legal tests. When prima facie unlawful practices
are at stake, the applicable threshold is one of plausibility. In other words, it is only in a narrow set
of circumstances that the firm(s) involved are able to rebut the presumption that the behaviour is
capable of having anticompetitive effects. Second, a threshold of likelihood, as defined by AG Kokott
in Post Danmark II, is relevant to evaluate the impact of practices subject to a ‘standard effects’
analysis, as well as concentrations within the meaning of Regulation 139/2004. Finally, the threshold
of certainty, or quasi-certainty appears to be the applicable one where the ‘enhanced effects’ test
defines the conditions against which the legality of conduct is assessed.
7.2.1. A plausibility threshold applies to prima facie unlawful conduct
An overview of the case law reveals that conduct that is prima facie unlawful irrespective of its
effects, such as cartel-like behaviour, is prohibited even when it is not particularly likely to have a
negative impact on competition. In that sense, one can rule out that a threshold of likelihood is
relevant in relation to these practices. It is sufficient to show that harm is a plausible outcome. In T-
Mobile, the referring national court expressed the view that the behaviour at stake in the case – a
single meeting where the reduction of remunerations paid to dealers was discussed among rivals –
could not qualify as an object infringement due to the fact that it was unlikely to have anticompetitive
47

effects. More precisely, the referring court was of the view that the ‘by object’ category would be
appropriate ‘where the actual detrimental effects are unmistakable and will occur irrespective of the
characteristic features of the relevant market’.173
The Court dismissed this view and held that the threshold suggested by the national court
(which hinted at certainty or quasi-certainty) is not the applicable one in relation to ‘by object’
conduct. These practices ‘must simply be capable in an individual case, having regard to the specific
legal and economic context’, of having anticompetitive effects.174 In T-Mobile, the Court used the
term ‘capable’ in a manner consistent with its literal meaning, which is consistent, in turn, with a
threshold of plausibility. Bananas confirmed that the requisite threshold of effects would be met in a
‘by object’ case where employees in an industry have bilateral discussions about pre-pricing
information.175 In Toshiba, it sided with the GC and concluded that a cartel-like arrangement
involving a group of potential competitors was capable of having restrictive effects insofar as barriers
to entry were not found to be insurmountable in its economic and legal context.176
Article 102 TFEU case law leads to similar conclusions. In AKZO, the Court held that pricing
below cost is capable of leading to the exclusion of equally efficient competitors.177 As noted in Post
Danmark I, however, below-cost pricing is not necessarily likely to have anticompetitive effects – in
particular where rivals would be able to cover the bulk of their costs. However, where the practice is
an element of a strategy aimed at excluding rivals, it will be deemed abusive even when exclusion is
no more than plausible.178 The case law on rebates leads to similar conclusions. Some rebate schemes
are prima facie abusive irrespective of their effects.179 Underpinning the legal status of these practices
is the idea that they can, or ‘tend to’, restrict competition.180 Finally, tying is also prima facie
173
T-Mobile (n 2), para 20.
174
Ibid, para 31.
175
Bananas (n 2), paras 111-135.
176
Toshiba (n 80), paras 40-48.
177
AKZO (n 37), para 72.
178
Ibid.
179
See above n 38.
180
Hoffmann-La Roche (n 38), para 90.
48

prohibited.181 As a result, it is sufficient for a prohibition to be grounded on the conclusion that
anticompetitive effects are at least plausible.182
The case law also gives an idea of what firms would need to show to rebut the presumption
that prima facie unlawful conduct is a plausible source of anticompetitive effects. First, and in line
with Toshiba and Generics, it would be possible for them to show that the behaviour is incapable of
having an impact on competition insofar as there are ‘insurmountable barriers to entry’.183 In such
circumstances, any actual or potential effect would not be attributable to the practice, but to the
regulatory context of which it is a part. Second, it would be possible for firms to argue that the
behaviour is objectively necessary to achieve a pro-competitive aim. They could show, for instance,
that an exclusive distribution agreement limiting both active and passive sales is incapable of having
anticompetitive effects insofar as, in its absence, market entry by the supplier would not occur.184 In
other words, the parties may be able to provide evidence to the effect that the practice can only
produce pro-competitive gains.185 Finally, Intel clarified that it is possible for a dominant firm to
provide evidence pertaining to the nature of the practice, its extent and the features of the relevant
market.186 The Court suggested that a dominant firm may be in a position to show that the exclusion
of an equally efficient competitor is implausible given the nature and scope of the practice in its
economic and legal context. Intel expressly refers to the ‘as efficient competitor’ test as a tool in this
regard. It is reasonable to infer from the judgment that, more generally, a firm would be able to
provide evidence showing that the practice does not deny rivals a minimum efficient scale.187
181
See above n 36.
182
See in this sense Microsoft I (n 58), para 1054.
183
Generics (n 32), para 45.
184
Guidelines on vertical restraints (n 24), para 61, mentioned above.
185
See also, in this same vein, Budapest Bank (n 42), paras 82-83.
186
Intel (n 38), paras 138-142.
187
This conclusion seems consistent with Tomra (n 155). In the latter, the Court held – at para 46, an in line with Intel (n
38), para 137 – that the Commission would not need to apply the ‘minimum viable scale’ test to establish that a loyalty
rebate scheme amounts to an abuse. Intel clarifies that, in spite of the legal status of the practice as prima facie unlawful,
a dominant firm may provide evidence showing that the practice is incapable of excluding equally efficient rivals.
49

