Michelin II: A per se rule against rebates by dominant companies?

Journal of Competition Law and Economics

Volumn 1, Issue 1, 2005, Pages 149-171

  Waelbroeck, Denis ^a,b  

^a FREE UNIVERSITY OF BRUSSELS   (Belgium)

^b EU International Relations and Diplomacy Studies Department   (Belgium)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 27844510507     PISSN: 17446414     EISSN: None     Source Type: Journal    
DOI: 10.1093/joclec/nhi001     Document Type: Review

References (82)

1. ECJ, Case 85/76 Hofmann-La Roche v Commission of 13 February 1979 [1979] ECR 461.
2. See, e.g., Case 322/81 NV Nederlandsche Banden Industrie Michelin v Commission of 9 November 1983 ('Michelin I') [1983] ECR 3461, at para 14
3. See also case C-310/93 BPB Industries plc and British Gypsum v Commission of 6 April 1995 [1995] ECR I-865
4. Case T-228/97 Irish Sugar plc v Commission of 7 October 1999 [1999] ECR II-2969.
5. CFI, Manufacture française des pneumatiques Michelin v Commission of 30 September 2003, Case T-203/01.
6. Commission decision 2002/405 relating to a proceeding pursuant to Article 82 of the EC Treaty, COMP/E-2/36.041/PO Michelin, OJ L143/1, 31.05.2002.
7. See the OFT draft guidelines on 'Assessment of conduct' (OFT 414a) from April 2004: 'The offering of discounts to customers is an important form of price competition and is therefore generally to be encouraged.'
8. Discounts off prices may well render firms pricing decisions less transparent. For example, the European Commission's guidelines emphasize that '...transparency may be low in a market where transactions are confidentially negotiated between buyers and sellers on a bilateral basis' (see para 50 of the European Commission's 'Guidelines on the assessment of horizontal mergers under the Council Regulation on the control of concentrations between undertakings', OJ C31/5, 5.2.2004).
9. See CFI, Case T-203/01 Michelin v Commission of 30 September 2003, see also Case T-219/99 British Airways v Commission of 17 December 2003.
10. OECD report on loyalty and fidelity discounts and rebates. DAFFE/COMP (2002) 21.
11. See at paras 98 and 100. An example of non-linear rebates would include rebates calculated by reference to the customer's purchases in the preceding time-period, or if the total discount applied across overall sales volumes increases as certain sales targets are met.
12. Admittedly, in Michelin II, the issue of discrimination had not been raised by the Commission and was hence not discussed by the Court (See para 63 of the judgment). It was, however, discussed in a number of other cases on rebates.
13. See, e.g., ECJ, Case C-163/99 Portugal v Commission of 29 March 2001 [2001] ECR I-2613, paras 52-53. In Michelin II, the Court's heavy reliance on Portugal v Commission is somewhat puzzling as the economic issues are rather different: One case involves discrimination in airport landing fees between the largest national airline and smaller airlines (i.e., involving the conditions of competition on the downstream market), whereas Michelin II related to the possible exclusion/ foreclosure of rivals at the same level of the market.
14. Ridyard, 'Exclusionary Pricing and Price Discrimination Abuses under Article 82 - an Economic Analysis' [2002] ECLR 286.
15. See also Bishop and Walker, The Economics of Competition law, pp. 201-202. "If the Commission wants discounts to be cost justified, the most that economic theory suggests should be meant by this is that incremental prices should be above incremental costs. Provided discount schemes or rebates do not lead to incremental prices below incremental cost of supply, they should be considered cost justified. A key point here is that, for firms with low marginal costs, even big price rebates can be 'cost justified' in the only sense that really matters from the point of view of economic efficiency, since they still tend to exceed the marginal costs of supply".
16. The fact that discrimination is often a feature of competitive markets has in fact been repeatedly recognized by the Commission itself, for instance, when it prohibited industry-wide agreements preventing companies to discriminate (See, e.g., Commission decision 74/292/EEC of 15 May 1974, and decision 75/497/EEC of 15 July 1975, IFTRA). See comments above in relation to rebates, etc. reducing the risk of coordination.
17. See, e.g., Muysert, 'Price Discrimination - An Unreliable Indicator of Market Power', 25(6) European Competition Law Review (2004).
