The Post Danmark II judgment: effects analysis in abuse of dominance cases
The facts of the case are old: rebates offered in 2007–08 by Post Danmark, the Danish postal incumbent, to its direct-mail customers, contributing to the exit of its only rival, Bring Citymail. Yet the principles set out by the ECJ in Post Danmark II have important implications for abuse of dominance cases in future. This case adds to other recent judgments—including Post Danmark I (2012) and Intel (2014)—in clarifying EU case law on exclusionary practices under Article 102 TFEU. There have been questions about the extent to which the criteria applied by EU courts differ from those proposed in the European Commission’s 2008 Guidance on Article 102, which sought to reform the approach to abuse of dominance. The policy debates about Article 102 are often couched in terms of effects-based versus form-based approaches. Which way does Post Danmark II point?
Form versus effect: the different philosophies
An old guiding principle in US antitrust law is that ‘[t]he successful competitor, having been urged to compete, should not be turned upon when he wins.’ In Europe, the presence of dominant companies has traditionally been seen as weakening the competitive process and reducing the economic freedom of other market participants. Dominant companies are like the proverbial bull in a china shop; they must be restrained to prevent further damage to their already fragile surroundings. As established in Michelin I (1983), a dominant company has a ‘special responsibility not to allow its conduct to impair genuine undistorted competition on the common market’.
This negative view of dominant companies is somewhat outdated. Economic theory and practical experience have shown that competitive dynamics can function well even in markets with some very large suppliers. Temporary positions of market power can improve competitive dynamics and provide incentives to innovate.
EU case law on abuse of dominance has historically followed a two-step approach: first determine dominance, then assess the form or nature of the conduct. Once a company was found dominant, its ‘special responsibility’ not to impair competition meant that it could not engage in certain forms of behaviour, such as pricing below variable cost, tying products, or offering loyalty rebates. Little consideration was given to the likely effects of these practices on competition and consumer welfare in a given case.
In 2008, however, the Commission’s Guidance endorsed an approach that focuses on examining the economic effects of practices. It builds on the principle that competition law should protect the competitive process in the market, not individual competitors. Aggressive commercial behaviour by dominant companies should not be ruled illegal merely because it makes life difficult for competitors. Rather, the analysis should focus on the effects on competition and consumers. If a particular business practice is unlikely to foreclose competition in a significant part of the market, or if it generates efficiencies that benefit consumers, intervention may not be required, even if individual competitors are harmed by the practice. In this regard, US case law requires there to be proof of a ‘dangerous probability of success’ in monopolising the market. It is not sufficient to provide evidence of the form and the intent of the exclusionary conduct.
Post Danmark I and Intel: opposite directions
The ECJ seemed to give the Commission’s reform efforts a boost when it referred to the importance of effects analysis in Post Danmark I in 2012. Like Post Danmark II, this ruling concerned a disputed determination by the Danish competition authority that the postal incumbent had abused its dominant position by offering targeted discounts to specific customers. There was evidence that these discounted prices were discriminatory and below average total cost. Yet the ECJ held that this in itself was not sufficient to establish an abuse:
Thus, not every exclusionary effect is necessarily detrimental to competition…Competition on the merits may, by definition, lead to the departure from the market or the marginalisation of competitors that are less efficient and so less attractive to consumers from the point of view of, among other things, price, choice, quality or innovation.
Article 82 EC [now Article 102 TFEU] must be interpreted as meaning that a policy by which a dominant undertaking charges low prices to certain major customers of a competitor may not be considered to amount to an exclusionary abuse merely because the price that undertaking charges one of those customers is lower than the average total costs attributed to the activity concerned, but higher than the average incremental costs pertaining to that activity…In order to assess the existence of anti-competitive effects in circumstances such as those of that case, it is necessary to consider whether that pricing policy, without objective justification, produces an actual or likely exclusionary effect, to the detriment of competition and, thereby, of consumers’ interests.
