Bart Mayens Thesis EMLE

8/6/2019 Bart Mayens Thesis EMLE

http://slidepdf.com/reader/full/bart-mayens-thesis-emle 1/54

DETTERENCE OF COLLUSION IN EU COMPETITION LAW: AN ECONOMIC

APPROACH

BART MAYENS

Thesis submitted to obtain the degree of

European Master in Law and Economics (EMLE)

Promoter: Dr. B. Targanski

- 2011 -



Bart Mayens

DETTERENCE OF COLLUSION IN EU COMPETITION LAW: AN ECONOMIC

APPROACH

Abstract:

Collusion refers to agreements made between otherwise independent firms with

the goal of raising joint profits hence disturbing the competitive market

equilibrium. The current legal approach to such anti-competitive conduct is one

where the Commission seeks to find evidence of communication and

agreements between undertakings rather than proving actual outcomes on the

market. This thesis demonstrates how such a treatment is unsupported by

economic theory which is incapable of proving the role of communication as a

vehicle to implement collusive prices on the market. As such, a divergent legal

approach is suggested which is coherent with formal economic theories on

collusion.

Promoter: Dr. B. Targanski

- 2011 -



I

Acknowledgements

This thesis is the result of a full year of interesting studies at the European Master in

Law and Economics (EMLE), but would never have been possible without the help of

other people.

Foremost, sincere gratitude is hereby extending to my promoter and assistant Dr.

Bartosz Targanski who guided me through the writing process. He never ceased to offer

help, appealing remarks and very useful input and inspirations throughout the

elaboration of this thesis. All of this made it after all a pleasant work.

I also want to thank my parents, offering me the gratefully appreciated opportunity of

participating at the European Master in Law and Economics (EMLE) program, most

likely the starting point of a hopefully interesting future career.

Finally, I am deeply indebted to all my friends, supporting me in difficult moments and

offering me the chance of changing minds by which new inspirations could arise.

Bart Mayens,

Bologna, July 2011



II

Table of contents

Acknowledgements ........................................................................................................... I Table of contents .............................................................................................................. II General Introduction .......................................................................................................... 3 Chapter 1 Economic analysis of collusion ................................................................. 6

1.1 About collusion ................................................................................................. 6 1.2 Stability of the collusive outcome ..................................................................... 9 1.3 The negative welfare effects of collusion ........................................................ 11 1.4 Conclusion ....................................................................................................... 14

Chapter 2 Collusion and competition law: legal responses ...................................... 15 2.1 Introduction: Article 101(1) and collusion ...................................................... 15 2.2 Scope: what is illegal and what is not ............................................................. 17

2.2.1 Agreements .................................................................................................. 17 2.2.2 Concerted practices ..................................................................................... 19

2.3 Judicial treatments of price fixing ................................................................... 24 2.3.1 By object versus by effect ........................................................................... 24 2.3.2 Application of Article 101(3) ...................................................................... 26

2.4 Conclusion: standards of proof ........................................................................ 27 Chapter 3 An economic view on the legal view ....................................................... 29

3.1 Introduction: setting the stage ......................................................................... 29 3.2 Is the current legal approach economically justified? An overview ...............30

3.2.1 Collusion and its game-theoretic aspects: formal economic theory ............31 3.2.2 Experimental and empirical literature ......................................................... 34

3.3 The right approach? ......................................................................................... 35 3.4 The alternative approach ................................................................................. 37 3.5 Policy implications .......................................................................................... 40 3.6 Conclusion ....................................................................................................... 44

General conclusion .......................................................................................................... 45 List of Figures …………………………………………………………………………..47 References……………………………………………………………………………...48



3

General Introduction

The capitalistic market economy based on perfectly competitive markets is driven by an

invisible hand towards a Pareto-optimal allocation of resources. This theoretical insight

is well known as the first theorem of welfare economics. It results in profit

maximization by firms on the one hand, and utility optimization of consumers on the

other hand as prices are the lowest and quantities produced the highest as is

economically feasible (De Bondt, 2006).

Competition authorities are installed to safeguard this theoretical insight. They act as the

principal patron of consumers and efficient operating firms to prevent them from failing

markets, diverging from an optimal outcome through higher prices and a restricted

output. This is where antitrust, or competition law arises. It provides legal foundations

in order to protect optimal market competition by convicting anti-competitive business

practices which are to the detriment of society.

Collusion is one of these and is characterized by undertakings agreeing to coordinate on

prices and/or output. By doing so, the competitiveness on the market is reduced which

results in higher prices on the market. Such conspiracies are often secret in build-up and

can accordingly prevail undetected while considerably harm consumers.

Metaphorically, it is frequently referred to as ‘cancers on the open market economy’ or

‘supreme evil of antitrust’ .1 Fighting collusion is accordingly one of the top priorities of

the European Commission (“Commission”) in Europe. It seeks to establish an effective

deterrence which is one of the biggest goals in its combat against anti-competitive

conduct.

In its legal answers, the Commission aims to hold a hard stance against collusion. The

legal approach as currently in place focuses on the detection of conspiracies in order for

firms to be found liable instead of proving actual market outcomes. Hence, if parties are

found to be make an agreement they are found to infringe competition law while the

Commission should not proof the actual implementation of it.

1See Monti (2001) and Justice Scalia in the Verizon Case; 540 US 398 (2004) respectively.



4

Although such an approach is fairly uncontested, practical case law has repeatedly

made clear that the Commission tends to neglect economic considerations. In

Polypropylene2 for instance, the Court of First Instance pointed to the lack of economic

evidence shortly after the Commission made its decision. Likewise in Flat Glass3 the

Court held that more economic analysis was necessary in order for the defendants to be

found liable. Consequently, one should ask itself whether it is acceptable for the

Commission to be released from the burden to engage in an economic impact study on

the actual effects of collusion on the market.

In accordance, this thesis seeks to answer this question by taking an interdisciplinary

approach. It aims at reviewing economic literature on collusion and asks itself whether

the current legalistic approach is theoretically consistent with the the micro-economic

aspects of collusionary behavior. In so doing, it is questioned whether the legal

approach as currently in place is in line with economic theory. More precise, the role of

communication as an instrument to implement collusion on the market turns out to be a

blurred one.

This thesis is in line with former work on collusion but differs from it by engaging in a

multi-sided view. While most earlier work focused on the economic and legal aspects

separately (see for instance Ivaldi et al. (2003) or Motta (2003)), this work aims to

interlink both views and searches for consistency between both. As such, it takes an

economic view on the legal view, in line with Whinston (2003).

Consequently, this thesis aims to, (i) clarify the basic economic building-blocks of

collusion and analyze how it negatively harms welfare; (ii) presenting the legal

approach as currently adopted by the Commission; (iii) comparing the legal view with

the economic view and arguing whether the economic analysis validates the legal view;

(iv) presenting a divergent approach from the current one which is more in line with the

economic view.

2

Commission Decision 86/398/EEC, OJ 1986, L230/1.3

Commission Decision 89/93/EEC, OJ 1980, L33/44.



5

This work will be structured as follows. The first chapter revises the economic

mechanisms that lie behind the functioning of collusion. It revises the equilibrium

characteristics and analyses how it is able to harm consumers. The second chapter lines

out how the Commission aims at deterring collusion by focusing on the scope and

judicial methods of the legislation. Chapter 3 then compares the insights gained in both

previous chapters. It elaborates the economic view by reviewing the literature and

investigates whether it is consistent with the legal view as presented in Chapter 2. It

equally presents an alternative approach to deal with collusion and presents some brief

policy implications based on practical case law.



6

Chapter 1 Economic analysis of collusion

In order to proceed to the main body of this work, a profound economic understanding

of collusion is a necessary first step approach. Accordingly, this chapter aims at

fulfilling this need by summarizing the economic ingredients of the functioning of

collusion. The next chapter will then focus on the legal aspects of coordinative

behavior.

This chapter will be structured as follows. The first section provides a definition, lines

out the scope of different related concepts and puts forward the economic process

behind collusionary practices. The second section explains why collusion is typically

unstable and thus difficult to sustain. Finally the third section clarifies the detrimentaleffects of collusion on society, pointing out to the welfare losses that arise.

1.1 About collusion

Collusion refers to coordinative behavior between legally independent firms aiming at

raising their collective profits by explicitly agreeing to coordinate on prices or output

(Connor, 2008). The conspiracy aims at moving up the market price away from the one

that would prevail in the competitive setting since joint decision making (instead of

independent ones) internalizes the competitive burden that would be present absent the

agreement (Martin, 2001).

Coordination can either be done explicitly as is the case with cartels, or can be done in a

more secret and subtle way. Participants in a cartel communicate with each other by

organizing meetings, making use of associations or even sign contracts. On the other

extreme, coordination can also be done with fewer forms of explicit communication

(informal), or even without communication at all (so called ‘tacit collusion’), see Figure

1 depicting the level of formality of the coordination.

Tacit Informal agreements Cartels

Figure 1: Tacit collusion versus cartels



7

More generally speaking capturing both explicit and less overt agreements, one can

define collusion as any situation where firm’s prices turn out to be higher than the

competitive benchmark (Motta, 2003; Whinston, 2003).4 Hence, the elevated price can

be reached by organizing explicitly by means of a cartel or can even be sustained by not

meeting up at all. In the latter case, tacit collusion, participants thus do not communicate

directly with each other (Ivaldi et al.,2003).5

Besides price arrangement (explicit or tacitly), firms may collude on sharing the market

and each serve a specific part of it. As a consequence, local monopolies are created and

thus the ability to increase prices (cf. ‘market allocation’). Equally, they can allocate

quotas as to lower the quantity sold, hence again lifting up prices (Motta, 2003).

