Competition Commission of South Africa v Media 24 (Pty) Limited

CONSTITUTIONAL COURT OF SOUTH AFRICA Case CCT 90/18 In the matter between: COMPETITION COMMISSION OF SOUTH AFRICA Applicant and MEDIA 24 (PTY) LIMITED Respondent Neutral citation: Competition Commission of South Africa v Media 24 (Pty) Limited [2019] ZACC 26 Coram: Mogoeng CJ, Basson AJ, Cameron J, Dlodlo AJ, Froneman J, Goliath AJ, Khampepe J, Mhlantla J, Petse AJ and Theron J Judgments: The Court: [1] to [4] Goliath AJ: [5] to [123] Cameron J, Froneman J and Khampepe J: [124] to [139] Theron J: [140] to [187] Mhlantla J: [188] to [193] Heard on: 22 November 2018 Decided on: 3 July 2019 Summary: Competition Act — section 8(c) — predatory pricing — cost standards Jurisdiction — arguable point of law — matter of general public importance —interests of justice ORDER On appeal from the Competition Appeal Court of South Africa (hearing an appeal from the Competition Tribunal of South Africa). The following order is made: 1. 2. Leave to appeal is granted. The appeal is dismissed with costs, including the costs of two counsel. JUDGMENT THE COURT The first judgment in this matter was written by Goliath AJ and concurred in by Mogoeng CJ and Dlodlo AJ. The first judgment held that this application raises a constitutional issue and an arguable point of law of general public importance within this Court’s jurisdiction. The first judgment further held that leave to appeal should be granted and the appeal must succeed. The second judgment in this matter was written by Cameron J, Froneman J and Khampepe J, and concurred in by Petse AJ. The second judgment held that it is not in the interests of justice to grant leave to appeal and that the application should be dismissed with costs, including the costs of two counsel. Theron J wrote a judgment (third judgment) concurring in the first judgment to the extent that this application raises an arguable point of law of general public importance within this Court’s jurisdiction and that leave to appeal should be granted. Theron J held that the appeal must be dismissed with costs, including the costs of two 2 THE COURT/GOLIATH AJ counsel. Basson AJ concurred in the judgment penned by Theron J. Mhlantla J wrote a judgment in which she concurred in the third judgment to the extent that this application raises an arguable point of law of general public importance within this Court’s jurisdiction and that leave to appeal should be granted. Mhlantla J concurred in the first judgment in respect of the merits and that the appeal must succeed. The effect of these four judgments is that six members of the Court held that this application raises an arguable point of law of general public importance within this Court’s jurisdiction and granted leave to appeal against the judgment and order of the Competition Appeal Court. Six members of the Court did not uphold the appeal. There is thus a majority decision that this application raises an arguable point of law of general public importance within this Court’s jurisdiction, that leave to appeal should be granted and that the appeal must be dismissed with costs, including the costs of two counsel. GOLIATH AJ (Mogoeng CJ and Dlodlo AJ concurring): Introduction “Predatory pricing is a paradoxical offense. Although antitrust law values low prices and abhors high ones, the ‘predator’ stands accused of charging too low of a price – of doing too much of a good thing. Society considers predation socially harmful because the artificially low prices of today drive out competitors and allow the high prices of tomorrow.”1 This matter is the first of its kind dealing with South African law prohibiting predatory pricing. Predatory pricing is conduct that involves a dominant firm setting prices for goods or services at such a low level that (a) the firm incurs losses relative to alternative non-predatory conduct in the short term (referred to as “sacrifice”); (b) the pricing has the effect of eliminating, or being likely to eliminate, one or more of the 1 Crane “The Paradox of Predatory Pricing” (2005) 91 Cornell Law Review 1 at 2-3. 3 GOLIATH AJ firm’s actual or potential competitors; and, in turn, (c) the pricing has the further effect of strengthening or maintaining the firm’s market power, thereby causing consumer harm. The principles underpinning the paradox of predatory pricing must be carefully balanced. An accurate test for predation must be capable of balancing the fine line between over-enforcing the prohibition, with the possibility of higher prices for goods and services, and under-enforcing it to the benefit of large firms aiming to monopolise the market. This case enjoins the Court to grapple with legal elements of the economic concept of predatory pricing. This involves establishing what must be proven when predatory pricing is alleged to have occurred. In making this evaluation, the Court is required to deal with two aspects. The first is whether it is in the interests of justice to grant the applicant leave to appeal to this Court. The second relates to the legal test for establishing predatory pricing. Parties The applicant is the Competition Commission of South Africa, established in terms of section 19 of the Competition Act.2 The respondent is Media24 (Pty) Limited (Media24), a company which acts as the print media arm of the South African media company Naspers Limited. 2 89 of 1998. Section 19 of the Competition Act, concerning the establishment and constitution of the Competition Commission, states: “(1) There is hereby established a body to be known as the Competition Commission, which— (a) (b) (c) has jurisdiction throughout the Republic; is a juristic person; and must exercise its functions in accordance with this Act. (2) The Competition Commission consists of the Commissioner and one or more Deputy Commissioners, appointed by the Minister in terms of this Act.” 4 GOLIATH AJ Background In 1971 and 1983, two community newspapers called Vista and Forum were established respectively. These two papers were fierce competitors in the Welkom community newspaper market. Both papers were acquired by Media24 which, by doing so, acquired a monopoly in the community newspaper market in Welkom. When Media24 took over Vista, some staff chose to leave. One of these staff members was Ms Leda Joubert, who went on to establish Gold Net News (GNN). Ms Joubert appointed veteran Welkom newspaper editor, Mr Hans Steyl to run GNN. While GNN started strongly with loyal advertising support, it ended up struggling to attract advertising and eventually closed down in April 2009. Nine months later, in January 2010, Media24 closed down Forum. Shortly before GNN closed, Mr Steyl filed a complaint on behalf of GNN with the Competition Commission that Media24, through Vista and Forum, had abused its dominant position in the Welkom area by drastically cutting the rates that it charged advertisers for the period between 2004 and 2009. It was alleged that the rates were so low that they were below market price. Mr Steyl asserted that this anti-competitive behaviour forced GNN to close as it could not compete with the low prices and continue to meet the profit margins it needed to survive. In 2011, the Competition Commission, after investigating the complaint, referred a case of predatory pricing on the part of Media24 to the Competition Tribunal (Tribunal). The Tribunal’s decision3 and the subsequent decision on appeal to the Competition Appeal Court4 are the foundation of the case before us. Predatory pricing Because low prices are generally encouraged by competition law, costs standards have been developed to indicate the line between competitive price cutting and unreasonably low prices that are predatory. Predatory pricing is prohibited in two 3 Competition Commission v Media 24 (Pty) Ltd [2016] ZACT 86 (Tribunal decision). 4 Media 24 (Pty) Ltd v Competition Commission 2018 (4) SA 278 (Competition Appeal Court judgment). 5 provisions in the Competition Act. Section 8(d)(iv) creates a specific prohibition against pricing below two specific cost standards. This section provides: GOLIATH AJ “It is prohibited for a dominant firm to— . . . (d) . . . engage in any of the following exclusionary acts, unless the firm concerned can show technological, efficiency or other pro-competitive, gains which outweigh the anti-competitive effect of its act: (iv) selling goods or services below their marginal or average variable cost.” Section 8(c) of the Competition Act contains a general prohibition against anti- competitive behaviour and is considered the “catch-all” provision to protect against abuses of dominance. The section provides: “It is prohibited for a dominant firm to— . . . (c) engage in an exclusionary act, other than an act listed in paragraph (d), if the anti-competitive effect of that act outweighs its technological, efficiency or other pro-competitive, gain.” An “exclusionary act” is defined by the Competition Act as “an act that impedes or prevents a firm from entering into, or expanding within, a market.”5 Section 8(c) should be read to prohibit pricing that is below other appropriate cost standards besides the two listed in section 8(d)(iv). There are five cost standards that are important to understand for the determination of this case. They are the cost standards that are used in various foreign jurisdictions to determine whether prices are predatory, and they are the standards pleaded in this case. These are: marginal cost; average variable cost; average avoidable cost; long run average incremental cost; and 5 Section 1 of the Competition Act. 6 GOLIATH AJ average total cost. The first two are the standards applicable to complaints brought under section 8(d)(iv), while the latter three are of relevance to complaints brought under section 8(c). The various standards are defined as follows: Marginal cost refers to a comparison between the cost of producing an additional unit of output and the price that the firm is charging per unit of output. If the price exceeds marginal cost, then it makes sense to produce an additional unit. If it costs more to produce a unit of output than the firm can charge, then the rational firm would not produce that unit. Consequently, if a firm prices below marginal cost it is presumed to be participating in predatory pricing. Establishing marginal cost is extremely difficult; therefore, a second and similar test was created.6 Average variable cost is this second test. It is referred to as the Areeda-Turner test in various local judgments and foreign commentaries.7 This is the sum of the firm’s variable costs (labour, electricity, et cetera) divided by the number of units produced.8 Variable costs are costs that vary as output increases, as opposed to fixed costs, which do not. If a firm charges a price which does not cover its variable costs, it intentionally suffers a loss on every sale and is presumed to be participating in predatory pricing.9 Average avoidable cost refers to the average costs that the firm could have avoided if it had not produced particular units of output. This includes variable costs and any fixed costs that are specifically incurred in the production of the particular product being examined. Average avoidable cost is generally equal to average variable cost plus product-specific fixed costs.10 6 O’Donoghue and Padilla The Law and Economics of Article 82 EC (Hart Publishing, Oxford 2006) at 237. 7 Id at 240. 8 Id at 237 and 245. 9 Id at 240. 10 Id at 241. 7 GOLIATH AJ Average total cost is equal to the total cost of producing a product divided by the number of units produced. This number includes the costs associated with average variable cost, the product-specific fixed costs, and a proportion of the common costs of the company that have been apportioned to the particular product if the firm produces multiple products. Common costs are costs that the firm incurs regardless of how many different products are produced. Long-run average incremental cost refers to the total value of costs that are needed to enter and start supplying a specific product, represented as an average over output. This test is generally used in industries which have high barriers to entry but low operating costs, as it allows companies to average out the high sunk costs11 that they must recover with the low operation costs.12 Litigation history Competition Tribunal The case referred to the Tribunal was based firstly on a contravention of section 8(d)(iv), and in the alternative on a violation of section 8(c). The Tribunal found that Media24 had not priced its advertising below its average avoidable cost, and thus the Competition Commission failed to establish that Media24 priced below the lower standards of average variable cost or marginal cost.13 As a result, the Tribunal made its determination using section 8(c) and not section 8(d)(iv). The case presented by the Competition Commission was that Media24 had used Forum as a fighting brand. This means that it was kept in the market with the express purpose of charging prices that were lower than its competitors, who were then forced to leave the market. Once this task was completed, the fighting brand was closed down. The Tribunal held that the Competition Commission had established that the average total cost was an appropriate 11 “Sunk costs” refer to money that is spent by businesses upfront and cannot be recovered or refunded. For example, once money is spent on rent or expensive equipment which value will deflate over time that amount is not recoverable and is thus sunk. See Baumol “Fixed Costs, Sunk Costs, Entry Barriers, and Sustainability of Monopoly” (1981) 96 Quarterly Journal of Economics 405 at 406. 12 O’Donoghue and Padilla above n 6 at 242-3. 13 Tribunal decision above n 3 at para 211. 8 GOLIATH AJ cost standard to use to evaluate predation in this case;14 that Media24 had charged advertising prices for Forum below its average total cost; that it had intended to predate GNN; that Media24 had the ability to recoup what it had lost during this predation period and that GNN had not been excluded due to its relative inefficiency. It found that Media24’s actions had an anti-competitive effect and that there was no evidence of pro- competitive gain which outweighed this effect.15 Consequently, it held that Media24 had contravened section 8(c) of the Competition Act. Additional evidence The Tribunal held that average total cost could be an appropriate costs standard for a finding under section 8(c) of the Competition Act when accompanied by additional evidence of predation.16 The Tribunal evaluated the additional evidence under the following four headings: direct intention to predate, indirect intention to predate, recoupment, and the equally efficient competitor test. Direct intention Direct intention relates to the subjective intention of Media24 as evidenced by witness statements. The Tribunal examined the factual evidence that was placed before it which consisted of oral and written testimony by persons who had been employed by Media24 during the complaint period. The evidence pointed to the fact that internal communication had taken place in terms of which employees had stated that they were attempting to undercut GNN, and that they intended to use Forum as a stopper in the market to keep business away from GNN. A similar sentiment was expressed at a Media24 strategy meeting. The Tribunal found that this constituted more than merely fighting words and indicated a direct intention to predate GNN.17 Moreover, it held that 14 Id at para 221: the reason for this finding was based first on an understanding that South Africa’s economy is characterised by high barriers to entry and second, that an information asymmetry exists between large, established firms and small and mediums ones. Information asymmetry refers to the way that large firms can manipulate their financial records to exacerbate the imbalance between them and their smaller rivals. 15 Id at para 621. 16 Id at para 222. 17 Id at para 383. 9 GOLIATH AJ Media24 had not done enough to rebut the evidence that had been put forward. It thus accepted that Media24 had acted with predatory intent and moved on to evaluate this finding in the light of other evidence of exclusion.18 Indirect intention Indirect intention is inferred from objective evidence and involves drawing economically pertinent inferences from certain behaviour.19 The Tribunal examined three areas of evidence. These were: firstly, evidence that Forum was consistently making a loss for nine years before it eventually closed down. Secondly, the Tribunal examined the circumstances around the closing down of Forum such as the fact that Media24 closed Forum nine months after GNN left the market. It concluded that Media24 had intended to close Forum as soon as GNN had left the market but retained it to make them look less suspicious and ensure that they were not brought before the Commission. Finally, the Tribunal examined the cannibalism factor. This relates to the fact that during the period that Forum was maintained, it prevented some advertisers from advertising in Vista. Generally, allowing a product to draw customers away from another in the same firm would not be rational. The Tribunal held that this points to a separate strategic reason for maintaining Forum. The strategic value was that Forum was able to predate GNN. Recoupment The Tribunal accepted that since GNN’s exit from the market, Vista had increased its prices by 17%. This, it held, was substantial when compared to the price increases of previous years. Moreover, because no other competitors had entered the market after GNN and Forum exited, Vista was able to continue charging high prices to the detriment of consumers in an unchecked fashion. These factors taken together amount to strong evidence of recoupment. 18 Id at paras 619 and 621. 19 O’Donoghue and Padilla above n 6 at 252. 10 GOLIATH AJ Equally efficient competitor test Media24 contended that the reason for GNN’s exit from the market was because it was not an equally efficient competitor20 and so could not survive in a true competitive market. The Commission contested this. The Tribunal held that it was unlikely that GNN left the market simply because it was inefficient. It based this finding on the fact that it was run by a veteran of the Welkom newspaper market who knew how to run a profitable business. Moreover, the fact that GNN was able to break into the market and gain market share pointed to the fact that it was efficiently run.21 All of this evidence, taken together with the fact that Media24 priced below average total cost, led the Tribunal to hold that Media24 was guilty of predatory pricing in terms of section 8(c) of the Competition Act.22 Competition Appeal Court The Competition Appeal Court summarised the Commission’s case as: “(a) Appellant operated Forum solely as a ‘fighting brand’ in order to exclude GNN from the market; (b) Forum’s average revenues did not cover the average total costs of producing and publishing the paper; (c) Forum’s average revenues did not cover its average variable costs over a 12 month period (where all costs were considered variable and thus could have been avoided if Forum had ceased publication at this point); and/or (d) Forum’s incremental costs (calculated as Forum's total revenues reduced by the proportion of its revenues which would have been diverted to Vista if Forum had exited the market) did not cover its incremental costs, calculated as the total costs that were incremental to the operation of Forum based on two definitions: (i) all costs that could have been eliminated over a period of one year; and 20 Elzinga “The goals of antitrust: other than competition and efficiency, what else counts?” (1977) Vol 125:6 University of Pennsylvania Law Review 1191 at 1192. 