Harmony Gold Mining Company Ltd Durban R

The judgment

The court's own file

1 THE COMPETITION TRIBUNAL OF SOUTH AFRICA CASE NO: 13/CR/FEB04 In the matter between: HARMONY GOLD MINING COMPANY LTD First Complainant DURBAN ROODEPOORT DEEP LTD Second Complainant and MITTAL STEEL SOUTH AFRICA LTD First Respondent MACSTEEL INTERNATIONAL BV Second Respondent Panel: D Lewis (Presiding Member), N Manoim (Tribunal Member), and M Holden (Tribunal Member) Heard on: 15 March – 25 April 2006, with argument heard on 29-30 November 2006 Delivered on: 27 March 2007 REASONS LEWIS PM: The Complaint

[1]Harmony Gold Mining Company Ltd and Durban Roodepoort Deep Ltd (henceforth ‘the complainants’ or ‘Harmony’) have filed a complaint against Mittal Steel South Africa Ltd (“Mittal SA”) and Macsteel International Holdings BV (“Macsteel International”) (‘the respondents’) relating to the respondents’ conduct in the manufacture and distribution of flat steel products in South Africa.

[2]Harmony alleges that Mittal SA is a dominant firm in the domestic market for flat primary steel products and that it has abused this dominance by charging, in contravention of Section 8(a) of the Competition Act, excessive prices for its flat steel products. 2

[3]The complainants also allege that Mittal SA has contravened section 8(d)(i) of the Act, in that it requires or induces customers to not deal with a competitor. 1

[4]The complaint was lodged with the Competition Commission (“the Commission”) in terms of section 49B of the Act on 19 September 2002.

[5]The Commission, in addition to investigating the alleged contraventions of Sections 8(a) and 8(d)(i), also considered a possible contravention of Section 9(1) of the Act, which proscribes price discrimination. This investigation centred around the question of whether the differentiation by Mittal SA in its pricing of flat steel products sold locally relative to the price it charged for these products on the export market, as well as Mittal SA’s practice of granting incentives to promote the export of its steel in value added form and additional rebates to certain industries, constituted price discrimination.

[6]The Commission concluded that there was no evidence of a contravention by Mittal SA of either sections 8(a), 8(d)(i) or 9(1) of the Act. On 6 January 2004, the Commission subsequently issued a ‘notice of non-referral of complaint’. 2

[7]On 27 February 2004 the complainants then lodged the current complaint with the Tribunal. The Hearing

[8]After an extensive discovery process, the hearing of evidence concerning the present complaint commenced on 15 March 2006 and concluded on 25 April 2006.

[9]The hearing heard oral evidence from the following witnesses called by the complainants: Mr Gerhard Nicolaus; 3 Mr Bernard Swanepoel; 4 Mr Alistair Lang; 5 Mr Gary Bell; 6 Mr Stephen Leatherbarrow; 7 Mr Roy Cohen; 8 Mr Neil Senior; 9 Mr Peter 1 We will for purposes of this judgment also refer to these two alleged abuses as the ‘section 8(a) complaint’ or the ‘excessive pricing abuse/complaint’, and to the ‘section 8(d)(i) complaint’ as the ‘inducement abuse’. 2 See Pleadings File, pages 114-117. 3 Senior Manager of the Metals Directorate at the Department of Trade and Industry (“the DTI”). 4 Chief Executive Officer (CEO) of Harmony. 5 Director of Nampak Ltd 6 CEO of Bell Equipment (Pty) Ltd. 3 Fish; 10 Mr Errol Classen; 11 Mr Henry Pretorius; 12 Mr Gavin Jacobsen; 13 Dr Zavareh Rustomjee; 14 Professor Harvey Wainer; 15 and Professor Simon Roberts. 16

[10]The respondents’ witnesses were Mr Marthinus Schoeman; 17 Mr Phillip Tomlinson; 18 Mr Charles Dednam; 19 Mr Rudolph Torlage; 20 and Dr Mike Walker. 21

[11]A number of additional witness statements were filed by both parties.

[12]On 26 April 2006 and 8 May 2006, the complainants served Notice of Application to Amend the relief sought. The application to amend was opposed by both respondents on various grounds. The Tribunal heard the amendment application on 31 May 2006, and subsequently delivered its judgment on 19 June 2006. 22 This application to amend is discussed when we consider the remedies to be imposed.

[13]The Tribunal heard argument concerning the present complaint, on 29 and 30 November 2006. The Complainants

[14]The complainants are listed gold mining companies, registered and incorporated in terms of the company laws of the Republic of South Africa (“RSA”). They are consumers of a range of flat steel products manufactured by Mittal SA and sold by Mittal SA through a number of steel merchants from whom the complainants source their steel requirements. The flat steel products purchased by the complainants are shaped or pressed into items such as hoppers or skips, and used 7 Planning Director, Barloworld Robor Tube (Pty) Ltd. 8 Director, Conveyor Manufacturers Association (“the CMA”). 9 Joint Managing Director of SENET cc (“SENET”). 10 An independent expert. Managing Director of a consulting firm, MEPS International Ltd (MEPS”). 11 Export Purchasing Manager of Volkswagen South Africa (“VW South Africa”). 12 Senior Vice-President for Product Development and Procurement of Toyota SA. 13 Financial Director, Global Roofing Solutions (Pty) Ltd. 14 Independent Consultant. He was previously the Director General of the DTI. 15 An Independent Expert. 16 An Independent Expert from Johannesburg Economics. 17 General Manager: Technology of Mittal Steel SA. 18 An Independent Consultant of the consultancy group, CRU Strategies. 19 Mittal SA’s Manager: Market Strategy and New Business Development. 20 Mittal SA’s General Manager: Company Controlling. 21 An independent expert, Vice President at CRA International, an economic consultancy group. 22 See our amendment application judgment, Harmony Gold Mining Company Ltd and Another vs Mittal Steel SA Ltd and Another, Tribunal Case Number: 13/CR/FEB04. 4 in the complainants’ gold mines. According to the complainants, the prices at which the merchants sell to the complainants are determined by the prices quoted by the steel mills, to which the merchants add a small trading margin. It appears that the merchants also perform several value-adding functions – for example, cutting the steel received from Mittal into sizes required by their customers – for which they naturally levy a charge. 23 The Respondents

[15]The South African Iron and Steel Corporation (“Iscor”) was incorporated on 5 June 1928 in terms of the Iron and Steel Industry Act 11 of 1928. It was owned by the South African state and was converted into a public company under the Companies Act 61 of 1973 by the Conversion of Iscor Limited Act 7 of 1989. Iscor was selected to lead the then South African government’s privatisation programme, ushering in a new era in the company’s history with its listing on the Johannesburg Securities Exchange (“JSE”) on 8 November 1989.

[16]On 1 March 2001, Iscor announced the restructuring of the company which was completed on the 26 November 2001 with the separate listing on the JSE of Kumba, which contained the mining assets previously owned by Iscor, leaving Iscor as a focused steel company. 24

[17]On 23 November 2001, the Iscor board announced that it had concluded a Business Assistance Agreement (“BAA”) with LNM Holdings B.V. (“LNM”), then the world’s second largest steel producer (with worldwide steel making operations). In addition, LNM bought 34,81% of Iscor’s issued share capital. At the end of November 2002, the IDC held 39 167 364 shares in Iscor, representing 8,79% of the number of Iscor shares in issue. Iscor shareholders approved the BAA on 15 January 2002, that is, at the time of the unbundling of Iscor. 25 The BAA was concluded with a view to 23 See the Pleadings Bundle, page 10, paragraph 10.4. 24 In terms of the restructuring: Iscor would transfer its mining companies and interests to Kumba or subsidiaries of Kumba, save for ownership of 6,25 million tons per annum of iron ore produced by Sishen, which was to be retained by Iscor; the Kumba shares would be distributed to the Iscor shareholders pro rata to their existing holdings in terms of the unbundling legislation; the IDC would inject additional equity into Saldanha Steel equivalent to 50% of its net debt; the IDC’s shareholding in Saldanha Steel would be acquired by Iscor; the IDC would contribute half of the anticipated operating funding requirements of Saldanha Steel for the financial year ending 30 June 2002; and post the Kumba unbundling Iscor would undertake a rights issue of R1,67 billion, to be fully underwritten by the IDC. 25 By the end of April 2002, the R1,67 billion rights issue was successfully completed. 5 assisting Iscor in improving efficiencies and cost-savings. By receiving new technology and skills from a global partner it was believed that Iscor could participate more effectively in the global steel industry. LNM, in terms of the BAA, provided business, technical, purchasing and marketing assistance to Iscor. As part of the BAA, LNM undertook to invest in Iscor shares and in February 2003, LNM increased its shareholding to 47%, following an offer to minority shareholders. In terms of the BAA, Iscor remunerated LNM at the end of 2003 for its technical assistance in ensuring that Iscor achieved the specified threshold cost saving levels. 26 Pursuant to this remuneration LNM, in terms of the BAA, acquired a further shareholding in Iscor, which resulted in LNM holding 50% of Iscor’s issued share capital. It is through this transaction that LNM gained control over Iscor, a transaction which was approved by the Tribunal on 8 June 2004. 27 Subsequent to the Tribunal’s approval of this transaction, Iscor changed its name to Ispat Iscor and then to Mittal Steel SA.

[18]The Mittal multinational subsequently merged its interests with those of Arcelor, then the world’s largest steel producer. Since 16 August 2006 Mittal Steel SA has been controlled by Arcelor Mittal, the world’s leading steel producer. 28

[19]Mittal Steel SA is a listed iron and steel manufacturing company registered and incorporated in terms of the company laws of the RSA. Mittal SA is the primary producer of both long and flat steel products in South Africa with four production facilities, viz., Vanderbijlpark Steel, Saldanha Steel, Newcastle, and Vereeniging Steel. The former two plants – Vanderbijlpark and Saldanha - produce flat finished steel products whilst the latter two plants produce long finished steel products.

[20]The current product range of Vanderbijlpark Works consists of a variety of flat steel products including hot rolled sheet, hot rolled plate, hot rolled strip, cold rolled sheet, electro-galvanised sheet, tinplate, hot dip galvanized sheet and colour coated sheet, all of which are available in a wide variety of sizes and specifications. These products are sold locally and are also exported to global destinations in Europe, the Middle and Far East, North and South America, Canada, Australia and Africa. 26 LNM was initially remunerated in the form of Iscor shares but this was amended in December 2003 to provide for payment in either shares or cash. See Iscor Limited Group report 2003, page 98. 27 See LNM Holdings N V / Iscor Ltd [2004] 2 CPLR 311 (CT). 28 See our decision, Mittal Steel Company N.V. and Arcelor SA, Case No.: 53/LM/Jun06. 6

[21]The Saldanha plant is located at the deep-sea port of Saldanha Bay on the west coast of South Africa and is largely focused on the export market. The plant commissioned its first hot rolled coil (HRC) in late 1998 and is currently producing at its designed nameplate capacity of 1,2 million tonnes per annum. In addition, the plant is distinguished by merging leading edge technologies to produce high quality ultra thin hot rolled coil (UTHRC).

[22]The Saldanha plant was initially controlled by a Joint Venture in which Mittal SA (or Iscor, as it was then known) held a 50% share with the remainder held by the Industrial Development Corporation (‘IDC’), a state-owned financial institution which provides loan and equity capital in support of industrial development. The IDC’s 50% share was purchased by Mittal SA in 2002. This transaction is fully described and assessed in the Tribunal’s previous decision in the merger of Iscor Limited and Saldanha Steel (Pty) Ltd. 29 Suffice for the present to note – and the significance of this observation will become apparent – that during the period in which Saldanha was controlled by the JV, an agreement between Iscor and the JV provided that all Saldanha output was to be exported. In other words, a market sharing agreement provided that Saldanha output would not compete with Iscor in the domestic market for flat steel products. Note too that the Tribunal’s approval of this merger was conditional upon the termination of an arrangement whereby Duferco, a firm also located at Saldanha and which performed certain value-adding functions on flat steel product purchased from Saldanha Steel, also undertook not to market its output in South Africa.

[23]Mittal SA’s website describes its production activities: The flat steel operations at Vanderbijlpark and Saldanha together produce 5.1 million tonnes of liquid steel per annum making it the largest supplier of these commodities in Africa. Vanderbijlpark produces 3.8 million tonnes of liquid steel per annum, which constitutes some 81% of South Africa’s flat steel requirements. Saldanha is one of the world’s most technologically advanced and environmentally friendly steel mills, producing ultra thin hot rolled coil for stringent applications in the domestic and select export markets. The state-of-the-art plant produces 1.2 million tonnes of steel per annum. 29 Case No.: 67/LM/Dec01. 7 The company's Newcastle and Vereeniging operations, services some 50% of the local market for long steel products, while maintaining a firm footing internationally. The two mills account for total annual sales of 1.9 million tonnes, half of which is exported due to the limited demand of the RSA market: 1,57 million tonnes is rolled profile products, 90 000 tonnes is seamless tube and 20 000 tonnes is forged products.”

[24]It concludes that Mittal SA is ‘… the largest steel producer on the African continent, producing 7,3 million tonnes of liquid steel per annum.

[25]The website also notes that Mittal SA is a modern, highly competitive supplier of steel products to the domestic and global markets.

[26]And that it enjoys an industrial presence in 27 countries across Europe, the Americas, Asia and Africa, Arcelor Mittal has a balanced geographic diversity within all the key steel markets, both developing and developed.

[27]It further avers that The company’s ability to generate profits and cash throughout the fluctuations of the steel cycle is testimony to the success of years of intensive business re-engineering and the cultivation of a continuous improvement culture that has embedded Mittal Steel South Africa’s position among the world’s lowest cash cost producers of steel.” 30

[28]Macsteel International, the second Respondent, is a joint venture company owned in equal parts by Mittal SA and Macsteel Holdings (Pty) Ltd (“Macsteel Holdings”). Macsteel International was established in the Netherlands pursuant to an agreement concluded on 27 June 1995 between Mittal SA and Macsteel Holdings. Macsteel International conducts all of the export sales of Mittal SA and deals with 30 See Mittal SA’s website, under ‘company overview’. Last visited on 19 January 2007. 8 other international transactions. 31 Macsteel Holdings also wholly owns a steel merchant that operates in the domestic market.

[29]Note that the joint venture is not confined to trading in Mittal SA’s steel which apparently constitutes roughly half of its business. Nor is the joint venture trivial from Mittal SA’s point of view - approximately 40% of its flat steel is traded through the joint venture. 32

[30]The agreement between the first and second respondents is described and analysed in some considerable detail below. The excessive pricing complaint

[31]The complainants allege that Mittal SA is in contravention of Section 8(a) of the Competition Act by charging an excessive price to South African consumers of its flat steel products. Much of this decision is naturally concerned with an interrogation of the difficult concept of an ‘excessive price’. It is fair to say that our conceptual approach, and so the evidence used to prove or disprove an alleged contravention of Section 8(a), parts company in crucial respects with those of both the complainants and the respondents. It is thus important that we summarise briefly the approaches of the adversaries in what has become a trial of fairly mammoth proportions.

[32]The complainants’ approach has relied upon a series of comparisons of prices in different markets. Hence they have compared the list price for Mittal SA’s flat steel products, the price which the complainants and most other South African consumers of these products are charged, with  Prices charged for the same flat steel products to a select number of Mittal SA’s domestic customers who receive varying degrees of rebate off the list price;  Prices charged for Mittal SA’s long steel products;  Prices charged by Mittal SA for flat steel products to its export customers;  Prices charged by other steel producers of flat steel products in a variety of markets across the world, and with  Mittal SA’s costs of production. 31 See the Amendment Application Bundle, i.e., Mr Peter Jones’ Answering Affidavit, pages 59-64. 32 The precise proportion of Mittal SA steel traded internationally varies with the South African economic cycle and consequent level of domestic demand for steel at the domestic price stipulated by Mittal SA. 9

[33]The complainants have sought to use these comparisons to demonstrate that those South African consumers who are charged the Mittal SA list price pay a price that is relatively excessive in relation to the prices charged to the other purchasers of steel listed above. As may easily be imagined this approach has entailed the presentation of massive quantities of empirical evidence regarding steel prices across the globe and in every conceivable market segment in which flat steel products are consumed. This approach – the use of comparators in other markets – finds echo in a number of decisions of the courts of the European Union and those of its member states. Many of these decisions are referred to below.

[34]Mittal SA, for its part, has not engaged much with the approach of its adversaries and, hence, with much of the voluminous evidence presented in support of the price comparison approach. It has taken a quite different approach to the question of excessive pricing. In essence Mittal SA has argued that a charge of excessive pricing can only be sustained if the complainants can demonstrate that this is reflected in excessive profits. This has entailed a detailed excursion into the complex world of profit measurement, a concept which has different meanings for economists, on the one hand, and, on the other, for the accountants and auditors who are charged with preparing the accounts of companies. We have heard the deeply contending views of a spiraling group of learned academicians and practitioners on, inter alia, the measurement of profit and the cost of capital and on the correct approach to the question of depreciation. This too has entailed the presentation of reams of empirical data.

[35]This, as may be imagined, has given rise to some rather bizarre testimony, with Mittal SA’s expert economist, Dr. Mike Walker, attempting to persuade the Tribunal that, his client, far from profiting excessively from its pricing practices, is, the conventional wisdom of the investment community notwithstanding, a firm in dire commercial straits, indeed is a firm whose very future existence is placed in doubt. This judgment is rendered all the more peculiar because it is contradicted by Mittal SA’s current performance and its own bullish, public assessments of its future prospects. Counsel for the complainants lost little time in pointing out that were Dr. Walker’s criteria to be applied to other companies, most of the blue chip companies listed on the Johannesburg Securities Exchange would suffer from a similarly negative assessment. 10

[36]Moreover, because Dr. Walker rightly conceded at the outset that his contentions regarding profit assumed the ‘efficiency’ of the firm in question – an inefficient firm may charge excessive prices and still not show exceptional profits – we have also had to consider the question of efficiency and its various measurements, also an issue that has necessitated the presentation of volumes of empirical evidence and much conceptual debate. This has also meant that because of Dr. Walker’s concession regarding efficiency, he was forced to argue that his client was simultaneously efficient and commercially unsuccessful.

[37]We, for our part, have taken a quite different view of the question of excessive pricing. We will not attempt a summary of our decision here – that is the subject matter of the pages that follow. We will simply emphasise that, in our view, the arguments of both the complainants and Mittal SA would effectively have the competition authorities adopt, by virtue of Section 8(a), the methodologies of price regulation. This is not our approach. While, as will be seen, we do not shy away from the responsibility imposed on us by Section 8(a) to pass judgment on the pricing practices of monopolies or, what we have termed, ‘super-dominant’ firms, we do so using principles and methodologies firmly rooted in the practice of competition law and economics. Although we have found that Mittal SA is indeed charging excessive prices, and is thereby in contravention of Section 8(a), we have not reached this conclusion by assuming the mantle of a price regulator. Mittal SA’s price setting methodology

[38]We set out below the basis upon which Mittal SA establishes the price that it charges for flat steel products in the domestic market. We note that until 1984 Mittal – or Iscor as it then was- was subject to price control with prices determined on a ‘cost-plus’ basis. From 1984 until about 1992 Iscor’s prices simply followed the domestic inflation rate. Mr. Dednam testified that by 1992 this pricing policy resulted in ‘imports coming into the country’ (presumably because of a relatively high domestic inflation rate) and so from then on the import parity price principle was applied. 33

[39]Mr Dednam further testified in his evidence-in-chief that the import parity price had formed the basis of Mittal SA’s price formation until the end of November 2005 at 33 Mittal SA Heads of Argument, para 11.10 citing Mr. Dednam’s testimony at transcript page 2112. 11 which time the pricing basis changed from import parity price to one based upon a basket of domestic prices prevailing in selected domestic markets. We comment on this claim below.

[40]In brief, since about 1992 and, on Mr. Dednam’s version, until late 2005, Mittal SA had arrived at its domestic price by establishing an FOB price based on one or other European price ( the prevailing Black Sea price was often referred to), adding on the relevant logistical costs of transporting the product to South Africa, such as the shipping, the stevedoring, the handling, and the port costs, as well as a commission of 2.5% onto the price, and an import duty of 5% to the price itself, and finally adding on to that the South African logistical cost for port and railage delivered into the Gauteng region and converting the price from a dollar price to a rand price based on the prevailing exchange rate. It is worth recounting at some length Mr Dednam’s version of this methodology: “MR DEDNAM: The calculation of the international price parity discounts was done as follows. We determining (sic) the FOB global price for a specific commodity and we basically benchmarked 3 basic commodities in the steel range. We benchmark against the prices that we achieve in the international market ourselves. We look at what are the published prices through CRU, Metro Bulletin research and World Steel Dynamics and what they are saying. We are also engaging into an in-house weekly conference call on Mondays where we learn from the other Mittal companies in the group what are the international prices doing in the different regions. So out of the intelligence that we actually gather from all these sources, we arrive at a FOB global price for a specific commodity. We do it for hot rolled coil, for cold rolled coil and for galvanised products. Then we add on the relevant logistical costs such as the shipping, the stevedoring, the handling, and the port cost. When we did these import parity calculations then, we added in a commission of 2.5% onto the price, we added in an import duty of 5% to the price itself. We added onto that the South African logistical cost for port and railage delivered into the Gauteng region. We converted this US Dollar price at the latest exchange spot rate to a Rand price and we compared this price with the actual prices in the pricelist itself 12 and then we determine from that the discounts applicable to the marketplace to reflect the difference in the international price and the price in the pricelist.

