- Market power — Economics
2.01 The current sociology of EU competition law Like other branches of corporate law, EU competition law is no longer the sole province of lawyers.1 Instead, lawyers collaborate on a daily basis with economic consultants. Authorities and courts are not to be outdone either. In 2003, the Commission announced the creation, within DG Competition (DG COMP), of the position of chief economist, responsible for heading a unit of economists who are to provide analytical support to the legal team and to review draft decisions for their social welfare implications.2 A year later in France, an eminent professor of industrial economics acceded to the highest judicial office, with Frédéric Jenny’s appointment a member of the French Cour de Cassation.3
(p. 60) 2.02 The origins of the economics of competition law The involvement of economists in the competition law process is a sufficiently important phenomenon for us to take a few moments to consider it.4 To appreciate its scope, we must go back in time and beyond the borders of the EU. In the United States, from the 1950s on, economists and law professors, versed in classical and neoclassical price theories, became interested in antitrust policy as a practical application of their theories. Their studies relied on complex economic arguments and quickly gave rise to a new discipline: competition economics.5
2.03 The resistance of EU competition law In the EU, the influence of competition economics (and incentives and effects-based approach) was marginal for a long time.6 The Treaty’s competition rules were designed by law professors7 and were formally enforced by DG COMP.8 Cooperation agreements between undertakings were deemed incompatible regardless of the market power of the undertakings for the businesses involved: it was thought that by committing to an agreement, undertakings illegally limited their freedom of action on the market and hence the motivations to do so could not be benign. The standard for assessing mergers was therefore whether they ‘create or strengthen a dominant position’, regardless of any efficiency gains that might derive as a result.9
2.04 The breakthrough of competition economics into EU law It was not until the end of the 1990s that competition economics started exerting some influence on the Commission and the EU Courts.10 The decisions of the Commission and the judgments of the EU Courts, particularly as regards vertical agreements, were then harshly criticized as being poorly in line with economic theory.11 They were contrasted with the decisions of the US courts and authorities, which were considered to be more sensitive to the teachings of economic theory.
2.05 Breaking with the formalistic approach which had so far prevailed, the Commission engaged in reforming the assessment of vertical agreements to be more in line with economic analysis.12 The new approach focused on the ‘market power’ of the parties to an agreement.13
(p. 61) 2.06 The legal revolution But the assessment of vertical agreements was not the only area of EU competition law to be reformed by the Commission. It also sought to modernize its approach in the areas of horizontal cooperation agreements, technology transfer agreements, control of concentrations, abuse of a dominant position, etc.14 In the end, no area of competition law was spared from what is now referred to as the ‘economic approach’ or the ‘effects-based approach’, albeit to varying degrees.15
2.07 The scholarly debate Despite the many efforts led by the Commission, the economic approach received a cold reception from a number of scholars who considered that economics, traditionally the ‘servant’ of competition law, had now become its ‘master’.16 In fact, the conflict between these two interpretations does not necessarily have to be resolved. What is new with competition economics is that the rule of law has more recourse to economic tools in its customary implementation. On the other hand, the economic purposes pursued by competition law have existed since the adoption of the Treaty of Rome, with the qualifications already mentioned in Chapter 1.
2.08 Scholars also recognized that competition law economics poses problems of a practical nature as well. While its aim is laudable (to serve as a tool to help in decision making, whether individual or regulatory, and to help in evaluating the law), it should be borne in mind that effects-based analysis when applied improperly (eg in the absence of a unified economic theory) can provide misleading results. Moreover, even when applied well, it is rarely the case that sufficient data are available to reach a foolproof conclusion. The inevitable ambiguities must, by necessity, be resolved by a judgment regarding the perceived risks and/or severity of the potential harm. Such judgments naturally imply some degree of legal insecurity for those subject to the legal system.17
2.09 The remainder of the chapter Today, competition economics is generally described as a composite discipline borrowing from different economic schools of thought. In what follows, we describe the leading schools of thought in regards to competition economics as they (p. 62) have evolved over the years (Section II).18 Then we get to the methodological aspects of competition economics or, more concretely, the instruments and concepts on which competition economics rely (Section III).
A. Classical and Neoclassical Competition Economics
2.10 Introduction Classical and neoclassical economists were the first to focus on competition issues. The classical economists of the seventeenth century and before, particularly Adam Smith, saw competition as a behavioural process .19 The common view was thus that competition is an individual behaviour by which each actor (firm or person) will play to their own advantage so as to defeat rivals. Thanks to the ‘invisible hand’ of the market, the sum of these selfish individual behaviours naturally results in a goods price that is close to the costs of production.
2.11 With the neoclassical economists of the late nineteenth century and early twentieth century came a structural interpretation of competition,20 which is still a feature of the economics of law today.21 Neoclassical theorists constructed two theoretical models of market structures which are meant to constitute the alpha and omega of social welfare. First, ‘perfect competition’, that is, a market made up of a multitude of buyers and sellers, optimizes the operation of the law of supply and demand and consequently provides a positive contribution to social welfare (see Section B). In other words, the more actors that are present in a market, the more likely it is that the invisible hand will function to the benefit of society. Second, monopoly markets (or cartelized markets) disrupt the invisible hand (the operation of the law of supply and demand) and by doing so greatly impair social welfare (Section C). These neoclassical conclusions, however, must be assessed with some care (Section D).
2.12 A brief recap of the law of supply and demand As is well known, if ‘demand’ for a product increases, the price increases;22 if ‘supply’ of a product increases, the price decreases.23 Hence, when coal is in high demand its relative scarcity of supply makes each unit more valuable (purchasers compete to buy it) and its price rises logically. This may have the consequence that the market price will rise above the reservation price of some buyers who will thus no longer be able to acquire the product. Analogously, when the harvest of certain agricultural products is plentiful (so that scarcity is reduced) their prices collapse. Only certain sellers, those who offer the lowest prices (notably the most efficient ones), will survive to sell their products on the market.
2.13 This intuition was formalized by Alfred Marshall, the father of neoclassical economics, in a graphic representation showing the prices on the Y-axis and the quantities on the X-axis (see Fig 2.1).24 As regards demand, the quantities sought decrease as the price of the good increases. The downward sloping curve of demand (D) illustrates this phenomenon. The quantities placed on the market, on the other hand, rise as the price increases. The upward sloping supply curve (S) illustrates this phenomenon.
2.14 Figure 2.1 illustrates the difference between the incentives of buyers and sellers.25 When the price is low (in relation to ‘equilibrium’, eg P L), only a few sellers are willing to place quantities on the market, but not enough to meet demand. In such a case, the result is what is termed excess demand (or insufficient supply). As a result, allocation of resources is not optimal because customers who would benefit from purchasing the good are not supplied
(p. 64) and must either to turn to other products or do without altogether. When the price is high (in relation to ‘equilibrium’, eg P H), many sellers are ready to place quantities on the market, but not enough consumers are willing to pay the price, hence not all the quantities produced are bought. In such a case, there is what is known as excess supply (or insufficient demand). In this case, too, allocation of resources is not optimal since the sellers have expended resources to produce goods that will not be purchased, resources that would have had a more productive use elsewhere. Marshall shows that there is an equilibrium price that exactly matches supply and demand.26 Its level is situated where the supply and demand curves intersect, resulting in P * and Q * in the graph.27
(2) The theory of ‘perfect competition’
2.15 The virtues of supply and demand The first neoclassical competition theorists were convinced that when competition is ‘perfect’, it contributes positively to social welfare as it causes the equilibrium price and quantity pair as presented by Marshall to emerge.
2.16 Conditions of perfect competition The perfect competition equilibrium, however, is not automatic. Five criteria must be satisfied for perfect competition to emerge (referred to as ‘necessary conditions’ by economists).28 First, the market must be characterized by a large number of sellers and buyers, no single one of which can influence prices.29 Second, the products must be homogeneous. In the eyes of the buyers, the products of each of the sellers must be viewed as strictly identical to one another. Third, the information of the buyers and sellers must be perfect .The purchase and sales terms on the market (especially the price) must be immediately known by everyone. Fourth, the entry (and exit) of new producers on the market must be free. Any economic agent may start up new productive operations without incurring any economic or regulatory obstacles (typically referred to as ‘barriers to entry’). Finally, the acquisition of a product must not result in any transportation cost, which would restrict the free movement of goods. Figure 2.2 summarizes these five necessary criteria.
2.17 Virtuous effects of perfect competition According to classical theorists, perfectly competitive markets provide for a number of virtuous effects. First of all, a single price is created in the market. No one seller is capable of fixing a price that differs from his competitors. As soon as one of the suppliers raises his price, he immediately loses all his customers to his competitors and is forced to exit the market. As soon as one supplier reduces his price, he finds that his profits decline relative to his competitors.
2.18 The second virtuous effect of perfect competition described by Marshall also relates to the equilibrium price. The price paid by the consumers ‘is driven down’ to its lowest sustainable level: it is simply equal to the marginal cost of production of all the producers. The explanation is relatively simple. Since none of the sellers can individually fix the price, each must ‘accept’ the price imposed by the market.30 That said, the only level guaranteeing that each(p. 65)
producer can stay in the market is the marginal cost of production: if price is less than the marginal production cost, the producer is no longer profitable and has to exit the market;31 if it is greater than the marginal production cost, the producer loses his customers to his rivals, who are selling at marginal production cost.
2.19 Conclusions When competition is perfect, the law of supply and demand works well. Perfect competition leads not only to an optimal allocation of resources through the equilibrium price, but also to the elimination of supra-competitive profits.32
(3) The monopoly and the cartel
2.20 Imperfections in the law of supply and demand A monopoly exists when a market is sup-plied by one single undertaking.33 Whether this results from the commercial success (stemming from, eg, particular ‘business acumen’) of a firm, the existence of natural barriers (eg the tremendous expense of building a tunnel linking France and Great Britain) or a regulatory decision (eg the restricted certification process for the provision of legal services), a monopoly disturbs the welfare benefits of the law of supply and demand.
2.21 Unlike perfect competition where no seller can influence prices (sellers are called price-takers), the monopolist is a price maker. Since it controls all the production capacities, the monopolist enjoys a ‘scarcity rent’; it has the power to reduce the quantities offered on the market. With the supply of the product restricted to a level below that demanded, prices naturally rise to a level greater than the costs of the monopolist.34 As long as the gains made (p. 66) from raising the price are greater than the losses resulting from reduced quantities sold (and hence not meeting some of the existing demand), the monopolist’s incentives are clear: prices are higher under monopoly than under competition.
2.22 Analogy with the cartel A cartel, which is nothing other than a monopoly comprised of several undertakings working in concert organized by means of an agreement to restrict production and keep prices high, produces analogous effects.35 A cartel, however, is an industrial organization which is less stable than the monopoly, requiring as it does trust and cooperation among the participants.
2.23 The three principle inefficiencies of a monopoly (or a cartel) Neoclassical theorists consider that a monopoly generates at least three inefficiencies. The first of these is its allocative inefficiency, of which we have already had a brief glimpse in Chapter 1. Certain customers who value the product at a level above the cost of producing it and are willing to compensate the monopolist for those costs are nonetheless not supplied because the monopolist reduces the quantities offered on the market in relation to the competitive level (see QM in Fig 2.3 and the arrow from QC, representing the competitive quantity supplied). This quantity reduction leads to a ‘suboptimal’ allocation of resources, as defined by Vilfredo Pareto, because some amount of surplus is left unclaimed.36 The monopolist could improve the situation of the customers excluded from buying the product without incurring a loss (his costs would be covered), but he would sacrifice his own profits to do so. Customers not supplied by the monopolist must turn to other products which they value less or simply do without. The loss of welfare (in value) caused by the monopolist is represented in Figure 2.3 by the triangle ABC and referred to by economists as the ‘deadweight loss’ of monopoly.37
2.24 The second inefficiency is the monopoly’s productive inefficiency. As John Hicks, Nobel prize winner for economics, explained, ‘the best of all monopoly profits is a quiet life’.38 Exempt from any competitive discipline, the monopoly undertaking is free to let its production costs drift upwards. Its shareholders do not take action against the monopoly’s poor cost control performance because they are not able to compare the monopolist with any competing undertakings.39 It is no surprise, then, to find among the great monopolists massive undertakings with breathtakingly high cost curves. The landline telecommunications companies, which were originally protected by natural barriers to entry in the form of substantial network building and later protected by government regulations, are a classic example of this phenomenon. Of course, it is important to understand that the natural barriers to entry in(p. 67)
telephone service were the reason that the telephone monopolies emerged in the first place. This point reinforces the difficulty involved in disentangling any productive inefficiencies from the inherently high costs frequently associated with monopolies.
2.25 The third inefficiency of a monopoly is dynamic. In the same vein as discussed, some have argued that monopolists have few, if any, incentives to innovate.40 The idea here, as expressed most prominently by Kenneth Arrow, is that the monopolist recognizes that any innovations it introduces will only cannibalize its current sales, and moreover that it will do so at the cost of the investment required to reach the innovation. It makes little sense to expend resources when few net gains in profit are anticipated to result. Hence, economists supporting dynamic inefficiencies arguments extended the notion of the ‘quiet life’ of the monopolist to research and development (R&D) and innovation as well.
2.26 This third dynamic inefficiency has been, and still is, extremely controversial. One of the more prominent attackers of the theory was the famous Austrian economist Joseph Schumpeter.41 In brief, Schumpeter argued that innovation is primarily done by monopolists. The connection between market power and innovation resulted, Schumpeter claimed, from the more powerful undertaking’s ability to finance and develop innovations and to bring them to market. Smaller, weaker undertakings, on the other hand, lacked the resources to conduct research or to develop its fruits. Moreover, under Schumpeter’s view, monopolists are condemned to innovate in order to maintain the uniqueness of their market position, not because they face rivals today but because rivals can emerge from unforeseen quarters at any moment, sweeping in to replace the incumbent. This view led to Schumpeter’s best known quote regarding the ‘perennial gale of creative destruction’.
(p. 68) 2.27 The debate over innovation—what defines it, who is best at creating it, what factors spur its creation, and so on—is one that continues to rage today.42 Even casual observers of high technology industries must recognize the tremendous amount of innovation that has emerged from small, start-up undertakings in the Silicon Valley of California.43 Thus far, the empirical evidence remains mixed with some support for both Schumpeter’s view that monopolists can be highly innovative as well as contrary views that smaller, nimbler, hungrier undertakings produce the most and best innovation. The truth is probably far too complicated to be explained by one variable alone.
