COMPETITION TRIBUNAL OF SOUTH AFRICA
Case No: 82/LM/Oct06
In the matter between:
Lexshell 668 Investments (Pty) Ltd Acquiring Firm
And
Wakefield Investments (Pty) Ltd Target Firm
Panel : N Manoim (Presiding Member), M Holden (Tribunal
Member) and M Madlanga (Tribunal Member)
Heard on : 18 April 2007 and 25 June 2007
Order issued on : 27 June 2007
Reasons issued on : 1 August 2007
Reasons for Decision [NonConfidential Version]
Introduction
1] In this merger between two coal firms it is common cause that in the near future,
the target firm will be able to charge more for its coal than it does presently. The
question for us to determine is whether this pricing opportunity comes about as a
result of the merger giving it the market power to do so, as some objectors to this
merger contend, or a change in supply conditions in the coal industry, as the
merging parties contend.
Transaction
2] The acquiring firm, Lexshell 668 Investments (Pty) Ltd (‘Lexshell’) is purchasing
the entire equity in the target firm, Wakefield Investments (Pty) Ltd. 1 Both the
1 Although Lexshell has been renamed Shanduka – we will to refer to it as Lexshell to avoid
1
acquiring firm and the target firm are engaged in the coal mining industry.
3] Wakefield is controlled by the diversified mining group, Metorex Limited, which
holds 70% of the equity, the balance being held by its empowerment partner,
Umnotho Wesizwe. Wakefield’s owns four collieries – Leeuwfontein, Banfontein,
Lakeside and Middelburg Townlands, all situated in the Kendall district in
Mpumulanga.2
4] Lexshell is owned (70%) by Glencore Investments BV a wholly owned subsidiary
of Glencore International AG (Glencore), and Shanduka Resources (Pty) Ltd
(30%). Glencore is an international commodity trading firm incorporated in
Switzerland. Shanduka Resources is a BEE owned mining company with at
present a stake of 15% percent in Kangra Mining, also a coal mining firm, and a
30% stake in Lexshell. 3 Glencore owns many assets worldwide, but of relevance
to this transaction is its stake in the diversified mining group Xstrata plc (presently
[ ]%). 4 Xstrata, inter alia, owns coal mining assets in South Africa, as well as being
a 21 % shareholder in the Richards Bay Coal terminal. 5 Glencore accepts that for
the purpose of this transaction it can be considered to control Xstrata. This means
when we consider the increase in concentration brought about by the merger we
add Wakefield’s market share to the market share of Lexshell and Xstrata. Despite
this concession, we are informed there are no management overlaps in respect of
Lexshell and Xstrata’s interests locally, and both account to Glencore separately.
5] Lexshell itself presently owns one coal mine, Graspan which it acquired in 2006.
Graspan is located in the Middelburg area contiguous to one of the Wakefield
collieries, Middelburg Townlands. 6 We consider below the implications of the
merger on the neighbouring mines. All the mines are situated in the coal rich area
merger on the neighbouring mines. All the mines are situated in the coal rich area
of Witbank/ Middelburg where 49% of South Africa’s coal is produced.
confusion with its shareholder which is Shanduka Resources (Pty) Ltd.
2 See Record file 1, pages 495498, extract from SA Coal Statistics report dated 2006.
3 Shanduka previously held 40% of Kangra and 60% by Graham Beck but it in 2007 both diluted
their respective stakes to15% each, selling the balance to a Spanish firm, Fenosa).
4 There is a rather complicated explanation concerning the evolution of Glencores’ stake in Xstrata
which in view of the concession that it can be considered as controlling Xstrata Plc we need not
consider further.
5 See Commission recommendation page 15.
6 See 25 June transcript, page 92.
2
Rationale
6] Lexshell is in the process of building a coal mining group. It recognises that there
are assets in the market that are trading below their potential and that by acquiring
them, and turning them around, it can increase the value of the assets. Wakefield
has the largest of its four collieries, Middelburg Townlands Colliery, situated next
door to Graspan, so for Lexshell it was a logical asset to buy. Lexshell was also
looking for mines that did not have an export allocation, because if they got an
allocation this would enhance the value of the assets. 7
7] Given the trading experience of its controlling shareholder, Glencore, and the fact
that the South African coal market is rapidly becoming more of an export market, it
is better placed than many to achieve its ambitions. Metorex was not a pressured
seller. The original offer from Shanduka was declined, but the present offer seems
to have made it worth its while to sell and to use the proceeds in investing further
in its other mining assets. 8
Coal Market
8] There are various types of coal whose utility is determined by the extent of their
carbon content. In a recent decision we went into the technical aspects of these
differences in more detail than we will in this decision. 9 For our purposes we are
only concerned with bituminous coal because this is the type of coal mined by the
7 By export allocation we mean rights to export through the Richards Bay Coal Terminal which we
explain more fully below.
8 The first indication that Metorex had received an offer from Shanduka appears in June 2005. See
Metorex discovery page 274. A second offer was made in about May 2006. See Metorex discovery
288
9 See Main Street 333 (Pty) Ltd And Kumba Resources Limited. Case No: 14/LM/Feb06
paragraphs 1821. “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, subbituminous, bituminous and anthracite. Lower rank coals (lignite and subbituminous
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. Bituminous and Anthracite are the two types of coal mined in South Africa.
3
merging parties. There are two types of bituminous coal – thermal (steam) and
metallurgical (coking coal). The merging parties’ mines produce both types of coal.
Thermal and metallurgical coal are not considered substitutes because they have
different uses. As we observed in Kumba:
a. Thermal coal is used in power generation and also has certain industrial
uses
b. 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.” 10
9] For this reason we will examine the impact of the merger on these products
separately. This has also been the approach of the Commission and the merging
parties.
