Trustees for the time being of Children's Resource Centre Trust and Others v Pioneer Food (Pty) Ltd and Others (050/2012) [2012] ZASCA 182; 2013 (2) SA 213 (SCA); 2013 (3) BCLR 279 (SCA); [2013] 1 All SA 648 (SCA) (29 November 2012)

About SAFLII
Databases
Search
Terms of Use
RSS Feeds
South Africa: Supreme Court of Appeal
SAFLII
>>
Databases
>>
South Africa: Supreme Court of Appeal
>>
2012
>>
[2012] ZASCA 182
|

|

Trustees for the time being of Children's Resource Centre Trust and Others v Pioneer Food (Pty) Ltd and Others (050/2012) [2012] ZASCA 182; 2013 (2) SA 213 (SCA); 2013 (3) BCLR 279 (SCA); [2013] 1 All SA 648 (SCA) (29 November 2012)

Links to summary

THE SUPREME COURT OF APPEAL OF SOUTH AFRICA
JUDGMENT
REPORTABLE
Case no: 050/2012
In the matter between:
THE TRUSTEES FOR THE TIME BEING OF THE
CHILDREN’S RESOURCE CENTRE TRUST
.....................
First
Appellant
THE TRUSTEES FOR THE TIME BEING OF
THE BLACK SASH TRUST
.................................................
Second
Appellant
CONGESS OF SOUTH AFRICAN TRADE
UNIONS
....................................................................................
Third
Appellant
NATIONAL CONSUMER FORUM
....................................
Fourth
Appellant
TASNEEM BASSIER
...............................................................
Fifth
Appellant
BRIAN MPAHLELE
................................................................
Sixth
Appellant
TREVOR RONALD GEORGE BENJAMIN
....................
Seventh Appellant
NOMTHANDAZO MVANA
.................................................
Eighth
Appellant
FARIED ALBERTUS
..............................................................
Ninth
Appellant
and
PIONEER FOOD (PTY) LTD
...............................................
First
Respondent
TIGER CONSUMER BRANDS LTD
..............................
Second
Respondent
PREMIER FOODS LTD
......................................................
Third
Respondent
LEGAL RESOURCES CENTRE
............................................
Amicus
Curiae
Neutral citation:
Children’s Resource
Centre Trust v Pioneer Food
(50/2012)
[
2012] ZASCA 182
(29
November 2012)
Coram:
NUGENT, PONNAN, MALAN, TSHIQI et WALLIS
JJA.
Heard
: 7 November 2012
Delivered
: 29 November 2012
Summary:
Class action – when permissible –
requirements for commencement of class action – cartel in bread
industry fixing
bread price – entitlement of non-governmental
community organisations and individual consumers to institute action
on behalf
of all consumers – requirements in regard to cause of
action and representation – definition of class – cause

of action raising a triable issue – common issue –
representation.
ORDER
On appeal from:
Western Cape High Court, Cape
Town (Van Zyl AJ sitting as court of first instance):
1. The appeal against the refusal to certify a class
action in respect of the national complaint and the class 2 claimants
is dismissed.
2. The appeal against the refusal to certify a class
action in respect of the Western Cape complaint and the class 1
claimants is
upheld and the application is remitted to the high court
for determination in accordance with the principles in this judgment.
3. The order of the high court is set aside and replaced
with the following order:
(a) If the applicants choose to pursue the application
they are granted leave to supplement their papers within two months
of this
order by delivering supplementary affidavits, to which are
annexed a draft set of particulars of claim in respect of their
delictual
claim against the respondents, embodying such further
evidence as they deem meet in amplification of that claim.
(b) The respondents are to deliver such further
answering affidavits as they deem meet within four weeks of the date
for delivery
of the affidavits referred to in para (a) of this order.
(c) The applicants are afforded two weeks thereafter to
deliver their replying affidavits, if any.
(d) The costs of the application are reserved.
4. Each party is ordered to pay his, her or its own
costs of this appeal.
JUDGMENT
WALLIS JA (NUGENT, PONNAN, MALAN et TSHIQI JJA
concurring)
[1] When may a class action be brought and what
procedural requirements must be satisfied before it is instituted?
These two questions
confront this court in litigation arising from an
investigation by the Competition Commission (the Commission) into the
bread producing
industry, initially in the Western Cape, and later in
four other provinces in South Africa. In the light of the outcome of
that
investigation the appellants applied to the Western Cape High
Court for the certification of a class action in which they proposed

to pursue a claim for damages against the respondents. That
application was dismissed and leave to appeal was refused. Such leave

was granted on petition to this court. The determination of the
appeal requires that we address the two questions I have described.

