Seminar 4: rules and standards in antitrust (30 October)

This session looks at rule design in competition law. We will focus on Article 101 and Section 1 of the Sherman Act. if you are unfamiliar with the legal texts, start there. The debate in both jurisdictions is roughly similar but plays out differently. the US statute prohibits all agreements restricting trade. Literally, every contract has some restrictive effect (if I agree to sell you an apple, that apple is not available to others) but the Supreme Court since the early days held that the statute should be construed more narrowly by considering the effects of the agreement on the market.  However, an effects-based approach has its own problems because it forces the plaintiff/prosecutor to engage in costly assessment of restrictive effects. The result has been that the Supreme Court has suggested that some agreements are restrictive of competition ‘per se’ while others are to be reviewed under a ‘rule of reason’.  The judgment below indicates that the Justices are looking for a middle way – consider why this might be so.

In the EU a similar debate existed in the past, when the Court was asked to apply Art 101(1)’s  test of restrictions by object or effect using a ‘rule of reason’.  This was the result of a concern that the Commission (the sole prosecutor) was reading the notion of a restriction of competition too widely.  Now this debate is less pressing, not least because the Commission is taking fewer cases under Article 101 that are not cartels, and with cartels, the anticompetitive nature of the agreement is clear. However, what happens when the welfare effects of an agreement are ambiguous? Must the Commission prove the effects, or can it sometimes find that an agreement is restrictive by object? in the past six years or so, this issue has troubled the courts on numerous occasions. In September, the Court, aided by an able Advocate General, tried to close this discussion in the judgment set out below.

The paper by Colin Diver is one of the most cited papers on rule design; the one by Kerber & Christiansen offers an economic reading of the discussion.

Reading

Christiansen and Kerber ‘Competition policy with optimally differentiated rules instead of per se rules vs. rule of reason’ (2006) 2(2) Journal of Competition Law and Economics 215

Diver ‘The Optimal Precision of Administrative Rules’ 93 Yale L.J. 65 (1983)

California Dental Association v Federal Trade Commission 526 U.S. 756 (1999) (please read both Opinions)

Case C‑67/13 P Groupement des cartes bancaires (CB) v Commission, judgment of 11 September 2014 (please read also the Opinion of the Advocate General)

For Discussion

We can discuss the merits of the cases – what are the competition concerns, what are the countervailing arguments?

More generally from the judgment we can discuss the scope of the rules, in teh EU context the role of the notion of restrictions by object, and in the US the role of the so-called ‘quick look’ rule of reason

Then we can generalise further by considering the role of rule making from the perspective of economics and general administrative law patterns.

 

29 comments on “Seminar 4: rules and standards in antitrust (30 October)

  1. Fabrizio says:

    Both Kerber, his co-author and Diver do not take seriously judges and administrative bodies. In both papers there is a failure to recognize how the law is shaped within its application process in “full blown case-by-case analysis”. (Or at least in analysis where the interested parties are allowed to give any reason they want.)
    In the former work the authors consider the distinction between rules and decision in single cases as a matter of substance rather than one of degree. They do not consider that deciding a specific case requires to motivate the decision and this motivation plays a normative role for future cases. Admittedly, this role changes from system to system. However, even the law and economics literature, for example of settlements, considers decisions public goods in as much as they clarify the content of law.
    In the latter analysis, Diver is too quick in liquidating the ‘rule-exeption regulatory technique’ (see pp. 70-71). The argument is that establishing exemptions is problematic. But this does not seem compelling. For example, Article 101 TFEU adopts a rule-exemption approach. Diver himself observes that the content of a rule can have a “subsequent process of elaborating” (p. 75). Focusing more on the role of exeptions would have lead the author to give more importance to what he calls incongruence.

    • giorgiomonti says:

      I am not sure I follow the critique of Kerber: consider an antitrust case where the judges are asked: is this a per se offence? Now, if the case is one which the judges believe is too complex for per se treatment they have to motivate that decision and say ‘No, it is a case for the rule of reason because…’ However if it is a price fixing agreement and there is no obvious efficiency that results then the judge can simply dismiss as unfounded the argument that the per se rule is inapplicable. At any rate the judge’s duty to motivate seems to be of little consequence to their approach, if you look at the Supreme Court judgment it isn’t too far from motivating the coninuum between per se and rule of reason in the way the authors work.

      I sort of agree with your second point but for a different reason. at pp.70-71 Diver is simply making the preliminary point that devising an exception to a very precise rule leads you to have to think about the cost of devising the exception. He hasn’t yet got to his analyticasl framework at that stage. so the critique is premature. However, my sense is that Diver later seems to argue that a rule can only be one of the several types, so for instance I could not find an example where a regulator would design a rule that has two limbs: one like Model I and one like Model II for example.

