About SAFLII
Databases
Search
Terms of Use
RSS Feeds
South Africa: Competition Tribunal
SAFLII
>>
Databases
>>
South Africa: Competition Tribunal
>>
2011
>>
[2011] ZACT 58
|
|
Economic Development Department and Others v Wal-Mart Stores, Inc and Another (73/LM/Dec10) [2011] ZACT 58 (15 August 2011)
COMPETITION TRIBUNAL OF SOUTH AFRICA
Case No: 73/LM/
Dec10
In the matter between:
The Economic
Development Department
….............................................
1
st
Applicant
The Department of
Trade and Industry
….................................................
2
nd
Applicant
The Department of
Agriculture, Forestry and Fisheries
…......................
3
rd
Applicant
And
Wal-Mart Stores, Inc
…............................................................................
1
st
Respondent
Massmart Holdings Ltd
….....................................................................
2
nd
Respondent
In the merger between:
Walmart Stores Inc
…...............................................................................
Acquiring
Firm
And
Massmart Holdings Limited
….......................................................................
Target
firm
Panel
: N Manoim (Presiding Member)
Y
Carrim (Tribunal Member)
A
Wessels (Tribunal Member)
Heard
on : 25 March 2011
Order
issued on : 25 March 2011
Reasons
issued on : 15 August 2011
Reasons: Discovery Application and pre-hearing
directions
PART A
Discovery Reasons
Background
This is an application
for discovery brought by three of the intervenors in this merger.
The merger is between Walmart Inc, an
international retailer based
in the United States of America and Massmart – a South African
retailer. Walmart seeks to
acquire a controlling stake in Massmart
through an offer for 51% of its shares.
The intervenors are
three government departments represented by a single legal team.
They are Economic Development Department,
Department of Trade and
Industry and the Department of Agriculture, Forestry and Fisheries.
We will for convenience refer to
them collectively as the
government.
The
Competition Commission in its investigation obtained extensive
documentation from the merging parties. Neither the Commission
nor
any of the other intervenors (several trade unions and a small
business organisation) sought additional discovery. The government
however was of the view that the discovery had been insufficient and
so they sought additional discovery and information to brief
their
expert economist and to cross- examine the merging parties’
witnesses.
1
The discovery
application took place late in these proceedings. The merging
parties were resistant to discovery at this late stage
as they were
concerned that it might delay conclusion of the proceedings.
However, the merging parties were relying on evidence
from an
economist and the government was the only intervening party
intending to call an expert economist. Since the debate on
the
economic issues would be important to any finding in this case, we
decided to permit the government to bring a discovery
application,
despite the lateness of the hour, and for this reason allowed it to
be brought with less formality and within shorter
filing periods
than would usually be the case.
The
application took the form of a letter from the government requesting
certain documents and a reply from the merging parties
responding to
the request.
2
Despite its expedition
and lack of formality, we consider that in argument matters were
sufficiently addressed in order for us
to make a decision on the
request.
Rather than dealing with
this request item by item, we will deal with them thematically as
the parties did.
Broad approach
This discovery
application seeks documents and information relevant to public
interest issues in terms of section 12A(3) of the
Act.
Self-evidently these relate to issues with a broad sweep.
Two factors distinguish
an approach to discovery in such an application from one in more
conventional adversarial litigation.
In the first place the public
interest canvas is much broader than it would be in conventional
litigation, where the factual
dispute in issue more narrowly frames
the issues. But whilst the canvas is narrower in conventional
litigation, individual documents
are more significant, because so
much turns on the resolution of specific factual disputes to which
the documents sought may
be relevant. In public interest disputes
potentially many issues can be said to be relevant. Since relevance
is the usual filter
for assessing discovery claims, it is less
useful to the adjudicator in such cases in determining what
documents ought to be
produced. But whilst more documents might be
deemed relevant in a public interest case, at the same time the
probative value
of individual documents is less compelling than in
conventional litigation, if their focus is too microscopic.
Therefore to avoid
an overwhelming number of documents being
required for production we must consider other filters in addition
to relevance to
determine an application.