The most prominent approach however to increase joint profits is to directly agree on

the prices charged to consumers, i.e. price fixing .6 That is, instead of each setting the

price for the good separately by which competition between both would prevail, one

single agreed upon price will be set which is relatively higher than the one absent the

coordination. In doing so, participants aim at achieving the monopoly profit instead of

the competitive benchmark.

Figure 1 captures this insight by comparing the situation with and without collusion.

Think of two pizza shops in town. Without coordination, firms compete with each other

by which the competitive equilibrium arises, i.e. the market price equals marginal costs,

say . Thus, in terms of the figure, IJJ . Indeed, a pizza shop’s

4The competitive benchmark is the situation absent collusion. In technical economic terms, collusion

would arise if the price is higher than the one-shot Bertrand equilibrium price when firms decide on

prices, or if the quantity is lower than the one-shot equilibrium quantity if firms compete à la Cournot and

thus decide on quantities. See for example Tirole (1988) for a general approach.

5Since this thesis is concerned with the outcome of coordination rather than the way it is achieved, both

notions, cartel and collusion, will be used as synonyms throughout the remainder of this work. This

approach is equally followed throughout most of literature.

6Henceforth, this thesis shall only refer to price-fixing cases, though it being understood that other

similar practices do exist. Consequently, collusion can be read as a synonym to price fixing in the

remainder of this work.



8

ability to charge a higher price (hence, its market power) is limited by the presence of

the other shop since nobody would spend more for the same product. On the figure this

is indicated as the “Competitive Price and Output Equilibrium” and can be seen as the

point where the marginal cost (H) curve intersects with the demand curve (or the

average revenue ( ) curve).

Figure 2: Price fixing7

Now consider both shops to coordinate on the price they charge. Since decision making

is done jointly the competitive constraint that was present absent the agreement is

internalized. Hence, the firms can agree on charging the monopoly price (say ),

naturally to the detriment of consumers. On the figure this is denoted as the

“Monopolist profit maximizes here” equilibrium which leads to JJJ

and is indicated by the point where the marginal revenue (H) curve crosses the H

curve. Collusion makes firms thus to decide on prices as would a monopolist do.

Accordingly, collusion can equally be described as the outcome where firms set prices

close enough to the monopoly price (Kühn, 2001). Note that the example above

assumes the collusive price to equal the monopoly price. In more realistic settings

however with imperfect information and communication between participants and with

7

Taken from http://www.proprofs.com/flashcards/cardshowall.php?title=MonopolyOligopolyPerf-Comp,

10/05/2011.

http://www.proprofs.com/flashcards/cardshowall.php?title=MonopolyOligopolyPerf-Comp






9

not all firms participating in the cartel, participants will often not be able to achieve the

full collusive price (Motta, 2003).

More precisely, firms can make mistakes and chose a collusive price which is not

jointly optimal since it is often hard to perfectly communicate with each other. In the

example above, imagine that each pizza shop thinks that the other would set a price of

instead of for example because there is lack of communication or because

external demand and supply shocks make it difficult to estimate correctly the full

collusive price. Then again collusion arises, but now at a level suboptimal for both firms

(Motta, 2003). Therefore, the collusive price can be anywhere in the region {H

with denoting the monopoly price.

Which economic mechanisms lie behind tacit price-fixing? Under tacit collusion, the

participants do not communicate directly with each other, yet they achieve a higher

equilibrium price on the market relative to the competitive setting.

Firms can accomplish such an outcome by observing and anticipating conduct of its

rivals. If one firm makes profits by selling at a price above its marginal costs while

spare capacity is available on the market, its rival can follow up and equally raise its

price to the same level. Hence, the announcement of a price raise by one firm can act as

a signal to other firms to equally raise prices rapidly. Indeed, competitors know that the

price-raising firm cannot maintain such a price if the others do not follow up since all

consumers would buy at the lowest price (Rodger & MucCulloch, 2004).

All being said, it results from the above that the conspiracy suffers from an internal

instability problem. Indeed, each firm will have the incentive to lower the price below

the collusive one since then it would increase its turnover. So how is collusion remained

in place? And what factors will influence the degree of stability of the agreement? The

next section will provide an answer to those questions.

1.2 Stability of the collusive outcome

Using the example above, imagine that both pizza shops meet on the market place and

decide to fix prices. Although enjoying a higher collusive price, both shops will have

the urge to deviate from the agreement. That is, cutting prices would increase the



10

individual profit of the cheater since turnover would increase while still benefiting from

a high unit margin (Motta, 2003). In such a case, collusion would break and competition

would prevail again.

Stated otherwise, both firms have the incentives to cooperate with its rival by which

joint profits are maximized. But they also have a strong incentive to cheat, i.e. to lower

the price beneath the collusive one. The collusive outcome thus tends to be a highly

unstable one (Gwartney et al, 2009).

The acknowledgement that any participant has the urge to break the collusive outcome

points out two general features of collusion: the ability of detecting deviations and the

degree in which the competitors can punish the cheater through lowering his own price

to the detriment of the profits of the punisher (Ivaldi et al., 2003; Whinston, 2003;

Motta, 2003). The easier it is to detect and punish the deviator, the more stable the

collusive outcome will be (Motta, 2003; Whinston, 2003).

This is an intuitive result. Indeed, if the deviation is observed, the other participant will

naturally try to counter it by equally lowering his price. That is, he will retaliate on the

cheater by which a price war occurs, reducing profits of all participating firms. Hence, if

cheating is detected and punished as well, any collaboration in the future is made

impossible. Punishment thus keeps the members honest, since they anticipate to lose out

on the profitability of the collusion in the remaining periods.

In more formal words, each participant makes the trade of between cheating and reaping

full monopoly profits for one single period and nothing after versus staying honest and

sharing the market ad infinitum. Therefore, collusion is sustained if:

-

- %

-

- - % - 4 - 0

with denoting the collusive profit and the discount factor (Ivaldi et al., 2003).8

8Note that it is assumed both firms (two in this modeling) to have the same discount factor and price

competition yielding a price equal to marginal cost, thus competing à la Bertrand. See Ivaldi et al.

(2003).



11

It is generally accepted in literature that some industry factors can increase the

likelihood of collusion based on the two aforementioned features, see for example

Feuerstein (2005) for a summary or Grout & Sonderegger (2005) for an empirical

approach.

Consider for example the number of firms on the market (i.e. concentration). Other

things equal, the more firms active on the market the less sustainable collusion will be

since the gain of deviating from the collusive price will be immense by capturing the

whole market. On the other hand if only two firms would be active on the market

colluding with each other, both get half of the market. Accordingly, gains from

deviating would be relatively much lower by which collusion becomes more attractive

(Motta, 2003).9

To illustrate another factor, think of the frequency and regularity of orders. Other things

equal, collusion will be harder to sustain if firms interact less frequently (less recurrent

orders) since punishment would only follow after a longer time period, thus being less

effective. Cheating thus becomes more attractive (Motta, 2003).

The analysis above clearly indicated the economic factors behind collusion regarding

businesses. Fixing prices makes the participants able of raising profits through lowering

competition on the market but it suffers from an internal stability problem, i.e. the

temptation to deviate from the collusive price. Yet, little attention has been given to

how this practice affects consumers. The next section will revise how collusion lowers

consumer welfare.

1.3 The negative welfare effects of collusion

As clear from now, a successfully implemented cartel raises profits of the participants

by means of charging a higher price to consumers. Consequently, consumers will be

9Therefore collusion is often a much bigger threat in oligopolies: industries where only a few producers

are active. This is even more the case when dealing with tacit collusion: if spare capacity is available on

the market and entry barriers exist reducing outside competition, firms will be much more likely to follow

the price raise of a competitor. In accordance, tacit collusion is often referred to as the ‘oligopoly

problem’.



12

worse off since they have to spend more resources compared to the situation absent of

collusion.

Figure 2 depicts in depth how welfare is negatively affected. The competitive

benchmark is again the situation where the demand curve crosses the marginal cost

curve H. Output on the market is while price equals marginal cost; H . As

is well known, this perfect competitive setting is Pareto-efficient: prices are the lowest

and quantities produced the highest as is economically feasible (De Bondt, 2006). If

now a cartel is formed encompassing all firms in the industry, they will maximize joint

profits and will set a price equal to the monopoly price, i.e. . This increase in

price has two distinct effects on welfare: an overcharge and a deadweight loss (Connor,

2008).

Figure 3: Welfare effects of collusion10

Regarding the former, a part of consumers will pay the monopoly price for the product

instead of the competitive price.. This effect is actually nothing more than a shift of

welfare from consumers to producers. Absent coordination, producers made zero profits

since price equaled marginal cost. When colluding, profits mount up to . I .

In Figure 2 this overcharge is depicted as the green rectangle and can be interpreted as a

sort of monopoly tax on the good consumers buy (Connor, 2008).

10Taken from http://www.ken-szulczyk.com/lessons/production_lecture_03.html, 22/05/2011.

http://www.ken-szulczyk.com/lessons/production_lecture_03.html





13

The latter, deadweight loss, refers to the fact that some consumers will not be able

anymore to buy the product at the higher price. They considered of buying the good but

will leave the market since they cannot afford the product anymore at the higher

monopoly price. On Figure 2 this is denoted as the triangle “Deadweight Loss” and is a

pure economic waste. Indeed, for all consumers in the area { on the horizontal

abscess their willingness to buy as measured by the demand curve exceeds the marginal

cost of the product by which transactions would be efficient. Because of the monopoly

price charged however, these set of consumers will not buy the product at all and seek

an inferior substitute (Connor, 2008).