2.28 Other inefficiencies Other forms of inefficiencies are also associated with monopoly, some purely formal. For instance, the term ‘X inefficiency’ is sometimes used in the competition economics literature. This expression, coined by Harvey Lebenstein, in fact covers inefficiencies that are not allocative, but rather are productive and dynamic.44 The competition economics literature also refers to the notion of technical inefficiency, which is a form of productive inefficiency (linked to the idea that the monopolist chooses suboptimal production techniques). Along the same lines, another term often used is internal or organizational inefficiency, which tends to indicate that the internal structure of the monopolist is not optimal (duplications in certain units, absence of control mechanisms, etc). The concept of ‘managerial inefficiency’, a more recent concept, is a variant of organizational inefficiency. Shielded from any competitive pressure and exposed to superficial control by shareholders (who, as noted, are incapable of comparing the performance of their undertaking with other operators on the market), executives running monopolies are not purely concerned with profit. Provided they meet the minimum profit requirement they are free to make decisions motivated instead by individual interest,45 such as in-kind bonuses (luxury offices, business cars, etc), initiatives aimed at boosting their personal stature (creation of foundations, charity work, etc), nepotism, and so forth.
2.29 Lastly, a newly emerging concept is that monopolies exhibit distributive inefficiency.46 This concept relates to the distribution of benefits across undertakings and individuals. As explained, monopolists charge prices in excess of their marginal costs. While the customers supplied at this monopoly price only pay the monopolist if the price set is less than their reservation price for the good, the fact remains that they are paying more than they would have under competitive conditions. As a result, monopolies transfer resources from consumers to themselves (see red zone in Fig 2.4). This distribution of resources to the company from consumers is said to be a new form of inefficiency. Hence, it is suggested that competition law should focus on controlling prices on concentrated markets. The question of distributive efficiency, which has been a bone of contention for economist for decades, is a political one. After all, the optimal distribution of resources among consumers and undertakings, the latter of which employ individuals (who are also consumers) and are often held by individual shareholders (who are consumers as well) is a subjective matter. Profits earned by a monopoly can be redistributed to shareholders (by dividends), to employees (by profit-sharing(p. 69)
mechanisms), and to society in general (by tax mechanisms), or even kept within the undertaking but reinvested in new R&D or other productive endeavours that contribute to social welfare.
(4) Evaluation of classical and neoclassical competition theories
2.31 Perhaps one of the most important aspects of the neoclassical models is that they illustrate quite clearly the effects of competition (and its antithesis, monopoly) on social welfare. For the reasons outlined hereafter, however, these models fail to provide a precise guide for intervention by the competition authorities.
2.32 Failure to adapt to contemporary economic realities With the exception of a few economic sectors, classical and neoclassical theories focus on market structures which largely disappeared in the twenty-first century. Nowhere do the features characterizing perfect competition seem to exist.47 Moreover, it is difficult to identify modern-day monopolies—except perhaps statutory monopolies—that are so stable that no firm is likely to enter the market on which they operate. Even Microsoft, which is often considered as a stable monopoly, is regularly challenged by new operators (Linux, Apple, and most recently Google) and hence only imperfectly fits the description of neoclassical theory. That said, neoclassical theorists did offer a glimpse of the more complicated market structures of today: Augustin Cournot’s theory focused on the oligopoly, a market structure where only a handful of sellers operate and thus represents a midpoint between perfect competition and monopoly. (Oligopolies are markets where there are only a few sellers.48) In the model known as Cournot competition, oligopolists fix their production at a level that leads to a price greater than perfect (p. 70) competition but less than the monopoly price. Cournot’s duopoly is then characterized by a ‘mild’ competition where each of the operators makes economic profits.49
2.33 Another problem comes from the fact that, because the classical and neoclassical theories are so abstract, they produce inexact and unrealistic results. As Friedrich Hayek notes, perfect competition paradoxically results in ‘the absence of any competitive activity’ since the price is constantly unified.50 The assessment made against the monopoly should also be corrected since, as we shall see, it acts in certain circumstances in a way that is allocatively efficient (through price discrimination), productively efficient (by realizing economies of scale), and dynamically efficient (by financing innovation).
2.34 Useful, albeit limited While it is true that the lack of realism of the models described have made practical enforcement based upon them impossible,51 neoclassical models nevertheless offer useful benchmarks for determining the direction of effects, even if they cannot be relied upon to determine the magnitudes.52
B. The Normative Economics of Competition
2.35 Introduction Immediately after the Second World War, competition economics became more normative. Based on empirical economic analyses, scholars from Harvard University formulated ambitious policy recommendations as a guide for competition authorities (discussed in Section 1). A few years later, members of the University of Chicago proposed a different interpretation of the economic theory of the legal implications, an approach that was diametrically opposite from Harvard’s (see Section 2). The teachings of these two schools were both criticized by a new modern movement of thought which appeared in the 1980s, often referred to as Post-Chicago School (see Section 3). The debate between the Chicago and Post-Chicago viewpoints is one that is still carried on today.
2.36 A new methodology In the 1960s, a group of economists from Harvard led by Professor Joe Bain started reflecting, as the neoclassical theorists had done, on the relationship between the structure of a market and its performance. But any similarity between the two approaches stops there. Rejecting the arguments of neoclassical theory, which were felt to be too abstract, the Harvard economists used an empirical method based on making factual observations of branches of industry.53
2.37 The ‘SCP’ paradigm Their work highlighted a causal relationship between the structure of a market, that is to say, its endogenous characteristics (the number of producers and buyers, barriers to entry, degree of product differentiation, etc), the conduct of the undertakings (price setting, investments, publicity policies, etc), and the performance of the industry, that is, its contribution to ‘welfare’, which is understood not only as economic efficiency but, more generally, as social progress.54 Professor Bain argued that the more concentrated the structure of a market is, the more suboptimal the behaviours of the undertakings on that market are (supra-competitive price policies) and the more negative the performance of the industry is (the supra-competitive profits of the undertakings are to the detriment of consumers).
2.38 Like any equation, the Structure-Conduct-Performance paradigm contains an unknown factor: the explanation for the relationship of cause and effect between the size of a firm and the profits it makes. Professors Carl Kaysen and Donald Turner suggested one answer, an elementary one: the existence of high profits is proof that large firms have ‘unreasonable market power’—the power to raise price above the costs—which they exploit to the detriment of consumers.55 Of course, this is just one possible answer. High profits may also signify high (and risky) upfront investments, such as often required with research-driven products and services; high profits ex post are required to recoup risky outlays made ex ante.
2.39 Structuralist recommendations The originality of the Harvard School is illustrated in the fact that its authors moved from economic theory to formulate policy recommendations. They felt that government must closely monitor the structures of concentrated markets.56 However, as regards the legal remedies to be used, the members of the Harvard School differ on which approach is best. Some, like Professor James Rahl, argue in favour of a strict application of Section 2 of the Sherman Act, which prohibits ‘monopolization’.57 Others, like Kaysen and Turner, recommended that a special law be passed which would allow (p. 72) concentrated markets to be deconcentrated at any time.58 On the other hand, the ex ante control of concentrations between undertakings was not endorsed by the Harvard School, which thought such a policy would be too limited and incapable of correcting concentration that emerged through organic growth.
2.40 From theory to practice The radical proposals put forward by Bain, Kaysen, and Turner found political favour at the end of the 1960s. The White House Task Force on Antitrust Policy produced a report known as the Neal Report59 that recommended the adoption of a special law entitled the Concentrated Industries Act which would have allowed the implementation of a policy to deconcentrate concentrated markets.60 The Attorney General would have been responsible for identifying all concentrated US industries and then for taking legal proceedings against firms holding more than 15 per cent of the identified market.61 The set objective was to use structural remedies to reduce the undertakings’ market share to below 12 per cent.62 In 1972, Senator Philip Hart proposed the adoption of a similar law evocatively entitled the Industrial Reorganization Act.63 Neither of these proposals was ever made into US law.
2.41 In fact, enthusiasm for the Harvard proposals quickly faded. The legislative proposals met with a lot of harsh criticism64 and the drafts got buried when reviewed by the parliamentary committee. Reviewers, for instance, pointed to the fact that the relationship between market concentration and firm profitability was weak and that such important considerations as examining the potential for barriers to entry had been ignored. Furthermore, there was little to suggest that the analyses had confirmed whether deconcentration trends held in the long term or if they merely exhibited short-term fluctuations.
(2) The Chicago School (or the ‘behaviouralist’ movement)
2.42 Introduction Coincident with the demise of the Harvard School of Thought, the years 1960–70 saw the birth of a new movement at the University of Chicago.65 The view there (p. 73) broke with the stance that concentrated market structures had to be dissolved. The Chicago School differed from the Harvard structuralists both methodologically and ideologically.66 First, methodologically, it rehabilitated the theoretical analysis of the markets (by invoking neoclassical competition theory).67 Next, ideologically, it assigned to competition (and to the policy underlying it) the single objective of promoting economic efficiency, in its three forms (allocative, productive, and dynamic).68
2.43 Another interpretation of the ‘SCP’ paradigm The difference in approach led members of the Chicago School to propose a different interpretation of the SCP paradigm which looks more favourably on industrial concentration.69 In their view, the condemnation of concentration, without other anticompetitive conduct, overlooks the efficiencies that can come with concentration in the form of economies of scale, cost reductions, and product improvements.70 When such efficiencies are present in an industry, undertakings may logically be encouraged to grow in size, thus accentuating the degree of market concentration. Large firms, which tend to be more efficient than smaller ones, thus sell products that are less costly and of better quality.71 And since their costs are lower, their profits naturally surpass those of small undertakings.
2.44 In general, the difference between the Harvard and the Chicago Schools lies in their interpretation of the SCP paradigm. For the Chicago School, the existence of profits in concentrated industries does not automatically come from the exercise of ‘unreasonable market power’ but can be explained by the greater efficiency of large undertakings.72
2.45 Minimalist legal recommendations Codified in work by Robert Bork evocatively entitled The Antitrust Paradox,73 the advice given by the Chicago School to government is the very (p. 74) opposite of the structuralist directives of the Harvard School. Competition authorities should not be concerned with market structure. Most of the time, the oligopolistic concentration of the markets is a rich source of efficiencies. The Chicago School challenged the Harvard proposals for administering structural measures on concentrated markets.74
2.46 Instead, the Chicago School argues that authorities should limit their intervention to the detection of collusive behaviours (which show no efficiency) and, where necessary, to punish them when they appear.75 But the need for competition authorities to intervene to prevent collusive behaviour may in fact be limited. According to George Stigler, horizontal collusion is rare in practice since it is intrinsically unstable—the participants nearly always have incentives to deviate from the agreement.76 As to vertical collusion (eg retail price-fixing agreements between suppliers and distributors), it can often be explained by large efficiency gains.77 In truth, according to the Chicago School, collusion should only be a concern for concentrated markets. That is, governments need only look into collusive behaviours when markets are oligopolistic. Thus both the Harvard and Chicago Schools were concerned with concentrated markets, but for the Harvard School concentration was enough on its own while the Chicago School focused on anticompetitive behaviours within concentrated markets.
2.47 In the final analysis, the Chicago School therefore advocates a ‘minimalist’ competition policy. In line with these precepts, the Reagan Administration of the early 1980s limited antitrust intervention to its very minimum (‘small antitrust’).78
(3) New industrial economics (or ‘Post-Chicago’ School)
2.48 Introduction On reflection, the Harvard and Chicago Schools may be closer than is ordinarily thought. If one were to pinpoint what they had in common, one would say that each paints too-simplistic a caricature of the way markets operate. It was too restrictive to conclude that all concentrated markets were problematic, as the Harvard School did; on the other hand, it was too sweeping a generalization to conclude that all monopolies and oligopolies were efficient and should be left to their own governance, as the Chicago School did. When not taken to their extremes, both schools offer useful insights.
2.49 The key concept of ‘strategic behaviour’ In the 1970s and 1980s, some economists drew the same conclusion that markets are more complex than suggested by the Chicago School.79 (p. 75) These economists focused on oligopolistic markets, which constitute most contemporary industrial structures, albeit under very specific circumstances. As Professors Dennis Carlton and Jeffrey Perloff explain, on oligopolistic markets interdependent undertakings sometimes adopt so-called ‘strategic’ behaviours which ‘try to reduce the competition exercised by [their] current or potential rivals’ with the ultimate goal of ‘harming [their] competitors’ and ‘thus increasing [their] profits’.80
2.50 The Post-Chicago economists, adopting a ‘behaviouralist’ approach like the Chicago School,81 have identified a number of non-cooperative strategic behaviours which negatively affect market performance and, where necessary, justify the intervention of competition authorities.82 In particular, they have identified some circumstances under which oligopolistic firms have incentives to engage in predatory pricing,83 ‘limit pricing’,84 raising rivals’ costs of production,85 strategic investments (eg advertising expenses,86 or R&D), etc.87 It is important to note that just like the Chicago School models, the Post-Chicago models rest on some strong assumptions and therefore can only be applied to policy with great caution, something that the authors themselves have been quick to point out.88 Moreover, these models are often difficult to implement in practice.89
(p. 76) 2.51 The Post-Chicago economists have also developed a theory of strategic cooperative behaviours. Undertakings sometimes take decisions which make it easier for them to coordinate and which ultimately limit competition.90 Unlike non-cooperative behaviour, where the undertaking that initiates the behaviour tries to reduce the profit of its rivals, cooperative behaviour is aimed at increasing all participants’ earnings, and thus it increases the profit of the initiators’ competitors as well. Cartels provide the clearest example of this type of behaviour. Other types of cooperation that are a priori less harmful but may nevertheless reflect strategic cooperative behaviours are covered as well, for instance the exchange of statistical information, early price announcements,91 the practice of facilitating cartel agreements (such as the widespread use of preferential customer clauses in vertical relations),92 etc.
2.52 Assessment In sum, the main contribution of the ‘Post-Chicago School’ is to rehabilitate the idea that intervention by the competition authorities is legitimate under certain circumstances. How frequently those certain circumstances are relevant in practice remains a subject of debate.93
2.53 Conclusions While it has more explicitly introduced the complexity of the economic realities into competition analysis, Post-Chicago economics has done so at the cost of conceptual unity and, where applicable, the normative force of the Harvard and Chicago Schools. In the final analysis, we are left with individualistic evaluations, entirely dependent upon the circumstances at hand in a particular case, which brings with it a serious problem of legal uncertainty. This is probably a contributing reason why, in various parts of the world, courts and competition authorities continue, according to their decisions, to draw on the teachings of the Harvard and Chicago theories.94
2.54 As is probably apparent from our review of all three economic movements, beneath the apparent clarity of the concepts of strategic cooperative behaviours (illustrated in Art 101 TFEU) and non-cooperative ones (illustrated in Art 102 TFEU) there is a fair amount of complexity and uncertainty. First, regardless of whether we are talking of Chicago or Post-Chicago, modern scholars are divided about the specific standards that need to be applied to the assessment of strategic behaviours. Many—not necessarily consistent—tests have been proposed to assess exclusionary abuses of dominance, such as predatory pricing or anticompetitive rebates. With respect to such practices, EU law is still in flux. Moreover, the analytical tools used by industrial economics today borrow from virtually the whole range of modern economic theories, including, in no particular order, the theory of contestable (p. 77) markets,95 game theory,96 principal–agent theory,97 transaction cost theory (sometimes also known as theory of the firm),98 signal theory,99 the theory of externalities,100 behavioural economics,101 etc.
