---

COMPETITION TRIBUNAL OF SOUTH AFRICA

Case No: 14/LM/Feb06

In the matter between:

Main Street 333 (Pty) Ltd Acquiring Firm

And

Kumba Resources Limited Target Firm

Panel: N Manoim (Presiding Member), Y Carrim (Tribunal Member) and U Bhoola (Tribunal Member)

Date of hearing: 24 - 25 July 2006

Order issued on: 15 August 2006

Reasons issued on: 14 September 2006

Reasons for Decision

APPROVAL

1. On 15 August 2006, the Competition Tribunal approved without conditions the merger between Mainstreet 333 (Pty) Ltd and Kumba Resources Limited. The reasons for approving the transaction follow.

THE TRANSACTION

2. In terms of the transaction which has been dubbed “Project Pangolin,” the acquiring firm is Mainstreet 333 (Pty) Ltd or “BEE Holdco” which is directly controlled by a newly formed entity, Eyesizwe SPV. According to the parties, Eyesizwe SPV’s shareholding will be held by Anglo American (11%), an Employee Trust (10%), PWC (4%) and BHP (9%). Anglo American’s shareholding and BHP’s shareholding will entitle them to one (1) director each on the Eyesizwe SPV’s board. The remaining 66% in Eyesizwe SPV will be held by Eyesizwe Mining (66%), a black empowerment company with interests in the coal industry.¹ The coal mining activities of Eyesizwe Mining are currently conducted through a subsidiary, Eyesizwe coal, in which Anglo American has an 11% shareholding interest and BHP Billiton enjoys a 9% shareholding interest.² effect, post merger the shareholding of Eyesizwe SPV will mirror the current shareholding of Eyesizwe Coal. Their current shareholding entitles them to one (1) director each on Eyesizwe coal’s board. The pre-merger shareholding of Eyesizwe Group is depicted below:

3. The target firm Kumba Resources Limited (“Kumba”) is a publicly traded South African company that was formed in 2001 pursuant to the unbundling of Iscor Limited’s (now Mittal Steel South Africa) mining division. Kumba’s Iron ore activities are conducted through Sishen Iron Ore, its coal activities through Kumba Coal, its base metals business through Kumba Base Metals and its heavy minerals business through Ticor. Anglo American controls Kumba and nominates five (5) out of the fifteen (15) directors that sit on Kumba’s board.

4. The current shareholding of Kumba is graphically illustrated below:

5. Project Pangolin involves a number of complicated steps, which need not be to reproduced here as ultimately these transactions lead to the transformation of Kumba into two companies: Kumba Iron Ore Limited and Exxaro Resources Limited.

6. It is intended that Kumba’s coal, heavy minerals and base metals operations and assets will be combined with the coal operations and assets of Eyesizwe Coal within a newly created company, Exxaro. Exxaro will be controlled by BEE Holdco which will own approximately 55% of the company. According to the parties, Anglo American will initially own approximately 17% of the shares of Exxaro and it is proposed that Anglo American will have one representative on Exxaro’s board.

7. Kumba’s iron ore assets which are currently housed in Sishen Iron Ore Company (“SIOC”) will be sold to a wholly owned subsidiary of Kumba, to be called Kumba Iron Ore. Kumba Iron Ore will hold 74% of SIOC, while Exxaro will retain a 20% stake in SIOC. An Employee Share Option Plan and a Community SPV will hold 3% each post merger. Ultimately though, Anglo American will continue to exercise control over SIOC through its 66% shareholding in Kumba Iron Ore. Anglo American will also have an indirect economic interest in SIOC through Exxaro, through its 20% stake.

8. The merging parties provided the following post merger diagram which illustrates the relevant ownership structures after the Pangolin transaction:³

*The shaded boxes are relevant to the co-ordinated effects discussion later

THE RATIONALE FOR THE TRANSACTION

9. Anglo American is the key driver behind this deal. The reason for that is the controversy generated by Anglo’s original bid for control over Kumba Resources in 2002. When we wrote our decision approving that merger, we noted the battle for control over those assets between Anglo and the IDC.⁴ The battle was over whether a historically privileged mining house should be permitted to take control over another class of mineral asset. Anglo, no doubt sensitive to this criticism, had entered into an understanding with government at the time, in which inter alia it undertook to ensure its holding in Kumba remained below 50%, and to ensure the company remained listed. ⁵

