#89 - Article 82 Theory - Competition Law

Theory: Article 82

Assessment of Approach to Abuse pre-effects based approach

• Market definition:

  • United Brands decision clearly wrong: Court said that because there were some core/locked in consumers, the banana market was unique. Wrong: the aim of the SSNIP test is to identify other products that can be used as alternatives in the case of a price rise, which means you need to look at marginal consumers.

  • Microsoft case: Is an operating System really distinct from a media player? CFI conceded that changes to understanding/products occur over time, and said it was looking at it from a 1997 perspective when the alleged abuse occurs. If an OS is distinct from all the programmes on it (internet explorer, word processor etc) then there would be nothing left on the OS! It comprises these ‘complementary products’.

• Market power:

  • Wrong to use opportunity cost of switching away from more profitable routes in British Midlands/Aer Lingus case. The fact that there are fewer competitors in a market with less demand is an example of the market working perfectly, with price and conditions being constrained by the potential for other airlines to move onto the currently ‘less profitable route’ in response to a rise in prices.

  • Wrong to use integration as an indicator per se. If the downstream firm could feasibly integrate and put itself in same position as the integrated one, then it is wrong to conclude that the integrated firm is dominant. There is potential upstream competition, which may stop it from behaving independently to an appreciable extent etc

  • Wrong to use technological advancement: This could chill innovation

• Abuses:

  • Attitude to rebates is v formalistic: A loyalty inducing rebate (and especially individualised ones) are deemed to force consumers to deal with them alone and are therefore abusive per se. Courts don’t actually look at effect on consumers.

    • In Virgin/BA the court considered whether the rebates would foreclose the sales of tickets of other airlines from the market but NOT whether this would actually harm consumers: In an industry with huge fixed costs, any extra income that can be earned and contribute to paying off those costs is desirable for future consumers. Also the rebates would ensure that the planes were always full, improving allocative efficiency. It also treated retroactivity of rebate as indicating abuse, even though retroactive rebates help lower costs to customers/consumers the most.

    • Loyalty-inducing rebates are presumed anticompetitive unless objectively justified (in Michelin the CFI said anti-competitive effects could be presumed in loyalty rebate cases)- NB this isn’t quite a per se unlawfulness approach, but is a variant of it (given how strict CFI/ECJ/Commission have been in assessing objective justifications). This doesn’t reflect economics: Rebate is another form of price competition and where one firm offers a price that is lower than another, the other will necessarily be excluded or foreclosed from custom. If the contract or sales are envisaged over the long term, this could be deemed to be ‘loyalty inducing’ even though it’s just the market working properly.

    • Anticompetitive effect of rebates is increased cost of switching to other suppliers once the rebate period has begun (since a rebate, and particularly a retroactive rebate) deters the switch even if another competitor offers a lower unit price for fear of losing the rebate and hence whether the effect is truly exclusionary to the detriment of the consumer depends on whether the effective price can be matched by an equally efficient competitor. Just saying ‘loyalty inducing’ doesn’t capture this, as it doesn’t consider cost measures in the way that predation does. (Rebate is a ‘non-linear’ form of pricing).

    • Commission doesn’t consider minimum orders necessary to enter/stay in the market and as such doesn’t really measure exclusionary effect.

  • Tying and bundling:

    • Test is made out even where the tying/bundling helps consumer welfare e.g. Microsoft case where consumers could still choose to get other products and were given the tied product free. Also the definition of distinct products in this case is formalistic and unconvincing.

    • Tying and bundling can have benefits for consumers: It may involve offering a rebate to take both (lowering price) and saves transaction cost (e.g. of travelling to different places to buy the tied/bundled goods). Also by supplying both products the supplier achieves economies of scale or scope. However these are rarely considered by the courts/commission.

  • Refusal to supply: Risks of demanding supply: compulsory dealing undermines contractual freedom, which in turn reduces efficiency of dominant company (since contracting parties know when it is in their best interest to deal e.g. suppose the other party is a bad debtor or there’s a shortage of stocks); it allows free riders to take advantage of innovations of others e.g. IP or ‘essential facilities’, which chills incentive to be the first to innovate or make an investment; loss of contractual freedom leads to a loss of efficiency for dominant company where it is forced to enter bad bargains, and therefore may lead to higher prices. Courts and Commission don’t consider these sufficiently e.g. in Commercial Solvent, where ICI was being deprived of the benefit of its investment in creating a full distribution system, which probably had a more competitive downstream operation than competitors.

  • Unfair prices: Makes no sense- every price is above perfectly competitive price, since there are no perfectly competitive markets in the world. At what point does this become unfair? ECJ’s test of comparison to other markets in Pompes funebres is unsatisfactory: still doesn’t explain cut off points, while markets covering LEDCs will obv have lower prices than markets covering MEDCS

• Defences:

  • Courts thus far haven’t really considered efficiency defences- v rarely made out BUT commission in Wanadoo considered the availability of efficiency defences (though, again, not present on facts of the case). Bad that IP rights can’t constitute a defence (per CFI in Microsoft).

Assessment of Effects Based Approach

• Main idea: Only punishing those practices which actually harm consumers, unlike the formalistic approach above, which tends to make Type II errors (over-inclusion). Commission Guidance on Exclusionary Conduct (the guidance paper announcing the economic approach) says that it will generally only prosecute abuses that will eliminate an equally efficient competitor from the market. If an equally efficient competitor could compete then there is no abuse. In price-based exclusionary conduct (predation, rebates etc) the conduct is presumed not to be abusive if above LRAIC, and presumed abusive if below AAC, while the price level in between is a ‘grey area’. Commission recognises the validity of ‘objective justification’ and efficiency defence in the style of 81(3) where the conduct realises efficiencies, the conduct is indispensible to realise those benefits, the efficiencies must outweigh any negative impacts on consumer welfare, and it must not eliminate all effective competition. Also Commission says that a market share of below 40% is unlikely to enable dominance.

• New way of dealing with

  • Predation: Below AAC is abusive, above LRAIC is fine, in between is a grey area. In grey area, no need to show actual foreclosure- reduction of competition from the dominant firm being able to ‘discipline’ the competitor into following price set by dominant company

  • Rebates: Only abusive if they would eliminate an equally efficient competitor from the market. Thus only where the effective price is below AAC (or sometimes grey area of below LRAIC but above AAC) of the dominant undertaking will it be abusive. To determine effective price:

    • The effective price=price a rival would have to offer to compensate the customer for the loss of the conditional rebate if latter were to switch part of his demand (the ‘relevant range’) away from the dominant undertaking. This is the normal price of the dominant undertaking minus the value of the rebate, over the relevant range of sales and time period.

    • Relevant range = depends on facts of each case. For incremental rebates this is the incremental sales, whereas for retrospective rebates it’s how much of a customer’s purchase requirements can realistically be switched to a rival in the specific market context (the ‘contestable share’ or ‘contestable portion’). This contestable share will be larger when the customer is willing and able to switch and the competitor has sufficient capacity to meet new orders. NB a larger contestable share means the rebate is divided among more sales and hence effective price is higher (and easier to meet) so that its less likely that the rebate will be abusive. This test has recently been used by the commission in the Intel case, though it remains to be seen whether the CFI will accept it on appeal.

  • Refusal to supply: Commission recognises the potential free rider and deterrent to invest/innovate problems of obligations to supply. In terms of refusal to supply where the dominant company refuses to supply to a downstream competitor (or constructively refuses) or margin squeezes, this will be unlawful...