This is the manner in
which we have approached this decision. While documents might be,
arguably, relevant to a microscopic issue,
we ask if they are
relevant to better informing us on macroscopic issues. Even if they
may relate to macroscopic issues, we have
to weigh the value of the
information yielded to the process, against the burden to the party
required to produce it. Where the
yield is minimal or uncertain, but
the burden great, this would favour denying production.
Items dealing with
Walmart’s relationship with its employees
The government has
sought several documents that deal with Walmart’s
relationships with its employees in particular disputes
with
employees. By way of example the intervenors seek in item 2.14:
“
The
total number of individual cases brought against Wal-Mart relating to
any matter involving employment matters, including discrimination,
dismissal, victimisation, retrenchment and non-appointment. Including
the number of persons in total affected by such cases, for
the period
2000 to 2010, in all countries it operates in and in each of the US,
Brazil, Chile, Mexico, India and China as detailed
in paragraph 28
and 67 of the Witness Statement by Bond
.”
The
government also seeks documents pertinent to wage increases and the
split in employment between full-time and part-time employees
in
Chile.
3
We have refused
discovery of these items; there are two reasons for this.
In the first place the
government intervenors are not best placed to deal with Walmart’s
direct relationships with its employees.
That issue is not pertinent
to the primary issue on which they seek to intervene, which is the
effect of the merged firm’s
post merger procurement policy on
South African manufacturers and producers. In particular the
government is concerned whether
the effect of Walmart’s
assumed superior purchasing powers and logistics in international
markets may lead to a displacement
of local suppliers by imported
goods with a consequent employment loss in South Africa.
The
relationship between the merged firm and its employees is an issue
which is adequately represented in these hearings by one
of the
other intervenors the trade union South African Commercial, Catering
and Allied Workers Union (‘SACCAWU’),
which is the union
currently recognised by Massmart in its various divisions. Further,
SACCAWU is supported in these proceedings
by an international trade
union solidarity movement which can be assumed to have direct
knowledge of Walmart’s labour practices
in other countries. To
the extent that there are issues pertinent to Walmart’s labour
practices they are better placed
to bring these to the Tribunal’s
attention.
4
The widespread nature of
the government’s request suggests it has no
a priori
knowledge of these labour practices and is simply making a request
on reading of the witness statements filed thus far, to see
what
they will yield.
Secondly,
Walmart points out that it employs 2,1 million workers in 15
countries, the likelihood, with such a large and widespread
workforce that some conflict will have taken place between employer
and employee, is irrefutable. The question then is what information
would, given this vast and widespread workforce, prove informative
to us in our consideration of this merger insofar as its impact
in
South Africa is concerned. It is unlikely that the request, as
formulated, would provide us with any useful information for
our
purposes. It is microscopic not macroscopic. How much information
would represent a trend for us to take note of is not clear
nor does
the government, which seeks this information, seem to know. Granted
we will hear of numerous cases of labour disputes,
but we do not
know if they represent a generalised trend, are conclusive ( as is
typical of many disputes of this kind, many
may not be resolved,
others may be settlements for which no admissions of wrong doing are
made)
5
or are historic.
The burden on the
merging parties to discover this material which we are told is not
available to it centrally, will be extensive
and will delay
proceedings to review it. The burden and delay of its production
seems to outweigh any possible utility in receiving
the material.
Material related to
Walmart’s procurement
More central to the
issues that the government wishes to address are the procurement
practices of Walmart. To this end they have
requested a wide range
of documents.
Broadly the documents
fall into two categories. Those that are relevant to Walmart’s
experience in other countries in which
it trades and those relevant
to a post merger scenario in South Africa.
Some
documents in respect of the international comparison, which relate
to Chile have been tendered and we need not consider this
request
further.
6
One
item, which refers to a buying program implemented by Walmart’s
subsidiary in Chile, known as the JBP, and which is
relied upon in
the merging parties’ witness statements, was partially
tendered and we have ordered more specific production
for the sake
of certainty.
7
The
remaining item in this category, item 2.3, is contested. In his
witness statement Andy Bond, the only Walmart executive who
will be
testifying, claims that in Mexico, India and Chile, Walmart sourced
more than 90% of its products locally.