Stated otherwise, before collusion consumer welfare equals the big rectangle above the

H curve and below the demand curve. Producer surplus is zero and total surplus is the

sum of both. After collusion, the green shaded triangle transfers into producer surplus

while the pink rectangle is neither consumer neither producer surplus anymore

(deadweight loss).

Some surveys have aimed at estimating welfare losses arising from collusion. Werden

(2003), using a dataset of 13 cartels found an average overcharge of 21%, while Posner

(2001) found a mean of 49% investigating a similar number of cartels. Using a larger

dataset, Griffin (1989) estimates the average overcharge around 46%. In contrast to

overcharges, less is known about dead weight losses since they require knowledge about

the elasticity of demand (Peterson & Connor, 1996). Yet, in most industries it can be

approximated to mount up between one-fifth and one-tenth as large as the overcharge

(Connor, 2001).11

These figures clearly give an indication of how society should seek to prevent collusion

to arise. Accordingly, competition law attempts to safeguard the well being of

consumers by prohibiting coordinative practices which will be the topic of the next

Chapter.

11While the negative welfare effects are clear-cut, collusion could in theory also lead to pro-competitive

effects, see for example Whinston (2003) who provides an example. Nevertheless, it is unlikely for such a

situation to arise.



14

1.4 Conclusion

This chapter explained the economic mechanisms behind collusion, i.e. the coordinative

behavior between undertakings in order to fix prices on a level higher than the

competitive benchmark. By doing so, the participants jointly decide prices and

accordingly act as a monopolist would do.

Such a collusive equilibrium is an unstable one since each firm has the temptation to

cheat on the agreement, i.e. to lower the price below the collusive one. Thus, the less

possibility to detect and punish deviators the more unsustainable collusion will be.

Collusion is able to considerably harm consumers. Not only will some consumers pay

more for the same good relative to the competitive situation, some other consumers will

also leave the market all together since they cannot afford the respective product

anymore. Both effects, the overcharge and the deadweight loss, have a negative impact

on consumer welfare.



15

Chapter 2 Collusion and competition law: legal responses

The first chapter presented the economic functioning of collusion: it raises collective

profits of the participants by fixing the price charged to consumers on a higher level

relative to the competitive benchmark. Consumers suffer from higher prices by which

society incurs a welfare loss.

Therefore, authorities naturally seek to prevent these detrimental effects by installing

proper laws ruling out anticompetitive behavior by firms. This chapter aims at summing

up the current approach taken by the European Commission in preventing collusion to

arise, hence protecting consumers on the market.

The first section introduces the topic by presenting Article 101(1) of the Treaty on theFunctioning of the European Union (TFEU) prohibiting collusion and discusses the

necessity for a further elaboration on the concepts used in it. The second section aims at

lining out the scope of the article by taking a look at former case law in order to know

which different modes of collusion are illegal and which are not. The third section then

gives insight in the legal treatment of price fixing cases. Knowing the scope and legal

treatments, the following section draws conclusions on the standards of proof needed to

be found liable. The last section finally sums up the current legal approach.

2.1 Introduction: Article 101(1) and collusion

European competition law prohibits agreements between two or more firms who restrict

competition, as drafted in Article 101(1). Article 101(1) reads as follows:

‘The following shall be prohibited as incompatible with the internal market: all

agreements between undertakings, decisions by associations of undertakings and

concerted practices which may affect trade between Member States and which have as

their object or effect the prevention, restriction or distortion of competition within the

internal market: (a) directly or indirectly fix purchase or selling prices or any other

trading conditions, (b) […]’ 12

12Consolidated version of the TFEU, Article 101(1) (ex Article 81(1) TEC), OJ C 115, henceforth

“Article” (markings added). Note that the Article consists of three paragraphs, each referred to as 101(1),

101(2) and 101(3). Paragraph 1, as written down, lines out the scope of liable practices. Paragraph 2



16

Although price fixing is in general clearly captured by Article 101(1) (see subsection

(a)), one can immediately notice the broad and loose concepts employed. Indeed, it is

difficult to see what exactly is meant with ‘agreements’ and ‘concerted practices’. Yet,

as we know from the first chapter, collusion can arise in many different forms: firms can

make a formal agreement, firms can restrict themselves to more informal concertation

or they can even not communicate at all by which they collude tacitly. It is however a

priori unclear which kind of collusion is captured by Article 101(1).

As a consequence, one should line out the scope of Article 101(1) by examining its

contents in more detail in order to know what forms of collusion are prohibited by the

Article. In this respect, the concepts agreements and concerted practices used in the

Article should be specified more clearly by looking at former case law. In doing so, one

can then gain insight in what forms of collusion are illegal and which are not by

knowing the precise denotation of the concepts ‘agreement’ and ‘concerted practice’.

The scope of the Article will be discussed in Section 2.2.13

A second point of attention concerns the legal treatments of cases brought before court.

Put different, should the plaintiff show actual outcomes on the market by looking at

pricing behavior of the defendants, or conversely is intent of collusion sufficient for

being found liable? In this context, the concepts object or effect as can be read in the

Article gives us indications of whether the former (intent, i.e. by object ) or the latter

(market outcomes, i.e. by effect ) approach is followed when dealing with price-fixing

cases. This will be dealt with in Section 2.3.

In lining out the scope and the judicial treatments, one can then summarize more

concretely what the standards of proof are in order for the defendants to be found liable.

describes the consequences of breaching the practices as described in 101(1): the agreement is declared

void. Paragraph 3 allows under certain strict conditions the inapplicability of 101(1) which is however not

relevant regarding price fixing cases, see further under Section 2.3.2.

13Note that we do not devote separate attention to the notion decisions by associations of undertakings

since its interpretation is quite straightforward. It is easily understood that associations who act as an

intermediary can be found liable as well as each related company, see for instance UEFA Champions

League, OJ 2003 L291/25.



17

Keeping in mind the diverse modes of collusion, is only an explicit agreement on prices

illegal, or can firms tacitly colluding infringe competition law as well? Additionally, are

firms merely agreeing on prices in a meeting liable of collusion, or should they also

implement the higher prices on the market? In accordance, Section 2.4. concludes the

current legal approach by summing up standards of proof.

Before proceeding however, note that the Commission attempts to combat collusion

also in different ways apart from competition law ex-post, being the subject of

discussion here. Leniency programs for example encourage cartel members to

collaborate with authorities in return for total immunity from fines or a reduction of

fines which the Commission would otherwise have imposed on them.14 As such, self-

reporting is expected to make enforcement of collusion more effective while saving

resources, see Motta and Polo (2003) and Aubert et al. (2003) for how such a scheme

affects cartel members’ behavior.

Next to granting leniency programs, authorities could equally attempt to anticipate ex-

ante. Motta (2003) for example, argues how the Commission should focus more on a

structural approach, i.e. to adapt the market structure in a way less suitable for collusion

to arise. In doing so, authorities would then focus on those factors who facilitate

collusion to sustain (see Section 1.2 above) and accordingly try to ban them (e.g. resale

price maintenance which makes prices more observable, see Motta (2003) who even

calls for such a ‘black list’ of practices).

2.2 Scope: what is illegal and what is not

2.2.1 Agreements

The concept of ‘agreement’ is a hazy one. One could interpret it very narrowly, by

which it would imply only formal agreements such as a written contract. Conversely, a

broad understanding would entail even oral and informal arrangements.

14

See Commission Notice on Immunity from fines and reduction of fines in cartel cases, OJ 2006,

C298/17.



18

In Polypropylene,15 the Commission found non-binding oral agreements which lacked

any kind of enforceability to infringe Article 101(1). Likewise, in ACF Chemifarma v.

Commission16

an unsigned ‘gentlemen’s agreement’ was equally found liable. Above

former case law clearly indicates a generous interpretation which is thus not restricted to

a traditional view where an agreement should be legally binding to apply (Ritter &

Braun, 2005). As noted by the Court of Justice:

‘The minimum requirement for there to be an “agreement” is an expression of

joint intention of the parties involved to conduct themselves on the market in a

specific way […]’ 17

Hence, all that matters is the expression of intentions of reaching consensus between

parties in order to act in a non-unilateral way. In the extreme, even seemingly unilateral

behavior can be imputed to an agreement when tacit acceptance is observed. As The

Court stated in BAI and Commission v. Bayer ,18 it is necessary for agreements to be

concluded by tacit acceptance that:

‘the manifestation of the wish of one of the contracting parties to achieve an

anti-competitive goal constitutes an invitation to the other party, to fulfill that

goal jointly.’ 19

In the same token, participating in meetings without verbalizing any intent of

cooperation suffices to prove involvement in an agreement (Ritter & Braun, 2005).

The Commission thus clearly attaches a very broad scope to the concept of an

agreement. Indeed: there is no need for the agreement to include enforcement

procedures, neither it should be written, neither it should be physical: verbal or even

15. Cited above.

16Commission Decision E.C.R. 661, Cases 41/69 [1970].

17 Jaeger/Opel Norge, EFTA Court Of Justice, 1998, E-3/97 at 35-36.

18

Cases C-2 and 3/01, ECR I-23 [2004].19

Cases C-2 and 3/01, BAI and Commission v. Bayer , ECR I-23 [2004], at 102.



19

tacit understandings are sufficient just as e-mails and phone calls (Ritter & Braun,

2005).