[41]In summary, then, in the import parity pricing regime Mittal SA sets its base prices for flat steel products in the domestic market by calculating the notional cost of importing those products. It then adds a 5% ‘hassle factor’, essentially a reflection of the additional costs or ‘hassle’ entailed in importing over the advantage of utilising a domestic supplier. The import parity price is determined monthly by Mittal SA and is conveyed to customers as a discount or surcharge off a list price that is published every three months.

[42]According to Mittal SA IPPD is calculated as follows: 34 MITTAL STEEL SA’S CALCULATION OF IPPD 1. The FOB overseas price is determined for the specific commodity, by looking at a basket of prices, including Iscor’s own export price, import price information and published international prices for the different international regions; 2. Relevant logistical costs, such as shipping, stevedoring, handling and harbour costs are added; 3. Agents commission of 2,5% is added; 4. The SA import duty has historically been added; 5. The South Africa fob costs, such as harbor and railage are then added; 6. This US dollar price is converted with the latest exchange spot rate to a rand price; and 7. The rand price is compared with the list price and the difference is the IPPD.

[43]The complainants allege that as a rule Mittal SA charges its large customers precisely the IPP it has calculated for basic or standard products. In fact the evidence suggests that in certain periods – sometimes quite lengthy periods of some 8 months - Mittal has charged those of its domestic customers who are not members of any of the rebated schemes (discussed later in this decision) above import parity while at other times it has charged slightly less than import parity. 35 The complainants submit 34 See Mittal Steel SA’s heads of argument, pages 80-81, paragraph 11.16. 35 These deviations from the import parity price are partly explained by exchange rate volatility. However, they may also be explained by Mittal SA’s efforts to fine tune its domestic price in order to 13 that the IPP is an artificially established price rather than a price determined through effective competition in the domestic market. We note – and the significance of this will become apparent – customers who received a price below import parity, be this the rate charged in the international market or the rates charged to those who qualify for one or other of the rebate schemes, were contractually prevented from redirecting this discounted product into the higher priced domestic market. At very least they were, before receiving the rebate, obliged to prove that the rebated steel had been used precisely for its intended purpose, largely for exporting or competing against imports, and no other purpose.

[44]In his opening address, Mr Loxton – senior counsel for Mittal SA – indicated that the terrain has changed because “Mittal [SA] no longer employs either import parity pricing, nor even international parity pricing insofar as it resembles import parity pricing and instead has moved to and is moving to as a result of its discussions with government, to a position where it bases its prices upon a basket of domestic prices of net export in countries”. 36

[45]Mr. Dednam further testified: This methodology, as I said just now, changed as we’ve also indicated in December last year where we’ve implemented the basket of domestic prices to be the determinator (sic) for the level of pricing that we are actually doing for the domestic market. And the difference is basically that we look at the domestic prices in comparable countries elsewhere in the world. We look at the absolute price level that we are charging the domestic customers in South achieve a level as close as possible to its profit maximising price in the relevant geographic market – the domestic market – in which it is super-dominant or, expressed otherwise, an effective monopolist. We will show – and this is the core of our argument – that Mittal SA has deployed its super-dominant position to engage in ancillary conduct which effectively allow it to restrict domestic supply, that is, which,enable it to move its domestic supply curve leftwards along a downward sloping demand curve. In other words, given domestic demand, supply is determined by price, rather than price being determined by the supply conditions - by what we will term ‘cognisable competition considerations’ – that prevail in the relevant geographic market, the domestic market. Hence the outcome is a pre- selected target price labelled the’ import parity price’ or, as discussed immediately below, what Mittal SA now claims, is the price determined by compiling the average of a basket of prices in a range of other national domestic markets. The point is that both the import parity price or the basket of international commodities are targeted because of their close approximation to the monopolist’s profit maximising price. 36 See Mr Loxton’s opening address, transcript of 15 March 2006, page 40. 14 Africa. And we point blank put that price at that particular level without taking into consideration any of these notional costs that’s been illustrated in these bullet points over here.” 37

[46]A lengthy debate ensued between the complainants’ counsel and Mr. Dednam regarding this claimed change in Mittal SA’s pricing basis, an argument that, in our view, the complainants had much the better of. 38 Suffice to say that in response to a direct question from the Tribunal regarding the claimed change in the pricing basis Mr. Dednam averred that the new pricing regime had been announced in mid-December 2005 and had been implemented from January 2006. The hollowness of this claim was thoroughly exposed under cross-examination - indeed it appears that it had not been part of the announcement of 15 December 2005. On further examination and questioning from the Tribunal Mr. Dednam acknowledged that the claimed shift in the pricing basis had had no discernible impact on the actual price charged. Indeed it is disappointing that a witness who we generally found to be helpful and, on some important points, candid – we note particularly his honest responses to the role of the anti-arbitrage provisions in Mittal SA’s agreements with its discount customers and its export merchant, Macsteel International – was prepared to blatantly mislead the Tribunal on this point. That it appears to form part of Mittal SA’s ‘offer’ to the Department of Trade and Industry suggests that it is willing to mislead the public as well.

[47]In any event the point is of no great moment. The argument that will be developed in this decision holds that a non-excessive price is one that is determined by competitive conditions in the relevant market. The manner in which the IPPD pricing basis works is to determine the price of flat steel products in South Africa by reference to demand and supply conditions that prevail in an arbitrarily selected market abroad (for example, the ‘Black Sea price’) markets and then to add to that price the notional costs of ‘importing’ the product to South Africa. The ‘basket’ approach that Mittal SA now claims to have adopted effectively uses demand and supply conditions – that is, competitive conditions - in an arbitrary array of other selected national markets to determine prices in the South African domestic market. It falls foul then of the same argument that we will use to condemn the targeting of import parity as the basis for setting the domestic price. We have no idea of what 37 See transcript, pages 1653-1654. See also Mittal SA’s heads of argument, pages 80-81. 38 See Mr Dednam’s evidence-in-chief, transcript of 5 April 2006, page 1654. See transcript of Mr. Dednam’s cross examination on this point from page 1744 to 1889. 15 competitive conditions prevail in the arbitrarily selected and diverse range of countries that Mittal SA claims to place in its basket. Suffice to say that it is a very peculiar way of settling on a price in our market which, we will insist, must, in order to be non-excessive, be set by reference to competitive conditions in the relevant market which is the South African market for flat steel products. As we will show the key competitive conditions in our market are Mittal SA’s structural super-dominance plus ancillary conduct aimed at maintaining the segmentation of differently priced markets, the cumulative effect of which is to produce a price that is not influenced by any competition considerations whatsoever and is, because of this, adjudged to be excessive. The relevant market

[48]The complainants contend that the relevant product market is that for flat steel products.

[49]Mittal SA has been less clear in its identification of the relevant product market. At the outset of the case, Mr Dednam argued that defining the relevant product market as that for primary flat steel products is an oversimplification. 39 He averred that the relevant product market for the purposes of this complaint is for steel products utilised in the gold mining sector. He argued that within the broad category of flat steel products Mittal SA produces thousands of different products, which for purposes of this complaint, are limited to seven different broad classes, namely, slabs, plates, hot rolled, cold rolled, galvanized, tinplate, and colour coated. 40 He averred that each of these seven categories predominantly attracted different buyers from different industries with Mittal SA’s customers in the mining industry purchasing plates and hot rolled steel. He further claimed that in each of the broad categories to which he referred there is a different degree of beneficiation and value-add. He argued that in the different categories different possibilities of substitution apply. For example, galvanized steel supplied to the building industry competes with, inter alia, roofing tiles as a possible substitute. For hoppers and skips utilised in the mining industry, the stainless steel product known as 3CR12 or aluminium could be a substitute. He contended that Mittal SA monitors its sales to specific industries, and 39 See page 132, 145-147 of the Pleadings Bundle, paragraph 1.6. and 6.4 of Mr Dednam’s answering affidavit. 40 See Annexure “CD2”, page 216 of the Pleadings Bundle. 16 takes cognizance of signs of declining sales due to competition from the use of substitutes or the importation of steel.

[50]Harmony however averred that there is a basic distinction at the production stage between flat and long steel products, which are typically produced in different types of plants. 41 Mittal SA recognises this distinction in a number of its internal documents, including in documents dealing with pricing policies and sales reports and when it reports its annual results.

[51]With the exception of the arguments advanced by Mr Dednam, nowhere did Mittal SA or any of its witnesses challenge the classification of the product market as that for flat steel products. Indeed much of Mittal SA’s analysis appears to be premised on the existence of a market for flat steel products. Mr Dednam himself confirmed this distinction in his evidence-in-chief when he testified about the production process of flat steel products at Vanderbijlpark. 42 Dr Mike Walker, an expert witness called by Mittal SA, based his analysis on a market for flat steel products. 43 Mr Tomlinson – one of Mittal SA’s witnesses - referred frequently in his evidence-in-chief to flat steel products as the relevant product category when assessing Mittal SA’s prices. His testimony relied on defining flat steel - with hot rolled coil as the base product – as a relevant product market for a number of conclusions he sought to draw. When discussing pricing Mr Tomlinson noted: “The prices I will quote today are for hot rolled coil. Flat products account for around 50% of world production of all steel products, the remainder being divided amongst so-called long products such as concrete reinforcing bar, wire rod and merchant bar and structurals. Of that 50%, 40% of the total is in former sheet products, which pass through a hot rolled coil production stage”. 44 41 As we have already pointed out Mittal SA produces flat steel products at Vanderbijlpark and Saldanha whilst its long steel products are produced at Newcastle and Vereeniging. See also, Prof. Roberts’ Interim Report, pages 15-16. 42 See transcript, pages 1641-1647. 43 See Dr Walker’s Final Report, at 224-240,

[53]– [78]. 44 See transcript, pages 1470-1471. 17

[52]Dr. Simon Roberts, an economic expert retained by Harmony, testified that while a number of different end products may be made from flat steel products there is extensive supply side substitutability. 45 This was not challenged by Mittal SA.

[53]Mittal SA’s counsel did not indicate in his opening address that the product market was in dispute. Nor is this indicated in Mittal SA’s heads of argument. Indeed in Mittal SA’s ‘concise heads’ handed up on the first day of argument, there is repeated reference to the product market for flat steel products. In emphasising Mittal SA’s position on the geographic market, counsel stated . “It is clear that the conditions prevailing in one regional market, including price, will affect other regional markets. Consequently the answer to the question what the relevant market is for Mittal’s flat steel products is: it depends upon the prevailing domestic and regional and upon the domestic demand/supply equation, but it is not an exclusively South African market. It is certainly not clear that it will be so in the future”. 46

[54]In our view the complainants have correctly identified the relevant product market as that for flat steel products.

[55]There is however clear contention surrounding the identification of the relevant geographic market. Harmony contends for a national geographic market. While Mittal SA has generally avoided pinning its colours to any clearly delineated geographic market – its expert simply decried, for reasons that are elaborated below, any attempt to delineate the relevant geographic market as a ‘mugs game’ – its insistence, which is a cornerstone of its case, that the prospect of import competition constrains its pricing power, suggests that it views the relevant geographic market as an international market. However this implicit identification of the relevant market is substantially undermined by Mittal SA’s own description of its South African market as one that is ‘naturally protected’ – and by this is meant protection by dint of its distance from competing producers of steel and the high cost of transportation and is thus geographic in nature – and by explicit concessions in the heads of argument of Mittal SA’s counsel which concede Mittal SA’s market power. 47 45 See Prof. Roberts’ testimony, transcript of 30 March 2006, pages 1122-1124. 46 See Mittal SA’s Concise Heads, page 18, paragraph 3.14. Our emphasis. 47 Mittal HOA para 2.7 ‘While Mittal may have market power in an economic sense…’ Note too Mr Dednam’s concession to the effect that Mittal enjoys a ‘naturally protected’ domestic market. See the transcript of 6 April 2006, page 1810. 18

[56]The reason why Dr. Walker, Mittal SA’s expert, refused to get drawn into the ‘mugs game’ of defining the relevant geographic market for flat steel products in a situation – which no-one contests – where a regional (that is, a South African) monopoly operates is, of course, because of the operation of the ‘cellophane fallacy’. This important analytical contribution to the identification of relevant markets is outlined with characteristic clarity in the book co-authored by Simon Bishop and Dr. Walker himself. 48 In essence the application of the hypothetical monopolist test – the standard test utilised in identifying anti-trust markets in merger cases – is substantially complicated when applied to abuse of dominance cases because the dominant firm in question may already be charging the monopoly price. Hence it is widely accepted that in the case of United States v E.I. du Pont de Nemours and Co. 49 the US Supreme Court erred in basing its judgement of the relevant market on evidence submitted by the respondent purporting to show that cellophane competed with other flexible packaging material because were the price of cellophane to be increased beyond the prevailing price it would be substituted for by other packaging materials. As Bishop and Walker explain: The key implication of the cellophane fallacy is that the identification of substitutes at existing prices does not necessarily identify those products that are effective substitutes at the competitive price, which is the relevant benchmark for defining markets in most non-merger cases. Evidence that products are effective substitutes at current prices merely identifies those competitors that constrain the prices of the firm or firms under investigation from increasing above the current level. It does not necessarily provide information on whether those products are constraining prices to the competitive level. 50

[57]Bishop and Walker cite the relevant notice of the European Commission: Generally and particularly for the analysis of merger cases, the price to take into account will be the prevailing market price. This might not be the case where the prevailing prices have been determined in the absence of sufficient 48 Simon Bishop and Mike Walker The Economics of EC Competition Law (Sweet and Maxwell, 2002). 49 1956 351 U.S. 377; 76 S. Ct 994 50 Bishop and Walker para 4.37, page 99 (our emphasis). 19 competition. In particular for investigation of abuses of dominant positions, the fact that the prevailing price might already have been substantially increased will be taken into account. 51

[58]Bishop and Walker note that, while the cellophane fallacy does not completely eliminate the value of the hypothetical monopolist test, in these cases: Certainly, it would be incorrect to ignore the implications of the cellophane fallacy when defining markets in non-merger cases since this will tend to lead to markets being defined too widely. 52

[59]Of course in this instance the consequence of ignoring the implications of the cellophane fallacy is to widen the market from a national to an international market. The arguments for ignoring the cellophane fallacy are frankly risible. Consider the following:  The scale of transport costs involved in the importation of a commodity like steel. One estimate is that transport costs may constitute as much as 47% of the cost of flat steel products imported into South Africa; 53  That there is no ‘international’ price for steel. While the existence of an international market does not depend on the existence of a single quoted world price, the regional variations are, as conceded by Mr. Tomlinson for Mittal SA, ‘very considerable’. 54 Indeed it seems that the most that can be said is that the prices of steel in the various geographic markets tend to ‘harden’ or ‘soften’ in tandem with each other; 55 51 Bishop and Walker para 4.39, page 100 (emphasis in the original). 52 Bishop and Walker para 4.40, page 101. 53 See Harmony Heads of Argument, para 85. Reference is also made to Mr. Tomlinson’s concession that for flat steel products ‘transport costs are relatively high in relation to value.’ We note that in the ISCOR/Saldanha Steel merger the merging parties argued that the scale of transport costs from the west coast of South Africa to its inland industrial heartland would even prevent Saldanha Steel from competing with Iscor in the lion’s share of the domestic market although it still chose to ensure this by means of an agreement that prevented Saldanha Steel from competing in the domestic market. 54 Transcript, page 1557. 55 We note parenthetically that much of the recent hardening in the steel market and the forecast that this will continue into the future is a consequence of the escalating price of iron ore, from which Mittal SA is largely immunised thanks to its agreement with Kumba. 20  The fact that there are significant periods in which Mittal SA has, managed to sustain prices above the import parity price without encountering significant importation of steel; 56  That in its calculation of its import parity price Mittal SA explicitly incorporates a 5% premium – which recall is the bottom limit used by the US Department of Justice in its application of the hypothetical monopolist test - for the ‘hassle’ of importing, thereby acknowledging in the most material of terms that it is able to raise its price by at least as much as 5% above what it claims to be the competitive price without encountering a threat from its claimed competitors;  That there is no evidence of importation of steel on any discernible scale except in the case of products not produced by Mittal SA. Where a South African firm, Bell Equipment, did attempt to engage in importation of steel, Mittal SA dropped its price after Bell had secured, as it was obliged to, a significant quantity of steel, not so much as an attempt to meet the import price, but as an apparently punitive or retributive – Mittal SA’s counsel described it as ‘spiteful’ - response to anyone who had the temerity to consider importation; 57  That customers consistently testified that but for the most exceptional sustained increase in prices above the import price they would not consider importation of flat steel products; 58  That Mittal SA itself acknowledged that the considerable matter of exchange rate volatility and cash drain posed ‘difficulties’ for importers; Notably, GRS already imports product. It has an established relationship with a foreign producer. While there are undoubtedly difficulties associated with importing, such as the fluctuating exchange rate and the cash drain associated with importing, we submit that Mr. Jacobsen’s evidence was ultimately not that imports were not an option for his company. 59 56 See in this regard, Mr Dednam’s concession on this point during cross-examination by Harmony Counsel. That is, transcript of 10 April 2006, page 2043. See also, Professor Roberts’ final report of 3 March 2006, paginated page 105. 57 See transcript of 17 March 2006, page 326. See also Mr Gotz’s cross examination of Mr Bell, transcript page 334. 58 For example, see Mr Bell’s evidence-in-chief, page 317. 59 Mittal SA Heads of Argument, para 11.32 (our emphasis). These admitted difficulties are presumably the effective basis for the 5% ‘hassle factor’ added to the import parity price formulation. 21  That South African importers of steel would inevitably be small customers of very large producers located in distant markets whose reliability as suppliers of an absolutely vital input would therefore be questionable. 60  That although several of the witnesses at these hearings testified that, if requested, local agents were willing to procure steel on the international market for importation into South Africa, we also note that 3 out of 7 of Mittal SA’s largest customers are steel merchants and we observe that it would be a brave local trading agent who chose to stand up to a Mittal SA intent upon preventing importation of competing products into South Africa. 61

[60]It is indeed remarkable how frequently the efforts of Mittal SA to establish the ease with which local fabricators are able to import portray precisely the converse. A local steel fabricator, Mr. Stephen Leatherbarrow testified: We don’t see any reason to import steel. We can buy a good quality product, at a good price, well not a good price but a price that we can pass onto our customers and the delivery is fine. So there’s no need for us to look at importing raw material. 62

[61]Given this we understand perfectly well why Dr. Walker would not have wished to compromise his expertise and independence by actually arguing for an international rather than a national market. If the notion of a relevant geographic market is to have any meaning in anti-trust, then the market in which the complainants and the vast majority of other consumers procure flat steel products is 60 Note Mittal SA’s averment that ‘most of the firms that export flat steel, or could plausibly export flat steel, to South Africa are large companies for whom exports to South Africa could only ever account for a small proportion of their production.’ (Para 11.33). While this appears intended to portray easy access on the part of South African steel consumers to international steel supplies it of course simply serves to portray the precise opposite, viz., that these South African purchasers would be minute players in the international market and their custom and hence requirements will figure very small in the calculations and custom of steel surplus countries and firms. 61 Reference was made in the papers filed in the hostile merger between Mittal/Arcelor to ‘exclusivity agreements that exist between locally based steel mills and the three largest steel merchants in South Africa (being Macsteel, Trident and Kulungile)’. These agreements were cited by Arcelor in order to support their contention that it was difficult for Arcelor to serve South African customers from their plants located overseas. When the hearing took place the merger had become friendly, that is, Arcelor was no longer opposed to merging with Mittal. Its attorneys then denied the veracity of the statements that they themselves had made. We did not have to get to the bottom of this and the matter was not raised in these hearings and so is accorded no weight in this decision. We simply observe that in a market where, as we shall elaborate below, the imposition of conditions by the steel producer on merchants is a common occurrence, the possibility that traders would easily become import conduits in the face of a resistant Mittal SA would have to be investigated carefully before much reliance was placed on it either way. 62 Mittal SA Heads of Argument, para 12.8 (our emphasis). 22 the South African market. 63 The import parity price is not the competitive price in this geographic market. It is simply the price that Mittal SA has selected because of its close approximation to its profit maximising monopolist’s price. However, as we will show, in order to realise this pre-selected price Mittal SA is forced to deploy its super- dominance in order to restrict the supply of domestically produced steel to its domestic market.

[62]It is then difficult to disagree with Harmony’s counsel’s summary of the evidence adduced by Mittal SA concerning the constraining influence of imports: So, we do submit that it’s very difficult on a conspectus of the facts of this case to look at the fragments of evidence around imports, simply to suggest that that is really going to discipline the market, because all that we are really examining is one or other permutations of the cellophane fallacy. And therefore we suggest and submit to you that as far as imports are concerned, they are not a significant constraint upon the pricing power of Mittal. 64

[63]It is our view then that the geographic market in which the complainant, Harmony, engages with Mittal SA is indeed the national South African market for flat steel products, the market in which a great many of its customers meet Mittal SA and in which its pricing power is effectively unconstrained by any competing suppliers, either in another country or from a product that could substitute for steel. This is consistent with the decisions made by the panels of the Tribunal on several previous occasions.