7.2.2. A threshold of likelihood applies to practices and transactions subject to a ‘standard
effects’ analysis
A threshold of likelihood applies to practices and transactions subject to a ‘standard effects’ test,
which includes mergers examined in accordance with Regulation 139/2004. Post Danmark II comes
across as the most obvious starting point, not only because it illustrates the idea effectively, but
because it is the background against which AG Kokott formalised the requisite threshold of
likelihood. An overview of the facts in Post Danmark II unambiguously shows that the rebate scheme
at stake was, in and of itself, a plausible means to exclude competition. The scheme had indeed been
implemented by an incumbent in a partially liberalised industry with a very large market share; in
addition, the rebates were retroactive, and the relevant reference period was of one year.188 In spite
of these considerations, the Court held that it was necessary to consider the likely impact of the
practice in light of a number of factors pertaining to the relevant economic and legal context. Other
Article 102 TFEU cases where the similar threshold applied include Post Danmark I, Deutsche
Telekom, TeliaSonera and MEO, all discussed above.
In the context of Article 101 TFEU, it seems clear that, once an agreement is found not to be
restrictive by object, showing that anticompetitive effects are plausible (or that there are no
‘insurmountable barriers to competition’) is not enough. In such circumstances, it would be necessary
to establish the likely effect of the practice. Suffice it to come back to Delimitis (and the rulings that
embraced its approach, such as Maxima Latvija) to illustrate the point. The practice at stake in that
case – exclusive dealing – is known to be at least a plausible source of anticompetitive effects. A
network of exclusivity agreements, alone or in combination with others, can lead to the exclusion of
equally efficient suppliers. However, the Court devised a test requiring an authority or claimant to
188
Post Danmark II (n 9), paras 30-46.
50

show, in light of an in-depth assessment informed by the features of the market, how the foreclosure
of new entrants would likely result from their implementation in a given economic and legal context.
This same conclusion follows, from an analysis of the way mergers are scrutinised in the EU
regime. That the applicable threshold is one of likelihood was already apparent from Kali & Salz, in
which the Court concluded that the Commission had failed to establish, to the requisite legal standard,
that the transaction was likely to lead to the strengthening of a collective dominant position.189 The
authority had identified several indicators suggesting that such an outcome was at least plausible in
the post-merger scenario. However, the Court found that the joint market share of the parties, or that
the structural links between them, did not point conclusively to the strengthening of a collective
dominant position.190 Other factors, such as the decline in the demand for the product concerned by
the transaction, suggested that an anticompetitive outcome was unlikely.191
The threshold of likelihood was put to the test – and confirmed – when the GC evaluated the
Commission decisions in GE/Honeywell192 and Tetra Laval.193 In the two cases, it did not dispute the
authority’s conclusion that the conglomerate effects of the transactions could lead to the extension of
a dominant position from one market to a neighbouring one. In the economic and legal context of
which the transactions were a part, such an outcome was found to be at least plausible.194 However,
the GC concluded that it was not a likely one. For that reason, it concluded that the Commission had
erred in law in the two cases. In GE/Honeywell, for instance, the GC took the view that the strategy
through which the alleged conglomerate effects would be manifested went against the ‘modus
operandi’ of the sector.195 Thus, an ‘additional commercial effort’ would be required from the merged
entity for exclusion to become a reality.196
189
Kali & Salz (n 77), para 170.
190
191
Ibid, para 238.
192
Case T-210/01 General Electric Company v Commission, EU:T:2005:456.
193
Case T-5/02 Tetra Laval (n 117).
194
General Electric (n 192), para 404; and Case T-5/02 Tetra Laval (n 117), paras 192-199.
195
General Electric (n 192), para 415.
196
General Electric (n 192), para 423.
51

7.2.3. A threshold of certainty (or quasi-certainty) applies to conduct subject to a ‘enhanced
effects’ test
As explained above, a refusal to deal within the meaning of Bronner is abusive if it can be shown to
relate to an input or platform that is indispensable and, in addition, that it would lead to the elimination
of ‘all competition’ on the relevant adjacent market.197 An input or platform is indispensable,
according to the case law, where there are no ‘alternative solutions’ to enter the relevant adjacent
market and where, in addition, duplicating it would be ‘impossible or unreasonably difficult’. These
conditions, which are notoriously demanding in practice, amount in effect to setting a threshold of
certainty for third parties requesting access to the said input or platform. In addition, the condition
that the refusal eliminate ‘all competition’ on the adjacent market implies that no alternative input or
platform is already in place.
8. Analysis and discussion
8.1. Questions addressed in the case law
8.1.1. The notion of anticompetitive effects has clear boundaries
The analysis above suggests that it is possible to discern, from the case law, a specific meaning of the
notion of anticompetitive effects. Starting with competition and the counterfactual, the Court made it
clear, from the outset, that competition comprises both its inter-brand and intra-brand dimensions. In
addition, the case law is consistent in taking into account both the ex ante and the ex post aspects of
the counterfactual. In this sense, the Court does not simply assume that the pro-competitive gains
197
See above n 48.
52