10. Anglo had always indicated that its main interest in Kumba was its iron ore holdings and not its other mining assets. Project Pangolin resolves all these difficulties for Anglo. The creation of a significant Black owned and controlled resource company, valued at approximately R 24 billion, which has the assets and balance sheet to make it attractive to list on the JSE, resolves the problem of the undertakings made to government. ⁶ Splitting the iron ore business off, allows Anglo to retain a significant stake, approximately 66%, in the part of the Kumba business in which it is most interested. By giving the newly created Exxaro a 20% stake in the iron ore asset company SIOC, it bulks up the latter’s’ BEE profile, towards compliance with the goals of the mining charter. (Note that Anglo claims that together with interests held by employees and the local community, SIOC’s empowerment credentials will already be Charter compliant)

11. Exxaro is therefore a very ambitious project, and crucial to its early success is the fact that Anglo, and to a lesser extent BHP Billiton via Ingwe, remain invested in it. Anglo is responsible for a large financial commitment to the success of the venture that is disproportionate to its equity interest. For this reason it seeks not only equity in the venture, but board representation at operating company and shareholder level. This desire, which as we will see later, becomes a source of controversy with the Commission, is driven, says Anglo, by a need to protect its investment and its reputation, which requires the new venture to succeed. Anglo also maintains that its partners in the venture want it on board to give the group credibility in the market in its formative years. Not least in making these suggestions, Anglo claims, is its erstwhile foe in the Kumba scrap, the IDC, who it seems, has kissed and made up with Anglo, and supports its role in the present structure.

THE COMMISSION’S RECOMMENDATION

12. As will be discussed later, the Commission was of the view that the implementation of the merger would, as a result of coordination, have the effect of substantially lessening or preventing competition in the affected markets. In an effort to address their concerns, the Commission recommended the imposition of conditions which essentially sought to prohibit Anglo American from having representatives on the boards of either Exxaro or Eyesizwe SPV.

13. In light of the fact that the merging parties were unwilling to accept these conditions, it became necessary to conduct a formal hearing.

THE HEARING

14. A pre-hearing was held on the 21^st June 2006. The main hearing was held on the 24^th and 25^th July 2006. The Competition Commission did not call any witnesses. The merging parties, however led the following witnesses:

    1. Dr Robert Stillman, an economist from CRA International; and

    2. Mr Phillip Michael Baum, the chairman and chief executive officer of the Ferrous Metals and Industries Division of Anglo American.

15. Mr Reint Dykema from Solidarity Union and Mr Jeffrey Magida from NUM also made submissions. These will be dealt with later under the section on “public interest.”

COMPETITION ANALYSIS

The Parties’ activities and the Relevant market

16. BEE Holdco and Eyesizwe SPV are special purpose vehicles and have not previously engaged in any commercial activities. Eyesizwe Mining and Eyesizwe Coal are active in the exploration and extraction of coal. Kumba is active in the exploration and extraction of coal, iron ore, base metals and industrial minerals. Kumba’s controlling shareholder, Anglo American interests in gold, platinum, diamonds, coal, base metals, industrial minerals, ferrous metals and industry and forest products. Project Pangolin therefore results in a horizontal product overlap in the market for the exploration and extraction of coal. ⁷

17. Coal is an internationally traded commodity. According to the CRA economic report filed by the merging parties (hereinafter referred to as the “CRA report”), 27% of the coal produced in South Africa is exported and very little is imported. The rest is consumed domestically. We therefore agree with the Commission that the relevant geographic market is national. This is consistent with our previous findings in this market.

The Exploration and Extraction of Coal

18. Coal is a differentiated product that is categorised according to the degree of transformation of the original plant material to carbon. The ranks of coal from lowest to highest are lignite, sub-bituminous, bituminous and anthracite. Lower rank coals (lignite and sub-bituminous coals) are typically softer and are characterised by high moisture levels and low carbon content. Higher rank coals (bituminous and anthracite) contain less moisture, more carbon and have a higher calorific value.

19. Bituminous and Anthracite are the two types of coal mined in South Africa. Neither Kumba nor Eyesizwe produce anthracite and this product will not be discussed further. Bituminous coal can be further segmented into thermal or steam coal and metallurgical or coking coal.

20. Thermal coal is used in power generation and also has certain industrial uses while, metallurgical coal is used in the production of iron and steel. Because of differences in calorific values, thermal coal is significantly less expensive than metallurgical coal. According to CRA, the average price in South Africa in 2004 was less than 25% of the average price of metallurgical coal. Substitution of thermal for metallurgical coal is limited to PCI (Pulverised coal injection) ⁸ coal, of which CRA submits, there is limited use in South Africa. The parties argue that since there is a limited ability to substitute thermal for metallurgical coal in the steel industry and in other uses of metallurgical coal, the two sub markets should be distinguished as separate. We have previously accepted this delineation of the bituminous coal market as well as the distinction between thermal and metallurgical coal and see no reason to depart from this.⁹

21. According to the CRA report thermal coal, accounts for the vast majority of domestic coal production, consumption and exports.¹⁰ CRA derived the following table from the South African Coal Statistics 2005 Marketing Manual, August 2005.