To that end we heard detailed argument in this and a related
application over two days and were furnished with copious reference

materials. That has assisted in illuminating the path for us in this
novel area of procedural law and it is appropriate at the
outset to
express our gratitude to counsel for their assistance.
Background
[2] A brief sketch of the
Commission’s investigation and the functioning of the bread
market is necessary to provide the setting
for the present
litigation. In 2006 the respondents,
1
to whom I will refer as Pioneer,
Tiger and Premier respectively, were the three largest bread
producers in the Western Cape. At
that time and for a considerable
period prior to that they, together with Foodcorp,
2
were the four largest bread producers
in South Africa. In December 2006, the Commission received complaints
in relation to the apparently
co-ordinated implementation of price
increases in the Western Cape, in conjunction with apparently
co-ordinated changes in the
terms upon which the producers dealt with
bread distributors, who supplied the informal sector of the bread
market. It then commenced
an investigation in relation to the Western
Cape in terms of the Competition Act 89 of 1998 (the Act). Premier
came forward and
disclosed details of anti-competitive conduct in
which it had engaged together with the other three bread producers,
not only in
the Western Cape but also in other parts of the country.
It sought and was granted leniency in terms of the Commission’s
corporate leniency policy.
3
[3] The disclosures by Premier led to the Commission
instituting a further investigation in relation to other parts of the
country,
which was referred to, somewhat misleadingly, as the
national complaint. Tiger entered into a settlement agreement with
the Commission
in relation to conduct in both the Western Cape and
under the national complaint. Foodcorp entered into a similar
agreement in
relation to the national complaint only. Both
settlements were confirmed in orders of the Competition Tribunal (the
Tribunal).
They involved the payment of administrative penalties of
nearly R99 million in the case of Tiger and about R45 million

in the case of Foodcorp. The complaints in respect of Pioneer were
referred to the Tribunal for adjudication. At the end of a lengthy

hearing it was found to have perpetrated anti-competitive conduct in
relation to both the Western Cape and the national complaint.