resulting from a given practice would have existed independently of any restraints that seemingly
amount, from an ex post perspective, to a restriction. What is more, the system provides for
mechanisms to take the counterfactual into consideration. Not only is it possible for firms to argue
that any ex post restraints are inextricably linked to the pro-competitive aspects of the practice and
thus do not have anticompetitive effects; some legal tests are specifically crafted to incorporate the
ex ante dimension of the counterfactual. The choices made by the Court in this regard are captured in
Figure 6.
Inter- and intra- brand
Inter- and intra-brand
Ex ante and ex post

Dimensions of competition
Inter-brand only
Ex post only Ex ante and ex post +
Dimensions of the counterfactual
Fig. 6: The dimensions of competition and the counterfactual in the case law
It is also possible to discern, from the case law, what effects are. The probability threshold,
which appears to vary depending on the applicable legal test, can also be identified. The different
combinations around two variables are depicted together in Figures 7, 8 and 9. First, only appreciable
effects are relevant in EU competition law. Second, effects amount to more than a mere competitive
disadvantage or a limitation of a firm’s freedom of action but (save a limited exception) to less than
53

consumer harm. In addition, it is clear that – at least as a matter of principle – the departure of
inefficient firms is understood to be the natural consequence of the operation of the competitive
process. Against this background, it would appear that effects can be defined as those that impact on
the ability and/or incentive of one or more firms to compete, and this to such an extent that
competitive constraints are reduced as a result.
One can identify three probability thresholds, each corresponding to a particular legal test. To
begin with, a threshold of plausibility applies to conduct that is prima facie unlawful irrespective of
its impact on competition (including ‘by object’ conduct under Article 101(1) TFEU). Second, when
practices and transactions subject to a ‘standard effects’ test are at stake, it is necessary to show that
the practice is ‘more likely than not’ to restrict competition (that is, a threshold of likelihood). Finally,
certainty, or quasi-certainty, is required when the impact of practices subject to an ‘enhanced effects’
test (both under the Bronner and the Magill doctrines) is at stake. The indispensability and the
‘elimination of all competition’ conditions inevitably amount to such a threshold.
Certainty
Likelihood
Plausibility Prima facie

- unlawful conduct

+
Fig. 7: Anticompetitive effects and prima facie unlawful conduct (where effects are presumed)
* (or competitive disadvantage)
54

Certainty
‘Standard
Likelihood effects’ test
Capability
-

+
Fig. 8: Anticompetitive effects under the ‘standard effects’ analysis’
‘Enhanced ‘Enhanced
Certainty
effects’ test effects’ test
(Bronner) (Magill)
Likelihood
Capability
-

+
Fig. 9: Anticompetitive effects under the ‘enhanced effects’ analysis
55

8.1.2. The approach to the analysis of effects is the same across provisions
A corollary to the conclusions above is that the analysis of effects is the same across the board. When
the ‘standard effects’ test is applicable, the assessment does not seem to vary depending on whether
Article 101 TFEU, 102 TFEU or Regulation 139/2004 is at stake. This conclusion would not have
been obvious to draw during the formative years of the discipline. In particular, it has not always been
clear whether evidence of anticompetitive effects is required to establish an abuse of a dominant
position and whether, if indeed required, the assessment was comparable to that undertaken in the
context of Article 101 TFEU and merger control. The evolution of the case law, in particular
following Deutsche Telekom, TeliaSonera, Post Danmark I and II and Intel, seems to have dissipated
any doubts in this respect. In these judgments, the Court clarified that the practices at stake were only
caught by Article 102 TFEU insofar as they were likely to have anticompetitive effects (other
practices, such as pricing below average variable costs, remain prima facie unlawful). Crucially, the
Court’s assessment is consistent with the framework captured in Figure 8.
From a normative standpoint, the application of a single approach, across the board, to the
analysis of effects seems reasonable and, arguably, inevitable. This is so, first and foremost, because
practices and transactions implemented by a dominant firm can be examined under all provisions
considered. Both Articles 101 and 102 TFEU can simultaneously apply to the same practice, as cases
like Generics and Delimitis/Intel show. Similarly, the likely impact of a leveraging strategy may be
examined under either under Articles 101 or 102 TFEU or, if it results from a conglomerate merger,
under Article 2 of Regulation 139/2004. Just to mention a clear example, similar tying concerns were
considered in Microsoft I (an Article 102 TFEU case) and Microsoft/Skype (a merger case).198 Insofar
as the underlying economic and legal context is essentially the same, it is not obvious to see what
198
Microsoft/Skype (Case COMP/M.6281) Commission Decision of 7 October 2011. See also Case T-79/12 Cisco
Systems, Inc. and Messagenet SpA v Commission, EU:T:2013:635.
56