8
In paragraph 2.3 of the
request the government seeks documents that underpin this
contention. We have no difficulty with that
part of the request
since it is Bond who makes this assertion and the government is
entitled to see on what facts it is based.
However the government
request is more widely phrased than this. Although Bond makes the
claim in respect of ‘local procurement’
the government
wants to widen the request so that it differentiates between that
which is ‘locally sourced’ and that
which is ‘locally
produced’. In other words a product may be sourced locally,
for instance through a locally based
import agent, but may not be
manufactured locally.
What is meant by
‘locally produced’ is a subject of dispute between the
government and Walmart. Some products are
locally manufactured from
locally sourced inputs – these, everyone agrees, would be
defined as locally sourced.
But many products do not
fit neatly into this category. Sometimes local firms manufacture,
but source some or all of their inputs
from overseas. Still others
comprise a portion of local and foreign manufacture. If one accepts
that locally produced may include
some portion of imported inputs,
how does one weigh the local added value to the value of the whole
good? Nor is this proportion
constant. An input for the same good
may sometimes be locally sourced and at other times be produced from
imported components.
The government is alive
to this difficulty and have attempted a definition of “locally
produced” in their application.
This definition states:
“
Locally produced” is defined as products
which involve local production and value-add (even if some
components are imported
in the process) and does not include
products that are purely imported through locally based agents.”
This definition is
problematic as it does not define the extent of local content to the
value of the whole good. The language
used suggests it includes the
T-shirt where only the label is sewed on locally, but the rest of
the product is imported. Presumably
the government does not have
this category in mind as locally sourced, but this is the problem
with the definition.
The definition could of
course be improved upon by deeming something ‘locally
produced’ if the local value-add makes
up some fixed
percentage of the sale price. But that makes the exercise less
reliable and more assumption based.
But even if the problem
of definition can be cured, which we suggest it cannot on the
current definition, the greater problem
is whether the merging
parties can fairly be expected to produce such documentation.
Walmart states that whilst it can evaluate
local sourcing it cannot
determine local production. It does not require this information
from its suppliers. Indeed suppliers
may not have this information
since this requires them to drill down to source in an input value
chain, something that they are
not likely to do.
There is nothing to
gainsay what the merging parties say on this point about not having
this information. No evidence in the record
suggests they use local
production as a metric internally nor is it likely that they would
be particularly interested in this
information as retailers.
We have thus confined
the request in 2.3 to information that supports Bond’s claim.
The same problem of
defining local production arises in respect of item 5.1.This request
states:
“
For
the top and bottom 10 locally produced products by Rand value
purchased by Massmart in 2010 in each of the categories listed
in
tables 7 and 8 of the RBB report, the ex-factory price (ex VAT) paid
by Massmart to the local producer and the likely lowest
delivered
price to South Africa (provide the ex-factory price and likely per
unit transport costs to South Africa) from Wal-Mart’s
global
suppliers.”
There are 15 categories
listed in the two tables in the RBB report. This means that the
exercise needs to be performed in respect
of 300 products. Thus
Massmart would first have to develop a guideline for what locally
produced means and then determine which
products from the 15
categories would fall into the top and bottom 10. This would mean
examining considerably more products than
the 300 required as they
would have to eliminate any product not falling within the ‘locally
produced’ definition.
Massmart like Walmart says it does not
make this classification of its suppliers nor does it know if they
can do it themselves.
Of course Massmart could
put the burden of the task on suppliers who it thinks may fall into
the category. This runs the risk
that the information may be
unreliable or deliberately distorted since suppliers may have an
incentive to exaggerate local content
if they suspect this may
threaten future custom with Massmart. Then of course the practical
problem is of how many firms need
to supply this information and in
what time frame they can be expected to be given to produce this
information.
But nor is getting the
information from the Walmart side free from problems, i.e.
determining the likely lowest delivered price
to South Africa from
Walmart’s global suppliers. Walmart is not presently supplying
into South Africa so it is being asked
to perform an exercise for
which it has no present information. Of course it could state for
what price it sources comparable
goods and add on a hypothetical
transport cost. But it argues that different goods are sourced
sometimes from different countries
depending on their destination.