21 Tribunal decision above n 3 at para 557. 22 Id at para 621. 11 (ii) all Forum’s costs.”23 GOLIATH AJ The Competition Appeal Court held that the test envisaged in section 8(c) determines whether specific conduct amounts to an exclusionary act as defined in the Competition Act. This, it held, is an objective test. The court held that subjective evidence of intent should not be examined in proving predatory pricing and that, once this evidence was disregarded, average total cost was not an appropriate cost standard to illustrate that predatory pricing occurred. The Competition Appeal Court concluded that: “[T]here is no escaping the conclusion that predation must focus on the likely economic effect of pricing below a particular cost measure to determine whether the low prices are due to a lawful competitive response to rivals or to predation and unlawful behaviour rather than on the intention with which a pricing strategy is adopted.”24 Having rejected average total cost plus intention, the Competition Appeal Court concluded that the only appropriate benchmark that had been relied upon by the Commission in their pleadings was average avoidable cost. This does not mean that it is the only appropriate benchmark to apply to section 8(c) in all cases, but rather that in this particular case it is the only appropriate test that remains. In the light of its rejection of the average total cost plus intent test, the Competition Appeal Court did not have to consider the balance of the evidence concerning the intention of Media24. In the result, the appeal by Media24 was upheld as it could not be established that Media24 had violated section 8(d)(iv) or 8(c). Leave to appeal Both the Competition Commission and Media24 contend that this case raises a constitutional issue. This is because the Competition Act was enacted as transformative legislation. Section 2(e) and (f) of the Competition Act states that part of the purpose 23 Competition Appeal Court judgment above n 4 at para 17. 24 Id at para 56. 12 GOLIATH AJ of the Competition Act is to ensure that small and medium sized businesses have equitable opportunities to participate in the economy and to promote a greater spread of ownership in the economy by those who were disadvantaged by Apartheid. These purposes implicate the right of equality contained in section 9 of the Constitution. Section 9(2) enjoins the State to take legislative and other measures to advance the equality of previously disadvantaged people and section 2(e) and (f) of the Competition Act is a legislative measure of this kind. Section 8 of the Competition Act is an important component for the achievement of the purposes contained in section 2(e) and (f). Section 8 prohibits dominant firms from abusing their power to the detriment of the market and consumers. Consumer welfare lies at the heart of the matter before us. The prohibition of abuses of dominance is recognition of the fact that dominant firms attained this status due to our exclusionary history. The protection section 8 provides applies specifically to small and medium sized firms which are at the greatest risk of being excluded from the market by abuses of dominance or monopolies. Consequently, when a Court interprets the extent of the protection contained in section 8, it must do so in line with the purpose of the Competition Act and the Constitution as abuses of dominance and the resultant monopolies have the potential to perpetuate historic patterns of exclusion and inequality. Beyond this, this Court held in S.O.S and Hosken, that the ambit of the Commission’s investigatory powers, in terms of section 21(1)(c) of the Competition Act, is an inherently constitutional issue.25 This section empowers the Commission to investigate alleged violations of Chapter 2 of the Competition Act. The case before us involves an interpretation of the extent of the powers held by the Commission to prosecute alleged violations subsequent to an investigation. The effect of the 25 S.O.S Support Public Broadcasting Coalition v South African Broadcasting Corporation (SOC) Limited [2018] ZACC 37; 2019 (1) SA 370 (CC); 2018 (12) BCLR 1553 (CC) (S.O.S) at para 21; and Competition Commission of South Africa v Hosken Consolidated Investments Limited [2019] ZACC 2; 2019 (3) SA 1 (CC) 2019 (4) BCLR 470 (CC) (Hosken) at para 31. 13 GOLIATH AJ Competition Appeal Court judgment is to limit the tools that the Commission has at its disposal to prove a violation of section 8(c). This is because on the test it applied, it was unnecessary to examine certain evidence regarding predation. While evaluations under section 8(c) are economic in nature, a limitation of the investigative and prosecutorial powers of the Commission may have adverse consequences. Surely, if the ambit of the Commission’s investigatory powers raises a constitutional issue, the ambit of its ability to bring prosecutions based on findings made during investigation is, by parity of reasoning, similarly constitutional. What would be the purpose of the strong investigatory powers of the Commission if it cannot present all of the evidence that it has gathered in order to prosecute an offence? Aside from raising a constitutional issue, this matter engages the Court’s jurisdiction because it raises an arguable point of law of general public importance which this Court ought to consider. The question whether pricing above average avoidable cost but below average total cost amounts to predation requires the interpretation of the law relating to exclusionary acts and the consideration of which tests can be used to equitably establish predation. This is unquestionably an arguable point of law, especially when the contradicting views in foreign jurisprudence, and between the Tribunal and Competition Appeal Court are considered. Moreover, given the impact that this determination will have on the interpretation and implementation of section 8(c) of the Competition Act going forward, it is certainly of general public importance and this Court ought to consider it. One problem exists in this regard, however: the Competition Act in section 62 states that for all matters, aside from those that raise constitutional issues, the Competition Appeal Court has exclusive and final appeal jurisdiction. Can this Court then hear competition cases which raise arguable points of law of general public importance? The answer must be yes. Section 62 was enacted prior to the seventeenth amendment of the Constitution, which imbued the Court with the power 14 GOLIATH AJ to hear arguable points of law of general public importance, and as such could not have predicted that the Court would one day have the jurisdiction to decide such cases. Section 62 of the Competition Act must be read in the context of section 167 of the Constitution. Section 167(3)(b)(ii) establishes that this Court can hear arguable points of law of general public importance and section 167(3)(c) states that this Court makes the final decision as to whether a matter falls within its jurisdiction.26 If the Competition Act was read to finally exclude arguable points of law of general public importance from the jurisdiction of this Court, it would then be in conflict with the Constitution. Legislation should not be read to derogate from the constitutionally enshrined powers of this Court, instead a reading which aligns with the Constitution should be favoured. A contextual reading of the Competition Act therefore leads to the conclusion that the Competition Appeal Court has exclusive jurisdiction to hear appeals on all matters except those which fall into the constitutionally protected jurisdiction of the Constitutional Court. Resultantly, this Court has jurisdiction to hear this matter as it raises a constitutional issue, and in any event raises an arguable point of law of general public importance which this Court ought to consider. Failure to approach the Competition Appeal Court Appellate jurisdiction in competition law matters is regulated by section 62 of the Competition Act. Section 62(1) stipulates which matters fall within the exclusive jurisdiction of the Tribunal and the Competition Appeal Court.27 Section 62(3) of the Competition Act states that the Competition Appeal Court is the final court of appeal 26 Section 167 of the Constitution states: “(3) The Constitutional Court— … (b) may decide— … (ii) any other matter, if the Constitutional Court grants leave to appeal on the grounds that the matter raises an arguable point of law of general public importance which ought to be considered by that Court.” 27 The matters that fall within the exclusive jurisdiction include the interpretation of chapter 2 of the Competition Act. Section 8(c) is located in chapter 2. 15 GOLIATH AJ on all matters regulated by section 62(1). However, the appellate jurisdiction of the Competition Appeal Court is not final in competition matters which raise constitutional issues.28 Resultantly, and crucial to this case, the appellate jurisdiction of the Competition Appeal Court is only final when the interpretation of chapter 2 does not raise a constitutional issue. The Competition Act states that where a constitutional issue is raised, both the Supreme Court of Appeal and this Court have the jurisdiction to hear it.29 An application for leave to appeal from the Competition Appeal Court to either of these courts is regulated by section 63(2) of the Competition Act. However, the right to appeal to the Supreme Court of Appeal and this Court is made conditional on a party first applying to the Competition Appeal Court for leave to appeal. Section 63(2) reads: “An appeal in terms of section 62(4) may be brought to the Supreme Court of Appeal or, if it concerns a constitutional matter, to the Constitutional Court, only (a) with leave of the Competition Appeal Court; or (b) if the Competition Appeal Court refuses leave, with leave of the Supreme Court of Appeal or the Constitutional Court, as the case may be.” (Emphasis added.) The interpretation of section 63(2) has been the subject of extensive debate. The question is whether the use of the word “only” bars an applicant from applying to either the Supreme Court of Appeal or this Court without first approaching the Competition Appeal Court for leave, or whether the failure to approach the Competition Appeal Court is merely an obstacle that can be overcome in certain circumstances. This dichotomy was set out in the majority decisions in both 28 Section 62(3)(b) of the Competition Act. 29 Id at section 62(4) which states: “An appeal from a decision of the Competition Appeal Court in respect of a matter within its jurisdiction in terms of subsection (2) lies to the Supreme Court of Appeal or Constitutional Court, subject to section 63 and their respective rules.” 16 GOLIATH AJ Loungefoam and Yara.30 However, this Court in those cases did not make a final pronouncement on whether a bar or an obstacle is created.31 It has become necessary to decide this issue. In doing so, the judgments in Loungefoam and Yara, which I fully endorse, are instructive. These judgments both hold that the failure to seek leave to appeal from the Competition Appeal Court before appealing to the Supreme Court of Appeal and this Court creates an obstacle to the application which can be overcome if it is in the interests of justice for this Court to grant a direct appeal.32 The restriction contained in section 63(2) of the Competition Act, and the right to appeal in section 62(4) are both qualified by section 63(1)(a).33 Section 167(6) of the Constitution establishes that: “National legislation or the rules of the Constitutional Court must allow a person, when it is in the interests of justice and with leave of the Constitutional Court— (a) to bring a matter directly to the Constitutional Court; or (b) to appeal directly to the Constitutional Court from any other court.” With regard to this, Cameron J in Yara noted the following: “[T]he right of appeal is expressly subject to ‘any law’ that ‘specifically grants’ a right of appeal. Plainly, section 167(6) of the Constitution is a law of this kind. It provides 30 Competition Commission v Loungefoam (Pty) Ltd [2012] ZACC 15; 2012 JDR 1119 (CC); 2012 (9) BCLR 907 (CC) (Loungefoam) at paras 22-3 and Competition Commission v Yara South Africa (Pty) Ltd [2012] ZACC 14; 2012 JDR 1118 (CC); 2012 (9) BCLR 923 (CC) (Yara) at para 20. 31 Loungefoam id at para 24 and Yara id at para 21. 32 Loungefoam id at para 24 and Yara id at para 68. 33 Section 63(1)(a) states: “The right to an appeal in terms of section 62(4)— (a) is subject to any law that— (i) (ii) specifically limits the right of appeal set out in that section; or specifically grants, limits or excludes any right of appeal.” 17 that national legislation or the rules of this Court ‘must allow’ a litigant, ‘when it is in the interests of justice and with leave of the Constitutional Court’, to bring an appeal directly to this Court.”34 (Footnotes omitted.) GOLIATH AJ The provisions of section 63 of the Competition Act must be interpreted so that they comply with the Constitution35. It appears that the section itself, by providing that the right of appeal is subject to any law that specifically grants a right of appeal, subordinates its requirements to the right of direct appeal to this Court contained in section 167(6) of the Constitution.36 In light of the provisions of section 167(6) of the Constitution, section 63(2) read with section 63(1)(a)(ii) cannot be said to create an absolute bar. As the minority in Yara held: “[T]he fact that a litigant has not sought the leave of the Competition Appeal Court would be pertinent to its application to this Court, but will not disable it. This Court is at liberty to grant direct access, as the Constitution requires, and nothing in the legislation need be read to detract from its power.”37 The word “only” should be interpreted to mean that litigants are barred from approaching this Court directly only when the interests of justice do not warrant the granting of a direct appeal.38 In such cases, leave of the Competition Appeal Court must first be obtained. Furthermore, on whether the interests of justice warrant the granting of leave to appeal directly to this Court appears later in this judgment. Failure to approach the Supreme Court of Appeal It must be established whether the Supreme Court of Appeal can still be approached to hear appeals relating to competition law. If it retains its appellate jurisdiction in such matters, this creates difficulties for the Commission as it failed to 34 Yara above n 30 at para 65. 35 Id at para 66. 36 Id. 37 Id at para 70. 38 Id at para 68. 18 approach the Supreme Court of Appeal before appealing to this Court. The seventeenth amendment to the Constitution amended section 168(3)(a) of the Constitution to read: GOLIATH AJ “The Supreme Court of Appeal may decide appeals in any matter arising from the High Court of South Africa or a court of a status similar to the High Court of South Africa, except in respect of labour or competition matters to such an extent as may be determined by an Act of Parliament.”39 (Own emphasis.) The question is how to interpret the express wording of the Constitution. The Commission contends that it should be interpreted to mean that the Supreme Court of Appeal’s appellate jurisdiction has been excluded by this amendment, while Media24 contends that until an Act of Parliament is enacted limiting the Supreme Court’s jurisdiction in these matters, it should continue to hear competition matters. Brand JA in Computicket40 embraces the conclusion by Froneman J in National Lotteries Board which states: “As a result of the Constitution Seventeenth Amendment Act of 2012, this right of appeal to the Supreme Court of Appeal no longer exists.”41 In Democratic Alliance, Zondo J, stated that the phrase “to such extent as may be determined by an Act of Parliament” in section 168(3), qualifies the words “except in respect of labour or competition matters” and not “the High Court of South Africa or a court of a status similar to that of the High Court”. Resultantly, Zondo J held that appeals concerning all matters except competition and labour matters still lie with the Supreme Court of Appeal.42 By implication, it appears that the seventeenth amendment 39 Section 168(3)(a) of the Constitution. 40 Competition Commission v Computicket (Pty) Ltd [2014] ZASCA; 185 JDR 2507 (SCA) at para 10. 41 National Union of Public Service & Allied Workers obo Mani v National Lotteries Board [2014] ZACC 10; 2014 (3) SA 544 (CC); 2014 (6) BCLR 663 (CC) (National Lotteries Board) at fn 26. 42 Democratic Alliance v African National Congress [2015] ZACC 1; 2015 (2) SA 232 (CC); 2015 (3) BCLR 298 (CC) at paras 26-8. 19 GOLIATH AJ did indeed remove the appellate jurisdiction of the Supreme Court of Appeal in respect of labour and competition matters. It may be that further Acts of Parliament will qualify this exclusion. However, the amendment itself constitutes an Act of Parliament and its express wording must be respected. Consequently, it was not necessary for the Competition Commission to approach the Supreme Court of Appeal before appealing to this Court. Even if this was not the case, there are instances where it is unnecessary for litigants to approach the Supreme Court of Appeal before approaching this Court. As Cameron J and Yacoob J held in Yara: “The Act envisages appeals directly from the Competition Appeal Court to this Court. In matters involving the constitutionality of the Commission’s interpretation and exercise of its statutory powers, this Court is necessarily the final Court. The questions at issue here are so important that it is nearly inevitable that, whoever succeeds in the Supreme Court of Appeal, this Court will be asked to have the last word. The issues do not involve matters of common law, on which this Court particularly values the views and experience of the Supreme Court of Appeal. It is true that even where common law matters are not at issue, this Court values the views of the Supreme Court of Appeal. Nevertheless, the largely statutory and constitutional nature of the questions at issue counts against requiring an appeal to the Supreme Court of Appeal first.”43 (Footnotes omitted.) The matter before us is an important one involving largely statutory interpretation and constitutional questions. This weighs against it being necessary for the Competition Commission to first approach the Supreme Court of Appeal. Besides, a proper reading of that excerpt from Yara relied on by the second judgment would reveal that Yara is no authority for the proposition that this Court ought to defer to the Supreme Court of Appeal in relation to the common law or to the Competition Appeal Court as a specialist court. Yara does no more than recognise that 43 Yara above n 30 at para 62. 