	Thermal Coal	Metallurgical coal	Total
Domestic production (1)	236.8	7.8	244.6
Domestic consumption (2)	171.4	8.4	179.8
Import (3)	-	2.0	2.0
Exports (4)	65.4	1.4	66.8
(1) Calculated as Domestic Consumption + Exports – Imports. (2) This is the sum of local consumption of domestic production from Figure 20 (page 28) and imports; (3) Figure 64 (page 69) for metallurgical coal (equals import of coking coal plus metallurgical coal). No evidence of thermal coal imports found; (4) Figure 61 (page 66); figures relate to export capacity for steam coal and metallurgical/coking coal.

22. Anglo American, Kumba and Eyesizwe Coal are producers of thermal coal. However, only Kumba is active in the metallurgical coal sector and is the largest producer of the product in South Africa.¹¹ We will therefore limit our analysis to the thermal coal market.

23. Eskom and Sasol consume approximately 87% of the thermal coal used in South Africa - some 107.33 million tonnes and 41.05 million tonnes respectively in 2004.¹² Eskom obtains nearly all of its coal supplies through long-term contracts from mines that are adjacent to its power stations. However, Eskom’s current coal requirements sometimes exceeds the contractual volumes covered by these supply agreements and in these cases, Eskom would look to obtain additional supply from either extending an existing contract or purchasing extra coal on the spot market or through short term contracts from other coal suppliers. These are generally done on a tender and offer basis. Therefore competition to supply thermal coal to Eskom is primarily with regard to supply of any new power plants or shortfalls in respect of existing power plants.

24. Most of the coal required by Sasol’s coal gasification and chemicals plants is obtained from mines owned and operated by Sasol Mining. The merging parties submit that Sasol has adequate reserves to meet its coal requirmets for many years. This despite a recently concluded 20-year supply agreement with Anglo Coal. Sasol has begun to sell coal on the domestic market. According to CRA, in 2004, Sasol sold approximately 1 million tonnes to Eskom.

The Impact on Competition in the market for Thermal coal

Unilateral Effects

25. Unilateral effects occur when a merged entity has the ability to profitablyraise prices and restrict its output, without any co-operative action/reaction from its competitors. In other words, the merger leads to the creation or enhancement of market power for the merged entity.

26. The first step in assessing unilateral effects is the examination of pre- and post- merger market shares. These are often a prima facie indicator of likely unilateral effects.

27. In its report, the Commission provided a pre-merger and post merger picture of “..domestic, export and total sales and shares of thermal coal by South African coal producers, 2004”:

Pre- merger
Sales (Million tonnes)				Share of sales (%)
Producer	Domestic	Export	Total	Total
Anglo American Anglo Coal Kumba	53.44 34.79 18.65	19.88 18.78 1.10	73.32 53.57 19.75	38.27
BHP Billiton	35.00	22.14	57.14	29.82
Eyesizwe	41.15	2.50	43.64	22.78
Xstrata	2.85	10.92	13.77	7.18
Total Coal SA	0.58	3.81	4.39	2.29
Kangra Coal	0.95	2.00	2.95	1.53
Wakefield Investments	1.85	0.20	2.05	1.07
Graspan Colliery	2.00	-	2.00	1.04
Kayusa	1.30	-	1.30	0.67
Anker Holdings	1.00	0.20	1.20	0.62
Others	6.91	2.63	9.54	4.98
Total	128.38	63.18	191.56	100

28. In the table above, the Commission has combined Kumba’s sales with that of Anglo Coal as it argues that both are part of a single economic entity that is Anglo American.

Post merger
Sales (Million tonnes)				Share of sales (%)
Producer	Domestic	Export	Total	Total
Anglo American	34.79	18.78	53.57	27.96
BHP Billiton	35.00	22.14	57.14	29.82
Eyesizwe	41.15	2.50	43.64	22.78
Xstrata	2.85	10.92	13.77	7.18
Total Coal SA	0.58	3.81	4.39	2.29
Kangra Coal	0.95	2.00	2.95	1.53
Wakefield Investments	1.85	0.20	2.05	1.07
Graspan Colliery	2.00	-	2.00	1.04
Kayusa	1.30	-	1.30	0.67
Anker Holdings	1.00	0.20	1.20	0.62
Others	6.91	2.63	9.54	4.98
Total	128.38	63.18	191.56	100

29. In its post merger table the Commission excludes Kumba’s activities from the production capacity for Anglo American since it argues “Kumba is to be subsumed so as to form part of a single economic entity that is Eyesizwe.” Therefore the production capacity of Eyesizwe is inclusive of Kumba.

