It is never a good sign when an advocate general’s opinion warns the European Court of Justice (ECJ) not to be swayed by “ephemeral trends” or the “Zeitgeist” of economic analysis, but instead to stick to the “legal foundations on which the prohibition of abuse of a dominant position rests in EU law”.
Advocate General Kokott’s opinion in Post Danmark offers old-school analysis of Post Danmark’s rebates.1)See http://curia.europa.eu/juris/document/document.jsf?text=&docid=164331&pageIndex=0&doclang=en&mode=lst&dir=&occ=first&part=1&cid=58656.
The relevant market was the bulk mail market, in which Post Danmark had a monopoly share. Its standardised volume rebate scheme applied to all customers, with a series of tiers starting at 30,000 letters. The scheme offered increasing rebates starting at 6% (for 30,000 letters) and increasing by 1% increments until the final two tiers, which increased by 2% increments. The final tier rebate was 16%. Post Danmark has a 95% share and its only rival a 5% share.
Although the advocate general said that as an efficient competitor (AEC) cost-based test was not necessary to analyse the schemes, an AEC would have immediately revealed a problem with the scheme. The rule for tiered rebates is to take the highest increment (2%), divide it by the contestable share (using the rival’s 5% share as a proxy) and add the second-to-last rebate in the scheme (14% – which is 2% less than the final tier rebate of 16%). Here that means a competitor faces an effective discount of (2/5 x 100) +14 = 54%. It is difficult to believe that a state monopolist postal system could achieve margins in excess of 50%. As a result, this scheme was always likely to be illegal under a price-cost test.
The advocate general offered a traditional view of rebates. They can be either quantity based, passing on cost savings through economies of scale, or of the loyalty-inducing kind. The label given to the scheme is irrelevant; “what matters is whether the dominant undertaking grants rebates which are capable of producing on the relevant market an exclusionary effect“.
The exclusionary effect is a matter of assessing the scheme in the round. The assessment should take account of the “rules governing the grant of the rebate… the conditions of competition prevailing on the relevant market and the (closely associated) position of the dominant undertaking on that market”.
The advocate general found that the retroactive nature of the scheme tended to exclusivity:
“An indication that the rebate scheme is abusive is present where… the effect of that scheme is not purely incremental, in the sense that the achievement of each new rebate threshold not only triggers a reduction in the price of all further orders placed but is also retroactive, thus also cutting ex post facto the cost of all orders already placed during a given reference period.”
The advocate general came close to saying that any retroactive scheme will have an unlawful loyalty-inducing effect: “such a rebate scheme usually has a loyalty-building effect.”
However, the following paragraph suggests that retroactivity alone is not enough. If the rebates are high in value and the reference period is long, an exclusionary effect is more likely. The facts of this case are extreme. The advocate general noted that the exclusionary effect was:
“further enhanced by the fact that the rebates applied without distinction both to the contestable part of demand and to the non-contestable part of demand, that is to say, in particular letters up to 50g, which are covered by Post Danmark’s statutory monopoly.”
Is an objectively applied scheme less culpable?
Post Danmark’s claims that the scheme was an objective, standardised rebate scheme, applicable to all and without exclusionary intent, fell on deaf ears. The advocate general found that intent might be an evidential-plus factor. However, its absence did not absolve Post Danmark: “an exclusionary intent or strategy is not, however, a mandatory precondition for a finding of infringement… since abuse of a dominant position is an objective concept.” Nor did the objective, standardised volume nature of the scheme. Discrimination might make “the abusive nature of the rebates applied by the dominant undertaking… particularly manifest”, but:
“the existence of abuse hangs less on a distinction between individualised and standardised rebates and more on an examination of whether the rebate scheme in question is actually capable of producing exclusionary effects on the market.”
The advocate general considered market shares, the size of rivals, foreclosing factors for entrants and whether network effects favoured the incumbent. In this case, the market conditions indicated a capability to foreclose. Post Danmark controlled “an overwhelming 95% share” of the market, while its statutory monopoly covered more than 70%. It had one geographically constrained competitor with a 5% share. The advocate general noted the greater impact of retroactive schemes where rivals are very small:
“Such a wide divergence between the dominant undertaking’s market share and that of the competition may promote the emergence of exclusionary effects, since it is particularly difficult in those circumstances for competitors of the dominant undertaking to undercut the rebates offered by the latter, which are based on its overall volume of sales.”
The creation of a postal infrastructure and network involved substantial network effects and barriers to entry, which would be difficult for any rival to overcome. Finally, at least two-thirds of customers faced losing their Post Danmark rebates if they switched business to the smaller rival, “indicative of a high exclusionary potential on the part of the rebate scheme operated by Post Danmark”.
The advocate general concluded that the scheme was likely to be abusive on this basis.