would justify attaching a different meaning to the notion of effects depending on the applicable
provision.
8.2. Open questions
8.2.1. The definition of appreciability
Even though it is clear from the case law that effects must be appreciable, the EU courts have not had
the occasion to clarify how to assess the question on a case-by-case basis. On the one hand, it is clear
from Völk and Expedia that, where the market power of the parties is insignificant, the appreciability
threshold is not met. On the other hand (and as a corollary to the first point), the effects of a practice
implemented by a firm that holds a dominant position (that is, a substantial degree of market power)
will, when established, be appreciable. The difficulty in practice has to do with the identification of
the point at which the degree of market power enjoyed by the firm(s) is significant enough to meet
the appreciability threshold (that is, the boundary between de minimis and appreciable effects). This
practical difficulty has two dimensions. One dimension relates to the problems that are inherent in
establishing market power in the first place. It is an inquiry that involves considering a broad range
of factors, and that is typically assisted by using proxies, in particular market shares. A second
dimension is the identification of the appropriate degree of market power.
In order to address these difficulties, the system relies, at present, upon the various proxies
devised over the years, both in the context of Article 101 TFEU and merger control. At best, these
proxies give an idea of the approximate point at which the degree of market power falls below the
threshold. In other words, these proxies do little more than flesh out, in a more precise way, the
principles defined in Völk. Where the practice or transaction is of a horizontal nature, questions about
appreciability (or rather, its absence) start where the joint market share is below 25%, which is the
57

threshold defined in the Horizontal Merger Guidelines.199 Where the practice or transaction is of a
non-horizontal nature, the threshold is set at 30%.200 As can be seen in Figure 10, these instruments
do little more than narrow down, in a tentative way, the ‘grey area’ between Völk and dominance.
The narrowing of this gap is tentative in the sense that nothing prevents a finding of appreciability
below the threshold, but also in the sense that most proxies are derived from a series of soft law
instruments issued by the Commission.
The difficulty that comes with the assessment of appreciability became apparent in CK
Telecoms. This is the first merger case in which the EU courts are confronted with unilateral effects
in the absence of dominance. As explained above, effects in such circumstances are to be established
by reference to factors such as market shares and the closeness of competition between the parties. In
addition, quantitative instruments may be used as a screen to filter out unproblematic
concentrations.201 It is apparent throughout the GC judgment in the case that it may not be easy to
define ex ante the point at which the impediment to effective competition becomes ‘significant’ (and
thus appreciable). It is a context-dependent exercise that requires an evaluation of the degree of
competition between the merging parties prior to the transaction and the features of the relevant
market. In this sense, CK Telecoms suggests that any guidance may only slowly emerge from a
succession of cases addressing the ‘grey area’ between de minimis conduct and transactions, on the
one hand, and those giving rise to dominance, on the other.
‘Grey area’
De minimis Dominance
Fig. 10: The definition of appreciability
199
See for instance Hutchison 3G Austria/Orange Austria (Case COMP/M.6497) Commission Decision of 12 December
2012, in particular paras 90-100.
200
Guidelines on the assessment of non-horizontal mergers (n 26), para 25; and Guidelines on vertical restraints (n 24).
201
CK Telecoms (n 101), paras 253-259.
58

8.2.2. The relative lack of guidance about the assessment of (exclusionary) effects
While the case law gives a sufficiently precise idea of what anticompetitive effects are (as Figures 7,
8 and 9 sought to capture), it remains relatively scant on examples setting out how the ‘standard
effects’ test is to be conducted in practice. This reality leaves several questions unanswered. The
absence of concrete illustrations of the implementation of the applicable framework becomes
particularly apparent when exclusion and exploitation are the mechanisms through which effects are
manifested. To focus on the former, it is true that the Court has already identified a number of factors
to consider in the assessment of anticompetitive effects resulting from exclusion and which comprise,
in particular, the extent of the market power enjoyed, the coverage of the practice, the economic
features of the relevant market and the nature of the product.
However, the Court has not been given sufficient chances to explain, in full, how these factors
are put into operation. As is true of appreciability, the case law is useful to identify the instances in
which anticompetitive effects are in principle implausible, on the one hand, and when they are
particularly likely, on the other. At one end of the spectrum, the Court has confirmed that above-cost
prices are in principle incapable of excluding equally efficient competitors. The same can be said of
prices that would allow rivals to cover the bulk of their costs. The ‘as efficient competitor’ test is an
expression of this same idea. At the other end of the spectrum, the Court has confirmed – in relation
to ‘margin squeeze’ conduct – that anticompetitive effects are at least likely where the relevant input
is indispensable, or where the ‘margin squeeze’ is negative.
Beyond these filters, which are useful to identify instances that safely lie at the two ends of
the spectrum of liability, the case law only sheds limited light on how to establish the likely
anticompetitive effects of a practice or transaction. For instance, while the coverage of the practice is
a relevant (and sometimes a fundamental) factor, there is no indication of the level below which the
59