Metallurgical coal
10] Compared to thermal coal, metallurgical coal occupies a modest position in local
coal production. 11 This does not obviate the need to evaluate the merger’s effect
on that product market. Presently, Wakefield sells all its production of metallurgical
coal to Xstrata. Accordingly, the merging parties argue the merger will have no
impact on competition in the market in respect of this product. Recall that as
Xstrata and Lexshell share a common controller in the form of Glencore, we treat
them as part of the same concentration. This means that even if post merger
Lexshell continues to supply Wakefield’s coal only to Xstrata, third parties will be
no worse off in obtaining supply, than they were before the merger. But even if we
do not treat Wakefield as already self supplying the merged entity, the merger
does not lead to a significant accretion in the metallurgical coal market as the
figures below suggest:
10 See Kumba paragraph 21. In that case according to the merging parties’ economist’s report the
average price for thermal coal was less than 25% of that of metallurgical coal. See the CRA
average price for thermal coal was less than 25% of that of metallurgical coal. See the CRA
International report entitled Project Pangolin: Competition Analysis, dated 6 February 2006, page
22.
11 According to SA Coal Statistic figures only 2,8 million tons of metallurgical coal were sold in the
domestic market in 2005. By comparison 243 tons of thermal coal was sold.
4
Table1
Producer Local Metallurgical Coal
Sales (Low Phos)
Market Share (%)
Anglo 640,000 13
Graspan 180,000 4
Kangra 0 0
Xstrata 0 0
Wakefield 270,000 6
Total Merged Entity 450,000 10
Anker 360,000 8
Endulwini 180,000 4
Eurocoal 100,000 2
Ingwe 720,000 15
JenshaEastside 180,000 4
Kumba 1,500,000 32
Total Coal 550,000 12
Woestalleen 70,000 <1
Total Coal 4,750,000 100
(Note that the premerger HHI was 1682 and the post merger HHI is 1730.)
11] We therefore conclude that the merger will not lead to a substantial lessening or
prevention of competition in respect of metallurgical coal, and no more on the
subject need be considered in this decision
Thermal coal
12] The effect of the merger on the thermal coal market requires more
consideration. Again, market definition has not proved controversial. Both the
merging parties and the Commission are in agreement that:
• The relevant product is thermal coal, of all grades, in sizes of 0 35mm;
5
and
• The geographic market is the residual domestic market.
13] This, as the language suggests, is a rather nuanced market definition and so
we need to explain how we have arrived at it.
14] In the industry thermal coal is subject to classification based on the size of the
coals and their calorific qualities. Coals are classified in size groupings ranging
from just above zero to 35mm. These groupings have different names, ranging
from the smallest known as duff, through to peas, and finally small and large
nuts. Until fairly recently the price quoted for coal differed depending on its size.
Recently, the difference in size has ceased to matter in the export market, and
thermal coal is treated as one commodity regardless of the size of the individual
coals, as we consider later in this decision. 12
15] The grade of coal matters. Only Grades A and B are considered exportable.
Grades also have a bearing on the use of the coal in the domestic market.
However the value of even a lower grade coal can be enhanced to be of export
quality, through a process known as washing. The process improves the yield
of the coal by removing the ash content. Washing however increases the
expense of producing coal and thus the decision whether or not to wash a lower
grade coal or to sell it as a lower grade coal at a lower price depends ultimately
on the price that can be realised for the washed coal. If that price is sufficiently
high, mines will have the incentive to wash coal and to sell it to customers who
need the higher grade coal. The technical requirements of a customer’s
production process will inform the grade of coal they can utilise. Not all
customers need higher grade coal. Eskom, to a large extent, is able to utilise a
lower grade of coal and accordingly pays much lower prices for it.
16] Other domestic customers, the cement industry being one of them, require a
16] Other domestic customers, the cement industry being one of them, require a
higher grade of coal. One might think that for this reason the customer who
12 For instance earlier price lists from Graspan to its clients quote different prices for grains and
peas (August 2005). In September 2006 the price list gives them a uniform price. (Compare page
27 with page 29, Commission additional file.)
6
buys export grade coal domestically would pay an export parity price. 13 If they
did not, the coal company would make more by exporting the coal. However
this has not been the case in South Africa until very recently. We consider
below why that has been the case.
17] In 2005, the last year for which we have figures available, South Africa
produced 243 million tons of thermal coal. Of this figure –
• 72 million tons (approximately 30%) were exported, with most of the
exports (96%) going through the Richards Bay Coal terminal (‘RBCT’). 14 Of
this exported amount, about 87% was exported to the European Union or
what is referred to as the Atlantic Basin market. In this market South Africa
is largest supplier with about 38 % of the market.
• 171 million tons were consumed locally. Out of this local consumption 107
million was consumed for power generation, largely by Eskom (106 million)
and the balance by municipalities (1million). The next major local consumer
is Sasol which in 2005 consumed 41, 5 million tons. The steel industry is
also a large consumer of steam coal (2,7 mt)
18] This leaves potentially 18 million tons of thermal coal for consumption by
other customers in the domestic market. This is what we mean by the term
the ‘residual domestic market’ which we used above. Customers in the
domestic market by and large do not have tied sources of supply nor are
they party to long term supply contracts. As a result they pay prices in
excess of what customers such as Eskom pay, by more than 20%.
19] Although the Commission and merging parties are agreed that for the
purpose of analysing the effects of the merger the relevant market to be
considered is this domestic residual thermal coal market, they are not in
agreement on what the size of this market is and hence their respective
agreement on what the size of this market is and hence their respective
conclusions differ on what the concentration in the market is, post merger.