[64]These are conceivably not the only markets in which Mittal SA participates as a supplier of flat steel products. There are particular uses, and this will be further elaborated below, in which flat steel products are substitutable by other products – for example, it appears that plastic is substitutable for metal in the production of cans for liquid products. And there are certain purchasers of flat steel products who are 63 As we note below, there are distinct segments in the domestic market because in selected applications steel is substitutable by other products such as aluminium, cement and plastics. Moreover, in one market – the auto manufacturing market – it is alleged that the existence of international supply chains enables the auto multinationals to import finished steel-intensive components for local assembly. As we shall elaborate at length these are the market segment to which Mittal SA grants varying degrees of rebate off the IPPD list price. Effectively these markets are delineated by demand elasticities somewhat greater than that applicable in respect of the majority of South African steel consumers who pay Mittal SA’s full target price. We examine below, in considerable detail, the mechanisms employed by Mittal SA to maintain the segmentation of its various markets. 64 Transcript, page 2344. 23 allegedly capable of sourcing, without incurring significantly onerous transactions costs, the end product for which steel is an essential input from producers in other geographies – for example, the auto industry with its well established international network of suppliers is allegedly capable of sourcing its steel panels from suppliers located elsewhere.

[65]In the first instance cited – the case of metal cans – the product market may, on closer examination, be that for materials used in the production of containers for liquid products and it may include producers of flat steel products as well as certain plastic products. In the latter case, the product market may be that for the supply of flat steel products to the auto industry and the geographic market may extend beyond South Africa’s borders because, of the relative ease with which South African based auto producers may allegedly source from alternative, non-South African based suppliers of auto panels. 65

[66]That Mittal SA is confronted by distinctive structural conditions in these market segments is suggested by the fact that it is precisely customers in these markets who receive discounts off Mittal SA’s list price.

[67]It may well be established that while Mittal SA is dominant as per the Act’s definition within these market segments – as we shall elaborate, the manner in which it establishes its discounts is certainly strongly suggestive of considerable market power – the degree of its dominance in these sub-markets may fall short of the exceptional degree of dominance that we consider necessary to engage in excessive pricing. We will however argue later that, for the purposes of this decision, the interest in these market segments derives from the lengths to which Mittal SA is 65 Note however that in the Trident Steel/Dorbyl Limited merger, which was precisely concerned with the supply of steel panels to the auto OEMs, the Tribunal, which explicitly considered the ability of local auto assemblers to source through their international supply chains and the impact that this factor had on the geographic market, cast considerable doubt on the Commission and the merging parties’ contention that this factor rendered this an international market. It noted (1) Imports were not always clearly substitutable. (2) Customer preference based on considerations other than price influence the degree of substitutability between domestic and foreign supply. (3) Considerable customer scepticism regarding the potential for foreign competition to constrain domestic producers. (4) The barriers to entry created by tariff and incentive schemes undermine the competitive position of foreign competitors. 24 prepared to go to maintain the segmentation, to immunise from the general domestic market the consequences of any discount granted to these producers.

[68]We note that in the present matter it has never been suggested that the complainant or any other customers, other than those that receive rebates, are capable of substituting their steel purchased from Mittal SA with alternative product. The insistence that steel imports restrain Mittal SA’s pricing is only at the point where Mittal SA’s domestic price exceeds the landed price of imported steel in South Africa. If Mittal SA’s domestic price were, for a sustained period, to exceed, by a significant margin, the landed cost of imported flat steel products plus ‘add-ons’ like the 5% ‘hassle’ factor, then, and only then, would the incentive to import become a realistic one for domestic consumers of these products. This is, in effect, the point at which the proverbial customers of cellophane would consider substituting other packaging materials. It is the sole basis for Mittal SA’s contention that despite the ‘natural protection’ which it concedes that it enjoys, the geographic market is a seamless, borderless international market.

[69]Note too that it appears that the share of Mittal SA’s sales of flat steel products that are sold at the list price significantly exceeds sales at rebated prices. The panel’s approach to allegations of excessive pricing

[70]There are few practices condemned by the Competition Act in terms as unambiguous as that identified in Section 8(a) which, in language of crystal clarity, provides that It is prohibited for a dominant firm to – (a) Charge an excessive price to the detriment of consumers

[71]An overly fastidious defence counsel may wish to make something of the subordinate phrase ‘to the detriment of consumers’ though none have attempted to do so here. What, after all, could more clearly inure to the detriment of consumers than an ‘excessive price’? We will, without further consideration, as, implicitly, have the defence counsel, treat this phrase as simply a superfluous description of an excessive price rather than a qualifier of its likely effects. 25

[72]Although, as shall be elaborated at length, the hurdles, particularly regarding the extent of dominance, that must be cleared by a complainant in order to prove excessive pricing are, in our view, exceptional, the repugnance attached to this offence is reinforced by the fact that an administrative penalty can be levied for a first time contravention.

[73]However, although this is frequently misunderstood by the broad public which, rightly, views excessive prices as the most likely and egregious consequence of monopoly, the theory and practice of competition law and economics is dominated by an equally unambiguous maxim that asserts that the task of a competition regulator does not extend to the determination and fixing of prices.

[74]The reluctance of competition practitioners to assume a price regulating function does not only derive from the truly massive technical difficulties entailed in determining the ‘right’ or, for that matter, the ‘wrong’ price, but from the founding principle underpinning the world view of the practice of competition law and economics that holds that price determination is best left to the interplay of independent actors engaging with each other in the market place. The fundamental task of competition regulators is then to promote and defend competitive market structures and to guard against conduct on the part of market participants which seeks to undermine the promise of those competitive structures to deliver quality goods and services at competitive prices.

[75]Core to competition enforcement is the recognition that the promise held out by competitively structured markets may be denied by co-operation between notional competitors. It is additionally recognised that a number of factors ranging from the acquisition of market share by pro-competitive means through to past or present governmental support and subsidy, may result in single firm domination of markets. Faced by single firm domination the principal function of competition enforcers is to guard against ‘exclusionary conduct’, that is, unilateral conduct of the dominant firm that has as its objective the reproduction of this dominance through the exclusion of actual or would-be competitors from the market.

[76]Price determination is thus not characteristically part of the armoury of competition enforcement. And yet Section 8(a)’s proscription of the charging of an excessive price appears, on the face of it, to assign us a role that precisely requires us to determine whether existing price levels are ‘right’ or ‘wrong’ (non-excessive or 26 excessive) and, if ‘wrong’ (excessive) to determine and impose the ‘right’ (non- excessive) price.

[77]Confronted by these two unambiguous, but manifestly contradictory, requirements – the one which appears to condemn a particular price level and require us to impose another lower price, the other which resists the administrative determination of a price – one might predict that attempts to enforce Section 8(a) would immediately run into serious conceptual difficulties. And this is indeed the case. Notwithstanding the linguistic clarity of Section 8(a), the interface of price regulation with the approaches and principles of competition law and economics is complex, to say the least.

[78]Although these difficulties are widely acknowledged, they do not permit us to ignore a clear legislative injunction against excessive pricing. As the United Kingdom Competition Appeals Tribunal (CAT) observed when it too was faced with adjudicating an excessive pricing allegation ‘the fact that the exercise may be difficult is not, however, a reason for not attempting it.’ 66

[79]But by the same token, nor, however, does the prohibition of excessive pricing in a competition statute provide the competition authority with the license – or the powers and resources – to convert itself into a regulator of steel or any other prices. If excessive pricing is to be identified and remedied by a competition authority rather than a duly empowered and appropriately resourced price regulator, then it must do so by recourse to its standard approaches and instruments.

[80]The proscription of excessive pricing is but one of the abuses of dominance described in the Act. Though often distinguished in the anti-trust literature as an ‘exploitative’ abuse by contrast with the ‘exclusionary’ abuses otherwise described in Section 8 of the Act, the requirement to enforce the proscription of excessive pricing is not accompanied by the sort of powers and practices normally associated with price regulation. If the legislature had intended Section 8(a) to convert a competition authority into a price regulator then it would surely have provided us with the powers and resources appropriate to that considerable task. Consider the process by which the sector regulators – each with their own statutory foundation and specialist powers and resources – determine and police pricing in the telecommunications and 66 See Napp Pharmaceutical Holdings Limited and Subsidiaries and Director General of Fair Trading, Case No. 1001/1/1/01, paragraph 392. 27 electricity markets and then consider whether the legislature can possibly have intended that this be replicated in the steel or any other industry by way of the insertion of a single nine word clause in the Competition Act. This cannot be so and this is why we insist that our approach to Section 8(a) allegations should employ the analytical framework and instruments that govern competition enforcement generally.

[81]The standard approaches and instruments of competition enforcement comprise interventions in the structure of the affected markets and in the conduct of its participants so as to produce outcomes that are, as far as possible, unsullied by the possession or, rather, the abuse, of market power. As already noted, there are compelling conceptual and practical reasons why a competition authority should eschew a price regulation role and if it is possible – and we believe in this instance it is - to prove and remedy excessive pricing without resort to the methodologies of price regulation, then this is the approach that must be favoured. 67

[82]It is a fair generalisation that in that branch of competition law and practice where the competition authority functions as an ex ante regulator – namely merger control – the preferred remedies in the face of a likely lessening of competition are structural although conduct remedies are, on occasion, employed when the decision maker is persuaded that they are sufficient in order to protect and promote pro- competitive outcomes. By contrast, in that branch of competition law where the enforcement and adjudicative powers of the competition authorities are invoked ex post – namely in the event of conduct or ‘restrictive practices’ that produce anti- competitive outcomes – the remedies most commonly employed seek to regulate behaviour except in those instances in which only a structural intervention is thought capable of producing the desired outcomes. It is instructive to note that although our 67 A competition authority may conceivably be called upon to act as a as a price regulator in instances that may be characterized as price ‘gouging’. For example were Section 8(a) to be invoked in the event of a natural disaster which had given rise to a temporary monopoly in some or other unregulated product or service that was vital to the life of the affected community, say ambulance services or fuel for heating, and this was exploited to effect a significant temporary price rise, the competition authority could easily assume the role of a temporary price setter. This would not only demand urgent action but it would also be a relatively simple technical task – the excess would simply be determined by reference to the price that prevailed immediately prior to the disaster and the ‘non-excessive’ price would be set accordingly. However where excessive pricing is alleged to flow from a systematic and systemic abuse of dominance in a complex market, the price setting task becomes infinitely more complex and unsuited to the powers and resources of a competition authority. We stress that where, in these latter cases, it is possible to isolate underlying structural conditions and ancillary behavior that enables the setting of a price in excess of that which would prevail in the absence of that anti- competitive structure and conduct, then, in the first instance at least, this is what should be addressed by the competition authority. Of course if a competition authority is not able to carry out its excessive pricing mandate in this manner then it may have to resort to the fixing of a price but, we stress, in our view this should only be done as a final resort. 28 power to impose structural remedies in order to cure anti-competitive conduct is generally limited, a proven allegation of excessive pricing is one of the few instances in which we are empowered to impose structural remedies in the case of a first offence.

[83]How then do we, as a competition authority, approach this allegation of excessive pricing?

[84]If we are to approach this allegation in the manner of a competition authority, we must first ask ourselves whether the structure of the market in question enables those who participate in it to charge excessive prices. As we will indicate, we believe this to be a significantly higher hurdle than those that must be cleared in order to establish ‘mere’ dominance. It requires ‘super-dominance’, a structural condition the characteristics of which are elaborated below. If that higher hurdle is cleared, we must then ask ourselves whether Mittal SA has engaged in conduct designed to take advantage of – to ‘abuse’ – those structural opportunities by imposing excessive prices on its customers. If the second question is also answered in the affirmative, the excessive pricing must be proscribed by imposing a remedy which addresses the underlying structural basis for the offending conduct and/or any ancillary conduct arising from the structural advantage that enables the firm in question to charge a price in excess of that which would have prevailed in the absence of the anti- competitive structure and/or the ancillary conduct. 68 Only if both forms of these remedies are impossible to devise should an actual price level be specified. In short, we treat excessive pricing as a phenomenon that may arise from a particular structure and that itself may be the basis for ancillary conduct that is utilised in order to sustain supra-competitive prices, to sustain, as per the definition of the Act, ‘a price for a good or service which (aa) bears no reasonable relation to the economic value of that good or service; and (bb) is higher than the value referred to in subparagraph (aa)’. 69

[85]This definition appears to be drawn directly from the court’s decision in the leading European case of United Brands. 70 Its elements are dissected in considerable detail later in this decision. 68 As will be elaborated at length, in this case Mittal SA does rely on its super-dominant structural position as well as on ancillary conduct in setting the price that it charges. 69 Competition Act Section 1(1)(ix). 29

[86]However, for present purposes, we note the considerable similarities between the manner in which the problem of excessive pricing is characterised in United Brands, on the one hand, and, on the other, our approach outlined above. In the oft- cited words of the court in United Brands: It is advisable therefore to ascertain whether the dominant undertaking has made use of the opportunities arising out of its dominant position in such a way as to reap trading benefits which it would not have reaped if there had been normal and sufficiently effective competition. 71

[87]This – the cumulative impact of structure and practices - is the standard approach for dealing with allegations of abuse of dominance. However, although a species of abuse of dominance, this approach has, in certain critical aspects, not always been carefully followed through by the European competition authorities when dealing with allegations of excessive pricing, particularly, although not exclusively, in the manner in which the Europeans approach the question of remedying excessive pricing. Faced by allegations of excessive pricing the European authorities, particularly the various national authorities, have, possibly because of their overriding emphasis on the creation of a single market and their concomitant focus on the elimination of intra-European price discrimination, too readily assumed the role of price regulator both in their analysis of the very existence of excessive pricing, but particularly in the remedies that they have constructed in order to cure it. This price regulation methodology has effectively been advocated by Mittal SA, the first respondent in this matter, who appears, at one stage, to insist that we are somehow obliged to follow European jurisprudence. 72 The tenor of much of the complainant’s evidence also effectively assumes that we will take on the methodologies and role of a price regulator although they – the complainants – do evidence a considerably deeper appreciation and elaboration of the underlying structural conditions and ancillary conduct that underpin excessive pricing. 70 See United Brands Company and United Brands Continentaal BV v Commission of the European Communities [1978] 1 CMLR 429, at paragraph 250 where the Court held that “…charging a price which is excessive because it has no reasonable relation to the economic value of the product would be such an abuse”. 71 Ibid para 249 72 By this we do not, of course, mean that Mittal SA has actually asked for its prices to be regulated. However, the approach that it advocates we that we adopt in deciding whether or not its prices are excessive is effectively a methodology adopted by price regulators. 30

[88]In consequence of the assumption – by the complainant and, though to a lesser extent, the respondent alike – that we will follow Europe and adopt the methodologies of price regulation, we have been presented with reams of evidence of prices in a myriad of markets and of costs of production in the manufacture of steel in every corner of the globe. We have been introduced to the arcane – and thoroughly unresolved – debates surrounding the question of profitability and its measure. This mimics the approach of litigants before the European authorities and courts who, when confronted by allegations of excessive pricing, have been willing to engage with this evidence precisely because their competition enforcers and adjudicators have been willing to assume the methodologies of price regulation.

[89]However, for the reasons outlined above, we eschew the role of price regulator, and so the vast quantum of the evidence and much of the argument submitted to us is simply irrelevant. This is not to suggest that our enquiry is any less fact-based than the approach of those competition authorities willing, if not necessarily able, to transform themselves into effective price regulators. However, the facts with which we engage go to the question of structure and conduct, traditional fare in the working life of competition enforcers and adjudicators. They do not go to questions of price regulation which effectively use analytical tools intended to simulate competition in order to arrive at the outcomes that that idealised state would putatively dictate. In short our response to proven allegations of excessive pricing is, wherever possible, to remove those structural and behavioural conditions that inhibit competition and so generate excessive prices; it is not, in contrast with the approach taken in many of the European cases, to simulate an alternative structure and then to impose outcomes associated with that ‘virtual’ alternative. 73 Dominance

[90]The charging of an excessive price is a contravention of the Act when it is levied by a dominant firm. 73 Not least of the difficulties of assuming a price regulator’s role, is that if the price is determined without intervening in either the underlying structural conditions and the ancillary conduct which cumulatively gives rise to the excessive price, the competition authority will have to maintain its regulatory role because the administratively determined price cannot be ‘right’ for all time. This, of course, precisely describes the modus operandi of an ex ante price regulator but is antithetical to that of an ex post regulator that responds to conduct that is allegedly in breach of a statutory obligation. 31

[91]The criteria for establishing dominance are stipulated in Section 7 of the Act which provides: A firm is dominant in a market if – (a) it has at least 45% of that market; (b) it has at least 35% but less than 45%, of that market, unless it can show that it does not have market power; or (c) it has less than 35% of that market but has market power

[92]The section 7 enquiry requires a definition of the relevant market and then the computation of market shares. If the market share of the respondent is found to be less than 45% - the level at which dominance is presumed – we must ask whether the respondent is nevertheless dominant by virtue of its possession of market power which is, in turn, defined by Section 1(1)(xiv) of the Act as the ‘the power of a firm to control prices, or to exclude competition or to behave to an appreciable extent independently of its competitors, customers or suppliers’.

[93]The identification of the relevant market and the determination of market shares and the presence or absence of market power, though often complex and controversial, is standard fare in competition analysis and is aided by a wealth of international and domestic jurisprudence and scholarship. However the importance of these enquiries should not be underestimated. In particular, they are critical in determining the extent of dominance, in our view a critical factor in the assessment of the complaint presently before us. As we shall elaborate below, while a firm that is presumed dominant by virtue of a market share that exceeds 45% presumptively also possesses market power and hence the ability to control prices, a dominant firm’s ability to unilaterally fix a price, its market power, is rarely devoid of all constraint. Market power, in other words, is seldom absolute. However, these limitations on market power do not prevent a dominant firm from enjoying and exercising a degree of pricing power, power, that is, that enables it to price at a level above that which would prevail in a perfectly competitive equilibrium.

[94]The Act clearly contemplates the continued existence of dominant firms, of firms possessed of market power. And thus, by extension, it contemplates a measure 32 of private pricing power in those markets in which there is a dominant firm. Indeed, the ability of a dominant firm to exercise pricing power is the critical incentive to compete robustly for market share. Hence interference with that ability, with that power, runs the considerable risk of restraining competition. This consideration then accounts for anti-trust’s pre-occupation with ‘exclusionary conduct’, conduct that is designed to preserve dominance by securing the exclusion of competitors through mechanisms other than ‘competition on the merits’. Hence we must, at the outset, agree with Mittal SA that, because the Act clearly contemplates the continued existence of dominant firms, it also contemplates the existence of the pricing power that accompanies dominance. There is, in other words, statutory recognition of the existence and legitimacy of prices that exceed marginal cost, that point that would manifest the complete absence of pricing power. 74

[95]However while the Act clearly contemplates the existence of pricing power, it effectively proscribes the exercise of ‘excessive’ pricing power as manifest in the clear prohibition of the charging of ‘excessive prices’. But while Section 7 and the relevant statutory definitions specify the conditions under which a firm is (permissibly) dominant and thus possessed of a degree of pricing power, it is silent on the standard that would permit us to identify ‘excessive’ dominance which, following the Act’s schema for dealing with exclusionary conduct, would produce ‘excessive’ pricing power and the possibility of an ‘excessive price’. This requires a close examination of the structure of the market and an examination of the scholarly literature that has grappled with the question of excessive pricing.