These goods are sourced in the currencies of the buying country
adding to the complexity. Nor
can pricing be assumed to be constant
over time. Then there is the issue of whether we are comparing like
goods. Is the South
African locally produced good the same one as
the one sourced from China which it is vulnerable to be displaced
by? There
inter alia
may be appreciable quality differences.
Assuming that all these
logistics from both the Massmart and Walmart side could be resolved
and we suspect not easily, is the
exercise worth its return in
effort? Apart from the fact that it is likely to be inaccurate and
imprecise (the definitional problem)
it is not explained why taking
the top and bottom 10 purchases of Massmart and comparing them with
the cost Walmart could import
them into South Africa at, is a useful
guide to measuring the potential displacement problem. We do not
know for instance:
What level of difference
in price is likely to influence the replacement of a local supplier
with an importer;
Whether, even if there
is a cheaper good that can be procured overseas that this will
occur. Consumers may resist the replacement,
imports may be
logistically difficult, security of supply may be a consideration,
to name but a few possibilities;
If we can assume
substitution will take place, how and whether this translates into
an adverse employment effect. Massmart is
not a dominant procurer
locally in most categories, on the evidence we have thus far; nor
does it mean that cheaper sourcing
via Walmart would lead inevitably
to substitution. It arguably could simply improve Massmart’s
bargaining position with
the local supplier as it has a credible
threat to turn to an overseas supplier.
Nor has the government
explained how the analysis, if it was possible and yielded some
robust conclusions, would translate into
a possible remedy.
For this reason we have
refused this request. We are unsure if the request can be complied
with and, even if it could, whether
the probative value of the yield
is worth the considerable effort in compiling it, which almost
certainly would have delayed
the hearing of this merger.
Related to this request
is another set of requests for documentation indicating the
respective procurement capabilities of Walmart
and Massmart. The
object of this request, as we understand it, is to gauge the
respective efficiencies of the procurement systems,
so as to infer
whether local procurement is likely to be displaced. It was not made
clear why comparing the respective sizes
of staff and assets of
Massmart and Walmart’s procurements (see item 2.8) would lead
to an answer to this question. Whilst
the number of offices and
personnel can be provided, what does knowing this tell us? Doubtless
that Walmart has a much bigger
procurement operation than Massmart.
But why do the number of staff it employs and their respective
office spaces translate themselves
into data from which one can
infer a future displacement of imports for locally produced goods?
Walmart sources internationally
for a huge operation while Massmart
sources for a much more modest one, operating in fewer countries. We
would expect the latter
to be larger than the former, but this does
not answer the displacement problem the government seeks to
identify.
The request is for
information that is burdensome to collect and if collected of little
probative value. Accordingly we refused
it.
The same applies to item
4.5, a request for Massmart to provide information on its use of
suppliers in particular SMME and historically
disadvantaged
suppliers. Not only is this request imprecisely framed, but it
suffers from the same problems of identifying source
as did the
‘locally produced’ category.
On the other hand we
have allowed the request for item 4.2 which is a high level strategy
document on local procurement strategy,
on which Pattison relies in
his witness statement. This is an example of a macroscopic rather
than microscopic document.
Information on German
operations
In item 2.13 the
government seeks: “
Representations and objections to
Wal-Mart’s entry into Germany and copies of all minutes of
board and management meetings
relating to its decision to exit from
the German market.”
Walmart it is common
cause exited the German market after a few years of operation.
Absent any further motivation it is unclear
why this information is
pertinent to any of the public interest issues that the government
wishes to raise in this merger. The
request is refused.
Documents whose
production we ordered
To the extent not
already referred to above the documents whose production we have
required relate to documents referred to in
merging party witness
statements either expressly or by implication or are high level
strategy documents which are macroscopic
in nature and whose
production does not appear to be unduly burdensome.
General
The final order is
annexed hereto. This order, as noted above, was given on 25 March
2011. Our reasons have only been furnished
now as they were not
requested before.
PART B
Pre hearing directions
reasons
We have also been asked
by the government intervenors to give reasons for what are described
as ‘scheduling decisions’.