According to the merging parties the total size of this market is 14,9million
13 Strictly speaking this is a net export parity price as it excludes the cost of transporting the cost
from the mine to the port and the related costs associated with this.
14 The balance went through Maputo (1, 1 million) and Durban (1, 1 million).
7
tons and of this the merging parties account for the following:
Xstrata 10,7 %
Graspan 8,0%
Wakefield 11,6 %
Total merged entity 30,3% 15
20] The Commission which interrogated the figures supplied for the production
of the various mines concluded that certain mines should be excluded from
the amount available to the residual domestic market, for reasons that need
not concern us now, and came to the conclusion that the size of this market
was 13, 5 million tons. On this basis the Commission recalculated market
shares and came to the following conclusion:
Xstrata 11,8%
Graspan 8,8 %
Wakefield 12,7 %
Total merged entity 36,3% 16
21] As nothing turns on the difference between these two calculations we don’t
need to decide which of these calculations is the more reliable, so we will
assume, for the purpose of our analysis that the Commission’s calculation is
the correct one, and hence, post merger, the merged firm will have 36,3% of
the relevant market. 17
22] Coal that is exported from South Africa goes to two primary offshore markets.
The first is known as the Atlantic market where South Africa is the largest
producer with about 38% of this market over the past three years for which we
have figures. 18
15 See Commission recommendation pages 212
16 See Commission recommendation page 26.
17 See Commission recommendation page 25 for its reasons. Some of the reasons include that
firms are not suppliers of thermal coal but anthracite or are committed to contracts with customers
such as municipalities.
18 See Xstrata figures for years 20042006, Lexshell discovery page 422.
8
23] There is also the Pacific Rim market where some South African coal is
exported, but unlike in the Atlantic, our geography counts against us and we
are a minor player here. According to SA Coal Statistics we are the third
largest seaborne coal supplier but lag at some distance behind Indonesia and
Australia.19 In addition, coal companies are trying to enter new markets such as
China and India, but at the moment they are not yet of significance, because
they either have sufficient domestic supplies or have more closely situated
sources. Future demand may require them to source further afield, including
considering South Africa.
24] Traders quote prices for the different export markets based on a weekly fixing.
The material is thus replete with a Richards Bay price which represents the
price in dollars for sale to the export Atlantic market.
25] Historically, the domestic residual price has been well below the export parity
price and hence, in past decisions, we have defined the export market as a
separate one from the domestic one. 20 Ordinarily if the export price was higher
than the domestic one, net of transport costs, one would expect firms to export
rather than sell to the domestic market. This has not happened primarily
because of constraints in exporting coal from South Africa. The optimal route for
exporting large amounts of coal from South Africa is through the RBCT. Whilst
Durban and Maputo are alternatives for coal producers who are predominantly
in the Witbank/Middelburg area, the transport costs are prohibitive and in
practice very little coal has been exported through these ports.
26] The South African coal industry’s ability to export is constrained by the extent
of loading capacity available at Richards Bay. It is further constrained by the
ability of Spoornet to transport coal from the inland to Richards Bay. To the
ability of Spoornet to transport coal from the inland to Richards Bay. To the
extent that the coal industry produces coal in excess of these constraints, that
surplus coal will have to be sold in the domestic market.
27] RBCT is a privately owned terminal. RBCT presently has a nameplate capacity
19 See Record page 393, SA Coal statistics 2006.
20 See for example Kumba
9
of 72 million tons. 21 Its shareholders are the major coal producers who by
arrangement between themselves have allocated the capacity into quotas
which the respective firms are entitled to. The quotas vary in size and depend
on the firm’s contribution to the project. Recently the shareholders have,
following criticism that the terminal is club for the “haves”, released 4 million
tons of capacity and made them available to firms with an empowerment
profile. This initiative, known as Project Quattro, was phased in over a few
years and it is only now the other firms have taken up near to all this capacity.
It has however created a new club. Although the quotas are up for allocation
annually according to Metorex, they followed the lines of the previous
allocation.
28] For this reason RBCT has ambitious plans to expand by another 19 mt to 91 mt
by mid 2009. 22 This means far more capacity will be available to other players in
the industry and hence the clubby nature of RBCT is likely to be a thing of the
past.
29] However it is not only the capacity of RBCT that is a bottleneck. In order for
coal to reach RBCT it must travel by rail. This in turn is dependant on the sole
suppler of rail capacity, Spoornet, to make and maintain capacity to meet the
demand. The evidence thus far is that Spoornet has not fully met these needs,
although it is doing a lot to improve the position. Whilst the privately owned
RBCT is confident it will reach its target capacity of 91 mt by 2009, the publicly
owned Spoornet is less sanguine in its expectations. 23 Unless Spoornet makes
the equivalent rail capacity available, RBCT can expand all it likes, but the coal
won’t reach it. Mr Van der Merwe, who is the man responsible for coal transport
in Spoornet, testified that Spoornet will only have this capacity by 2011 thus two
in Spoornet, testified that Spoornet will only have this capacity by 2011 thus two
years later than the RBCT projections, unless Spoornet get contracts from coal
producers before then.