2.55 Introduction As we shall see, the main focus of study of competition economics is ‘market power’ (Section A). Market power can enable behaviours with pernicious effects on economic efficiency (which competition law, from a purely economic angle, is supposed to be responsible for protecting). Making the same finding, EU competition rules are today based, if not wholly at least mainly, on the concept of market power.102 Economists have designed instruments to help authorities, courts, and undertakings to identify and measure market power and its possible abuses and these are listed in Section B.
2.56 Concept of ‘market power’ Whether an undertaking merges with a competitor, resorts to the strategic behaviours already mentioned, or enters into a cartel agreement, the concern of competition economics is that the undertaking may acquire, strengthen, or exploit market power. Market power describes the capacity of undertakings, by reducing their production, to set higher prices than those which would prevail in a situation of perfect competition (ie, prices exceeding marginal cost)103 and, at least in the short term, to thus make supracompetitive profits.
2.57 Differing degrees of market power Defining market power in these terms might lead one to fear that any market would, at the end of the day, find itself in the sights of competition authorities. Obviously such a situation would not be desirable. Market power is not harmful on its own, rather it is the abuse of market power that raises concerns. Emphasizing this point is the fact that in all markets in which products/services are not perfect substitutes104—that is, all the markets with the possible exception of those for fungible commodities105—undertakings can (p. 79) fix their prices higher than their marginal cost.106 Despite this price setting ability, it quite often happens that on these same markets the undertakings engage in fierce competition, hence making the intervention of competition authorities unnecessary.107
2.58 Market power and competition are therefore not necessarily antithetical. Economists tell us today with one accord (and such consensus is sufficiently rare as to be underlined here) that there are degrees in market power.108 Only certain undertakings enjoy ‘significant’ market power, that is, those that are able to raise their price to a level that is substantially greater than their marginal cost, in the medium, or even long, term. The crux of the problem is obviously to determine at what level of market power anticompetitive concerns begin to appear. And it is on this matter, which calls for a necessarily arbitrary response, that structuralist and behaviouralist economists are frequently split.109
2.59 Extending the concept of market power Focused on the power a firm has over prices, our definition of the concept of market power is incomplete. Apart from quantities, it says nothing about the fact that a powerful undertaking on the market is as often capable of influencing other parameters such as the quality of products/services, which it can make worse while holding price and quantity constant, or innovation, which it can stall .110 The concept of market power therefore encompasses other dimensions.111
2.60 Power to exclude Others, without in any way denying the relevance of power over price, consider that the concept of market power also covers a power to exclude rivals or entrants. This is the theory put forward by Professors Thomas Krattenmaker, Robert Lande, and Steven Salop.112 A firm has power to raise prices above the competitive level (or to prevent any price decrease) when it is able to raise the costs of its competitors and limit their production. An established undertaking can, for example, buy a large amount of shelf space in supermarkets while its competitor develops an intense advertising campaign. Unable to obtain adequate or appropriate display space, and hence unable to make the sales potentially triggered by the advertising campaign, the production of the competitor stagnates. Similarly, (p. 80) an established undertaking can dissuade customers from turning to the products of a new entrant113 by offering them financial benefits (rebates, customer loyalty programmes, etc) or by imposing penalties114 (threatening to stop supplying another key product or, in the case where the established undertaking supplies complementary products, by making competing products incompatible). All these practices may have the effect of raising rivals’ costs. If one takes the example where an undertaking stops supplying a complementary product or service that is essential to consumers, the new entrant will also have to penetrate this new market if it wants to supply the consumers in the first market. These practices clearly limit production. If one takes the example of customer loyalty programmes, the new entrant can be dissuaded from marketing new capacities. The increase in production, which should have led to a price reduction, does not occur.
2.61 Practical scope This less classical form of the concept of market power has been recognized by European authorities.115 For instance, in the area of merger control, authorities traditionally investigate the future risk of significant market power being created or strengthened. In their investigations, authorities now tend to test behavioural theories of anticompetitive behaviours, which are directly inspired by the works mentioned earlier. They will, for instance, try to determine whether the merged entity will engage in behaviours that raise rivals’ costs. Such scenarios are frequent in vertical mergers, that is, the merger of two undertakings operating at distinct stages of the production process. When one of the parties to the operation controls certain resources that are key to downstream operators who are competitors of the other party to the operation, the newly merged entity is likely to increase the costs of its downstream competitors by making it more difficult for them to be supplied with the input at prices and on terms identical to those which would have prevailed if there had been no merger.116 117
Illustration: behavioural theory of anticompetitive conduct—the Alcoa/Reynolds case
In 2000, the two US aluminium producers, Alcoa and Reynolds, notified the Commission of a planned merger.117 This merger had a vertical dimension. Reynolds was the main producer of P0404, a specific quality of primary aluminium used in the manufacture of aluminium alloys for the aerospace industry. Downstream, Alcoa produced alloys of this type. Its sole competitor was McCook Metals, previously supplied by Reynolds. The Commission was concerned that post-merger McCook would no longer be supplied with the raw material by the merged entity—or would be supplied on disadvantageous conditions—and would therefore be excluded from the downstream market for aluminium (p. 81) alloys for the aerospace industry.118 During the proceedings, Alcoa offered to share with an independent third party part of a foundry producing P0404 and, where necessary, to maintain competitive supplies of P0404. The Commission held that these commitments were appropriate, and cleared the transaction.119
2.62 Preliminary remarks Looking for an absolute instrument with which to measure market power has always come to naught. Despite the efforts of many economists, no instrument manages clearly to measure whether or not market power exists. At the end of the day authorities, courts, and undertakings must rely on a range of instruments, concepts, and criteria meant to proxy market power, as listed below.
(1) Direct measurement of market power
(a) The Lerner index and the semantics of cost
2.63 The Lerner index As market power has been defined as the power of a firm to fix prices at a level above that which would prevail in a perfectly competitive market, that is, above marginal costs, Abba Lerner considered that the best way to evaluate market power would be to compare the level of prices in a given market with the level of prices which would exist in the same market in a situation of perfect competition.120 According to Lerner, the gap between the price of the seller and its marginal cost of production reveals its degree of market power. This measure is essentially the undertaking’s gross margin.
2.64 While this comparison is easy to grasp conceptually, the level of the gap between the two figures is not necessarily representative of the extent of market power. For example, in some markets, a gap of several hundred euros on an extremely expensive product (eg a luxury car) may not identify meaningful market power, while a gap of just a few euros on an inexpensive good (eg a food product) may. To account for the problem of levels, Lerner designed an index that expresses relative values. The index (L) measures the ratio between (i) the price (P) less marginal cost of production (Cm) (eg the absolute price gap) and (ii) the price (P).The marginal cost of an undertaking (sometimes referred to as incremental cost), is defined as the cost incurred to produce an additional unit of output (in actual fact, the cost of the last unit of output produced).121 Thus, Lerner’s index measures the proportion of a price offered on the (p. 82) market which is greater than its costs. The undertaking has market power if the ratio is greater than 0. The closer the index gets to 1, the larger the undertaking’s market power.
2.65 Marginal cost, theoretical standard Although attractive in theory, in practice Lerner’s index only provides limited indications.122 First, note that it is a short-term measure because it is based on marginal costs. In industries with high fixed costs that must be recovered through apparently high prices (in that they exceed short run marginal costs), the Lerner index will suggest greater market power than actually exists.
2.66 Second, the index glosses over real-world complications that can have a material impact on the analysis. Let us imagine a simple example in which a firm produces electricity. It costs €500 million to build a nuclear power plant. It then costs €1,000 to produce each megawatt of electricity (the productive capacity of modern electric nuclear generators is estimated to be between 500–2,000 MW). The total cost of the first unit is €500,001,000; the total cost of the second is €500,002,000; the total cost of the third unit is €500,003,000; and so on. In our example, the marginal cost of the first unit is €500,001,000–500,000,000, that is, €1,000; the marginal cost of the second unit is €500,002,000–500,001,000, that is, €1,000; and the marginal cost of the third unit is €500,003,000–500,002,000, that is €1,000. Here, marginal cost is not only easy to identify but it also remains constant as production rises. That is not always the case. Indeed, it is typically quite hard to calculate incremental costs in practice.123
2.67 First, with the exception of certain sectors, undertakings do not typically produce one additional unit of output but will rather produce series of additional units. In other words, production does not increase in a smooth linear fashion but instead increases in steps. Under such circumstances, how should one calculate the cost of a single marginal unit? Second, sometimes a single input is used to produce more than one good at a time, such as when the lights in a factory shine equally on two distinct production lines. How should the utility bill be allocated to each product’s cost?
2.68 Besides the problems of isolating incremental costs in the first instance, analysing marginal costs poses a range of additional issues. The electricity example was obviously simplistic. The marginal cost of an undertaking is rarely constant. As their production increases, undertakings achieve efficiency gains (effects of training, gains in productivity, etc) or, on the other hand, incur new fixed costs, when, for example, their output capacity reaches saturation level (need to build a new power plant, exhaustion of pollution emission quotas, etc) or their production tool deteriorates or suffers an incident following heavy use (breakdowns). Depending on the situation, the marginal cost can either decrease or increase as production rises, and can do both over a large enough range of output.
2.69 The difficulties in cleanly isolating costs are the likely explanation for what has been called the ‘marginalist controversy’. In the 1940s, a study by Richard Lester challenged the ‘marginalist’ premises of neoclassical price theory by demonstrating that undertakings (p. 83) themselves do not calculate their marginal cost.124 Despite its imperfections, however, marginal cost remains a useful and frequently employed concept in economics.
2.70 Productive inefficiency and high costs of monopolies Aside from the complications of accurately measuring marginal cost, the Lerner index also underestimates the market power of the most powerful undertakings. In the most serious cases of market power, the gap between price and costs may shrink away to very little. This can occur, as noted earlier, because firms with substantial market power often have little incentive to keep costs low or production efficient and may indeed use revenues for non-monetary profit taking, in the form of lavish offices say, which of course show up on the accounting books as costs.125 If price margins do not keep pace with rising (inefficiently so) costs, then the Lerner index will be misleadingly small.
2.71 Marginal cost and non-manufacturing industries In many sectors, and this is particularly so in the IT sector, the marginal costs of undertakings are minimal or close to zero.126 In such industries, firms incur large fixed costs, the size of which bears no correlation to the volume they intend to produce. For example, the invention of a new piece of software, production of a film, development of a biotechnology research tool, design of a financial product, etc, all involve substantial upfront costs but low to no incremental costs of use.127 While upfront costs are high in these markets, ‘variable costs’, that is, the cost that depends on the volume of production, are low.128 Burning a software programmes on a disk, granting intellectual property (IP) licences for broadcasting a film, licensing a biotech research tool, or marketing a financial investment—all these actions involve only reproduction or distribution costs and perhaps some enforcement costs for protecting IP rights, but do not involve sizeable recurring charges analogous to a manufacturer’s variable costs. In these non-manufacturing sectors, the marginal cost of putting an additional unit of the products/services on the market will therefore be insignificant.129 In this case, if prices were set equal to marginal costs, the undertakings would be unable to recoup their large, and typically risky, upfront investments and would thus leave the market. But applying the Lerner index to such sectors would lead to the erroneous conclusion that all market players hold significant market power. (p. 84) This problem would also be present in all other industries where fixed costs are high and variable costs low: energy, air transport, telecommunications, rail transport, the pharmaceutical industry, etc.
2.72 Conclusions What must be well understood is that the problems we have just identified present a significant obstacle to reliable and accurate measurement of market power by competition authorities and courts based on the Lerner index: if undertakings do not know their marginal cost, how could third parties (the competition authorities and courts) be expected to measure it accurately,130 especially since there are asymmetries of information and risks of the firms overestimating their costs in order to escape investigation?131 In light of this problem, competition lawyers and economists have tried to design other measurement tools in the never-ending search for better guidance on the market power question.
(b) Measuring profits
2.73 Introduction Intuitively, it might be tempting to infer a causal link between the high profits observed in a sector and the holding of significant market power. This, in any case, is the link upon which many litigants relied in the 1970s in the United States to prove the existence of illegal monopolies. After all, a monopoly can typically charge supra-competitive profits. There were therefore prima facie good reasons for making profits the legal ‘paradigm’ of market power.
2.74 Type I and II errors Making profits the benchmark for measuring market power is not always appropriate. As already noted, monopoly undertakings which enjoy market power are sometimes inefficient, and so their reported profits may be limited. Moreover, in the event of an abrupt fall in demand, even a monopolist will post heavy losses. Similarly, relying on profits is likely to result in a type II error (or false negative), assuming some undertakings do not have market power when if fact they do. Thus a type II error appears when an authority fails to control a conduct which harms economic efficiency.132
2.75 In addition, large profits pose a similar problem of identification of costs133 since the level of profits can be explained by the achievement of major efficiency gains (decreasing marginal costs) that cannot be replicated by other smaller scale operators (eg new entrants).134 Focusing intervention on firms making significant profits, although the firms may just be reaping the rewards of their industrial and commercial superiority, would lead here to a type I error (or false positive).135 The repercussions of this enforcement failing can distort both the (p. 85) immediate market (by hindering an efficient undertaking) and long-term competition (by reducing the expected rewards of investing in efficient production in the first place).
2.76 Accounting profits or economic profits? As noted, litigants in the 1970s usually relied on documentation showing large profits to persuade authorities and courts of the existence of significant market power. In this respect, they usually cited the ‘accounting profits’ of the undertakings, pointing to the values shown on the assets side of the annual balance sheet of the undertakings in question. However, in the early 1980s, economists showed that this was a profound mistake for at least two reasons. First, accounting profits are calculated without taking into account the opportunity cost of mobilizing capital.136 The problem is that a firm, which decides to invest x euros as capital, takes a risk as it gives up other market opportunities (to allocate its resources differently). Hence, this firm incurs an opportunity cost (the cost of what it gives up) that is in no way reflected by the book value of the capital. Second, accounting profits do not inform on the timing of investments or enable economists to identify easily the incremental profits from a particular action. For example, the depreciation of past investments (such as when a piece of equipment loses production capacity over time and with use), the allocation of taxes and costs, and so forth will all prevent the final accounting profit figure from being economically predictive.