  2. Marita says:

    While to my mind Diver’s exposition of the trade-offs facing rule makers and administrators is very clear and perhaps at places even bordering on explaining the obvious, I must say that I find Christiansen and Kerber’s analysis. Maybe it is just a problem with the terminology they use, or perhaps it goes deeper than that. The authors use the definition of differentiation used by Kaplow to define it as the number of complexity of distinctions incorporated in the set of rules. There is nothing controversial in that as such, but then they put per se rules on one side of the differentiation spectrum, on the ground that they require minimum case-specific analysis. However, differentiation can occur at two levels, abstract and case specific. As Christiansen and Kerber observe themselves, theoretical models have undergone a refinement since the early days. This could point to increasing differentiation of those models based on “the renewed recognition of the fact that markets are much more varied and complex” (Hovenkamp). So, even an object focused, rule based approach can be highly differential at the abstract (or general) level, as exemplified by model III in Diver’s article.
    Christiansen and Kerber’s analysis seems to be focusing on developing an approach to the depth of assessment at a case-specific level. However, it is not just an analysis of effects of a particular conduct that might be costly or involve a high degree of differentiation, fitting a conduct into an abstract category (so asking oneself whether the case fits within the rule) might be equally burdensome. It just shifts the analysis into a slightly differently formulated question, grounded at a level of generality. In effect it might be misleading to present the problem as a two-sided one. The three sided basic model presented by Diver seems to be better representative of the various options available to the regulators.

    I find Christiansen and Kerber’s article objectionable on one more ground. My criticism concerns the conclusions they derive from their mathematical model. In particular, the authors appear to be very hasty in reaching their conclusions about the rate of change of certain costs. From there they conclude that a more rule based approach is a superior solution. I consider this conclusion to be premature since it seems to be based on pure guesswork. The conclusions on the superiority of rule based approach might be unjustified, especially when they seem to coincide with underestimating of the importance of the cost of false positives and false negatives in the equation (when error costs have traditionally been a major concern of antitrust law) or considering only one type of error when in fact both under-and over- enforcement is possible. Even if they got the interplay of factors in the balancing exercise right, at most the analysis suggests that the optimum is a question of balance. Difficulty of ascribing values to the factors, makes it impossible to tell how far we are on the differentiation spectrum. This in turn means that it is impossible to tell what the optimum level of differentiation of rules actually is. Moreover, I find if doubtful whether there exists one optimum that fits all circumstances.

  3. Marita says:

    While to my mind Diver’s exposition of the trade-offs facing rule makers and administrators is very clear and perhaps at places even bordering on explaining the obvious, I must say that I find Christiansen and Kerber’s analysis troubling. Maybe it is just a problem with the terminology they use, or perhaps it goes deeper than that. The authors use the definition of differentiation used by Kaplow to define it as the number of complexity of distinctions incorporated in the set of rules. There is nothing controversial in that as such, but then they put per se rules on one side of the differentiation spectrum, on the ground that they require minimum case-specific analysis. However, differentiation can occur at two levels, abstract and case specific. As Christiansen and Kerber observe themselves, theoretical models have undergone a refinement since the early days. This could point to increasing differentiation of those models based on “the renewed recognition of the fact that markets are much more varied and complex” (Hovenkamp). So, even an object focused, rule based approach can be highly differential at the abstract (or general) level, as exemplified by model III in Diver’s article.
    Christiansen and Kerber’s analysis seems to be focusing on developing an approach to the depth of assessment at a case-specific level. However, it is not just an analysis of effects of a particular conduct that might be costly or involve a high degree of differentiation, fitting a conduct into an abstract category (so asking oneself whether the case fits within the rule) might be equally burdensome. It just shifts the analysis into a slightly differently formulated question, grounded at a level of generality. In effect it might be misleading to present the problem as a two-sided one. The three sided basic model presented by Diver seems to be better representative of the various options available to the regulators.

    I find Christiansen and Kerber’s article objectionable on one more ground. My criticism concerns the conclusions they derive from their mathematical model. In particular, the authors appear to be very hasty in reaching their conclusions about the rate of change of certain costs. From there they conclude that a more rule based approach is a superior solution. I consider this conclusion to be premature since it seems to be based on pure guesswork. The conclusions on the superiority of rule based approach might be unjustified, especially when they seem to coincide with underestimating of the importance of the cost of false positives and false negatives in the equation (while error costs have traditionally been a major concern of antitrust law) or considering only one type of error when in fact both under-and over- enforcement is possible. Even if they got the interplay of factors in the balancing exercise right, at most the analysis suggests that the optimum is a question of balance. Difficulty of ascribing values to the factors, makes it impossible to tell how far we are on the differentiation spectrum. This in turn means that it is impossible to tell what the optimum level of differentiation of rules actually is. Moreover, I find if doubtful whether there exists one optimum that fits all circumstances.