However, the Article 102 pendulum has not swung as far towards an effects-based approach as the Commission might have wished. Indeed, it received a strong pull in the other direction in the General Court’s Intel judgment in 2014. Intel had long been the dominant chip-maker, with more than 70% of the global market. In its 2009 decision, the Commission carried out a detailed assessment of the effects of Intel’s rebates to major computer manufacturers. In line with its 2008 Guidance, the Commission assessed whether an as-efficient competitor would be able to match these rebates, and concluded that it would not. The Commission also followed the more form-based case law that condemns rebates that are conditional on exclusivity. Thus, both the effects and form of the rebates pointed to the same conclusion in the Commission’s assessment in this case. However, the General Court held that the Commission’s analysis of effects was unnecessary:
The question whether an exclusivity rebate can be categorized as abusive does not depend on an analysis of the circumstances of the case aimed at establishing a potential foreclosure effect.
The court held that ‘exclusivity rebates granted by an undertaking in a dominant position are by their very nature capable of restricting competition.’ It does not matter whether these rebates result in foreclosure of a significant or only a small part of the market. Nor does it matter that competitors are free to sell to other customers, since they ‘must be able to compete on the merits for the entire market and not just for a part of it’. The court also stated that even if as-efficient competitors are able to match the rebates, there could still be foreclosure effects, since access to the market is made more difficult. Equally irrelevant, in the General Court’s view, was the fact that AMD, Intel’s main competitor, was commercially highly successful and grew its market share during the period of abuse.
The Intel judgment thus favoured a form-based approach, unlike Post Danmark I before it, and went against the grain of the Commission’s 2008 Guidance. This has generated much debate among commentators, revisiting many of the arguments on form versus effects that were used ten years earlier at the start of the Commission’s reform process.
Form in Post Danmark II: clarity and legal certainty?
In the area of rebates and discounts, EU case law has tried to provide clarity by distinguishing three categories. The first category is quantity rebates. These are linked solely to the quantity of purchases, and are generally deemed benign since they tend to reflect cost savings from higher volumes.
The second and third categories are the potentially problematic ones: loyalty rebates and exclusivity rebates. The form of loyalty rebate that has received most attention in competition law is the retroactive rebate, which applies not only to the customer’s incremental purchases above the target, but retroactively to all purchases. This makes it attractive for customers to purchase any incremental requirements from the dominant company as this generates large additional rebates—a loyalty-enhancing mechanism called the ‘suction effect’. Exclusivity rebates, on the other hand, are conditional on the customer obtaining all or nearly all of its requirements from the dominant company.
The distinction between these categories is not clear-cut, and the Post Danmark II ruling does not help in this regard, as discussed below. Exclusivity rebates are closely related to loyalty rebates. Exclusivity rebates also come in degrees: in the extreme version they are explicitly conditional on full exclusivity; in softer versions the condition refers to some majority proportion of the customer’s requirements (Intel’s rebate schemes were of the latter kind). A form-based approach is therefore not necessarily straightforward. The Intel judgment contains a long discussion, running over many pages, of emails, internal presentations and other factual evidence to establish whether Intel’s customers had actually understood the form of the rebates—i.e. that they were conditional on exclusivity.
The rebates offered by Post Danmark were retroactive over a one-year period. Prices were set at the start of the year based on estimated volumes, and adjusted at the end of the year based on the actual volume of mail sent. Volume thresholds were set for discounts on a scale from 6% to 16%, and applied across all customers—i.e. they were standardised. This contrasts with previous rebate cases where thresholds were set individually based on a customer’s purchases in previous periods, which was seen as enhancing the loyalty-inducing effect. In this regard the Post Danmark rebates were of a less anticompetitive form than other retroactive rebates.
The ECJ restated the principle that quantity rebates linked solely to the volume of purchases are not anticompetitive. It considered Post Danmark’s rebates to be different from quantity rebates since they were not granted with respect to individual orders (thus not corresponding to cost savings made by the supplier), but on the basis of the aggregate orders placed over a given period. This reasoning does not necessarily hold, since cost savings from greater volumes may arise not only per individual order, but also from a customer’s aggregate purchases over the year.
Post Danmark’s rebates did not fall into the category of exclusivity rebates, as they were not conditional on the customer obtaining all or a given proportion of their demand from the incumbent. Somewhat confusingly, the ECJ stated that they were therefore different from loyalty rebates. The term ‘loyalty rebates’ is usually applied in a broader sense to capture retroactive rebates as well (as mentioned above).
In all, it is difficult to sustain the argument that the form-based approach to rebates by dominant companies—with some forms being allowed and some being ruled out per se—provides substantially greater clarity or legal certainty than a more effects-based approach. It is not always clear which category a rebate scheme falls into, and within each category there can be significant differences in the anticompetitive nature of the rebates.