20 GOLIATH AJ obvious benefit that this Court generally stands to derive from allowing courts with general jurisdiction and specialist courts to first express themselves on issues that fall within their jurisdiction before we deal with a matter. The apex court of the Republic of South Africa ought never to abdicate its constitutional responsibility of providing guidance to all courts in this country regardless of how complex or specialised the area of law under consideration might be. To do so could be misunderstood as a confession by this Court, of its incompetence and would probably undermine the jurisdiction and authority of this Court in relation to matters of a specialist nature. As a matter of fact, many complex issues that require speciality in certain areas of the law have served before this Court, but we have never deemed it appropriate to defer to any of the lower courts, for good reason. Interests of justice The fact that this Court has jurisdiction to hear the matter is, however, not decisive and leave may still be refused if it is not in the interests of justice to hear the appeal. There are a number of elements to this case that indicate that it would be in the interests of justice for this Court to hear a direct appeal to it. First, the Commission pointed out that this matter involves statutory interpretation which ought to be developed in the legal framework of the Constitution and not parallel to it. This Court regularly undertakes interpretative exercises such as this in relation to the Labour Relations Act.44 The analogy between the two areas of law, Media24 submits, is misplaced because this case seems to raise more specialised economic issues which require expert evaluation. This may be true, but this Court has the benefit of the expert opinions of both the Tribunal and the Competition Appeal Court in making its decision. Were this matter to confront the Court with a factual question, it might then be in the interests of justice to defer to the 44 See South African Commercial, Catering and Allied Workers Union v Woolworths (Pty) Ltd [2018] ZACC 44; (2019) 40 ILJ 87 (CC) (Woolworths) at para 20; National Union of Metalworkers of SA v Intervalve (Pty) Ltd [2014] ZACC 35; (2015) 36 ILJ 363 (CC); 2015 (2) BCLR 182 (CC) at para 25 and National Education Health and Allied Workers Union v University of Cape Town [2002] ZACC 27; 2003 (3) SA 1 (CC); 2003 (2) BCLR 154 (CC). 21 GOLIATH AJ specialist courts but as the third judgment points out that, “this question entails critically examining the policy and normative implications of the various standards for predatory pricing”45, this Court is in a position to reach the best determinable conclusion in line with the interests of the public. Davis JP, in the introductory paragraph of the Competition Appeal Court judgment, places great emphasis on the fact that this matter is the first of its kind.46 As a result, in interpreting the Competition Act’s prohibition of predatory pricing, this Court must establish how the prohibition aligns with, and fulfils, the imperatives of the Constitution. Competition matters impact on the interests of the public, especially considering South Africa’s evolving and transforming market economy. The need to provide the country with free and fair guidelines for an equitable competitive market is crucial, and something that this Court is qualified to do. Second, the implications of this case will be far-reaching. Media24 contested this by arguing that the Competition Act is being amended and amendments minimise the importance of this Court’s decision. While it is true that the Competition Act is being amended and that the new sections may limit the implications of this judgment, section 8(c) will still remain in the amended version of the Competition Act. The question of what kind of evidence can be used to prove a violation of that section will continue to affect parties beyond those before us. The potential amendment of the Competition Act does not negate or counteract the importance of the issues raised in this matter. Third, the Competition Commission has reasonable prospects of success on appeal. Both the Tribunal and the Competition Appeal Court are specialist bodies, and in this case they have arrived at opposite conclusions. Ngcobo J in NEHAWU states that— 45 Third judgment at [144]. 46 Competition Appeal Court judgment above n 4 at para 1. 22 GOLIATH AJ “[A]n important factor in considering the prospects of success in this application is the fact that members of the Labour Appeal Court and the Labour Court are divided on the proper construction of section 197. This factor alone suggests, at least prima facie, that there are prospects of success. It is true that the Labour Appeal Court, like all courts, is bound by the doctrine of precedent, and should not depart from its own decisions unless it is satisfied that they are clearly wrong. Nevertheless, given the clear division amongst the labour Judges, it is desirable for this Court to consider the issue.”47 (Footnotes omitted.) The division between the specialist Judges at these bodies, indicates that interference with the decision of the Competition Appeal Court may be warranted. In these circumstances, as is the case with labour law matters, this Court will not shy away from intervening and interpreting the law in a way that best caters for the interests of justice and the public at large. Consequently, it is in the interests of justice that leave to appeal directly to this Court from the Competition Appeal Court be granted. What must be decided now is whether the appeal should be upheld. The appeal Aims and purpose of the Competition Act It is a fundamental principle of competition law that competition between firms is desirable. Competition generally encourages efficiency, innovation and the charging of lower prices by firms.48 Competition on the merits49 includes price competition which involves firms lowering their prices to attract business away from their competitors. However, competition law also acknowledges that exceptionally low 47 National Education Health and Allied Workers Union v University of Cape Town [2002] ZACC 27; 2003 (3) SA 1 (CC); 2003 (2) BCLR 154 (CC) (NEHAWU) at para 26. 48 Bishop and Walker The Economics of EC Competition Law: Concepts, Application and Measurement 1 (Sweet & Maxwell Ltd, London 2002) at 11. “What 49 OECD at http://www.oecd.org/competition/mergers/37082099.pdf. “Competition on the merits” refers to competition between a dominant firm and its competitors which has the effect of increasing consumer welfare. the merits?” competition available 2006), (June on at is 1 23 GOLIATH AJ prices may be harmful to competition and thus, in the long run, consumers. This is because below-cost pricing has the effect of forcing competitors out of the market which increases the likelihood of the formation of monopolies50. Monopolies are harmful to consumer welfare and should therefore be regulated.51 The South African Competition Act acknowledges that the unjust distribution of wealth during apartheid has led to concentrations of power and capital in the market. One way of combatting this concentration is through the regulation of firms which hold dominant shares in the market. The Competition Act specifically regulates the behaviour of these firms to ensure that they are not abusing their dominant positions to the detriment of competition and consumers.52 Section 8 of the Competition Act prohibits two forms of abusive behaviour by dominant firms: exploitative abuse and exclusionary abuse. Exploitative abuses result in direct harm to consumers.53 Exclusionary abuse, on the other hand, is conduct that attacks the dominant firm’s competitors and as a result, indirectly harms consumers.54 The prohibition of predatory pricing contained in section 8(c) and (d)(iv) of the Competition Act is an example of the regulation of exclusionary abuse. This regulation is not an easy task. Approaches to evaluating whether predation has occurred are often criticised as being over-inclusive, in that they implicate innocent firms in predation, or under-inclusive, in that they fail to identify firms which are genuinely participating in predation, to the detriment of competition.55 Any approach taken to prevent predatory pricing must strike a balance between over and under enforcement, while 50 Bishop and Walker above n 48 at 51 at para 6.69. 51 Id at 21 at para 2.21. 52 Section 8 of the Competition Act. 53 Section 8(a) of the Competition Act states that “it is prohibited for a dominant firm to charge an excessive price to the detriment of consumers”. 54 Section 8(c) and (d) of the Competition Act prohibit examples of this kind of behaviour. 55 Mackenzie “Are South Africa’s predatory pricing rules suitable?” (September 2014) at 3, 5 and 6, available at http://www.compcom.co.za/wp-content/uploads/2014/09/Neil-Mackenzie-Predatory-Pricing-in-SA.pdf. 24 GOLIATH AJ simultaneously fulfilling the purpose of the Competition Act. This case therefore requires a careful balancing act. Section 2 of the Competition Act sets out the purpose for which it was enacted. It states: “The purpose of this Act is to promote and maintain competition in the Republic in order— (a) (b) (c) to promote the efficiency, adaptability and development of the economy; to provide consumers with competitive prices and product choices; to promote employment and advance the social and economic welfare of South Africans; (d) to expand opportunities for South African participation in world markets and recognise the role of foreign competition in the Republic; (e) to ensure that small and medium-sized enterprises have an equitable opportunity to participate in the economy; and (f) to promote a greater spread of ownership, in particular to increase the ownership stakes of historically disadvantaged persons.” This provision seemingly envisages the desired balance mentioned above. While competitive prices and product choices are important, so is the protection of small and medium-sized businesses. Therefore, any approach to prohibiting predatory pricing must avoid over-inclusion as this will harm the ability of firms to set competitive prices, while simultaneously not being under-inclusive as this makes it easy for dominant firms to force small and medium-sized competitors to exit the market. Examining predatory pricing in the Competition Act Predatory pricing is prohibited by sections 8(c) and 8(d)(iv) of the Competition Act. Section 8(d)(iv) contains a specific prohibition against predatory pricing. It presumes that prices which are below a firm’s marginal cost or average variable cost will have exclusionary effects and are not motivated by a rational business rationale beyond the gains that a dominant firm makes by excluding competitors. Section 8(c), on the other hand, is a catch-all provision which prohibits unjustified exclusionary acts 25 GOLIATH AJ not listed in section 8(d)(iv). This appeal requires this Court to determine how best to interpret section 8(c) to ensure that predatory pricing that falls outside of section 8(d)(iv) is prohibited. The approach to evaluating claims in terms of section 8(c) was established by this Court in Senwes. Here the Court held: “Plainly the section requires the presence of three conditions in order to establish that an abuse of dominance has occurred. First, the act in which the dominant firm was engaged must be an ‘exclusionary act’ as defined in the empowering legislation. . . . Second, the act in which the dominant firm was engaged must fall outside the scope of section 8(d). Third, the anti-competitive effect of that act must outweigh its technological, efficiency or other pro-competitive gain.”56 However, when examining section 8(c), both parties before us argued that the section creates a two-stage test. This is likely because it is common cause that the exclusionary act falls outside of section 8(d). The Commission contends that first, it must be determined whether an exclusionary act had occurred, and second, whether the anti-competitive effect of that act outweighs some form of gain listed in the section. Media24 characterised the test for predation as pricing which is low enough to exclude an equally efficient competitor, which is done without a rational business justification. There is not much difference between the tests that the parties each propose. Moreover, while there are two stages to each test, they appear to be interrelated. This is because an act that leads to the exclusion of a competitor but does not have an anti-competitive effect, or has a rational business justification, will not be predatory and conversely an act that has anti-competitive effects but does not result in the exclusion of a competitor is not prohibited by section 8(c). 56 Competition Commission of South Africa v Senwes Limited [2012] ZACC 6; 2012 JDR 0579 (CC); 2012 (7) BCLR 667 (CC) (Senwes) at paras 27-8. 26 GOLIATH AJ Both stages of the test are effect-based and objective.57 Whether an act excludes competitors is a question of factual causation. Once it has been determined that an act excluded a competitor, the extent of the anti-competitive effect of this exclusion is established by examining whether there has been actual harm to consumers or a foreclosing effect on the market.58 Actual harm to consumers occurs when, upon a competing firm’s exit from the market, the dominant firm raises the price that it charges for goods and services in order to recoup the losses that it made while charging below- cost prices. This is known as recoupment. It is often difficult to prove that recoupment has taken place.59 Consequently, anti-competitive effect can also be proven by showing that there has been a substantial effect on the structure of the market and that this is likely to indirectly harm consumers. The exit of a firm from the market, where this exit results in the elimination or reduction of competition, is the kind of structural change that is envisaged. Proving predatory pricing There is a debate about whether predatory pricing even exists. Some commentators suggest that pricing below-cost, even in the short term, is irrational as it runs contrary to the principle of profit maximisation.60 However, it is now settled that predation is a tactic used by some dominant firms to cement their ability to maximise profits in the future.61 This is because by making short term losses which drive a competitor from the market, the dominant firm is able to set the higher price at which they can recoup their losses and continue to charge in the medium to long term.62 The principle of short term profit maximisation still underpins market strategy. 57 See definition of “exclusionary act” in [14]. 58 European Commission Communication from the Commission: Guidance on the Commission’s Enforcement Priorities in Applying Article 82 of the EC Treaty to Abusive Exclusionary Conduct by Dominant Undertakings (2009/C 45/02, February 2009) at para 71. 59 O’Donoghue and Padilla The Law and Economics of Article 102 TFEU 2 ed (Hart Publishing, Oxford 2013) at 314. 60 Bolton, Brodley and Riordan “Predatory Pricing: Strategic Theory and Legal Policy” (2000) 88 Geo Law Journal 2239 at 2244. 61 O’Donoghue and Padilla above n 6 at 237. 62 Id at 236. 27 GOLIATH AJ Consequently, a determination that a firm is purposefully loss-making, and thus sacrificing some of the money it could have been earning, may be indicative that greater scrutiny of its pricing model is required.63 The cost standards that are central to the dispute before this Court become relevant here. Costs standards identify the various expenses in creating a product and running a firm. Pricing below a particular cost standard indicates that a firm is not recovering all of the costs incurred in producing a product.64 This means these standards are useful to establish whether predation has taken place because they point to profit sacrifice on the part of the dominant firm.65 However, all costs measures have short-comings. Marginal cost is criticised as being too difficult to calculate.66 Average variable cost, which was developed to supplement these difficulties, is difficult to determine when the line between fixed and variable costs is blurred.67 Average avoidable cost, which is generally considered to be the most readily ascertainable and calculable cost standard, still encounters problems where predation has occurred over a long period of time and it is difficult to determine which costs are avoidable and which are not.68 Long run average incremental cost is accused of being biased against firms that produce only one product, as the calculation will not include common or joint costs.69 With regard to average total cost, there are often rational business justifications for pricing below this standard and determining the costs included in it is practically difficult due to the element of discretion present in the allocation of common costs to products in multi-product firms.70 63 Id at 237. 64 Bishop and Walker above n 48 at 17-9. 65 Id at 12. 66 O’Donoghue and Padilla above n 6 at 240. 67 Tribunal decision above n 3 at para 82. 68 O’Donoghue and Padilla above n 6 at 243. 69 Id. 70 Bishop and Walker The Economics of EC Competition Law: Concepts, Application and Measurement (Sweet and Maxwell, London 2010) at 305. 28 GOLIATH AJ This case requires the Court to evaluate the merits of two of these cost standards which were pleaded to illustrate that Media24 was participating in predatory pricing. The Competition Commission contends that pricing below average total cost could amount to predatory pricing if it is accompanied by additional evidence of intention to predate or a predatory scheme. Media24, on the other hand, submits that relying on average total cost plus additional evidence is an unreliable way of evaluating whether predation has occurred. Instead, Media24 proposes that average avoidable cost be used as the appropriate costs standard below, which prices charged indicate a violation of section 8(c). The Tribunal found in favour of the Competition Commission, while the Competition Appeal Tribunal agreed with Media24 and excluded average total cost plus additional evidence as a relevant cost measure. However, because of the difficulties inherent in all cost standards, it would not be wise for this Court to tie itself too closely to any particular cost standard. The ultimate concern in cases such as these should be whether pricing below a cost standard could lead, or led, to the exclusion of a competitor, which exclusion had anti- competitive effects that are not outweighed by the gains listed in section 8(c). Profit sacrifice One of the crucial indicators of predatory pricing is profit sacrifice. This refers to a dominant firm deliberately incurring losses relative to alternative non-predatory conduct in the short-term by pricing its products too low. The profit sacrifice test assumes that a firm would not rationally engage in exclusionary conduct unless it considers the short-term loss of profits to be less than any expected gains achieved as a result of excluding or discouraging rivals.71 When a firm prices below average avoidable cost (or average variable cost), sacrifice is easily established because the firm is actively making a loss on a particular product line. This is because the product- specific costs incurred are not being covered by the price that is charged for the product. 71 O’Donoghue and Padilla above n 59 at 227. 