30. Although the Commission found that the affected market was highly concentrated,¹³ it nevertheless was of the view that the merged entity would not possess the capacity to exert market power as a result of the implementation of the merger transaction. This was based on inter alia the fact that its investigations revealed an abundance of opportunities regarding the acquisition of alternative supplies of coal.

31. The Commission’s table above was criticised by Dr Stillman, for the merging parties, as being incorrect in that it included domestic, export and total sales and shares of bituminous coal. This would include metallurgical coal as well. Although Dr Stillman concedes that the metallurgical coal data is “small relative to the totality”¹⁴ we agree that it is more appropriate to use the table contained in the CRA Report pertaining only to thermal coal.

32. The CRA thermal coal table of market shares was sourced from the 2005 Coal Statistics Manual. The table is based on total domestic sales of thermal coal for 2004 (i.e. production less exports) and excludes estimates of coal sales to Eskom and consumption by Sasol. The net result is an estimate of the supply of thermal coal available to small customers.

33. Producer (excl. Sasol)	Thermal coal sales (excl. Eskom and synthetic fuel sales (million tonnes)	% Share
    Kumba Coal	1.85	12 %
    Xstrata Coal SA	1.71	11 %
    Graspan Colliery	1.70	11 %
    Kuyasa Mining	1.20	8 %
    Kangra Coal	0.95	6 %
    Bisicht/Endulweni	0.90	6 %
    Anglo Coal	0.80	5 %
    Eyesizwe Coal	0.80	5 %
    Wakefield Investments	0.80	5 %
    Anker Holdings	0.60	4 %
    Ingwe Coal Corporation	0.45	3 %
    Sumo Collieries	0.45	3 %
    Small Junior Miners and Total Coal SA	3.40	22 %
    Total excluding Sasol	15.61	100 %

    It would appear from the table above, that the thermal coal industry is not highly concentrated. Post merger, Exxaro’s share of this market will be approximately 17%. There are several other players and two very large consumers, Eskom and Sasol. We agree with the merging parties, that the increment in market share does not confer market power on the merged entity and it is therefore unlikely that a combination of the coal assets of Eyesizwe and Kumba would have any material adverse effect on customers’ costs of coal supply.

34. We now turn to consider the area of contention between the Commission and the merging parties and the one that led to the condition the Commission recommends.

Co-ordinated Effects

35. The Commission’s case is that, while the merger will not lead to any anticompetitive effects as a result of unilateral conduct by the merged firm, it will make the market more conducive to what are termed ‘co-ordinated’ effects.

36. In merger control the term ‘co-ordinated effects’ is used in contradistinction to the term ‘unilateral effects’. In a unilateral effects case, as we noted earlier, we analyse whether the merger gives the merging parties the ability to substantially prevent or lessen competition as a result of the elimination of competition between them. In a co-ordinated effects case we look at how the merger will affect the behaviour of rival firms in the market. The competition concern is that although the merger will not lead to conditions in the market that will give the firm significant unilateral market power, it will generate new industry conditions that will enhance the scope for collusion. This collusion, be it explicit or tacit, could lead to an anticompetitive outcome. ¹⁵

37. According to international practice, a merger may give rise to co-ordinated effects concerns in two instances.¹⁶ In the first instance, it can strengthen an existing co-ordination. In this instance there would need to be evidence of an existing co-ordination, and secondly, that the merger is likely to strengthen that co-ordination. The second instance is that the merger increases the likelihood that firms will co-ordinate. Here there may be no evidence of an existing co-ordination, but evidence that post merger, it will be probable.