Post Danmark II: a small step towards effects analysis?
Post Danmark’s rebates clearly had an effect on competition. Bring Citymail, a subsidiary of the Norwegian postal incumbent, had entered the Danish market for business mail, including direct mail, in 2007. It was the only serious competitor to Post Danmark, but exited the market in 2010 after suffering heavy losses. Post Danmark had a market share of more than 95% in bulk mail. Barriers to entry and expansion were considered to be high due to economies of scale and the national coverage of Post Danmark’s delivery network. In addition, in the relevant time period, more than 70% of the bulk mail market was subject to statutory monopoly. This meant that Bring Citymail could compete for only 30% of the market.
The ECJ stated that it was necessary to take into account the extent of Post Danmark’s dominant position and the particular conditions of competition prevailing in the relevant market. This is in line with the economic logic that the degree of dominance matters when assessing effects. The ECJ also considered it relevant that the rebate scheme covered 25 of Post Danmark’s largest customers, representing half of the direct mail market:
the fact that a rebate scheme, such as that at issue in the main proceedings, covers a majority of customers on the market may constitute a useful indication as to the extent of that practice and its impact on the market, which may bear out the likelihood of an anti-competitive exclusionary effect.
Could an as-efficient competitor match Post Danmark’s rebates without incurring a loss? This question is at the heart of the as-efficient competitor test for exclusionary conduct. The ECJ considered it relevant that Post Danmark had 95% of the market, and that 70% of the market was uncontestable due to statutory monopoly. The ECJ also noted that the suction effect from the retroactive rebates was further enhanced by the fact that they applied across the contestable and non-contestable parts of the market. Due to its statutory monopoly, Post Danmark was an unavoidable trading partner, thus making it even more attractive for customers to purchase all of their requirements from the incumbent in order to obtain a larger discount.
After considering these factors, it would have been a small step for the ECJ to apply the price-cost formula of the as-efficient competitor test: what is the effective price that Post Danmark charged for the contestable share of the market and that a competitor would therefore need to match to attract customers? With a 95% share of the market, and 70% of the market being uncontestable due to statutory monopoly, it is not difficult to conclude that any material discount would be difficult to match. For example, if a customer purchased direct mail services worth 100 krone at the full price, and the discount of 16% applied to all purchases, the effective price charged in the contestable part of the market would be 46.7 krone, a discount of more than 53%. This would be difficult for a competitor to match.
Yet the ECJ stated that it is not necessary to apply the as-efficient competitor test to show that a rebate scheme is anticompetitive. It also noted that, in a market protected by high barriers to entry and economies of scale, even a less-efficient competitor might contribute to intensifying competition. This is correct in theory, but it is a reason not to use the term ‘as-efficient’—and not a reason to reject the logic of the test outright. In previous regulatory contexts, the as-efficient competitor test had often been adjusted to account for the fact that new entrants were ‘not yet as efficient’.
The ECJ stated that the anticompetitive effect of a rebate scheme must be ‘probable’, there being ‘no need to show that it is of a serious or appreciable nature’. This is not a very clear criterion. How probable is ‘probable’? How does this threshold compare to that of the ‘dangerous probability of success’ used in the USA, as mentioned above? This is typically a matter of degree and judgement. At least the ECJ clarified that the anticompetitive effect must not be ‘purely hypothetical’. Again the ECJ referred to the need to assess all relevant circumstances, including the nature of the rebate schemes and the number of customers concerned. From an economic perspective, applying the as-efficient competitor test can add rigour to the assessment here. The ECJ did not disregard the as-efficient competitor test for rebate cases altogether, but considered it as ‘one tool among others for the purposes of assessing whether there is an abuse of a dominant position in the context of the rebate scheme’.
The debate on form versus effects in abuse of dominance cases is likely to continue. A next milestone will be the ECJ’s judgment in Intel, following the appeal against the General Court’s judgment. Meanwhile, Post Danmark II has provided some greater clarity: the as-efficient competitor test is not a necessary condition when assessing retroactive rebates by dominant companies, but it can be a useful tool, and there is still room to ask some effects-based questions, including in relation to the extent of the dominance and the proportion of customers covered by the rebates.