[96]In summary then our approach is to follow the schema of the Act and the standard approach to allegations of abuse of dominance which, as we have seen, derives dominance from specified market shares and the possession of market power. Following this approach, it reasonably holds that the power to price ‘excessively’ is the preserve of firms of overwhelming size relative to the market in which they are located and which are, in addition, markets characterised by unusually high entry barriers. That is, the market share enjoyed by the firm in question should approximate 100% and there should be no realistic prospect of entry – in other words the market should be both uncontested and incontestable. The 74 We cite again the well known statements of the United Brands court and the Napp Pharmaceutical tribunal where equally the phrases ‘normal and sufficiently effective competition’ (United Brands) and ‘prices higher than would be expected in a competitive market’ (Napp) similarly do not in any way appear to connote perfect competition as the appropriate norm or comparator. 33 concept of ‘super dominance’ and the special responsibilities that attach to this privileged status is well recognised in scholarly work 75 and in the decisions of competition adjudicators. In Napp Pharmaceuticals the UK Competitions Tribunal held: ‘We for our part accept and follow the opinion of Mr. Advocate General George Fennelly in Compagnie Maritime Belge , cited above, that the special responsibility of a dominant undertaking is particularly onerous where it is a case of a quasi-monopolist enjoying ‘dominance approaching monopoly’, ‘superdominance’ or ‘overwhelming dominance verging on monopoly’ [2000] ECR 1-1356 at paragraphs 132 and 137. In our view Napp’s high and persistent market shares put Napp into the category of ‘dominance approaching monopoly’ – i.e. superdominance – and the issue of abuse in this case has to be addressed in that specific context’ 76

[97]Our approach is also consistent with that taken by a number of scholars who have examined the question of excessive pricing. Although different standards are proposed by the various writers, it is generally accepted that mere dominance is an insufficient structural basis for the charging of excessive prices. Hence Evans and Padilla in their article ‘Excessive Prices: Using Economics to Define Administrable Legal Rules’ suggest that firms should, as a general rule, be free to charge prices above cost and the only exception to this rule should be ‘…situations where the dominant firm enjoys a legal monopoly and the excessive prices charged by the goods and services offered by the legal monopolist are likely to prevent the launching of new products or the emergence of adjacent markets’ 77

[98]Motto and de Streel write With regard to exploitative excessive prices, we suggest that the Commission intervenes in cases of very strong dominance (confined to a monopoly or near 75 See Richard Whish Competition Law 5 th Edition (LexisNexis, London 2003), pages 189-190. 76 Napp Pharmaceuticals at para 219. 77 David S. Evans and A. Jorge Padilla, “Excessive Prices: Using Economics to Define Administrative Legal Rules”, CEMFI Working Paper No. 0416 (September 2004), page 5. 34 monopoly) that are caused by past or current legal entry barriers, whenever market forces alone are unlikely to lead to competitive results. 78

[99]Motto and de Streel continue Competition rules cannot be applied in newly liberalised markets in exactly the same way as they have been applied in ‘normal’ sectors because the market structures and the risks for competition are substantially different’ 79

[100]Evans and Padilla are, understandably, particularly wary of excessive pricing rules that rely on the establishment of competitive benchmarks ‘…in dynamic industries where investment and innovation play a paramount role’ 80

[101]Mittal SA has taken up this latter theme by its reference to the right of a patent holder to extract a monopoly price: ‘...the patented price is always higher than the economic value of the product for good reason. Nevertheless, a patent holder has a legal right to the exclusive economic exploitation of an innovation (and the market power which that brings), for a limited period. Accordingly, it is not unreasonable for a patent holder to charge a price which bears no relation to the economic value of the product for the economic duration of the patent.’ 81

[102]How does our test stack up against those proposed here? We agree with Evans and Padilla that excessive pricing allegations should be particularly carefully scrutinised in dynamic industries characterised by investment and innovation. 82 78 M Motta and A de Streel, “Exploitative and Exclusionary Prices in EU Law”, Paper presented in the 8 th annual European Union Competition Workshop, Florence (June 2003), page 27. 79 Ibid, page 28. 80 See Evans and Padilla, page 7. 81 See Mittal SA Heads of Argument, para 9.3. We deal with the question of innovation and the extent to which it is reflected in ‘excessive prices’ below. Suffice for the moment to note that where a monopoly is temporarily granted by patent in to promote the innovative process, a critical element of the competitive process itself, this may provide the necessary ‘reasonableness’ in the relationship between a monopoly price and the goods economic value, although, as effectively suggested in Napp, ‘patent abuse’, rather than the patent itself may, in certain circumstances, be found. 82 Although note that the Tribunal in Napp Pharmaceuticals, a case which did concern a dynamic ‘innovation market’ and patent issues, was able to take account of these issues and apply an excessive pricing rule with coherence and relative ease. 35 However this dynamism, as Mittal SA’s expert, Dr. Walker, conceded, does not characterise the production of flat steel products. Nor is Mittal SA’s invocation of the rents earned on patents of assistance in the matter before us. Again there is no claim that Mittal SA’s pricing is rooted in the extraction of any innovation rents or patent rights. 83

[103]While the Mittal SA monopoly is not a legal monopoly and has, to our knowledge, never operated under exclusive license, it was capitalised by the state and between its establishment in 1928 and its privatisation in 1989, it was owned and controlled by the state. It must also be borne in mind that until as recently as 1984 the state saw fit to regulate the price of steel. Moreover its privileges did not terminate with its privatisation. It is worth reciting at some length the following exchange between Harmony’s counsel and Dr. Zavareh Rustomjee, a previous director general of the Department of Trade and Industry and industry expert: “ADV UNTERHALTER: Yes. Now if I could refer you to paragraph 12 of your witness statement, you refer to a range of Government incentives and support measures, which were offered to industry over the decade. I wonder if you could just briefly indicate what they were and what the point of these benefits was. DR RUSTOMJEE: Yes, some of these schemes were in place from 1994, at the start of my tenure as Director-General and some of them were instituted later and some were adapted over the period. These schemes that I’ve highlighted here were mainly relating to support for investment. So they were investment promotion schemes. The philosophy behind all of them did underpin that cleavage between upstream and downstream. So wherever incentives were given on the upstream, and I will elaborate in a minute on some of these, the intention was and it was sometimes translated into commitments by the parties who received these benefits, that they link the upstream to the downstream. So, if we take the General Export Incentive Scheme, this was an inherited scheme that was started in 1992. It was extremely wasteful in my view. We terminated it prematurely. I have no doubt if the current Auditor- General were around at the time he would have indicted us for wasteful expenditure for continuing this scheme for as long as we did, but that’s not 83 Transcript, at page 2321. 36 be. R21 billion were absorbed over a period of 5 years. That was essentially a scheme, which gave a cash payout to anybody who exported anything, whether it be upstream steel or finished complex manufacture on a sliding scale. So you got more if it was a complex manufacture. Then there was the accelerated depreciation allowance under Section 37E of the Income Tax Act. Now this scheme essentially allowed for a capital intensive product and it’s very useful for capital intensive product. It allowed a write-off of capital expenditure earlier. So effectively it’s the time value of money that you save here, but during the difficult period of a start-up of a project it’s quite important. That was offered between 1991 and 1993. I believe only 13 firms received that benefit, including Iscor. I can’t remember exactly on which projects, but I know Saldanha was one of the major recipients of this. Now in all of those 13 recipients, all of them committed to in terms of the agreements that were struck at the time to making available the products of those investments to downstream users who would use that to export at a price, if I remember the wording, at a price that did not give the supplier of that product a greater benefit, a greater profit than they would have obtained on their exports. So that applied … I stand to be corrected, but I’m almost certain it applied to all 13 of the recipients of that one. Then we had the regional industrial development program, which was inherited and effectively ends at the downstream side of the industry, smaller investments, more labour intensive investments. In some cases there were certain additional benefits to be gained, if the project was more labour intensive. And that was modified over the years. It became the small and medium manufacturing development program. That was then extended to non-manufacturing enterprises through the small medium enterprise development program and I think that still is in existence today. Then more recently, between 2000 and 2004, there was a strategic investment program scheme that was offered, which, as I understand it, it was different from 37E in that it was almost a tax write-off. It allowed you a tax write-off. It was kept at 3 billion and it was absorbed by projects, which mainly confine themselves to the metal and chemicals sectors. ADV UNTERHALTER: Yes. 37 DR RUSTOMJEE: The same for 37E incidentally. Most of those projects were in the metals and chemical sectors.” 84

[104]Mittal SA has sought to question the relevance of this evidence. While it certainly does not dispose of the complex question that we are here required to deal with, it is enlightening from two points of view. First, while this firm may not have been a legal, licensed monopoly, it enjoyed, even after it was privatised, a degree of public largesse that would, in the assessment of the scholars cited above, qualify it for membership of that small universe of companies in whose pricing practices the state is entitled to take an active interest. And again we stress that until 1984 the state’s interest in Iscor extended to actual price regulation. Certainly Motta and de Streel should have little difficulty – particularly given the extent of post-privatisation subsidy testified to, without challenge, by Rustomjee - characterising this as a ‘newly liberalised market’ in which ‘competition rules’ could not be easily applied and in which ‘the risks for competition are substantially different’. Secondly, it is also instructive that, as testified by Dr. Rustomjee, public policy was clearly concerned that the considerable subsidies it provided translated, not into the improved bottom lines of the direct recipients, but were rather used to support consumers of the vital intermediate products, the producers of which received public support.

[105]While a licensed or legal monopoly is undoubtedly sheltered behind very high entry barriers indeed, there are clearly circumstances where the entry barriers established by historical circumstance and technological and commercial considerations are, in effect, at least as insurmountable as those that are constituted by law or license. Just as legal monopolies are established by statute or by administrative fiat, so too can their monopoly status be undermined by the same process. Hence while an exclusive license to collect garbage in Johannesburg may conclusively prevent the entry of others into the protected market, were a second license to be issued the erstwhile monopolist may have little left by way of entry 84 See the transcript dated 24 March 2006, pages 773-775 (our emphasis). Note that of course one of Mittal SA’s most significant advantages is the price at which it receives a large proportion of its iron ore, a price that, it appears, was agreed in the process of restructuring the erstwhile Iscor into a separate steel company and an iron ore company. While this may not necessarily be a subsidy conferred by the state it certainly, as we note below, sets up a considerable barrier in the way of a new company that may wish to enter the South African market and that will have to pay the prevailing international price for iron ore rather than the low price at which Mittal SA receives this vital input. This too is outlined in Dr. Rustomjee’s witness statement. Harmony also points to the low price – relative, that is, to international prices – of Mittal SA’s electricity and natural gas costs. (See Harmony Heads of Argument, para 62). This does not appear to have been challenged by Mittal SA. Again it is not clear whether this arises from relatively low electricity and natural gas prices in South Africa or whether Mittal SA has, for some or other reason, managed to extract a particularly low price for these inputs. 38 barriers in order to sustain its monopoly. On the other hand, a long established, licensed fixed line telephone monopoly may find itself as effectively protected by commercial and technological barriers as by its license – a telecommunications regulator may issue a new operator’s licence at the stroke of a pen, but the newly licensed operator may nevertheless experience considerable difficulties in overcoming the advantages that the incumbent derives from its established network.

[106]We should of course add that the question of excessive pricing – or, at least, the possibility of challenging pricing conduct – is unlikely to arise in the case of a legal monopoly precisely because, in the current economic policy environment, such an institution will, invariably, be subject to regulation. As already noted, it is our view that Section 8(a) is precisely intended to apply to those rare markets that are uncontested (monopolised or ‘super-dominated’), incontestable (subject to insurmountable entry barriers) and unregulated (not subject to price regulation). The South African market for flat steel products is, the evidence shows, just such a market, and this is why the proposal of Evans and Padilla and other writers that the powers of competition authorities to intervene in pricing conduct be reserved for the most exceptional circumstances is, in our view, strictly adhered to in this decision even though we do not require it to be restricted to a case of a legal monopoly.

[107]In the present case, dominance of the relevant market is indeed absolute, that is, there are, within the boundaries of the relevant market, no meaningful constraints on the first respondent’s ability to unilaterally determine price – its market share is persistently vast and there is no prospect of new entry at all, and certainly not within any time-frame that anti-trust jurisprudence and enforcement practice would regard as constituting an effective competitive constraint. Moreover, the firm in question was owned by the state, for much of its life its prices were regulated by the state, and certain of its current advantages derive from advantages accrued from the period of state ownership as well as subsequent subsidisation.

[108]In short, the first respondent, Mittal SA, is no mere ‘dominant firm’ – it is ‘super dominant’, a ‘monopoly’ in the parlance of US anti-trust law. It is, to all intents and purposes, an uncontested firm in an incontestable market. This is a market structure that is rarely encountered in competition analysis, possibly as rare as its opposite number, a market that meets the conditions of perfect competition. As already noted, while even a super-dominant firm, a monopolist pure and simple, 39 remains constrained by the existence of a ceiling in the price that it may charge, this limitation is not imposed by, indeed is in no way influenced by, the pricing practices of competitors, actual or potential, in the relevant market, or, even as a last resort, by the ability of the customers, to forego use of the product in question.

[109]Mittal SA’s overall share of sales of flat steel products in the South African market clearly establishes its overwhelming dominance. Below are the target market shares of total flat steel products, which are computed by Mittal SA. We have also extracted from the table provided by Mittal SA the actual market share that they managed to sustain. These resemble market share figures of Mittal SA from the second quarter of 2002 to the second quarter of 2005. Source: Exhibit 2: The market share figures computed by Mittal SA on a product-by-product, category-by-category basis .

[110]Mittal SA’s market share clearly establishes its dominant position and, hence, the jurisdictional basis for a Section 8 complaint. Indeed, the overwhelming market share and its striking stability over time establish that this is effectively an uncontested market. This was confirmed by the candid reflections of Mr. Dednam, a leading Mittal SA witness, on the negligible extent of competition offered by Highveld Steel, the second largest participant in the South African market for flat steel products: “ADV UNTERHALTER: Yes. Just one other issue on that Mr Dednam, you are exporting considerable amounts of product at prices that are considerably less than what you are obtaining in the domestic market. MR DEDNAM: We do. ADV UNTERHALTER: Yes, and so to an extent the market share that Highveld sustains at present pricing levels, should be an opportunity for you, because if you were trying to grab market share from them, then you would Quarter 2Q 2002 3Q 2002 4Q 2002 1Q 2003 2Q 2003 3Q 2003 4Q 2003 1Q 2004 2Q 2004 3Q 2004 4Q 2004 1Q 2005 2Q 2005 Flat Products Market Share % 84.1 85.9 85.1 83.1 81.8 80.1 77.9 79.9 81.8 79.5 82.7 82.6 79.0 Target Market Share % 81.0 81.0 81.0 81.0 81.0 81.0 81.0 81.0 81.0 81.0 81.0 81.0 81.0 40 be offering prices, which may be better than Highveld’s, but are certainly a lot better that your export prices. MR DEDNAM: Mr Unterhalter that particular proposition we think about every time in all the marketing discussions that I have with the team. We talk about it. ADV UNTERHALTER: Yes and why don’t you… MR DEDNAM: And I think that Mittal Steel South Africa, being the dominant player in the market, has the muscle and the power to take out those guys [Highveld Steel] in the market place.” 85

[111]Dednam’s remarks clearly suggest not merely that Highveld offers negligible competition to Mittal SA, but that it only participates in the market at all at the pleasure of Mittal SA. 86 Mittal SA has, for whatever reason, been prepared to allow Highveld to retain a certain small market share. 87 The evidence shows clearly that Highveld’s pricing simply follows Mittal SA’s lead. Indeed the evidence is strongly suggestive of active collaboration between the two producers, although clearly with the terms of co-operation dictated by Mittal SA. 88

[112]We repeat then, that the size and persistence of Mittal SA’s share of the South African market for flat steel products and the extreme weakness of its token opposition qualify it for designation as a ‘super-dominant’ firm or monopoly.

[113]We have already noted the existence of some effective competition from alternative products – for example, from plastic in the production of cans used to hold liquid – and from alternative geographies. The auto industry, where the potential 85 See transcript dated 6 April 2006, page 1816 (Our emphasis). 86 Indeed Mittal SA’s own counsel describes Highveld as ‘hapless’. See Mittal SA Heads of Argument, para 11.5. 87 Mr. Dednam for Mittal SA explains the stability of the respective market shares of Mittal SA and Highveld in the following terms: Mittal ‘feel(s) comfortable at a certain level of market share and we will not aggressively attack the market to take market share out of the market, because we are the dominant player in the market. And that will provoke all kinds of other complaints I assume’. 88 This is borne out by the following statement of Mr. Dednam: ‘I also hold the opinion that in any price war, I don’t think that any producer is winning. I think everybody is losing in a price was situation, except for the end user. The end user will definitely win in a price war situation. And I assume that is basically the reason and if Highveld is thinking the same it holds the same type of opinion about it.’ See also the testimony of Mr. Bernard Swanepoel at transcript pp. 188-91 where Swanepoel testifies as to how closely Highveld prices track those of Mittal SA and concludes at p191 as follows: ‘Mr. Unterhalter: Yes. Now can I ask you in general terms in your experience as a purchaser of steel, have you any experience as to whether Highveld is a competitive discipline upon Iscor’s pricing? Mr. Swanepoel: I think certainly everybody in my organization believes that whatever Mittal does gets followed by Highveld within a few days. I don’t think that there is any competition evident or visible from where we sit in the chain’. 41 allegedly exists for the importation of body panels from other parts of the multinational auto assemblers’ supply chains, exemplifies the latter. These are the customers who enjoy varying level of rebate off the Mittal SA list price. We note however, that the extent of the rebates varies greatly. Certain of the rebates offered are, by any measure, insignificant, and so result in an ultimate price for the rebated product that differs only slightly from the list price at which the vast bulk of Mittal SA’s output of flat steel products is sold.

[114]Mittal SA argues that its rebate programmes represent its responsiveness to those of its customers who receive them and therefore an absence of market power on its part. As an absolutely clear matter of law the existence of these discounted segments has no impact whatsoever on the assessment of market power. That is to say, Mittal SA’s share of the South African flat steel market clearly establishes its presumptive dominance in terms of Section 7 of the Act.

[115]In any event, the complainants, for their part, have much the better of an argument with their adversaries in which they aver that the manner in which Mittal SA grants its rebates is, if anything, further confirmation of its market power. There is, insists Harmony, and the testimony of customers supports this, little evidence of a genuine negotiating process – rebates are granted, are bestowed, rather than agreed. They are subject to capricious and sudden unilaterally determined amendment which, several customers testified, made pricing of his product, the downstream product, a complex task. 89 89 See testimony of Mr. Stephen Leatherbarrow at transcript pages 358-9:’Mr. Leatherbarrow: Okay. During 2002 we approached Mittal to look at reviewing the way we could go forward with secondary exports for tube and pipe. We made a strategic decision to try and grow that business and we needed to obviously base the future on a firm foundation and we weren’t comfortable with the existing rebate arrangement. Mr. Unterhalter: Why not? Mr. Leatherbarrow: At that time the rebate arrangement was more retrospective than forward looking. It was on a quarterly basis. So it didn’t take into account changes in the steel cycle month-on-month. It didn’t take into account monthly exchange rate fluctuations. We wanted to be more in control of our destiny to be able to build on and grow an export business. So we needed something that was more transparent where it also put a certain amount that we were prepared to take on our side, rather than leave Mittal to work out a rebate’. And then further Leatherbarrow emphasises: ‘We felt that it (the computation of the rebate)was left very much to a few people in the organisation with a view. We weren’t privy to the details of the calculation and how it was worked out and that it wasn’t taking a forward view and the steel cycle and the exchange rates adequately enough’. The lack of transparency in the computation of the rebate and the capriciousness of changes in the rebate also comes through clearly in the lengthy examination of Mr. Cohen, the conveyor belt manufacturer and exporter. See particularly transcript pages 404-5. Cohen also bemoans at page 409 of the transcript the impact that this has on the ability of a Mittal customer to price in its own, downstream market: Mr. Cohen: When we submit a price to an end user, if we are fortunate enough to be granted an escalation, it will be based upon our increase in costs and the increase in costs will be…we have to pay a higher price for steel. We have to pay a higher price for labour. We have to pay higher prices for nylon to produce the seals or for rubber to produce the belt, 42

[116]This is true even of the rebates granted to the auto original equipment manufacturers (OEMs), those of Mittal SA’s customers who may reasonably be thought to possess the greatest countervailing power vis a vis Mittal SA. Mittal SA explained that the auto rebates are driven by the prospect that, should domestic flat steel prices diverge too greatly from those available in other parts of the world, the domestic auto OEMs may move the production of significant steel components offshore by utilising their well established international supply chains, with the South African plant then used only for purposes of assembling the components produced offshore. This, insisted Mittal SA, is evidence of its responsiveness to its customers. However, we have noted elsewhere in this decision that the ease with which even the auto OEMs are actually able to rely on imported steel components has been strongly questioned in another Tribunal decision which concluded that, precisely because of the difficulties in the path of relying on imported steel for the auto industry, the relevant geographic market for steel used in the South African auto industry was national and not international. 90 Testimony from auto industry representatives at these hearings vindicates that view. Certainly even the auto rebates are erratic in the basis of their formulation and are, in their amounts, also subject to capricious and significant and sudden change. An auto executive, Mr. Errol Classen, testifed that the industry was recently simply advised by letter from Mittal SA of immediate increases in the prices of steel of as much as 25%. 91

[117]At best for Mittal SA, what the evidence on the auto rebates suggests to us is that at a certain point in the auto OEMs’ decision making process Mittal SA may be somewhat vulnerable to the OEMs’ international supply chain but that once an OEM has committed itself to the production of a particular model at a particular site and etc. This system has no relationship to the costs, and that’s what’s making this thing complicated. This system is based upon the price that Mittal can get in the marketplace based upon a base price, which they are telling the market is going to prevail in 3 months time and a rate of exchange. But the rate of exchange hasn’t actually affected Mittal. It hasn’t changed t heir costs. It only changed their potential to earn money if the exchange rate has weakened. And that’s why it’s become so difficult for an exporter, because an exporter can’t say here’s the price that I’m going to pay, it’s based on the cost of iron ore or it’s based on the cost of steel. We’re saying here’s the price that you’re going to pay, but we cant hold it firm for longer than 3 months, because if we export it in the 4 th month, there could be a different price pertaining at that stage in time and that is based on the international prices and has nothing to do with the costs that we’re faced with. That’s why this thing is so difficult to understand and to implement and requires us to use consultants to monitor and administer it. (Our emphasis) 90 See Trident Steel (Pty) Ltd and Dorbyl Limited, [2001-2002] CPLR 302 (CT). 91 Transcript, page 687. There is an allegation, denied by Mittal SA, that the increase in question was in violation of a contractual commitment by Mittal SA to maintain particular pricing levels. This is not for us to decide and is, for our purposes, not relevant. However what appears undeniable is the vulnerability of even the auto industry to large and sudden – and largely unexplained – adjustments in the price of their vital steel input. 43 has established its supply chains accordingly, the bargaining power then shifts right back to the super-dominant steel producer who, as the instance cited immediately above indicates, does not hesitate to use it.

[118]Moreover, in our view, from the perspective of the decision that we are called upon to make here, the most important learning to be derived from the Mittal SA rebate system, is not its implications for the assessment of market power. We emphasise that its massive market share is vastly in excess of the threshold of 45% and this establishes that it is presumptively dominant and thus possessed of market power. However, what the rebating system and the manner of its operation does eloquently evidence are the lengths to which Mittal SA is prepared to go in order to immunise the lion’s share of, on the one hand, its domestic market in which its customers pay the full list price, from, on the other hand, those markets, be they domestic or export, in which it discounts or rebates off the list price. We will return to this critical aspect of our reasoning in relation to the rebates below.

[119]Note too that even were Mittal SA to refuse to discount to these select customers this would not necessarily mean the importation of flat steel products into South Africa. A refusal to discount may curtail the production of metal cans in South Africa or it may result in the importation of steel auto panels from beyond South Africa’s borders. It may, in other words, result in a shrinking of the South African market for flat steel products, but it would not reduce Mittal SA’s share of that market.