29 GOLIATH AJ It is presumed that there can be no rational justification for making such a loss and thus, the pricing is predatory. However, a firm that prices below average total cost is still participating in profit sacrifice even if it is not actively loss-making. Common costs are allocated to product-lines in multi-product firms on some rational accounting basis. However, management makes the conscious decision to sacrifice part of the contribution that the firm could, and in accounting terms, should be making to common costs in favour of charging lower prices. While profit sacrifice is relied upon in many cases of predatory pricing, criticism of the test does exist. Academics question whether the profit sacrifice test requires a firm to opt for the most profitable business strategy to avoid a finding of exclusionary conduct. It is also unclear what degree of sacrifice is required to establish exclusionary conduct, or whether the rule is strict in that any profit sacrifice indicates abuse.72 In response to these criticisms, some jurisdictions, such as the United States, have adopted a variant of the test called the “no economic sense” test.73 This test establishes that conduct is not exclusionary unless it would not make economic sense for a dominant firm to pursue this behaviour save for its tendency to eliminate or lessen competition.74 By combining the usual test for profit sacrifice with the “no economic sense” test, other jurisdictions have been able to develop an approach to predatory pricing that ensures that only profit sacrifice which is undertaken exclusively to eliminate competitors is prohibited. This is essentially the same approach which the parties before us argue section 8(c) requires. It is clear from the discussion above that there is no one cost standard that best illustrates profit sacrifice in all cases. Instead, the circumstances of each case must be considered to determine which cost standard is most appropriate to prove that sacrifice has occurred. 72 Id at 229. 73 Id at 230. 74 Id. See also Verizon Communications, Inc. v. Law Offices of Curtis V. Trinko, LLP 540 US 398 (2004). 30 GOLIATH AJ Anti-competitive effects Once profit sacrifice is established, it must be determined whether the sacrifice excluded or could exclude a competitor with anti-competitive effects.75 The anti- competitive effect element can be determined based on whether the below-cost pricing was for a legitimate business reason, or whether the only rationale was to drive competitors out of the market. In SAA, the Competition Tribunal held: “If the conduct meets the requirements of the definition, we then enquire whether the exclusionary act has an anti-competitive effect. This question will be answered in the affirmative if there is (i) evidence of actual harm to consumer welfare or (ii) if the exclusionary act is substantial or significant in terms of its effect in foreclosing the market to rivals. This latter conclusion is partly factual and partly based on reasonable inferences drawn from proven facts. If the answer to that question is yes, we conclude that the conduct will have an anti-competitive effect. Whichever species of anti- competitive effect we have, consumer welfare or likely foreclosure, we have evidence of a quantitative nature and hence we can return to the scales with a concept capable of being measured against the alleged efficiency gain.”76 There are many legitimate business reasons why a firm might choose to price a product below average total cost. Two such examples come immediately to mind. First, the levels of competition in particular markets fluctuate over time. It may be, therefore, that new entrants can afford to charge lower prices and the dominant firm is forced to decrease its prices in the short-term to meet the prices that competitors are charging.77 Second, when dominant firms want to encourage customers to become familiar with a product, they will drop the price of the product in the short-term as a promotion.78 75 O’Donoghue and Padilla above n 59 at 227. 76 Competition Commission v South African Airways (Pty) Ltd [2005] ZACT 50 (SAA) at para 132. 77 O’Donoghue and Padilla above n 59 at 285. 78 Id at 291. 31 GOLIATH AJ Neither of these price reductions are predatory because they are done with the intention of attracting customers and not of excluding competitors from the market.79 Because of the multiple justifications that exist, pricing below average total cost is considered lawful in some jurisdictions, but competitors may adduce evidence to prove that, in the circumstances of the specific case, the pricing was predatory. The Tribunal relied on the decision of the European Court of Justice in AKZO, which is the seminal European authority for the contention that pricing below average total cost must be regarded as abusive if it is determined to be part of a plan for eliminating a competitor.80 In terms of the rule established in AKZO, prices that are above average variable cost (or average avoidable cost) but below average total cost are only abusive when they are part of such an eliminatory plan, meaning that in all other instances, below average total cost pricing is considered to be lawful.81 In terms of this decision, a plan is usually shown first through direct evidence of intent arising from the dominant firm’s documents, and second through indirect factors which, taken together, show an anti-competitive intention underlying the price-cutting.82 Europe is not the only jurisdiction that relies on the presumption that pricing below average total cost is lawful, but can be shown to be otherwise with additional evidence. The United States has not endorsed any particular cost standard and instead holds that predatory pricing must be shown to be below some measure of incremental cost.83 This rule was articulated by the Supreme Court of the United States in Brooke Group, where the Court held that only below-cost pricing may be predatory.84 In defining below-cost pricing, the Court held that this referred to pricing that was below 79 Id at 284. 80 Case C-62/86 AKZO Chémie BV v Commission of the European Communities [1991] ECR I-3359 (AKZO) at para 72. 81 O’Donoghue and Padilla above n 6 at 249. 82 Id. 83 Bolton, Brodley and Riordan above n 60 at 20. See also Brooke Group Ltd v Brown and Williamson Tobacco Corp 509 US 209 (1993) (Brooke Group) at 222-3. 84 Brooke Group id at 223. 32 GOLIATH AJ the average total cost of producing a product.85 Notably, the Court did not endorse average total cost but rather held that pricing anywhere below average total cost, including below the lower cost standards of marginal cost and average variable cost, may be predatory. Pricing below a measure of incremental cost is only the first stage of the inquiry used by the Court. The second stage requires that the competitor alleging that a dominant firm is engaged in predatory pricing show that there is a reasonable prospect or dangerous probability of the dominant firm recouping more than the losses incurred through below-cost pricing.86 While the requirement of recoupment has made successful prosecution of predatory pricing in the United States nearly impossible, the above is confirmation of the fact that other jurisdictions accept that pricing below average total cost may be predatory if it is accompanied by other evidence of a predatory scheme or predatory effects. Average avoidable cost or average total cost plus intent This case neither requires the Court to determine the correct test to prove predatory pricing nor to create a new test to be applied to section 8(c). Instead, this Court is called upon to decide whether the Competition Appeal Court’s decision to exclude a cost standard from the section is correct. If there are instances where the use of average total cost plus additional evidence may be the best test for the Competition Commission to rely upon to prove predation, then the appeal must succeed. Average avoidable cost has gained prominence in the past few decades as a reliable standard to determine whether predatory pricing has occurred. Both American and European academics recommend it as the cost standard of choice, and EU Guidance states that average avoidable cost should be used as the starting point in investigations 85 Hovenkamp Federal Anti-trust Policy: The Law of Competition and its Practice 3 ed (Thomas/West, St Paul 2005) at 346. 86 Brooke Group above n 83 at 224. 33 GOLIATH AJ into predatory pricing.87 It is true that average avoidable cost is generally accepted as the most readily ascertainable and reliable cost measure.88 However, as discussed above, it is not without its problems. The most pressing of these exists in multi-product firms where assets are transferred or redeployed from one operation to another. It is not clear in those instances which costs are avoidable and which are not.89 The Competition Commission asserted that average avoidable cost may be difficult to determine as it is inherently uncertain. Resultantly, average avoidable cost should not be used dogmatically as the test for predation in all alleged cases of predatory pricing. Media24 characterises average avoidable cost as never considering non- avoidable common costs. Non-avoidable common costs refer to non-product specific fixed and sunk costs. It also appears that average avoidable costs are calculated from the dominant firm’s perspective and not from the perspective of the firm that is the victim of predation. However, pricing below the cost of recovering fixed costs or sunk costs may, in some instances, exclude competitors. Multi-product firms The Competition Commission points to a few instances where fixed and sunk costs may contribute to the exclusion of a competitor. The first is where a multi-product firm is competing with an equally efficient single-product firm. Multi-product firms are firms which produce multiple product lines. The firm incurs product-specific fixed and variable costs for each product line but also has unavoidable costs that would be incurred regardless of how many different product lines the firm produced. These unavoidable costs are also referred to as common costs. The multi-product firm can spread out its common costs, with each product line contributing something to the recovery of these costs, while the single-product firm must incur and recover all of its costs from money made by one product. In order to compete with the multi-product 87 European Commission above n 58 at para 64. 88 O’Donoghue and Padilla above n 6 at 242 read with fn 19. 89 O’Donoghue and Padilla above n 59 at 248. 34 GOLIATH AJ firm which spreads out its common costs, the single-product firm would consistently be making a loss as it is not covering its fixed and sunk costs. In this situation the Competition Commission contends that average total cost may be the appropriate standard to use to evaluate whether predation has occurred because average total cost is the only standard which considers a portion of the company’s common costs. The length of the period over which predatory prices were charged is relevant to determining whether pricing above average avoidable cost but below fixed and sunk costs had exclusionary effects. Alleged predatory prices that last only one month may not cause an equally efficient rival to lose any money by not exiting the market unless those prices are lower than the short-run costs incurred by operating on a month-to- month basis.90 However, fixed and sunk costs become very relevant when low pricing occurs over a long period of time. Pricing that lasts for one or ten years will cause an equally efficient rival to lose money, relative to exiting the market, if the price does not allow it to cover the fixed costs of producing anything, like its overhead costs, the following year. Moreover, the longer a dominant firm prices at a low price, the less likely competitors would be able to make future capital payments for sunk costs like long-term property rent or maintenance of facilities.91 In these cases, it would be cheaper for the competitor to exit the market than remain in it and attempt to compete with the dominant firm’s prices. Some academics, including Elhauge, believe that the difficulties associated with average avoidable cost stemming from its exclusion of fixed and sunk costs can be rectified by simply changing the perspective from which costs are examined when calculating average avoidable cost. They contend that the common costs faced by the dominant firm should be viewed from the perspective of the competitor and would thus be avoidable.92 When deciding whether to sign a new long-term lease or exit the market, 90 Elhauge “Why Above-Cost Price Cuts to Drive out Entrants Are Not Predatory - and the Implications for Defining Costs and Market Power” (2003) 112 Yale Law Journal 681 at 708. 91 Id. 92 Id. 35 GOLIATH AJ the lease has yet to be signed and is thus avoidable to the competitor. Therefore, the cost of rent could be factored into the average variable cost below which the dominant firm cannot price. This approach, while intuitively appealing, faces a few obstacles. The first is that other commentators contend that it would be nonsensical to accuse a dominant firm of predatory pricing every time a competitor finds itself unable to make a profit.93 Moreover, expecting a dominant firm to price above the average avoidable cost of its competitors is inherently uncertain as to which costs are avoidable as it depends on the circumstances that the competing firm finds itself in. It is unclear how a dominant firm would ascertain a competitor’s average avoidable cost to ensure that pricing is above it. Due to this uncertainty, a cautious firm would price its goods based on the assumption that the firm that will accuse it of predation is a competitor who is yet to enter the market. This prospective competitor’s average avoidable cost would be equal to its average total cost as it has yet to incur any fixed or sunk costs, making all costs avoidable. Media24 argues that it is dangerous to require firms to price above their average total cost as it will harm the economy in the long run. It is unclear, however, why the criticisms that Media24 level against average total cost would not equally apply to the situation described here. The final problem with this approach stems from the fact that Media24 pleaded average avoidable cost based on the costs incurred by the dominant firm, and not those from the perspective of the competitor. This approach was not challenged by the Competition Commission or criticised by either the Tribunal or the Competition Appeal Court. It appears therefore that, at least in the context of the present case, it is common cause that average avoidable cost is examined from the dominant firm’s perspective. Average total cost is the one cost standard that allows regulators to look at fixed and sunk costs in order to determine whether pricing below these costs resulted in the exclusion of a competitor. In his scathing critique of the average total cost standard, 93 Bishop and Walker above n 70 at 306. 36 GOLIATH AJ Baumol argues that there is, in reality, no such thing as a single-product firm.94 And as a result, the reliance on average total cost in instances where a single-product firm competes with a multi-product firm is only ever hypothetical. Even if single-product firms are indeed rare, it is far too much of a generalisation to assume that they do not exist outside of the pages of textbooks. Indeed, in the case before us, GNN was running a single-product firm for the duration of the alleged predation period even if it initially started out as a multi-product firm. The generally difficult task of calculating fixed and variable costs becomes more complex when a firm produces multiple products.95 This is because the firm may have fixed and variable costs that are incurred jointly between the various products.96 There is no unambiguous way of calculating what proportion of these common costs are incurred by any particular product.97 There appears to be two approaches to the problem created by common costs in multi-product firms. The first is to ignore these costs altogether, while the second is to undertake the complicated task of allocating the costs between products.98 The first approach essentially is to exclude common costs from all considerations of cost standards.99 This approach is the one that is favoured by Media24 in that they argue that average avoidable cost, the cost measure which excludes common costs, is the most appropriate one to prove predation. This means that the only relevant concern is whether a price enables a firm to recover the product-specific costs of producing it, and not whether it is able to cover these costs and make a contribution towards common costs.100 The problem with this approach is that it disadvantages competitors who 94 Baumol “Predation and the Logic of the Average Variable Cost Test” (1996) 39 Journal of Law and Economics 49 at 59. 95 O’Donoghue and Padilla above n 6 at 260. 96 Id. 97 Id. 98 Id at 261. 99 Id at 260. 100 Id. 37 GOLIATH AJ produce only a single -product. Even if the firm is as efficient as the dominant firm, it may still be forced to exit the market as it must recover all of its costs from one product without the luxury of recovering common costs from different product lines.101 This is the problem that the case before us presents. The second approach is to undertake to allocate the common costs between product lines in order to determine the average total cost of a particular product.102 It is true that the calculation of this amount is complicated and cumbersome but other jurisdictions have developed guidelines for the allocation. In Claymore Dairies, the UK Competition Appeal Tribunal set out general guidance for the allocation of common costs.103 The Tribunal held that— “[t]here are conventional accounting methods for making such allocations (e.g. by volume, value, time, etc.) but the most appropriate yardstick to use may be debateable. One approach, shared by the expert witnesses in the present case, is to seek to identify “the cost drivers”, i.e. to determine the factors that cause the costs to be incurred and then make allocations appropriately. So far as possible, cost allocations should reflect the underlying business reality. A reasonably detailed understanding of the nature of the business, and how costs arise, is generally necessary when determining how particular costs should be allocated. Similarly, how the business itself treats the costs in its internal management accounts will normally be an invaluable source of information. . . . However the allocations are ultimately to be made, it is important in our view that the investigation is grounded on a firm and reliable assessment of what the total costs are, cross-checked as far as possible against the dominant undertaking’s statutory and management accounts.”104 101 Id at 261. 102 Id. 103 Claymore Dairies Limited and Arla Foods UK PLC v Office of Fair Trading [2005] CAT 30 (Claymore Dairies). 104 Id at paras 210-1 and 216. See also O’Donoghue and Padilla above n 6 at 264. 