    • giorgiomonti says:

      I’m not sure about some of this. I take your point with Hovenkamp’s quote that if knowledge is variable then this makes the search for optimality more complicated, but I don’t agree with your main point. Consider for instance their assessment (p.228) of how to deal with reasale price maintenance. at the time of writing in US antitrust this was per se ilegal. they ask themselves, should we move to a more nuanced approach? and they say first that this depends on the stat of knowledge in economics, then they say: if the per se rule leads to over enforcement (Type II errors), then we need to estimate the costs of that, but take into account that the per se rule is easy to enforce (more on this below). if we move to a more sophisticated analysis we would have less over enforcement (reduce Type II errors) but we need to factor the risk of under enforcement (Type I errors), and the cost of doing more legwork (e.g. considering economic expertise to decide if the conduct merits condemnation). In taking this stance I don’t think they are arguing for a particular solution (rules are better), they simply say that differentiated rules are probably better than always going for a rule of reason.
      I think the problem you note about their assumption of the rate of cost change is a good one, what they might ignore are the high costs of per se rules. That is, just saying that RPM is per se illegal may be too simplistic if people can argue that what they are doing is not RPM, for example prices are just recommended to the downstream retailer.

  4. Jotte says:

    Free riding arises where the investments of an undertaking create externalities that other firms can use to their benefit. For example, a bar requires equipment to serve beer. Brewers can subsidise these investments since they will benefit from the resulting sales. However, if the equipment is used to sell rival brands, the brewer will have less incentive to support these investments. The free riding problem thereby points towards instances where there is a risk of under provisions of investments because of the free riding of competing firms on those investments. In the CB case the free rider problem is as follows:

    Issuer <—— Acquirer

    I I

    Cardholder Merchant

    Here you have the user of the card (cardholder), the bank that issues the card (issuer), the seller or products or services that accepts the card (merchant) and the bank that provides services to the merchant, which accepts the card as payment (acquirer). The acquirer makes investments to offer merchants services such as a payment machine that also allows payments of other cardholders than the one of the acquirer. A ‘normal’ credit card scheme solves this free riding problem through so-called interchange fees. An interchange fee is a wholesale fee paid to the issuer of a card by the acquirer whenever a merchant accepts that credit card for payment. The wholesale fee paid to the issuer internalises the network effects (reflecting the value of the instalment of a user base). Acquirers pass the interchange fee plus a margin to their merchants as a merchant fee calculated as a percentage of the value of the transaction to cover the costs of providing acquiring services. As such the free riding problem is solved because the acquirer gets compensation even if a payment is made with a card from another issuer. Starting from this conventional fee structure in credit card schemes, the payment structure in the CB case appear strangely and suspiciously idiosyncratic. In effect, the payment structure in CB puts a premium on every issuance of a card in order to compensate the investments made by the acquires in the payment services of merchants. That seems like a highly inaccurate measure to address the alleged free riding issue compared to the conventional payment structure that exists with regard to credit card schemes. That is because the premium on the issuance of a card is not necessarily connected to the free riding issue as it is not clear beforehand if and how many times payments are made with that card at merchants of the acquirer. In stead, in the credit card scheme a fee is only paid whenever there is an actual payment made at the acquirers merchant. On first sight therefore, the premiums put on the issuance of cards by banks seem not to pursue the free riding issue at all and instead seem to be solely concerned with penalising attempts to increase market shares by issuers, which has a clear anticompetitive object.

    • giorgiomonti says:

      Interesting perspective, may be worth looking at how the ECJ deal with the more conventional systems in the twin case Mastercard.
      On your point, the question I have is ‘how is this relevant to analyse the CB judgment?’ On the one hand I read you as saying: this looks like a restriction by object: the scheme looks patently unfit for purpose. Is this what you get out of it?

  5. Christopher Johnson says:

    Christiansen and Kerber’s article provides an interesting framework within which to consider the two opinions in the California Dental Association decision.

    The majority in CDA favour a more discretionary, rule of reason based approach. They argue that, although limits on advertisements generally tend to reduce consumer information and distort competition, specific attributes of the dental market make such restrictions procompetitive in nature.

    The dissenting judges adopt a position closer to the “per se illegality” end of the spectrum. They largely oppose the highly sector-specific arguments of the majority, instead likening the restrictions on advertising in the dental market to restrictions on advertising more generally.

    Applying Christiansen and Kerber’s reasoning, the Supreme Court should have aimed to maximise total welfare through minimising the sum of decision errors and regulation costs. In practice this would have required them to consider (amongst other, more difficult to quantify costs) the cost to the FCC of a more detailed investigation versus the cost to the market of overregulation. Without any figures on which to base this, I believe that such an analysis would be likely to support the majority. The cost of a more vigorous investigation may be high, but the likely welfare benefits of the possibly procompetitive advertising policy across such a large market will probably exceed those costs.

    • giorgiomonti says:

      One thing to bear in mind is that the approach of the dissent actually shifts costs. They way: the burden on the FTC is to say simply whether with modest economic analysis the practices in question look illegal. This shifts the burden to the defendant to say: no, there is a justification. as they say: the defendant bears the burden of proving pro-competitive justifications.
      This links nicely to Diver’s point later in the article that an agency like the FTC has little incentive to improve its rule design when the costs are exteralised (in this case, to the courts).

  6. Leticia says:

    Continuing with Marita’s critique of Christiansen and Kerber, I will add some minor point on one of the aspects of the article. I have been impressed by the analytical work of the authors and their design of a model covering all the variables that could be relevant in determining the degree of differentiation of competition rules. However, I would be cautious in including in this model the impact of political influences and rent seeking, especially regarding the author’s statement that beyond a certain degree of differentiation the negative effects of rent seeking outweigh the positive ones.