ability and incentive of rivals to compete would be unlikely to be affected. At most, Post Danmark I
suggests that, where there is contemporary evidence showing that rivals were able to remain on the
market in spite of the anticompetitive potential of a practice, this fact can be sufficient to rule out a
finding of anticompetitive effects.
8.2.3. Capability, likelihood, certainty: between semantic and substantive issues
As explained above, there appears to be a gap between the probability thresholds as declared by the
Court and as actually applied by it. The law as declared may not give a clear idea of the requisite level
of probability – the words ‘capable’ and ‘likely’, sometimes used indistinctly, convey different
meanings. The law as applied, on the other hand, gives a more precise idea in this regard. The gap
between the law as declared and the law as applied is a potential source of legal uncertainty, which
may be exploited by stakeholders. It may also be a source of inconsistencies if the confusion trickles
down into the law as applied. Inconsistencies may emerge within provisions – which would arise, for
instance, if both plausibility and likelihood apply to potentially abusive practices subject to a
‘standard effects’ test – and across provisions – which would arise, for instance, if the assessment of
tying practices were assessed differently under Article 102 TFEU, on the one hand; and merger
control, on the other.
It is submitted that the risk of uncertainty and inconsistency could be addressed by bringing
in line the law as declared and the law as applied. The current gap between the two can be explained
by the fact that it has not always been clear which practices were deemed prima facie unlawful
irrespective of their effects and which practices were subject to a ‘standard effects’ analysis. As the
law stands, and following the evolution of the past two decades, it is now possible to tell one group
of practices apart from the other. For the same reason, it would come across as reasonable to define
more clearly and explicitly the respective scopes of the ‘capability’ and ‘likelihood’ thresholds, on
60

the one hand, and to use the vocabulary in a consistent way, on the other. The threshold of ‘capability’
seems appropriate, both in form and substance, for behaviour that is prima facie unlawful irrespective
of its effects; ‘likelihood’, in turn, for practices and transactions subject to a ‘standard effects’ test.
8.3. Areas of friction in practice
The open questions described above, together with the relative absence of detail regarding certain
aspects of the assessment, can be expected to give rise to frictions, in the sense that disagreements
about the meaning and/or operation of some concepts are likely to arise before courts and authorities.
One can think of six main areas of friction in practice. The first is the tendency to conflate
appreciability and effects. The second concerns the assessment of effects, and more precisely what
the evaluation entails in practice. Third (and in part as a result of the vocabulary used in the case law),
questions about the role and relevance of the counterfactual in practice are likely to emerge. The
fourth relates to the application of the application of the principle whereby only the exclusion of
equally efficient firms is relevant in the analysis. Fifth, there is a tendency to conflate the legal test
and the standard of proof. Finally, the time dimension and the probability threshold tend to be
confused.
8.3.1. The conflation of appreciability and effects
As explained above, the definition of effects and the question of whether such effects are appreciable
are different questions. The former issue involves making a choice along the continuum that ranges
from a competitive disadvantage (or a limitation of firms’ freedom of action) to harm to consumers.
The second question, in turn, relates to the market power enjoyed by the firm(s) involved in a practice
or transaction. Accordingly, market power can exist without effects. It is not necessarily the case that
61

every practice or transaction implemented by a firm with significant market power will have a
restrictive impact on competition. This is, after all, what the Court has consistently held since
Deutsche Telekom. Where effects are an element of the legal test, an actual or potential impact on
competition must be established (and not simply inferred from the market power enjoyed by the firm).
It is not difficult to see, however, why and how the two concepts can be conflated in practice.
The point is easily illustrated by reference to the case law on rebates and exclusivity under Article
102 TFEU. As explained by the Court in Post Danmark II and Intel, the market coverage of the
practice is one of the key factors when evaluating whether such practices have actual or potential
effects. Where the coverage is limited, it is reasonable to expect firms to argue that their schemes are
incapable or unlikely to have a restrictive impact on competition. In such circumstances, the ability
and incentive of equally efficient rivals to enter or remain on the market may be unaffected. However,
an authority or claimant may attempt to counterargue (conflating appreciability and effects in the
process) that there is no appreciability threshold in the context of Article 102 TFEU and therefore
arguments about the limited coverage of the practice are irrelevant.
8.3.2. The tendency to equate every competitive disadvantage with an anticompetitive effect
Experience shows that a competitive advantage – even an unparalleled one – does not necessarily
lead to anticompetitive effects. However, an analysis of the administrative practice suggests that
stakeholders tend to equate the former and the latter. In particular, evidence that rivals are placed at
a disadvantage tends to be used as conclusive proof that a practice has exclusionary effects. This idea
is aptly illustrated in light of the analysis conducted by the Commission in Google Shopping. The
assessment in the decision revolves around the finding that the practice inflicted an competitive
disadvantage on rivals (it decreased traffic from Google’s general search results pages to rivals, and
62