It is better to think of
scheduling decisions as part of the pre-hearing directions of which
they form part.
Pre-hearing directions
are case management issues and are not considered matters for which
we are required to give reasons. The
so-called scheduling decisions
formed part of the pre-hearing directions in this case. At
pre-hearings, issues are discussed
with the parties attending, views
are given and after hearing everyone present, directions are given
by the presiding member.
This is what has taken place in the present
merger.
The first pre-hearing
took place on 18 February 2011 and present were representatives of
the merging parties, some of the union
intervenors and the
Competition Commission. The government departments were not on
record at that stage. The Economic Development
Department (the
“EDD”’) however was sent notice of the
pre-hearing, but it did not attend.
Subsequent to this the
EDD, in a letter on 24 February 2011, indicated that it wanted to
make representations in respect of the
transaction and the
Chairperson of the Tribunal advised the Department, in light of the
urgency of the matter, to submit new
information not already on
record by 28 February 2011 and to make further submissions by 9
March 2011.The department complied
with this direction.
The pre-hearing
directions and related correspondence are part of the record and we
make no further comment on them as they are
self-explanatory.
A second pre-hearing
took place on 22 March 2011 after the hearing of the transaction was
stayed on application by the union intervenors.
The hearing was then
converted to a pre-hearing and all the intervenors, including the
government, were present.
The
only issue that requires comment from us, in view of the review, is
the suggestion that proceedings were unfairly curtailed.
Whilst it
is correct that the Tribunal’s directions imposed limitations
on all parties, this was done in consultation with
those who
attended the respective pre-hearings, including the subsequent one
held on 22 March 2011, and our impression is that
no one present, at
that stage, had any objection. The government was represented at
this subsequent pre-hearing and actively
took part in the
discussions that informed our decision making. In fact, as the
transcript shows, counsel for the government
remarked:
9
Adv Bhana: “
Chair,
what we’ve done with the witnesses is we’ve kept the
existing timetable, subject to the following changes. Mr
Baker of RBB
we thought one hour for cross-examination may be too little and we’ve
suggested 2.5 hours for cross-examination,
30 minutes re-examination.
We’ve allowed the same period for the EDD expert, 2.5 hours for
cross-examination and 30 minutes
re-examination. Other than that, we
think the order can be retained, except we perhaps can do the two
experts at the end. So, we
will just move Mr Baker and we think that
fits comfortably in the 5-day timetable.”
Given the broad brush
nature of the hearings, which we discussed earlier in our discovery
reasons, we were of the view that curtailing
proceedings would not
detract from consideration of the key issues. Had proceedings been
left open-ended and there was no curtailment
of the number of
witnesses to be led or cross-examination time, these proceedings
could have been so prolonged as to be substantially
unfair to the
interests of the merging parties. The intervenors filed 17 witness
statements, which together with the four of
the merging parties
totalled 21 statements. Had all these witnesses who gave statements
been called to give oral testimony, with
no limits imposed on
cross-examination, the hearing would have required several weeks to
conclude; instead it took 6 days. Limiting
time focussed parties on
the key issues and curtailed possible diversion into any number of
potential collateral disputes, which
would ultimately either not
have been capable of resolution or necessary for us to reach a
conclusion on for the purpose of considering
the issues raised by
the merger. Recall that the hearings were limited to the public
interest considerations in terms of section
12A(3) as it was common
cause that the merger raised no section 12A(2) issues .
History of the process
We first consulted
parties about the likely length of the hearings, the number of
witnesses to be called and the likely nature
of their evidence. Once
in consultation we had determined the length of the hearing, we
asked them to submit witness statements
and thereafter we would
indicate which witnesses we would like to hear oral testimony from.
In the case of the merging parties,
limited witnesses were offered
and all were called.
The intervenors provided
17 witness statements. No limitations were placed on the length of
these statements so it was possible
to ascertain the evidence the
witness was likely to be able to give oral testimony on by perusing
them. We indicated, after considering
these statements, which
witnesses we wanted to hear oral testimony from, based on the
probative nature of their evidence and
whether they could testify
from their own knowledge or expertise or whether such knowledge was
derivative. We also preferred
witnesses who could testify to a
broader range of issues as opposed to some who had a very narrow
focus. By way of example the
witness statements of Saccuwu’s
Mbongwe and Sactwu’s Vlok were indicative of personal
knowledge as opposed to some
of the other union witnesses, and hence
they were preferred for giving oral evidence.