18. OECD report on loyalty and fidelity discounts and rebates. DAFFE/ COMP(2002)21 at p. 7.
19. Even from a strict legal point of view, it is unclear that such differences in prices are themselves 'discriminatory' in the sense of Article 82(2)(c) EC. Different prices indeed mostly reflect different elasticities in demand and in United Brands, the ECJ itself found that a difference in elasticity of demand ('density of competition') could justify a different pricing. It is a different treatment but of different situations. The fact that a firm charges different price-cost margins on different transactions is thus not necessarily a form of discrimination even in EC law, and will often be the most efficient pricing policy, also from the point of view of the overall welfare. Since different customers will inevitably have different demand elasticities, any attempt to impose a uniform price will arguably result in a suboptimal output (see also the principles often described as 'Ramsey pricing', which justifies cross subsidiations from product A to product B, and hence price discriminatory as leading to higher economic welfare). For this reason, it is unclear that creating a constant price-cost ratio over all transactions should even be attempted (see Ridyard, 'Exclusionary pricing and Price Discrimination abuse under Article 82 - an Economic Analysis' [2002] ECLR 286: 'A per se prohibition on price discrimination - i.e., a requirement that dominant firms should earn equal price-cost mark ups on all their transactions - would be unduly restrictive and almost certainly lead to grossly inefficient outcomes in the context of fixed cost recovered industries'). In its September 1999 Guidelines on Assessment of Individual Agreements and Conduct, the OFT therefore rightly stresses that 'Price discrimination raises complex economic issues and is not automatically an abuse. There are marry areas of business where it is a usual and legitimate commercial practice. For example, it might be objectively justified in industries where there are large fixed costs and low marginal costs'.
20. Guidelines 3.8.
21. John Vickers, 'Abuse of market power' Speech to the 31st Conference of the European Association for Research in Industrial Economics, Berlin, 3 September 2004.
22. See, e.g., Spector, 'Loyalty rebates and related pricing practices: When should competition authorities worry?', Chapter 10, Global Competition Policy: Economic issues and impacts, LECG, 2004, pages 325, 326 and 329.
23. Example: Staff training, possession of certain machinery, vehicles and know-how, quality of facilities, etc.
24. Example: Twice as many points could, in certain cases, be granted for the assessment 'excellent' as for the assessment 'good'.
25. Companies grant rebates for many different reasons, e.g., because it is difficult to anticipate demand to appropriately allocate costs over time, or because it is a mechanism to reduce prices without reducing 'face' prices, or simply to respond to requests from customers, or still because it is a standard industry practice, etc. ....
26. See, e.g., Vertical guidelines, OJ no C291/1, 13 October 2000, para 116, para 153 ff.
27. Although there is no equivalent in Article 82 to Article 81(3), there is indeed no a priori reason not to apply a similar balancing test in Article 82 cases. Even in horizontal mergers, the Commission has accepted that the efficiency defence applies (See Commission guidelines on the assessment of horizontal mergers under the Council Regulation on the control of concentrations between undertakings, OJ no C31, 5 February 2004, p. 5, at paras 76-88).
28. Except possibly in the Commission's recent Microsoft Decision, where however the Commission dismisses the argument for the reason that 'Microsoft has failed to supply evidence that tying of WMP is indispensable for the alleged pro-competitive effect to come into effect' (decision of the Commission of 24 March 2004, in case COMP/C-3/ 37.792, Microsoft, at paras 955 ff not yet published). Moreover, in that case, several remarks could be read as entailing an 'efficiency offence', e.g., in para 847 where it criticizes Microsoft's statement that integration of new functionalities into the operating system 'reduces the price customers have to pay for costly add-ons from other vendors that may not be needed'. It could be that the Commission here sees the efficiency as affecting negatively competition, and thus as 'bad'. See also para 816 where the Commission sees no problem in bundling as long as Microsoft's media player had no streaming media capability. A problem occurred only in 1999, when 'Microsoft tied a product that matched other vendors' products'!