2.77 Measuring ‘economic profits’, that is, the total income less the cost of labour, input, and capital and including the opportunity cost will actually reflect the profits of an undertaking.137 This definition means that it is possible for an undertaking to achieve significant profits in the accounting sense, but for profits to be low, or even zero, in the economic sense; hence, the often condemned unreliability of profit-measurement tools.138
(c) Price comparisons
2.78 Introduction If it were possible to identify the competitive level of prices that would otherwise prevail, it would then be fairly simple to determine if an undertaking exercises significant market power by simply comparing existing prices to the competitive prices. According to some, this type of approach would be possible, and desirable, by introducing some degree of empiricism,139 based on what economists call ‘natural experiments’. Such experiments essentially look within a given sector for data on the market in time (observation of price differences on the market at various comparable times), in space (observation of price differences on distinct but comparable markets), or in sectors that are close but distinct (observation of price differences on geographical markets).140 At the end of such an investigation, it may (p. 86) be possible to ‘reconstruct’ one (or more) market(s) that appear structurally competitive(s).141 All that is then needed is to compare the price of the undertaking in question with the estimated price on those benchmark markets (‘competitive benchmark price’).
2.79 Mobile phones in Ireland Such natural experiments can be illustrated by the following example. In 2007, Professors Gregory Sidak and Jerry Hausman published an analysis focused on the mobile phone market in Ireland, which was composed of two operators controlling nearly 94 per cent of the market.142 The national regulatory authority had deemed that these operators held significant market power. The authors proposed an alternative analysis of the market based on the method just presented. In particular, they focused on the prices on the British mobile phone market, which the Irish national authority deemed to be competitive. The UK would appear to be a valid benchmark for Ireland since, in terms of costs, technology, and regulations, the British market shared many characteristics with the Irish market. Sidak and Hausman proved empirically that mobile prices were actually higher in the UK, where competition was presumed, than in Ireland, where only two operators essentially shared the market.
2.80 Impracticality Because it is conceptually simple, the empirical benchmark approach is undeniably appealing. The sector analysed by Professors Sidak and Hausman was, however, exceptional in that the presence of authorities controlling the day-to-day operation of the markets offered accurate and up-to-date information. Moreover, due to the harmonization Directives adopted at Community level, the regulatory requirements imposed in both Member States were relatively close. The problem with this approach is that finding two markets that are sufficiently comparable may be extremely difficult. As Sidak and Hausman seem to acknowledge,143 it will not always be possible to identify two appropriate comparators, with the possible exception of adjacent retail markets.144 As a result, this approach can be but one tool in the box; it will not always be feasible.
(d) Price-elasticity of residual demand
2.81 Other techniques enable market power to be measured directly—for example, by measuring the price-elasticity of residual demand faced by an undertaking.145 The price-elasticity of demand in the market measures the reaction of the quantities that are wanted in response to a price increase—if quantities remain stable, demand is said to be inelastic; if they decrease considerably, demand is said to be elastic. The concept of residual demand faced by an undertaking is the total demand of the market less the quantity supplied by other undertakings.146 To put it more simply, it is the fringe of demand not supplied once other undertakings have satisfied their customers and sold the quantities they have.147 One might say that this is the natural customer base of the undertaking in question.
(p. 87) 2.82 If, on this portion of demand, the undertaking is unable to increase prices without incurring a decrease in demand (customers stop consuming or turn to its competitors), demand is elastic and the undertaking is deemed not to have market power. Conversely, if its order book is kept full in spite of a price increase, demand is inelastic and the undertaking has market power. Although attractive in theory,148 measuring market power using residual demand faced by an undertaking poses significant problems in terms of collecting information and conducting the econometric calculations.149
2.83 Preliminary comments Given the limits to marginal cost, residual demand, and other direct approaches to measuring market power, competition economics traditionally has taken recourse to other instruments. For instance, Professors Landes and Posner have expressed the view that market power should be estimated using a formula that takes into account market shares, as well as the general elasticity of demand and supply.150 Today, competition authorities typically rely on multi-factor analysis, combining a review of market shares (Section 1) with other indices (Section 2).
(1) Structural measurement
(a) Market shares
2.85 Value or volume? According to economists, measuring market shares in terms of revenue (ie, sales of each firm in relation to sales of the sector) certainly provides the most reliable indirect assessment of the power of operators on the market.151 To reconsider the mobile phone example, a legitimate question is whether it is a good idea to evaluate market shares in terms of volume (number of subscribers of each operator) when it is possible that some only rarely use their telephone. A measurement in terms of revenue is a truer reflection of the economic importance of each operator. Even so, revenue-based market share is not always representative either. On markets where there are a lot of firms, input reserves and capacity can be indicative of market power. For instance, anticipating reserves of raw materials (eg in the steel industry, where the stocks of scrap metal ore must be bought in advance) or where firms have production capacities in reserve, an undertaking that has low reserves and is therefore incapable of influencing quantities/prices in the mid and long term, will only have weak market power, even if, at the time of evaluation, it makes most of the sales in the sector.152 The European authorities, as a matter of principle, adopt the opposite approach, placing more weight on current sales.
(p. 88) 2.86 Characteristic of economic operations Assessing market shares is intrinsically linked to the way a particular market works. In some markets where transactions are infrequent and irregular in terms of size (eg the aeronautical sector) market share can vary greatly from one year to another. In fact, in any industry in which ‘design wins’ are important, ‘share’ can fluctuate considerably over time. Thus care must be taken to ensure that assessment of market share is performed over a sufficiently long period.
2.87 Similarly, when the focus is on the future market power of an undertaking (eg in the prospective area of control of concentrations) it is crucial to take account of anticipated changes. For example, suppose an undertaking currently has a ‘large’ share, but due to a recent contractual win is at production capacity. If other sales contracts are expected to come up for bid in the short term, other undertakings clearly will be better positioned to win them.
(b) Measuring concentration
2.88 Concentration ratios Simply counting the number of firms on the market is not always a true reflection of the actual size of an oligopoly. Certain markets with a large number of operators can in fact be controlled by a handful of undertakings. The risk of tacit collusion is then greater than suggested by simply counting the undertakings that are active in the market. This issue can be addressed by looking at market concentration ratios. A concentration ratio (CR) is the sum of the market shares of the m largest firms (CRm). Thus, the ratio CR4 is the sum of the market shares of the four largest firms in a market. A market composed of ten operators, but where the ratio CR4 is 80 per cent (the four largest firms hold 80 per cent of market shares) can be described as oligopolistic.
2.89 Two instruments are generally used to measure concentration: concentration rates, which are nothing other than the market share held by a given group of undertakings, and concentration indices. The best known index is the HHI (Herfindahl-Hirschman Index, named after the two economists who invented it), which equals the sum of the square of the market shares of each of the undertakings present in the market.153 The index ranges from zero (perfectly competitive) to 10,000 (monopoly); it goes up when the number of undertakings goes down and when asymmetries of market share between operators increase.154
(2) Additional measurement tools—assessment of obstacles to entry and expansion
2.90 Introductory comments Market share and the measurement of market concentration are, in the view of many economists, ‘poor’ indicators of an undertaking’s market power.155 To our knowledge no economist would dare to define an absolute threshold for presuming significant market power based on a level of market share/concentration. At best, market shares serve as a mechanism for ‘filtering’ market power. How then can authorities and courts measure market power more precisely? Economic theory tells us that in addition to the structural measurement tools, the focus must be expanded to other parameters, that is, the obstacles (usually referred to as ‘barriers’) to entry and expansion.
(p. 89) 2.91 Understanding barriers to entry The ‘contestable market theory’, which was developed in the early 1980s,156 embodies the idea that faced with the threatened intrusion of potential entrants (or the expansion of current competitors), a firm is, in principle, deterred from cutting its production or raising its prices.157 On the other hand, when a firm is protected by barriers to entry, it will be able to exert a certain influence on prices and, hence, will enjoy market power. If the existence of barriers to entry and expansion is not in itself sufficient to conclude that market power exists,158 it is nonetheless a crucial parameter in the analysis undertaken by competition authorities, regardless of the practice/operation in question (determination of a dominant position, prospective analysis of the effects of a concentration, etc).
2.92 Entry or expansion? Economists are equally interested in barriers to entry (ie, obstacles to the entrance of operators that are not currently present on the market) and in barriers to expansion (ie, obstacles to the growth of the production capacities of incumbent competing operators). In fact, whether it is an undertaking that is likely to enter the market (referred to as a ‘potential competitor’ or ‘new entrant’) or an undertaking already present which might expand its activities (‘current competitor’), the competitive pressure being exerted on the market power is inherently the same.159
(a) What is a barrier to entry?
2.93 Harvard vs Chicago, again From a simple commonsense point of view it would be tempting to say, as the OECD does, that a barrier to entry is ‘any factor which makes it harder for potential applicants to enter a market’.160 But this definition would ignore a significant economic controversy between the Harvard and Chicago economists.161 The Harvard economists, whose views are expressed through Joe Bain, define a barrier to entry as
any advantage of incumbent firms in an industry over potential entrants, reflected in the extent to which the incumbent undertakings can raise their prices above the competitive level without encouraging new competitors to enter the industry.162
Bain also indicates the three main factors that contribute to the presence of a barrier to entry: absolute cost advantage, economies of scale requiring major investments, and product differentiation. Bain therefore proposes an empirical definition rooted in examining the prices practised on the market.
a production cost (over all possible production levels or over a subset of the latter) which must be borne by the firms seeking to enter an industry but which is not (or has not been) incurred by incumbent firms.163
Stigler therefore proposes a definition more closely aligned with neoclassical theory and focusing on the costs incurred by the undertakings. Of course, one of the three advantages that Bain lists are cost advantages, so the two definitions have some fundamental aspects in common.
2.95 Illustration A brief example taken from the electronic communications sector highlights the differences and, in some cases, the analogies, between these two approaches. If the entrants and the incumbent operators have identical access to the technology, Stigler’s view is that economies of scale and other productive, financial, advertising, and other investments (costs already realized by the incumbent operators) do not constitute barriers to entry, whereas Bain thinks that they do.164 On the other hand, if the incumbent was ‘offered’ its telecommunications network by the authorities, the (costly) construction of an alternative network by the new entrants constitutes a barrier to entry according to Stigler’s definition.165 Similarly, the incumbent operator generally enjoys a ‘first-mover advantage’ which would act as a barrier to entry according to both Stigler’s and Bain’s definitions: the incumbent operator, which has for a long time had no competitors on the market, has been able to build a customer base for itself without undertaking any significant marketing. In order to capture customers new entrants must, on the other hand, spend large amounts for marketing. These (asymmetrical) investments constitute a barrier to entry.166
2.96 Assessment Bain’s definition is certainly more inclusive than Stigler’s. Of the various different criticisms that have been made of this definition the most persuasive is that Bain’s definition covers elements which concern the efficiency of the undertakings already present in the market, such as, for example, a cost advantage or a profitable product differentiation strategy.
2.97 On the other hand, Stigler’s definition has been accused of being too narrow in that (i) it excludes from barriers to entry irrecoverable costs (which are also incurred by the incumbent undertaking)167 and (ii) it does not take into account the absence of similarity of the cost structures between the undertakings.
(p. 91) 2.98 Finally, Bain’s and Stigler’s detractors have accused them of not being interested in ‘strategic barriers to entry’, that is, behaviours adopted by the incumbent undertaking to stave off the entry of competitors (eg making threats of exclusion by public announcements, making a pre-emptive acquisition of the capacities of marginal operators eyed by the entrant, establishing excessive production capacities,168 registering ‘dormant’ or ‘immaterial’ IP rights (for blocking purposes), etc).169
2.99 Recent discussions of the definition of barriers to entry tend to show that a static structural approach to barriers to entry is not sufficient; there must also be a more dynamic and behavioural approach. For instance, even if an entrant’s and an incumbent’s costs are similar, there may be a barrier to entry if a new entrant must incur those costs over a shorter period of time.170 Likewise, economies of scale and investment costs, as identified by Bain, may represent barriers if they delay entry and reduce social welfare.171
(b) Characteristics of a barrier to entry
2.100 Status of problem As interesting as it may be in theory, the debate relating to the definition of the concept of barriers to entry seems to have had a minor impact in practice. What is important is not so much the formal content of the concept of barriers to entry but the ability of the obstacle in question to deter the probable entry (or expansion) of a competitor which would be timely and sufficient .172
2.101 Preventing ‘timely’ entry It would be wrong to think that only obstacles that are insurmountable in the long term for potential entrants are barriers to entry. On the contrary, since the authorities are interested in the competitive pressure exerted on undertakings that are already present in the market, anything that on the day of examination, or in the near future, constitutes an obstacle to entry deserves to be considered as a barrier to entry.173 Following the same line of thought, long-term entry (eg with a time line three or four years down the road) is not likely to constrain the current price policy of incumbent operators.
2.102 Preventing ‘probable’ entry The aim of analysing barriers to entry is to determine the existence and the extent of the market power of an undertaking that is already established on the market. This is why the analysis must not be simply theoretical: an answer must be found to the question of whether entry (or expansion) is not only possible within a suitable period in order to put pressure on the incumbent undertaking, but also whether it is probable. In other words, it is not a matter of finding out whether an undertaking might enter a market (p. 92) but of knowing whether it is likely to do so174 because it has a commercial interest in doing so.175
2.103 Preventing entry of ‘sufficient’ competition Because only entry (or expansion) that is sufficiently intense is likely to constrain the incumbent undertaking, the extent and importance of the entry (expansion) that is envisaged must be examined as well.176
(c) Types of barrier to entry
2.106 Absolute cost advantages These advantages comprise the favourable economic conditions enjoyed by firms already present on the market which new entrants cannot enjoy (so that new entrants would be unable to replicate the prices offered by incumbent firms).179 The question of innovation in the market is crucially important here: in a market that is exposed to a sustained rate of innovation, the cost advantage enjoyed by an undertaking may, in the medium term, prove to be particularly precarious.
2.107 Economies of scale With an identical cost structure, undertakings already present on the market may benefit from economies of scale resulting from a high level of production. The undertakings can then allocate their fixed costs (eg the cost of obtaining their premises) over a relatively high number of units produced and thus reduce their average production cost.