38. This is a useful way of approaching the analysis and it seems perfectly compatible with our legislation. Section 12 A (2) of the Act reads:

“ When determining whether or not a merger is likely to substantially prevent or lessen competition, the [competition authority] must assess … the probability that the firms in the market after the merger will behave competitively or co-operatively, taking into account any factor that is relevant to competition in the market including…..the history of collusion in the market.” (Our emphasis and our edits)

39. The tests we have referred to above are a conclusion we make about the effects of a merger after considering all the evidence. However, the tests do not answer the question of what the conditions are for successful co-ordination to take place. Economists, at least seem to agree that the following conditions are a prerequisite for co-ordination to be possible. Would be participants to a co-ordinated strategy must:

1. Be able to reach agreement;

2. Be able to monitor whether the agreement is being adhered to;

3. Be able to punish deviation so as to make it costly; and

4. Believe that co-ordination is feasible. Co-ordination will not be feasible if there are enough firms in the market who are not part of the co-ordination, or if enough firms can enter the market to make it unprofitable for the firms contemplating co-ordination.¹⁷

40. We have not cited these conditions to suggest they should constitute the test for the probability of co-ordination in future merger cases. It is not necessary for the purpose of this decision to be that categorical. Indeed, all those who have favoured some adherence to these conditions, have been anxious to explain that they are not to be rigidly applied. The ICN stresses that the conditions are a starting point and should not be applied as a checklist. ¹⁸ In a recent commentary on its horizontal merger guidelines, the US agencies stress that co-ordination need not be perfect and that to the contrary the agencies will assess whether co-ordinated conduct will be sufficiently successful following the merger to result in anticompetitive effects.¹⁹

41. More recently, in the Sony/ Bertelsman decision,²⁰ where the European Court of First Instance had a chance to comment on the subsequent application of its requirements for successful co-ordination laid down in Airtours it cautioned against a dogmatic application of the test when it held that the issue is:

“ .. the assessment of the risk that a concentration would create a collective dominant position and not, as in the context of the first part of the present plea, of the determination of the existence of a dominant position.” (Paragraph 249) It follows that in the context of a determination of a dominant position although the three conditions defined by the Court of First Instance in Airtours …which were inferred from a theoretical analysis of the concept of a collective dominant position, are indeed also necessary, they may, however, in the appropriate circumstances, be established indirectly on the basis of what may be a very mixed series of indicia and items of evidence relating to the signs, manifestation and phenomena inherent in the presence of a collective dominant position.” (Paragraph 251)

42. These prerequisites are useful therefore, not as a basis for determining what our own legal position on these issues should be, but to help as a method of analysing the theory of harm advanced in this case.

43. The challenge to would be co-ordinators to find mechanisms that serve to facilitate the reaching of agreement, monitoring compliance and punishing those who cheat, is what economists sometimes talk of the as the ‘cartel problem’. In order to resolve the cartel problem, firms may utilise a variety of mechanisms. In this specific case, we are asked to see whether interlocking directorships, between competing firms, can resolve, at least some, of the cartel problem.

44. There is authority for this.

45. Cristina Caffara, an economist with CRA International, has suggested that links between firms can assist information exchanges needed for agreement, and additionally provide an opportunity for monitoring adherence, because of the speed and accuracy with which cheating can be detected.

“Having a minority share in B might provide A with information on B’s plans, costs etc. which it would otherwise not have. Clearly this will depend on issues such as board representation.” ²¹

46. The Organisation for Economic Co-operation and Development (OECD) in a paper on oligopoly notes:

“There are a host of ways falling short of actual ownership links which leading firms employ to make themselves more similar and transparent to rivals and simultaneously credibly commit themselves to a more co-operative relationship with them. They include: cross directorships, [further practices are then cited]….. To a greater or lesser degree all nine of the above listed practices could be justified as innocent means of improving efficiency and potentially benefiting consumers. Nevertheless, because of their effects on transparency and enhancing the ability to credibly commit to enhancing co-operation, they also raise the probability of co-ordinated interaction. “ ²²

47. According to the European Union’s horizontal merger guidelines:

“Co-ordinating firms may, however, find other ways to overcome problems stemming from complex economic environments short of market division. Publicly available key information, exchange of information through trade associations, or information received through cross shareholdings or participation in joint ventures may also help firms reach terms of co-ordination…. Structural links such as cross shareholding or participation in joint ventures may also help in aligning incentives among the coordinating firms…. Cross directorships, participation in joint ventures and similar arrangements may also make monitoring easier.” (Our emphasis) ²³

48. From this literature, it would seem that cross-directorships provide at least two solutions to the cartel problem. Firstly, they provide a forum for the exchange of information in a setting conducive to an innocuous explanation. Secondly, they provide a highly efficient and expeditious mechanism for monitoring compliance with the terms of the co-ordination.

49. We now turn to the facts of this case in the light of this theoretical background.

50. The Commission argued that because Anglo American was permitted to appoint one director on to the Exxaro board and one on to the Eyesizwe SPV board, this would provide an opportunity for an exchange of commercially sensitive information that would facilitate co-ordinated conduct in the thermal coal market in South Africa. This is because Anglo American owns 100% of Anglo Coal a competitor of Exxaro.

