Dernière mise à jour : 17 avr. 2021
Perhaps you have seen the last European Commission's draft (dated on December, the 14th) , which circulated a report to the Parliament on the transposition of Directive 2014/104/EU into the laws of the Member States? [source]
This directive, which hit the headlines in France when it was transposed into national law in the spring of 2017, defines practices, rules and methods for claiming compensation for damage suffered as a result of anti-competitive practices (hereafter "private enforcement actions") such as price and volume cartels (cartel practices, boycotts, etc.), or abuses of a dominant position.
The proven objective of the directive was to facilitate private enforcement actions, which were previously underdeveloped in Europe, even though the Union had made them one of the cornerstones of its antitrust policy (cf. Articles 101 and 102 of the Treaty on the Functioning of the European Union).
According to the state of play drawn up by the Commission, the directive has indeed had the desired effect, with private enforcement actions increasing from 50 cases at the beginning of 2014 (since the adoption of Articles 101 and 102 in 2001) to 239 cases in 2019.
How can this increase be understood? Have private enforcement actions really been simplified or is it a fad? Why were these actions so complex? And do these changes now allow victims to obtain substantial reparations for their harm?
Elements of answers from the point of view of the financial expert in charge of quantifying.
A insightful case : Orange c. Digicel
Let's start with the last question: on 17 June 2020, the Paris Court of Appeal validated Digicel's €250 million claim for damages, victims of anti-competitive practices (market foreclosure) by Orange in Martinique and French Guiana.
This rather impressive sum is commensurate with the sanctions that the Competition Authority can pronounce in public enforcement proceedings (which have greatly increased since the creation of the Authority in 2008).
This rebalancing completes the system for fighting anti-competitive practices, particularly in that it allows the guilty parties, beneficiaries of the Authority's leniency policy, to be brought to justice and sets up a two-bladed sword - public sanction and private reparation - that is particularly dissuasive.
Let us digress for a moment: the rise of private enforcement enshrines competition law as a bulwark for individual freedoms within the meaning of Article 16 of the Declaration of the Rights of Man and of the Citizen: "Any society in which the guarantee of rights is not assured, nor the separation of powers determined, has no constitution. »
However useful it may be, the logic of sanctioning anti-competitive practices (public enforcement) meets a criterion of damage to the economy (cf. Article L464-2 of the Commercial Code), dependent on ideals such as the Pareto optimum or pure and perfect competition.
These ideals are questionable: to remain in a liberal vein, the Austrian school, Murray Rothbard in particular [more details here] , integrates cartel practices and monopolies into a dynamic understanding of the economy (monopolies and cartels are only transitory states, intrinsically unstable and necessary to economic life).
On the other hand, the infringement of anti-competitive practices on the freedom of enterprise - the founding value of our society - is as serious as it is indisputable. This philosophy of private enforcement is important in that it is indicative of a fundamental trend in the modernization of economic law to the advantage and support of entrepreneurs.
For the time being, let's take a closer look at French law: the transposition of Directive 2014/14, adopted by the European Parliament and the Council on 26 November 2014, into French law in March 2017, entails :
- the adjustment of the burden of proof thanks to three presumptions ;
- the definition of a general framework for the assesment of harm related to anti-competitive practices.
The three presumptions
First (irrebuttable) presumption: an entity is deemed to have committed cartel practices when it no longer has an ordinary means of recourse against a challenge by decision of the Competition Authority or any other competent court at the level of the European Union (Article L481-2 of the French Commercial Code). Second presumption: the agreement between competitors is presumed to cause harm (article L481-7 of the French Commercial Code). Third presumption: the victim of cartel practices is presumed not to have passed on the additional costs incurred to his or her own customers (article L481-4 of the French Commercial Code).
The first two presumptions clearly refer to the concepts of fault and causality, which constitute two of the three elements of proof to be provided when making a claim for compensation. These presumptions are essentially intended to simplify follow on actions (claims for compensation based on a prior sanction from the Authority), such as the one conducted in the Orange-Digicel case mentioned above.