2.108 Economies of scope An incumbent producing a range of products can allocate its common costs over all the units of the various products and, if need be, cut the average production cost of the different products. In other words, economies of scope allow an undertaking to crosssubsidize production. A potential entrant who is not present on these various other markets would not be able to replicate the prices of that undertaking.
2.109 High investment costs The amount and cost (interest) of capital required to enter into a market may constitute a barrier to entry. Financing terms can differ substantially across investment projects depending upon their relative risk levels, and interest rates can vary (p. 93) across undertakings even within the same industry due to the differing ability of firms to repay loans.
2.110 Product differentiation and advertising investments Product differentiation and advertising investments are often considered to be barriers to entry180 for they often go hand in hand with loyalty to the brand of the incumbent. A potential entrant may decide not to enter a market if it has to make major expenditures in order to advertise its products and encourage the customers of the incumbent to change their consumption habits. It should be noted, however, that when it is permitted (ie, in most of the economic sectors) advertising can have a positive effect on entry. By improving the information provided to consumers, it can increase sales opportunities.181
2.111 Reputation effects The notoriety acquired by undertakings already present in the market can constitute a barrier to entry to the extent that the new entrant must make special efforts (eg investments) to attract customers who consider reputation to be an important factor in their purchasing decision.182
2.112 Switching costs183 These are the costs incurred not by the entrant but by a consumer who switches from one supplier’s product to another’s. When this cost is high, elasticity of demand is weak and the new entrant is faced with a barrier to entry.184 An example helps to clarify. Previously when a consumer switched his mobile phone operator, he lost his phone number, necessitating the often time-consuming process of notifying all his contacts of the change. This switching cost made it more difficult for customers of an operator to switch to another operator, even in the face of mobile service price increases. Regulations requiring number portability have eliminated this switching cost.185
2.113 Network effects Direct network effects appear when the usefulness that one customer derives individually from the consumption of a good/service increases with the number of other people who collectively consume it. Hence the individual usefulness of being connected to the telephone network increases with the number of subscribers collectively connected to the network. Indirect network effects appear when the size of the network affects the quality or cost of the good/service. For instance, an individual does not benefit directly when she purchases a widely used printer, but because it is widely used she is likely to find that service and repair is easier to find in her area and the cost of replacement toner cartridges may be lower as well. An undertaking wishing to enter a market which has network effects will tend to encounter barriers that may even be virtually insurmountable,186 whilst the (p. 94) incumbent undertaking can take advantage of its customer base to expand the number of its customers.
2.114 Regulatory barriers Notwithstanding its merits or how opportune it is, the existence of a regulation, law, or any other legal instrument that imposes conditions on undertakings within a given market can also constitute a barrier to entry.187 The practice of setting quotas to regulate access to certain professions (eg the legal and medical professions) is a good example of this.
2.115 Barriers to exit These are all costs that an undertaking might incur if it leaves the market. These barriers therefore include sunk costs188 (costs that cannot be recouped, eg advertising investments), investments that have no salvage value (eg highly specialized machinery), or undertaking dissolution costs (eg the costs of unwinding established contractual relations). Economists consider that exit costs are likely to deter new entrants from entering a market in the first place.189 What is more, to the extent that the fear of incurring large exit costs is also shared by the incumbents on the market, the latter may be all the more determined to prevent any market entry by using strategic behaviours. Barriers to exit therefore can double as barriers to entry for potential entrants.190
2.116 ‘First-mover advantages’ First-mover advantages can take several forms, most of which are related to other potential barriers. Most simply, being first in a market can create reputation effects that later entrants have difficulty overcoming. In markets with network effects, firstmovers can benefit from switching costs. Or, if the market is characterized by learning effects, experience or the acquisition of know-how (eg production experiences a learning curve that takes a ‘positive exponential’ form) then it may be impossible for undertakings which were not present from the start to compete on equal cost footing in the short, or perhaps even in the medium, term with the incumbent undertaking.
2.117 Vertical integration Vertical integration can be a source of strategic advantages. Vertically integrated undertakings, for example, may enjoy security in terms of upstream supplies or, more simply, they may achieve substantial cost savings, particularly relating to transaction costs since they may not need to look for co-contractors to distribute their products, negotiate the terms for distributing their products, or monitor proper performance of the contractual obligations. Finally, on markets where undertakings are vertically integrated, new entrants risk being foreclosed due to preferential treatment between the upstream and downstream operations of vertically integrated firms.191
2.118 These are barriers intentionally erected by incumbents. Unlike structural barriers, they are not obstacles inherent to the market in question but rather barriers created by market participants. While many of these barriers are analysed by economists as measures of the market position of an undertaking (eg its dominant position), lawyers are generally interested in these barriers in relation to the lawfulness of the behaviours of the undertaking in question.
2.119 Strategic tariff barriers A distinction is usually made between three types of strategic tariff barriers: (i) strategic financial barriers, the purpose of which is to block/limit the access of a new entrant (or a current competitor that wishes to expand) to financial resources (capital); (ii) strategic barriers with a signalling effect the object of which is to influence the perception that a potential entrant has of the profitability of the market; and (iii) strategic barriers with a reputation effect which the incumbent undertaking uses to reinforce an aggressive image vis-à-vis new entrants.192
2.120 Overinvestment in capacities By investing in excess production capacities an incumbent can discourage new entrants which may be concerned that once they have entered the market, the incumbent will flood the market with huge quantities, inducing a downward trend in prices that makes entry unprofitable.193 This phenomenon becomes more acute when a potential entrant must incur sunk investment costs in order to compete with the incumbent undertaking.194
2.121 Loyalty discounts and group rebates Provided it enjoys sufficient market power, an incumbent can create an obstacle to the entry (or expansion) of competitors by offering discounts to its customers. The latter can take several forms. First, it can set up a system of loyalty discounts: customers are encouraged to obtain their supplies for their contestable volumes (those which might be satisfied by competitors) from the incumbent by making the grant of a discount conditional upon satisfying a volume or minimum percentage of orders.195 The incumbent can also make the grant of a discount on a product conditional on the purchase of other products.
2.122 Tied sales The dominant undertaking may also erect a barrier to entry on the market by jointly offering several products. If customers respond positively, it will be hard for (p. 96) a potential entrant to penetrate the market without itself adopting this approach. Entry on multiple fronts is, however, more costly than entry on one single product market.196
2.123 Exclusive commercial arrangements An incumbent supplier may use exclusive supply agreements with its customers to limit, or even prevent, competitors from entering the market. These agreements may erect barriers to entry to the extent that the contestable share of the market for new entrants is reduced in proportion to the extent/term of said agreements.197
2.124 Patent hoarding If the existence of intellectual property rights and, where applicable, exclusive operating rights can constitute a barrier to entry on a market, these rights logically fall within the category of regulatory barriers. What is involved here is a strategy used by some undertakings which develop a portfolio of intellectual property rights with the intention of blocking the entry (or expansion) of rival undertakings. Incumbents may, for example, file an array of patent applications in related fields with the competent authorities (related to sometimes minor innovations). The risk of infringing any one of these patents in the event of entry onto the market (and costly court actions that can follow) can deter the entry of new operators.198 These barriers to entry must be examined very carefully for it is particularly tricky to separate what falls under a legitimate intellectual property right from an illegally obtained intellectual property right.
2.125 The key role of production costs in analysing competition Examining production costs in competition law is useful from two points of view. First, a focus on production costs may facilitate the identification of competition law infringements by competition authorities. Prices that are significantly above the production costs of a dominant firm, for instance, may amount to exploitation. Conversely, prices that are below certain measures of costs may be considered exclusionary or predatory. In either of these extreme cases, production costs will be the benchmark against which prices will be assessed.
2.126 Second, undertakings can sometimes justify practices that are deemed suspect by the authorities by formulating counterarguments based on their production costs. One example of this is a merger that generates efficiency gains, particularly when it results in cost savings (eg by means of returns to scale or economies of scope).
2.127 Introduction In the pure and perfect competition model, we have seen the market price drops to the level of the marginal cost of the undertakings. Since it offers a theoretical yardstick of significant market power, if data permitted it would be sufficient to measure the marginal cost of an undertaking against its prices to reach conclusions regarding the undertaking’s capacity to harm consumer welfare.199 Inversely, according to the teachings of neoclassical price theory, when undertakings set prices aligned to marginal cost, the market is deemed to be operating competitively.
2.128 Definition As explained earlier, marginal cost is the cost which would be borne by a firm to make one additional unit of a given product/service.200 Industrial economics, which is interested in the decisions of firms, teaches that this cost is the one which entrepreneurs would calculate to decide whether it would be opportune to produce an additional unit (ie, to choose their level of production), but since in competition law the focus is on the behaviour of undertakings that have already made their decisions regarding production, the marginal cost of an undertaking is represented by the cost of the last unit produced.201
2.129 Controversy Phillip Areeda and Donald Turner, both representatives of the Harvard School, have argued that although marginal costs are worth looking at, they are very hard to calculate in practice.202 We know from the work of economists in the 1930s that in practice undertakings do not calculate this cost—ever. How could competition authorities therefore base their investigations on marginal cost when this information does not exist and must be estimated imperfectly? In fact, modern economists teach us that undertakings calculate their prices based on three considerations: production costs, obviously, but also the commercial objectives of the firm in question (maximizing result, developing a reputation on the market, obtaining share, etc) and consumer behaviour (sensitivity to price, etc). Given this latter parameter, it is possible for consumers to be responsive to high prices that are not aligned to costs—luxury goods come particularly to mind. In such markets, pricing is not generally aligned to costs but seeks to reflect the value perceived by the consumer. What is more, even if access to information about the marginal costs were available, this cost only concerns the production of an additional unit/the last unit, so that it is hard to draw conclusions about the behaviour of an undertaking for anything other than that unit.
(a) Average total costs
2.130 Definition Average total cost (ATC) is the total cost incurred by an undertaking divided by the total number of units. It is perhaps in order to overcome the problems associated with calculating marginal cost that economists and lawyers are, in practice, often tempted, by trial and error, to rely on ATC. And on the face of it, ATC, representing as it does the ‘typical’ cost of a unit produced (per unit cost), has the virtues of a significant yardstick.
2.131 Customary practical applications ATC is often used when assessing the pricing conduct of a dominant firm and, in particular, predatory strategies. EU case law thus considers that certain pricing practices that lead to prices below ATC may, under certain conditions, have a predatory effect.203
2.133 Limited significance of ATC Prices below ATC are not necessarily anticompetitive.204 For instance, firms that conduct what is called ‘loss-leading’ or ‘penetration pricing’ strategies set prices, in the short term, that forgo recovery of all of their ATC (eg they give up some of their fixed costs) to ensure that their products take hold on the market quickly. Grocery stores often follow a loss-leader strategy, setting some popular item’s price below cost in an effort to induce shoppers into the market who will then purchase other, higher margin, items so as to offset the loss.
2.134 Multi-product firms and the allocation of common costs Measuring ATC can prove to be complicated. This is the case when looking at multi-product firms, that is, undertakings which are active in markets that are distinct but which use common production factors (labour or capital) to supply their diverse products. It is a question of both measuring and then assigning common costs, which are defined as the costs associated with two or more activities within the same undertaking.205 Consider an extreme example: what proportion of the remuneration of a postman should be imputed to the cost of delivering one unit of express mail compared to the normal mail that he also delivers during his round? A more common example lies in the cost represented by establishing an HR or legal department responsible for the activities of the various group subsidiaries.
2.135 Economists tell us there must be an allocation of the common costs over the various products/services in question. There are, however, many different methods of distributing (p. 99) common costs over the different activities (and determining the proportion that should be attributed to the supply of a specific product/service). All these methods involve varying degrees of complicated analyses and there is no consensus among analytical accounting experts about any of them. Generally, therefore, even with ATC (as opposed to the more controversial marginal cost), a certain degree of arbitrariness remains.
2.136 Strategic behaviours There has been a lot of talk about cost-allocation issues in the recently liberalized sectors. For instance, in the postal sector, incumbents faced with the new challenges of opening up to competition have sometimes taken to transferring to the accounts of their (State remunerated) universal service activities a large part of their common costs so as to lower only the reported costs on the newly competitive activities (eg express mail). These accounting strategies, also known as ‘cross subsidies’, can be illegal.
2.137 Definition These are production costs which do not vary depending on the quantity of goods or services, such as property taxes, rent, and infrastructure expenses in some cases. The level of these costs remains ‘fixed’ regardless of the number of units produced. In the air transport sector, for example, the expenses incurred to build up a fleet are, at least in the short and medium term, fixed costs.
2.138 Economic law Undertakings with high fixed costs will in principle be able to achieve ‘returns on scale’, that is, they will be able to reduce their per-unit fixed costs by producing more units. Fixed costs are then allocated over a greater scale of production and have less impact on each unit produced. With returns to scale the average fixed cost decreases as production increases.
2.139 Practical impact In practice, the principle of spreading fixed costs over larger production levels directly influences the strategies of firms: undertakings have every interest in adopting behaviours that enable them to increase their production and, if necessary, in exploiting all of these returns on scale. Economics thus sheds a new light on behaviours which competition law sometimes penalizes blindly. A particular instance of this is price discrimination or discounts granted by dominant undertakings: where a single price system necessarily may exclude certain consumers from buying the product (those whose reservation price is lower than the price set) and therefore does not allow the quantities produced to be maximized, discrimination or volume-related rebates allow these customers to be satisfied and raise the scale of production.
2.140 Similarly, Ronald Coase’s famous theory of the firm indicates that if economies of scale are achieved by operators situated downstream or upstream, it may be efficient for a firm to delegate the performance of a task (eg the distribution of its product or the production of an input) to a specialized third party. This is why car manufacturers delegate tyre production to third party specialists. These ‘transactional’ reasons provide objective justifications for the cooperation agreements between undertakings which also concern competition law.
(p. 100) 2.141 Confusion between fixed costs and sunk costs Lawyers and public decision-makers sometimes confuse the concepts of fixed costs and sunk costs (also called ‘irreversible costs’). Sunk costs are, as we have seen, costs which have been incurred and which cannot be recovered in case of exit (advertising expenses, certain innovation expenses, etc). Although it is true that sunk costs are, if not exclusively at least primarily, fixed costs,206 the opposite is not true: a fixed cost is not necessarily irrecoverable. To take another example in the air transport field, acquiring the fleet is a fixed cost which may be recovered in case of exit since planes can be resold on the second hand market or reallocated to other routes.