    Firstly, the authors defend that the lobbies had strong effects on the decision-stage of anti-trust cases, and that ‘per se’ rules offer fewer possibilities for lobbying. I would question the statement that rule-making is less exposed to lobbying; in fact, in Spain there currently is a hot debate on the opposite direction. Some political economists believe that the very detailed Spanish (general) legislation, which leaves scarce room of manoeuvre to courts, has been easier to adjust to certain private interests. This was so because the information costs at the decision-making stage were much higher than what the general population could afford; whereas in the judicial stage the parties involved would have got more incentives to find the relevant information and ‘counterbalance’ the position of the other party. Additionally, I think when rent-seeking reaches the rule-making stage the social costs can be much more damaging than those of particular mistaken decisions, given that rules will apply to an indefinite number of cases. Further research would be needed in order to determine whether, in the EU competition law context, lobbying does indeed affect the investigators more than if affects the legislature.

    Secondly, even assuming that lobbying takes place more strongly in the absence of ‘per se’ rules, it is not obvious to me how this should equate rent seeking. The involvement of the private/political actors affected by the investigation can in fact bring valuable information so as to decide on the anti-competitive object/effects of the actions investigated. This would be particularly so in cases where welfare effects are ambiguous, which are in fact cases where a more ‘differentiated’ approach would be advisable.

    • giorgiomonti says:

      Good points. In response to 1, I suppose this depends on the rules in question. Your example is very detailed legislation, a bit like model III in Diver’s paper.
      In response to 2, you are right: the issue is perhaps to devise procedures that channel information in one way and lobbying in another. For example in commitment decisions you ‘market test’ by securing the views of stakeholders, but it is also not impossible for one to lobby the decision-maker directly outside this process.

  7. Maria H says:

    Comments on California Dental Association v Federal Trade Commission 526 U.S. 756 (1999)

    The Supreme Court accepts that a ‘quick-look’ analysis is appropriate in cases of agreements that are restrictive per se or more specifically “when an observer with even a rudimentary understanding of economics could conclude that the arrangements in question have an anticompetitive effect on customers and markets”. However, where anticompetitive effects of a given restraint are “far from intuitively obvious” these agreements have to be reviewed much more thoroughly under the rule of reason. In the present case, the Court holds that the advertising restrictions could also very well have procompetitive effects or even no effect on competition whatsoever. The Court explains this with the particularities of the market for dentists. The Court makes a very good case that the question of whether these advertising restrictions is not one that has an ‘obvious’ answer or that an observer with a rudimentary understanding of economics could answer. The concern that the Court has in this case seems to be the impreciseness of the ‘quick-look’ analysis, that that analysis might be too ‘quick’, too broad-sweeping to take into account different circumstances in different cases. At the heart of this might lie the Court’s worry of the risk of over-deterrence.

    The dissenting opinion puts the case more into context. It argues that from the FTC’s findings it is clear that the advertising restrictions are anticompetitive in the dentistry market. It denies the importance given by the Court to the ‘obviousness’-requirement; instead, the FTC needs to show that the restrictions had anticompetitive tendencies, that procompetitive justifications were put forward and that the parties have enough market power to make a difference. The minority argues that the Commission correctly evidenced all of these elements and that the Court of Appeals therefore came to an “unexceptional” decision. The dissenting Opinion puts forward a concern that counterposes the one of the Supreme Court: administrative burden. The Court does not even consider this issue, which is a shortcoming of its decision. The dissenting opinion seems more persuasive to me as it refers to more specific findings about the effects of CDA’s restrictions, whereas the Supreme Court considers relatively broad claims about the dentistry market.

    • giorgiomonti says:

      Note that in the first paragraph of the dissent it says that ‘a traditional application of the rule of reason… requires affirming the Commission’ so the 4 step test set out in Para I is the rule of reason, not the quick look, it would seem,

  8. Dental vs. FTC – first opinion focuses on “there can be competing claims about the pro or anti-competitiveness” and as such no abbrevaited test should have been applied as it wasn’t obvious. The dissenting opinion finds that the Commission and the court reviewing its decision had enought evidence to find anti-competitive conduct.
    The dissenting opinion also found it obvious that there is a real effect on the market, on the face of an average of 75 % market share with up to 90 % in some areas. Additionally, it held that

    Regarding the CB judgement – while reading I got the impression that while of course maybe the General Court could have referred to more case law or formulated things differently, it almost seems like the Court is not merely doing a legal review but a review of the facts or a review of how the General Court applied the law to the facts.

    In general, I think there is good reason to have administrative procedures first in many proceedings, but there should still be a case by case analysis by the administration and appeals (to a court) can’t be barred.

    • giorgiomonti says:

      on your point on CB it is also the case that the court erroneously states that it is engaged in full review, while such review is only for fines.
      Is your concern in CB mitigated by the fact that the GC was now asked to re-decide?