[120]Nor does it appear that there is any realistic prospect of new entry. Certainly it has never been suggested by any witness that this is a feasible prospect. 92 When the giant LNM – now the Arcelor Mittal multinational – chose to enter South Africa it did so by way of acquisition of ISCOR rather than by way of green-fields investment in steel producing plant. Arcelor had, prior to its merger with Mittal, been a de minimis participant in the South African market supplying, as far as we can ascertain, only grades of steel not produced in the Mittal SA plants. There was some indication that it, Arcelor, had intended setting up steel service plants in South Africa in order to 92 Note that Mittal SA characterises the view of the UK Competition Appeal Tribunal (“CAT”) in terms that would certainly suggest that the CAT would find that the underlying conditions in the South African market for flat steel products would render Mittal SA a strong candidate for pricing excessively: ‘The CAT thus appeared to suggest that what distinguishes excessive pricing from reasonable monopoly pricing is that the former takes place in circumstances where there is no threat of competitive entry to restore prices to competitive levels, with the result that the monopolist can earn supra-competitive profits in a completely uninhibited manner’. Mittal SA Heads of Argument, para 7.22. 44 better serve, by way of imports, the steel requirements of South African auto manufacturers. However these remained at the level of intention only and, with the merger of Mittal and Arcelor, an intention that will now certainly not be realised. 93 We note too that any new entrant would want to serve the domestic market and it would have to confront the massive advantage bestowed on Mittal SA by its extraordinarily favourable arrangements with respect to iron ore pricing.

[121]In summary then, Mittal SA is, for the purposes of the Act, clearly dominant in the relevant market, the South African market for flat steel products. However, as already elaborated, in order to establish the structural basis for charging excess prices, something more than mere dominance is required. In our view Section 8(a) demands a showing of extraordinary or ‘excessive’ market power, the power to price at a level beyond that available to a mere dominant firm. The extent of Mittal SA’s market share taken together with the height of entry barriers and its recent history of state support easily establishes its status as a super-dominant firm within the relevant market. It has been proved that it is indeed an uncontested firm within an incontestable market. Excessive prices in the South African market for flat steel products

[122]There may well be a credible argument for ending our enquiry here. The regulation of competition is underpinned by the observation – supported by an elaborate theoretical framework and a wealth of empirical evidence - that different market structures produce distinctive outcomes as measured by output prices and production costs or efficiency, including, in a dynamic product market, levels of innovation. Along a spectrum that begins with the perfectly competitive paradigm of price-taking, fragmented, small producers and ends with a single, price fixing monopolist, the structures with which competition enforcers engage most regularly, and for which much of competition law is designed, are characterised in the literature as ‘imperfectly competitive’ and ‘oligopolistic’. These markets are characterised by a relatively small number of producers whose engagement may be underpinned by a rivalry aimed at acquiring market share through producing the best quality goods and services at the lowest possible price, or, in dynamic markets, at producing a product sufficiently differentiated from those of its rivals to accord it a degree of pricing power. 93 See Mittal Steel Company N.V. and Arcelor SA, Tribunal Case No.: 53/LM/Jun06. 45

[123]However, at the risk of repetition, this imperfectly competitive structure is threatened by the prospect of profit maximising strategies other than ‘competition on the merits’, the term commonly used to describe the pro-competitive process elaborated in the previous paragraph. One possible strategy is through the merging of rival producers, hence the merger review powers characteristically extended to competition authorities; a second is through the development of co-operative mechanisms between nominally competitive firms, hence competition law’s particular concern with horizontal agreements. A third possibility is that, through legitimate pro- competitive conduct, or, alternatively, through a history of government subsidy and support, a single firm may come to dominate a market and may then choose to focus its profit maximising strategy on devising mechanisms to exclude rivals from the dominated market, hence the preoccupation with exclusionary conduct in abuse of dominance provisions.

[124]Abuse of dominance is preoccupied with exclusionary conduct because of the well-founded view that holds that most markets are, to a greater or lesser degree, contestable, that is, that potential entrants will respond to anti-competitive conduct on the part of a dominant firm and the privileged margins that it enables by entering the dominated market thus making substitute products available to the dominant firm’s customers. In much the same way then that anti-trust law is not directed at producing perfectly competitive market structures, it is, conversely, not generally directed at eliminating dominant firms. Its reluctance to disallow dominant firms is predicated on a fear that it may, by so doing, inadvertently chill competition on the merits which is legitimately directed at the attainment and maintenance of a dominant position vis a vis competing producers. Dominant firms are thus viewed as beasts best dealt with by entry, or even just the prospect of entry, into the dominated market, and this can be ensured by a vigilant competition regulator enforcing rules designed to prevent conduct directed purely at excluding actual or potential rival producers.

[125]However, the existence of incontestable markets – and hence of monopolies capable of being sustained without resort to impeachable exclusionary conduct – is recognised. In particular it is recognised that the public authority may choose or may be compelled to govern entry through the assumption of licensing powers. In these – and other circumstances in which markets are, for one or other reason, incontestable, at least by the unilateral decision of a potential new entrant – the state usually establishes a regulatory authority mandated to simulate competitive 46 outcomes tempered by whatever ‘public interest’ or non-competition objectives are deemed appropriate. In summary then, the cases of incontestable markets – and hence sustainable monopoly – are rare and where they occur they are generally dealt with by regulation. The case of a sustainable monopoly in an unregulated market is a particularly rare phenomenon the more so in a large economy, the context for which much competition law is written and in which most jurisprudence has been developed. Hence the relative lack of attention paid by contemporary competition law to what are termed ‘exploitative’ abuses – it is, we repeat, generally assumed that exploitation of a dominant position will either be dealt with by new entry or, where that is not possible, it will be dealt with by regulation.

[126]Clearly though the corollary of the recognition that the potential for exploitation is limited by potential entry (or, conversely, by an absence of exclusionary conduct) or, where necessary, regulation, is that, absent contestability or regulation, exploitation – expressed either as excessive prices or gross inefficiency – is a rational profit maximising strategy. There is, in these rare circumstances, no incentive to behave otherwise. It is indeed this insight that prompted a wide array of states to withdraw from direct involvement in the economy in the last quarter of the 20 th century. Incontestable markets monopolised by state owned enterprises engaged in ‘profit maximisation’ strategies that were most often characterised by chronic inefficiency, by optimising the ‘quiet life’, rather than by excessive pricing. However where these inefficient enterprises were privatised or part-privatised into incontestable and unregulated markets, shareholder pressure ensured that the likely profit maximisation strategy was the charging of ‘excessive prices’, prices, that is, that accord with the extraordinary degree of market power, the super-dominance, enjoyed by the firm in question. This precisely describes the conditions of Iscor’s privatisation and its continuing passage from inefficient state owned enterprise to profit maximising monopolist. We note again, that for several years after its privatisation, it appears that the state saw fit to regulate the price of Iscor’s product. However, it was then left to its own devices and so, shortly after its privatisation, a state owned monopoly had effectively been transformed into a privately owned and unregulated monopoly.

[127]The insertion into the Act of an ‘exploitative abuse’ in the form only of Section 8(a)’s proscription of excessive pricing – all the other Section 8 abuses refer to exclusionary abuses – is then to cater for those rare beasts who are subject neither to the constraining presence of a regulator or of a potential 47 entrant. A firm in a market that is both uncontested and incontestable and unregulated is unconstrained by law or competition – it can exploit its structural advantage without fear of competition, actual or potential, and therefore without necessary recourse to impeachable exclusionary conduct and it is unconstrained by regulation. While Mittal SA’s counsel has referred us to the draconian efforts to curb monopoly power made – in 301 A.D and 483 A.D. respectively - by the emperors Diocletian and Zeno, he may have been better advised to heed the words of William Shakespeare who in Measure for Measure tells us ‘O, it is excellent to have a giant’s strength: but it is tyrannous to use it as a giant’

[128]Section 8(a) is intended to constrain the conduct of these giants, to constrain ‘tyrannical’ conduct, conduct that is enabled by their ‘gigantic’ strength.

[129]However the complainants (and anti-trust generally), have effectively taken the view that, Shakespeare’s injunction notwithstanding, giants of commerce do not voluntarily constrain the exercise of their ‘tyrannical’ powers. Harmony insists that it would be irrational – an ‘obvious absurdity’ is the term used by its counsel - for Mittal SA to extract a price lower than that which its structural super-dominance enables it to charge. 94 It will, in other words, as a matter of profit maximising rationality, price to the full limit of its market power, and because the extent of its market power is unrestrained by any competitive consideration, the price that it charges will rationally exceed those whose pricing power is constrained by the likely responses of their actual or potential competitors or, if it were subject to regulation, by the price-setting powers of the regulator. Because, given the unusual character of its uncontested and incontestable market, Mittal SA’s pricing power extends beyond that even enjoyed by a mere dominant firm and because its price is not subject to regulation, we should be able to conclude, without further, that the price in question may well, as a matter of profit maximising rationality, be reasonably construed as excessive because, in the oft quoted words of the United Brands judgement, it does not derive from structures compatible with any notion of ‘normal and effective competition’.

[130]There is much about this argument that is compelling. At the very least, it would suggest that given Mittal SA’s structural capacity to price without restraint, and 94 Harmony Heads of Argument, para 120. 48 the ‘absurdity’, within a profit maximising framework, of voluntary restraint, it is for Mittal SA to establish that it has foresworn profit maximisation in favour of pricing behaviour that would effectively see it leaving money on the table every time it sold any output in the domestic market.

[131]However, prudent adjudication suggests that proof of structural super- dominance is necessary, but not sufficient, to find excessive pricing. It may, after all, be argued that all abusive conduct, indeed all anti-competitive conduct, is rational, albeit illegal, profit maximising conduct. Just as dominant firms bear, relative to non- dominant firms, particular responsibility to restrain their unilateral conduct to ‘competition on the merits’ or risk prosecution, so are ‘super-dominant’ firms effectively required by Section 8(a) to leave a certain amount of money on the table by restraining their pricing below that which their respective markets may bear or face prosecution for excessive pricing even though a rational approach to profit maximisation may dictate that they price to the limit of their market power. 95 Of course, the constraints operative in the case of ‘mere’ dominant firms may well derive less from voluntary restraint and fear of prosecution than from the predicted response of their competitors or potential competitors in contestable, albeit dominated markets, while the privileges of super-dominant firms are not, by virtue of their super- dominance, similarly constrained by considerations of competition, actual or potential.

[132]But it is, on the other hand, conceivable that even a super dominant firm like Mittal SA, given the degree of expressed public and governmental anxiety with regard to its pricing practices, may well voluntarily recognise some sort of obligation arising from its super-dominant status and restrain its own pricing conduct, and so, we cannot, simply by reference to the rational conduct of super-dominant firms, conclude that Mittal SA is indeed in contravention of the prohibition on excessive pricing. Indeed, Mittal SA’s claim – which we have rejected – that it has moved from an import parity basis for price determination to one based on a basket of domestic prices in a number of comparator countries is precisely meant to suggest that it is engaged in a degree of voluntary restraint. Periodic media reports that speculate on the prospect for agreement between Mittal SA and government on a pricing policy for steel may also be intended to portray a monopolist willing to engage in voluntary 95 Harmony argues that ‘as a super dominant firm Mittal has a special and more onerous obligation not to abuse its position’. Harmony Heads of Argument, para 245. 49 restraint. 96 In short, while our examination of the relevant market and market shares and entry barriers is a sufficient basis for our conclusion that Mittal SA is one of those rare firms endowed with sufficient market power to charge excessive prices, it may be an insufficient basis for finding that it has actually deployed that power in contravention of Section 8(a).

[133]Moreover, all of Section 8 is ultimately directed at conduct. And we must reasonably conclude that this is no less true of Section 8(a) than of any of the other abuses described in the other sub-clauses of Section 8. This is true even though structural remedies are available in respect of a first Section 8(a) offence. However were we to derive our conclusions regarding the alleged existence of excessive pricing solely by reference to structure, we would effectively be concluding that certain structures – uncontested firms in incontestable and unregulated markets – were prohibited per se. We do not believe that this accords with the character of Section 8. It is conduct that abuses a structural advantage – dominance or, in Section 8(a)’s case, ‘super-dominance’ – that is prohibited. It is not the underlying structure that is prohibited. To paraphrase Shakespeare, while it is permissible (we hesitate to use ‘excellent’) to be a giant, it is the ‘tyrannous’ conduct that is enabled by this advantageous state that is prohibited.

[134]Hence, in addition to examining the structural features of the market in question, we must examine evidence which suggests that Mittal SA has engaged in conduct designed to abuse its ‘super dominant’ position by charging an excessive price. This is an enquiry mandated by the principles and practice of competition law and economics. We emphasise that we will not approach this enquiry by considering that evidence relating to actual price levels which effectively requires us, first, to identify a particular level as unlawful (‘excessive) and then to impose a level of price that would be lawful (‘non-excessive’). This, we stress, is an approach consistent with the practice of price regulation – it is not commonly found in the principles and practice of competition law and economics.

[135]We must underline that the approach outlined above conforms to the somewhat opaque statutory definition of excessive pricing. 96 Indeed Mittal SA avers that ‘we submit that the Department of Trade and Industry’s intervention and persistence has also had a constraining effect on Mittal’s pricing’. Mittal SA Heads of Argument, para 13.1. 50

[136]Our Act defines an ‘excessive price’ as: a price for a good or service which – (aa) bears no reasonable relation to the economic value of that good or service; and (bb) is higher than the value referred to in subparagraph (aa).

[137]We have already noted that this definition appears to borrow from United Brands, the leading European case on excessive pricing. We should re-iterate at once that the obvious borrowing from United Brands in no way requires us to adopt uncritically all elements of a European approach to excessive pricing despite the rather curious attempt by Mittal SA to insist that this is what is required of us. 97 The Competition Appeal Court – and indeed the Supreme Court of Appeal – have precisely cautioned us against this sort of slavish adoption of international jurisprudence. 98 We hasten to add that given, as shall be elaborated below, the high degree of confusion surrounding the efforts of both the various European courts and enforcement agencies to apply United Brands, the cautious approach to international jurisprudence commended by our own superior courts is particularly apposite.

[138]Furthermore any attempt to uncritically apply United Brands should at once recognise the striking differences between, on the one hand, the European treatment of pricing offences as contained in Section 82(a) of the Treaty of Rome and, on the other, the manner in which Section 8 of our Competition Act treats various pricing offences. Section 82(a) proscribes a litany of ‘unfair pricing’ and other pricing practices and has been used to tackle, under the rubric of unfair prices, pricing that is too low as well as pricing that is considered too high or excessive. The term “excessive pricing” is in fact not to be found in the Treaty of Rome at all, but is a product of case law, which decided that a species of the ‘unfair’ pricing targeted by Section 82(a) of the Treaty was to charge an excessive price and the test in United Brands is an attempt by the court to define when a price is excessive. The same section which prohibits unfair prices also condemns prices that are too low or predatory. In contrast in our law we have a provision – Section 8(a) - dedicated to the practice of excessive pricing while predatory pricing is dealt with separately under 97 See Mittal SA Heads of Argument, paragraph 4.10. 98 See Mondi Ltd and Kohler Cores and Tubes (a division of Kohler Packaging Ltd) v Competition Tribunal [2003] 1 CPLR 25 (CAC) at 35J-36B; and Federal-Mogul Aftermarket Southern Africa (Pty) Limited v The Competition Commission and another [2005] 1 CPLR 50 (CAC) at 53A-E. See also, Standard Bank Investment Corporation Ltd v Competition Commission and Others; Liberty Life Association of Africa Ltd v Competition Commission and Others 2000 (2) SA 797 (SCA) at 814F- 815A. 51 section8(c) and 8(d)(iv). Moreover, in contrast with the European treatment, our Act explicitly defines an excessive price – it does not leave this up to case law, except to the extent that case law reads meaning into the language provided in the statute.

[139]Because we have a statutory definition of what constitutes excessive pricing as opposed to a term such as unfair which then relies on case law to give it meaning from time to time, our interpretive efforts must surely initially focus on the language in our definition section. What did the South African legislators mean by the definition of excessive price that it inserted into the statute? This enquiry must surely take precedence over an uncritical borrowing from the decisions of a foreign court. We emphasise that this is the approach that our superior courts have commended to us.

[140]Turning then to the statutory definition of excessive pricing, the concepts that give rise to interpretive difficulty are contained in the phrase ‘bears no reasonable relation to the economic value’.

[141]The choice of words in the definition, ‘bears’, ‘reasonable’ and ‘relation’ - all words connoting imprecision rather than exactness – immediately suggests that the legislature did not intend that the relationship between the alleged excessive price and its economic value be capable of precise calculation, that is to say, it is not intended that we should find that the price is 34,5% over the ‘economic value’ and then find that, on some or other basis, that this differential is excessive. Rather the choice of language directs us to have regard to the ‘relationship’ between, on the one hand, pricing that is alleged excessive, and, on the other hand, a notion of the good or service’s economic value and then to judge whether that relationship is reasonable.

[142]As we have already outlined at length, since this is a competition statute one can safely assume that a price that is the subject of functioning market forces will not be deemed excessive or unrelated to the economic value of the good in question. After all the critical premise of competition law is that functioning markets determine what prices are reasonable. In other words, the judgement then that we are required to make is not of the price level itself but rather of the market conditions that generated the price level. In other words, we must ask ourselves whether the relevant market in question is capable of functioning in a manner that is likely to produce a reasonable relationship of price to economic value, or, rather, whether the structure of the market and, conceivably, ancillary conduct that depends on that anti- 52 competitive structure, forestalls the effective functioning of the market – forestalls ‘normal and effective competition’ in the words of United Brands - thus generating a price, the level of which, is unrelated to, is not influenced by, any cognisable competition considerations.

[143]Thus the use of the phrase ‘reasonable relationship’ requires us to examine the origin and extent of the dominant firm’s pricing power and then to enquire whether the price of the good or service in question derives purely from an anti- competitive market structure and, which is the case here, from ancillary conduct that relies upon the existence of that anti-competitive structure which, for the purposes of founding a claim of excessive pricing, we have termed ‘super-dominance.

[144]We adopt the same approach when defining the vexed concept of ‘economic value’. Just as the practice of law is comfortable with terms like ‘reasonable’, terms which have no precise meaning and intrinsic content but are given meaning by their context, so with the term ‘economic value’ in the discipline of economics. It too has no intrinsic, quantifiable meaning. It is not a fixed level capable of prior specification. That is, there is no fixed point that reflects the intrinsic ‘economic value’ of a good or service. ‘Economic value’ like ‘reasonableness’ is also a product of context, and that context is competition.

[145]Evans and Padilla note: There is no generally accepted definition of what an “unfair” price is. For Marxist economists, the “fair” price of a product is equal to the value of labor involved in its production. Classical economists like David Ricardo also held a cost-based theory of value. For neo-classical economists, the “fair” value of a good or service is given by its “competitive’’ market price, which is the equilibrium price that would result from the free interaction of demand and supply in a competitive market. This was also the interpretation given to the notion of “fair” prices by Scholastic economic thought, and is also the interpretation used by the ordo-liberal school of economic thought, which had a major impact on the development of competition policy in Europe. 99 99 See Evans and Padilla supra, at page 5 53

[146]We should at once dispel the notion that the term ‘economic value’ in our Act is intended to impute a cost-based theory of value, much less one that is rooted in any particular version of cost because if the legislature intended economic value to mean marginal cost or average variable cost it would have said so since it uses these terms explicitly in 8d(iv). That is to say, in assessing predatory pricing the legislature intends us to use a cost-based test and so Section 8(d)(iv) explicitly guides us in the cost measurement that is central to an evaluation of an allegation of price predation. However Section 8(a) and its accompanying definition make no reference at all to the relationship between an excessive price and cost. The reference is rather to the relationship between price and economic value.

[147]The concept of economic value consistent with the principles and practice of competition law and economics is, in the words of Evans and Padilla, ‘the equilibrium price that would result from the free interaction of demand and supply in a competitive market’ or the ‘competitive market price’. As we have already suggested then, our judgement of the relationship between price and economic value rests on our evaluation of the market conditions that underpin the price. If the examination of the structure of the market and any relevant ancillary conduct reveals that price is indeed determined by, what we have termed above, cognisable competition considerations, then that price will bear a reasonable relationship to the economic value of the good in question. However, if the price is the product of a market structure and of ancillary conduct that reflects precisely the absence of cognisable competition considerations then that price will be excessive in relation to the economic value because it will not have been determined by ‘the free interaction of demand and supply in a competitive market’. As we are careful to explicate below, ‘cognisable competition considerations’ or a ‘competitive market’ do not necessarily equate to conditions of perfect competition.

[148]Having rejected the view that the concept of economic value suggested by the Act is cost-based, we note the obvious point that this does not mean that cost does not play a major role in determining the absolute level of the competitive price or, what is the same thing, the economic value of a good or service. Even if the market for high performance cars is vigorously competitive and that for bicycles is monopolised, a high performance sports car will always have a higher price or economic value than a bicycle, and this for the simple reason that the underlying costs of producing a high performance sport car would not allow a manufacturer of these products to stay in business if he sold his product at the same price as a 54 bicycle. But to extend this analogy, it may well be that price and economic value are satisfactorily aligned in the pricing of high performance sports cars (that is, there is no excessive pricing) if their prices are the product of competitive market conditions, while the price of bicycles may be found to be excessive in relation to their economic value if they are priced under conditions of pure monopoly.