51. The Commission therefore recommends the imposition of the condition to prohibit Anglo from being able to propose a director at either Exxaro or Eyesizwe SPV level, in order to close off this conduit of possible co-ordination. This, it argues, is a proportionate remedy because it eliminates some of the competition concerns, whilst preserving the empowerment objectives of the merging parties.

52. The merging parties vigorously opposed the imposition of the condition. They contend that the merger would not give rise to a concern about co-ordinated effects and they led expert testimony in support of this contention. As an alternative argument, they contended that the merger would not make the possibility of co-ordinated effects any more likely than it was pre-merger, given that cross directorships between Eyesizwe, as it was then, and Anglo and Ingwe exist. The merging parties led evidence that the cross directorships were innocuous and were justified on commercial grounds, related to the empowerment ambitions of the deal.

53. One of the difficulties for the Commission in this case was, to express it colloquially, whether the merger had made a bad situation any worse. The Commission has not been that confident on this point. This emerges in its recommendation where it is stated:

“The Commission assumes this view on the basis that were there to have been a likelihood that Anglo American, Eyesizwe Coal and BHP Billiton coordinated their conduct on the affected market prior to the proposed implementation of this merger transaction as a result of the prevalence of cross-directorships amongst them, the likelihood of such coordinated market conduct being sustained post merger is a reasonable likelihood.”²⁴

54. Unsurprisingly, given that this was its initial stance, the Commission’s legal argument was that even if a merger did not of itself lead to a substantial prevention and lessening of competition, if the merger perpetuated or sustained an anticompetitive structure, this was sufficient to justify the imposition of remedial conditions.

55. This has not been the manner in which we have interpreted the Act thus far. Granted, we have not yet been called upon to decide the matter definitively but we do not find that this occasion justifies a departure from that interpretive approach.

56. The test of harm to be applied to this case is the first one we referred to earlier - does the merger strengthen an existing co-ordination? (It does not seem on the facts that the second instance is of application here.) Because we have found that the present merger does not do so, and indeed appears to be weakening an existing co-ordination, assuming, it to be in existence, it is not necessary for us to decide the Commission’s point of law.

57. We now go on to explain why we have come to this conclusion.

Prior to the merger –

58. Anglo American directly or indirectly –

1. Held 11% of Eyesizwe;

2. Had the right to appoint one (1) director to the board of Eyesizwe;

3. Held 66% of Kumba;

4. Had the right to appoint five (5) directors to the board of Kumba;

59. Kumba was a company owning iron ore, coal and other mineral assets (base metals and heavy minerals). Eyesizwe was only a coal company. The Eyesizwe board provided a forum for Anglo, Ingwe and Exxaro directors to meet.²⁵

60. BHP Billiton, via its coal subsidiary Ingwe, directly or indirectly held –

1. Held 9 % of Eyesizwe; and

2. Had the right to appoint one (1) director to the board of Eyesizwe.

Post merger -

61. Anglo American directly or indirectly –

1. Holds 20.3% of Exxaro (17% directly and 3.3% indirectly);

2. Appoints one (1) director to the board of Exxaro;

3. Appoints one (1) director to the board of Eyesizwe SPV;

4. Controls Kumba iron ore and indirectly the operating company SIOC; and

5. No longer controls the remaining Kumba businesses now part of Exxaro.

62. Exxaro is a mineral company of which coal is only one business. The Exxaro board provides a forum for Anglo and Exxaro but not Ingwe directors to meet.

63. Ingwe –

1. Holds 9 % of Eyesizwe SPV which is an indirect interest of 2.75% in Exxaro;

2. Appoints one (1) director to the board of Eyesizwe SPV; and

3. Is not represented on the Exxaro board.

64. The essential differences then are that Anglo has a reduced holding in its erstwhile Kumba coal assets. It is reduced to having only one director at Exxaro level and another at the SPV level. Ingwe, once represented at operational company level, is now relegated to shareholder status at Eyesizwe SPV level. The Anglo and Ingwe appointees still meet at a board, but this is now not at operational level, but in an investment company two steps removed from the operational company. It is not clear, but unlikely, that Exxaro coal executives would be represented at this shareholder level.

65. Exxaro is a very different company to Eyesizwe. Whereas Eyesizwe was a dedicated coal company, Exxaro is a mineral company, with a coal division, but coal is by no means its most important asset. Indeed it would appear that coal is now a small part of its business. Of an enterprise value of R 24 billion the coal assets represent R 1, 6 billion. The most valuable asset of Exxaro is its 20% stake in iron ore company SIOC (R3, 8 billion).