2.142 Dogma Lawyers sometimes say that sunk costs are not taken into account by firms when defining their pricing policy. In other words, firms would not seek to recover those costs from their customers.207 This claim is obviously going too far. All undertakings, for example those which have incurred major innovation expenses, seek to recover some of those costs over the long term. If an undertaking does not expect to recoup R&D costs through sales in the market, for instance, it will not invest in that R&D. This point is especially important in highly innovative sectors, such as pharmaceuticals and IT.
2.143 Definition These are costs which vary depending on the production volume. The costs of inputs or energy used in the production process are in principle variable costs. In the transport sector, fuel expenses are a variable cost. Because measuring them tends to be easier, economists often use average variable costs (AVC) instead of marginal costs (which depend solely on variable costs since fixed costs do not vary based on production).208
2.144 Economics Economic theory teaches that an undertaking will stay on the market as long as it is able to cover its variable costs in the short and medium term (ie, its unit price is higher than or equal to its average variable costs) since recovering fixed costs is in fact a long-term concern. On the other hand, if an undertaking does not manage to cover its variable costs it is, in principle, forced to exit the market. This is why an airline which has high fixed costs and low variable costs continues to fly its fleet of aircrafts even when the occupancy rates of its planes are low.209
2.145 Practical impact AVC has practical applications in the two previously mentioned areas of identifying conduct violating competition law and justifying conduct that may be deemed in violation of competition law. First, when it comes to identifying an infringement, AVC is traditionally used by competition authorities in the context of predatory pricing tests.210 In short, a dominant undertaking which prices its products or services at a price that is lower than its AVC is undoubtedly seeking to squeeze out its competitors. It is not acting rationally (p. 101) since each unit that it sells represents a net loss. Although it is often considered to be a satisfactory approximation of marginal cost, using AVC has been, however, criticized so that authorities seem to be moving away from it in favour of average avoidable costs (AAC). AAC are defined as those costs that are entirely avoided if production ceases.
2.146 Next, looking at ‘justification’, dominant undertakings sometimes refer to their AVC to justify discriminatory pricing behaviour or price cutting, which the authorities suspect is anticompetitive. The idea is simple: a dominant firm which grants a price reduction to some customers and agrees to sell its product above AVC is not necessarily seeking to foreclose its competitors. Since any price higher than the variable cost enables it to recover at least some of its fixed costs (the difference between the price and the average variable cost), serving these customers, including at a knock-down price, is more efficient than failing altogether to satisfy their demand. This is why it is usually said that price discrimination, rebates, or discounts result in more rapid recovery of fixed costs.
2.147 When do costs cease being fixed A fixed-cost item may, in time, become variable, just as a variable cost item may become fixed. Outsourcing is one example. When an undertaking itself produces an intermediate component which it incorporates into final goods, it must incur certain fixed costs, in particular fixed assets (equipment, etc) associated with that component. But if that undertaking decides to outsource production of that same component, it no longer incurs a fixed cost but a variable cost corresponding to the purchase of the component from a third party.211
2.148 Similarly, any fixed cost becomes variable in the long term. For instance, the cost of building a nuclear power plant, which is a fixed cost in the short term for any energy producer (regardless of whether it produces 0 or 1,500 MW) necessarily becomes a variable cost when it is looked at over the long term. Once a power plant reaches its capacity limit, another one must be built to increase production. The construction cost of a power plant is therefore linked to the relatively large increase in the quantities produced.
2.149 Definition As mentioned briefly above, AAC represent all average costs incurred by an undertaking for the supply of a specific product/service and which are recoverable over a given period .They therefore include not only the AVC but also the part of the fixed average costs which (i) is specific to the production of the good/service in question and (ii) can, over a certain period, be recovered by the undertaking. The term used is ‘avoidable cost’ as a way of simply referring to what the undertaking would have saved by discontinuing the production in question.212 One of the main advantages of the notion of avoidable costs is that it overcomes the shortcomings that go along with the measurement of variable costs: those of fixed costs which, as we have seen, influence the pricing decisions of firms (those they want to recover eventually) are taken into account here;213 common fixed costs that cannot be imputed to the good/service in question are not taken into account.
(p. 102) 2.150 Practical impact The concept of AAC has been used increasingly in competition law in recent years,214 particularly with regard to characterizing exclusionary or predatory price policies of dominant undertakings, that is, abusively low prices. The variable cost standard, in fact, previously allowed firms that incurred large fixed costs to escape any criticism under competition law despite the potentially anticompetitive efforts of their practices. The avoided cost standard, which is less broad and more closely tied to business decisions, is undoubtedly going to become the common legal standard in EU competition law.215
2.151 Definition Average incremental costs are made up of all costs specifically linked to an increment, that is, an increase, of production. This concept therefore covers not only variable costs but also fixed costs, whether recoverable or not. The difference with marginal cost lies in the fact that average incremental cost corresponds to the average cost of a set of additional units and not of a single one, which is more realistic since firms do not decide to produce by unit but by ‘a set’ of units. As with AAC, average incremental cost does not include the common costs relating to other products, which makes it a standard that is useful for multiproduct firms.
2.152 Practical impact Long-run average incremental cost (or LRAIC) is regularly used in the newly liberalized network industries. These industries have two characteristics that make the LRAIC a cost standard that argue for its use in reaching conclusions about the existence of a competition law infringement (eg predatory prices). First, undertakings that are active in these sectors often have multiple activities, including what are called ‘reserved’ activities, that is, where they enjoy a monopoly behind which they can attempt to conceal the costs of competing activities.216 Second, LRAIC is used within a relatively long-time frame in order to take account of the often considerable investments made in these sectors (infrastructure investments) which are only incurred by firms that take a long-term view.217 However, as recognized in the Article 102 Guidance Paper, LRAIC is not applicable when products and services can be sold in bundles.218 In that case, the relevant comparison is not whether incremental revenues cover incremental cost, but rather is whether the bundle as whole is priced (p. 103) in a predatory fashion. Moreover, sometimes costs are common across the provision of several goods, which requires an allocation.
2.153 Definition These are irrecoverable costs of a specific type since they only appear in certain industries. They can be defined as the costs incurred by an undertaking due to the universal service obligation (eg picking up the mail and distributing it to households at least once each business day) which the undertaking has and which it would not have incurred if the activity to which the costs relate had been a competitive activity.219
2.154 Practical impact Incumbents with universal service obligations have often taken advantage of significant stranded costs in order to (i) challenge the price reduction obligations imposed by the sector regulators based on lower cost measurements (eg average variable costs) and/or (ii) justify their high price policies by the need to recover their investments from the customers.220
2 See L.-H. Röller, ‘Economic Analysis and Competition Policy Enforcement in Europe’ in P.A.G. van Bergeijk and E. Kloosterhuis (eds), Modelling European Mergers: Theory, Competition Policy and Case Studies (Cheltenham: Edward Elgar, 2005), 14. It should also be noted that approximately 200 of the 700 civil servants who make up DG COMP have training in economics. Around 20 civil servants hold a doctorate in economics. M. Monti, Competition Commissioner between 1999 and 2005, was Professor of Economics at the University of Bocconi in Milan.
5 See eg the appointment of D. Turner in 1965 as Assistant Attorney General for Antitrust at the Department of Justice. The office of the US judge is also becoming more economics-based. Lawyers with advanced economic training, such as R. Posner or F. Easterbrook exercise judicial functions.
6 See D. Gerber, ‘Law and the Abuse of Economic Power in Europe’ (1987) Tulane L Rev 57. D.B. Audretsch, W.J. Baumol, and A.E. Burke, ‘Competition Policy in Dynamic Markets’ (2001) 19(5) Int’l J Industrial Organization 614: ‘If there is any body of law that owes its existence to economics, it is surely antitrust law.’
7 See our arguments on the origin of the Treaty’s competition rules in Chapter 1.
10 However, in the past European judges and the Commission, did take, from time to time, economic analyses into account. The Wood Pulp II judgment is a case in point. The Court ordered two opinions by economic experts for the purpose of determining if the price parallelisms observed in the wood pulp market could be explained by the oligopolistic characteristics of the sector. See CJ, C-89/85, C-104/85, C-114/85, C-116/85, C-117/85, and C-125–129/85 Ahlström Osakeyhtiö et al v Commission, 31 March 1993  ECR I-1307.
12 See Green Paper—Community competition policy and vertical restrictions, COM(96) 721, 22 January 1997, at para 65: ‘For policy purposes, it is necessary to translate the conclusions from economic analysis into workable tools that are both consistent with EC competition rules and relatively easy to implement with the necessary legal certainty for undertakings’ (emphasis added).
13 In order to assess market power, the Commission uses market share thresholds. Below certain thresholds, practices are presumed to be compatible. Above certain thresholds, the practices must each be subject to an in-depth examination. A legitimate question is whether the market share approach is actually much of an economic improvement over the prior process though. Market share analysis has been the subject of considerable debate. For an example of the criticisms levelled, see Robert H. Lande, ‘Market Power Without a Large Market Share: The Role of Imperfect Information and Other “Consumer Protection” Market Failures’, American Antitrust Institute Working Paper No 07-06, 14 March 2007. Available at SSRN <http://ssrn.com/abstract=1103613>.
16 In other words, many lawyers expressed concern that economics, heretofore a tool for implementing the rule of law, had become the very purpose of the rule of law. See L. Vogel, L’économie, serviteur ou maître du droit (Paris: Litec, 2004), at 605.
17 Interdisciplinarity has been the subject of some heavy criticism. See eg A. D’Amato, ‘The Interdisciplinary Turn in Legal Education’, Northwestern University School of Law Public Law and Legal Theory Series No 06-32, at 66:
as to the question of emergence, there is little evidence that economic analysis of law has changed these areas in any innovative way. Indeed, the focus on the quantitative aspects of antitrust—such as in Robert Bork’s reductionism of antitrust to the goal of delivering the lowest prices to the consumer—has had a distorting effect on the field. The original impetus (not the only motive, of course) for antitrust legislation—combating an incipient fascist tendency of huge corporate combinations to overwhelm and run the government—seems to be an inconvenient memory for those who would like economic analysis to quantify everything in dollars.
18 Of course, there are disagreements between the various movements that we discuss in this chapter. See R.J. van den Bergh and P.D. Camesasca, European Competition Law and Economics: A Comparative Perspective (Antwerp/Oxford: Intersentia/Hart Publishing, 2001), at 16, who emphasize the great divergences among economic schools since classical theory. However, competition economics has been nourished by these interactions between movements, schools, and doctrines so that today it is relatively stable and reliable, at least in those applications with a longer history.
19 See P. McNulty, ‘A Note on the History of Perfect Competition’ (1967) 75 J Political Economy 395, who mentions the influence of other writers in the seventeenth century; P. McNulty, ‘Economic Theory and the Meaning of Competition’ (1968) 82 Quarterly J Economics 639.
20 In fact, classical economists such as A. Smith and J. Stuart Mill spoke of competition but never gave a precise definition of the concept. It was not until the British economist S. Jevons, one of the co-founders of the neoclassical movement, that the adjective ‘perfect’ began to be heard. The neoclassical economist Y. Edgeworth was subsequently the first to attempt to define what perfect competition might be. But it was not until 1921 and the book Risk Uncertainty and Profit by F. Knight that economists offered an explicit statement of the five conditions that must be met for a market structure to qualify as perfect competition. J. Boncoeur and H. Thouement, Histoire des idées économiques de Walras aux contemporains, 3rd edn (Paris: Armand Colin, 1921), ch III, at 39–55. V. Pareto subsequently explained why perfect competition is an ideal to be attained.
26 Note that Marshall studied what is known as ‘partial’ equilibrium, ie, he showed how prices are determined on a particular market, not for the economy as a whole. L. Walras studied the much broader concept of general equilibrium by imagining the situation of an auctioneer who by trial and error determines the equilibrium prices on all markets.
27 See Guesnerie, n 24, at 6.
31 When the price is equal to the marginal cost, sellers are nonetheless profitable since they do not make a loss and therefore remain on the market. The important notion to bear in mind here is the difference between ‘accounting profits’ and ‘economic profits’. The marginal cost that economists use includes opportunity costs, or the cost of using resources to produce the good at hand instead of moving those resources to their next best alternative. Hence zero economic profits, as occurs when price equals marginal cost, enables firms to earn positive accounting profits equal to the competitive return for the particular industry or market.
32 See Glais and Laurent, n 28, at 6.
People of the same trade seldom meet together, even for merriment and diversion, but the conversation ends in a conspiracy against the public, or in some contrivance to raise prices (A. Smith, An Enquiry into the Nature and Causes of the Wealth of Nations, 1776, at 127–8).
36 V. Pareto argues that a situation is optimal when it is no longer possible to improve the situation of an economic agent without making that of another (economic agent) worse. This condition is referred to as Pareto Optimality. See V. Pareto, Course of Political Economy (Lausanne: 1896).
37 This triangle is also sometimes referred to as the ‘Harberger triangle’. Its value is equal to (QC−QM) × (PM−PC)/2. See A. Haberger, ‘Monopoly and Resource Allocation’ (1954) 44(2) Am Economic Rev 77. Note that the fact that the holder of a monopoly makes economic profits is not the source of major concern for neoclassical theorists. Such profits are only one of the collateral consequences of monopoly power.
42 The relationship between competition and innovation is one of the most hotly debated issues in economic theory. See P. Aghion, ‘A Primer on Innovation’, Bruegel Policy Brief 04, Issue 2006/06, October 2006.
43 See Guesnerie, n 24, at 78 who points out that the greatest technological leaps in the twentieth century originated in the tight networks of small undertakings in the Silicon Valley.
50 Incidentally, a market structure organized on the basis of atomicity of producers is generally inefficient. In a large number of sectors, the existence of economies of scale results in efficiency gains. Exploiting these economies of scale implies an increase in the size of the undertakings active in the market and a reduction in their number. These scale efficiency gains go hand in hand with substantial improvements in ‘welfare’ for both manufacturers and consumers (reduction in price, improvements in the quality of the products and services, etc). This observation subsequently led to the development of more pragmatic theories, eg the theory of ‘workable competition’. See J. Clark, ‘Towards a Theory of Workable Competition’ (1940) 30 Am Economic Rev 241. The theory of monopolistic competition is another such theory. Proposed by E. Chamberlin in 1933, it is based on four conditions which to some extent echo perfect competition, but differ slightly from it: (i) undertakings sell products that are of the same type but are imperfectly substitutable (differentiated products); (ii) each firm may choose the price of its product; (iii) the number of undertakings that are part of the industry in question is high and each is negligible in relation to all the others; and (iv) there is free entry and exit on the market. In this model each firm sells a specific product the price of which it sets. On its own (narrow) market, it holds a limited monopoly. See E. Chamberlin, The Theory of Monopolistic Competition (Cambridge, MA: Harvard University Press, 1933).