2.2.2 Concerted practices

The term ‘concerted practice’ has equally been given a very broad interpretation. It was

first defined by the Court in Dyestuffs:20

‘[…]a form of co-ordination between undertakings which, without having

reached a stage where agreement properly so called has been concluded,

knowingly substitutes practical co-operation between them for the risks of

competition.’ 21

As noted, the term ‘concerted practice’ can be read as a ‘meeting of the minds’ whereby

parties coordinate with each other without reaching the stage where a full-fledged

agreement is made. However, it does enable the parties to anticipate with greater

certainty the future conduct of their competitors, thereby reducing the competitive

burden (Van Bael & Bellis, 2005). In other words: a concerted practice refers to mutual

cooperation or direct/indirect contact in order to influence actual or future conduct on

the market (Jones & Suffrin, 2008).

Indeed, in Suiker Unie v. Commission,22 the Court stated:

‘[…]any direct or indirect contact between such operators, the object or effect

whereof is either to influence the conduct on the market of an actual or potential

competitor or to disclose to such a competitor the course of conduct which they

themselves have decided to adopt or contemplate adopting on the market.’

23

The concept of ‘concerted practice’ and its according scope as defined above allows for

some observations which will be dealt with below each in turn.

20Case 48/69, ICI v. Commission ECR 619 [1972].

21Ibid, at 64.

22

Cases 40/48, 50, 54/56, 111 and 113/73 ECR 1663 [1975].23

Ibid, at 174.



20

2.2.2.1 Agreements and concerted practices

One can observe the close connection between ‘agreements’ and ‘concerted practices’

where the latter obviously refers to a more looser type of agreement. Nonetheless, it is

hard to see where an agreement stops and a concerted practice begins (Rodger &

MacCulloch, 2004).

By adding ‘concerted practices’ to the scope of Article 101(1), the Commission clearly

aims to capture all forms of coordination between firms: if not found liable under the

concept of ‘agreement’, then still the safety net of ‘concerted practice’ exists. It can be

argued that in practical case law no real distinction is made between both.

2.2.2.2 Tacit collusion and concerted practices

Since the concept is broadly defined the Commission is given space to intervene in

cases where coordination is less overt, i.e. tacit collusion. Recall that tacit collusion

exists where the coordinating firms do not communicate directly with each other, let

alone they form an agreement (see Section 1.1.). Hence one can ask itself how exactly

the Commission intervenes in cases where there is no explicit proof of direct

communication. Therefore: what is the legal treatment of tacit collusion?

Seminal legal precedents can be found in Dyestuffs and Suiker Unie v. Commission.24 In

both cases, the defendants communicated price increases to the market after which

competitors followed suit and found themselves operating on a higher equilibrium price.

As a result, the Commission observed prices to jointly raise and thus change in a similar

way. Dealing with so called ‘price parallelism’, the question accordingly arises if

parallel conduct infringes Article 101(1). Put different: can market behavior by which

prices are raised in the hope that other competitors will follow be seen as a concerted

practice and hence infringe the Article (Rodger & MacCulloch, 2004) ? In Dyestuffs, the

Court said:

‘Although parallel behavior may not by itself be identified with a concerted

practice, it may however amount to strong evidence of such a practice if it leads

24Both cited above.



21

to conditions of competition which do not correspond to the normal conditions

of the market […].25

It went on to say that:

‘Although every producer is free to change his prices, taking into account in so

doing the present or foreseeable conduct of his competitors, nevertheless is is

contrary to the rules on competition contained in the treaty for a producer to co-

operate with his competitors.’ 26

In Suiker Unie v. Commission, the Court of Justice held that:

‘[…] this requirement of independence does not deprive economic operators of

the right to adapt themselves intelligently to the existing or anticipated conduct

of their competitors.’ 27

Hence, in both cases the Court did not deprived firms of the right to adapt themselves to

market behavior of their competitors, as long as such parallel conduct can be explained

by the normal forces of the market. Indeed, it would be insensible to convict firms just

because prices are observed to follow the same evolution. Such price parallelism can be

caused by other factors apart from the intention to collude, think for example of an

exogenous shock such as a raise in mutual costs due to inflation.

Yet, the above mentioned cases clearly demonstrate the ability of the Commission to

intervene in such conscious parallelism cases. Indeed, in Soda Ash28 the Commission

literally said:

‘[…]there is no need for an express agreement in order for article [101] to

apply. A tacit agreement would also fall under Community competition law.’ 29

25Cited above, at. 66.

26Ibid, at. 118.

27

Cited above, at 173-174.28

Soda-Ash-Solvay, ICI, OJ 1991 LI 52/1.



22

Former case law thus shows the Commission’s capability to intervene in cases of tacit

collusion under the Article.30 It should however prove that the underlying reason for the

parallel price levels is a concerted practice since such a practice automatically leads to

price parallelism while the opposite does not hold. The way this is usually done is by

making use of so called ‘plus evidence’ (‘plus factors’), indicating that parallel prices

are the result of collusion rather than normal forces of the market (Rodger &

MacCulloch, 2004).

Such circumstantial evidence can refer to facilitating factors for collusion to hold (see

Section 1.2.), the defendant’s record of past violation of collusion-related competition

law, industry factors that facilitate collusion or more blurred elements such as the

presence of rational motives to collude.31 Nevertheless, this approach has been heavily

critized since it is hard to see what is ‘enough’ circumstantial evidence and what not,

see for example Kovacic (1993).

Therefore, many scholars have doubted the appropriateness of the Article concerning

tacit collusion, see for example Motta (2003) and Turner (1962). The main argument

raised is that tacit collusion is the result of structural factors on the market who make

tacit collusion easy to arise rather than behavioral factors, i.e. communication between

parties (Rodger & MacCulloch, 2004).

Accordingly, the question arises whether there exist alternative approaches to deal with

tacit collusion. A more sensible approach would be to impose structural remedies, i.e. to

maintain a very competitive market by which collusion is highly unlikely to arise, see

also above. As such, entry barriers could be seeked to remove, allowing new firms to

29Ibid, at 55.

30Nevertheless, in Wood pulp (C-89, 104, 114, 116, 117 and 125-129/85) [1993] 4 C.M.L.R. 407) ,

equally a case dealing with price announcements, the Court limited the concept of ‘concerted practice’ by

noting that ‘[…], the system of quarterly price announcements on the pulp market is not to be regarded as

constituting in itself an infringement of Article [101].’ Still, concertation has been found easily in many

other cases. One can conclude that it is still not fully clear how price announcements are legally dealt with

(Korah, 2007).31

See Kovacic (1993) for an elaborated list of plus evidence.



23

enter the market and raise competition.32 Equally, merger controls could help to block

mergers who considerably increase concentration on the market (Korah, 2007).33

Finally, sophisticated economic analyses (making use of econometric tools) can be used

in order to filter out price effects who are the result of collusion and which are not, see

also Chapter 3.

2.2.2.3 Exchange of strategic information and concerted practices

Often, firms will seek to share strategic information with competitors in order to

coordinate their behavior on the market. Such information can for example refer to

individual price data, discounts, costs, output levels, capacities and quantities; all of

which would usually be held strictly confidential (i.e. ‘strategic’) (Van Bael & Bellis,

2005).

Although there is no direct link between conspiring and sharing strategic information,

the exchange of it is illegal under Article 101(1). Indeed, sharing strategic data between

competitors makes firms able to anticipate with greater certainty future conduct of their

competitors, by which collusion is strongly facilitated. Therefore, the concept of

concerted practices equally captures the exchange of strategic data. Accordingly, firms

who take part in such practices will be found liable, see for example COBELPA34 and

UK Tractors.35 As noted in the Guidelines on the Applicability of the Article:36

‘communication of information among competitors may constitute an agreement

[or a] concerted practice (…) with the object of fixing, in particular, prices or

quantities.’ 37

32One should note however that the removal of entry barriers often conflict with political forces (Korah,

2007).

33As a result, it has been debated to deal with tacit collusion under Article 102, referring to collective

dominance.

342 C.M.L.R. D28. [1977].

35 UK Agricultural Tractor Registration Exchange, OJ 1992, L68/19.

36Guidelines on the Applicability of Article 101 of the Treaty on the Functioning of the European Union

to horizontal co-operation agreements, 2011/C11/01, henceforth “Guidelines”.37

Ibid, at. 59.



24

2.3 Judicial treatments of price fixing

The judicial treatment of cases brought under Article 101 consist of two consecutive

steps. The first step, under Article 101(1), considers the anti-competitive nature of

agreements or concerted practices either by their object or their effect. The second step,under Article 101(3), becomes relevant only when the anti-competitiveness is

determined within the scope of Article 101(1) (i.e. see above). It involves an assessment

of the likely pro-competitive effects of a certain practice and balances those with the

anticompetitive effects. If detrimental effects outweigh pro-competitive effects, the

defendant will be found liable.38 Each will be dealt with in turn.

2.3.1 By object versus by effect

We repeat Article 101(1):

‘[…] and which have as their object or effect the prevention, restriction or

distortion of competition within the internal market.’

As can be seen, the Article makes a distinction between agreements and concerted

practices that have the object of restricting competition and others that have the effect of

restricting competition, often referred to as per se and rule of reason respectively. This

distinction is vital since it determines the legal treatment of anti-competitive agreements

brought before courts. It is also essential in order to proceed to the remainder of this

work, where it will be debated which of these approaches is most adapted to deal with

price-fixing cases.

Let us first consider concertation that is dealt with by object , i.e. per se. Cooperation

that distorts competition by object refers to those anti-competitive actions that are in

their very nature potentially harmful. Such arrangements are set up with the sole

purpose of achieving a certain outcome which harms consumers and accordingly lowers

welfare.

38Guidelines, at 20.