increased traffic to Google’s own services202), which cannot be fully offset by other means of
obtaining traffic.203 In this sense, the scenario was not fundamentally different from that at stake in
Microsoft I and Microsoft/Skype. The Android decision also revolves around a similar theme and
displays the same tendency to equate an unparalleled advantage with anticompetitive effects.204
An unparalleled competitive advantage suggests, at most, that the restrictive impact of the
practice is plausible. In and of itself, it is insufficient to conclude that anticompetitive effects are
likely. As discussed above, the experience of cases like Microsoft I (Media Player) and
Microsoft/Skype reveals that even an advantage that rivals cannot match may fail to have a negative
impact on rivals’ ability and incentive to compete. Competitors may be able to exploit their own
strengths, develop counterstrategies or improve their products in response to the practice. In other
words, an unparalleled advantage may spur rivalry (rather than reduce it) and, insofar as it does, firms’
incentives to invest and innovate. In Post Danmark I, for instance, the dominant firm enjoyed unique
advantages as the incumbent operator in a partially liberalised industry. 205 As mentioned above, its
main rival was able to gain back the customers initially lost as a result of the practice.206
8.3.3. The role and assessment of the counterfactual
As explained above, the case law considers both the ex ante and ex post dimensions of the
counterfactual. In practice, however, the ex ante dimension, which accounts for firms’ incentives to
invest and innovate, may be downplayed or disregarded. This is so because ex post restraints are
observable whereas ex ante gains are typically assumed to exist. In other words, the pro-competitive
effects of a practice a practice tend to be taken as a given, without considering that they may be
202
Google Shopping (n 12), paras 452-501.
203
204
Google Android (Case AT.40099) Commission Decision of 18 July 2018, in particular paras 896-963.
205
Post Danmark I (n 8), paras 3-4.
206
63

inextricably linked to the observable ex post conduct (and thus that they may not have existed in the
absence of the latter). This tendency is more likely to be displayed where the ex ante dimension is not
enshrined in the legal test – that is, where the practice is deemed prima facie unlawful or where it is
subject to a ‘standard effects’ test. In such circumstances, claimants and authorities, on the one hand,
and defendants, on the other, can be expected to disagree about whether the pro-competitive gains
resulting from the practice would have existed in their absence.
Just to mention an example, frictions might arise about whether an agreement has, as its
object, the restriction of competition. Such agreements are prima facie prohibited irrespective of their
effects. However, the evaluation of the object of the agreement cannot disregard the ex ante dimension
of competition, as explained above. Accordingly, one can expect disagreements to emerge around
whether the practice improves the conditions of competition that would otherwise have existed (and
is therefore not prima facie unlawful). Generics and Budapest Bank make it clear beyond doubt that
defendants in a case can challenge, in light of the counterfactual, claims that an agreement has, as its
object, the restriction of competition. In the coming years, frictions can be expected to arise around
whether such claims meet the requisite legal standard.
8.3.4. Anticompetitive effects and equally efficient firms
As already discussed at length, EU competition law is only concerned with firms that are as efficient
as the firms involved in a practice or transaction – at least so as a matter of principle. The principle
has two main implications, also addressed above. First, a firm’s departure from the relevant market
would not be attributable to a practice or transaction (that is, the requisite causal link would not exist)
if it is the consequence of the firm’s inability to prove attractive with consumers. Second, such an
outcome would be a natural manifestation of competition on the merits and would thus not amount
64

to anticompetitive effects. The fact that above-cost prices are prima facie lawful and the ‘as efficient
competitor’ test is used as a filter in rebate cases follow logically from the above.
The practical difficulty that emerges in relation to this principle has to do with its practical
implementation. Filters such as the ‘as efficient competitor’ test are particularly suited for instances
where the price is the relevant parameter of competition. In such instances, evaluating whether an
equally efficient firm would be able to withstand competition is relatively straightforward. The
assessment is likely to be more complex where the practice does not involve prices directly and/or
obviously (suffice it to think of exclusive dealing, tying or an outright refusal to deal). It may also be
more complex where price is not the relevant (or not the most relevant) parameter of competition at
stake. The fact that the filters designed by the Court do not seem relevant or immediately applicable
in such instances does not mean that the fundamental underlying principle, whereby only competition
from equally efficient firms is relevant in the system, does not apply.
Even though the difficulty of implementing the principle is not a reason to depart from it, one
can expect frictions to arise in practice. For instance, a firm may argue – in line with the logic
underpinning the case law – that there is no causal link between a practice and its actual potential
effects, since the potential or likely exclusion of rivals would be attributable to the latter’s inability
to offer attractive goods or services, not to its own conduct. In the same vein, a firm may claim that
anticompetitive effects are implausible (and thus unlikely) insofar as the contentious behaviour is
incapable of denying rivals a minimum efficient scale. On the other hand (and this is the reason why
frictions might arise), the authority (or claimant) may retort that such considerations are irrelevant in
the context of the case, or that they are incapable of calling into question a finding of anticompetitive
effects.
65