38 GOLIATH AJ The above shows that while there may be difficulties with the allocation of common costs, it is neither an impossible task nor one that courts in other parts of the world are unwilling to undertake. Importantly, the use of any specific cost measure is motivated by the facts of the case.105 For instance, where a multi-product dominant firm is operating in a market which has high barriers to entry, the long-run average incremental cost standard would likely be most appropriate and thus be used instead of average total cost.106 In cases where a multi-product firm is predating an equally efficient single- product firm, it is the fact that the multi-product firm can spread out its common costs that contributes to the exclusion of the competitor. It would make sense therefore to consider common costs, and thus use average total cost, when evaluating whether predation has occurred in multi-product firms. The second example provided by the Competition Commission of a situation where fixed and sunk costs might lead to the exclusion of a competitor is where a less efficient competitor is the only firm competing with a dominant firm which would, but for that competitor, hold a monopoly in the market. As the firm is less efficient, it will not be able to compete with the prices levied by the dominant firm. While competition law is generally not concerned with protecting less efficient competitors, for the sake of protecting the market from monopoly, pricing below total costs should be considered as possibly exclusionary in this case.107 The above illustrates that there are instances where it is possible to have profit- sacrificial conduct below average total cost which excludes an equally efficient competitor and thus, while pricing below average total cost should not be determinative 105 O’Donoghue and Padilla id at 261. 106 Case COMP/35.141 Commission Decision of 20 March 2001 relating to a proceeding under Article 82 of the EC Treaty-Deutsche Post AG, 2001/354/EC. See also O’Donoghue id at 261-2 and Case 66/86 Ahmed Saeed Flugreisen and Silver Line Reisebüro GmbH v Zentrale zur Bekämpfung unlauteren Wettbewerbs eV [1989] ECR 803. 107 European Commission above n 58 at paras 23-4. 39 GOLIATH AJ of predation, it should be available to competition authorities. While the Competition Appeal Court is an expert body, the categorical exclusion of average total cost plus additional evidence of predation as a standard of proof is incorrect as it unduly limits the prosecutorial powers of the Competition Commission. This finding does not lay down a new standard, but rather makes the determination of the appropriate standard for predation dependent on the specific facts of each case. Evidence of predation before the Tribunal Because there are often legitimate business justifications for pricing above average avoidable cost or average variable cost but below average total cost, the Tribunal held that pricing below average total cost alone was not sufficient to justify a finding of predation.108 The Tribunal held that average total cost plus intent was an appropriate cost standard to examine in this case to establish whether predation occurred. While the test that it used has been colloquially referred to as “average total cost plus intent”, it would be more accurate to describe it as “average total cost plus additional evidence including evidence of intent”. This is because the Tribunal did not only examine evidence of subjective intention to predate. It relied on direct, subjective evidence of intention to predate, indirect and objective evidence of predation, evidence of recoupment and the equally efficient competitor test to show that the pricing below average total cost was executed for the purpose of eliminating GNN from the market.109 The Tribunal specified the meaning and content of both subjective intention110 and objective intention.111 This judgment endorses the Tribunal’s interpretation of these terms.112 108 Tribunal decision above n 3 at paras 233-8. 109 Id at paras 242-8. 110 See [23]. 111 See [24]. 112 The Tribunal decision above n 3 at paras 260-1 held that— “Direct intent is the type of intent that we have discussed up until now - that which emanates from the documents of the accused firm and statements by its employees. Indirect intent has a greater economic basis. Examples cited would be reputational reasons for predation, targeting of specific customers and competitors reliant on external funding. 40 GOLIATH AJ Intent The Competition Commission argues that evidence of subjective intention to predate is not determinative of predatory pricing, but rather that this evidence has probative value when coupled with other evidence, including objective evidence that corroborates the subjective evidence of intention. This argument is a considerable departure from the arguments that the Competition Commission presented before the Tribunal and the Competition Appeal Court. Before those bodies, the Commission contended that average total cost plus evidence of subjective intention was definitive proof of predation. The angle that the Commission’s argument before us takes in terms of subjective intent is in line with the manner that this form of intent is treated in other jurisdictions. The European Court of Justice has examined what constitutes direct evidence of intention to predate. In AKZO, the Court relied on documentary evidence of a detailed plan to eliminate the competitor.113 In Tetra Pak II, the Court relied on a report by the company’s board of directors indicating that financial sacrifices should be made to fight competition.114 In Wanadoo, the Court held that subjective intent could be proven by looking at both formal presentations by management and informal remarks made by staff members but that the former had higher evidential value.115 The European Courts do recognise the difficulties with direct or subjective evidence of intention to predate. In AKZO, the Commission noted that it does not As the name suggests, indirect intention is not reliant on direct intention as expressed by the impugned firm but circumstantial economic evidence that may give rise to an inference that conduct is predatory. Indirect intention received more attention in a document that preceded the Guidance, which is referred to as the Discussion Paper, yet the approach to analysing indirect intention still forms part of the present Guidance.” 113 AKZO above n 80. See also O’Donoghue and Padilla above n 6 at 249. 114 Case T-83/91Tetra Pak International SA v Commission [1994] ECR 11-765. 115 Case COMP/38.233 Wanadoo Interactive Commission Decision 16 July 2003. See also O’Donoghue and Padilla above n 6 at 250. 41 GOLIATH AJ consider an intention to prevail over competitors as unlawful.116 Moreover, commentators suggest that the enforcement of competition law should not be dependent on whether a firm uses “commercially correct” language in its correspondence.117 Resultantly, evidence of subjective intent is treated with caution even in Europe, being the jurisdiction which relies most heavily on intent as an element of predation. In response to these problems, reliance has been placed on indirect evidence of intent, which is really just objective evidence that assists a court to make inferences about the explanation for the below-cost pricing.118 This encompasses evidence of a plausible predatory strategy in a particular market.119 Examples of indirect evidence include the actual or likely exclusion of a competitor, whether certain customers are selectively targeted, the scale, duration, continuity of the low pricing, the possibility of the dominant firm off-setting its losses with profits earned on other sales and the possibility of recoupment of losses through a future return to high prices.120 This approach was endorsed both in AKZO and Tetra Pak II.121 However, when reliance is placed on indirect evidence, there must be a strong evidential basis for alleging that the below-cost pricing had no rational business justification.122 If pricing below average total cost only makes commercial sense as part of a predatory strategy and there are no other reasonable explanations, indirect evidence will be sufficient to show that predation has occurred. The European approach to pricing below average total cost appears to be in line with the only South African authority on predatory pricing in terms of section 8(c) – the Competition Tribunal decision in Nationwide.123 The Tribunal held: 116 AKZO above n 80 at para 81. 117 O’Donoghue and Padilla above n 6 at 250. 118 Id. 119 Id. 120 Id at 251-2. 121 AKZO above n 80 and Tetra Pak II above n 114. See also, O’Donaghue above n 6 at 252. 122 O’Donaghue id. 123 Nationwide Airlines v South African Airways (Pty) Ltd [2001] ZACT 1 (Nationwide). 42 GOLIATH AJ “[I]f a complainant, relying on section 8(c), can show that a respondent’s costs are below some other appropriate measure of costs not mentioned in [section 8(d)(iv)] it may prevail provided it adduces additional evidence of predation beyond mere evidence of costs.”124 While the decision of the Competition Tribunal in Nationwide is obviously not binding on this Court, both the Competition Tribunal and the Competition Appeal Court referred to the extract approvingly, despite their disagreement over what additional evidence should entail.125 Nationwide has therefore been accepted as a useful starting point by both of the specialist bodies on matters of competition law. In my view, average total cost may be an appropriate cost standard in some instances and courts should be given the power to investigate pricing that is below average total cost, including examining the additional evidence that is adduced to illustrate predation. The critiques of subjective evidence are compelling and justify an approach that illustrates that the management of a firm had the subjective intention to predate does not mean that this conduct had exclusionary effects. It appears that section 8(c) focuses on the effects that conduct has on competition and there is no reason why objective evidence cannot be presented to prove that pricing that excluded a competitor had anti- competitive effects that outweighed pro-competitive gains. This evidence would include objective, indirect evidence. Media24 raised a number of potential harms that a finding endorsing average total cost plus additional evidence as an appropriate cost standard would have on the South African economy. Most notably, they argued that dominant firms would, because of fear of allegations of predation, primarily charge prices above average total cost and that this would harm consumers because they would have to pay high prices for products. The second concern was that multi-product firms would be discouraged from 124 Id at 10. 125 Competition Appeal Court judgment above n 4 at paras 45 and 55. See also the Tribunal decision above n 3 at paras 100-1. 43 GOLIATH AJ growing their firms and running different product lines because those lines would be less profitable. This, it was submitted, would harm consumers as it would limit their product choices. Both of these eventualities are extremely undesirable. It is important, therefore, that the approach taken strikes an appropriate balance between under and over- enforcement. Pricing below average total cost that is to the benefit of consumers and competition in general should not be discouraged. As a result, pricing below average total cost should be considered lawful. However, competitors may raise additional evidence to prove that the profit sacrifice undertaken by the dominant firm was only beneficial to the dominant firm insofar as it excluded competitors. This is a high evidential burden. If pricing behaviour only makes commercial sense as part of a predatory strategy, and there are no other reasonable explanations, this will meet the burden. The Competition Commission presented various pieces of evidence to the Competition Appeal Court, all of which potentially pointed towards proving that predation had occurred. Some of the evidence proves the intention to predate. However, the balance of the evidence was objective and allegedly pointed towards predation being the only rational reason for the low prices charged and towards the anti- competitive effect of these prices. The Competition Appeal Court dismissed all of the additional evidence out of hand on the basis that intention has no role to play in the interpretation of section 8(c). This meant that it failed to examine any of the other evidence that, while framed as indirect evidence of intent, is in actual fact objective evidence to establish the purpose for which the below cost prices were charged. Equally efficient competitor test Media24 alleged before the Tribunal that the reason for GNN’s exit from the market was not related to Forum’s pricing but instead occurred because GNN was a less 44 GOLIATH AJ efficient competitor that was unable to survive in a volatile market.126 It was contended that average avoidable cost and not average total cost is the appropriate cost standard to use in cases involving section 8(c) because pricing above average avoidable cost will not lead to the exclusion of an equally efficient competitor. It is a fundamental principle of competition law that firms should be allowed to maximise their profits by charging prices that are lower than those charged by their competitors.127 In general, “less efficient firms should not receive any protection from aggressive competition since consumers are best served by more efficient firms”.128 This is because competition law is concerned with protecting competition and not competitors.129 It should be noted that the second example that the Competition Commission points to in respect of which average total cost may be an appropriate cost standard to rely upon is where a less efficient competitor is the only competitor preventing a dominant firm from becoming a monopoly. The Tribunal held that GNN and Forum were equally efficient but even if that was not the case, this may not be fatal to the Competition Commission’s case.130 There are concerns in academia about the equally efficient competitor test. The first concern is that less efficient competitors can increase consumer welfare if the increased competition that their presence in the market creates outweighs their relative inefficiency.131 In Europe this has led to an acceptance that the duty to protect firms from predation is not limited to protecting equally efficient firms, but also, in exceptional circumstances, less efficient firms.132 126 Tribunal decision id at paras 511-2. 127 Kelly et al Principles of Competition Law in South Africa (OUP, Cape Town 2016) at 2-3. 128 O’Donoghue and Padilla above n 6 at 233. 129 Id. 130 Tribunal decision above n 3 at paras 607-9. 131 O’Donoghue and Padilla above n 6 at 232. 132 Id. 45 GOLIATH AJ The second concern with the equally efficient competitor test arises in cases where economies of scope and scale dominate the market and competitors only compete for a limited portion of the market.133 The European courts seek to overcome this obstacle by putting aside the advantages that the dominant firm derives from being an unavoidable competitor for a portion of the market and focusing instead on equal efficiency only for that portion of the market that is contestable.134 It is important that economies of scale exist and that we do not prohibit legitimate competition on the merits. The protection of equally efficient firms only does not appear to be a hard-and-fast rule that must be applied in all investigations into potentially abusive pricing conduct. In terms of the definition of an “exclusionary act” in the Competition Act,135 there is no textual basis which mandates the protection of equally efficient competitors only. The definition speaks merely of “a firm” being actually or potentially impeded or prevented and not an equally efficient firm.136 Moreover, as discussed above, there are instances where the benefit that a competitor brings to the market through enhancing competition outweighs its inefficiency. The clearest example of this is where a dominant firm would be a monopoly but for the existence of one smaller, less efficient firm. The less efficient firm’s existence protects consumers from monopoly pricing and thus benefits them greatly. International authorities appear to support the relaxing of the equally efficient competitor test in these circumstances.137 This is the very instance that the Competition Commission relies upon to illustrate that pricing below average total cost may lead to exclusion in some cases. 133 Id. 134 Id at 233. 135 See definition in [14]. 136 Section 1(1) of the Competition Act. 137 European Commission above n 58 at para 24 and International Competition Network “Predatory Pricing Analysis” in Unilateral Conduct Workbook (International Competition Network, Rio de Jeneiro 2012) at paras 35-7. 46 GOLIATH AJ Further, in many industries the variable costs of large-scale production are cheaper than those associated with small-scale production. Here, new entrants are easily driven out of the market before they have had the opportunity to expand and reach their potential as an equally efficient competitor.138 In markets which are already dominated by one or a few firms, new entry by firms is desirable, as is their allowance to blossom into large scale producers.139 Given that the aims of the Competition Act include giving small- and medium-sized enterprises an equitable opportunity to participate in the economy, it would be inappropriate to rule that pricing below average total cost can never be predatory simply because it would not exclude an equally efficient competitor. Recoupment The Tribunal found that there was evidence of recoupment by Media24.140 Recoupment refers to the raising of prices by the dominant firm once the competitor has left the market.141 The Commission alleged that once GNN had left the market, Media24 raised the prices that it was charging advertisers for advertising space in Vista. O‘Donoghue and Padilla suggest that recoupment can be tested using two methods: either by examining the conduct itself or by looking at the structure of the market.142 The examination of conduct refers to determining whether the predation has paid off in that it has enabled the dominant firm to raise its prices. Such an increase must be to a level that enables the firm to recoup the losses that it made previously.143 To show that the structure of the market has changed, a number of factors can be examined. If the market share of a predator has grown, post predation, then it has gained power over a greater share of the market than it had before the predation period; in that 138 Bishop and Walker above n 70 at 306. 139 Id. 140 Tribunal decision above n 3 at paras 595-6. 141 O’Donoghue and Padilla above n 6 at 253. 142 Id at 253. 143 Id. 47 GOLIATH AJ sense there has been recoupment.144 Other factors which indicate changes to the market structure include examining what barriers of entry have been created and whether there are significant constraints on the capacity of competitors to enter and survive in the market.145 The Tribunal held that there was evidence of conduct of predation in that Media24 had raised its prices and there had been changes in the market structure as a result of GNN exiting the market.146 Treatment of evidence by the Competition Appeal Court The Competition Appeal Court held that average total cost plus intent was not an appropriate test to use to determine whether a violation of section 8(c) had occurred. It held that the test in section 8(c) for an exclusionary act is an objective one that examines the effect of conduct and not the intention behind it.147 Moreover, the Court held that because all firms competing with one another on the merits intend to harm their competitors, with the potential result of their exit from the market, it is difficult to distinguish predatory intent from the intention to compete.148 These criticisms are valid. However, they are largely directed at the subjective evidence of intent, and do not address the rest of the evidence which was objectively presented and aimed to show that the conduct had exclusionary effects. The Competition Appeal Court held that once the subjective evidence had been excluded, average total cost alone was not appropriate to illustrate that predation occurred.149 The Court was correct to hold that unsupported evidence of subjective intention should be excluded from investigations into predatory pricing. However, the 144 Id at 254. 