Recently, in the lift cartel case, the European Court of Justice extended this logic by confirming that persons who are neither suppliers nor buyers on a market affected by a cartel must be able to claim compensation for the damage they may have suffered as a result of the cartel if a causal link is reported. source link] [source link]
The third presumption naturally concerns the last piece of evidence, namely the damage. It allows, on the one hand, the burden of proof relating to the part of the damage which is passed on to the defendant (the repercussions are a reduction of the damage suffered) so that the entire judicial effort required does not rest on one of the parties; and, on the other hand, a better circulation of information between the two parties (also reinforced by an article concerning the modalities of production of documents - Art. L483-1).
From a more pragmatic point of view, this presumption fleshes out the tactical options of the initiator of the action, who can thus adapt his estimates according to the level of information in his possession and the overall judicial strategy pursued.
Finally, it should be noted, in the alternative, that the reform provides for a further lightening of the work previously required of the plaintiff, namely the distribution of compensation for damages claimed between the various parties at fault in the event of cartel practices.
From the paradigm of quantification of anti-competitive practices
This second set of measures outlines the general framework for assessing the harm suffered as a result of anti-competitive practices (Article L481-3 of the Commercial Code). It is essentially based on the main lines of the detailed technical note produced by the European Commission on the valuation of the impact of additional costs in 2016. [source]
(i) determination of the additional cost,
(ii) study of the repercussion of the extra cost (pass on),
(iii) study of the loss on sales induced by a possible repercussion.
Consistent with the third presumption, this approach makes it possible to phase the assessment of the loss in two or three parts and, possibly, to leave it to the defendant to prove the repercussion.
The plaintiff remains free to take the opposite tack and deal with the three sequences itself, either through successive work or through a simultaneous - so-called holistic - approach, generally based on economic modelling.
Like the presumptions, the spread of these methods of quantification weighed in the Orange-Digicel case. In particular, the Paris Court of Appeal validated the conformity of the method of calculating the loss (by spatial and temporal comparison - or double difference) with the methods recommended by the Commission in its costing guide.
However, it should be noted that the Court continues to use the traditional categories of law (loss of profit and loss incurred) to qualify the quantification carried out, proof that the financial expert has every interest in making an effort of navigation and pedagogy between his valuation work and the usual categories of qualification of the loss.
The fact remains that, even with the effective weapons forged by the transposition of the Directive, assessing a cartel loss is not an easy task.
II) The hyperbolic case of cartel practices
The complexity of the cartel
The Orange-Digicel case, on which we have based ourselves at this stage, involves a complex practice of abuse of a dominant position (foreclosure of the mobile telephony market in the West Indies and French Guiana) by several Orange subsidiaries, but not cartel practices on the scale of an entire market.
Beyond Épinal's image of a small private individual facing a coalition of multinationals, private enforcement procedures relating to cartels remain cumbersome, due to the very nature of cartels. Each cartel is indeed unique, due to :
Its context: Every cartel is the realisation of a market opportunity and as such is part of a particular socio-economic context. The development of the cartel of rough diamond producers around the Beers multinational in 1925 cannot, for example, be seen independently of British imperialism at the end of the 19th century or the developments in the South African market at the beginning of the 20th century.
Moreover, even if two cartels pursue the same objectives, they evolve in very different ways depending on their context. Thus, the OPEC concertations did not have the same effect at all during the first oil shock of 1971 as during the oil counter-shock of 2009. In both cases, the impact of the cartel is irreducible from its geo-economic context (end of the Bretton Woods agreements and the Yom Kippur war for the former, global recession following the Subprime crisis, and development of shale gas exploitation for the latter).