30] The evidence about RBCT’s expansion was given by its chief Executive officer
Mr Kusani Dlamini. Mr Dlamini exuded confidence and optimism, inasmuch as
21 See transcript dated 18 April 2007, page 20.
22 See transcript dated 18 April 2007, page 7.
23 See transcript dated 18 April 2007, page 39.
10
Mr Van der Merwe from Spoornet was cautious and sceptical. What do we
make of this contrasting outlook on the prospects for coal exports from two
people equally wellplaced to make an informed assessment? It is certainly
correct that RBCT has been the subject of much hype in the past. Press
cuttings in the record reveal how the gap between dream and reality has
dogged this project previously 24. On this basis prospects for expansion may
need to be much more conservative and hence RBCT may have a bottleneck
for years to come. 25The implications of the bottleneck for the domestic market
are significant. If the bottleneck remains, then the domestic coal price will
remain below export parity as it has in the past. The coal surplus would have to
be absorbed in the local market and prices would reflect this oversupply be
tending downward from export parity. In 2004 the average residual domestic
price was 38% less than the net export parity price according to an economist
report that was filed in the Kumba matter. 26
31] If the merger would enable the new entity to control supply to a degree in the
domestic market then the merger might have an effect on competition. Although
the merged entity will not be considered a major South African coal producer –
it lags considerably behind the majors such as Anglo Coal and Ingwe –
nevertheless once we consider the residual domestic market, the merged entity
is a significant player with 36% of that market. Wakefield has to date supplied
largely to the residual domestic market. One informant told the Commission that
24 See for instance an article entited RBCT expansion expected soon dated 5 May 2005 record,
file 1 page 603, where the following appears, “The expansion to 82 million tons a year has been
delayed numerous times since the announcement of the project in 2001 mainly due to differences
between the various stakeholders.”
25 It is a curious feature of RBCT, which is privately owned by its major shareholders that it should
have any incentive to expand supply and offer this expanded capacity to rivals, when a supply
increase from RBCT may well impact on prices in the Atlantic Bay market. Just how sensitive the
Atlantic market is to the most minor mishap at RBCT, or on the way to it, is clear from the Glencore
material where incidents such as derailments are reported on weekly. Perhaps the motive for
expansion is politically rather than commercially driven as emerges albeit elliptically from a speech
by the Minister of Mineral and Energy where she remarks in a speech at a sod turning ceremony at
Richards Bay in September 2006, that “The Coal Industry Task Team was formed …with the aim
of sharing this coal terminal that at the time was for shareholders only. This happened in
anticipating the implementation of the Mineral and Petroleum Products Development Act(MPRDA)
as an increase for coal export infrastructure for the new BEE coal mines was anticipated as
compliance with the MPRDA made sure that there are new entrants who would be able to
participate in all levels in this industry.” See record file 1 page 635.
26 See CRA International report entitled Project Pangolin; Competition Analysis, dated 6 February
2006, page 23 footnote 49. The data relied upon by CRA is for 2004.
11
the merged entity will control virtually almost all of the coal used in the
manufacturing process in South Africa. 27 If supply constraints at RBCT do not
get resolved then the acquisition of Wakefield could lead to two possible effects
on the domestic market:
1. The merged firm could reduce supply to the residual domestic market by
decreasing production to increase prices; or
2. The merged firm could reduce supply to the domestic market by diverting
production to exports, because it was in a better position to overcome
supply constraints than Wakefield could on its own.
32] The merging parties’ response to this theory of harm is straightforward. They,
as we noted earlier, do not deny that prices may rise in the future, post merger,
but argue that this outcome is unrelated to the merger, but rather as a result of
structural changes going on in the coal industry. Their submission is that
whether Wakefield is run by a Metorex or a Lexshell, prices will be the same.
Since the merger was filed they argue the trend in prices to tend towards export
parity has become even more marked.
Table 2
27 See record file 1 page 678, submission by [ ].
12
33] In the table above, supplied by the merging parties, we see that this trend is
already taking place before the merger, and before the expansion of RBCT. The
figures are based on the merging parties own sales and not an industry
average, but nevertheless as prices in the market are relatively transparent, we
can assume that they are a reflection of price movements more broadly.
34] We can note from Table 2 that the most significant increase occurred in about
October 2006. Since RBCT has not increased its capacity in this period what
accounts for the steep rise in domestic prices? The merging parties’ explanation
is that until 2006 RBCT capacity was under utilised as we see from Table 3
below.
13
Table 3
Official budget Capacity Actual Capacity utilised
Share
holders
Project
Quattro
Total Spoornet
scheduled
Share
holders
Project
Quattro
Total
Under
utilization
against
budget
2002 68 0 68.00 67.23 66.05 0.00 66.05 1.95
2003 68 0.33 68.33 71.20 68.03 0.28 68.31 0.02
2004 69 1.35 70.35 72.42 64.83 1.11 65.94 4.41
2005 69 3.00 72.00 75.84 66.57 2.61 69.18 2.82
2006 69 3.78 72.78 72.84 62.87 3.61 66.48 6.30
35] The merging parties say that what has happened recently is that capacity is
being fully utilised, because there has been an increase in the export price of
coal. Firms with under utilised quotas are either selling quotas to firms without
or buying coal from them to fill their own quotas. 28 This has diverted capacity
from the domestic market and so led to a shortage which has resulted in
increased prices.
36] Before we consider the validity of this thesis let us consider the objections to the
merger from domestic customers
Eskom
37] Eskom is the largest domestic customer for the coal industry consuming close
on 120mt annually. When the Commission was investigating the merger,
Eskom filed a submission dated 29 March 2007 in which the following
statement was made:
“The proposed transaction should not have an impact on the mining industry
28 See transcript dated 18 April 2007, page 21. The questions are based on Table 7 of the
Commissions’ recommendation which shows an underutilization of capacity from 2002 to 2006 with
the greatest under utilization taking place in 2006, when the facility was under utilized by 6,3 million
tons. An exporter told the Commission that the unusual shortfall in 2006 was twothirds attributable
to Spoornet and onethird to loss of production, due to unexpectedly heavy rains. (See
recommendation page 17).