  9. Maria says:

    The two papers by Diver and Christiansen and Kerber present an interesting framework from which to consider the design and enforcement of competition law, as discussed in the two decisions. The need for an intermediate approach between per se prohibition rules and rule-of-reason has been discussed both at European and American level. The American and European debate may differ to a certain extent in pace and focus, but it remains substantially the same.
    While per-se prohibition rules are generally considered to result in more transparent and reliable decisions in a less costly process, these very aspects are said to prevent them from taking into account the many nuances and not necessarily anticompetitive effects of behaviours generally classified as infringements, as opposed to an effects-approach in the application of competition law, which on its turn, is considered too costly and somewhat obscure. Next to this dichotomy, more and more voices call for an intermediate approach in the design of rules in order to achieve the “optimal degree of differentiation” that represents the minimum of total costs of the rule.
    The paper by Diver and Christiansen provides a mathematical model for this intermediate approach, which in other words can be said to look for the balance between right degree of generality and specifity of rules –new terminology for old problems-.
    A balance is certainly needed, but I am not so sure whether a purely mathematical approach, however well-intentioned, may be incorporated in the intricate process of rule drafting, or if a more differentiated prohibitive rule wouldn’t result in a more “differentiated” debate, in the sense that the diverging interpretations of the economic effects of a given behaviour become more and more nuanced as we re-calibrate the generality of the rule, without ever arriving at tackling the problem of the costly interpretation and assessment of the circumstances of each specific case. At the end of the day, this intermediate approach merely moves the need for an economic assessment of a given behaviour from the decision-making level (by competition authorities, courts), to the rule-making level.
    This is a complex debate where very different variables and factors play: the generality of rules, the design of exemptions, the costs of enforcement and rule-design, but also the very legitimacy and justification of competition rules. Echoing debates in previous sessions, these readings have make me wonder again an old question: must the design of competition rules only reflect economic efficiency concerns?

    • giorgiomonti says:

      Your main point is not too dissimilar to Marita’s. You are right in saying that they shift the debate to the rule-making level. In fact this is their point in the final part of the paper: there are two economic approaches: one is to apply industrial organisation and work outt he welfare effects of a practice; the other is to apply techniques that economise the legal process. they suggest you need to do both.
      Your last point: well, Diver has a go early on at asking: should we test the optimal precisin of rules by reference to ‘justice’ (e.g. equality) and finds that unhelpful. Maybe there’s more to it than he thinks?

  10. stavros says:

    To my understanding Christiansen and Kerber’s most important contribution is the conceptual framework they offer for thinking about competition rules. In particular, they correctly argue that the role of economics is (a) to determine the optimal degree of differentiation of competition rules in the level of rule designing, and (b) to inform competition analysis, when it comes to specific cases. In this respect, their version of the utilitarian calculus suggests that the optimal precision of rules aims at minimizing the sum of welfare costs caused by decision errors (type I and type II) and regulation costs. Following Divers’ analysis, they argue that the optimal degree of differentiation is reached when marginal reduction of the sum of error costs (the marginal benefit of differentiation) equals the marginal cost of differentiation (the regulation costs) and accept that the optimal precision varies from rule to rule (or category of practices).

    Adopting such a normative standard for designing and interpreting competition rules moves Antitrust away from a deontological principle of justice according to which (a) a certain behavior could be inherently harmful for competition and (b) the perpetrator of such conduct shall be sanctioned. On the contrary the said standard implies that not every anticompetitive practice shall be condemned. Another inference of this approach could be that a competition rule that cannot be administered effectively (i.e. the increase of the regulation costs exceed the reduction of the error costs) should not be applied.

    In addition, such an approach allows for an economically informed categorical thinking in Antitrust: a business practice is not prohibited or allowed per se (formalistic analysis), nor due to its actual or expected welfare effects in a specific case (full-fledged welfare analysis), but on the basis of the average effects of the category it belongs and after taking into account the specific context. Consequently, certain practices will be entitled to a qualified per se illegality, while others to a qualified per se legality, while the evidentiary rules on the burden of proof will permit individual assessments.

    In light of the above, under Christiansen and Kerber’s conceptual framework we are able to compare different variants of structured rules of reason (e.g. Easterbrook, The Limits of Antitrust; Hovenkamp, the Antitrust Enterprise) and avoid formalistic rules or unstructured rules or reason (e.g. EAGCP’s proposal of an unstructured rule of reason standard in An Economic Approach to Article 82, July 2005). Moreover, such an approach calls for an optimal design of antitrust remedies (“tailoring the offense to the remedy”), since an intense remedy would be allowed on the basis of the strength of the authority’s conclusion about anticompetitive effects and its ability to measure them.

    In this regard, Christiansen and Kerber’s conceptual framework could explain the methodological common ground in CDA v FTC that (a) there is no categorical line to be drawn between “obviously anticompetitive restraints” and those that call for more detailed treatment, while (b) the quality of proof required varies with the circumstances. Likewise, EU practice is not far from such methodology: ECJ’s definitions of restrictions by object and restrictions by effect (Allianz and others), article 101(3) defense and article 102 “objective justification” dialectic, as well as ECJ’s annulment of General Court’s decision in C-67/13 on the grounds that the GC failed to establish the sufficient degree of harm for the practice at stake, so as to qualify as a restriction by object (para 69) constitute indications towards such a inference.