25

Dealing with such cases, the Commission holds the burden of proof to establish the anti-

competitive object of a certain case. In so doing, it is however not compulsory for the

plaintiff to show actual effects of the suspected practice. Put different: the mere finding

of the existence of the anti-competitive conduct suffices in order to be found liable:

there is no need to proof the consequences of it on the market. Hence, uncovering a

practice which has an anti-competitive goal automatically leads to conviction since

there is no need to engage in a market impact study. In other words: intention is what

matters, not the actual outcomes.39

Next we consider concertation that is dealt with by effect , i.e. rule of reason. Contrary to

a by object -approach, the determination of the anti-competitive object of a practice is

not sufficient in order for the defendant to be found liable. Instead, the Commission

holds the burden of proof to actually show the detrimental effects on the market, both

potential and actual. Accordingly, a market impact study is crucial in order to show how

exactly welfare was lowered.40 Following this approach, a profound economic analysis

of the market is therefore crucial (Rodger & MacCulloch, 2004).

How are price fixing cases dealt with? The Guidelines state:

‘Price fixing is one of the major competition concerns arising from

commercialization agreements between competitors. Such agreements are

therefore likely to restrict competition by object.’ 41

This by object -approach is overly applied in practical case law. In Polypropylene42 for

instance, there was clear evidence that parties were colluding yet there was hardly any

effect to be found on the market. The Court of First Instance however held that the

intent to collude is sufficient, regardless of actual effects. This approach does not differ

39Guidelines, at. 24-25.

40Guidelines, at. 26.

41

Guidelines, at. 234.42

Cited above.



26

whether collusion is found by means of an agreement, or whether a concerted practice is

found to exist. Concerning the latter case, see for example Shell 43 and Rhône-Poulenc.

44

Hence, when dealing with price fixing cases, the Commission has to proof whether the

defendants had conspired, either done by means of an agreement or either done by

means of a concerted practice. Whether the concertation was also implemented on the

market (as shown by its impact on prices, i.e. higher market prices in the period of

collusion) is irrelevant.

On the one hand this approach seems logic: the anti-competitive effects of collusion are

clear-cut and well-known (see Section 1.3), so why spend additional resources on the

investigation of market outcomes? Yet, as will be argued in the following chapter, some

economic counter-arguments on this legal view can be brought forward as well, raising

attention to the unsustainability of this approach and arguing for an impact-based

treatment, see Chapter 3.

2.3.2 Application of Article 101(3)

Once the anti-competitive object or effect is determined, the second step consists of

investigating whether a possible exemption can be granted of Article 101(1) under

certain conditions (improving production or distribution of goods or promote technical

or economic progress, ensure that consumers receive a fair share of these benefits, not

containing any indispensable restrictions and not substantially eliminate competition).45

The Commission will then weight anti-competitive effects of the practice with its

efficiency gains to the benefit of welfare. If pro-competitive effects outweigh

detrimental effects, the defendant will not be found liable. Note that the burden of proof lies with the defendants in order to show these beneficial effects.

Clearly, in order for Article 101(3) to apply, noticeable efficiency gains have to be

proven, considerably benefiting consumers. As we know from Section 1.3., collusion

43 Shell v. Commission, T-11/89, [1992], E.C.R.

44

Rhône-Poulenc v. Commission, T-1/89, [1991], E.C.R.45

Guidelines, at. 46-47.



27

does not lead to such a result and if it would, this situation would be highly unlikely. As

such, Article 101(3) is completely irrelevant when considering price fixing cases. In

other words: there is no possibility to invoke Article 101(3) as a defence since collusion

is de facto always to the detriment of consumer welfare.

2.4 Conclusion: standards of proof

From the above analysis we should now be able to summarize the standards of proof

regarding collusion, i.e. to fix prices on a level higher than the competitive benchmark

by which overall welfare is reduced. Any evidence of either an agreement or concerted

practice, both within the meaning and scope as lined out above, is sufficient to infringe

Article 101(1).

Agreements can be written (such as e-mails or letters) or can be oral (i.e. meetings). The

formality of these agreements is irrelevant: even very informal notes can constitute as

evidence, as well does communication on a golf court or a gentlemen’s agreement. Note

also that there is no need to show a formal consent: parties acting in accordance are

supposed to agree. If, for instance, an undertaking participates in a meeting without

expressly agreeing, it will be found liable unless it publicly distances itself from what

have been said (Ritter & Braun, 2005).

Concerted practices are found whenever prices are observed to follow the same trend,

accompanied by plus evidence showing that this is the result of collusion instead of a

normal functioning of the market. Equally, firms sharing strategic information by e-

mails, letters, oral communication, etc. will be found liable.

There is no need for the Commission to proof any visible effects on the market. If undertakings are found to make an agreement or a concerted practice they will be found

to infringe competition law without further investigations. Hence, whether prices did

indeed stabilized at a higher level is irrelevant. In other words: merely the fact that firms

communicate leads the Commission to believe that collusion will also be implemented.

Accordingly, Article 101(1) is put in place to prohibit such communication.

Even though this legal approach seems at first sight acceptable, it is still questionable if

this legal view is also accepted by economists. Is a by object approach where the focus



28

is on intentions rather than what is priced on the market a right way to deal with price

fixing cases? The next chapter will argue that this approach might be inappropriate by

making use of an economic view instead of a legal one. As such, it will be argued that a

by effect approach might be more suitable.



29

Chapter 3 An economic view on the legal approach

The first chapter gave us insight on the economic functioning of collusion. It was

presented how such a concertation raises prices but in the same time is inherently

unstable. The second chapter presented the legal view as currently adopted, showing

how courts deal with price-fixing cases brought before court. The present chapter

combines the insights gained in the preceding ones. It aims at arguing whether the

current legal approach is the adequate one by making use of economic theory of

collusion. More precise, it is questioned whether economic literature validates the legal

treatment of price fixing cases.

This chapter will proceeds as follows. The first section introduces the debate byrecalling the current legal treatment together with its according advantages. The second

section reviews the economic literature by which the misalignment between the legal

and economic view is presented. As a result, the third section questions the current

approach while the fourth section presents an alternative one in conformity with

economic insights. The last section finally points out some policy implications.

3.1 Introduction: setting the stage

In the previous chapter the legal responses to collusion were presented. Firms

communicating with each other and agreeing to fix prices, say by meeting in the Hilton

Hotel or by exchanging e-mails, will be found guilty without examining the effects of

the conduct, see for instance Polypropylene.46 As such, the legal view on price-fixing is

summarized by stating that such behavior is illegal by object : what matters is the anti-

competitive goal rather than the actual outcomes on the market.

It is believed and generally accepted that such an approach is best suited to deal with

price-fixing cases. Since the pro-competitive effects of it are rare and highly unlikely to

arise while the anti-competitive consequences are well-known, it is then best to deter

collusion per se.

The theoretical possibility of a possible pro-competitive justification of collusion makes

little sense to doubt about the appropriateness of the per se rule. Indeed, if it is rare for

46Cited above.



30

beneficial conspiracies to exist, authorities can better bar them altogether since doing

otherwise would make price fixing cases a lengthy and costly affair. The validation of a

by object approach is thus in fact nothing more than the application of a statistical

decision rule (Whinston, 2003).

In addition, such a per se prohibition has some clear advantages as well. First of all, the

ban on agreements without any assessment of its impact on competition provides for a

clear rule to businesses. As such, firms face a legal environment that reflects

predictability and certainty. Moreover, courts do not face the tough task of analyzing in

detail the economic consequences of the conduct by which resources and time are saved

(Ministry of Economic Development, 2010).

The arguments raised above are reasonable. Nevertheless, so far nothing has been said

whether formal economic theory connects with this legal side of collusion. As such, this

chapter aims at analyzing whether the economic elements of collusion (as presented in

Chapter 1 and here further elaborated and reviewed) justify the legal approach as it is

currently in place (as presented in Chapter 2). Put different: does the economic view

supports the legal view?

In so doing, we will observe a complex economic reality lying behind an apparent

uncontested legal one by which it will be argued the arguments above to be insufficient.

Despite the seemingly uncontroversial status of the by object approach, the analysis

below will make clear that some challenging questions remain unanswered. As a result,

it will be argued that an impact-based prohibition could serve as a better approach in

dealing with collusion.

3.2 Is the current legal approach economically justified? An overview

Undertakings coordinating on prices are found liable by competition authorities based

on the proof of concertation rather than on the economic effects of the conduct. As such,

Article 101 prohibits communication and agreements on prices since they raise prices

and lower welfare.

The statement above departs from the assumption that the communication between

parties (either written or oral) and concurrence on fixing prices automatically leads to



31

the implementation of this conduct on the market. As such, it is believed that talking

acts as a vehicle in order to charge higher prices to consumers. Accordingly, since we

know that such conduct lowers welfare, authorities should ban them.

All of this appears to be straightforward, yet it is not. In particular, it is peculiar to see

how uncontested the above assumption is. Does communication really lead to the actual

implementation of collusion on the market (i.e. higher prices)? And if we believe this,

should we not proof the effects of it on the market?

After all, we know from the economic analysis in Chapter 1 that every collusive

equilibrium is highly unstable. As a result, participants are tempted to cheat on the

agreement by which no higher prices on the market would prevail. Accordingly, when

deviated from the equilibrium, prices on the market would not raise since the conspiracy

is not actually executed on the market.

In accordance, it remains to be seen what economic literature has to say about this. Can

we find theoretical proof whether talking yields higher prices on the market? Or

differently: in what way does the prohibition on talking prevents anti-competitive

pricing on the market and how does it thus raises welfare?

To answer this question, we first take a look at the game-theoretic foundations of the

collusive equilibrium. In so doing, we formally review the function of talking in order

to reach collusion and observe that its role is imprecise. Second, we explore both

experimental and empirical economic literature. Equally, there is few proof that leads us

to believe that communication is the driving force as to implement collusion on the

market.