14
and supply of thermal coal in South Africa as a whole. However the proposed
transaction may have a more significant impact on supply in the Middelburg
coal region. Eskom’s understanding is that the new owners of [Middleburg]
Townlands Mine and the owners of Graspan Colliery, in the same area, are
one and the same. The need for coal in the Middelburg area as well as
Middelburg’s proximity to Arnot, Hendrina and Duvha power stations may lead
this group of collieries to utilise the opportunity to further inflate the market
price of coal in the region” 29
38] Eskom did not apply to intervene in the merger or to make further
representations. However, on the basis of the strong sentiment expressed in
the paragraph quoted above the Tribunal subpoenaed Eskom to testify. Mr Rob
Lines, the General Manager of Generation and Primary Energy at Eskom came
to testify. He was assisted by two of his colleagues from the legal department.
Mr Lines did not get to testify in April when the merger was first heard and
testified on its resumption on 25 June 2007.
39] On that day Mr Lines was to provide the hearing with a surprise, distancing
himself from his earlier written submission to the Commission. It appears that in
the time the hearing was adjourned Eskom representatives (but not Mr Lines)
met with representatives of the merging parties and discussed the transaction.
Mr Lines, having had time to reflect on the extent of the transaction with his
colleagues, who were part of the consultation, was now of the view that his
initial concerns had been based on incorrect information. He now considered
the transaction modest and of no concern to Eskom. Lines went on to state that
when they looked at their purchases from these mines they were not large and
that there were other mines in that area that they could contract with. 30 He also
that there were other mines in that area that they could contract with. 30 He also
conceded that Graspan does not supply the power stations nearest it, but
Camdon which is further away. 31 He described his initial response as “a little
knee –jerk” . 32
29 See Record, Commission additional filing, page 205.
30 He mentions Blackwattle, Woestalleen, Koringfontein and Kleinkoppie. See transcript dated 25
June 2007, page 28
31 25 June transcript, page 32
32 25 June transcript, page 28
15
40] Despite Eskom’s shift in stance it was still necessary for us to evaluate whether
Eskom may be affected by the merger. For this reason we put various
questions to Mr Lines concerning Eskom’s coal supply situation.
41] Mr. Lines testified that Eskom, not unlike other domestic buyers is facing coal
prices that escalate well above inflation.
42] According to Mr Lines, Eskom is presently purchasing about 120 million tons of
coal per year. 33 Eskom purchases 80% of its requirements through contracts
with mines and the remaining 20% in the spot market. Pricing is either
determined according to a base price that is escalated in terms of a preagreed
mechanism, referred to as a fixed price arrangement, or is premised on a cost
of mining plus a fee basis, referred to as a cost plus arrangement. Lines was
not able to express an opinion on which was more favourable to Eskom.
Contractual terms are he said a product of the time at which they are entered
into. 34
43] Its contracts historically were over long periods approximately 3040 years to
cover the life of a power station that was being supplied by a particular mine.
More recently, due to increased demand, contracts have been of shorter
duration. The contracts with two of the merging firms are an example of this. 35
44] Its current contract with Graspan is for two years and it recently terminated a
three month contract with Middelburg Townlands Colliery. 36 The latter contract
has not been renewed apparently because Metorex wanted to wait for the
outcome of this merger to be decided before renewing it. 37 The costs of coal
have been going up for Eskom. In the last two years Eskom has seen prices
33 25 June transcript, page 7
34 25 June transcript, page 9
35 25 June transcript, page 6
36 25 June transcript, page 5
34 25 June transcript, page 9
35 25 June transcript, page 6
36 25 June transcript, page 5
37 See 25 June transcript, page10. In cross examination it is put that “ the parties simply didn’t
come to terms” To this rather vague submission Mr. Lines responded with equal vagueness that “ I
don’t know the details , but I believe the you’re correct in general statement” ( 25 June transcript,
page 32)
16
going up somewhere between [ ] and [ ] per annum. 38
45] Lines stated that prices for export coal are presently around 50 dollars per ton,
but that the Eskom price is both below that and that of the domestic residual
market. Although it was difficult to say what the price was on average in the
residual domestic market, he estimated that they are about 2030% higher than
Eskom prices.
46] According to Lines what Eskom pays for its coal in terms of its fixed price
contracts and spot market purchases is influenced by prices in the domestic
market.
LINES: When you look at a fixed price type contract, that price is generally
informed by the market, that market as it exists in South Africa. And in
many cases bears no resemblance to the actual cost of mining.
CHAIRPERSON: Yes but it bears a resemblance to the market price?
MR LINES: To the market price, correct. 39
47] Lines attributed these cost increases to various causes including the costs of
new capital equipment for coal mines, but he added that because Eskom was
burning more coal than initially anticipated, “the supply demand relationship has
come under a bit of strain.”
48] Eskom is expanding its capacity by restoring to production power stations
previously mothballed, and by building new capacity. According Lines in the
next three years the only capacity to be coming into the market will be the
mothballed stations (three of them) and although they will have a capacity to
burn 18 million tons of coal he estimates that they may only need 8 million. 40
38 25 June transcript, page11. Although Mr. Lines wanted to give this evidence on these figures in
camera, the publicly stated figure given by his finance director Bongani Nqwababa is that Eskom
had seen its major input cost, diesel, steel and coal “rocket by about 20% on average this year and
our tariffs are unsustainable” (See Financial Mail special report on Infrastructure, June 22 2007,
page 18.)