The nature of its practices: Depending on the structure of the market under consideration, the agreements alternately aim at general objectives of rent, predation or eviction. However, these strategies result in very different respective practices: market sharing, imposing high tariffs on the common customer, maintaining an eviction price against the common competitor, boycott are some examples.
the victim's sensitivity to the practices: identical cartel practices can have very different effects depending on the victim's dependence on the market impacted, his potential for repercussions, his development strategies or his negotiating power.
This characterization effort concerns not only the upstream market, the focus of the cartel, but also the downstream market in order to decide on a possible repercussion of the cartel (or vice versa: it is possible for the victim of repercussions to launch an action that allows him or her to climb to the epicenter of the cartel).
Let's go to the simplest: let's consider a case benefiting from all three presumptions, located in a market with few problems, and devoid of any notion of loss of opportunity (which is nevertheless fundamental if the project in question is cancelled due to the sharp increase in the acquisition cost of its main component under the influence of the cartel, or if the cartel was set up precisely to oust the competition).
In this very case, nine phases of work remain essential for the costing, namely :
(i) Determining the scope of the cartel's impact;
(ii) Analysis of the impact on the company's profitability;
(iii) Assessing the impact of the additional costs related to the cartel's practices, i.e. modelling a counterfactual upstream market;
(iv) The evaluation of its own position within this counterfactual market;
(v) The estimation of the pass-on of the additional costs to its own consumers;
(vi) Modelling a counterfactual downstream market;
(vii) The study of a possible contraction of the volumes of this market (estimation of lost sales on the market);
(viii) Correcting these lost sales for the counterfactual evolution of the firm's competitive position (e.g. entry of new competitors in the counterfactual scenario...);
(ix) The quantification of an overall loss, its capitalization at the appropriate rate and its relevant distribution among cartels.
While the third presumption places the burden of steps v) and vi) of the quantification on the defendant, it is nevertheless essential to lend oneself, at least roughly, to the exercise (if only to assess the financial interest of launching a private enforcement procedure).
In this article, we will insist on the two phases of analysis, which are particularly interesting to determine the extent of the prejudice. The other phases are purely technical quantification phases for which we will limit ourselves to providing a few tools.
The analysis phases: establishing the extent of the injury
In addition to the points mentioned in the previous paragraph (context of the cartel, nature of the practices, sensitivity to the practices), these phases of analysis focus on the extent of cartel practices, which can be usefully distinguished by their impacts.
The direct impact of the practices concerns the margin level of the victim of the practices, i.e. the overcharges (reduced by the repercussions and increased by the loss of volumes linked to the repercussions) over the period of injury. This is the core of the injury (possibly increased by an inertia effect).
The indirect impact of the practices includes all other harmful effects, in particular, (i) loss of opportunity due to crowding out or abandonment of a project, and (ii) umbrella effects.
This impact analysis exercise presupposes a multidisciplinary approach: not all these impacts appear with the same importance to the lawyer, economist, strategist or financier, who see all the cartel under a different prism. The lawyer sees them as an infringement of free competition, the economist as a distortion of the market, the strategist as a change in the competitive balance of power, the financier as a change in the distribution of value.
The core of the damage
The core of a prejudice caused by cartel practices corresponds to the overcharges (the additional supply costs, excluding pass-through, and increased by the lost gain in terms of the volume of sales lost as a result of the pass-through) over the period of the practices. In a follow on procedure, this is the period of the practices sanctioned by the Competition Authority.
This core of damage, which is modelled on the concept of damage to the economy used by the Competition Authority, is classically schematized as follows (the version reproduced below is that of the European Commission's guide cited above):
Area A: Overcharges
Area B: Gain due to price effects (pass-on)
Area C: Amount of sales lost as a result of impacts
Area D: Compensation area considering pass-on and lost sales (= Damage)
Damage = Overcharges - pass-on + Lost sales
= Area A - Area B + Area C
However, a critical analysis, and possibly a modification of this scope is essential because the Competition Authority's perspective is that of sanctioning proven cartel practices from a legal standpoint, and not the more economic perspective of assessing the impacts of said practices.