Sound legal criteria must give certainty to businesses and their advisers about what is allowed and what is not. It would be unworkable for each abuse of dominance case to require an exhaustive economic analysis of all the positive and negative effects on competition and consumers. However, that is not what many proponents of an effects-based approach have in mind, and nor does the Commission’s 2008 Guidance imply this. It can sometimes be sufficient to ask a few simple effects-based questions that take the analysis beyond dominance and form: what is the degree of dominance? What is the degree of foreclosure? Is the exclusionary strategy likely to succeed? Post Danmark II, like Post Danmark I before it, acknowledges the relevance of these questions.
Contact: Dr Gunnar Niels
 Case C-23/14 Post Danmark A/S v Konkurrencerådet, judgment of 6 October 2015 (Post Danmark II).
 Case C-209/10, Post Danmark A/S v Konkurrencerådet, judgment of 27 March 2012 (Post Danmark I).
 Case T-286/09 Intel Corp v Commission, judgment of 12 June 2014.
 European Commission (2008), ‘Guidance on the Commission’s Enforcement Priorities in Applying Article 82 EC Treaty to Abusive Exclusionary Conduct by Dominant Undertakings’, December.
 US v Aluminum Co of America 148 F 2d 416, 430 (2nd Cir. 1945).
 Case 322/81 Nederlandsche Banden-Industrie Michelin NV v Commission  ECR 3461 (Michelin I).
 Spectrum Sports, Inc. v McQuillan, 506 U.S. 447 (1993).
 Case C-209/10, Post Danmark A/S v Konkurrencerådet, judgment of 27 March 2012 (Post Danmark I), para. 22.
 Case C-209/10, Post Danmark A/S v Konkurrencerådet, judgment of 27 March 2012 (Post Danmark I), para. 44.
 Intel (Case COMP/C-3/37.990), Decision of 13 May 2009.
 Case T-286/09 Intel Corp v Commission, judgment of 12 June 2014, para. 80.
 Case T-286/09 Intel Corp v Commission, judgment of 12 June 2014, para. 85.
 Case T-286/09 Intel Corp v Commission, judgment of 12 June 2014, para. 132.
 See, for example, Wils, W. (2014), ‘The Judgment of the EU General Court in Intel and the So-Called “More Economic Approach” to Abuse of Dominance’, World Competition, 37:4, pp. 405–34; and Rey. P. and Venit, J.S. (2015), ‘An effects-based approach to Article 102: A response to Wouter Wils’, World Competition, 38:1, pp. 3–30. See also Oxera (2015), ‘Reform of Article 82: where the link between dominance and effects breaks down (revisited)’, Agenda, April.
 Case C-23/14 Post Danmark A/S v Konkurrencerådet, judgment of 6 October 2015 (Post Danmark II), para. 46.
 Total purchases are 100 krone at full price. 30% of purchases are contestable. A 16% discount equals 16 krone. Attributing the entire discount to the contestable part of the market—which the as-efficient competitor test does—means that the effective price in that part is 14 krone—i.e. a discount of 53.3% (16/30). The next step in the as-efficient competitor test would then be to compare this effective price of 14 krone with the avoidable costs or long-run incremental cost of the service. If the effective price is lower than the costs, the as-efficient competitor test is failed and the rebate is considered anticompetitive.
 See, for example, Ofcom (2008), ‘Complaint from Energis Communications Ltd about BT’s Charges for NTS Call Termination’, August.
 Case C-23/14 Post Danmark A/S v Konkurrencerådet, judgment of 6 October 2015 (Post Danmark II), para. 74.
 Case C-23/14 Post Danmark A/S v Konkurrencerådet, judgment of 6 October 2015 (Post Danmark II), para. 65.
 Case C-23/14 Post Danmark A/S v Konkurrencerådet, judgment of 6 October 2015 (Post Danmark II), para. 60.
Understanding and influencing behaviour: economics vs science
Have you heard the terms behavioural science and behavioural economics floating around? Individuals, corporations and governments are turning to both in the quest to understand and change human behaviour—but does their application add value? And what is the difference between these two similarly named disciplines? With reference to our… Read More
Contemporary infrastructure business cases: assessing impact in the round
Recent developments in infrastructure business cases reflect a complex and volatile environment. Writing infrastructure business cases now comes with new challenges—but how can we tackle these challenges in such a rapidly evolving environment? We explore the key points that factor into this, including the Five Case Model, value-based decision-making, uncertainty,… Read More