The advocate general declined the opportunity to steer rebates case law towards a price-cost test, such as that advocated by the EU enforcement priorities paper. The exclusionary effect of the rebate scheme was analysed by assessing whether a rival could remain above cost if it discounted its prices over a small quantity of contestable volume to match the dominant company’s rebates. She held that “a reorientation of the case law concerning Article 82 EC warrants some scepticism, on a number of grounds”. It is resource intensive, difficult to interpret and depends on unreliable data:
“There are many other factors, such as the specific modus operandi of a rebate scheme and certain characteristics of the market on which the dominant undertaking operates, that may also be relevant to a finding of abuse. In fact, they may be much more informative than a price/cost analysis.”
The advocate general concluded that the authorities can use such a test, but there is no need to do so – and there may even be markets where the price-cost test is inappropriate. For example, a market characterised by high barriers to entry, network effects and a statutory monopoly may simply not allow a second competitor to enter at a scale that would allow its cost base to approach the efficiencies of the incumbent. She summarised the position as follows:
“Article  does not require the abusive nature of the rebate scheme operated by a dominant undertaking to be demonstrated by means of a price/cost analysis such as the as-efficient-competitor test, where its abusive nature is immediately shown by an overall assessment of the other circumstances of the individual case…
“the authorities and courts dealing with competition cases are at liberty to avail themselves of a price/cost analysis in their overall assessment of all the circumstances of the individual case…
“unless, on account of the structure of the market, it would be impossible for another undertaking to be as efficient as the dominant undertaking.”
Causation and de minimis
The advocate general also looked at two additional difficult areas of Article 102 case law.
First, what degree of causation must be proven to show that a scheme is exclusionary? Older case law effectively dispensed with a need for any analysis of actual effect. It sufficed if the scheme was ‘capable of’ or ‘tended to’ exclude. The advocate general decided that this was unsatisfactory. As with any antitrust burden, the regulator must show “on the basis of an overall assessment of all the relevant circumstances of the individual case, the presence of the exclusionary effect appears more likely than its absence”. This is a limited, if not substantial improvement in legal clarity. However, she tempered this conclusion by offering greater hope of leniency in less clear-cut abuse cases when it comes to fines: “the degree of likelihood as to the presence of exclusionary effects may at most have a bearing on the extent of any penalties.”
Second, does Article 102 admit of a de minimis exception? If only one small customer is affected by an exclusionary practice – leaving the remaining 99% of the market open to competitors – how can this be culpable conduct? Both Tomra and Intel dicta suggest no de minimis safe harbour. The advocate general confirmed prior case law and found a new way of rationalising that a category of lawful, negligible exclusionary conduct must exist:
“The use of a de minimis threshold for the purposes of assessing the exclusionary effects of a dominant undertaking’s commercial conduct seems to be unnecessary for two further reasons: first, as I have already said, the exclusionary effects referred to above are to be determined on the basis of a specific examination of all the relevant circumstances of the individual case and their presence must be more likely than their absence. Secondly, the prohibition of abuse contained in Article  is in any event directed only at conduct capable of affecting trade between Member States. Each of those two considerations is in itself entirely sufficient to rule out the proposition that the prohibition of abuse laid down in Article  may apply to conduct the anti-competitive effects of which are only hypothetical or of wholly negligible significance.”
It is often said the ECJ more readily advances the law’s development through national court references than through European Commission case appeals. However, Kokott warned the ECJ against judicial development which would usher Article 102 rebate cases into the economic age. It is disappointing to see economics on the ‘too difficult’ pile for regulators and courts. The Article 102 priorities paper offers an intellectually rigorous method for filtering harmful conduct from pro-competitive price conduct.
The advocate general did not entirely exclude price-cost economics from the assessment. It remains open to authorities to use this tool to assess rebate schemes. Failing a price-cost test is likely to indicate illegal exclusion; but passing the test is no safe harbour. If the mechanics (large, retroactive rebates) and broader economic factors (high shares, small rivals, network effects and entry barriers) indicate exclusion, then this will suffice.
Much turns on the facts of the case. Intuitively, a case involving a statutory monopolist and state incumbent with an overwhelming market share squeezing out its only nascent rival was unlikely to win sympathy. In an open, less regulated, dynamic market, a less intrusive approach may have been applied. The UK Competition and Markets Authority recently dismissed a rebates case in the pharmaceutical sector precisely by applying the AEC methodology which the advocate general has warned against.2)See https://assets.digital.cabinet-office.gov.uk/media/558c2743e5274a1559000004/Pharmaceutical_sector_investigation_closure_statement.pdf.
If followed by the ECJ, Kokott’s views do not bode well for the Intel appeal. It is likely that Intel will also fail an ‘overall assessment’ type analysis, as advocated by Kokott.
Finally, the advocate general de minimis conclusions, named ‘negligible significance conduct’, are welcome. It was an aberration of earlier dicta that saw culpable abusive conduct in the most insignificant arrangements.
This update was first published in the Competition & Antitrust Newsletter of the International Law Office – www.internationallawoffice.com
References [ + ]