3.2.1 Collusion and its game-theoretic aspects: formal economic theory

Recall from Section 1.2. above the instability of collusion. Since no enforceable

contract can be signed (as it is illegal!), the collusive equilibrium is a shaky one: each

participant would rather want to cheat by which it captures the whole market, yet it

knows that any cooperation in the future will then be lost since trust is broken.



32

This intuitive trade-off can be formalized by making use of game theory as lined out by

Whinston (2003). Consider the pay-off matrix in Figure 4, formally depicting the

coordination between Firm A and Firm B. Each firm simultaneously set either a low

price (‘Low’) or a high price (‘High’) on the market. Note that at this moment there is

no communication between both firms, since each decides on his price at the same time.

In doing so, they anticipate on the conduct of their competitor by which 4 different

outcomes can arise.

Firm B

High Low

Firm A High (8,8) (-1,10)

Low (10,-1) (1,1)

Figure 4: Pay-off matrix

The pay-offs are marked in each cell, with the left number referring to the pay-off of

Firm A and the right one to Firm B.47 If both firms set a high price, we end up in the

upper-left cell which is the cooperative outcome, i.e. (8,8). If one firm sets a high price

while the other one sets a low price, we end up with the unequal outcomes, i.e. (10,-1)

and (-1,10). Here, the firm being cheated on ends up with the worst pay-off. If both

firms set a low price we arrive at the lower-right cell depicting the non-cooperative

situation where both firms compete, i.e. (1,1).

While the collusive outcome is definitely the best situation for each firm to obtain, it is

not a Nash equilibrium since both players can do better by deviating (10 > 8).

48

Theonly Nash equilibrium here is the outcome where both firms set a low price, on the

figure marked in bald.

47Note that the absolute level of the numbers are chosen randomly.

48

A Nash-equilibrium is any situation where no single player benefits from deviating unilaterally from it.

See Nash (1951) for the seminal paper.



33

Accordingly, both players know that ending up in the collusive outcome would be

better, yet they are ‘stuck’ in an equilibrium which yields lower pay-offs ( (1,1) versus

(8,8)). In literature this is referred to as a prisoners dilemma: the game-theoretic

condition characterized by the Nash equilibrium being Pareto dominated by another

outcome preferred by both participants (Farell and Rabin, 1996)

As Whinston (2003) notes, there is clear coordination problem: the self-interest of each

firm guides them to the collusive equilibrium where they can achieve higher profits.

Yet, not being Nash, they both risk ending up in the non-coordinative outcome.

Accordingly, the question arises how such coordination problem can be solved. In

particular, we investigate the role of communication: if both players are allowed to talk,

can they attain collusion?

Farell and Rabin (1996) studied whether communication can solve the prisoners

dilemma by which collusion can arise.49 Imagine that both firms are now allowed to

talk, unlike the setting above. Think of Firm A approaching Firm B, stating:

‘Let’s make a deal. I would like you to quote a high price on the market. If you

do that, I will also quote a high price. We can than achieve higher profits

instead of competing.’

Being the receiver, would you believe that statement and follow up on it? After all, if

Firm A really has the intention of quoting a high price he would get a pay-off equal to 8.

However, if he intends to quote a low price, he would get a pay-off equal to 10. As

such, a firm intending to collude would make that statement just as a firm who is not

since doing so would leave him with an even bigger pay-off. Accordingly, the

communication on intentions is meaningless for the receiver, i.e. it is cheap talk (Farell

and Rabin, 1996).

In the framework of Farell and Rabin (1996), the statement above is neither self-

signaling (Firm A would prefer Firm B to choose High whatever he does) nor is it self-

committing (Firm A has no incentive to follow up on the promise, whether or not she

49See also Crawford and Sobel (1982).



34

expects Firm B to believe her). As such, it is regarded as communication who is

uninformative: it is communication lacking direct pay off consequences (Whinston,

2003).

Note that industry factors influence the role of communication, as lined out in Chapter

1. If for instance detection of cheating is easy (think of publicly announced prices) or

the number of participants is low (a highly concentrated market) the likelihood that

talking will lead to the implementation of collusion will be higher. As such, industry

factors who facilitate collusion will obviously make it more likely for communication to

establish higher prices on the market.

Hence, is cheap talk able to solve the coordination problem? Put different: can

communication in the setting of a prisoner’s dilemma (as is the case with collusion),

theoretically make firms able to reach the collusive outcome? As revised above, it is

notable to see how little formal economic theory has to say about this. Maybe

experimental and empirical literature can give us some input on whether the ban on

talking has an effect on the prices charged on the market?

3.2.2 Experimental and empirical literature

Some experimental literature has aimed to investigate whether communication about

intentions matters as to achieve collusion, see Holt (1993) and Crawford (1998) for an

overview on this work.

Some papers were able to show that coordination is achieved to a great extent, while

others showed how talking did not matter a lot. As such, experimental literature is up till

now not able to show how cheap talk is exactly able for parties to reach collusion. Nonetheless, some of these papers differed what matters the set-up: in some of them for

instance communication was heavily regulated while in others it was less structured

(Whinston, 2003).

Empirical literature has equally seeked at exploring the effect of the ban on talking on

the price charged by undertakings. Just as above, there is surprisingly little evidence to

be found.



35

Sproul (1993) investigates 25 price-fixing cases in the period from 1973 to 1984. For

each of them, he constructs a ‘but for’ price (the price that would prevail absent

collusion, see later) by extrapolating a regression explaining the price of the good by

closely connected and related prices of other goods taken from the period prior to the

indictment.

Sproul (1993) then analyzes the ratio of the actual price over the ‘but for’ price in the

period subsequent to the indictment. Surprisingly, prices raised by approximately 7

percent after a filing for price fixing instead of an expected drop in prices. In addition, it

is found that a negative relationship exists between the price ratio and the severity of

punishment: in the period after 1976 when tougher penalties were adopted prices rose

less. Hence, there is little to conclude that collusion enforcement yields a reduction in

prices (Whinston, 2003).

Another study conducted by Block et al. (1981) looks at the market for bread. They

analyse prices in the period between 1965 and 1976 and compute a mark-up which they

then regress against measures of antitrust enforcement such as the budget of the

Department Of Justice. They found a negative result, but a small one and a statistically

insignificant one.

A paper carried out on the macro-level is done by Konings et al. (2001). Here, the

authors question whether a change in competition law has any effect on the price-cost

margins of firms. Using a 4-year dataset, they reveal that such an effect is inexistent. On

the other hand, Warzynski (2001) examined whether stricter competition law and policy

resulted in lower price-cost margins and found such a relationship to hold.

3.3 The right approach?

All being said, the above overview leads us to conclude that we can find little economic

evidence on how exactly the ban on talking makes prices to be lower and hence improve

welfare. Stated otherwise: few literature is able to show how communication is able to

fully implement collusion, i.e. to quote higher prices on the market.



36

It is therefore stunning to see how uncontroversial the per se approach on collusion as

currently adopted is. As Whinston (2003) states:

‘It is in some sense paradoxical that the least contested area of antitrust is

perhaps the one in which the basis of the policy in economic theory is weakest.’

Competition authorities remain overall silent about the underlying economic theories

analyzing the role of communication in collusion. This is no surprise, since such an

economic theory does not exist. We thus observe a fairly uncontested legal approach

which is in the same time not at all supported by the economic literature where a huge

loophole exists. This is quite startling, to say the least.

Yet, in the past cases have been made based on the premise that if people are caught

talking, they are seen as to implement collusion without fully acknowledging the micro-

economic aspects that lie behind all of this, see for example Wood Pulp,50 and also

further Section 3.5. Such condemnations can often have disastrous results.

Oxford Economics (2010) for instance investigates the follow-on impact of cartel fines

on investment and employment. It is found that a fine of 150 million € yields an average

loss of 503 jobs in the respective economy. In addition it is revealed that a higher fine

aggravates this effect.

This is quite intuitive since a fine is generally absorbed by a reduction in the amount it

spends on investment. As a result, the amount of jobs will be lowered together with a

decrease in purchases from suppliers. These suppliers will thus equally be hit by the fine

since they receive less revenues. There is thus a chain-reaction in the economy.

Fines accordingly have the ability to ruin companies with disastrous results of the

economy as a whole. Moreover, in the United States price-conspiracies are criminal acts

by which executives caught guilty are send to jail. On the other hand, as we have seen,

there is no clear evidence whether communication leads to higher prices. How can we

then be sure we are doing the right thing?

50OJ L 85/1 [1985].



37

Hence, before doing all this, should we not make use of a market impact study which

makes us able to at least proof that prices indeed rose? To be at least significantly sure

before we ruin companies and sentence people to prison? In accordance, the next

section proposes a different approach from the current one, focusing on the actual

outcomes of the market rather than on intent.

3.4 The alternative approach

The approach as currently adopted is one where the Commission has the burden of

proof to show that an agreement or concerted practice exists. If for example a video tape

is made of the discussions taking place in a meeting participants are immediately in

infringement under the per se rule. Nonetheless, this does not have to mean that the

agreement made will actually be implemented.

Accordingly, a divergent approach would be to give the defendants at least the

opportunity to show that their communication did not result in any harm on the market.

As such, one can allow them to use carefully designed econometrics showing that

communication did not surpassed cheap talk (Van Cayseele, 2010, personal

communication).

If the econometric techniques adopted are inspected thoroughly and the methods used

not disproven, one can abstain from conviction since prices are not shown to increase.

More precisely, one should make use of a counterfactual which indicates a ‘but for’

price; i.e. the price that would prevail absent collusion by controlling for demand and

supply shocks.