    However, Christiansen and Kerber’s basic conclusion in favor of the model of differentiated rules is not so innovative, since a per se rules inducing an irrebuttable presumptions could be deemed incompatible with EU law (Nazzini, The wood began to move: an essay on consumer welfare, evidence and burden of proof in Article 82 cases) and a full-fledge effects-based approach would transform antitrust into an unprincipled area of law. A shortcoming of such an approach could be that it may not take seriously corrective justice in private enforcement (namely, in the area where it is particularly important). In other words, Christiansen and Kerber’s optimally differentiated rule could become troubling in cases of naked restrictions when the victims of an anti-competitive practice are not compensated fully, for deterrent objectives have prevailed in the design and the application of the rules.

    • giorgiomonti says:

      I am going to pick up your second paragraph and your final one.
      In your second paragraph you claim that their approach movs away from a deontological approach. How far is it fair to say that antitrust is deontological? For exmaple it is fairly easy to make the claim that many rules of contract are deontological (promises must be kept); but are there any such deontological rules in antitrust contexts?

      On your final point: you are right in saying that their model does not map out well on the law in the EU – every agreement may benefit from art 101(3) for example. But I don’t get the link to corrective justice in the above.

  11. Elias says:

    Both articles by Diver and by Christiansen/Kerber illustrate the consequences and trade-offs between different degrees of rule precision in terms of efficiency, welfare, legal certainty and rent-seeking. Moreover, they underline the complexity of finding a social and economic optimum balance between over- and under-inclusiveness of rules (Type I and II errors). The California Dental Association case as well as the Groupement Cartes Bancaires case underline that this difficulty is one of the major challenges for competition enforcement.

    The reading of the majority opinion and dissenting opinions in the California Dental Association case suggest that the US Antitrust enforcement relies beyond the dichotomy between per-se rule and rule of reason also on a more differentiated “quick-look” approach advocated by Chrisiansen/Kerber. Thus, the majority in the CDA underlines that there is no fixed categorical line between per-se rule and rule of reason, but rather a sliding scale. However, the CDA case also points out the difficulty of grasping this sliding scale by means of a quick look approach. Thus, the majority rejects a quick-look analysis, whereas Justice Breyer seems to apply rather a per-se rule approach while advocating the quick-look approach.

    Contrary to US Antiturst law, EU competition at the first glance relies in the context of Art. 101 TFEU on a dichotomous approach which divides restrictions of competition under Art. 101 TFEU in a “by-object” box for hard-core restrictions and a “by-effect” box for more complex violations of Art. 101 TFEU. These two categories seem at the first glance to reflect the “per-se” and “rule of reason” approach of US Antitrust. The advantages of the formalistic “by object” approach in terms of predictability, legal certainty, procedural economy, enforcement costs and deterrence have been pointed out by AG Kokott in T-Mobile Netherlands and AG Wahl in para. 35 of his Opinion in Groupement Carte Bancaires. Despite this conceptual dichotomy, the dividing line between by-object and by-effect infringement of Art. 101 TFEU have not been very clear in the past, since the Commission and Courts often have carried out an analysis of the anti-competitive effect of the restrictions in order to corroborate a finding of a by-object infringement of Art. 101 TFEU.

    Therefore, the Court and AG tried to clarify the dividing line between both approaches in Groupement Carte Bancaires. They underlined that a finding of a by object infringement must be based on the analysis of the content of the provisions of the restriction, which must reveal a sufficient degree of harm, its objectives and its economic and legal context (para. 53 of the judgment). By this clarification, the Court excludes an analysis of the anti-competitive effects of the restriction within the by object category. Despite this clarification, the dividing line between both categories continues to be blurred since the Court underlined that the Commission and General Court should have taken into account the economically complex nature of the two-sided credit card markets when they analyzed the Carte Bancaire restriction within the by-object category. In my opinion, this type of analysis differs only insignificantly from an in-depth analysis of the economic effects of a restriction. Instead of establishing a clear cut distinction between “by-object” and “by-effect” infringements of Art. 101 TFEU comparable to the per-se rule and rule of reason categories, the judgment rather suggests that the by-object approach constitutes a certain form of a “quick-look” approach. Given the sliding scale of different competition law infringement and the apparent differentiated application of antitrust standards, it remains unclear how the “quick-look” or “by-object” category could be given a strict interpretation as required by the Court and AG Wahl in Groupement Cartes Bancaires and the majority opinion in California Dental Association.

    • giorgiomonti says:

      Two quick points:
      1) note that in the EU you cannot have a per se ‘box’ because you always have the right as a defendant to plead art 101(3)
      2) you;re right that the CB case moves to quick look (if one wants) and yes then this makes it odd to talk about a strict constructions. But on that point I think the ECJ exaggerates the error of the lower court.