Administrative penalties totalling nearly R 196 million were
imposed upon it. It appealed against that decision but the matter
was
resolved before the hearing of the appeal. We were not told the basis
for that resolution.
[4]
The
bread producers do not sell bread directly to the public. They
determine list prices at a national level. The retail market
has
three elements. They are the large national supermarket chains, which
purchase some 25 to 30 per cent of these bread producers’
total
production, smaller general retailers, and an informal sector that
obtains supplies of bread from resellers, who are distributors
who
purchase bread for onward sale to informal retail outlets. Strictly
speaking the resellers are wholesalers not retailers.
Each
producer’s list price provides the basis for negotiating the
prices at which they supply retailers with bread. With the
large
national customers these negotiations take place at a national level.
With other customers they take place at a regional
level, subject to
some constraints and a degree of national oversight. The price
actually paid by the retailers is determined on
the basis of a
discount, expressed as a percentage, of the list price. There is no
direct control by the bread producers of the
prices at which bread is
sold in the retail market.
[5] The following conduct gave rise
to the Commission’s investigation. On 6 December 2006
the respondents’
representatives in the Western Cape met and
informed one another of the increases in the list price of bread
determined by their
respective national head offices. A date for
implementation of the increases was agreed upon. At the same time it
was agreed that
discounts afforded to distributors would be
restricted to 90 cents per loaf in Paarl and 75 cents per
loaf in the Cape
Peninsula and that the bread producers would not
deal with one another’s distributors. The effect of this was,
indirectly,
to fix the price of bread and trading conditions in
contravention of ss 4(1)
(b)
(i)
and (ii) of the Act. The national complaint was more diffuse and less
clear-cut. It involved agreements in terms of which bakeries
were
sold by one large bread producer to another, resulting in the
purchaser achieving dominance in a particular region; meetings
on
various occasions and at various places where bread prices in
relation to particular areas were discussed or agreed, or the
date of
increases in bread prices in that region were agreed; and agreements
not to poach one another’s customers. It is
apparent from the
description of these meetings in the tribunal’s determination
of the Pioneer complaint that these anti-competitive
activities were
sporadic during a lengthy period; did not always involve all of the
bread producers and were frequently restricted
to relatively small
regions or even specific places.
4
Like the Western Cape they involved
contraventions of ss 4(1)
(b)
(i)
and (ii) of the Act. The determination by the tribunal that these
provisions of the Act were contravened provides the foundation
for
the claims that are sought to be advanced in the proposed class
action.
[6] Three of the appellants are NGOs that work among
children, the poor and the disadvantaged, of whom there are so many
in our
society. The fourth, COSATU, is the largest trade union
federation in South Africa. The other five are individuals who were
consumers
of bread in the Western Cape at the time of the conduct
that gave rise to the competition complaint. All of these individuals
had
limited means and would have been adversely affected by any
increase in the price of bread. In that sense they are typical of
many
consumers of bread in both the Western Cape and the country as a
whole.
The proposed class action
[7] Mr Solomon, the Centre
Coordinator of the Children’s Resource Centre, deposed to the
founding affidavit. He alleged that
the respondents’ unlawful
conduct had breached the rights of both bread consumers and bread
distributors in the Western Cape,
but expressly confined the scope of
the application to consumers.
5
In regard to the national complaint
he accepted that the Western Cape court lacked jurisdiction to deal
with it.
6
In the result there were no
allegations in his affidavit concerning the national complaint, its
consequences or the identity of
the persons injured by the conduct
giving rise to the national complaints. Reverting to the Western Cape
he alleged that:
‘
Every consumer who bought their products
during the period in question suffered damages as a result of the
unlawful price fixing
and other prohibited practices.’
Mr Solomon said that the proposed class action was to be
brought ‘on behalf of the consumers for compensation and
related
relief’. He said that most were not in a position to
afford to engage in litigation and that each individual’s claim

was too small to justify litigation as an individual, but that
collectively the claims of consumers were ‘for a very large
sum
of money’. That is hardly surprising, as he claimed that
‘literally millions of bread consumers in the Western
Cape’
had been affected by the unlawful conduct and that for practical
purposes this amounted virtually to the public at
large in the
Western Cape.
[8] Thus far the proposed action was expressed as one in
which the claims of bread consumers against the respondents would be
consolidated
and dealt with in a single action with the appellants
representing the interests of the consumers. However, there was an
important
shift in emphasis when Mr Solomon came to deal with the
relief to be claimed in the action. He said this:
‘
The damages which each individual bread
consumer suffered are of the nature of things, very small. If a
global sum of damages was
awarded in respect of the unlawful conduct
of the respondents, the cost of distributing to each consumer his or
her share of those
damages would be prohibitive and not viable. The
further problem which would arise would be to establish precisely how
much of
the respondents’ bread each individual consumer bought
during the period in question. For this reason, in this class action