As such, the only proof that matters is the simulated but for price as obtained by the

non-cooperative equilibrium using a regression analysis. This counterfactual is then

compared to the actual price in order to calculate the overcharge which is

mathematically the difference between the actual price and the but for price (Stangle,

2009). If no overcharge can be found, no conviction should take place.

Of course, the key input in such a modeling is retrieving an accurate counterfactual.

Literature puts forward different methods of retrieving the but for price, see for example



38

Stangle (2009). One of the most common and precise estimation methods is carrying

out a residual regression analysis. As such, data from the benchmark period is used in

order to predict prices during the collusive period controlling for demand and supply

shocks. Hence:

- # # - $ $ - - -,

where is the regressed but for price, # are the independent variables

influencing the counterfactual and the residual factor, see Finkelstein and

Levenbach (1983); Page (1996) and Fisher (1980) for a technical exposition and Sproul

(1993) for a practical application.

In Figure 5 for example, the area to the right of the dotted line refers to the conspiracy

period (88’-89’). By using the above approach, regression analysis is able to simulate

the ‘predicted’ price (but for price). A positive overcharge was found since the predicted

price was lying below the actual one. Accordingly, the defendants should be found

liable.

Figure 5: But for price by using regression analysis



39

The methodologies described above are not unknown for courts. They are widely used

to calculate private damages arising when private parties sue the defendants for having

paid too much. The reason is that private parties have the right to obtain relief from the

harm caused.51 As such, they have to bring up evidence and carry out economic analysis

which will then be used to calculate the right amount of damages.52

Accordingly, econometric models will come up for sure when large sums of money are

involved see for example Air Cargo53 where companies such as Philips and Ericsson

claimed 500 million euros from KLM-Air France. Hence one can argue why these

resources are not already used in the criminal part of the investigation (Van Cayseele,

2010, personal communication).

One could state such an approach would yield a too high administrative burden

increasing litigation costs significantly. Given that at the same time the pro-competitive

effects are most likely nil and resources are scarce, the argument could make sense.

Nonetheless, it is mistaken since such a point of view disregards the huge personal,

corporate and economic costs that fines generates on the respective economy which

equally have to taken into respect while there is no proof of harm on the market.

Moreover, such an approach can be harmful for an effective antitrust policy since price

fixing cases with harmful consequences can go undetected due to the lack of overt

agreements, while negligible ones not if ‘by accident’ agreements are found (Posner,

1976).

Being said, the above leads us to conclude that the viability of the alternative approach

boils down to comparing: (i) the costs of regulatory error under the current per se

51See the Guidance Paper ‘Quintifying harm in actions for damages based on breaches of article 101 or

102 of the Treaty on the Functioning of the European Union’, public consultation, European Commission,

2011.

52

See http://www.carteldamageclaims.com/damage%20quantification.shtml, 20/06/201053

See IP/10/1487 European Commission.

http://www.carteldamageclaims.com/damage%20quantification.shtml





40

approach (false positive) with; (ii) the costs of a more lengthy treatment under an effect-

based approach as suggested here, assuming that the latter does not contain legal errors.

While (i) can be roughly estimated from the past (see above), it is much more difficult

to gain insight on the total costs that would prevail when using a rule-of-reason

treatment in the future. This intrinsically depends on the development and adaptability

of econometric techniques who on their turn influence the easy by which economic

evidence is brought forward. Yet, in the past few years considerable improvements are

made especially in its application (note that these techniques until recently were

inexistent in private damage claims). One can assume that in the near future they will

somehow increase in importance while in the meantime being able to keep legal costs at

an acceptable level.

The approach suggested above is not an isolated one. It is in line with the generally

shared opinion among economists who do not believe that cases can be made solely on

the basis of intent, see also Whinston (2003) and Werden (2004). Posner (1976) for

instance equally presents an ‘economic approach’ to price fixing cases consisting of two

consecutive steps: (i) indentifying markets in which collusion is likely to show up; (ii)

determining whether collusive pricing is really present on those markets.

The issues raised above were up till now focused on a theoretical loophole. Nevertheless

the next section attempts to place the discourse above in a broader context by reviewing

practical case law. As such, it intends to put light on the practical policy implications of

this thesis.

3.5 Policy implications

Since decades the Commission struggles to accept sound economic analysis in its cases.

In both Polypropylene and Flat Glass54 for example the Court of First Instance critized

the lack of economic proof in the decision made. The Commission itself acknowledges

this caveat and accordingly has taken steps to move away from a fully legalistic

approach by for instance appointing a ‘Chief Competition Economist’ considerably

extending the team of economic experts.

54Both cited above.



41

Nevertheless, the per se doctrine concerning price-fixing cases leaves few space for

economic considerations to be made. If firms are caught concerting they are

immediately in infringement by which any economic analysis of its market impact are

excluded. Despite attempts to modernize the application of Article 101 by moving

more to an effect-based approach (and hence allowing for some economic analysis)

practical case law frequently makes clear that the Commission still sticks to a

formalistic and over-inclusive approach (Jones, 2010).

Several cases demonstrated the Commission’s focus on communication without

indulging in an impact-based analysis surpassing the possibility of cheap-talk cases. In

Polypropylene the Commission found documentary evidence of 15 firms active in the

polypropylene market forming an agreement in order to fix prices. There was proof of

regular contact and meetings held between the undertakings every two months on a

Community-wide basis.

In its investigation the Commission decided that all contact between firms can be seen

as a single agreement with the object of restricting competition. Accordingly, the

participants were found to infringe Article 101(1). Although there was no single

indication of the actual anti-competitive effects of the conspiracy on the market, a large

fine was imposed since the Commission held that there was no need to investigate the

effects of the agreement given that its object was clear.

On appeal, the CFI questioned the lack of economic input but upheld the decision made.

While the defendants raised the argument that the agreement did not caused any harm

on the market, the CFI agreed with the Commission that such an argument is

‘irrelevant’ (Rodger & MacCulloch, 2004).

In Wood Pulp55

undertakings were found liable by sharing information making use of

quarterly advance price announcements. The Commission was of the opinion that the

information exchange was set up with the intention of fixing prices since it observed

prices to change in a parallel manner.

55Cited above.



42

Nonetheless, a team of experts employed by the Court later found out that those

information sharing schemes were actually on the request of consumers who apparently

were asking for a more transparent and stable market. As such, the communication did

not resulted in any collusive prices on the market yet the Commission held the

participants liable. Accordingly the Court smashed the decision and stated that there

was no proof of collusion (Stroux, 2004).56

The Commission’s focus on documentary proof of an agreement is equally illustrated in

LdPE 57 and the Cement

58 case. In Cement the Court made clear that the Commission is

entitled to rely on even one single document in order to decide that Article 101(1) has

been infringed by the defendants (Van Bael & Bellis, 2005). In LdPE , evidence was

found of documents and minutes of (secret) meetings between some producers who

were, based on this, found liable of price-fixing. As was stated by the Commission:

‘It is inherent to the nature of the infringement […] that any decision will to a

large extent have to be based upon circumstantial evidence.’ 59

The decision was eventually annulled by the CFI who even had to pay all legal costs of

the case. (McGowan, 2010).

The cases above clearly indicate the limits of the per se approach. In Polypropylene

there never was no proof of harm on the market, yet personal and corporate costs as a

result of the conviction aroused. In Wood Pulp legal costs were spend on a mistaken

decision while in the same time considerably harming the defendants business

reputations. Likewise, in LdpE no market impact study was done by which a wrong

decision was reached.

56In a similar setting in the Petrol cartel (ruled in Italy) the decision made by the Supreme Administrative

Court was equally annulled based on the fact that it did not consider any other grounds for parallel prices

to arise, see Esso Italiano v. Autorita Garante della Concorrenza e del Mercaton n. 4053/2001.

57OJ L74/21 [1989].

58

Cimenteries CBR and others v. Commission; ECR II-491 [2000]59

As cited in Van Bael & Bellis (2005), page 53.



43

Taking into account the above, it has often been stated that the protection of consumers

is the main goal of the Commission. Neelie Kroes, former Commissioner for

Competition stated:

‘Consumer welfare is now well established as the standard the Commission

applies when assessing mergers and infringements of the Treaty rules on cartels

and monopolies’ .60

However, it is doubtful whether the per se approach on collusion is fully in line with the

statement above. While the welfare reducing object of price fixing is known, this thesis

made equally clear that it is less straightforward to believe that communicating parties

automatically would implement collusion on the market. As such, it is doubtful how the

by object approach is connected with consumer welfare.

Meanwhile, economic tools applicable to competition law keep on gaining in

importance. In contrast to the past, it is nowadays possible to make use of sophisticated

econometric techniques such as adequate regression analysis to calculate precisely the

impact of a certain conduct on the market. As explained above, estimating a but for

price can serve this role concerning price fixing cases. In times to come, one can expect

the role of economic instruments to increase and be even more reliable.

If the Commission wants to keep up with these developments, a rigid per se approach

will no longer be sustainable. As such, we argue for a more flexible impact based legal

treatment which is (i) able to make use of reliable economic tools adapted to the case at

hand; (ii) accordingly is in line with the goal of protecting consumer welfare; (iii) and is

coherent with economic literature on collusion.

60Neelie Kroes (2005).



44

3.6 Conclusion

This chapter analyzed how the economic view on collusion relates with the legal view

and has found both to be inconsistent with each other. Economic theory formalizes the

role of communication in collusion by making use of game theory. By doing so it isrevealed an ambiguous role of talking in order to reach the collusive outcome. Likewise,

empirical and experimental literature leave us with few proof on the consequences of

communication.