145 Id at 253-4. 146 Tribunal decision above n 3 at para 621. 147 Competition Appeal Court judgment above n 4 at para 111. 148 Id at para 38. 149 Id at para 111. 48 GOLIATH AJ exclusion of this evidence does not require reliance on costing below average total cost alone as it is still supported by indirect, objective evidence of exclusionary effects, objective evidence of efficiency and objective evidence of recoupment. The failure to examine this evidence, particularly the evidence of recoupment is noteworthy, as many jurisdictions treat it as crucial to the establishment of predation. While the Competition Appeal Court may have concluded that there was insufficient evidence to prove recoupment, or that the indirect evidence and evidence of efficiency did not exclude all explanations for the below cost pricing, it was important for the Court to scrutinise the evidence to ensure the protection of consumer welfare. By failing to examine this evidence and excluding the average total cost standard altogether, the Competition Appeal Court impermissibly limited the ability of the Competition Commission to investigate, prosecute and present evidence of violations of section 8(c). Remittal The Competition Appeal Court made no finding as to whether Media24 had participated in predatory pricing based on the average total cost plus additional evidence test. The court dismissed the evidence of intention and found that average total cost alone is not sufficient to ever prove predation. It is true that average total cost alone is not sufficient to prove predation but this Court has now held that, where below average total cost pricing is alleged, a complaint may be found to have statutory warrant if there is significant additional evidence that the pricing was purely to predate and had anti- competitive effects. The question whether Media24 participated in predatory pricing remains open and the parties requested that it be remitted to the Competition Appeal Court to make a final decision on this in the event that average total cost plus additional evidence was found to be an appropriate test. 49 GOLIATH AJ This Court has previously ordered that matters be remitted to lower courts when it would be just and equitable to do so.150 The question whether Media24 engaged in predatory pricing based on the test that this Court has formulated is a question of fact combined with a question of economics. The Competition Appeal Court heard extensive arguments about the additional evidence proffered by the Competition Commission despite opting not to make findings based on it. The Court has had the advantage of oral testimony and cross-examination that will best allow it to assess the expert evidence that was presented. Moreover, this Court does not have jurisdiction to determine pure questions of fact. The examination of evidence and a subsequent finding on the validity of the predation complaint would require this Court to investigate factual questions without having properly tested the evidence. The remittal of this matter to the Competition Appeal Court does not mean that Media24 will necessarily be found guilty of predation. As discussed above, the evidentiary burden placed upon a complainant alleging that predation has occurred due to pricing below average total cost is high. However, the remittal will compel the Competition Appeal Court to scrutinise the evidence presented and then come to a decision about whether predation has taken place. This is an additional accountability mechanism to ensure that dominant firms do not abuse their power and monopolise markets. This mechanism is important given that the Competition Act envisages the empowerment of small- and medium-sized businesses and a more equitable share in the market by previously disadvantaged persons.151 Conclusion The Competition Commission has wide investigatory powers. This Court has recently held that the protection of these powers is paramount as it empowers the 150 Qhinga v S [2011] ZACC 18; 2011 (2) SACR 378 (CC); 2011 (9) BCLR 980 (CC); Road Accident Fund v Mdeyide [2010] ZACC 18; 2011 (2) SA 26 (CC); 2011 (1) BCLR 1 (CC); and Merafong Demarcation Forum v President of the Republic of South Africa [2008] ZACC 10; 2008 (5) SA 171 (CC); 2008 (10) BCLR 968 (CC). 151 Section 2(e) and (f) of the Competition Act. 50 GOLIATH AJ/ CAMERON J, FRONEMAN J and KHAMPEPE J Commission to fulfil its envisaged role.152 However, the decision of the Competition Appeal Court limits its prosecutorial powers arising from investigations which concluded that abuses of dominance had occurred. The Competition Appeal Court has dismissed all evidence of intention as irrelevant, even where that evidence is objective and points towards the effects of conduct rather than the motivation behind it. The Court’s failure to look past the subjective evidence and consider the objective evidence relating to recoupment and efficiency has severely hamstrung the ability of the Commission to prosecute cases where the fixed and sunk costs included in average total cost have contributed to the exclusion of a competitor. The Commission should be empowered to plead whatever costs benchmark best suits the facts of the case. This should include allowing the Commission to plead average total cost when there is sufficient additional evidence to illustrate that the pricing was lowered for no rational reason besides to force the exit of a competitor. In the result, I would have made the following order: 1. 2. Leave to appeal is granted. The appeal is upheld to the extent that the finding by the Competition Appeal Court, that pricing below average total cost plus additional evidence is an inappropriate standard to establish a violation of section 8(c) of the Competition Act 89 of 1998, is set aside. 3. The matter is remitted to the Competition Appeal Court to determine whether Media24 engaged in predatory pricing. 4. Media24 is ordered to pay costs of the applicant, which costs shall include the costs of two counsel. CAMERON J, FRONEMAN J and KHAMPEPE J (Petse AJ concurring): 152 Hosken above n 25 and S.O.S above n 25. 51 CAMERON J, FRONEMAN J and KHAMPEPE J We have had the privilege of reading the judgments of Goliath AJ (first judgment) and Theron J (third judgment). Whilst we admire the breadth of the first judgment’s treatment of predatory pricing benchmarks and recognise the added justification for doing so in the third judgment, we do not think it is in the interests of justice for this Court to embark on an exercise of that kind. In short, we doubt whether we have jurisdiction to entertain the application for leave to appeal and, even if we have, we do not consider it to be in the interests of justice to grant leave. We would therefore dismiss the application for leave to appeal. The first judgment considerably extends the constitutional jurisdiction of this Court in respect of the workings of the Competition Act.153 Previous cases have found constitutional jurisdiction in determining the boundaries of the powers and functions of the public bodies of the competition institutions created by the Competition Act, a legality issue.154 This is not the starting point for finding constitutional jurisdiction in the first judgment. Understandably so, because section 8 of the Competition Act is not a provision that purports to delineate those boundaries. So the first judgment seeks the constitutional jurisdictional links elsewhere. The first purported link is equality: “Section 2(e) and (f) of the Act states that part of the purpose of the Competition Act is to ensure that small- and medium-sized business have equitable opportunities to participate in the economy and to promote a greater spread of ownership in the economy by those who were disadvantaged by Apartheid. These purposes implicate the right of equality contained in section 9 of the Constitution. Section 9(2) enjoins the State to take legislative and other measures to advance the equality of previously disadvantaged people and section 2(e) and (f) of the Competition Act is a legislative measure of this kind.”155 153 Above n 2. 154 Hosken above n 25; S.O.S above n 25; and Senwes above n 56. 155 First judgment at [31]. 52 CAMERON J, FRONEMAN J and KHAMPEPE J The effect of this is to make the interpretation and application of the entire Competition Act a constitutional matter for purposes of the jurisdiction of this Court. This would mimic the way that the interpretation and application of the Promotion of Administrative Justice Act (PAJA)156 and the Labour Relations Act (LRA)157 are innately constitutional matters because they give legislative expression to fundamental rights in the Constitution.158 But the PAJA and the LRA templates are not appropriate. First, the analogy is inapt. The Competition Act is not the legislation the Constitution mandated to give effect to the right to equality in the way that PAJA and the LRA are statutes specifically mandated by the Constitution in terms of, respectively, the rights to just administrative action159 and fair labour rights.160 The Competition Act is just legislation. It has no elite constitutional allure. All legislation must conform to the constitutional imperative of equality. Second, this roundabout way of seeking a legislative link in the Constitution’s equality clause is unnecessary. If the interpretation and application of the provisions of the Competition Act do infringe the equality clause, this Court would have constitutional jurisdiction in any event, by virtue of that infringement alone. But this was not the basis on which the Commission brought its application. And notably and rightly, the first judgment does not dispose of the merits of the matter on equality grounds. The next link is said to reside in the indirect impact on the investigatory powers of the competition authorities: 156 3 of 2000. 157 66 of 1995. 158 In respect of PAJA, see State Information Technology Agency SOC Ltd v Gijima Holdings (Pty) Ltd [2017] ZACC 40; 2018 (2) SA 23 (CC); 2018 (2) BCLR 240 (CC) at para 17. In respect of the LRA, see Woolworths above n 44 at para 20. 159 Section 33 of the Constitution. 160 Section 23 of the Constitution. 53 CAMERON J, FRONEMAN J and KHAMPEPE J “The effect of the Competition Appeal Court judgment is to limit the tools that the Commission has at its disposal to prove a violation of section 8(c). This is because on the test it applied, it was unnecessary to examine certain evidence regarding predation. While evaluations under section 8(c) are economic in nature, a limitation of the investigative and prosecutorial powers of the Commission may have adverse consequences. Surely, if the ambit of the Commission’s investigatory powers raises a constitutional issue, the ambit of its ability to bring prosecutions based on findings made during investigation is, by parity of reasoning, similarly constitutional. What would be the purpose of the strong investigatory powers of the Competition Commission if it cannot present all of the evidence that it has gathered in order to prosecute an offence?”161 Apart from the objection that this is a “tail wagging the dog” argument, in that the tools to prove a violation and not the violation itself become the issue, it also entails that any wrongful exclusion of evidence in proceedings before the Tribunal and Competition Appeal Court will necessarily have a constitutionally significant impact on the investigatory powers of the Commission. As a general proposition this has attraction, but closer examination reveals that it has nothing to do with the legality issue relating to the nature and extent of the Commission’s investigatory powers, which is a constitutional issue. Those investigatory powers are set out in section 21(1)(c) of the Competition Act. This section empowers the Competition Commission to investigate alleged violations of Chapter 2 of the Competition Act. A finding by the Tribunal or the Competition Appeal Court on what evidence may be led at a hearing into violations of Chapter 2 does not in any way diminish the legal competence of the Commission to investigate these violations. To hold that it does would be tantamount to saying, in other settings, that any ruling on excluding the admissibility of evidence in a criminal matter necessarily impacts on the powers of the South African Police Service to investigate 161 First judgment at [33] to [34]. 54 CAMERON J, FRONEMAN J and KHAMPEPE J crime. Any evidential matter then becomes a constitutional legality issue. That is not correct. This Court’s decisions in S.O.S and Hosken offer no support for the proposition. They both dealt with the competence of the Commission to investigate specified aspects of merger activity under the Competition Act, not the manner in which the investigations should be executed. Whether the Commission has the competence to investigate particular subject matter under the Competition Act raises legality issues; what evidence it may legally present before the adjudicative bodies established under the Competition Act after its investigation does not. The first judgment finds final jurisdictional solace on the ground that the Commission’s case raises an arguable point of law of general public importance, namely “whether pricing above average avoidable cost but below average total cost amounts to predation [which] requires the interpretation of the law relating to exclusionary acts and the consideration of which tests can be used to equitably establish predation”.162 The third judgment adds further justification for this jurisdictional ground. This does not, however, establish any persuasive basis for this Court to exercise its jurisdiction. It is only in those rare instances where purely legal issues of interpretation arise in the work of the Tribunal and the Competition Appeal Court that this Court should intervene. This is not one of those rare cases, because it does not raise a purely legal issue, but a mixed one of fact and law. And that is where the problem lies and remains, because mixed questions of fact and law require evaluative assessments,163 and it is precisely those assessments that it is not the function of this Court to tread into. 162 First judgment at [35]. 163 Media Workers Association of South Africa v Press Corporation of South Africa Limited [1992] ZASCA 149; 1992 (4) SA 791 (AD). See also Mukaddam v Pioneer Foods (Pty) Ltd [2013] ZACC 23; 2013 (5) SA 89 (CC); 2013 (10) BCLR 1135 (CC); Dikoko v Mokhatla [2006] ZACC 10; 2006 (6) SA 235 (CC); 2007 (1) BCLR 1 (CC); Betha v BTR Sarmcol, A Division of BTR Dunlop Ltd [1998] ZASCA 5; 1998 (3) SA 349 (SCA); National Union of Mineworkers v Free State Consolidated Gold Mines (Operations) Ltd – President Steyn Mine; President Brand Mine; Freddies Mine [1995] ZASCA 109; 1996 (1) SA 422 (SCA).. 55 CAMERON J, FRONEMAN J and KHAMPEPE J The adjudicative institutions under the Competition Act are expert bodies and due recognition must be given to this, also in determining the proper constitutional competence of this Court in relation to competition matters. In addition to the accepted deference given to other courts in relation to factual findings,164 this means a similar deference to the competition authorities as better qualified to determine economic issues. In Yara, Cameron J stated: “In considering whether it is in the interests of justice to allow an appeal directly to this Court, the nature of the matter is relevant. In the same way as the views of the Supreme Court of Appeal are particularly pertinent in common law matters, so are the views of the Competition Appeal Court, as a specialist body, important in competition matters that lie at the complex intersection of law and economics.”165 Determining the appropriate benchmark for predatory pricing under section 8(c) of the Competition Act inevitably depends on an assessment of the relative merits of expert evidence in that regard. That does not fall within the functional competence of this Court.166 It is not a purely legal interpretative exercise. Section 167(3)(b)(ii) requires that this Court “ought to” consider the legal point. As Paulsen trenchantly explained— “[i]f – for whatever reason – it is not in the interests of justice for this Court to entertain what is otherwise an arguable point of law of general public importance, then that point is not one that ‘ought to be considered by [this] Court’.”167 164 ST v CT [2018] ZASCA 73; 2018 (5) SA 479 (SCA); Matlou v Makhubedu 1978 (1) SA 946 (A) at 950 A-E; Protea Assurance Co Ltd v Casey 1970 (2) SA 643 (A) at 648E; R v Dhlumayo 1948 (2) SA 677 (A) at 705-6. 165 Yara above n 30 at para 71. 166 Hosken above n 25 at para 31; and S.O.S above n 25 at para 21. 167 Paulsen v Slip Knot Investments 777 (Pty) Ltd [2015] ZACC 5; 2015 (3) SA 479 (CC); 2015 (5) BCLR 509 (CC) at para 30. 56 CAMERON J, FRONEMAN J and KHAMPEPE J / THERON J And even if one clothes it as an assessment of the reasonableness of the Competition Appeal Court’s assessment, one cannot escape that, in reality, it involves second-guessing the relative merits of different expert views. The final proof of this is in the remedy favoured in the first judgment: referral back to the Competition Appeal Court. What option does the Competition Appeal Court have other than to allow evidence of intention now? And how can we avoid the blunt reality that by compelling the Competition Appeal Court to allow evidence of intention this Court is itself setting the appropriate benchmark – in specialist territory that the statute specifically entrusts to that Court? For these reasons we would dismiss the application for leave to appeal with costs, including the costs of two counsel. THERON J (Basson AJ concurring): I have had the pleasure of reading the well-crafted judgments of my colleagues Goliath AJ (first judgment) and Cameron J, Froneman J and Khampepe J (second judgment). I concur in the first judgment that we have jurisdiction to adjudicate the appeal as it raises an arguable point of law of general public importance. I also concur that it would be in the interests of justice to grant leave to appeal.168 I, however, part ways with the first judgment and find that the Competition Appeal Court was correct in holding that the adoption of a test for predatory pricing under section 8(c) of the Competition Act that is based on the average total cost standard and the intention of a dominant firm is inappropriate. In this regard, I am of the view that the interpretation adopted by the first judgment of section 8(c) of 168 See first judgment at [57]. 