29. See, e.g., CFI, case T-65/89 British Gypsum [1993] ECR II-389 at paras 65 ff.
30. See, e.g., CFI, case T-65/98 Van den Bergh Foods v Commission, 23 October 2003, para 159, where the Court repeats the statement in British Gypsum that exclusivity agreements 'are concluded in the interests of the two parties and cannot be prohibited as a matter of principle'. The Court nevertheless stresses that 'business conduct which contributes to an improvement in production or distribution of goods and which has a beneficial effect on competition' may be problematic in the case of dominant companies, at least in case such as Van den Bergh Foods, where it effectively forecloses '40% of outlets in the relevant market'.
31. Vertical guidelines, quoted above, in para 141, which accepts that non-compete obligations can be 'objectively justified' even for dominant companies within the context of Article 82.
32. Similarly to the efficiencies in vertical agreements, such 'pro-competitive effects' arise, inter alia, from encouraging investments and competition by preventing 'free-riding', providing security of sales and purchases, inducing lower prices, etc.
33. In para 75 of the Michelin II judgment, the Court summarizes the Commission's position as follows: 'the Commission infers that the quantity rebates are loyalty inducing from the following evidence: The fact that the discount is calculated on the dealer's entire turnover with Michelin and the fact that the reference period applied for the purpose of the discount is one year'. The Court agreed by stating, e.g., in para 85 of its judgment that 'it cannot be denied that the loyalty inducing nature of a system of discounts calculated on total turnover achieved, increases in proportion to the length of the reference period'. The Court concluded - at para 95 - that 'a quantity rebate system in which there is a significant variation in the discount rates between the lower and higher steps, which has a reference period for one year and in which the discount is fixed on the basis of total turnover achieved during the reference period, has the characteristics of a loyalty inducing discount system'. In such a case, the Court finds that if the dominant company cannot demonstrate (para 107) that the rebate 'rewards an economy of scale' (para 98) it is illegal, independently of its effects on the market (para 241).
34. Economies of scale arise where average costs fall with output, typically because fixed costs can be spread over a greater sales volume (until capacity constraints are reached). As a result, the decline in average costs for a given increase in output level falls as output increases, rather than falling by a constant amount or percentage.
35. Which moreover notes - beyond recovery of fixed costs - other cost efficiencies such as reducing supplier's sales variability thus allowing economies in smoothing its production and reducing its inventories.
36. In so doing, the Court ignores also that Article 2 of Regulation 1/2003 states clearly that it is up to the Commission to prove an infringement of EC Article 82. See para 58 according to which 'quantity rebate systems linked solely to the volume of purchases made from an undertaking occupying a dominant position are generally considered not to have the foreclosure effect prohibited by Article 82 EC [...] Quantity rebates are [...] deemed to reflect gains in efficiency and economies of scale made by the undertaking in a dominant position'.
37. Paras 98 ff of the judgment; see para 108: 'Far from establishing that the quantity rebates were based on actual cost savings [...], the applicant merely states generally that the quantity rebates were justified by economies of scale in the areas of production costs and distribution'.
38. It is normally only if the Commission can establish a prima facie that a given system poses a major problem under competition law (as was the case, e.g., in case Portugal/Commission) that the presumption will be reversed. No such proof was however put forward by the Commission in Michelin II. During the administrative procedure, the Commission never discussed the economic justification of the quantitative rebates, as the Commission itself did not deny in its pleadings. Nor was the point addressed anywhere in the Decision, contrary to what the Court tries to claim at para 101 of its judgment in an overstretched reading of paras 216 and 227 of the Decision.
39. See, e.g., Article 5 of Regulation No. 2790/1999.
40. See Case 85/76 Hoffmann-La Roche v Commission of 13 February 1979 [1979] ECR 461, para 91, emphasis added.
41. In Michelin II, the Court ignores - rightly - the Commission's allegation that the rebates were 'exploitative'. This odd statement was criticized in the appeal, as well as at the hearing by the judges.
42. See RBB Brief of February 2004, 'The special responsibility of dominant firms under Article 82: Do not compete on price'.
43. See CFI, case T-65/98 Van den Bergh Foods v Commission 23 October 2003, paras 98 ff (concerning Article 81), and para 160 (concerning Article 82 and concluding that the practice had a foreclosure effect). In Van den Bergh Foods, the CFI held that foreclosure of competition under Article 82 in an exclusivity-type case need not reach the level of complete foreclosure but indicated that it must not be insignificant. The Court in that case concluded that de facto 40% of the outlets were foreclosed, although it recognized that the foreclosure was not absolute as retailers were not prevented from buying a second freezer cabinet or replacing the one supplied by the dominant company (see at paras 149 and 160).