39 25 June transcript, page 12
40 25 June transcript, page 1314
17
49] Of the new projects he mentioned –
• Medupi will come on stream in 2011 with a full capacity of 11 million tons
• Project Bravo will come on stream in 2012 with a full capacity of 18 million
tones
50] Coal for Medupi will be coming from Exxaro’s Grootgeluk mine and contracts
have already been concluded.
51] Contracts for Project Bravo have yet to be concluded, but it would appear that
Eskom will not proceed until coal reserves in the area have been proved.
Eskom is dealing with one of the mining houses (not one of the merging firms)
in this regard at the moment.
52] Eskom does not seem concerned about supply shortages or escalating prices.
This is surprising given that the firm that is the biggest buyer of domestic
coal,has seen its prices in the past two year escalate “...substantially higher
than the going inflation rate” and needs to increase purchases by about 50 %. 41
53] There are various possible reasons for this. One, the answer given by Lines, is
that Eskom is presently faced with a situation where demand has outstripped
supply beyond anyone’s expectations. As he put it:
“Clearly looking into the future Eskom’s planning will be such that we don’t get
caught in that predicament again.” 42
54] In other words what Lines is saying is that Eskom has been caught in a short
term purchasing crisis that has weakened its bargaining position. Lines expects
this short term problem to last another 2 to 3 years. 43 Short term contracts are
much more expensive for Eskom than its traditional long term contracts. Under
a long term contract Eskom is able to offer long term supply security as trade off
for lower prices. Going forward it will be able to enter into better contracts as it
41 25 June transcript, page 20
42 25 June transcript, page 2021
43 25 June transcript, page 27
18
has a better understanding of its future needs.
55] Why would coal mines want to enter into long term contracts with unattractive
pricing, at a time when the domestic market is rocketing and firms appear to
want shorter contracts not longer ones?
56] Lines’ explanation for this is that Eskom is not wholly dependant on higher
grade or export quality coal. 44 The newer power stations utilise lower grade coal
– i.e. the non export variety, indeed, if they were to burn coal of too high a
quality, they would experience technical problems. Older power stations
however were designed to burn higher grade coal and thus would be more
vulnerable to prices in the export market. Lines does not tell us how much of
Eskom’s coal requirements fall into the respective categories.
57] From the supply side even if coal is considered low grade it can be beneficiated
to make it export quality, by a process known as washing. The decision of the
mine to beneficiate will depend on the price in the export market, the cost of
washing and the price to be obtained from Eskom for unwashed coal.
58] Thus notwithstanding the fact that Eskom for the most part uses a lower grade
of coal, the price at which this can be obtained is not independent of what is
going on in the export market and by extension the residual domestic market.
According to Lines:
“I believe obviously the higher the export price, the more pressure there is on
suppliers to see how much of the coal they have, they can beneficiate for
export. So there would be, as I say, in a certain quality range, there certainly is
an overlap or there could be an overlap” 45
59] He also testified that contrary to appearances Eskom’s reliance on the spot
44 However in his written submission where Lines in explaining why thermal coal prices have
44 However in his written submission where Lines in explaining why thermal coal prices have
increased he states as a reason, inter alia, “ Demand from overseas markets ( e.g. India) for low
grade Eskom quality coal at export prices has further put pressure on the price of coal.” See
record, Commission additional file page 204.
45 25 June transcript page 26
19
market (the most expensive coal for it) would in the future reduce from the
present 20%.
60] Another reason that Eskom may be less exercised about the price of coal than
a firm in its position might be expected to be, is that as a monopoly supplier of
electricity it is in a position to pass this price on to the consumer. Mr Lines
denies and says that because the price of electricity is regulated it is not in a
position to do so. Eskom’s tariffs are set by the regulator and are imposed for a
three year cycle. Eskom is supposed to estimate what its input costs will be
over the cycle accurately. If increases are beyond what is foreseen in the
application made to the regulator (in this case the National Energy Regulator of
South Africa (Nersa)) then Eskom would have to absorb these costs.
61] However the legislation provides that the Regulator may, in prescribed
circumstances, approve a deviation from set or approved tariffs. 46 But even if
consumers do not pay this increase in the short term, it seems likely that they
will in the longer term when the present tariff structure is up for review in March
2009. Clearly this is how Eskom sees the matter itself according to the Financial
Mail, which quotes Eskom managing director Jacob Maroga as saying:
“Prices have to go up in real terms. ‘says Maroga. Not only will tariffs rise, they
will need to be ‘ well above inflation,’ he says.” 47
PPC
62] The other firm concerned about the merger was PPC. PPC is a large cement
producer which purchases cement for its cement and lime plants. PPC is the
leading cement supplier in the country and presently purchases more than 1
million tons of coal per year. It expects that figure to increase in the future due
to the growing demand for cement. At present coal is the only energy input for
cement producers. In other countries natural gas would be a substitute, but that
cement producers. In other countries natural gas would be a substitute, but that
is not possible in South Africa as we have no source of natural gas presently.
46 See section 16(3) of the Electricity Regulation Act, Act no 4 2006.
47 See Financial Mail supplement supra page 16.
20
Cement producers are thus not able to substitute coal for another energy
source.
63] It is not hard to see why PPC had concerns about the merger. At present the
merging firms combined, constitute 55% of their present coal supply. During the
course of 2006 prices to customers had escalated above inflation. When PPC
made its submission to the Commission in 2006, it knew its contract with
Lexshell’s Graspan would expire in April 2007. In the course of this year PPC
duly renegotiated its supply contract with Graspan and its worst fears were
confirmed the price doubled from what it was previously in the period April
2006 – March 2007. 48
64] Two witnesses from PPC Mr Fleische and Mr Fenn testified during the hearing.