57 THERON J the Competition Act is contrary to the objectives of the Competition Act and may serve to severely undermine price competition and in so doing harm consumer welfare. Jurisdiction In my view, and for the reasons given in the second judgment, this case does not raise a constitutional matter. It does, however, raise an arguable point of law of general public importance that ought to be considered by this Court. In Paulsen, this Court explained that a point of law is one that is not factual.169 In K, this Court held (in the context of vicarious liability) that what renders an issue legal (and, in that context, potentially constitutional) as opposed to purely factual, is an enquiry into the social policy and normative content behind a rule.170 When this Court is called upon to examine the otherwise inarticulate premises in the normative implications of a rule that is, in my view, a legal enquiry.171 To view this examination, even in the context of where a common law rule is imbued with social norms as in K, 169 Paulsen above n 167 at para 20, which explains: “The point must be one of law; and it must be arguable. Starting with the first prong, quite axiomatically, the point must not be one of fact. This Court’s jurisprudence on purely factual matters, developed in the context of what constitutes a constitutional, as opposed to a factual issue, is an instructive guide on this.” 170 K v Minister of Safety and Security [2005] ZACC 8; 2005 (6) SA 419 (CC); 2005 (9) BCLR 835 (CC) at para 22, which reads: “Despite the policy-laden character of vicarious liability, our courts have often asserted, though not without exception, that the common-law principles of vicarious liability are not to be confused with the reasons for them, and that their application remains a matter of fact. If one looks at the principle of vicarious liability through the prism of section 39(2) of the Constitution, one realises that characterising the application of the common-law principles of vicarious liability as a matter of fact untrammelled by any considerations of law or normative principle cannot be correct. Such an approach appears to be seeking to sterilise the common law test for vicarious liability and purge it of any normative or social or economic considerations. Given the clear policy basis of the rule as well as the fact that it is a rule developed and applied by the courts themselves, such an approach cannot be sustained under our new constitutional order. This is not to say that there are no circumstances where rules may be applied without consideration of their normative content or social impact. Such circumstances may exist. What is clear, however, is that as a matter of law and social regulation, the principles of vicarious liability are principles which are imbued with social policy and normative content.” 171 Id at para 23. 58 THERON J as a factual enquiry over which this Court has no jurisdiction will “sterilise” the law and “purge it from any normative, social or economic considerations”.172 In this case, there is clearly a point of law: is it appropriate to determine whether prices are predatory in terms of section 8(c) of the Competition Act by considering the average total cost standard and the intention of a dominant firm to predate? Answering this question entails critically examining the policy and normative implications of the various standards for predatory pricing. There is nothing factual about this. On the contrary, the facts in this case are common cause: all parties agree that Media24 was pricing below average total cost (but above the standard of average avoidable cost). The question in this matter is legal. No facts are evaluated by this Court.173 I also agree with the first judgment’s argument that the Competition Act should not be read to deprive this Court of jurisdiction over arguable points of law of general public importance.174 In this regard, section 62(2)(b) and (3)(b) of the Competition Act states that the Competition Appeal Court does not have exclusive and final appeal jurisdiction for those matters that raise constitutional issues. In Soda Ash, albeit in a slightly different context, the Supreme Court of Appeal similarly interpreted the Competition Act to be consonant with the appellate jurisdiction that was conferred on the Supreme Court of Appeal by the Constitution (prior to the 17th Amendment).175 The approach adopted in Soda Ash is equally apposite in this context: where it would not be unduly strained, as in this case, section 62(2) of the Competition Act must be interpreted to accord with section 167(3)(b)(ii) of the Constitution which founds this Court’s jurisdiction. Our jurisdiction is accordingly engaged. 172 Id at para 22. 173 Even if this matter concerned a mixed question of facts and law, in Mtokonya v Minister of Police [2017] ZACC 33; 2018 (5) SA 22 (CC); 2017 (11) BCLR 1443 (CC) at paras 43-4, this Court endorsed the proposition that if a rule of law must be applied prior to the reaching of a conclusion, then that conclusion is necessarily one of law. It follows that we may have jurisdiction to adjudicate matters involving mixed questions of fact and law. 174 First judgment at [35]. 175 American Natural Soda Ash Corporation v Competition Commission of South Africa [2005] ZASCA 42; 2005 (6) SA 158 (SCA) (Soda Ash) at paras 13-15. 59 THERON J Leave to appeal Though this matter raises an arguable point of law of general public importance, it does not follow that it “ought to be considered” by this Court. In Paulsen, this Court explained that an arguable point should only be considered by this Court if it would be in the interests of justice to do so. This enquiry is identical to that employed by this Court in establishing whether leave to appeal should be granted in respect of a constitutional matter.176 Media24 posited two reasons why leave to appeal should not be granted. First, it argued that the Commission should have approached the Competition Appeal Court for leave to appeal before approaching this Court. Second, it argued that the Commission should have also first approached the Supreme Court of Appeal before appealing to this Court. Approaching the Competition Appeal Court for leave to appeal Media24 argued that section 63(2) of the Competition Act always requires the Commission to apply to the Competition Appeal Court for leave to appeal to this Court against a decision of that Court.177 It contended that the Commission’s failure to do so was fatal with the result that this Court cannot hear the Commission’s appeal. Section 63(2) of the Competition Act provides: “An appeal in terms of section 62(4) may be brought to the Supreme Court of Appeal or, if it concerns a constitutional matter, to the Constitutional Court, only— (a) with leave of the Competition Appeal Court; or 176 Paulsen above n 167 at paras 29-31. 177 Such matters are listed in section 62(2) read with section 62(3)(b) and this Court’s finding that section 62(2) must include appeals concerning “arguable points of law of general public importance”. 60 THERON J (b) if the Competition Appeal Court refuses leave, with leave of the Supreme Court of Appeal or the Constitutional Court, as the case may be.”178 Media24 submitted that— (a) the word “only” in section 63(2) implies that an application for leave to appeal in the Competition Appeal Court is a necessary condition for approaching this Court; (b) the purpose of a prior approach to the Competition Appeal Court for leave to appeal ensures that this Court enjoys the views of that specialist Court; and (c) section 167(6)(b) of the Constitution envisages the promulgation of national legislation to allow for a direct appeal from any other court,179 and the Competition Act is consistent with this. The Competition Act allows for a direct appeal to be made to this Court only after the procedural step of applying to the Competition Appeal Court for leave to appeal has been complied with. The challenge for Media24 is that the third leg of its argument cannot be accepted. This is for the reasons given by Cameron J and Yacoob J in their dissenting judgments in Yara180 and Loungefoam.181 Their reasoning in Yara is that the right to appeal to this Court under section 62(4) of the Competition Act is made subject to any law that specifically grants any right of appeal.182 Section 167(6)(b) of the Constitution 178 Section 62(4) reads that “[a]n appeal from a decision of the Competition Appeal Court in respect of a matter within its [non-exclusive] jurisdiction in terms of [section 62(2)] lies to the Supreme Court of Appeal or Constitutional Court, subject to section 63 and their respective rules”. 179 Section 167(6) reads: “National legislation or the rules of the Constitutional Court must allow a person, when it is in the interests of justice and with leave of the Constitutional Court— . . . 180 Above n 30. 181 Above n 30. (b) to appeal directly to the Constitutional Court from any other court.” 182 Section 63(1)(a)(ii) of the Competition Act. It would follow nonetheless that section 63(4) would be subject to and must be reasonably interpreted to accord with the Constitution. See section 1(2)(a) of the Competition Act, which requires that its provisions must be interpreted in a manner that is consistent with the Constitution, and 61 THERON J requires that national legislation or the Rules of this Court “must” allow a person to bring a matter directly to this Court when it is in the interests of justice to do so.183 Section 167(6)(b) specifically grants a right of direct appeal to this Court. It follows that the right to appeal provided for in section 62(4) of the Competition Act must be subject to section 167(6)(b) of the Constitution, which gives persons the right of a direct appeal to this Court if the interests of justice require.184 If a right to appeal in section 62(4) of the Competition Act is subject to section 167(6)(b) of the Constitution, then an applicant is entitled to rely on the pre-existing right to appeal directly to this Court in terms of rule 19 of this Court’s Rules. This pre- existing right does not require that an applicant first obtain leave to appeal from the Court below. Moreover, an applicant need not rely on the right in section 62(4) of the Competition Act, which is subjected to the restriction in section 63(2)(a) (that leave of the Competition Appeal Court to appeal against its judgment must be applied for). An applicant is empowered to appeal directly to this Court without first seeking the leave of the Competition Appeal Court. However, where the leave of the Competition Appeal Court is not sought, and the procedure in rule 19(2) of this Court’s Rules is invoked by an applicant, the matter is a direct appeal. In these circumstances, this Court must assess whether it is in the interests of justice to entertain the appeal without the Competition Appeal Court having had an opportunity to consider the application for leave to appeal.185 Factors which will Democratic Alliance v Speaker of the National Assembly [2016] ZACC 8; 2016 (3) SA 487 (CC); 2016 (5) BCLR 577 (CC) at para 33. 183 Rule 19(2) of the Rules of this Court provides for direct appeals as envisaged in section 167(6)(b) of the Constitution: “A litigant who is aggrieved by the decision of a court and who wishes to appeal against it directly to the Court on a constitutional matter shall, within 15 days of the order against which the appeal is sought to be brought and after giving notice to the other party or parties concerned, lodge with the Registrar an application for leave to appeal: Provided that where the President has refused leave to appeal the period prescribed in this rule shall run from the date of the order refusing leave.” 184 Yara above n 30 at paras 64-6; and Loungefoam above n 30 at para 23. 185 Yara id at paras 68-70; and Loungefoam id. 62 THERON J influence this assessment include the nature of the matter, the prospects of success of the appeal, the significance of the issues raised, whether the views of the Competition Appeal Court are useful or necessary to determine the appeal, and whether the subject matter of the appeal implicates complex questions of economics and law.186 Both Yara and Loungefoam were handed down on the same day by this Court, but were heard in different terms and have slightly different quorums. It is worth mentioning that the minority’s reasoning in Yara, while not endorsed by the majority in that matter, was considered to be “appealing” and assumed to be correct in the partial concurrence of Froneman J (with Skweyiya J and van der Westhuizen J concurring) in Yara.187 The issue whether the Competition Appeal Court must first be approached for leave to appeal was left open in the judgment of Zondo AJ (with Jafta J and Nkabinde J concurring).188 It follows that no Justice of this Court has objected to the minority’s reasoning in Yara and a majority found it to be, at the very least, “appealing”. As in Yara, the majority in Loungefoam, per Maya AJ did not consider it necessary to decide the issue regarding section 63(2) and approaching this Court without the leave of the Competition Appeal Court.189 The minority invoked their reasoning in Yara for why, in Loungefoam, the Commission did not need to approach the Competition Appeal Court for leave to appeal before this Court could consider the appeal.190 On the facts of this case, do the interests of justice permit the Commission to approach this Court directly? It is true that this matter involves complex questions of law and economics. This suggests that the Commission should be required to first approach the Competition Appeal Court for leave to appeal. On the other hand, this 186 Yara id at para 71; and Loungefoam id at para 36. 187 Yara id at para 77. 188 Id at para 21. 189 Loungefoam above n 30 at para 24. 190 Id at para 32. 63 THERON J Court has had the benefit of comprehensive argument. In addition, it has access to two judgments of the Tribunal and the Competition Appeal Court; these judgments each reach divergent conclusions on the relevant issues, which, along with reasons that will become apparent, indicate that there may be prospects of success;191 the matter primarily concerns statutory interpretation; and the determination of the proper legal test for predatory pricing is of immense significance. For these reasons, leave to appeal ought to be granted notwithstanding the Commission’s failure to approach the Competition Appeal Court for leave to appeal. Failure to approach the Supreme Court of Appeal Media24 argued that even if the Commission was not required to approach the Competition Appeal Court for leave to appeal, the Commission should first have appealed to the Supreme Court of Appeal before coming to this Court. Their argument was that section 168(3)(a) of the Constitution anticipates that legislation may limit the jurisdiction of the Supreme Court of Appeal in competition matters;192 Loungefoam held that sections 62 and 63 of the Competition Act confer appellate jurisdiction on the Supreme Court of Appeal in respect of matters mentioned in section 62(2) of the Competition Act;193 therefore sections 62 and 63 of the Competition Act confer on the Supreme Court of Appeal jurisdiction in competition matters falling within section 62(2) of the Competition Act. It follows that the Commission could have applied to the Supreme Court of Appeal. There is then no reason why it would be in the interests of justice to allow it to bypass that Court. The Commission agreed that section 168(3)(a) removes the Supreme Court of Appeal’s jurisdiction to hear appeals in respect of competition matters. The extent of 191 Similarly, this makes the point of law “arguable” as envisaged in Paulsen above n 167 at para 23. It is also trite that prospects of success are an important factor in determining leave to appeal. See S v Boesak [2000] ZACC 25; 2001 (1) SA 912 (CC); 2001 (1) BCLR 36 (CC) at para 12. 192 The section was amended in 2012 by the 17th Amendment to the Constitution to read: “The Supreme Court of Appeal may decide appeals in any matter arising from the High Court of South Africa or a court of a status similar to the High Court of South Africa, except in respect of labour or competition matters to such extent as may be determined by an Act of Parliament.” 193 Once again, section 62(2)(b) is now to be read as including arguable points of law. 64 THERON J such removal may be determined by an Act of Parliament. However, the Commission argued that sections 62 and 63(2), properly interpreted, do not give the Supreme Court of Appeal jurisdiction to hear appeals arising from the Competition Appeal Court that concerned constitutional matters under section 62(2)(b). Section 63(2) provides that an appeal “may be brought to the Supreme Court of Appeal or, if it concerns a constitutional matter, to the Constitutional Court”. In support of its argument, the Commission invoked the Supreme Court of Appeal’s decision in Computicket.194 In that matter, that Court held that section 63(2) of the Competition Act excludes the Supreme Court of Appeal from hearing an appeal which raises a constitutional matter as envisaged in section 62(2)(b). The Supreme Court of Appeal read section 61(2)(b) to mean that appeals concerning constitutional matters must be determined by this Court; they cannot “also” be entertained by the Supreme Court of Appeal.195 This approach would prevent a case which had already gone through two courts (the Tribunal and the Competition Appeal Court) from being considered by two further courts (the Supreme Court of Appeal and this Court) ostensibly without justification.196 On the face of it, Computicket appears to contradict Loungefoam. In Loungefoam this Court held that the Competition Act confers “appellate jurisdiction on both the Supreme Court of Appeal and this Court from the Competition Appeal Court in respect of constitutional and other matters listed in section 62(2)” and that this “is evident from the plain wording of sections 62 and 63”.197 This Court also held that the Commission had not demonstrated why the Supreme Court of Appeal should be bypassed – a finding that presupposes that the Supreme Court of Appeal was vested 194 Computicket above n 40. 195 Id at para 17. 196 Id. 197 Loungefoam above n 30 at para 19. 65 with jurisdiction to adjudicate the appeal in question.198 This Court could not have been clearer about its interpretation of the Competition Act: THERON J “Further, until the Legislature decides otherwise, the Supreme Court of Appeal also serves as a further filter in the appellate hierarchy, even in matters that do not explicitly involve the development of the common law.”199 The first judgment appears to pronounce on this question in favour of the Commission by finding that the Supreme Court of Appeal no longer has jurisdiction over matters envisaged in section 62(2)(b) of the Competition Act as contemplated in Computicket. Its terse reasoning is that the 17th Amendment clearly removes the Supreme Court of Appeal’s jurisdiction.200 It does not deal with this Court’s finding in Loungefoam that appellate jurisdiction is also conferred on the Supreme Court of Appeal from the Competition Appeal Court. It also fails to deal with whether section 63(2) nonetheless gives the Supreme Court of Appeal jurisdiction over section 62(2) matters as envisaged by the 17th Amendment.201 It also does not have regard to the further complication flowing from its finding that arguable points of law are also appealable under section 62(2)(b). This finding might imply that the Supreme Court of Appeal can no longer adjudicate appeals from the Competition Appeal Court concerning an arguable point of law. This approach could severely restrict the jurisdiction of the Supreme Court of Appeal in competition matters. At the same time, the first judgment considers whether it is in the interests of justice to hear the appeal without the Supreme Court of Appeal having adjudicated it first.202 It is accordingly unclear whether the first judgment’s finding that the Supreme Court of Appeal lacks jurisdiction is obiter. 