In concrete terms, the start and end dates adopted by the Competition Authority are based exclusively on irrefutable evidence (generally material: meeting notes, e-mail exchanges, logistical evidence, etc.): the cartel thus begins with the first proven meeting, only concerns the participants in formal meetings and ends with the last meeting conducted before dismantling.
A nuance, often anecdotal, to be brought to the work of the Competition Authority concerns possible cartel practices prior to the first proven meeting. The analysis of certain financial data (price increases, deterioration of margins) may in particular suggest the existence of such practices, although it is difficult to establish them.
From a conceptual point of view, the transition from the perimeter of sanctioned practices ("the core of the harm") to the perimeter of the real impact of the practices can be apprehended, for the most part, by analyzing the inertia and umbrella effects.
The inertia effect
Let's put on our economists' glasses: if the cartel is indeed a market distortion, to what extent does its dismantling mark a return to normal market equilibrium? The effect of inertia lies entirely in this measure.
At the very least, the return to equilibrium is subject to the time it takes for the market to integrate the information. This time, depending on market rigidities, can be substantial for sectors in which the negotiation cycles (for a price agreement), or the readjustment of the productive apparatus (for an agreement on quantities) are long.
However, a market parasitized by a cartel is not necessarily unstable. The De Beers company thus organised the world diamond market during the great majority of the 20th century (1925 - 1995) and the price of this product did not collapse (diamonds are certainly a safe haven) following De Beers' announcement to move towards a competitive model (in 2000).
It is therefore perfectly conceivable that the equilibrium resulting from the cartel practices will survive the dismantling of the cartel. If necessary, the readjustment of the market to a normative level would be provoked by exogenous factors (change in regulations, market penetration by new competition, breakthrough innovations...).
The umbrella effect
Let's put on our strategist's glasses from now on: the appearance of a cartel can be likened to the shift from a relatively competitive market to a relatively monopolistic one.
In other words, cartels tend to become price makers. In this configuration, it is in the interest of external companies (direct competitors and - to a lesser extent - producers of substitute or complementary products) to align themselves with the prices charged by the cartellists.
Still in this strategic perspective, the umbrella effect must also be considered in terms of possible repercussions.
As a windfall effect, out of a desire to preserve the coherence of its price scale, or out of a concern to limit the effects of substitution towards products on which it makes a lower margin, a company is likely to use the additional costs incurred as a lever to increase the prices of its entire range.
However, the intensity of the umbrella effect is closely linked to competitive relations and the negotiating power of each of the players in the market impacted by the umbrella effect
Loss of opportunity
Loss of opportunity may be added to the effects described above.
These losses are highly variable, and may represent the bulk of the damage (in the case of predatory pricing) as well as a subsidy (in the case of project abandonment following a cartelization of the market for its main component).
In any case, the diagram presented above is more concerned with a missed gain (lost sales: area C) than with a loss of opportunity, which, moreover, does not have as mathematical a relationship with the repercussions as the latter.
The area of effect of cartel practices may therefore be significantly larger than the core of the damage suggested by the legal elements. It remains to restrict this area of effect of cartel practices according to its impact on the various cases of species.
Here again, the impacts suffered may vary greatly, depending on their alternative supply, their margin of repercussion of additional costs, their non-price competitiveness or their bargaining power.
In this exercise of precision, a good knowledge of the company's processes is therefore essential, but the financial approach, in particular, provides an appreciable overview. Ultimately, the impacts noted above all translate into either a variation in margin levels or a decrease in sales volume.
In the case of a significant umbrella effect, the rates of increase in supply costs evolve more or less indifferently depending on whether or not the references impacted by the cartel are considered.
In the case of a significant inertia effect, the rate of price increases continues to accelerate after the date of dismantling of the cartel .