According to them, when PPC representatives queried these new prices, they
were told that as export prices for coal had increased significantly and the coal
they wanted was of export quality, they would have to pay export prices. 49
65] The reason that PPC is concerned with the merger is that post merger it fears
coal will be diverted from the domestic market into the export market where it
receives a premium.
66] PPC’s dependency on the merged firm is not related to a geographic
competitive advantage, as its plants are situated far from the coal fields in any
event. Although in cross examination, by counsel for the merging parties, other
sources of supply were suggested to them, the view of the PPC witnesses was
that they had written to other firms to tender and there had been little response.
The number of firms in theory did not amount to alternatives in practice. In cross
examination it was suggested that attempts to find alternatives were not
sufficiently robust and that had PPC wished to, it had other suppliers to go to.
67] Even if PPC is limited to a smaller pool of suppliers than the merging parties
67] Even if PPC is limited to a smaller pool of suppliers than the merging parties
suggest it seems clear that the price increases and drive to export parity have
48 25 June transcript, page 55
49 25 June transcript Page 55
21
already occurred premerger and that the merger will not make any difference
to this situation. When this was put to PPC specifically, Mr Fleisher, very fairly,
conceded this, saying:
“….we cannot answer that question with absolute certainty at this time” 50
68] Nothing in their remaining evidence suggests anything else.
69] We did get some perspective of how Graspan viewed the April 2007
negotiation, both before and after it occurred. The before scenario emerges
from an email addressed to Glencore head office by its local representative,
Clinton Ephron. Prior to the negotiations, Ephron mentioned to his head office
that he will be meeting PPC and what price he expected to obtain. He describes
the fact that [ ]. When questioned about this during his
evidence, Ephron stated that he was not able to get the price he predicted he
would get, but got something lower, i.e. lower than export parity, and secondly,
that the reference to PPC being [ ] was not a comment on its lack of
alternative suppliers, but the fact that it had come to negotiations very late for
that quantity of supply to be available to it in the market.
Other evidence
70] There is other evidence in the record that prima facie may raise concerns about
the merger. Minutes of Wakefield colliery suggest that prices at Middelburg
Townlands are influenced by prices at its neighbour Graspan.
71] Thus Wakefield minutes dated 2 Feb 2005 reflect the following entry to explain
why prices at Middelburg Townlands Colliery were not as high as they could be:
“This product was sold at an average price of [ ] negatively affected by
the low ratio of phos (sic) coal and pricing competition from Graspan. ” 51
50 25 June transcript, page 57
51 See Metorex discovery file page 355
22
72] The situation is reversed a year later after Lexshell have taken over Graspan
and here in a quarterly report for the Wakefield board on the results period
ending 31March 2006, Mr Spencer, the managing director and author of the
report, first observes that that international prices have gone from $ 45 to $55
and goes on to comment:
“On the domestic market prices were increased as of 1 April by an average of [
] to all customers, including [ ]. The demand for domestic coal is
extremely strong, one of the reasons being that Graspan Colliery is now
owned by Shanduka/Glencore consortium, and they have diverted the majority
of the coal from this colliery into the export market.” 52
73] Mr Spencer who gave evidence at the hearing attempted to diminish the
significance of these observations by saying that they were speculative. In need
of some explanation to give to the board to explain the price movements, he
found it next door. In hindsight he would say that this explanation was
incorrect.53
74] Whether or not the local rivalry was as intense as the minutes suggest, there is
little doubt that the escalations in prices at both mines, have occurred pre
merger and appear to be on the upward trajectory, notwithstanding the merger.
Thus whatever local rivalry existed between the two mines has been eliminated
premerger, a function of either a more aggressive pricing policy since Graspan
has been under the Lexshell leadership or the change in the structure of the
markets as export prices push up domestic prices. This is best exemplified in
the graph in Table 2 above, as well as the response of customers to increases
from Wakefield. One customer of Wakefield, a large manufacturing concern,
complained very vociferously about the size of the increases in 2006 and
compared them to CPI over the same period describing the increase as
“unacceptable”. 54
52 See Metorex discovery file page 388
53 See 25 June transcript, page 95.
54 Letter from [ ] to Wakefield dated 28 March 2006, Metorex discovery page 517. See also
Metorex discovery page 462, a Wakefield marketing report dated June 2005. The report notes that
[ ] has negotiated an agreement with them for 12 months at [ ] per ton. “It is clear by this
example that the prices we can achieve on the domestic market is very close and sometimes even
23
75] The best evidence however of premerger pricing power is an internal email
exchange between Wakefield management and their wholly owned marketing
company, which gives it pricing advice. In a letter dated 8 March 2007 Ms
Lynette Kruger, the marketing manager of Wakefield, recommends that in April
2007 Wakefield increase its prices – and further suggests that the firm
reconsider fixing prices for a certain period to enable it to adjust prices should
the need arise. 55 Lest there be any ambiguity about why they were doing this,
we asked this of Mr Spencer, who confirmed that the recommendation for
shorter rather than long term contracts was based on their sense that prices still
had room to move up. 56
76] In short, it would appear that notwithstanding the contrary impression gained
from the minutes, Wakefield, even under previous ownership, was increasing
prices in the direction of export parity.
77] Another factor to consider is whether Wakefield would be in a better position to
export if owned by Lexshell instead of Metorex and hence be able to divert
more domestic production to the export market with a concomitant impact on
domestic pricing.