51 Ibid, at 22.
52 While most economists recognized the practical limitations of neoclassical models, some ordo-liberal theorists later argued in favour of making perfect competition, with some adjustments, a policy guideline. See van den Bergh and Camesasca, n 18, at 21.
54 Understood as being the contribution of the commercial activity of the undertakings in the market to general material welfare, ie, not only allocative, productive, and dynamic efficiency but also in terms of distribution of revenue, growth, and employment. See F.M. Scherer and D. Ross, Industrial Market Structure and Economic Performance, 3rd edn (Boston, MA: Houghton Mifflin, 1990), at 4.
57 See the arguments of R. Joliet, Monopolization and Abuse of a Dominant Position—A Comparative Study of the American and European Approaches to the Control of Economic Power (The Hague: Martinus Nijhoff, 1970), at 106.
59 This was a Commission established by the White House to study the concentration of US industry. The Task Force, set up in December 1967, was chaired by Phil C. Neal, Dean of the Faculty of Law at the University of Chicago and comprised three practising lawyers, three economists, and six professors of law (William F. Baxter, Robert H. Bork, Carl H. Fulda, William K. Jones, Dennis G. Lyons, Paul W. Macavoy, James W. McKie, Lee E. Preston, James A. Rahl, George D. Reycraft, Richard E. Sherwood, and S. Paul Posner). See A.A. Foer, ‘Putting the Antitrust Modernization Commission into Perspective’ (2003) 51 Buffalo L Rev 1029, at 1039–41. The proposals for specific legislation were adopted by 11 out of the 13 members of the Task Force.
64 See Y. Brozen, ‘The Antitrust Task Force Deconcentration Recommendation’ (1970) 13 J Law and Economics 279. At the famous Airlie House conference in 1974, critics of the structuralist approach synthesized their arguments and research. The conference activities and subsequent work succeeded in discrediting the recommendations of the structuralist school.
66 See R. Posner, ‘The Chicago School of Antitrust Analysis’ (1979) 127 U Penn L Rev 925, at 944 who establishes the issue of economic concentration and deconcentration policies as the fundamental difference between the Chicago School and the Harvard School.
67 See van den Bergh and Camesasca, n 18, at 4. This methodology is no less capable of being criticized. While it is motivated by empirical observation, the analysis is based on theoretical models, often with restrictive premises far removed from reality.
68 See R.H. Bork, The Antitrust Paradox—A Policy at War with Itself (New York: The Free Press, 1993), at 91; N. Mercuro and S.G. Medema, Economics and the Law—From Posner to Post-Modernism (Princeton, NJ: Princeton University Press, 1997), at 53; see van den Bergh and Camesasca, n 18, at 4. The concept of economic efficiency refers to the optimization of allocative efficiency and productive efficiency. Allocative efficiency exists when productive resources are used for the purposes that consumers value most or, in other words, when all the demand on the market is satisfied. Productive efficiency exists when the productive resources are used optimally by the undertakings, ie, when they reduce their costs in an optimal fashion.
70 See Posner, n 66.
72 A positive causal link can certainly be established between the structure of a market and its performance (for the Harvard School, making a profit). The Chicago School finds that this link, however, does not come from the market power of the operators but, on the contrary, from the greater efficiency of the latter. See H. Demsetz, ibid, which explains the concentration of an industry either (i) by the commercial or productive superiority of an undertaking or (ii) by the greater efficiency of a market structure comprising a limited number of undertakings (eg due to economies of scale). It should be noted that Bain had already mentioned the possibility of concentration of markets being both the cause and the consequence of the efficiency of undertakings. See J. S. Bain, Industrial Organization (London: Wiley, 1959), at 424. Nevertheless, Bain felt that efficiency was only one of several explanations while the Chicago School considers that this is the primary explanation.
73 See Bork, n 68.
74 See Demsetz, n 71, at 9. Deconcentration policies include the risk of promoting the appearance of inefficient market structures. If, in certain cases, they can reduce the risk of collusion it is not certain that this positive effect is sufficient to offset the resulting efficiency losses. Ibid, at 4–5.
75 This doctrine is echoed in Germany, where Professor Hoppmann criticized the structure/performance link and came out in favour of a competition policy targeting behaviour. See van den Bergh and Camesasca, n 18, at 40 and 45.
82 See Carlton and Perloff, n 80, at 315–79; see van den Bergh and Camesasca, n 18, at 60. Intervention is justified where there is ‘welfare loss’. The concept is broader than the concept of efficiency established as the objective of competition policy by the Chicago School. Efficiency is the maximization of the sum of producer and consumer surpluses; thus equal transfers from consumers to producers create a zero net effect on total surplus. ‘Welfare’, however, is most often evaluated in terms of consumer surplus alone, so that gains to producers that have no direct impact on consumer welfare (eg when improvements in the cost of production are not passed on to consumers via lower prices) have a zero net effect on welfare. Second order effects, such as gains to the producer’s shareholders who also function as consumers, are generally not considered.
83 D.E. Waldman and E.J. Jensen, Industrial Organization, Theory and Practice (Reading, MA: Addison-Wesley Longman, 1997), at 245–8. Eg a strategy consisting of reducing prices in order to eliminate competitors, then raising them again later when rivals find entry difficult.
84 See for a full treatment of price limitation practices, Tirole, n 39, para 9.4, at 367ff.
86 See R. Schmalensee, ‘Advertising and Entry Deterrence: An Exploratory Model’ (1983) 91(4) J Political Economy 636; O.P. Heil and A.W. Langvardt, ‘The Interface between Competitive Market Signalling and Antitrust Law’ (1994) 58(3) J Marketing 81.
88 See eg Dennis W. Carlton and M. Waldman, ‘Theories of Tying and Implications for Antitrust’ in W. Dale Collins (ed), Economics of Antitrust (American Bar Association, forthcoming). Hart et al concede a similar point in their seminal paper on vertical integration simply stating that ‘Given these conflicting effects it is hard to deliver clear-cut prescriptions for antitrust policy on vertical mergers’, O. Hart et al, ‘Vertical Integration and Market Foreclosure’, 1990 Brookings Papers on Economic Activity: Microeconomics, 205, 213.
[w]hile [the post-Chicago methodologies] sometimes produce robust economic conclusions, testing them has proven difficult. Further, they are messier and more difficult to use and too often strain the fact-finding power of courts beyond the breaking point. ( H. Hovenkamp, ‘Post-Chicago Antitrust: A Review and Critique’ (2001) Colum Bus L Rev 257, 336 ).
91 For a complete description of these behaviours, see Carlton and Perloff, n 80, at 379–86.
92 Ibid. Clauses such as ‘most favoured nation’ offer a guarantee to any buyer that he is paying the lowest possible price. When its use is widespread in an oligopoly setting, each producer is able more easily to detect if one of its competitors is cheating on the implied cooperative agreement and the incentive of each to deviate from the collusive price decreases since a price reduction becomes very costly (it has to be extended to all customers).
94 According to Professor E. Elhauge, recent judgments of the US Supreme Court apply the precepts of the Harvard School. See E. Elhauge, ‘Harvard, not Chicago: Which Antitrust School Drives Recent Supreme Court Decisions’ (2007) 3 Competition Policy International 59.
95 See W.J. Baumol, J.C. Panzar, and R.D. Willig, Contestable Markets and the Theory of Industrial Structure (New York: Harcourt Brace, 1982). The theory of contestable markets appeared in the early 1980s with the publication of Baumol, Panzar, and Willig’s work. It summarizes and sets out the conclusions of the Chicago School. In short, it finds that concentrated markets are not problematic provided new operators can enter the market when incumbents try to extract supra-competitive profits (ie, the market is contestable). This theory therefore stresses the concepts of exit costs and irrecoverable investments, the effects of which deter potential entrants from entering a market and therefore ease the competitive pressures on the established operators.
96 See N. Petit, Oligopoles, collusion tacite et droit communautaire de la concurrence (Brussels: Bruylant-LGDJ, 2007), ch I. Game theory is a mathematical theory based on the premise that producers are players in a ‘game’ seeking to maximize their gains in light of the anticipated behaviour of others. Players therefore act in a strategic way. This theory allows predictions to be made about the behaviour of operators on the market. Eg it allows one to determine if operators in an oligopoly are going to become involved in parallel price hike strategies or if a price-fixing agreement will be adhered to by all the parties to the agreement.
97 This theory focuses on the ‘principal–agent’ relationship: a ‘principal’ grants authority to an ‘agent’ to conduct a task, the performance of which he cannot completely observe or control. When information is asymmetric, with the agent having greater information, the theory finds that ‘problems of opportunism’ appear. One such problem is ‘moral hazard’: the agent does not comply with the terms of the agreed contract but instead tries to defraud the principal, say by distributing a competitor’s products in addition to the principal’s. In order to limit the problems of opportunism, principals generally must incur two types of expenses: incentive expenses meant to align the agent’s incentives with the principal’s (eg profit sharing) and audit expenses meant to catch and punish the agent for any opportunistic behaviour and thus to limit its occurrence (eg unannounced visits to the point of sale). Agents, in an attempt to limit the principal’s concerns, incur expenses known as agency costs, such as monthly reporting which a distributor agrees to perform. See Tirole, n 39, at 51–5.
98 Transaction cost theory seeks to answer the following question: why do firms not conclude a contract with a specialist third party to produce a good or a service instead of handling diverse production tasks themselves, using internal employees over whom they have hierarchical control? Ronald Coase was the first to answer this question. He explains that to carry out an action, undertakings can choose between (i) the market, ie, they will conclude a contract with another operator (a decentralized mechanism based on prices) and (ii) the firm, ie, they will themselves carry out this activity (an organized mechanism based on authority). Consider an example: to fit its vehicles with tyres a car maker can choose either to buy them from major tyre producers or to become vertically integrated in the tyre manufacturing business. According to Coase, the decisive criterion between both these possibilities is the size of the transaction costs (from concluding a contract) and the internal organization costs (from vertical integration). Transaction costs are traditionally split between (i) research and information costs; (ii) negotiation and decision costs; and (iii) monitoring and control costs. See R. Coase, ‘The Nature of the Firm’ (1937) 4 Economica 386; O. Williamson, Market and Hierarchies: Analysis and Antitrust Implications (New York: The Free Press, 1975).
99 The decisions taken by economic operators regarding pricing and production send signals to the market. An operator can thus make considerable strategic investments (advertising, R&D, etc) in order, eg, to signal to potential entrants that it has major resources to defend its market share in the event of market entry. See O.P. Heil and A.W. Langvardt, ‘The Interface between Competitive Market Signalling and Antitrust Law’ (1994) 58(3) J Marketing 81. Signals are informative when there is imperfect information because they offer indirect information in the absence of direct information. Signal theory has interesting applications in the area of tacit oligopolistic collusion, eg although it originated in the area of labour economics. See A.M. Spence, ‘Job Market Signaling’ (1973) 87(3) Quarterly J Economics 355.
100 In very general terms, an externality, either positive or negative, is defined as the undesired and unintentional after effect on a third party of a specific behaviour, either an individual or several individuals linked by agreements, including those joined in a firm. Eg a car distributor creates a positive externality when he advertises his products. The other distributors in the network carrying the same brands, which are therefore potential competitors, benefit from the car brand names being circulated among potential customers. A steel producer, on the other hand, creates a negative externality when its factories generate pollution affecting, eg nearby agricultural activities.
101 Behavioural economics suggests that agents do not always, as neoclassical theory tends to suggest, make rational decisions (ie, decisions aimed at maximizing utility). On the contrary, in some situations, economic agents adopt behaviours that might appear irrational (contrary to what neoclassical theory would predict). Eg individuals or undertakings may have very high discount rates that result in myopic behaviour. Or, the decisions taken by undertakings may not always be motivated by monetary profit, eg if racial prejudice motivates some decisions. For individual behaviour, lab experiments show the sometimes irrational nature of operators’ decisions. See V.L. Smith, Bargaining and Market Behavior—Essays in Experimental Economics (Cambridge: Cambridge University Press, 2000).
102 The criterion of market power has, since the reform of the regulations applicable to agreements between undertakings, occupied a central place in EU competition law. See Commission Notice of 6 January 2001: Guidelines on the applicability of Article 81 of the EC Treaty to horizontal cooperation agreements, OJ C 3 of 6 January 2001, at 1, para 7: ‘Economic criteria such as the market power of the parties and other factors relating to the market structure, form a key element of the assessment of the market impact likely to be caused by a cooperation and therefore for the assessment under Article 81.’ Similarly, in its guidelines on vertical restrictions, the Commission considers that: ‘For most vertical restraints, competition concerns can only arise if there is insufficient inter-brand competition, i.e. if there is some degree of market power at the level of the supplier or the buyer or at both levels.’ See Commission Notice—Guidelines on Vertical Restraints, OJ C 291 of 13 October 2000, at 1, para 6.
103 See Tirole, n 39, at 284. Conceptually, market power is analogous with the notion of monopoly power, although a firm need not be a monopoly to possess market power.
104 See Motta, n 87, at 116. Such a situation, eg, exists when products are differentiated, exchange costs are large, or customers are loyal to particular brands.
105 See R. Posner, Antitrust Law, 2nd edn (Chicago, IL: University of Chicago Press, 2001), at 265. A product is said to be a commodity once any qualitative difference derived from its natural origin has disappeared. Wheat, chemicals such as sulphuric acid, refined metals, refined oil products, refined sugar, more elaborated manufactured products such as rails or standard electronic components, electricity, pass band, RAM processors, etc are generally viewed as commodities.
106 See D. Geradin, P. Hofer, F. Louis, N. Petit, and M. Walker, ‘The Concept of Dominance’ in D. Geradin (ed), GCLC Research Papers on Article 82 EC (Bruges: Global Competition Law Centre, 2005), at 9.
108 See R. Landes and R. Posner, ‘Market Power in Antitrust Cases’ (1981) 94 Harvard L Rev 937, at 939: ‘the fact of market power must be distinguished from the amount of market power’. See also Geradin et al, n 106, at 9; Motta, n 87, at 116.
109 See Motta, n 87, at 116.
[m]arket power is the ability to maintain prices above competitive levels for a significant period of time or to maintain output in terms of product quantities, product quality and variety or innovation below competitive levels for a significant period of time. In markets with high fixed costs, undertakings must price significantly above their marginal costs of production in order to ensure a competitive return on their investment. (Commission Notice—Guidelines on the application of Article 81(3) of the Treaty, OJ C 101 of 27 April 2004, at 97, para 25).