8.3.5. The confusion between the legal test and the standard of proof
It has become apparent in the past two decades that there is a tendency to conflate the applicable legal
test, discussed in this paper, and the standard of proof.207 Commentators may occasionally fail to
distinguish between the two; or they may refer to one when the substance of the discussion refers to
the other. This confusion is primarily due to the fact that that the two concepts tend to be expressed
in probabilistic terms. This is true, as explained above, of the threshold of anticompetitive effects.
The Court routinely refers to the ‘likely’ or ‘probable’ impact of a practice or transaction. The various
standards of proof are sometimes – at least so in some legal traditions – presented as encapsulating
different probability thresholds (suffice it to think of the expressions ‘balance of probabilities’ and
proof ‘beyond reasonable doubt’).208 In addition, the two questions may arise in the context of the
same case. A challenge against a decision, for instance, may relate both to the appropriate legal test
and about whether the facts have been established to the requisite legal standard.
8.3.6. The confusion between actual effects and the certainty of effects
The concept of actual effects may be interpreted in more ways than one. As explained above, the
Court has consistently used it in contrast with that of potential effect. In this context, it means that the
retrospective analysis of the impact of a practice on competition must consider the actual context in
which it was implemented as well as the contemporary evolution of the market. For instance, the
analysis of actual effects may take into consideration, as in Post Danmark I, that rivals did not lose
their ability and incentive to compete and were in fact able to gain back their main customers. In
practice, however, stakeholders may conflate the time dimension with the probability threshold. More
207
Kalintiri (n 43) 72.
208
For a discussion, see Fernando Castillo de la Torre and Eric Gippini Fournier, Evidence, Proof and Judicial Review in
EU Competition Law (Edward Elgar 2017), 34-36.
66

precisely, there may be a tendency to assume that actual effects are only established where the
expected harm to competition (collusion, exclusion or exploitation) has been fully manifested.
According to this view, anticompetitive effects would not be established unless the retrospective
analysis reveals that rivals have departed from the market.
As can be seen, stakeholders may assume that actual effects are synonymous with a certainty
of effects. This tendency may be displayed both by authorities (or claimants) and by defendants in
competition law disputes. On the one hand, defendants may be naturally inclined to equate the time
dimension with the probability threshold so as to escape liability. As suggested above, claims that
rivals were not excluded during the relevant period, or that one or several parameters of competition
were not affected, can be expected from firms facing or challenging a finding of infringement. On the
other hand, authorities and claimants may display a tendency to dismiss any contemporary evidence
contradicting their theory of harm. In this sense, they may argue that conduct may infringe
competition law even when it fails to fully display its impact on competition and, similarly, that
competition law considers both actual and potential effects. The Krka judgment, discussed above,
provides an example in this sense.
9. Conclusions
The impetus behind this article was to study the notion of anticompetitive effects. It does so by
structuring the case law around a framework that considers the relevant variables. The exercise, which
brings together several strands of the case law across all provisions, shows that it is possible to attach
a concrete and consistent meaning to the notion. Some central questions, including the role and
operation of the counterfactual and the threshold of effects have been answered by the Court. In
addition, it has long been clear that anticompetitive effects amount to more than a mere competitive
disadvantage and/or a limitation of a firm’s freedom of action. Something more, namely a reduction
67

of competitive pressure resulting from a negative impact on equally efficient firms’ ability and/or
incentive to compete, is required.
At the same time, several open questions remain. Since the number of cases in which the EU
courts have engaged in an in-depth evaluation of the impact or transaction is relatively limited, it is
also likely that additional frictions around the meaning and boundaries of the notion of
anticompetitive effects will emerge. In particular, there is a consistent tendency on the part of
stakeholders to conflate the issue of appreciability with the meaning of effects, to equate every
competitive disadvantage with an anticompetitive effect and to downplay the role of the
counterfactual in the analysis. These same tendencies, observed over the years, are likely to be
displayed before the EU courts. Some pending cases will provide an opportunity to shed additional
light on a notion that is central to the EU competition law system.
68

The case law considers both
Time dimension
actual (retrospective analysis)
and potential effects
Actual and/or potential effects
(prospective analysis)
Dimensions of competition
Both the inter-brand and
intra-brand dimensions of
Inter-brand and intra-brand
competition are relevant
competition
Counterfactual Competition means

The counterfactual has an ex
‘competition that would have
ante and an ex post
Benchmark against which existed in the absence of the
dimension
effects are assessed practice or transaction’
The ability and/or incentive

to compete is affected to such
Meaning of effects
an extent that competitive
pressure is reduced
Threshold of effects A threshold of plausibility