[149]Our approach is not dissimilar to United Brands’ concern with ‘normal and effective competition’ 100 or Napp’s fear that prices would be excessive in a market with high entry barriers and is devoid of ‘effective competitive pressure’. 101 Mittal SA is absolutely correct to insist that there is no warrant for reading ‘normal and effective competition’ or prices at ‘competitive levels’ as conditions of perfect competition and of prices that reflect that state of textbook grace. So too, we also emphasise that our enquiry into the market structure and conduct, into the presence or absence of ‘cognisable competition considerations’, that generated the price alleged to be excessive, similarly does not amount to a search for the holy grail of perfect competition. In other words, we, like United Brands and Napp, certainly do not hold that a price that bears a ‘reasonable relation’ to economic value, must be a price that is the product of perfect competition. Far from that, we have already clearly recognised that the statute explicitly admits of the possibility of dominance – hence Section 8 of the Act – and the pricing power that is a product of dominance. Hence there can be no presumption that a price that reflects a degree of market power will, for that reason, fall foul of Section 8(a) and its accompanying definition.

[150]Price formation under conditions of imperfect or oligopolistic competition is, to repeat the phrase we employ above, clearly influenced by ‘cognisable competition considerations’, even if it also reflects a degree of pricing power. We are, in other words, perfectly content to acknowledge that pricing in a market with a dominant firm and a number of other smaller players may reflect a degree of pricing power. Price formation in a duopolistic market, particularly if entry barriers are not insurmountable, may well reflect fierce rivalry and, hence, a competitive pricing outcome. In other words, albeit that these markets are ‘imperfectly’ structured, their pricing outcomes may well be the result of cognisable competition considerations. In these circumstances the competition authority will be alert to the prospect of exclusionary conduct or collusion but it will not attempt to second-guess the price because, in the absence of collusion or exclusionary conduct, the price will be determined by 100 See United Brands judgment, para 249. 101 Cited Mittal Heads of Arguments at para 7.21. 55 cognisable competition considerations and hence will bear a reasonable relationship to economic value.

[151]Mittal SA – and, indeed, many of the scholars that have examined the concept of excessive pricing – are concerned that the high priced product that is the subject of significant innovation will fall foul of the proscription of excessive pricing. However, while on a simple arithmetic interpretation of price and value along the lines conducted by the various European authorities, the relationship may be construed excessive or unreasonable, on our interpretation which requires an examination of the underlying market circumstances that produce the price in question, the price of an innovative product is unlikely to be at risk precisely because it will, by dint of the process of innovation and differentiation – a cognisable competition consideration - ‘bear’ a ‘reasonable’ relation to its value. By contrast, we emphasise again, where the price appears to have no explanation other than the pure exercise of monopoly power, then the price is not reasonable in relation to economic value. In other words what is relevant in our enquiry is not the arithmetic relationship between the price and some or other conception of cost. What is relevant are the underlying considerations that underpin the price level. Are these considerations founded in competition in its many degrees and guises or are they founded in pure monopoly?

[152]In this case our finding is that the price of flat steel products in the South African market is only explicable by reference to Mittal SA’s unusually high level of structural dominance which, in turn, supports ancillary conduct that maintains the price targeted by the monopoly steel producer. The ancillary conduct – which as we shall elaborate below is the enforced segmentation of separately priced markets – is a critical element of this decision because it demonstrates, as we shall elaborate, that even Mittal SA’s structural super-dominance was not on its own sufficient to guarantee that it actually achieved its unilaterally selected target price. Instead it was obliged to engage, in a clearly pre-meditated fashion, in ancillary conduct, conduct that is only available to a super-dominant firm, to achieve its desired price level in the domestic market. In other words, our finding of excessive pricing does not derive from an examination of the market structure alone; it also rests on the ancillary conduct upon which Mittal SA relied, ancillary conduct that itself depends upon the existence of structural super-dominance, in order to achieve its pricing ambitions. It is the cumulative impact of this structure and the ancillary conduct that puts Mittal SA’s contravention of Section 8(a) beyond doubt. 56

[153]In summary then, our examination as to the source of the pricing power is thus an examination into its reasonableness. Reasonableness in the context of a competition statute must mean ‘economically reasonable’. Economically reasonable in the context of a competition statute must mean having regard to the pro and anti – competitive considerations that we normally apply. As we go on to argue in this decision, the occasions where one can find no reasonable relationship between a price and the economic value underpinning it are few indeed. The circumstances giving rise to Mittal SA’s pricing power in respect of some of its domestic consumers depends on the existence of a range of factual issues that we do not encounter in the market place everyday, even in those markets habituated by long extant dominant firms.

[154]Nor is there any need to dwell on dictionary definitions of what excessive means. The term is a defined one and hence it is the statutory, rather than the dictionary, definition of the word that we apply. The statutory definition as opposed to the ordinary word ‘excess’ does not require one to conclude when a particular level of differentiation is sufficiently large to constitute excess. Rather it requires one to find a relationship between a price and economic value that admits of no reasonable explanation, that is, of an explanation that does not rely upon the exercise of the degree of market power that arises from super-dominance. The finding of an excessive price is then determined not by some arbitrary measure of difference but is rather an enquiry into the rationality of pricing. It thus condemns pricing for which unchallenged and incontestable monopoly is the only explanation as opposed to a price that may simply be high but for which innovation or even branding – that is, pro- competitive measures - provide the underlying rationale. 102

[155]For this reason we find that a reading of the Act that requires us to find precise levels for the economic value and then the actual prevailing price and then to 102 The question of branding and reputation answers the question posed by two economists in a newspaper column on the question of excessive pricing. They point out that a Porsche motor car is expensive, but asks whether this means that its price is excessive? The answer is no simply because, leaving aside questions of actual quality and cost of production differences with other automobiles, the manufacturer has created a brand image for which consumers are willing to pay. Indeed, the price of a Porsche may even be high relative to the price of other high performance sports cars, but as long as the market for high performance sports car is competitive, the likely explanation for Porsche’s success in persuading its customers to buy its unusually expensive car will probably be found in its branding strategy and history which has produced a highly desired object which is able to command a premium price. (see Johannes Fedderke and Volker Schoer ‘The price of attacking the wrong target’ Business Day, 27 September 2006) 57 correlate them to some notional competitive price to be overly mechanistic and contextually unsupported. This reading might have some validity if we were meant to act as price regulators and to order the price back down to the non-excessive level. We have already firmly rejected the implicit contention that the sparse wording of Section 8(a) is intended to convert us from an agency that promotes and protects competitive market conditions to an agency that determines price through the simulation of competitive market conditions.

[156]We should add that another considerable problem that arises from an approach that dictates that we must find three precise levels (namely, an excessive price, a fair price and an economic value) is that it is too static. Prices and costs are seldom static over time even when one may be dealing with the exercise of monopoly power. In order to come to a conclusion under 8(a) there needs to be some administrable durability to the pricing, one that will have some applicability over time. Thus price movements are likely during any reasonable period of measurement. Hence the Act, on our interpretation, is an enquiry into less precise, but more durable, relationships between the respondents’ alleged excessive price and its economic value or its costs. These relationships are the product of the structure of the market and the ancillary abusive conduct enabled by the structure and hence the emphasis in this decision on those aspects in the evidence.

[157]European excessive pricing jurisprudence strongly evidences the pitfalls of the many alternative conceptions of the meaning and measure of value and cost that are all, in one way or another, bedevilled by the mistaken notion that ‘economic value’ has intrinsically measurable content. United Brands suggests comparing ‘if it were possible’ the selling price and the costs of production which would then reveal the profit margin and then notes ‘the very great difficulties in working out production costs’. 103 Napp preferred to rely on a comparison between prices in different market segments although the CAT conceded that ‘there may well be other ways of approaching the issue of unfair prices’ including comparing prices to costs, Napp’s prices compared to the costs of its next most profitable competitor and Napp’s prices compared to its competitors. 104 The latter measure was favoured by the court in General Motors Continental NV, 105 while the court in British Leyland 106 preferred to 103 See United Brands judgment, paragraph 251. 104 See Napp Pharmaceutical judgment, paragraphs 391-392. 105 See General Motors Continental NV v Commission of the European Communities, Case 26/75. [1976] 1 CMLR 95 58 evaluate the extent of price movements over time. In the Port of Helsingborg case the European Commission cautioned that the difference between revenue and cost was not a sufficient basis for finding excessive pricing and suggested that ‘customers demand’ was also a relevant determinant of price. In this case the Commission attempted to compare the price charged for ferry services (an uncompetitive market) with the price charged by the same port for cargo services (a competitive market) but concluded that the services offered on the two markets in question were themselves not sufficiently comparable. 107 In Ministere Public v Jean-Louis Tournier the basis of comparison used by the court were fees charged in different member states of the European Union, the so-called ‘geographical comparison’ test. 108

[158]This brief survey offers a mere sample of the diverse measures attempted by the various courts of the European Union and some of its member states, all of which unfailingly emphasise the difficulties and shortcomings in the measures that they used. Indeed the most revealing comment probably comes from United Brands, which after a survey of a large range of possible measures of excessive prices notes ‘Other possible measures may be devised – and economic theorists have not failed to think up several – of selecting the rules for determining whether the price of a product is unfair’.

[159]It is clearly not possible to glean a single European conception of a pricing standard and measure of excessiveness, that is, an arithmetic relation between price and an intrinsic, measurable economic value. Indeed a survey of the European jurisprudence serves to confirm our view of the pitfalls of competition authorities assuming a price regulating function as part of their excessive pricing jurisdiction. It is precisely to avoid the confusion and uncertainty generated by the jurisprudential maze outlined above that price regulators are accorded a specific statutory basis which assigns them appropriate price determination powers and, indeed, often prescribes the specific price determination mechanism that is to be employed.

[160]We need not detain ourselves further with this. As indicated, our view of pricing is that a price that is ‘reasonable’ in relation to economic value is one that 106 See British Leyland Public Limited Company v Commission of the European Communities, Case 226/84. 107 See Scandlines Sverige AG v Port of Helsingborg, Case COMP/A.36.568/D3 108 Ministere Public v Jean-Louis Tournier, Case 395/87, [1991] 4 CMLR 248, at paragraph 38 59 emerges from, in the words of Padilla and Evans, the ‘free interaction of demand and supply in a competitive market’. At the risk of repeating ourselves, where that market structure admits – as it often does – of a degree of pricing power, the firm will rationally conduct itself in a manner designed to maintain and extend its structural advantages and hence its pricing power.

[161]As we have already elaborated much of this conduct is legitimate, pro- competitive conduct. Hence product innovation that seeks to differentiate the firm’s output from that of its rivals and so maintain its structural advantage and hence pricing power in a newly crafted market niche, exemplifies a pro-competitive quest for market power. Process innovation designed to achieve a position on the cost curve that permits a firm to lower price and increase market share and so consolidate or even extend its structural position is another standard category of pro-competitive conduct. All of this conduct produces market and, hence, pricing power – indeed if a patent is successfully registered with respect to any of these innovations it may produce monopoly market power. However, as elaborated, the price that results from the exercise of power so acquired may thereby meet the test of ‘reasonableness’ that is required by the Act’s definition because it is the product of competition as reflected in innovation.

[162]On the other hand, conduct by a dominant firm that is directed at excluding rivals, potential or actual, may, depending on the nature of the conduct in question, well fall foul of one or other of the exclusionary abuses of dominance described in Section 8 of the Act. As already indicated, the focus of abuse of dominance on exclusionary abuses is predicated on the view that, in the absence of specific exclusionary conduct, attempts to abuse a structural advantage by excessive pricing or by optimising the quiet life (inefficiency) will be met by new entry. Thus the price of the ‘merely’ dominant firm may well reflect a degree of pricing power but this will not be excessive because it is constrained by competition considerations, namely, the contestability of the market, and anti-trust is directed at ensuring, through the proscription of exclusionary abuses, the maintenance of this contestability in order to prevent an ‘exploitative’ abuse of dominance or ‘excessive pricing’.

[163]However in this case Mittal SA’s super-dominance, its uncontested control of an incontestable market, is the underlying basis for exceptional or ‘excessive’ pricing power in its domestic market. But even its ability to price to the very limit of its monopolistic power in the relevant geographic market is, super-dominance 60 notwithstanding, potentially undermined by the fact that its level of output exceeds the level of demand for steel in the South African economy at the monopolist’s target price which, in this case, is the import parity price, the price of procuring steel from beyond the boundaries of the relevant geographic market. And so, in order to maintain its target price in its domestic market it is compelled to dispose of this surplus output in markets – principally in the international market but also in certain limited domestic market niches - where competitive conditions oblige it to accept a market determined price (a ‘competitive price’) that is below its domestic target price. This immediately opens up the possibility for arbitrage between the separately priced markets and downward pressure on the price in the higher priced segment. Mittal SA’s response to this threat is to engage in the most fundamental and egregious monopolistic conduct – it has effectively withheld supply from the domestic market by studiously and purposefully preventing arbitrage, by maintaining the segmentation of its markets. Formally expressed, and as earlier noted, it has contrived to move its domestic supply curve leftwards along a downward sloping demand curve, the better to achieve a price target approximating to the profit maximising price of a domestic monopolist.

[164]This puts beyond doubt Mittal SA’s contravention of Section 8(a) – it has, by virtue of its super-dominance, the structural market power to pre-select a target price in its domestic market; it has imposed this pre- determined target price, in this instance the import parity price, on most of its domestic market; and to support this pre-selected target it has withheld supply from the domestic market, the most elementary and offensive of monopolistic conduct. This does not mean that Mittal SA’s offence is its super-dominance; nor does it mean that a firm that is not super-dominant is not entitled to manipulate its supply – if such a firm reduced output its competitors would quickly move to replace the lost output. However, it is the cumulative impact of its structural advantage (its super-dominance) and the conduct thereby enabled (its purposeful, unilateral withholding of supply from the domestic market) that results in a price that is unconstrained by any competitive considerations whatsoever. 109 By withholding output, an option only available to a 109 We use ‘purposeful’ advisedly. Note statements from the transcript cited below where Mr. Dedman acknowledges that the purpose of the market segmentation that is maintained through the Macsteel 61 super-dominant firm, it has assured its ability to charge its pre-selected target price, a price unconstrained by any competitive considerations in its relevant market, and thus has contravened the Act’s proscription of excessive pricing.

[165]Although it is not necessary for us to answer the question, it is nevertheless interesting to ask why Mittal SA has not produced precisely to the level necessary to accommodate its pre-selected domestic price. 110 There are a number of likely answers. Firstly, it is probably as difficult for a steel monopoly as it is for a central government planner to plan its output as closely as that. It would not want to risk getting into a sustained position of excess demand and serious steel shortages and rationing at its pre-selected price and nor would it want to overestimate demand for fear of imposing downward pressure on price. Moreover planning at this level of detail would require a degree of flexibility in managing output levels that is probably not permitted by the large integrated steel mill technology that it manages. Secondly, and most important, scale economy considerations probably dictate that, if Mittal SA wishes to achieve the cost position of an internationally competitive steel producer, it produce at a level of output that is in excess of local demand at its pre-selected domestic target price. 111 In other words, Mittal SA wants it both ways: it wants to produce at the lowest possible point on its cost curve, the better to engage profitably in international markets, but, in order to maximise its profits in its monopolised domestic market, it denies the fruits of these necessary scale economies to its arrangements and the various resale conditions contained in the rebate schemes is precisely to maintain the pre-selected target price. 110 Note that Mr. Dednam clearly understands that a reduction of total output is an alternative to its current approach which limits supply to the domestic market whilst simultaneously exporting at a price lower than that attainable on the domestic market:’ Mr Dednam: Yes its rather difficult then in trying to arrive at a mechanism that is good. So, the alternative that’s been put to us on the table is well why don’t you use your export parity price as the comparator to your domestic price levels. What will happen if we do not export at all? How do we determine our domestic prices then and how would we get to then the competitive price, if we had not had the capacity to export and we curbed our total production to such an extent that we are only geared to service the domestic market? Will we then say to the domestic market that the price we are going to charge you will be exactly the same as what other domestic consumers will pay in other markets?’ (Transcript, page 1868). The answer of course is that if Mittal cut back actual production to the extent of current domestic demand at the prevailing price and actually ‘hit the target’, then the outcome would be the import parity price, the price presently targeted, and which is achieved by denying the domestic market output by means of market segmentation rather than by cuts in total production. However, as explained in the rest of this paragraph, it would not cut back production principally because this would raise its costs per unit of steel produced thus reducing its margins in the domestic market and it would be denying itself access to an international market which, we believe, is, given its scale economies and cost advantages, clearly profitable. 111 In relation to its long steel products Mittal SA’s website explains that ‘the two mills {at which long steel products are produced} account for total annual sales of 1.9 million tones, half of which is exported due to the limited demand of the RSA market’ 62 domestic customers by restricting supply in the this market, the better to extend its margins in the domestic market. Of course, in addition to scale economies, the underlying cost advantages that Mittal SA enjoys in South Africa – these by virtue of, inter alia, its iron ore price and electricity price – also contribute to Mittal SA’s export competitiveness. Indeed, as we shall elaborate below, the evidence suggests that it is certainly profitable for Mittal SA to export even at the lower prices that it realises on the international market.

[166]In short, there is no magic in Mittal SA’s mechanism to achieve its targeted price in its monopolised domestic market. It does what any elementary textbook on competition law and economics will reveal in its opening chapter about monopolistic conduct – it exploits its structural power by reducing output in order to increase price. In this instance it does not physically reduce its level of output. But it contrives, as we shall elaborate below, to remove this ‘excess’ output from the domestic market in order to maintain its pre-selected target price, which closely approximates, though, as we have noted, occasionally deviates from the import parity pricy. This it achieves by resort to the ancillary conduct to which we have referred and which we now proceed to examine. Ancillary abusive conduct

[167]The ancillary conduct by which Mittal SA ensures its ability to price excessively comprises principally the segmentation which it studiously maintains between its export market and its domestic market. It markets its domestic output through a group of traders whose activities are confined, by agreement with Mittal SA, to the domestic market. And its markets its international output through a single trader, Macsteel International, the second respondent, whose trading activities are confined, again by agreement with Mittal SA, which is a 50% joint venture partner in Macsteel International, to the international market. Hence the pertinent terms of the agreement between Macsteel Holdings and Mittal SA that govern the joint venture, Macsteel International, the second respondent, are: 1. Mittal SA undertakes to market a specified range of steel products, which include flat steel, only through Macsteel International in the international market; (clause 30.1) 2. Macsteel International undertakes not to sell any of these steel products in the domestic market without Mittal’s consent; (clause 29.1) 63 3. The Macsteel Group also makes a similar undertaking to Mittal SA that as long as the agreement persists it will not market Mittal SA products in the international market other than through the joint venture (clause 31.2); and 4. Mittal SA undertakes to sell steel to Macsteel International at “international related market prices”. (clauses 29.2.1. and 30.2)

[168]The arrangement with Macsteel International, the second respondent, is the essential ancillary conduct whereby Mittal SA abuses its structural advantage to maintain its pre-selected price level. It is, of course, conduct that is only available to an uncontested firm in an incontestable market. If this were not the case, Mittal SA’s traders would be able to turn to alternative suppliers of flat steel products in order to meet demand that is unrealised at Mittal SA’s pre-selected domestic price level. Mittal SA of course wishes to create the impression that there is no unrealised demand. It wishes to create the impression that it satisfies all domestic demand and that its export activities are simply a vent – an unprofitable vent, it moreover alleges - for a surplus that it would much rather sell into the more lucrative domestic market. However this is, at best, only a half-truth and one that, when fully considered, does considerable violence to the whole truth – certainly, Mittal SA meets all domestic demand, but, and here is the crucial caveat, at its unilaterally targeted price level. If a would-be purchaser of steel for use in South Africa were to approach Macsteel International with an offer to purchase at a price below the prevailing domestic price but above that which Macsteel International could realise on the international market, the trader would, as a matter of profit maximising rationality, accept such an offer. However, it is by agreement with Mittal SA, prohibited from accepting the offer because to do so would, as Mr. Dednam candidly concedes, be to reduce the price of steel across the whole range of Mittal SA’s domestic sales. Its willingness to enter into such an agreement is, of course, predicated on Mittal SA’s super dominance – to express it crudely, a firm that wishes to trade in South African steel is obliged to accept Mittal SA’s trading conditions. The economics is disarmingly elementary – indeed it is the first principle of monopolistic conduct.

[169]The segmentation that is at the heart of the Macsteel International JV is not the only segmentation that Mittal SA maintains. We have already outlined how Mittal SA segments the bulk of its domestic consumers from consumers in those market segments in which it faces more competitive conditions. These are the markets in which Mittal SA discounts its price. As we have outlined above, from the perspective 64 of this decision, the hallmark of these arrangements is not the level of the various discounts relative to the domestic price, but rather the anxious efforts by Mittal SA to ensure that these market niches are segmented from the rest of the domestic market.

[170]There is no need, for present purposes, to describe or analyse the Mittal SA rebates in any great detail. A brief outline of some of its rebate programmes will establish how they clearly demonstrate Mittal SA’s overriding anxiety to ensure that as much of its domestic market as is possible pays the pre-selected target price.

[171]There are a range of rebates on offer. Mr. Dednam characterised them as follows: The strategic rebates, we have basically 3 types of strategic rebates that we do. The one is for replacement of value-added product imports. The other one is basically for value-added secondary steel exports, and the other one is basically to be competitive against substitute products in the marketplace itself. 112

[172]These rebates have already been referred to in passing and there is little reason to elaborate the detailed rationale for, and operation of, each of them. They are, in fact, reasonably well-described by their names. Hence when the domestic market of a local manufacturer of a steel intensive product is being displaced by imports of the product, Mittal SA will consider a rebate on the steel input in order to retain the domestic market for the domestic fabricator provided that it, Mittal SA, is persuaded that it is a differential in the steel price that is at the root of the decline in the competitiveness of the South African producer. Or similarly where a domestic producer’s export market is being eroded by fabricators elsewhere, Mittal SA will also consider rebating the steel input into the exported product in order to retain the South African manufacturer’s international market. Competition in the instances described may not only come from cheaper steel but from substitute products such as plastic or aluminium or cement.