198 Id at para 26. 199 Id. 200 First judgment at [47]. 201 One cannot assume that it does not do so only because it was passed before the 17th Amendment. 202 First judgment at [48] to [50]. 66 THERON J In my view, the issue is best left open. I am prepared to assume in Media24’s favour that the Commission could have approached the Supreme Court of Appeal. Even then, this is not fatal to the Commission’s case. The interests of justice permit this Court to decide a matter without the Supreme Court of Appeal having heard it first. This is for the same reasons why it was not necessary for the Commission to seek the leave of the Competition Appeal Court to appeal the judgment.203 Functional competence The second judgment suggests that this matter falls outside the functional competence of this Court as it entails an evaluative assessment of economic policy issues. I agree that the first judgment’s interpretation of the Competition Act implicates economic policy considerations. This Court has, however, repeatedly adjudicated over complex questions of public policy. In my opinion, there is no sound reason why public policy considerations of an economic nature suddenly swing the interests of justice in the opposite direction. The matter is properly before us and we should not shy away from our duty to determine the policy-laden issue of general public importance that it raises.204 For these reasons, leave to appeal should be granted. Merits On the merits, the first judgment reaches two conclusions: (a) A firm can be guilty of predatory pricing if it charges prices above average avoidable cost;205 and 203 See [151] to [156] above. 204 See S v Makwanyane [1995] ZACC 3; 1995 (3) SA 391 (CC); 1995 (6) BCLR 665 (CC) at para 187. 205 This standard is based on the costs a firm could have avoided by not engaging in a predatory strategy. The costs standard includes an element of fixed costs known as product-specific fixed costs. These are the fixed costs associated with the product or service that would not be incurred by the firm if it were to avoid producing the product or providing the service. See Competition Appeal Court judgment above n 4 at para 19. 67 (b) A firm can be guilty of predatory pricing if it charges prices below average total cost206 and has an intention to predate. THERON J I agree with the first judgment on its first finding.207 There may be instances where a firm charges above average avoidable cost but it is still guilty of predatory pricing. There is no need to spell out these circumstances in any detail. The Commission before us has pinned its case on one circumstance in which a firm would be guilty of predatory pricing even though it charged prices above its average avoidable cost. The Commission argued that where a firm with an intention to predate prices between average total cost and above average avoidable cost, then that firm is guilty of predatory pricing. I disagree that either average total cost or an intention to predate can ever be relevant to determining predatory pricing. This is borne out from a proper interpretation of section 8(c) of the Competition Act. For the reasons that follow, the Commission’s appeal should be dismissed. Proper approach to predatory pricing under section 8(c) of the Competition Act To determine the proper approach to prohibited predatory pricing under section 8(c) of the Competition Act, we are enjoined to interpret section 8(c) purposively, within its proper context, and in accordance with the ordinary grammatical meaning of its words.208 A key objective of the Competition Act is to provide consumers with competitive prices.209 A dominant firm may adopt a strategy of reducing its prices to exclude competitors from the market. This strategy may have short term pro-competitive effects for the economy in that prices for the goods or services in question are reduced. In the long term, however, the adoption of such a predatory pricing strategy may empower the 206 The average total cost standard advocated for by the Commission is calculated by dividing the total costs incurred by Forum in the production of its product by the number of units of the product that Forum produces. In this regard, Forum’s total costs include both its variable and fixed costs. 207 To the extent that the Competition Appeal Court found otherwise, it must be overturned. 208 Road Traffic Management Corporation v Waymark (Pty) Limited [2019] ZACC 12 at para 29. 209 Section 2(b) of the Competition Act. 68 THERON J firm to further abuse its position of dominance with exclusionary anti-competitive effects. The adoption of this predatory pricing strategy would not only lead to the undermining of competitive prices, but would violate the Competition Act’s objective of reducing “excessive concentrations of ownership and control within the national economy”.210 The regulation of predatory pricing poses significant challenges. The Competition Act recognises that firms have a right to compete on price. In this regard, the adequate protection of the competitive process between firms is essential to the achievement of the Competition Act’s objective of providing consumers with competitive prices.211 This inherent tension that is associated with predatory pricing implies that the approach adopted to determine the existence of exclusionary abuses of dominance under section 8(c) of the Competition Act should be carefully developed to avoid two types of errors, (commonly referred to as type 1 and 2 errors, respectively): (a) first, errors in which there is false condemnation of a dominant firm’s pricing behaviour as constituting prohibited predatory pricing;212 and (b) second, errors in which there is a failure to condemn predatory pricing by a dominant firm. In Nationwide, the Tribunal held that section 8(c) requires the use of an appropriate measure of costs to assess the prices that are charged by a firm accused of adopting a prohibited predatory pricing strategy.213 This must be correct because section 8(c) of the Competition Act, when interpreted in line with the object of the Competition Act to provide consumers with competitive prices, requires that a firm 210 See the preamble to the Competition Act. 211 See the Competition Appeal Court judgment above n 4 at para 27. 212 For example, where a firm might have low prices (and is competing on the price) but is still labelled as predatory by the regulation in question. 213 Nationwide above n 123 at 10. 69 THERON J should not be penalised for merely reducing its prices. Additional evidence would also be required to demonstrate that the general requirements for prohibited exclusionary abuses of dominance in section 8(c) are met. Section 8(c) of the Competition Act does not prescribe the application of a specific test to determine whether a dominant firm has engaged in prohibited predatory pricing. Section 8(d)(iv) of the Competition Act, however, states that the selling of goods or services by a dominant firm below the marginal or average variable costs is a prohibited exclusionary act unless the firm is able to show that the technological, efficiency or other pro-competitive gains associated with the act outweigh its anti- competitive effect. In Senwes, we held that an exclusionary act must fall outside the scope of section 8(d) for it to be prohibited by section 8(c).214 It follows that a complaint of predatory pricing brought in terms of section 8(c) of the Competition Act may not be founded on the cost formulas prescribed in section 8(d)(iv). I intend to demonstrate that the consideration of either average total cost or predatory intent under section 8(c) will lead to the type 1 and 2 errors referred to and consequently undermine the objectives of the Competition Act. Is average total cost an appropriate standard to measure alleged unlawful predatory pricing? The key question raised in this case is whether it is appropriate to measure prices against the standard of average total cost. The answer is no. This is for four reasons. First, a firm could be pricing below average total cost, which includes fixed costs, but still be engaging in legitimate competition. For example, where the price is sufficient to cover the avoidable cost for producing the unit, the firm may still be acting competitively by producing that unit even if the price is less than the average total cost. 214 Senwes above n 56 at para 28. 70 THERON J To hold otherwise would prevent firms from producing new products in circumstances where they would be deriving more benefits than costs from such production. Second, the use of a dominant firm’s average total cost in the determination of whether it has engaged in prohibited predatory pricing may operate as a price floor. This is because, if the average total cost standard is endorsed, a dominant firm that wishes to avoid being found to have engaged in a prohibited exclusionary abuse under section 8(c) may seek to avoid pricing its goods or services below its average total cost. This might encourage firms to raise their prices, especially in light of the potentially severe sanctions which they face under the Competition Act for engaging in prohibited exclusionary abuses of dominance. This price floor would undermine the objective of the Competition Act by inhibiting the achievement of competitive prices. Third, the average total cost standard without more can easily be manipulated by multi-product firms (which the respondent clearly is in this case, and which many dominant firms will be).215 This is because in multi-product firms, there are fixed costs that are shared across different products. For example, the fixed cost of renting a factory may be common to all the different products produced by the firm in that factory. The issue is that there are multiple ways of apportioning these common costs to the various products. The Commission accepts that under their approach, it is up to the dominant firm to decide how to rationally apportion the common costs. Without additional guidance, this makes the standard susceptible to manipulation: a firm can account for common costs through one product and not another in order to avoid regulation aimed at preventing anti-competitive conduct.216 Finally, the average total cost standard is inappropriate because it forces multi- product firms to ignore economies of scale.217 In practice, firms can add new products 215 Baumol above n 94 at 59. 216 As the Competition Appeal Court put it in its judgment above n 4 at para 53, the average total cost of a firm is “a figure that is undefinable and unmeasurable in a multi-product firm and must therefore be rejected as part of any legitimate test of predatory pricing”. See further Bolton, Brodley and Riordan above n 60 at 2272. 217 Elhauge above n 90 at 718. 71 THERON J to their lines of production and set prices close to or at marginal or avoidable cost because other products are funding fixed costs. This may be for the benefit of the consumer: it allows the lowering of prices. Given the finding that a dominant firm’s average total cost is an inappropriate benchmark for predatory pricing under section 8(c), the Commission has failed to demonstrate that the prices charged by Forum violated section 8(c). This is because the only standard against which the Commission tests Forum’s prices in this application is its average total cost. Is the intention of a dominant firm a relevant consideration? Section 8(c) of the Competition Act prohibits a dominant firm from engaging in an exclusionary act in the form of predatory pricing if the anti-competitive effect of that act outweighs its technological efficiency or other pro-competitive gain.218 An act is exclusionary if it impedes or prevents a firm from entering into, or expanding within, a market.219 The determination of whether a dominant firm has engaged in prohibited predatory pricing under section 8(c) of the Competition Act accordingly involves: (a) assessing whether a dominant firm’s pricing of its goods or services operates to prevent another firm from entering into or expanding within the market; and (b) weighing up the anti-competitive and pro-competitive effects of a dominant firm’s pricing of its goods or services. I endorse the finding by the Competition Appeal Court that these tests are objective in that they are outcome based.220 A dominant firm alleged to have engaged in predatory pricing under section 8(c) will only commit a prohibited exclusionary act where its pricing decisions operate to exclude competitors, and the associated anti- 218 Senwes above n 56 at paras 27-8. 219 Section 1 definition of “exclusionary act” in the Competition Act. 220 Competition Appeal Court judgment above n 4 at para 52. 72 THERON J competitive effects outweigh the associated pro-competitive effects. In this regard, the meaning of section 8(c) conveys the Legislature’s chosen approach to the regulation of predatory pricing by dominant firms. As noted by the Competition Appeal Court in Mittal, a court “may not eschew the text to promote its own theory, however attractive the latter may appear to be. In the event that the language of the text is unable plausibly to support the advocated theory, then it is for Parliament, if it so wishes, to reconsider the text”.221 I also endorse the Competition Appeal Court’s finding that when a competitive firm reduces its prices, it is possible that it intends to— “increase its market share by taking away customers from its rivals. In a real sense, such a firm intends to ‘harm’ its rivals but in a way permitted by competition policy. The firm may even hope that a prolonged price war may drive its rival from the marketplace.”222 It follows that the consideration of a dominant firm’s intent when assessing whether it has engaged in prohibited predatory pricing under section 8(c) is inappropriate. A firm’s intention is an “unreliable guide to proving predation” that does not assist in the evaluation of the likely economic effects of a dominant firm’s decision to price below a particular cost measure.223 I am unpersuaded by the Commission’s reliance on European jurisprudence in support of the use of predatory intent as a requirement for predatory pricing. This requirement was adopted by the European Court of Justice in AKZO224 in relation to allegations that a firm had violated Article 102 of the Treaty on the Functioning of the European Union by cutting its prices below average total cost (but above average 221 Mittal Steel South Africa Limited v Harmony Gold Mining Company Limited [2009] ZACAC 1 at para 28. 222 Competition Appeal Court judgment above n 4 at para 56. 223 Id. 224 Above n 80 at para 72. 73 THERON J/MHLANTLA J avoidable cost).225 Although Article 102 was held to be capable of incorporating a consideration of a dominant firm’s intent to predate, it does not resemble section 8(c) of the Competition Act. As warned by Kriegler J in Du Plessis, our Constitution does not “warrant the wholesale importation of foreign doctrines or precedents”.226 We may, of course, have regard to foreign law when interpreting section 8(c), but this should not displace the express meaning of the legislation regarding the appropriate test to be applied under the section. Conclusion Unlike in respect of a complaint of prohibited predatory pricing pursued against a dominant firm under section 8(d)(iv), the Commission bears the onus under section 8(c) of demonstrating that a dominant firm has engaged in an exclusionary act by implementing a predatory pricing strategy. The Commission is, however, afforded significant scope under this catch-all section to advance an appropriate cost standard against which to measure a dominant firm’s pricing practices. In the present matter, the Commission has failed to advance such a test. For these reasons I would grant leave to appeal but dismiss the appeal with costs, including the costs of two counsel. MHLANTLA J: I have had the benefit and pleasure of reading the judgments of my colleagues Goliath AJ (first judgment), Cameron J, Froneman J and Khampepe J (second 225 Article 102 of the Treaty (which was Article 82 of the Treaty Establishing the European Community at the time AKZO was decided) provides in relevant part: “Any abuse by one or more undertakings of a dominant position within the internal market or in a substantial part of it shall be prohibited as incompatible with the internal market insofar as it may affect trade between Member States. Such abuse may, in particular, consist in— (a) directly or indirectly imposing unfair purchase or selling prices or other unfair trading conditions.” 226 Du Plessis v De Klerk [1996] ZACC 10; 1996 (3) SA 850 (CC); 1996 (5) BCLR 658 (CC) at para 144. 74 MHLANTLA J judgment), and Theron J (third judgment). After due consideration, I find myself in the difficult position of agreeing with the third judgment on the issue of jurisdiction and leave to appeal only, and with the first judgment on the merits. I too am of the view that this matter does not raise a constitutional matter, but rather raises an arguable point of law. The first judgment notes that “[t]he case before us involves an interpretation of the extent of the powers held by the Competition Commission to prosecute alleged violations subsequent to an investigation”,227 which the first judgment notes is – amongst other reasons – indicative of the issue before us being a constitutional issue. I disagree with this. I wholly support the third judgment’s formulation of the point of law at issue.228 Further, I commend and support the third judgment’s excursus of the issue of leave to appeal, as well as its conclusion that it is in the interests of justice to permit the Commission to appeal directly to this Court without seeking leave to appeal from the Competition Appeal Court. Finally, I agree that whether the Commission could have approached the Supreme Court of Appeal should have been left open.229 I do not believe that this matter renders it necessary to decide whether or not the Supreme Court of Appeal’s jurisdiction in competition matters should be restricted. However, I am unable to agree with the third judgment’s conclusion and reasoning on the merits.230 I agree with the approach of the first judgment to the issue of the merits. In my view, the first judgment correctly recognises that this Court should be hesitant to tie itself to a cost standard to show predatory pricing pursuant to section 8(c) and 227 See [32]. 228 See [144] (para 5 of the third judgment). 229 See [163] (para 24 of third judgment). 230 See [166]-[187] (para 27-48 of the third judgment). 75 MHLANTLA J recognises the dangers in doing so.231 However, it also recognises the dangers in not allowing various costs standards to be employed to determine predatory pricing pursuant to section 8(c).232 In doing so, in my view, it does not commit to average total cost being used as a cost measure, but restricts it to instances where it is necessitated by the facts of each case. I agree with this approach. For these reasons and subject to the qualifications set out, I concur in the order made by the first judgment. 231 See [71]. 232 Id. 76 For the Applicant: For the Respondent: J Wilson SC, G D Marriott and T Ngcukaitobi instructed by Gildenhuys Malatji Incorporated W Trengove SC, M Norton SC and M Mbikiwa instructed by Werksmans Attorneys