78] There is certainly evidence to suggest that for a variety of reasons Wakefield
has not succeeded to date in getting an export allocation out of RBCT. 57
Various explanations are given for this. In the first place it struggled with the
bureaucracy of RBCT. It never had a quota initially, and when project Quattro
was launched it applied and was told by RBCT that it was too late to be
considered that year. The following year, RBCT informed Wakefield that no
quota was available, as Project Quattro’s previously successful applicants, were
all being given renewed quotas. 58 To add to its woes, Spoornet was just as
all being given renewed quotas. 58 To add to its woes, Spoornet was just as
much a stumbling block. Lakeside Colliery had a rail siding, but Spoornet was
better than the export price. It is essential that we remain a role player on the domestic market.”
55 There used to be price regulation of coal prices and prices were traditionally set by government
gazette in April. See 25 June transcript, page 97.
56 25 June transcript, page 100
57 See 25 June transcript, page 85.
58 See 25 June transcript, page 87.
24
unhappy with it from an environmental point of view and demanded changes
before it would transport coal from it. 59 Middelburg Town Colliery did not have a
siding and one would need to be constructed. 60 They had not tried to get an
allocation for 2007, because prices on the domestic market had firmed to the
point where it was the better place to be.
79] Unlike Middelburg Townlands, Graspan has a railway siding which is operating.
If the merger goes ahead it seems that no new siding will be developed and that
this will be used by the merged firm. The merger thus eliminates at least the
front end of the bottleneck to RBCT for Wakefield. It also seems that Lexshell is
more adept at managing allocations of supply through RBCT, than Wakefield.
Indeed Glencore has already been buying coal from Wakefield for export
although the return to Wakefield would be nowhere as good as if it were
exporting directly itself.
80] The merging parties submitted that Wakefield would be in as favourable a
position to export if the merger did not proceed. Ephron testified that it would
probably cost Wakefield about R 12 million to develop a siding at Middelburg
Townlands Colliery, but that this is not a risky investment given the present
returns on export coal. He testified that the reason this has not already
happened was that the merger did not make it worthwhile. If it was to proceed
this investment would be superfluous given Graspan’s rail capacity. We have no
reason not to accept this explanation – there seems little doubt that once it had
decided to sell Wakefield, Metorex had no reason to incur further investment
costs. It is also clear from the earlier minutes that Wakefield has always aspired
to export, albeit unsuccessfully to date. It is unlikely then that the merger will
make a difference in the medium term to Wakefield’s’ chances of becoming a
make a difference in the medium term to Wakefield’s’ chances of becoming a
better exporter of coal.
The merging parties’ theory of structural change
59 This problem appears to relate to Lakeside colliery. See the June transcript at page 105.
Townlands Colliery does not have a siding at present and to construct one according to Spencer
would cost between 1320 million. ( 25 June transcript, page 102)
60 Mr. Spencer who gave this evidence explained that after one had netted off all the costs of
exporting such as rail charges and port handling the local market was “a good place to be.”
25
81] Ephron testified that for a number of years there has been underinvestment in
the coal industry particularly by the major firms. 61 One of the reasons has been
the price of coal has not made it attractive and given the export constraint
through RBCT, realising an export premium was limited. However with the
promise of the expansion of the Richards Bay terminal by 2009, the
improvement in the rail logistics and the increase in export prices, firms are now
investing to increase capacity. In the future we should have enough coal to
supply the domestic market, but those customers are going to have to accept
that the opportunity cost of supplying them is the net export parity price. The
other trend he observes is that coal or certainly export coal is becoming more
commodified.
82] Thermal coal, as we observed earlier, comes in different size ranges. Whilst
previously each of these size ranges was quoted for separately, and still is to
the domestic market, it is now sold to the export market as blend. Despite their
size difference, the individual coals from a mine all have the same chemical
content. Now that coal is commodified customers are less interested in size of
individual coals, but the cost per kilogram of a load of coal, because it is the
energy of the coal, not the size of individual coals that matters for them.
Commodifying coal in this way makes entry into the market easier and exports
likewise. Nor is it necessary any longer to buy coal from a specific mine. 62
83] Prior to the commodification trend domestic companies such as the cement
industry were large buyers of duff. Because coal companies did not have an
outlet for duff or because of the RBCT constraint, they were in a weak
bargaining position in relation to the cement companies, who could dictate cost
bargaining position in relation to the cement companies, who could dictate cost
plus prices and the lengths of contract terms, typically three years. Graspan
prior to its takeover by Lexshell was such a mine. After the takeover its new
owners appreciating that coal could be sold to the export market, even if it was
duff, were no longer beholden to the domestic customer.
84] With the changes that have taken place in the industry – the debottlenecking of
RBCT, the improvement in export prices, and the further commodifying of coal
61 25 June transcript page 120
6225 June transcript, page 115
26
even small to medium size producers will be able to export. On this scenario, a
domestic coal price at a substantial discount to export parity is a thing of the
past. It is not necessary for us to conclude whether the industry changes will be
as far reaching as Ephron suggests. There is enough evidence to suggest that
the industry is at least changing sufficiently rapidly, and has already at the time
of the merger, to make the merger an inconclusive event in a much larger
picture of structural change. Foremost in bringing this about has been the
change in capacity at RBCT and even its promise, and not yet its fulfilment, is
already driving up prices.
85] In all likelihood the merged firm will continue to increase prices at all its collieries,
Wakefield and Graspan, post merger. But this ability to raise price will come
about not because the merger gave it the market power to do so, but because of
the changes to the structure of the coal industry we have noted. For this reason
we approve the merger without conditions.
Public interest
86] The merger raises no public interest issues which would alter our conclusion.
____________________ 1 August 2007
N Manoim Date
M Holden and M Madlanga concurring.
Tribunal Researcher: Malanee MuruganModise
For the merging parties: Adv David Unterhalter instructed by Gareth Driver
(Werksmans)
For the Commission: Simon Roberts and Hylton Peterson (Mergers &
Acquisitions)
27