[173]We have earlier described and analysed the auto sector rebate, far and away the largest of Mittal SA’s rebate system, from the perspective of market power. We have, in our discussion of the auto rebate, already noted that it is characterised by 112 See transcript of 5 April 2006, page 1658. 65 capricious shifts in the scale of the discount that it provides, a pattern which is indicative of Mittal SA’s market power and this is as, if not more, evident in the rebates granted Mittal SA’s weaker customers. In support of this we cited the evidence of Mr, Leatherbarrow, a pipe and tube manufacturer, and Mr. Cohen, a manufacturer of conveyor belt systems. What we have not yet mentioned is the extent to which the rebate programmes are characterised by a striking combination of, firstly, surgical precision in deciding which sales to discount and, secondly, the extraordinarily complex bureaucracy required to actually pass the rebate on to the carefully chosen beneficiary.

[174]In fact this combination of precision and bureaucracy is all of a piece. If a local fabricator is producing for the domestic market and the international market and it has successfully persuaded Mittal SA that it requires assistance to maintain or extend its presence on the international market, then Mittal SA’s overriding concern will be to ensure that the only steel that gets rebated is that which is exported – conversely expressed, Mittal SA is anxious to ensure that rebate does not undermine its ability to extract its pre-selected target price on the steel used by the fabricator for output sold on the domestic market because this is not where Mittal SA’s rebate is ‘required’.

[175]Consider the case of GRS, a firm producing steel roofing tiles, which was granted a rebate by Mittal SA in order to enable it to better compete with producers of cement roofing tiles. In the case of the hapless GRS, where Mittal SA is granting a rebate in order to enable a customer to compete more effectively with a roofing tile product produced with an alternative material, namely, cement, Mittal SA’s overriding anxiety is to ensure that the rebate does not extend beyond the precise competitive disadvantage of its customer. Not only did this give rise to wild fluctuations in the size of the rebate – its gyrations were, it seems, determined by both movements in the price of cement and the price of steel - the rigmarole involved in actually computing the precise level of the rebate ‘required’ could be considered quite comic, if it were not indicative of the deadly serious business of a super-dominant firm intent upon maintaining the careful segmentation of markets necessary to achieve its target price in a monopolised market. When Mr. Van der Westhuizen, a witness from GRS, was asked to explain why there had been a sudden significant decrease in the size of the rebate granted by Mittal, he testified: 66 I asked the question as to why and the reply I got was that the price of cement has increased and for that reason the differential that we require or the discount that we require to compete has reduced. 113

[176]In order to persuade Mittal SA to grant it a larger rebate Mr. Van der Westhuizen and his team then presented to Mittal SA ‘ ..a full study of a roof, the timber structure included as well as the covering, which could be the steel or the concrete plus all accessories, all the thrashing that necessary (sic) to finish off a roof and we presented the two total answers, in other words, what it will cost to do a roof in concrete and what it would cost to do a roof in steel tile. That we then presented to Mittal’. 114

[177]Our point here is not whether GRS proved capable of persuading Mittal SA to grant a larger rebate – it seems to have achieved partial success. The question is as stated above: is this the conduct of a large steel producer intent upon squeezing the last cent out of a small customer, or is it a large steel producer with a large domestic market intent upon protecting its default target price over the whole of that market? The answer is that it is clearly the latter. It is Mittal SA ensuring that its rebate is not ‘excessive’, that is, that it does not inadvertently price steel in the domestic market at a level below its default target price unless it is absolutely required to do so by highly particular and proven competitive considerations. It fears that a discount granted ‘unnecessarily’ may inadvertently start to exercise downward pressure on the pre- selected target price generally charged in the domestic market.

[178]Mr. Roy Cohen, whose firm, Melco, is a significant exporter of fabricated steel products, was asked by the complainants’ counsel to describe the process whereby his company gained access to Mittal SA’s rebate for secondary exporters. He answered: ‘We as a company have opted to use a consultant. A consultant is more familiar with the rules and the interpretation of the rules. We started off the first time we exported trying to submit the product ourselves, the application for rebate ourselves and found that it was that difficult and complex and required that much attention to the detail and an understanding of the 113 Transcript, pages 765-766. 114 Transcript, page 766. 67 interpretation of the rules, that it was easier for us just to pay a fee to the consultant and he undertakes that for us and submits the claims for us.’ 115

[179]This is representative of the experience of other recipients of Mittal SA’s rebate. We recount it not in order to establish that Mittal SA possesses market power, nor is it intended to evidence Mittal SA’s parsimony in the granting of rebates. It is rather evidence of the highly selective manner in which it granted any rebate at all and the care it took to ensure that rebating remained the exception, with the default target price strictly maintained as the norm. It also explains why Mittal SA used a system of rebating rather than discounting – before paying out the rebate it required the beneficiary of the lower price to establish that the steel subject to the rebated price had actually been used for its intended purpose and its intended purpose only. Conversely expressed, Mittal SA required the beneficiaries to demonstrate that by granting a lower price it, Mittal SA, was not putting its superior margins in the South African market at risk. Mr. Dednam clearly articulates the overriding purpose of the segmentation: ‘We are offering a very, very good price into the market that is equal to our worst export price that we are offering into the market. And the way that we are offering the price into the market is we would not like that particular price eventually to erode the rest of the business and other transactions that’s taking place in the domestic market at a domestic price level. For that reason we’ve introduced all these measures to ensure that the product that we are offering for that particular application will eventually end up in that particular application and will eventually be exported.’ 116

[180]Mr. Dednam characterises these as efforts to prevent the creation of a ‘grey market’ in steel. 117 115 Transcript, page 406. Indeed a reading of the evidence-in-chief of Mr. Cohen conclusively establishes the truly grotesque complexity of the rebate schemes. But for the fact that we found Mr. Cohen to be a particularly well-informed witness whose account of his dealings with the steel producers and merchants was not subject to serious challenge by Mittal SA’s counsel, his testimony could well have been received as a satirical lampoon of a particularly bizarre series of business dealings. See transcript pages 399-445. 116 Transcript, page 1663 (Our emphasis). 117 Transcript, page 1665. 68

[181]Mittal SA’s discounting policy is then fairly characterised as a privately conceived and managed industrial policy designed to maintain and promote the utilisation of steel products in the South African economy – South Africa is, after all, the geographic market in which Mittal SA earns its most attractive returns. But it is applied highly selectively and with excruciating care lest it infect the higher margin segments of its domestic market. Mittal SA’s margins in the South African market are a product of the price that it is able to charge over the significant bulk of its domestic sales and it will not compromise those margins, even if this permitted an expansion in the South African market for flat steel products because it clearly calculates that the additional volumes thus gained would be at the expense of a diminution of the margins over all of its domestic sales. Accordingly, where the utilisation of South African steel in clearly identifiable downstream market segments is under threat Mittal SA will isolate those segments from the broader market and it will offer a rebate, always ensuring that the rebate does not ‘contaminate’ those vast swathes of the South African market over which its domination and its pricing power is uncontested. While market segmentation and price discrimination is a hallmark of an efficient monopolist, its unintended consequences may be severe if it enables ‘non- deserving’ customers to take advantage of the available discounts, thus ‘cannibalising’ its high margin market segments.

[182]We should add that this practice of segmenting its differently priced markets in order to protect its ability to charge its targeted domestic price has, in the past, been bolstered by other conduct, equally offensive from a competition law perspective. Hence, as already explained, prior to the acquisition of Saldanha Steel by Iscor – Saldanha having previously been controlled by the Industrial Development Corporation and Iscor – Saldanha and Iscor concluded a market sharing arrangement in terms of which Saldanha agreed that it would not make its output available on the domestic market. Where it faced potential competition from another domestic steel producer, Mittal SA (or Iscor as it then was) sought to rely on a market sharing cartel to maintain its pre-selected domestic price. Having absorbed, through merger, its potential competitor, it no longer required an anti-competitive horizontal agreement to divide markets. It now relies on the unilateral imposition of anti- competitive conduct – effectively a withholding of supply from most of its domestic market - in order to price to the full extent of its uncontested market power. 69

[183]Nor is this the only market in which we have encountered this practice. In the Tongaat Hullet/TSB 118 sugar merger and in the Sasol/Engen liquid fuel merger, both of which were prohibited, the firms in question maintained the targeted import parity pricing despite the existence of excess supply at that price level. In both the sugar and fuel markets this was achieved, despite competitive market structures, by means of anti-competitive conduct – in these instances a combination of regulation and collusion – which effectively segmented the international and domestic markets. We observed in Sasol, and this has been much referred to in these hearings, that ‘import parity pricing of fuel – or BFP – in a fuel exporting economy can only be artificially maintained by administrative fiat…or by collusive agreement…or by the unilateral exercise of market power.’ 119

[184]This conclusion is confirmed by the facts of this case. Here the mechanism for maintaining the target price is neither regulation (there is none) nor cartelisation (there is a single producer), but through the unilateral action of a monopolist imposing upon those that trade in its product conditions that are designed to reduce the supply available on the domestic market.

[185]Mittal SA has argued that this conduct should have been attacked under other provisions of the Act. It has been suggested that the practice that is, in reality, under attack is that of price discrimination and that the complainants should have brought suit in terms of Section 9, the provision of the Act that proscribes price discrimination and that would have allowed Mittal SA defences not available under Section 8(a). It is also suggested that the conduct complained of is in reality the exclusive dealing agreement between Mittal SA and Macsteel International, a potential contravention of Section 8(c), which, in contrast with the prohibition provided for in the Section 8(a), requires the complainant to establish that the anti-competitive harm outweighs any pro-competitive consequences.

[186]We do not know whether the complainants have made their election on the basis that a Section 8(a) offence is easier to prove than a Section 8(c) or Section 9 offence, but simply observe that this constitutes perfectly legitimate and appropriate legal advice. We do, however, note that the extent of dominance that we require in order to prove a Section 8(a) offence requires that the complainant clears a 118 See Tongaat Hullet Group and Transvaal Suiker Bpk [1999-2000] CPLR 127 (CT). 119 Sasol Ltd and others / Engen Ltd and others [2006] 1 CPLR 189 (CT) at 233F-G. 70 particularly high hurdle in proving its case, considerably higher than the ‘mere’ dominance required by Section 7 in order to prove other abusive conduct or price discrimination.

[187]Moreover, it is clear from the evidence that the complainant and the other witnesses who are steel users believe that their commercial activities are disadvantaged in their respective market places because of the price charged for a ‘must have’ input, steel. This is what has caused them to invoke Section 8(a) of the Act - they are aggrieved by the absolute price charged for steel. Hence what we have been asked to pronounce upon is the inability of domestic producers to bargain over the price of Mittal SA’s excess supply of flat steel products. What clearly emerges is dissatisfaction – and a degree of puzzlement – at the notion that domestic steel prices are based on market conditions in distant markets rather than on supply and demand conditions in the South African market and that notional transport charges are levied on a product that is not, in physical reality, transported over the vast and costly distances that nevertheless constitute an important element of the domestic price. The complainants clearly believe – as, we should add, does the steel expert Mr. Fish 120 – that the price of steel should be determined by local demand and supply conditions which, in their estimation – and they are correct – would, if free of abusive conduct, produce a lower steel price. Hence the allegation has been one of excessive pricing.

[188]Comparative prices – and thus the question of discrimination – only entered the equation because European jurisdiction, at least in some of its variants, has approached the question of excessive pricing by comparing prices in different markets and, in the South African case, the obvious markets to compare are the domestic and international markets in which Mittal SA trades.

[189]We however have not approached the matter in this way. We have rather asked whether the structure of the market admits of the possibility of excessive pricing. This, as we have elaborated, requires a showing of exceptional or super- dominance. We have then examined the relationship between this super-dominance, on the one hand, and Mittal SA’s ability to price up to its pre-selected price target, particularly given the existence of excess supply at that target price. What we have found is that Mittal SA employs its super-dominance to achieve its target price by 120 See Mr Fish’s testimony, transcript of 22 March 2006, pages 520-521. 71 ensuring that the excess supply that exists at that price is removed from the domestic market and that it does not re-enter the domestic market again. And because Mittal SA has no domestic competition to speak of it does not have to fear new supply from a domestic source. This reduction of supply is the essence of its agreement with Macsteel International and it is further built into the manner in which it grants rebates off its list price to selected domestic customers. The result is a price in excess of that which would prevail in the absence of Mittal SA’s super-dominance and the ancillary conduct that it enables.

[190]We should add that a careful reading of United Brands, evidences an important conformity between our reasoning and that of the European Commission’s advocates before the court. Subsequent reference to this case in litigation involving excessive pricing allegations has tended to focus exclusively on the European Commission’s and the Court’s approach to pricing comparisons. In the confusion and obfuscation that arises from an examination of the European authorities’ decisions in the matter of price determination, it is often forgotten that in this leading case the respondent, United Brands, was accused – and found guilty – of a number of abuses of its dominant position. Principal amongst these was the imposition by United Brands of conditions of resale upon those firms that purchased its green bananas and ripened them for on-sale into the various national European markets. In commenting on these conditions of resale, the European court reports: According to the Commission these prohibitions and practices are both the essential constituent of an overall system enabling the applicant to control the entire marketing of its product and to restrict competition and also form the basis of the three other abuses for which ubc (United Brands) is blamed. 121

[191]We emphasise that one of the ‘three other abuses’ referred to is that of excessive pricing. This, the identification and elimination of the ‘overall system’ or the ‘basis’ of the abuses, rather than price determination, is the approach that we favour. 122 121 See para 137 (Our emphasis). 122 We are also, of course, at one with the European Commission’s summary of its position following United Brands and which is cited in the Mittal SA HOA: ‘The existence of a dominant position is not itself against the rules of competition. Consumers can suffer from a dominant company exploiting this position, the most likely way being through prices higher than would be found if the market was subject to effective competition. However, the Commission in its decision making practice does not normally control or condemn the high level of prices as such. Rather it examines the behavior of the dominant company designed to preserve its dominance, usually directly against competitors or new entrants who 72

[192]The other leading European case on excessive pricing is that of Napp Pharmaceuticals. Again subsequent citations from the Napp decision have focused on the question of a comparison between the price levels in the two markets in which the product in question was sold. However it is absolutely clear that the excessive price charged was a product of the careful segmentation of the two markets in question and the practice of predatory pricing in the lower priced of the two markets which had the effect of excluding new entrants to the higher priced market thus enabling the charging of ‘quasi-monopoly’ prices in that segment. Again we have in Napp Pharmaceuticals garden-variety monopolistic conduct with the respondent segmenting its market so as to produce very low, in fact clearly predatory, prices in one market which enabled it to limit entry – that is, control the supply available – to the second market that then resulted in extremely high prices in the second market. 123 In our view the courts in both United Brands and Napp would have better served competition jurisprudence by restricting themselves to identifying and eliminating the basis for monopolistic price determination rather than by would normally bring about effective competition and the price level associated with it. A dominant company therefore has a special obligation not to do anything that would cause further deterioration to the already fragile structure of competition or to unfairly prevent the emergence or the growth of new or existing competitors who might challenge this dominance and bring about the establishment of effective competition’ XXIVth Report on Competition Policy (1994), point 207 cited in Mittal SA’s Heads of Argument at para 5.9 (Our emphasis). Mittal SA cites this passage in support of its view that we focus on exclusionary abuses despite its concession that both South African and European law explicitly condemn the exploitative abuse of excessive pricing. However the position that explicitly informs our decision is at one with the view that examines and condemns the basis for excessive pricing rather than ‘the high level of prices as such.’ That practice, as we have emphasized at length, is not the exclusion of a competitor (there is none) or a new entrant (barriers to entry do not permit one) but rather the cumulative impact of Mittal SA’s super-dominance and its purposeful withholding of a portion of the supply of flat steel products from the domestic market. 123 Note the CAT’s summary at para 364 of the arguments of the Office of Fair Trading in Napp:’In essence, the Director General’s case is that Napp charges excessively low and/or discriminatory prices in the hospital segment and thereby sustains very high prices market shares in the community segment of the market. These two aspects are accordingly interlinked. Napp’s pricing practices have had the effect of placing significant obstacles against the successful entry of competitors, and in consequence serve to preserve its quasi monopoly position in the community segment of the market and enable it to continue to charge prices for MST higher than could be sustained in the absence of that quasi- monopoly position ie competitive prices…Accordingly the Director General does not seek to condemn the price in the community segment in isolation; in other words, if his case should fail as regards the exclusionary character of Napp’s pricing practice in the hospital segment, he does not contend that the prices in the community segment violate the Chapter II prohibition simply because of their absolute level’. The CAT continues: As stated above, it is the Director General’s case that Napp’s conduct has had the effect of excluding competitors from the hospital segment, thereby foreclosing the essential gateway for entry to the community segment. As a result Napp has been able to charge quasi- monopoly prices. In those circumstances, the charging of prices, which are higher than Napp would have been able to charge in a competitive market, constitutes an abuse.’ (Napp, paras 364-365) 73 supplementing the torturous European jurisdiction that has attempted to pass judgement on particular price levels and their differentials. 124

[193]We should emphasise that just as we have not approached the question of determining whether or not the price is excessive by a determination of the ‘correct’ or ‘incorrect’ level of the price, so, obviously, do we not approach the question of whether or not the price charged is ‘excessive’ or ‘unreasonable’ by reference to the scale of the difference between two price levels.

[194]Our approach to the question of establishing whether the difference is ‘excessive’ is consistent with our approach, as a competition authority, to the question of pricing in general. We have – following our Act and the principles and practice of competition law - sought an ‘economically reasonable’ explanation for Mittal SA’s pricing, that is an explanation that is not rooted in super-dominance, that is, in an absence, as the Napp Tribunal put it, of ‘significant competitive pressure’ and then in ancillary conduct designed precisely to manifest that super-dominance in a supra-competitive price level. That the price charged is clearly in excess of that which would have been charged had Mittal SA not been super-dominant and not taken advantage of its super-dominance to reduce the output of steel available to the South African market, is clearly conceded by Mr. Dednam. We have earlier cited his negative view of price wars, a potential reality of competitive life for all but the super- dominant and effectively discounted any possibility of a price war with Highveld Steel. We have also cited his explanation of the segmentation of Mittal SA’s rebated 124 Napp Pharmaceuticals produced a slow release morphine tablet for terminally ill patients which it sold to the public health authorities at a predatory price which effectively foreclosed this market, the hospital market, from other competitors. When patients who had, in their hospital treatment, been prescribed the Napp product left the hospital market and were prescribed the Napp product in the ‘community market’, it was only made available in this latter market at a significantly higher charge. In short, by charging a predatory price in the first market Napp effectively foreclosed entry by competitors into the hospital market which constituted a critical ‘gateway’ for participating effectively in the second market. Thus its predatory conduct in the first market effectively prevented entry by competitors into the second market , thus according Nap super-dominance which enabled it to charge an excessive price in the community market. The CAT decided the question of excessiveness on the basis of price comparisons between the markets in question and then, in addition to imposing an administrative penalty, ordered the price in the second market to be decreased by a seemingly arbitrary 15%. However the competition process may have been better served had CAT eliminated the underlying ancillary conduct (price predation) that led to Napp’s super-dominance in the first market, thus preventing entry through the ‘gateway’ to the second market. The upshot would have been an increase in prices in the relatively small hospital market which would have enabled competitive entry into that market. Entry into that ‘gateway’ market would then have enabled new entry into the much larger community market and this would, through the process of competitive entry, have resulted in a decrease in prices in the community market where prices were found to be excessive. 74 segment from the rest of its South African market as rooted in an effort to prevent the rebated price ‘..eventually to erode the rest of the business and other transactions that’s taking place in the domestic market at a domestic price level.’

[195]But, most important, Mr. Dednam also explicitly conceded that were Mittal SA not deliberately, through its agreement with Macsteel International, relying on its super-dominance to engage in the ancillary conduct of ‘shorting’ the South African market, that is, deliberately depriving South African customers of steel produced by Mittal SA, the price of its flat steel product would be lower than it is at present. Note the following exchange between the Tribunal and Mr. Dednam: Chairperson: I’m saying that my understanding is and my understanding might be wrong, so correct me, but that Macsteel International sells on the export market and only on the export market and let’s say the price it pays there is $400,00. ISCOR sells to its least preferred customers, if I could put it that way, on the domestic market at import parity price. That’s $600,00. The big merchants get a volume discount that allows them to purchase this product at $500,00, let’s say. I don’t know that there are any restrictions on the amount of that discount that they retain or the amount of that discount that they pass on, although I note that in the sort of pricing schedules what seems to be passed on is simply the ISCOR price, which kind of makes me think that they don’t pass on the volumetric discount that much and when they tell their customer why their price has changed, they simply reflect the change on the ISCOR price. And then you have a whole lot of special customers like the car industry and others that get their price at $450,00. What would happen? I mean obviously the particular issue is between Macsteel International and the IPP price, but there are all sorts of other prices in between. What happens if all of these were open to trading steel in the domestic market and in the international market? In other words, if Macsteel International wanted to trade in the domestic market and Trident wanted to trade in the international market and it was prepared to sort of buy 10,000 tons from ISCOR, 2000 of which it intended putting into the international market. Under those circumstances what would happen to price? 75 Mr. Dednam: Well, the lowest common denominator in the price would actually prevail. So, if you look at the prices that Macsteel International is basically paying and one should argue that that is the price that should prevail for the whole market, then prices will tend down towards that particular level. 125

[196]This is precisely our position. Mittal SA’s leading witness, Mr. Dednam, concedes that structural super-dominance and Mittal SA’s ancillary conduct in maintaining the segmentation of its markets, conduct which itself relies on this structural dominance, has placed a floor under prices, in the South African flat steel market. Conversely expressed, prices would be lower but for Mittal SA’s super- dominance which allows it to craft ancillary conduct that enables it to reduce supply to its domestic market. No more is required in order to prove a contravention of Section 8(a). In our view the temptation, to which the European competition decision makers have generally succumbed, to then attempt to divine the precise level that the competitive market, in its variety of degrees and characteristics, would find, is an incorrect approach for a competition authority to adopt. It is much preferable to identify the underlying bases for ‘maintaining prices higher than would be expected in a competitive market’ and then to eliminate these and then allow the competitive features of the particular market, ‘imperfect’ though they may be, to determine a competitive price level.