Consequently the current approach was questioned since an inconsistency was revealed

between the legal approach and economic theory. As a result, this chapter advocated for

a rule of reason approach where the Commission makes use of correct economic tools

able to demonstrate the real effects on the market.



45

General conclusion

This thesis has aimed to discuss the correct legal approach in order to deter collusion,

the anti-competitive conduct characterized by firms conspiring on prices in order to

raise mutual profits to the detriment of consumer welfare. The work was structured in

three chapters. The first two focused on respectively the economic and the legal view

while the third chapter intertwined these insights in order to reach a conclusion.

Chapter 1 illustrated the economic functioning of collusion and consisted of three

consecutive parts. The first part demonstrated the profit-increasing effect of price fixing

on the conspiring companies. It was shown that price coordination makes firms to act as

a monopolist by which higher mark-ups arise. The second part focused on the instabilityof the collusive equilibrium. Since each participant can do better by cutting prices below

the collusive one the outcome tends to be unstable. The last part explained how such a

practice lowers consumer welfare, giving rise to the necessity of legal interventions.

Chapter 2 focused on the legal side of collusion. It revised in depth how price fixing is

in violation with Article 101(1) of the Treaty. However, it was equally apparent that the

article used vague concepts which needed further clarification. As such, the first part

lined out the scope of ‘agreements’ and ‘concerted practices’ and concluded that both

require a very broad interpretation. In so doing, the Commission is given space to

intervene in cases dealing with tacit collusion, though there is lack of agreement

whether such intervention is justified. The final part raised attention to the legal

treatment of cases. It was shown how the Commission makes use of a ‘by object’

approach, hence proving conspiracy instead of actual effects on the market.

The third chapter made use of the insights gained in the prior ones. It builded on the

knowledge gained in the first chapter by extending the economic formalization of

collusion. Consequently it was seen how economic literature has surprisingly little to

say about the role of communication in order to implement collusion on the market. It

then coupled these insights to the information presented in the second chapter. More

precisely, it focused on the role of communication in order to implement collusion by

which it was revealed that talking does not necessarily lead to higher prices on the



46

market. In accordance, an alternative impact based approach was suggested which is

line with the aforementioned conclusion.

It goes beyond saying that this work has to be read in its particular context and

additional research could be done. This thesis does not want to claim that the current

per se approach is entirely wrong. As raised above, there are proper arguments who

support such a view. Nevertheless, it reveals that economic counterarguments exist as

well, namely that economic literature is not able to support the current approach. As

such, this thesis is a theoretical representation of the seemingly ignored disjointed

economic and legal view on collusion.

Hence, a more practical implication of this work could give us better insights of how

legislation precisely affects consumer welfare. Equally, an estimation of the additional

costs that would show up when making use of an effect-based approach could make us

able to compare more precisely the tradeoff with the costs arising from false positive

decisions under a per se approach. Finally, more support from past practical case law on

the arguments raised would be more than welcome.



47

List of Figures

Figure 1: Tacit collusion versus cartels………………………………………………..7

Figure 2: Price fixing…………………………………………………………………..8

Figure 3: Welfare effects of collusion………………………………………………...12

Figure 4: But for price by using regression analsysi………………………………..…37

.



48

References

Aubert C., Rey P., Kovacevic W. (2003). The impact of Leniency programs on Cartels.

Unpublished manuscript.

Block K.B., Nold F.C. and Sidak J.G. (1981). The Deterrent effect of Antitrust

Enforcement . Journal of Political Economy. Vol 89. Pages 429-445.

Connor J.M. (2001). Global Price Fixing: Our Customers are the Enemy. Kluwer

Academic Publishers

Connor J.M. (2008). Global Price Fixing. Springer.

Crawford V. (1998). A Survey of Experiments on Communication via Cheap Talk.

Journal of Economic Theory. Pages 286-298.

Crawford V. and Sobel J. (1982). Strategic Information Transmission. Econometrica.

Vol 50. Pages 1431-1451.

De Bondt R. (2005). Elementen van bedrijfseconomie. Alta Leuven.

Farrell J. and Rabin M. (1996). Cheap Talk . Journal of Economic Perspectives. Vol 10.

Pages 103-118.

Feuerstein S. (2005). Collusion in Industrial Economics: A Survey. Journal of Industry,

Competition and Trade. Vol 5(3). Pages 163-198.

Finkelstein M.O. and Levenbach H. (1983). Regression Estimates of Damages in Price-

fixing Cases. Law and Contemporary Problems. Vol 46. Pages 145-169.

Fisher F.M. (1980). Multiple Regression in Legal Proceedings. Columbia Law Review.

Vol 80. Pages 702-736.



49

Griffin J.M. (1989). Previous Cartel Experience: Any Lessons for OPEC? In Klein L.R.

and Marquez J. Economics in Theory and practice: An Eclectic approach. Boston:

Kluwer Academic

Grout P.A. and Sonderegger S. (2005). Predicting Cartels. London: Office of Fair

Trading

Gwartney J.D., Stroup R.L., Sobel R.S. and MacPherson D. (2009). Economics: Private

and Public Choice. Cengage Learning

Holt C. (1993). Industrial Organization: A Survey of Laboratory Research. In: Kagel J.

and Roth A. Handbook of Experimental Economics. Princeton University Press.

Ivaldi M., Jullien B., Rey P., Seabright P. and Tirole J. (2003). The Economics of Tacit

Collusion. Final Report for DG Competion, European Commission, March 2003

Jones A. and Suffrin B. (2008). EC Competion Law: Text, Cases & Materials. Oxford

University Press.

Konings J., Van Cayseele P. and Warzynski F. (2001). The Dynamics of Industrial

Markups in Two Small Open Economies: Does National Competition Policy Matter?

International Journal of Industrial Organization. Vol 19. Pages 841-859.

Korah V. (2007). Competition Law of Britain and the Common Market. Kluwer

Academic.

Kovacic W. (1993). The Identification and proof of Horizontal Agreements Under the

Antitrust Laws. Antitrust Bulletin. Vol 38. Pages 5-82.

Kühn K-U. (2001). Fighting Collusion by Regulating Communication between Firms.

Economic Policy CEPR, CES, MSH. Vol 16(32). Pages 167-204, 04.

Martin S. (2001). Industrial Organization: A European Perspective. Oxford University

Press.



50

McGowan L. (2010). The Antitrust Revolution in Europe: Exploring the European

Commission’s Cartel Policy. Cheltenham: Edward Elgar Publishing.

Ministry of Economic Development. (2010). Cartel Criminalization: Discussion

Document. Ministry of Economic Development New Zealand. Available online at

http://www.med.govt.nz/upload/70683/Cartel-Criminalisation.pdf .

Monti M. (2001). Why should we be concerned with Cartels and Collusive Behavior? In

Fighting Cartels: Why and How. 14,14. Swedish Competition Authority, 2001

Motta M. (2004). Competition Policy: theory and practice. Cambridge University Press.

Motta M. and Polo M. (2003). Leniency Programs and Cartel Prosecution.

International Journal of Industrial Organization. Vol 21(3). Pages 347-379.

Nash. J. (1951). Non-Cooperative Games. The Annals of Mathematics. Vol 54(2).

Pages 286-295.

Oxford Economics. (2010). An analysis of the Follow-on effects of Cartel Fines on

Investment and Employment. Available online at

http://www.oef.com/samples/cartels.pdf .

Page W.H. (1996). Proving Antitrust Damages. American Bar Association Chicago.

Peterson E.B. and Connor J.M. (1996). Consumer Welfare Loss Estimates in

Differentiated Food Product Markets. Review of Agricultural Economics. Vol 18.

Pages 233-246.

Posner R. (1976). Antitrust Law: An Economic Perspective. Chicago: The University of

Chicago Press.

Posner R.A. (2001). Antitrust Law. Chicago: University of Chicago Press

http://www.med.govt.nz/upload/70683/Cartel-Criminalisation.pdf

http://www.oef.com/samples/cartels.pdf

http://www.oef.com/samples/cartels.pdf

http://www.med.govt.nz/upload/70683/Cartel-Criminalisation.pdf



51

Ritter L. and Braun W.D. (2005). European competition law: a practitioner’s guide.

Kluwer Law International.

Rodger B.J. and MucCulloch A. (2004). Competition Law and Policy in the EC and

UK. Cavendish Publishing

Sproul M.F. (2003). Antitrust and Prices. Journal of Political Economy. Vol 101(4).

Pages 741-754.

Stangle B.E. (2009). Calculating Cartel Overcharges. Analysis Group: Economic,

Financial and Strategy Consultants.

Stroux S. (2003). US and EC Oligopoly Control. The Hague: Kluwer Law International

Tirole J. (1988). The Theory of Industrial Organization. MIT Press.

Turner D.F. (1962). The definition of agreement under the Sherman Act: Conscious

Parallelism and Refusals to Deal. Harvard Law Review. Vol 75. Pages 655-706.

Van Bael I. and Bellis (2005). Competition law of the European Community. Kluwer

Law International.

Warzynski F. (2001). Did Tough Antitrust Policy Lead to Lower Markups in the US

Manufacturing Industry? Economic Letters. Vol 70. Pages 139-144.

Werden G.J. (2003). The Effect of Antitrust Policy on Consumer Welfare: What

Crandall and Winston Overlook, EAG 03-2. Washington DC Economic Analysis Group

Antitrust Division US, Department of Justice

Werden G.J. (2004). Economic Evidence on the Existence of Collusion: Reconciling

Antitrust Law with Oligopoly Theory. Antitrust Law Journal. Vol 71. Pages 719-800.

Whinston M.D. (2003). Lectures on Antitrust Economics. MIT Press.



52

Bart Mayens Thesis EMLE

Documents