66. What are the implications of this for a post merger strategy of co-ordination? Assume for the moment in the Commission’s favour, that pre-merger, the board arrangement at Eyesizwe provided a legitimate meeting place for Anglo, Ingwe and Eyesizwe to meet, and hence, was a potential forum for co-ordination in some form.

67. Leadership in such a forum is likely to have come from Anglo and Ingwe since Eyesizwe was their creation, formed out of these firms’ coal assets. However, and this is the most significant difference – Ingwe is no longer a party to this forum as it is now only represented at two removes from the Board and will according to the testimony, receive only shareholder information. According to the testimony of Mr Baum, Ingwe had wanted board representation, but that Anglo had resisted it.²⁶ This is probably the most significant fact in the merger. Co-ordination if it is to be successful should at the very least involve Ingwe. Ingwe once part of a forum that could receive information and monitor performance is now no longer party to this information except to the extent that shareholders are given information on the performance of Exxaro as a whole.

68. In the second place, Anglo’s economic interest changes. It is less invested in coal than it was pre-merger with the sale of Kumba coal to Exxaro.

69. Thirdly, Anglo’s economic interest in Exxaro is now changed to an investment in a broad based mineral company and not a dedicated coal company. The type of executive who will be appointed to the Exxaro board will need to be a generalist not a coal industry insider. The fact that Anglos’ nominee to the board of Exxaro is not to be Mr Shout, its existing nominee to Eyesizwe, who is also on the Kumba board, but Mr Baum is an indication of this. While this creates a structural difference that is more nuanced than its is definitive, it does predispose this to be a board where information flows are qualitatively less detailed, immediate and transparent than on a company whose sole business is coal.

70. The reason why we have placed such emphasis on the structural differences pre and post merger is that we are not in a position to accept the merging parties’ argument that there is no evidence of any existing co-ordination. The burden of the parties evidence in this regard was that co-ordination is highly unlikely, as the international market is not susceptible to successful co-ordination by South African companies, their share of the market is too small, and the bulk of the domestic market for thermal coal (87%) is taken up by long term contracts with Eskom from its so called tied mines, where pricing is regulated. ²⁷ The balance of the domestic market is small, according to the parties, and there are a large number of players who compete for opportunities here. In this residual domestic market, Stillman argued Anglo, Exxaro and Kumba would have only 23% of the market. If Ingwe is added this would constitute 26% of the residual domestic market.²⁸ This analysis may well be correct, but there were a number of features of the evidence, which were curious, and we put it no higher than that, which the merging parties did not satisfactorily deal with.

71. Minutes of board meetings evidence the level of information that comes before them, and includes discussion of supply availability, prices, how firms had performed against budget predictions, future strategy etc. ²⁹ This is information normally useful to rivals in a commodity business. The merging parties did little to provide comfort that this was not the case beyond the evidence of Dr Stillman who could do no more than provide the theoretical model for why it should not happen. It would have been useful to allay concerns if the merging parties had led one of the directors who sat on both the Eyesizwe and Anglo boards, or Ingwe. It would also have been useful to hear from someone who could explain remarks in the business plans that were, at the very best for the merging parties, ambiguous on the possibility of existing co-ordination.³⁰ It would have been useful to hear from a witness who attended the Eyesizwe meeting, where its chief executive Mr Nkosi in the presence of the Anglo and Ingwe nominees, lamented the fact that there was an oversupply of A grade coal, whilst discussing the marketing report serving before the board.³¹ Whilst Dr Stillman made a valiant attempt to interpret them favourably to the merging parties, they were not his documents and he could provide no more light on them than we could ourselves.

72. This is also an industry where there is a high degree of co-operation between rival firms. The documents are filled with references to joint ventures that exist or are proposed by all the major players. Various players in the industry are also shareholders in the Richards Bay Coal terminal (“RBCT”), access to which is crucial for exports. Since the RBCT is a scarce resource the manner in which the parties allocate it, could serve a variety of functions in facilitating co-ordination.³² Whilst joint ventures are not unusual in mining for various logistical reasons, one should not be complacent about them either from a competition perspective. As Posner J a leading United States judge remarked in Hospital Corporation of America v. FTC, “a market in which competitors are unusually disposed to co-operate is a market prone to collusion”. ³³

73. It is for this reason that we have assumed that co-ordination exists, and we have asked whether the merger enhances the possibility of co-ordination i.e. scenario one that we referred to earlier in our discussion of the theoretical issues. Since we answer that question in the negative, it is not necessary for us to examine the prior assumption in any further detail. If the answer to that question were in the affirmative, we would have called for more evidence on the question of whether there was a pre-existing co-ordination and this would unavoidably have prolonged the hearing.