44. This tends to show that - at least so far - whenever foreclosure effects are shown to exist, they will be used in support of a finding that the practice is abusive but the opposite does not apply. See, e.g., Decision of the Commission of 14 May 1997 in Irish Sugar, OJ 1997, L 258, p.1, at para 129.
45. Cited above in paras 835 ff. In that case, the Commission recognizes that users can and do to a certain extent obtain third party media players through the internet, sometimes for free and that 'there are therefore indeed good reasons not to assume without further analysis that tying WMP constitutes conduct which by its very nature is liable to foreclose competition'. The Commission claims that foreclosure will occur as a result of 'indirect network effects'. The alleged foreclosure remains however unconvincing as WMP compares roughly with Realplayer and as they have a market share of 30% each (See para 907 of the decision).
46. Exclusive dealing analysis applies also where exclusivity is the practical effect of the practice under examination see Tampa Electric v Nashville Coal Co. 365 US 320 (1961).
47. Section 3 of the Clayton Act outlaws exclusive dealing arrangements where the effect of the arrangement 'may be to lessen competition substantially or tend to create a monopoly in any line of commerce'.
48. Prior to the passage of the Clayton Act, the courts had taken an extremely permissive approach under common law (see, e.g., Chicago, St Louis and New Orleans Railroad Co. v Pullman Southern Car Co. 139 US 79 (1891).
49. Standard Fashion Co. v Magrane-Houston Co. 258 US 346 (1922).
50. Oil Co. (Cal.) v United States 337 US 293 (1998).
51. In that case the contracts in question affected 6.7% of the relevant market and similar contracts with Standard Oil's six largest competitors covered 42.4%. Exclusivity has been condemned in cases involving contracts covering very low percentages of the market for example Brown Shoe Co. v FTC 384 US 316 (1966) involving only 1% of the retail shoe market.
52. Tampa Electric Co. v Nashville Coal Co. 365 US 320 (1961).
53. Electrics Corp. 100 FTC 68 (1982).
54. Gilbarco 127 F.3d at 1162.
55. Although the third, dissenting judge came to the opposite conclusion.
56. 3M Co. v Le Page's Inc., US, No. O2-1865, 10 June 2003.
57. In that case, 3M - maker of scotch tape and Post-it notes - had more than 90% of the US market in transparent tape and conceded it had a monopoly. It began competing with Le Page's in the small private label tape market. To some retailers, 3M offered rebates on total purchases of 3M tape, if the retailer purchased target amounts. To other retailers, 3M offered bundled rebates of up to 2%, depending on the number of customer-specific growth targets met in 3M's six broad product lines. The jury awarded Le Page's USD 68.5 million for 3M's unlawful maintenance of monopoly. The Court rejected 3M's argument that its conduct must be assessed solely under the Brooke Group rule, i.e., requiring proof a possible recoupment of the costs of the practice in the longer term. The Court held that the jury was entitled to find that 'the long-term effects of 3M's' bundled pricing 'were anticompetitive', but is was quite vague on why. The Court held that bundled pricing can exclude an efficient rival but did not hold that the jury could have found 3M's conduct would have excluded an efficient rival. A petition for review was rejected by the Supreme Court.
58. A different approach with regard to the burden of proof, but also towards rebate schemes in general, is nowhere better illustrated than in the treatment of Virgin Atlantic complaint about British Airways' incentive systems. The European Commission found in its decision 2000/74 of 14 July 1999, Virgin/British Airways, OJ L30/1, 4.2.2004, that British Airways abused its dominance on the UK market for the purchase of travel agent services by rewarding loyalty and by applying dissimilar conditions to equivalent transactions. In its judgement in Case T-219/99 British Airways v Commission, of 17 December 2003, the CFI dismissed an appeal against this decision. By contrast, when Virgin tried to achieve the same result in the US, the district court granted summary judgement in favour of British Airways finding that Virgin's claims lacked factual support [Virgin Atl. Airways Ltd. v British Airways Plc, 69 F. Supp. 2d 571, (S.D.N.Y 1999)]. The Court of Appeals agreed that there was insufficient proof, but also concluded that Virgin failed to show how British Airways' competition harmed consumers [Virgin Atlantic Airways Ltd. v British Airways Plc, 257 F.3d 256, (2d Cir. 2001)]. The Court of Appeals stressed, inter alia, that 'the decision to offer incentives was nothing more than an attempt to generate increased business on the whole by limiting profitability on selected sales' and that 'low prices are a positive aspect of a competitive marketplace and and are encouraged by the antitrust laws'.