Although
we limited the number of the union intervenors’ foreign
witnesses, they ultimately chose who should testify for
them. Due to
a change in hearing dates the overseas witness initially preferred
by the unions’ was not available and he
was replaced by
another foreign witness who had also given a witness statement.
Similarly, the merging parties’ initial
foreign witness was
not available and another was offered who could testify on the same
issue, namely the Walmart/Chilean experience.
10
There was no curtailment
of the number of government witnesses. They only sought to call an
expert economist who was allowed to
testify.
Because we had to fit
the hearing in the allocated number of days we had to ration the
time allowed to the respective parties
for oral testimony. We did so
by preparing a draft schedule which indicated our assessment of how
much time would be required
for cross-examination and re-examination
of witnesses. Because detailed witness statements were circulated in
advance, witnesses
were not lead in chief, with the exception of the
two experts. We did however calculate the total time permitted to
each side
of the case i.e. those who supported the merger and those
who opposed it. However the respective parties were free to allocate
this total time in any manner they felt appropriate and they were
not bound to follow our recommendations as they appear in the
schedules. Thus if we had recommended half an hour for cross
examination per witness of a certain category in the schedule, the
parties were free to re-allocate that time, so as to spend less time
on one witness and more on another, provided that they did
not
exceed their sides total time. We allocated the total time on a
basis that we felt was equitable, although it was not equal.
The
intervening parties were allocated approximately 2 hours more in
total time than the merging parties.
On the final day we
heard argument and again total time was equitably allocated to those
sides in favour and against the transaction,
although we had to
include time for two parties who had not lead or cross-examined
witnesses, the Commission and the South African
Small Medium and
Micro Enterprises Forum, which represents small business
organisations.
Conclusion
Although the merger
proceedings were curtailed in terms of our directions, this was done
after consultation with all the parties
involved in the process.
Secondly, given the nature of the enquiry before us, which was
limited to considering the impact of
the merger on the public
interest, curtailment of proceedings did not prejudice the proper
consideration of issues.
____________________
15 August 2011
Norman Manoim Date
Concurring: Yasmin
Carrim and Andreas Wessels
Tribunal
Researchers:
Rietsie
Badenhorst and
Kasturi
Moodaliyar
For
the merging parties: D Unterhalter SC, assisted by J Wilson, G.
Marriot, F. Pelser instructed by Webber Wentzel Bowens and Edward
Nathan Sonnenburgs
For
the Commission: P Mtshaulana SC assisted by G Ngcangisa instructed by
the State Attorney.
SACTWU:
P McNally SC instructed by Werksmans Attorneys.
SACCAWU:
P Kennedy SC assisted by M Le Roux, instructed by Cheadle Thompson &
Haysom
The
Government: R Bhana S.C. assisted by J Meiring, instructed by Deneys
Reitz
SMMES:
G. Engelbrecht instructed by Deneys Reitz
1
Counsel
for the government intervenors stated that they were also seeking
information. See transcript of discovery hearing dated
25 March 2011
at page 54.
2
The
request and the merging parties reply are set out as Annexures A and
B. Our order which we gave on 25 March is annexure C.
3
Item
2.5
4
For
instance, in Item 2.5 the Government is requesting employment
information in respect of Chile, this is data that was presented
in
the witness statement of one of the witnesses called by SACCAWU,
Alvarez, and not by the merging parties.
5
Witness
Statement of Andy Bond, Record page 33 of the Witness Statement and
Statement of Issues Bundle
6
Item
2.4
7
Item
2.6
8
See
Bond’s witness statement paragraph 32.1.
9
See
transcript of 22 March on page 36.
10
For
the unions, Nelson Lichtenstein was replaced by Kenneth Jacobs. For
the merging parties Debra Layton was replaced by Enrique
Ostale
Cambiaso.
17