74. Thus to the extent that co-ordination may have occurred or been possible pre-merger, it is less likely post merger. If co-ordination is taking place in this industry, and we reiterate that we are in no position to comment on whether it is, then the structure resulting from the merger does little to further facilitate it, and indeed, in the respects that we have identified, inhibits rather than promotes it.

75. We have also examined in detail through the discovered record and the evidence of Mr Baum, the rationale for the merger. We have previously cautioned against placing undue weight on the rationale for the merger:

“In the ordinary course, merger analysis does not draw heavily on the parties’ stated rationale for the merger. This usually amounts to little more than a statement of intent and is generally expressed in anodyne terms that do little to advance understanding of the competition implications of a merger transaction. In this instance, however, it is instructive to juxtapose the stated rationale with the record”.³⁴

76. A similar sentiment is expressed by Areeda and Hovenkamp:

“Merger inquiry is objective it rarely or never considers the merging firm’s manifested subjective intentions. Firms are assumed to be profit maximizers within whatever environment they find themselves.” ³⁵

77. In this case, however we have resorted to the rationale in order to test the veracity of the parties’ contentions that the interlocking directorships are not to promote an anticompetitive purpose. We are satisfied that in this respect Anglo’s rationale for wanting representation on the board of a company that has some interests rival to its own, is driven by a series of considerations that can be justified on grounds that have nothing to do with an attempt to co-ordinate the respective firms behaviour. These have been fully captured earlier in our section on the rationale for the merger and do not need to be repeated here.

Conclusion on co-ordinated effects

78. Given this finding it is not necessary for us to decide the point of law raised by the Commission. We have decided that even assuming the existing interlocking directorships between Anglo, Ingwe and Eyesizwe have created a mechanism for co-ordination, the merger does not meet the test required of strengthening the existing co-ordination. To the contrary, the merger inhibits this possibility, because it complicates the possibilities for the exchange of information and monitoring, and it changes the incentives of all the firms who may have been party to any pre-existing co-ordination.

Undertakings

79. This merger might have been cleared far more expeditiously if the merging parties had shown the same pragmatism during their first interactions with the Commission that Mr Baum demonstrated in the course of his testimony during the hearing. The panel, endeavouring to see if a via media could be found that would meet the Commission’s competition concerns in a manner that would still allow Anglo representation on the Exxaro board, explored various possibilities for a condition to be imposed on the merger by way of an undertaking from Anglo. Mr Baum, to his credit, showed a willingness to do so. We invited the merging parties and the Commission at the end of the hearing to see if they could come up with an agreed condition.

80. Regrettably, that did not happen and therefore we have had to decide the matter in the absence of an agreed undertaking, which on the facts of this case, ought not to have been difficult to reach. Whilst merging parties are of course under no obligation to give undertakings that they believe are not warranted, they cannot complain if this leads to hearings being prolonged.

81. The Commission is rightly concerned about the competition problems posed by interlocking directorships between rival firms. In the United States, this is illegal per se.³⁶ The courts there have explained the reason for this measure was,

“to nip in the bud incipient violations of the antitrust laws by removing the opportunity or temptation to such violations through interlocking directorates” ³⁷

82. The Commission’s job is to do precisely this. While ultimately we have not found in its favour we are satisfied that the issue absent a satisfactory undertaking from the merging parties, justified proper scrutiny.

PUBLIC INTEREST

83. During the hearing, Mr Reint Dykema from Solidarity Union and Mr Jeffrey Magida from NUM made submissions relating to inter alia the participation of employees in the Employee share option schemes. The merging parties however, confirmed that negotiations regarding ESOP’s in both SIOC and Exxaro were ongoing and that undertakings in respect of these ESOP’s were still being respected.

CONCLUSION

84. Based on the assessment above, we find that the merger is unlikely to substantially prevent or lessen competition in the relevant market. There are no significant public interest issues and we accordingly approve the transaction without conditions.

____________________

N Manoim

Y Carrim and U Bhoola concurring

Tribunal Researchers: M Murugan-Modise and R Kariga

For the merging parties: Adv. D Unterhalter SC and Adv. J Wilson instructed by Webber Wentzel Bowens

For the Commission: Adv. J Gauntlett SC and Adv. H Shozi instructed by the

Competition Commission (TM Kekana - Mergers & Acquisitions)

---