59. Jacobson, Exclusive Dealing, Foreclosure and Consumer Harm, 70 ALJ 311 at 349.
60. See appeal of the US department of Justice in United States v Dentsply, http://www.usdoj.gov/atr/cases/indx102.htm.
61. Paddock Publications Inc. v Chicago Tribune Co. 103 F.3d 42 (7th Circuit 1996).
62. See, e.g., RJ Reynolds Tobacco Co v Philip Morris Inc., 199 F. Supp. 2d 363 (M. D.N.C.2002), appeal docketed, No. 02.1595 (4th Cir. June 2002). In a very short unpublished opinion, the district court decision was upheld in appeal.
63. National Society of Professional Engineers v United States 435 US 679 (1978).
64. Guideline 5.1.
65. The OFT report on switching costs states in that in regard in relation to discounts: 'Another means of creating switching costs is through non-linear rebate schemes [...] non-linear schemes could potentially cause customer lock-in, although they may have other justifications in terms of recovery of fixed costs and increasing marginal sales. Such schemes can make it costly for customers near one of the thresholds for a rebate or discount to switch purchases to another firm and can mean that such incremental sales are sold below marginal cost. It is important to note that only non-linear discounts may cause problems - a linear discount (e.g., 1p back for €1 spent) is equivalent to a simple price reduction and, as long as it is not predatory, should not cause any foreclosure. However, a non-linear scheme, when practised by a dominant firm, can potentially exclude competitors competing for marginal sales. The analysis of the effects of such schemes is complicated, particularly where the product has high fixed costs'.
66. Guideline 5.2.
67. Guideline 6.10.
68. Decision CA98/20/2002, BSkyB of 17 December 2002, at points 638-646.
69. See, e.g., point 639 of Decision CA98/20/2002, BSkyB of 17 December 2002: 'The Director concludes that the (...) discount (...) has not infringed the Chapter II prohibition on the grounds that it did not and was not likely to foreclose entry to the wholesale market for premium channels in the period investigated'.
70. Decision of the OFT under section 47 relating to Decision CA98/20/2002, BSkyB of 29th July 2003.
71. See also Case No 1009/1/02 Aberdeen Journals [2003] CAT 11 para 350.
72. This also indicates that the burden of proof falls squarely on the authority alleging the abuse.
73. See point 164 of the Decision of the OFT under section 47 relating to Decision CA98/20/2002, BSkyB of 29 July 2003: 'Mixed bundling refers to the situation where two or more products are offered as unbundled (i.e., undiscounted) products and simultaneously offered together at a price less than the sum of the individual product prices (i.e., discounted)'.
74. See also Nalebuff, 'Bundling, Tying, and Portfolio Effects', DTI Economics Paper No. 1, February 2003, at point 2.2.
75. Decision of the OFT under section 47 relating to Decision CA98/20/2002, BSkyB of 29 July 2003 at point 192.
76. Decision of the OFT under section 47 relating to Decision CA98/20/2002, BSkyB of 29 July 2003 at point 194.
77. In other decisions, the OFT also found pricing at a level that could not be matched by competitors before going on to see if there were any effects on competition see, e.g., Decision of 11 July 2003, BT Broadband.
78. Decision of the OFT, BT/BSkjB broadband promotion, 15 May 2003.
79. Decision of the OFT, BT UK-SPN calls service alleged anticompetitive pricing, 23 May 2003.
80. Decision of the OFT, BT UK-SPN calls service alleged anticompetitive pricing, 23 May 2003, at point 48.
81. Decision of the OFT under section 47 relating to Decision CA98/20/2002, BSkyB of 29 July 2003, at point 170.
82. Decision of the OFT under section 47 relating to Decision CA98/20/2002, BSkyB of 29 July 2003, at points 178-181.