A threshold of likelihood
applies to prima facie
applies to conduct subject to
unlawful conduct (‘by
Probability threshold of effects a ‘standard effects’ test
object’)
Fig. 11: The notion of anticompetitive effects – relevant variables
Anticompetitive effects are However, firms may rebut the It is sufficient for firms to
presumed in relation to prima presumption that the practice provide ‘serious indicia’
facie unlawful (‘by object’) is capable of having casting doubts on the
conduct anticompetitive effects capability of effects
The presumption may be

rebutted in three ways
The restriction of competition

would be attributable to the The practice is objectively Anticompetitive effects are
regulatory context, not the necessary to attain a pro- implausible in the relevant
practice competitive aim economic and legal context
Fig. 12: Anticompetitive effects and prima facie unlawful (‘by object’) conduct
69

Identify the mechanism
through which effects would
be manifested: collusion,
exclusion or exploitation
Identify the dimension of

competition affected by the
practice*: inter-brand or intra-
brand
No effects where the practice No effects where there is no

Define the counterfactual:
creates the very competition causal link between the
conditions of competition in
it is said to restrict (objective practice and the effects (no
the absence of the practice
necessity) attributability)
Question 1: will equally

Establish the likely Question 2: will the practice
efficient firms’ ability and/or
anticompetitive effects against reduce competitive pressure
incentive to compete be
the counterfactual** as a result?
affected by the practice?
Once the likely effects

Establish the appreciability of This step is necessary in the
established, appreciability is
the anticompetitive effects context of Article 101 TFEU
assumed in the context of
and merger control
Article 102 TFEU
Fig. 13: Anticompetitive effects under the ‘standard effects’ test

* Practice is used as a shorthand for both practice and transaction
** For a detailed assessment of exclusionary effects, see fig. 5 above
70

SSRN-id3599407 2

Uploaded by

Copyright:

Available Formats

SSRN-id3599407 2

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

SSRN-id3599407 2

Uploaded by

Copyright:

Available Formats

Anticompetitive Effects in EU Competition Law

Pablo Ibáñez Colomo*

Electronic copy available at: https://ssrn.com/abstract=3599407

a prominent role before courts and authorities.

Electronic copy available at: https://ssrn.com/abstract=3599407

practice by the Commission is not wholly uncontroversial.7

Electronic copy available at: https://ssrn.com/abstract=3599407

of anticompetitive effects tend to be available in a fragmentary (and sometimes embryonic) manner

legal test, not the relevant provision.

both prospective and retrospective.

Electronic copy available at: https://ssrn.com/abstract=3599407

give rise to frictions in the medium to long term.

Electronic copy available at: https://ssrn.com/abstract=3599407

2.1. The time variable: actual and potential effects

Electronic copy available at: https://ssrn.com/abstract=3599407

intervention takes place.12

question in the case law is considered in Section 4 below.

2.2. The dimensions of competition and of the counterfactual

2.2.1. Inter-brand and intra-brand competition

It is commonplace to distinguish between the inter-brand and the intra-brand dimensions of

Electronic copy available at: https://ssrn.com/abstract=3599407

competition is universally understood to be more important, it can be expected to take precedence.

2.2.2. The ex ante and ex post dimensions of competition

Electronic copy available at: https://ssrn.com/abstract=3599407

competition (insofar as would-be rivals would be denied access to the market).

both the ex ante and ex post dimensions of the counterfactual.

2.3. The meaning of effects

2.3.1. Between freedom of action and harm to consumers

The assessment of the impact of a practice or transaction on competition is particularly sensitive to

Electronic copy available at: https://ssrn.com/abstract=3599407

are at the opposite ends of the spectrum.

or transaction.17 This is known to be a particularly demanding exercise.

Freedom of Market As efficient Consumer

action* structure competitor welfare

Fig. 1: The meaning of effects

Electronic copy available at: https://ssrn.com/abstract=3599407

2.3.2. The appreciability of effects

make an appreciable contribution to rivals’ departure from the relevant market.

Electronic copy available at: https://ssrn.com/abstract=3599407

2.4. The probability of the effects: plausibility, likelihood and certainty

practice or transaction is certain to have a negative impact on competition.

Electronic copy available at: https://ssrn.com/abstract=3599407

Fig. 2: The probability of effects

Electronic copy available at: https://ssrn.com/abstract=3599407

3.1. Mechanisms through which anticompetitive effects are manifested

3.1.1. Collusion or absorption of a source competitive pressure

3.1.2. Exclusion of a source of competitive pressure

Electronic copy available at: https://ssrn.com/abstract=3599407

3.1.3. Exploitation of market power

obligations, unrelated to the main transaction, upon them.29

Electronic copy available at: https://ssrn.com/abstract=3599407

downstream market (secondary-line effects of an exploitative nature).31

3.2. The place of anticompetitive effects in the various legal tests

A practice or transaction that is incapable of having anticompetitive effects is not subject to EU

Electronic copy available at: https://ssrn.com/abstract=3599407

with the internal market following a case-by-case assessment of their effects.

by Article 102 TFEU insofar as they have anticompetitive effects.

Electronic copy available at: https://ssrn.com/abstract=3599407

presumption that these practices are capable of restricting competition.43