In this sense, the return to financial data makes it possible to ensure the robustness of the analysis carried out upstream and provides an initial indication of the nature of the cartel, as well as the importance of the various sequences (additional costs, repercussions, lost sales) and related effects (inertia, umbrella, substitution).
Let us now put on our strategist's glasses: the appearance of a cartel is comparable to the shift from a relatively competitive market to a relatively monopolistic one.
In other words, cartelists tend to become price makers. In this configuration, it is in the interest of external companies (direct competitors and - to a lesser extent - producers of substitute or complementary products) to align themselves with the prices charged by the cartels or to pass them on.
III) Introduction to the main tools of quantification
Once the extent of the damage has been rigorously determined, the expert has several tools at his disposal to help him calculate the cost.
The first version of the Guidelines published by the European Commission in 2016 (see above) listed a number of econometric, statistical and financial tools that practitioners could use for their quantification.
The 2019 version of the Guidelines [source], a supplement more specifically dedicated to passing-on, however, reflects a slight change in the European Commission's approach, which now favours so-called "comparative" methods.
There are three comparative methods:
- the before-during-after approach, which consists of analyzing the market under consideration before, during and after the practices under consideration. Relevant in certain cases, this method presupposes, however, that the entire market distortion observed during the practice periods is attributable to the practices. In the absence of a comparison with an external market, this is a questionable assertion.
- the spatial comparison (cross-sectional approach) which consists of the analysis of two distinct but similar geographic markets over the practice periods; one of these markets being affected by cartel practices, the other not. Once again, this is a questionable assertion provided that the opposing party highlights other differences between the markets under consideration.
- the double difference (difference-in-differences approach), which consists in comparing the respective temporal differences presented by two similar markets (one being affected by the periods of practice, the other not). This is a highly appreciated method in that it allows for the control of parameters foreign to the practices on the basis of empirical data.
The Commission points out that these comparative approaches must be carried out separately from data previously adjusted for parasitic factors (e.g. tension on raw material markets which would lead to an increase in the price of the product).
These reprocessing may include linear regressions or index analysis.
While these methods tend to be favored by judges, they involve considerable and sometimes unsuccessful additional data collection and processing work.
Since the development of big data, there are more and more IT solutions capable of processing and retrieving a large volume of information (IDEA, Qlik Sense, ACL Analytics for example), which is no longer the prerogative of specialized players.
However, the exploitation of empirical data has many practical limits: exhaustiveness and homogeneity of the data in the case of the constitution of a synthetic index, representativeness of the data in the case of reasoning by inference, comparability of the data in the case of reasoning by double difference...
Pratically speaking, empirical data are rarely exhaustive because the information needed is usually fragmented, strategic and possibly archived.
The representativeness of the available samples is often undermined by the heterogeneity of the cartel's impacts according to the ranges or products affected (the impact of the cartel depends on downstream customer demand, as well as the sensitivity and reaction of competitors to the cartel, at the product level).
Comparability, for its part, is disqualified provided that the market is specific, that numerous exogenous effects impact the periods under consideration (economic, change in the legal framework, disruptive innovation disrupting the market, etc.) or that the importance of the cartelized companies (cartel and umbrella effect) weighs too heavily on the market.
Moreover, the classic form of comparison by double difference (low prices before, high during and low after) ignored defensive cartels, aiming at maintaining a pre-existing situation of domination (formation of a cartel to erect a barrier to market entry or to counter legal measures aimed at lowering prices in a market).
More fundamentally, the direct exploitation of microeconomic data loses its relevance in the exercise of constructing a counterfactual scenario (of supply or sales) which is indispensable in estimating the damage suffered.
Indeed, they transcribe the situation of society in a very specific context, and therefore have no real value for carrying out market reconstruction.
Data tracing the value/volume evolution of supply over the period of effect of the practices are obviously crucial.