A firm or group of firms may raise price above the competitive level or prevent it from falling to a lower competitive level by raising its rivals’ costs and thereby causing them to restrain their output (‘exclude competition’). Such allegations are at the bottom of most antitrust cases in which one firm or group of firms is claimed to have harmed competition by foreclosing or excluding its competitors. We denote this power as exclusionary or ‘Bainian’ market power.
115 This form of market power is already explicitly taken into account by a number of competition authorities, including the OFT which states that ‘[a dominant firm] can also use its market power to engage in anticompetitive conduct and exclude or deter competitors from the market’. See ‘The Chapter II Prohibition’, OFT 402, 1999, at para 3.9.
121 Neoclassical theory postulates that economic agents think about the margin by comparing the usefulness of an additional action to its cost. Additional units are produced until the marginal cost of a given unit is exactly equal to the marginal revenue expected from its sale.
123 See Motta, n 87, at 116.
125 See Motta, n 87, at 116.
128 Generally, these are the costs that vary depending on the volume of production. Costs of material and energy used in the production process, eg are variable costs. In the aviation sector, fuel expenses are a variable cost. Economic theory teaches that an undertaking will continue to produce as long as it covers its variable costs. This is why an undertaking that operates cinemas and which therefore has very large fixed costs but very low variable costs will remain open even if it only fills 5 per cent of its cinemas each night (in the short run, anyway—at some point the operator will choose to sell the assets to a higher value user). This logic can be applied to aviation transport as well.
129 See C. Ahlborn, V. Denicolò, D. Geradin, and A.J. Padilla, ‘DG COMP’s Discussion Paper on Article 82: Implications of the Proposed Framework and Antitrust Rules for Dynamically Competitive Industries’, at 13, para 3.1:
Innovative industries tend to have high fixed costs and low marginal costs of production. This is because developing a new, innovative product requires heavy investment, possibly in research and development. However, it may also be because innovative firms often need to invest in a physical or virtual network to create and distribute their products. Once these initial investments are made, the incremental costs of additional units are fairly low, sometimes close to zero.
130 See Landes and Posner, n 108, at 941: ‘[B]ecause marginal cost is a hypothetical construct—the effect on total costs of a small change in output—it is very difficult to determine in practice, especially by the methods of litigation.’
131 See Motta, n 87, at 116. What is more, undertakings can take advantage of their asymmetry of information with the authorities and courts to overestimate the cost data they send competition authorities.
134 See S. Peltzman, ‘The Gains and Losses from Industrial Concentration’ (1977) 20 J Law and Economics 229 which shows that profits in concentrated industries are larger not because of higher prices linked to a market upsurge, but because of the reduced costs of the large undertakings: H. Demsetz, ‘Industry Structure, Market Rivalry, and Public Policy’ (1973) 16 J Law and Economics 1.
135 See Christiansen and Kerber, n 132. A type I error (false positive) appears when the competition authority restricts a market practice or controls a market structure which contributes to economic efficiency.
138 Ibid. But measuring opportunity costs is a very inexact science. Allocating common costs for multiproduct firms is a sensitive issue for the competition authorities and courts.
140 See the definition given by D. Scheffman and M. Coleman, ‘FTC Perspectives on the Use of Econometric Analyses in Antitrust Cases’ available at <http://www.ftc.gov/be/ftcperspectivesoneconometrics.pdf>: ‘natural experiments try to exploit differences in data over space, time, and competitors to shed light on market definition, barriers, and the analysis of potential competitive effects’. See L.-H. Röller, ‘Economic Analysis and Competition Policy Enforcement in Europe’ in P.A.G. van Bergeijk and E. Kloosterhuis (eds), Modelling European Mergers: Theory, Competition and Case Studies (Cheltenham: Edward Elgar, 2005), 17.
142 See Hausman and Sidak, n 139.
146 See Carlton and Perloff, n 80, at 66.
147 See Motta, n 87, at 125.
148 Ibid. The reasoning generally ignores dynamic issues, such as the reaction of competitors who might decide to increase their own prices, and does it account for the introduction of new features or changes in quality?
149 Ibid, esp at 125–8.
151 See Motta, n 87, at 119.
153 The index varies therefore between 10,000 (in the case of a monopolistic market) and 1 (in the case of an atomistic market). See Tirole, n 39, at 221.
155 See Carlton and Perloff, n 80, at 644.
156 Although the question of barriers to entry goes much further back. See in particular D.H. Wallace, ‘Monopolistic Competition and Public Policy’ (1936) 26(1) Am Economic Rev, Supplemental Papers and Proceedings of the 48th Annual Meeting of the American Economic Association, 77.
158 The level of competition already present on the market is particularly indicative of the weight to be given to barriers to entry. OECD, Policy Roundtables, Barriers to Entry, 2005, available at <http://www.oecd.org/dataoecd/43/49/36344429.pdf>.
159 The European Commission placed particular emphasis on this point during the discussions on the modernization of the enforcement of Art 102. See DG COMP Discussion Paper on the application of Article 82 of the Treaty to exclusionary abuses, (2005), at 34–40.
161 Here we will pass over the most radical opinions, such as that of H. Demsetz who considers that the only real barriers to entry originate in the action of government power. See H. Demsetz, ‘Barriers to Entry’ (1982) 72 Am Economic Rev 47.
163 G. Stigler, The Organization of Industry (Chicago, IL: University of Chicago Press, 1968). In his definition, Stigler does not (at least not explicitly) propose taking into consideration the costs which have been incurred by the undertakings present on the market, but only those that they are incurring today or will incur tomorrow. Legal scholars are, however, unanimous in considering that Stigler’s definition should be understood as comprising all costs incurred by undertakings.
164 eg Bain argues that it will be more difficult for new entrants to find funds to increase their production and make investments than the incumbent undertaking. See also on this point, Carlton and Perloff, n 80, at 79. As a general rule, competition and regulatory authorities espouse the Bain approach of economies of scale. The economies of scale achieved by telecommunications operators that have a huge customer base (enabling them to reduce the rate of contribution to the fixed costs of each client) are traditionally perceived to be a barrier to entry by competition and regulatory authorities.
166 See Carlton and Perloff, n 80, at 128.
169 One can also cite the development of customer loyalty programmes aimed at raising switching costs or the engagement of massive investments for R&D (the infamous ‘R&D wars’). See S.C. Salop and D.T. Scheffman, ‘Cost-Raising Strategies’ (1987) 36(1) J Industrial Economics 19; Motta, n 87, at 554, note 59. Another possibility is the renegotiation of supply and distribution contracts. See Carlton and Perloff, n 80, at 298, who take the example of the Alcoa case in the United States (where the undertaking holding the aluminium monopoly had negotiated contracts with electricity generators stipulating that the latter would not supply energy to competing aluminium producers).
173 See D.W. Carlton, ‘Why Barriers to Entry Are Barriers to Understanding’ (2004) 94 Am Economic Rev, Papers and Proceedings, 466, 467. This analysis suggests competition authorities pay particular attention to the market dynamics and the costs of adjustments.
174 See OECD, Policy Roundtables, Barriers to Entry, n 158, at 86. See also J. Baker, ‘The Problem with Baker Hugues and Syufy: On the Role of Entry in Merger Analysis’ (1997) 65 Antitrust Journal 371. See also the guidelines of the Federal Trade Commission and the Department of Justice, 1992 Horizontal Merger Guidelines, 3.3.
178 To be clear, we treat sunk costs as one type of structural barrier to entry. It is possible to consider that some sunk costs are intentionally incurred by incumbent firms in order to create an obstacle to entry and, if need be, to constitute strategic barriers to entry (eg advertising investments in particular). For our purposes, however, treating them as a separate category is not warranted.
185 With regard to the exit costs in the electronic communications sector in France, see the Report on ‘the exit costs’, Mission entrusted to Philippe Nasse by the Minister for Industry (2005), available at <http://lesrapports.ladocumentationfrancaise.fr/BRP/054000619/0000.pdf>.
187 See OECD, Policy Roundtables, Barriers to Entry, n 158, at 74. It would be possible to consider that some barriers described as regulatory barriers in fact originate in the legal strategy of the undertakings already present on the market, so that these should fall within the category of strategic barriers rather than structural barriers.
188 See in particular T. Ross, ‘Sunk Cost and the Entry Decision’ (2004) 4 J Industry, Competition and Trade, Bank papers, 79, 80. These should be distinguished from fixed costs, some of which may be recovered, even if the two concepts can sometimes describe the same cost expended by the undertaking.
190 See Werden, n 186.
191 See eg the Guidelines on the assessment of non-horizontal concentrations under the Council regulation on control of concentrations between undertakings, Official Journal C 265 of 18/10/2008, at 49.
192 Two situations may be distinguished here. First, the undertaking can set its prices at a level lower than its costs for a certain limited period. Such predatory prices may in certain circumstances have crowding out effects and therefore be contrary to the competition rules. Second, an undertaking may set its prices at what is called the ‘price limit’ level—greater than costs, but sufficiently low so as to deter entry—with the aim of discouraging any potential entrant from entering the market. See in particular Tirole, n 39, at 367–75; see also R. Gilbert, ‘Mobility Barriers and the Value of Incumbency’ in R. Schmalensee and R. Willig, Handbook of Industrial Organization (Amsterdam: Elsevier, 1989); P. Milgrom and J. Roberts, ‘Limit Pricing and Entry under Incomplete Information: An Equilibrium Analysis’ (1982) 50(2) Econometrica 443–59. See also OECD, Policy Roundtables, Predatory Foreclosure, 2004, available at <http://www.oecd.org/dataoecd/26/53/34646189.pdf>.
195 See in particular OECD, Policy Roundtables, Bundled and Loyalty Discounts and Rebates, 2008, available at <http://www.oecd.org/dataoecd/41/22/41772877.pdf>.
196 The barriers to entry on the second market may also deter a potential entrant from entering the first market, when the provision of both products is seen as necessary to viably compete in either market. See in particular, B. Nalebuff, ‘Bundling as an Entry Barrier’ (2004) 119 Quarterly J Economics 159.
197 As a result, a potential entrant might not enter a market if entry would not allow it to reach a sufficiently large number of customers. See in particular P. Aghion and P. Bolton, ‘Contracts as Barriers to Entry’ (1987) 77 Am Economic Rev 388.
198 See OECD, Policy Roundtables, Intellectual Property Rights, 2004, available at <http://www.oecd.org/dataoecd/61/48/34306055.pdf>. See also, J. Lerner, ‘Patenting in the Shadow of Competitors’ (1995) J Law and Economics 489–90: ‘Firms with high litigation costs appear less likely to patent in the same [patent technology] subclasses as rivals’.
199 The theory that pricing aligned to marginal cost guarantees an economic optimum has been the subject of scholarly debate. See, in particular, W. Baumol and D.F. Bradford, ‘Optimal Departures from Marginal Cost Pricing’ (1970) 60 Am Economic Rev 265.
200 See US Department of Justice, Competition and Monopoly: Single-firm conduct under Section 2 of the Sherman Act (US Department of Justice: 2008), 62. See also P. Areeda and D. Hovenkamp, Antitrust Law, 2nd edn (Alphen aan den Rijn: Kluwer Law International, 2002), 753b3, at 367.
202 See Areeda, P. and Turner, D.F., ‘Predatory Pricing and Related Issues Under Section 2 of the Sherman Act’ (1975) 88 Harvard Law Review 697; Motta, n 87, at 116; DG COMP Discussion Paper on the application of Article 82 of the Treaty to exclusionary abuses (2005), at para 107.
203 See the case law on predatory prices, in particular CJ, C-62/86 AKZO Chemie BV v Commission, judgment, 3 July 1991, and for a recent application, Court of First Instance, T-340/03 France Télécom SA v Commission, 30 January, 2007 (appeal pending). See DG COMP Discussion Paper, n 202, at paras 127–33. Based on the case law of the Court and in particular the Compagnie Maritime Belge judgment, the Commission considers that a price exceeding average total cost would not be predatory except in exceptional circumstances. See CJ, Cases C-395/96 P and C-396/96 P Compagnie Maritime Belge, judgment, 16 March 2000.
207 This idea has resulted in some competition authorities refusing to incorporate sunk costs in their analysis of the allegedly excessive price policies of dominant undertakings and, by doing so, rejecting the justifications for the pricing based on the high sunk costs incurred by said firms.
208 P. Areeda and H. Hovenkamp thus define this cost as the cost resulting from the production of one unit at a period during which ‘the undertaking does not change its production assets generating fixed costs, such as its factory’; see Areeda and Hovenkamp, n 200, 735b1, at 365 and 735b3, at 367.
210 See in particular ICN, Unilateral Conduct Working Group, Report on Predatory Pricing (2008), available at <http://www.icn-kyoto.org/documents/materials/Unilateral_WG_3.pdf>.
211 This cost is called variable in that it varies depending on the number of final goods produced. It might, however, become a fixed item again if the undertaking had a contract with the outsourcing company providing for a flat fee regardless of the quantity bought.
213 In addition, with regard to predatory prices or rebates, a cost test based on AAC allows the sole focus to be on the costs incurred by the undertaking with regard to specific levels of output or supplying certain customers only. D. Ridyard, ‘Exclusionary Pricing and Price discrimination Abuses under Article 82—An Economic Analysis’ (2002) 23 European Competition L Rev 286, 295.
214 See first, J.A. Ordover and R.D. Willig, ‘An Economic Definition of Predation: Pricing And Product Innovation’ (1981) 81 Yale LJ 8; see also OECD, Policy Roundtables, Predatory Foreclosure, n 192, at 68–9.
215 See in particular W. Baumol, ‘Predation and the Logic of the Average Variable Cost Test’ (1996) 39 J Law and Economics 49, 59 and P. Bolton et al, ‘Predatory Pricing: Strategic Theory and Legal Policy’ (2000) 88 Georgetown LJ 2239, 2250. See also, ICN, Report on Predatory Pricing, n 210; Department of Justice, Singlefirm conduct under Section 2, n 200, at 65–7; DG COMP Discussion Paper, n 202, at 106–12.
217 Recourse to LRAIC as well as its place within cost tests are both still confused however. See, eg OECD, Policy Roundtables, Bundled and Loyalty Discounts and Rebates, European Contribution, 2008 (not yet published). Compare with DG COMP Discussion Paper, n 202, at 123–6; Department of Justice, Single-firm conduct under Section 2, n 200, at 63–4.
218 See Communication from the Commission—Guidance on the Commission’s enforcement priorities in applying Article 82 of the EC Treaty to abusive exclusionary conduct by dominant undertakings, OJ 2009, C 45/7, at para 27.