[197]Indeed we note that although Mr. Dednam contended that prices would ‘tend’ to the lowest price charged by Mittal SA (that is, currently the price at which Macsteel International purchases steel), he conceded that there were differences in grades of steel and in value added services provided to customers who required these services and that these would serve to maintain a degree of price differentiation in the market. However it would be differentiation based on services provided rather than on a reduction of supply to the market. We agree with this and will discuss it more fully below when we examine, in our discussion of the remedies to be imposed, Mittal SA’s contention that it would no longer be able to operate profitably were prices to decline in response to these competitive pressures. 125 Transcript of 6 April 2006, pages 1848-1849 (Our emphasis). All references to ISCOR can be read as references to Mittal SA. There is value in perusing the entire transcript between pages 1846-1880 which consists of an exchange between the Tribunal, the witness, Mr. Dednam and Mr. Unterhalter, counsel for Harmony. 76 Finding in respect of the excessive pricing complaint – alleged contravention of Section 8(a) of the Competition Act

[198]We find that Mittal SA has contravened Section 8(a) of the Act by charging an excessive price to the detriment of consumers. The Inducement Complaint

[199]The complainants have also alleged that the respondents have contravened Section 8(d)(i) of the Act by inducing its customers not to deal with a competitor. The complainants have asked: A. For an order declaring that the first respondent has engaged in abuse of dominance in terms of Section 8(d)(i) of the Act; B. For an order directing the first respondent to desist from such abuse; C. For an order levying an administrative penalty on the first respondent of 10% of its annual turnover for the financial year ended 30 June 2003 in the South African flat steel market; D. For an order directing those respondents who oppose the complaint to pay the costs incurred by the complainants in prosecuting the complaint.

[200]While we appreciate that the sheer volume of the heads of argument are not necessarily a measure of the merits of the case or the seriousness with which the complainants pursue it, it is tempting to find some significance in the fact that the complainants heads of argument in respect of the excessive pricing complaint run to 230 pages while those submitted in respect of the inducement complaint appear to require no more than 5 pages.

[201]The brevity of the complainants’ treatment of this matter may account for some of our difficulties in gaining a clear understanding of the facts at issue. Essentially the inducement complaint revolves around a single piece of evidence. In brief Mr. Bell of Bell Equipment, a leading South African producer and exporter of specialised heavy industrial vehicles, imported a certain amount of steel which it managed to obtain at a price lower than that on offer to Mittal SA’s domestic customers. As noted in our discussion of the excessive pricing complaint it seems that there are periods when Mittal SA has priced above its pre-selected target price, the import parity price. Bell clearly spotted this and decided to import. 77

[202]The import deal having been concluded, it appears that Mittal SA responded by offering steel at a discounted price to Bell. However Bell was unable to take up the offer because in importing it had to place an order that met its needs for a six month period. 126 It is not clear whether this discount was still available by the time Bell required further quantities of steel.

[203]It is then not at all clear why Mittal SA offered the discount at all. Indeed counsel for Mittal SA suggested that his client had ‘perhaps rather spitefully or otherwise dropped its price’ in the manner described by Bell. There are no domestic producers who compete in the South African market with Bell who were able to take advantage of the discount and so steal a march on Bell who had committed itself to import at what now turned out, as a result of Mittal SA’s belated offer of a discount, to be a premium price. Certainly had Bell been offered the discount prior to committing itself to the import order it would have been better placed to compete in both the South African and international markets with other producers.

[204]We are not satisfied that the facts as presented permit us to get to the bottom of this rather peculiar chain of events. We would want to proceed with particular care on an inducement allegation. After all, on the face of it, the practice of competition consists precisely in inducement. While we can envisage – as clearly does the Act – a species of anti-competitive inducement, the facts of this case are insufficiently clear to arrive at so far-reaching a conclusion.

[205]We should however comment on one aspect of Mittal SA’s response to the inducement complaint. Mittal SA insists that the argument that it has conducted itself so as to induce its customers not to import is tantamount to a concession that it competes with importers who constrain it in it pricing behaviour.

[206]However there is no merit in this argument whatsoever. We have dealt at length with the cellophane fallacy in our treatment of the excessive pricing complaint. This attempt by Mittal SA to effectively divine a concession on the relevant market from the framing of the inducement complaint is nothing more than a case of the 126 What this does evidence are some of the difficulties in relatively small purchasers importing product. Mr. Gary Bell explained that his company would naturally prefer to hold as little stock of steel as possible. However the ability of his company to do so is clearly limited if it has to resort to direct importation. 78 cellophane fallacy writ large. At the heart of the cellophane fallacy is the understanding that even a monopolist is subject to a price ceiling, that is, to a price at which its customers will find alternatives or cease purchasing the product altogether. The framing of the inducement complainant does not eliminate the cellophane fallacy – it draws attention to it and to the persistent failure or, more likely, refusal, of Mittal SA to come to grips with it. Finding in respect of the inducement complaint - alleged contravention of Section 8(d)(i) of the Competition Act

[207]The inducement complaint is dismissed. Remedies

[208]Having found that Mittal SA has contravened section 8(a) of the Act, we must now consider what remedy to impose. For reasons that become apparent later we will not be concluding the issue of remedies in this decision. We will however consider what remedies were sought by the complainants, how they evolved in the course of proceedings, and our competence to consider the remedies. We will not, for reasons explained below, impose the remedies yet. The relief sought

[209]When the complainants referred their complaint to the Tribunal in February 2004 it sought the following relief in respect of the ‘excessive pricing’ complaint. A. For an order declaring that Iscor’s practice of employing import parity pricing (as set out in paragraph 11.1.5 above) in the South African flat steel market amounts to an abuse of dominance in terms of section 8(a) of the Act; B. For an order directing Iscor to refrain from charging excessive prices in the South African flat steel market; C. For an order directing Iscor to levy factory gate prices in the South African flat steel market, irrespective of whether the product is intended for export or not; D. For an administrative penalty to be levied on Iscor of 10% of its annual turnover for the financial year ended 30 June 2003 in the South African flat steel market; 79 E. For those respondents that oppose the complaint to pay the costs incurred by the complainants in prosecuting the complaint. 127

[210]On 26 April 2006 Harmony brought to the Tribunal an application to amend the relief it had originally sought. The amendment sought related only to the excessive pricing allegation and not to the alleged ‘inducement abuse’. Essentially, the complainants sought the insertion of a remedy as an alternative to prayer C in the original relief sought. We set out below the original relief sought with the amendment sought in bold type. “…the complainants intend to apply to the above honourable Tribunal to amend their referral of complaint, form CT1, by substituting the relief sought in the referral in respect of the claim of excessive pricing with the following: “A For an order declaring that the first respondent’s practice of employing import parity pricing (as set out in paragraph 11.1.5 of the founding affidavit) in the South African flat steel market amounts to an abuse of dominance in terms of section 8(a) of the Act; B For an order directing the first respondent to refrain from charging excessive prices in the South African flat steel market; C For an order directing the first respondent to levy factory gate prices in the South African flat steel market, irrespective of whether the product is intended for export or not; C bis In the alternative to prayer C above, for an order directing that: 1 The first respondent may not itself, or with any natural or juristic person, or through any entity, vehicle, trust or other juristic person in which it has an interest, export flat steel products from South Africa; 2 The first respondent divest its interest in the second respondent to an independent third party or parties approved by the Tribunal 127 See Founding Affidavit of Ferdinand Dippenaar, Pleadings Bundle, page 23. 80 within such period and on such conditions as the Tribunal considers appropriate; 3 The first respondent may not: i. impose upon any customer of its flat steel products any condition in respect of the customer’s use or resale of those products; or ii. reach agreement on a condition with a customer of its flat steel products, or enter into any arrangement or understanding with such a customer, in respect of the customer’s use or resale of those products; 4 The first respondent waive in writing any condition in any agreement concerning the use or resale of flat steel products by a customer; 5 The first respondent make known in the public domain, at all times, its list prices, rebates, discounts and other standard terms of sale for flat steel products; D For an administrative penalty to be levied on the first respondent of 10% of its annual turnover for the financial year ended 30 June 2003 in the South African flat steel market; E For those respondents that oppose the complaint to pay the costs incurred by the complainants in prosecuting the complaint; F For an order granting further and/or alternative relief.” 128

[211]On 31 May 2006 the Tribunal heard argument in the amendment application, and subsequently delivered its judgment on 19 June 2006. 129 The Tribunal refused the amendment sought insofar as it related to the insertion of Clauses C(bis) (1) and (2) but it permitted the amendment insofar as the insertion of Clauses C(bis) (3),(4) and (5) is concerned.

[212]Having been granted leave to amend the relief sought by the Tribunal the relief now being sought by Harmony is the following: 128 See pages 1-4 of the amendment application. 129 See the Tribunal’s judgment in the amendment application, Tribunal Case No.: 13/CR/Feb04. 81 “A For an order declaring that the first respondent’s practice of employing import parity pricing (as set out paragraph 11.1.5 of the founding affidavit) in the South African flat steel market amounts to an abuse of dominance in terms of section 8(a) of the Act; B For an order directing the first respondent to refrain from charging excessive prices in the South African flat steel market; C For an order directing the first respondent to levy factory gate prices in the South African flat steel market, irrespective of whether the product is intended for export or not; Cbis In the alternative to prayer C above, for an order directing that: 1. The First respondent may not: (i) Impose upon any customer of its flat steel products any condition in respect of the customer’s use or resale of those products; or (ii) Reach agreement on a condition with a customer of its flat steel products, or enter into any arrangement or understanding with such a customer, in respect of the customer’s use or resale of those products; 2. The First Respondent waive in writing any condition in any agreement concerning the use or resale of flat steel products by a customer; 3. The First Respondent make known in the public domain, at all times, its list prices, rebates, discounts and other standard terms of sale for flat steel products; D For an administrative penalty to be levied on the first respondent of 10% of its annual turnover for the financial year ended 30 June 2003 in the South African flat steel market; E For those respondents that oppose the complaint to pay the costs incurred by the complainants in prosecuting the complaint. F For an order granting further and/or alternative relief.” 82 Our remedial powers

[213]The Tribunal is empowered to impose wide ranging remedies for contravention of the Act. Of relevance here, our remedial powers include interdictory relief, declaratory relief, the imposition of an administrative penalty, an order of divestiture and the power to void all or part of an agreement. These are set out in Section 58, the relevant clauses of which are set out below: (1) In addition to its other powers in terms of this Act, the Competition Tribunal may – (a) make an appropriate order in relation to a prohibited practice, including – (i) interdicting any prohibited practice; (ii) ordering a party to supply or distribute goods or services to another party on terms reasonably required to end a prohibited practice; (iii) imposing an administrative penalty, in terms of section 59, with or without the addition of any other order in terms of this section; (iv) ordering a divestiture, subject to section 60; (v) declaring conduct of a firm to be a prohibited practice in terms of this Act, for the purposes of section 65; (vi) declaring the whole or any part of an agreement to be void; (vii) ordering access to an essential facility on terms reasonably required;

[214]Section 59, insofar as it is relevant to us provides: (1) The Competition Tribunal may impose an administrative penalty only – (b) for a prohibited practice in terms of section 4(1)(b), 5(2) or 8(a), (b) or (d); (c) for a prohibited practice in terms of section 4(1)(a), 5(1), 8(c) or 9(1), if the conduct is substantially a repeat by the same firm of conduct previously found by the Competition Tribunal to be a prohibited practice; 83 (2) An administrative penalty imposed in terms of subsection (1) may not exceed 10% of the firm’s annual turnover in the Republic and its exports from the Republic during the firm’s preceding financial year. (3) When determining an appropriate penalty, the Competition Tribunal must consider the following factors: (a) the nature, duration, gravity and extent of the contravention; (b) any loss or damage suffered as a result of the contravention; (c) the behaviour of the respondent; (d) the market circumstances in which the contravention took place; (e) the level of profit derived from the contravention; (f) the degree to which the respondent has co-operated with the Competition Commission and the Competition Tribunal; and (g) whether the respondent has previously been found in contravention of this Act.

[215]Section 60 deals with the question of ‘divestiture’ and provides: (2) The Competition Tribunal, in addition to or in lieu of making an order under section 58, may make an order directing any firm, or any other person to sell any shares, interest or assets of the firm if – (a) it has contravened section 8, and (b) the prohibited practice – (i) cannot adequately be remedied in terms of another provision of this Act; or (ii) is substantially a repeat by that firm of conduct previously found by the Tribunal to be a prohibited practice. (3) An order made by the Competition Tribunal in terms of subsection (2) is of no force or effect unless confirmed by the Competition Appeal Court. (4) An order made in terms of subsection (1) or (2) may set a time for compliance, and any other terms that the Competition Tribunal considers appropriate, having regard to the commercial interests of the party concerned. 84 Competence to grant the relief sought

[216]Our reasoning has placed considerable emphasis on those underlying structural circumstances in the market and on the ancillary conduct that enables the first respondent, Mittal SA, to sustain the charging of an excessive price for flat steel products in the South African market. It is appropriate then that we should focus on remedies designed to cure this underlying structure and/or the ancillary conduct.

[217]Harmony has not asked for a structural remedy. It is however clear that a solution of the excessive pricing problem might be to order Mittal SA to divest itself of either the Vanderbijl or Saldanha plants. As already noted, in the period when the Saldanha plant was jointly owned by Iscor and the IDC there was an agreement in place which specifically prohibited Saldanha from marketing its output on the domestic market. Indeed the very location of the plant is clearly designed to favour exports, or, conversely, to dis-favour sales on the domestic market. This agreement nevertheless clearly evidences the threat that Iscor perceived to its dominance – and specifically to its extra-ordinary pricing power – in the domestic market even from a mill located as unfavourably, from the perspective of domestic market sales, as the Saldanha plant.

[218]However, it is not our intention to compel Mittal SA to divest itself of steel producing plant. We should record that our decision to eschew a divestiture remedy of the sort suggested – that is, divestiture of steel producing plant – does not imply acceptance of the argument advanced by counsel for the second respondent, Macsteel International, that would seek to limit our remedial powers to those sought by the complainant. 130 We do not impose an order of divestiture because it is our view that, under present market conditions, there are adequate alternative remedies available. However should these prove ineffective, or capable of circumvention, that is, should it not prove feasible to proscribe Mittal SA’s output-reducing and, hence, excessive pricing conduct in the future, then divestiture may constitute the only appropriate remedy because in the words of the Act cited above “the prohibited practice cannot be adequately remedied in terms of another provision of the Act.”. 131 130 Although we need not decide this point for the purpose of this decision, provided fairness was accorded to a respondent in advance that the Tribunal was considering a remedy not sought by a complainant, there seems to be no reason why the Tribunal is confined only to remedies contemplated in the pleadings. 131 Section 60(2)(b)(i). 85

[219]One of the remedies sought by the complainants is the imposition of an administrative penalty.

[220]Mittal SA contends that an administrative penalty cannot be imposed at the instance of a private complainant, that it is only the public enforcement agency, the Competition Commission, that is competent to ask for the imposition of an administrative penalty. We reject this argument. The appropriateness or otherwise of a remedy is derived from the public and statutory nature of the offence, rather than private nature of the complainant. We are, in other words, mandated to impose an appropriate remedy for conduct which has inured to the disadvantage of the consuming public, of whom the complainants are a small part. In any event, part of the purpose of the administrative penalty is to dis-incentivise future conduct aimed at maintaining an excessive price and we cannot see why the complainants would not have a direct interest in this objective. We note too that the imposition of an administrative penalty is specifically provided for in the Act – indeed the Act specifically provides for an administrative penalty in the case of a first contravention of Section 8(a). The Act also specifically provides processes for the prosecution of complaints by private complainants. Had the legislature intended that, in actions of this sort, the administrative penalty remedy was not to be applied then it would have said so. Instead, the Act allows the private complainant to step into the shoes of the Commission in the event of a non-referral, as in this case, without restricting the complainant to a lesser class of remedies. Indeed where the legislature has intended to place such a restriction on a complainant, it has done so in express terms as in section 49D where following a consent order the complainant is limited to applying for the remedies set forth in section 58(1)(a)(v) or (vi) or an award of civil damages.

[221]We note too – and this surely conclusively disposes of any question surrounding our power to impose a ‘public’ remedy at the instance of a private party - that the Supreme Court of Appeal in American Natural Soda Ash Corporation and Another v Competition Commission of SA and others has recognized that a private complainant can seek an interdict as a remedy in a competition case even when it has not shown injury to itself. It follows then that the principle that private complainants are limited to private remedies has been rejected by the SCA. In a telling passage in that case Cameron and Nugent JJA’s held: “Ansac likewise underscores that an applicant, to obtain interdictory relief under the Act, must place on the scale of risk to it of ‘serious or irreparable 86 damage’, but ignores the fact that obtaining such relief may not be an intervenor’s sole interest in the proceedings.” 132 Quantum of fine

[222]We have decided that it is competent for us to impose a fine in a complaint brought by a private complainant. We have not decided whether to exercise our discretion in favour of a penalty or if we do so, what the level of that penalty should be. Mittal has not addressed us on these issues at all in its Heads of Argument.

[223]Ordinarily, if a party failed to avail itself of the opportunity to lead evidence or to make submissions on a relevant issue, then at the close of the case it would have to face the consequences of this omission. In this case the situation is different. Prior to the close of evidence in this case in April 2006, Mittal advised the Tribunal that it had reached an agreement with Harmony’s legal team that evidence in this respect would only be led if the Tribunal had come to a decision on the merits and determined that a fine was competent. 133 From the record it appears that this matter was left in abeyance at the end of the hearing of the evidence. Much water has since flown under the bridge and it is not clear to the Tribunal whether this matter was further explored between the parties. It was however not brought to our attention in the intervening period between the termination of evidence and the commencement of final argument. Certainly, neither party advised the Tribunal after its decision on the amendment application of a need to resolve this issue prior to final argument.

[224]In the course of final argument it became apparent that there is a dispute between the parties as to whether such an agreement exists. Counsel for Harmony contends that the agreement was only to allow Mittal an opportunity to lead evidence on this aspect if it so wished, and claims that it has not availed itself of that opportunity. 134 Since we were not party to deliberations between counsel, we are not 132 See American Natural Soda Ash Corporation and Another v Competition Commission of South Africa and Others, Supreme Court of Appeal decision of 13 May 2005, Case No. 554/03, at paragraph 34. Our emphasis 133 See transcript pages 2147-2149. 134 See page 2424 of the transcript of 29 November 2006, where Harmony’s Senior Counsel responded to the Tribunal Chairperson’s question as follows: “Adv Unterhalter: Chair, on the last score it is said by Mittal that they are not dealing with the merits of the administrative penalty because they would have that traversed at a separate proceeding in due course. They refer to what they claim to be an agreement struck to that effect. We would just ask you to have regard to the record at page 2148 where this matter was traversed in proceedings before you and we indicated on that score that if there was evidence to be led, it might conveniently to be done when Mr Tomlinson’s evidence was heard and 87 in a position to decide this point. It is clear that Mittal has relied on this alleged understanding not to lead any evidence in mitigation during the course of the evidence. 135 Whether its reliance on this understanding is erroneous or not, Harmony has done nothing to clarify the matter on the record. Mittal will now be afforded the opportunity to lead this evidence. Harmony will be afforded an opportunity to lead evidence in aggravation. Note that whatever evidence led, if any, is strictly confined to this issue.

[225]The tribunal will convene a pre-hearing conference shortly to make arrangements for the further hearing of this evidence. Other remedies

[226]We could, of course, take a decision in respect of the remaining remedies sought, as the record in respect of these matters is complete. However we consider it undesirable to consider remedies in a piecemeal fashion. Accordingly, we will postpone considering the remaining remedies sought, until we have heard the further evidence in connection with the administrative penalty and then we will issue a decision in respect of all the remedies sought. ______________ D. Lewis Concurring: N Manoim and M Holden Tribunal Researcher: T Masithulela For the Complainants : Adv. DN Unterhalter SC (with him Adv. M Wesley) instructed by Cliffe Dekker Inc. For the First Respondent : Adv. CDA Loxton SC (with him Adv. G Pretorius SC, Adv. AG Gotz, and Adv. M Sikhakhane) instructed by Bell Dewar Hall Inc. For the Second Respondent : Adv. JJ Gauntlett SC (with him Adv. A Cockrell) instructed by Edward Nathan Sonnenbergs) certainly we have no agreement that the administrative penalty part of the relief would be heard at some later date. We invited them simply to put up the evidence and have it dealt with…” 135 See Mittal SA heads of argument, paragraphs 44.3-4.