Sales data, while useful for estimating a possible impact, are more dispensable. They are affected by stock variations (strictly speaking, the impact must be estimated on the resale of quantities purchased at the time of the events, and not on the resale of quantities sold).
The other methods cited by the Commission are essentially modelling methods (simulation approaches) based on econometric analyses (notably of elasticity) or behavioural economics theory (industrial).
In the light of its draft Guidelines on passing-on, the Commission now seems to fully consider them as fallback methods: in particular to make up for the lack of economic and financial data.
In fact, dozens of models exist, some of them relying heavily on empirical data and others being almost exclusively theoretical. As this article is not intended to be exhaustive, we propose, by way of illustration, to briefly present the most classical / characteristic of them.
They are models developed from general paradigms of industrial economics.
From the 1930s to the 1970s, the Harvard School developed the Structure-Behavior-Performance (SCP) paradigm, mainly on the basis of empirical work.
Models within this paradigm are based on a structuring hypothesis of causality: Market Structure determines the Behaviors of firms, which in turn determine Market Performance.
In other words, these models are based primarily on market assumptions. This paradigm was heavily criticized in the 1970s on the grounds of theoretical weakness, but it has nevertheless been refined thanks to the development of econometric techniques and the consideration of the reciprocal influence of the behavior of actors on market structure.
This approach still allows a modelling of cartel practices essentially based on market data. The model developed by Verboven and Van Dijk [source] precisely for the purpose of legal assessments, for example, is in line with this perspective.
Since the late 1970s, however, the initiative in industrial economics has passed to the Chicago school, which has developed its own approach based on game theory.
Unlike its Harvard counterpart, the Chicago school relies heavily on neoclassical economic theory (Austrian school of the analysis of market dynamics rather than equilibrium, duopoly of Cournot, Bertrand, Hoteling, Nash equilibrium...). For the most part, the models resulting from this paradigm are based on game theory and the notions of collusion and deviation equilibrium.
From this point of view, the cartel leads to a collusion equilibrium, which is sustainable as long as one of the cartelists does not deviate from the agreement; as soon as there is deviation, all the cartelists redirect their decisions towards non-compliance with the cartel (this is a Nash equilibrium: the cartelists are all winners if they do not deviate, but it is individually more rational to deviate in order to benefit from the market impact of the cartel while not being constrained by it).
In this sense, the cartel's impacts are determined by the stability of the collusion balance. Since the end of the 1980s, a number of specific models have been grafted onto this basis in order to develop specific hypotheses (tacit agreements, factors increasing collusion, nature of competition, products or demand, information framework, barriers to entry, impact of information asymmetries, etc.).
Whether the models are rather related to the Harvard or Chicago schools, they remain limited both by their underlying assumptions and by their meso-economic rather than microeconomic scope.
They do, however, have the three main advantages of reasoning ceteris paribus (isolating the cartel effect from other market effects is relatively difficult using empirical methods), of being rooted in the economic tradition from which the notion of competition derives, and of allowing for data economy, which is sometimes unavoidable.
IV) To conclude
European Commission was not mistaken: the increase in the number of private enforcement actions, as well as the refinement of modelling techniques visible between the two published guidelines, testify to the successful transposition of the directive.
They also show that anti-competitive practices are still very topical and that an impressive number of damages have unfortunately not yet been repaired. Although these procedures are still long (2 to 4 years), they are now more accessible and open the door to very significant damages being repaired.
In terms of figures and law, however, this is a specific practice which is subject to very frequent changes (at the instigation of the Commission in particular) and which benefits greatly from being dealt with by specialists in these matters.
At present, while the Commission is in favour of the twofold difference, the approaches adopted are often hybrid; partly based on raw statistical analyses, partly on more synthetic econometric extensions, and partly on the integration of market elements in an economic perspective (without, however, systematising the approach).
It should be recalled in this respect that the European guides only propose possible approaches and useful tools for the estimation of repairs: the aim is to facilitate the expert's analytical work, not to replace it.