  12. Theodosia says:

    After reviewing the material I would like to focus on the California Dental Association and the Groupment des cartes bancaires cases and examine the relationship between their findings. In particular:
    • In the Groupment des cartes bancaires case the Court examined to what extent the General Court’s assessment regarding the nature of the agreement under examination as restrictive “by object” or “by effect” was correct. The General Court took the view in this case that the types of agreements covered by article 101 TFEU do not constitute an exhaustive list of prohibited collusion and accordingly that the concept of an infringement “by object” should not be given a strict interpretation. The Court stated that the General Court erred in its assessment on this issue. It highlighted that the concept of restriction of competition “by object” can be applied only to certain types of agreements which reveal a sufficient degree of harm to competition. I consider that the Court’s view on this issue is crucial. If the concept of restriction of competition “by object” could be applied to any type of agreement, the Commission would have unlimited discretion to decide which agreement restricts competition “by object”. This may undermine legal certainty and may lead to over deterrence. Above all, if the concept of competition “by object” could be applied to any type of agreement, the Commission might have the incentive to expand the type of agreements which considers by their nature harmful in order to escape from the obligation to prove that a certain agreement restricts competition by effect.

    • Although the distinction between agreements that restrict competition “by object” and “by effect” may be useful for the reasons stated above, sometimes it may lead to undesirable results. In my opinion this type of risk is illustrated in the California Dental Association case. In general the Commission considers pricing or advertisement restrictions as per se violations of competition law. As a result the Commission is exempted from proving their actual effects on the market. This approach may be effective for certain types of markets but not for all. The market of medical services for example is characterized by a strong intellectual aspect, the fact that a high level of knowledge is required to tailor the service to the needs of the client. As a result, the professionals are typically in a far better position than their customers to judge the quality of a service. Due to the asymmetric information issue professionals may be unable to signal differences in their relative quality to consumers. The result is that professionals depressing quality are not punished by the market mechanism. Consequently the market price may reflect only the average quality level. In turn, this may lead to a reduction in the average level of quality perceived by customers. In this regard price or advertisement restrictions may be considered necessary restrictions for the protection of healthcare quality. On the other hand it may also be supported that the elimination of any type of price competition in the market may lead to higher prices and my deprive the healthcare participants from any incentive to invest in innovation. In this respect, in markets characterized by strong asymmetry of information a rule of reason approach may seem more appropriate. Again this may not be an easy exercise for mainly two reasons 1) It entails the use of considerable economic resources 2) quality claims cannot be easily evaluated by Courts. In my view the complexity of the task shows that in critical sectors, such as the healthcare sector, which are characterized by strong asymmetries of information, a sectoral guidance may be necessary. it would exempt the Commission from a case by case complex assessment and it may lead to safer results.

    • giorgiomonti says:

      On your first point: do you actually understand what the phrase ‘reveals a sufficient harm to competition means’? I find it just as unclear as any of the other forumulations.
      On your second – is your argument that while applying the full rule of reason may be beneficial, that there is a role for guidelines that.. do what? help structure how the rule of reason is to be applied?

  13. Marcos says:

    The ECJ seems to want to give better contour to the distinction between the agreements with anti-competitive objects and agreements with anti-competitive effects. The Court tries to do so by restricting the use of the by object approach by both the Commission and the General Court. According to its judgement, the reasons given by both institutions to apply the 101 (1) TFEU using the by object logic were not sufficient to justify their approach to the agreements.
    However, the ECJ is not very successful in establishing a clearer difference between both approaches, since it does not follow the AG opinion. According to the AG, the case law of the Court creates confusion by including in the evaluation of the anti-competitive objects the consideration of their economic and legal context (without this intrusive element, says the AG, the assessment of anti-competitive objects is a pure formalistic approach and, as such, must be limited to some certain cases in which experience shows that the a conduct is constantly harmful and, thus, prohibited). By ignoring these considerations of the AG, the Court seems to reaffirm the ‘quick look’ approach to the reasoning applied to anti-competitive objects. Nonetheless, the confusion between by object and by effect approaches is not necessarily addressed, since the ‘quick look’ and the evaluation of effects seem to be distinguishable only in terms of degree, but the reasoning is essentially the same (and the consequences to the extension of the analysis, justification and burden of proof might not be entirely justified)

    • giorgiomonti says:

      On your main point: perhaps the question is how much evidence is required for the quick look object versus how much for the effects test. then yes it is a matter of degree but that’s fine if you see the object/effect like a continuum. This ocntinuum is found in the Supreme Court and in the Kerber paper.
      On AG Whal – what is the approach he recommends then?

  14. Noguier Alice says:

    The case ‘Groupement des Cartes Bancaires’ and the recent ‘Lundbeck’ decision show a tendency for the Commission to use the ‘by object’ standard of review, even in non cartel 101 cases. In both cases the Commission considered that the agreements at stake were anticompetitive by object, even if she had never before issued cases on this type of agreements that require a complex economic assessment. [Lundbeck is a case on reverse payment settlements in patent litigation (these settlements arise during litigation on the validity of pharmaceutical patents: the pharmaceutical company pay the generic companies to settle the litigation and fix the generic drug’s date of entry). This practice is fairly new, at least in Europe, the Commission and national competition authorities do not have any experience in the field. The assessment of the anticompetitive harm is very difficult and it is not certain that all reverse settlements are restricting competition.]

    Practices constitutive of a restriction by object are usually flagrant infringements of EU competition law. Some agreements due to their known gravity are presumed to infringe competition law ‘by object’, competition authorities are exempted to demonstrate agreements’ anticompetitive effects in contrast to case of infringement ‘by effect’. The burden of proof is directly shifted to defendants who can under 101(3) TFEU demonstrate pro-competitive effects. However, the Commission can review cases under the by object standard, only if she knows by experience that the agreement is anticompetitive. Indeed, the point 21 of the 2004 guidelines on the application of Article 81(3) (now 101(3)) of the Treaty, the Commission states that ‘restrictions of competition by object are those that by their very nature have the potential of restricting competition. These are restrictions which in light of the objectives pursued by the Community competition rules have such a high potential of negative effects on competition that it is unnecessary (…) to demonstrate any actual effects on the market. This presumption is based on the serious nature of the restriction and on experience showing that restrictions of competition by object are likely to produce negative effects on the market (…)’. The Commission gives examples like, price fixing and market sharing cartels that are unanimously considered as detrimental to consumers following numerous studies and decades of cases with demonstrated harmful effects. Such cartels can thus be prosecuted on a by object basis without having to look at their negative effects.
    Moreover, in his opinion in ‘Groupements des Cartes Bancaires’ Advocate General Wahl confirmed that under EU competition law, the ‘by object’ infringement category is reserved for agreement that are by their very nature injurious to competition, and likely to produce negative effects on competition, based on experience.
    This lead to think that the Commission had not sufficient experience of agreement in the CB market and in reverse payment settlements to consider that there are ‘by object’ infringing competition law, like cartels would.
    When intervening is such complex markets, it is really difficult for the Commission to prove the anticompetitive effects of agreements. When the Commission is clearly determined to condemn these agreements, the ‘by object’ standard may be a good means to avoid the ‘effects demonstration’. In some cases the ‘per se’ rule may give to competition authorities and judges more discretion than a ‘rule of reason’.

    However, it should be noted that there is no real per se rule of review in European competition law. Agreements reviewed under the ‘by object standard’ are not considered as strictly illegal since they can be exempt by Article 101(3). The per se American rule of review is more stringent. A practice reviewed by American antitrust courts under the per se rule of review cannot benefit from exemptions. The only purpose of the trial will be to fix the penalties. In contrast under European law, an infringement ‘by object’ can be justified under Article 101(3) TFEU. The defendant will be able to show that there is pro-competitive efficiency justifications that can outweigh the anticompetitive effects of the practice. Nevertheless, some practitioners argue that the possibility of obtaining an exemption under Article 101(3) TFEU for an object infringement is at best extremely unlikely. Four strict cumulative conditions have to be fulfilled. Second, even if exemption under Article 101(3) TFEU were available, the Commission’s approach would shift the burden of proof to the parties, which would need to show that the conditions of Article 101(3) TFEU are satisfied. In other words, the plaintiff does not have to prove anti-competitive effects of the settlement to build a case.

    • giorgiomonti says:

      You are right in suggesting that the Commission has a habit of moving to using the by object category as a shortcut, and that in part the ECJ and the AG may want to avoid this.
      However, as you note, if you have an approach that applies the object criterion, all you do in the EU is shift the burden to the defendant. (In contrast in the US if you apply a per se approach the presumption is irrebuttable). in this context should the question to grapple not be one about where the burden of proof is best placed?
      Your answer to the above woudl appear to me that it should be on the Commission at least in new cases like CB or Lundbeck.

  15. Jonathan says:

    The Dental vs FTC case displays the problems which develop when trying to apply the quick rule analysis or the rule of reason. I believe that the court got the decision right by not applying the abbreviated test because the anti-competitive nature was not obvious. The court stated that the quick rule analysis should be used “when an observer with even a rudimentary understanding of economics could conclude that the arrangements in question have an anticompetitive effect on customers and markets”. However, the dissent also had a good argument as well. The dissent expressed that there was more than enough evidence to show anti-competitive behavior. The problem with the quick approach is that it is too quick, and the problem with the rule of reason approach is that it is too burdensome. It was interesting to read the different possibilities in fixing this problem. I do think that Christiansen and Kerber’s mathematical approach is a nice try to solve this issue, but I do not think this mathematical formula is practical.

    • giorgiomonti says:

      Recall that the dissent seems to have applied a rule of reason.
      Might the answer be to ask on whom the burden of proof should be best placed? so in a case like dentists it woudl be preferable for the defendant to explain the benefits of limiting price competition: the association would have better knowledge than the FTC. On the contrary the FTC has more knowledge of consumer markets in general and is able to assess the veracity of the justifications offered by the dentists.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s