the applicants will seek class relief, in the form of an order which
will require the respondents to pay the unlawful overcharge
into a
trust or trust or similar institutions to be established for the
benefit of the bread consumers who suffered damages.
This relief was adapted in the appellants’ heads
of argument and is dealt with in para 80 below.
[9] Having set out the basis for the proposed action in
these terms, Mr Solomon claimed that the conduct of the respondents
had
infringed the constitutional right of all people in South Africa
to sufficient food in terms of s 27(1)
(b)
of the
Constitution. He invoked the section on the basis that it embodied a
negative obligation on the respondents not to impair
the right of
access to sufficient food. His principal purpose in doing so appears
to have been to bring the claim squarely within
the provisions of
s 38
(c)
of the Constitution that provides for class
actions to be brought in respect of infringements of or threats to
rights in the Bill
of Rights. The reference to the right to
sufficient food did not add anything to the cause of action or the
contention that the
anti-competitive conduct of the respondents had
caused damage to consumers of bread and that this gave rise to claims
against the
respondents for damages.
The application in the high court
[10] Based on these allegations the appellants applied
as a matter of urgency for the issue of a rule nisi calling upon the
respondents
to show cause against an order;
‘
Declaring that all bread consumers in the
Western Cape Province (“
the
consumers”
) who were
prejudicially affected by bread prices in consequence of the
respondents’ breach of
section 4(1)(b)(i)
and (ii) of the
Competition Act … constitute
members of a class.’
The draft order went on to say that the class would be
an ‘opt out’ class and a declarator was sought that the
members
of the class would be bound by the judgment in the class
action, unless they notified the appellants’ attorneys that
they
wished to be excluded as members of the class. The other
significant orders sought were a declaration that the appellants,
duly
assisted by their attorneys, to the extent necessary, had the
requisite standing to bring the class action ‘on behalf of the

consumers for damages’ pursuant to the findings of the
Commission and the Tribunal and an authorisation for the commencement

of the action forthwith without waiting for the return day of the
rule nisi. The remaining portions of the proposed rule nisi were

largely procedural in nature and dealt with notice to the members of
the class, discovery and other matters that depended on the
grant of
the primary relief.
[11] The notice of motion was issued on 18 November 2010
and the application was heard on 23 and 25 November 2010. During the
hearing
the appellants were granted leave to amend the opening
paragraph of the prayer, quoted above, by inserting the words ‘or
elsewhere’ after ‘Western Cape Province’,
presumably with a view to incorporating the national complaint. No
additional affidavits were filed. In view of the urgency of the
matter only short affidavits were delivered on behalf of the
respondents.
That procedure was adopted on the footing that the
respondents could present their case on the return date. The order
dismissing
the application was made on 26 November 2010.
[12] The result of this is that there is no evidence
from the appellants in regard to the composition of the class in
respect of
the national complaint or the manner in which they propose
to conduct that litigation. All we have is the description of the
national
complaint in the Tribunal’s determination in relation
to Pioneer We do not have full answering affidavits from the
respondents
in regard to the factual issues in respect of both
complaints and the practical issues before the court. There is a
further complication
because the appellants accepted in their heads
of argument that the certification originally sought was deficient.
Instead they
are now seeking a final order certifying two different
classes described as follows:
‘
Class 1; All persons who purchased the
bread of the first, second or third respondents in the Western Cape
Province during the period
18 December 2006 to 6 January
2009.
Class 2: All persons who purchased the bread of the first, second or
third respondents in Gauteng, Free State, North-West or Mpumalanga

Province during the period 1 September 1999 to 6 January 2009’
In the course of oral argument these orders were further
amended by inserting the words ‘for personal consumption’
after
the word ‘respondents’ in each class. In addition
the end date of the period in respect of class 1 was changed to 14

February 2007 and in respect of class 2 to 8 May 2008.
[13] We were thus asked to answer entirely novel
questions, having potentially far-reaching implications for the
respondents and
our procedural law, in the absence of a complete set
of affidavits and to grant relief different from that sought from the
high
court. This will not matter if the respondents’ arguments
prevail and we are satisfied that this is not a case that can, on
any
basis, be permitted to proceed as a class action. In that event the
appeal must fail. However, if there is some merit in the
appellants’
application, but it does not satisfy all the requirements we now
prescribe for the commencement of a class action,
it would be
patently unfair to deny them relief on a final basis without
affording them the opportunity to address matters that
would have
been addressed had they been aware of those requirements. On the
other hand it would be equally unfair to uphold their
appeal and make
a final order before the respondents have had an opportunity to
present their case fully. It would also be inappropriate
for the
court to set in train this type of novel litigation without all the
relevant facts before it. That must affect the manner
in which we
dispose of the appeal. The court raised with counsel the possibility
of an order remitting the matter to the high court,
with leave to
file further affidavits and we received submissions on that. Whether
that is an appropriate order is a matter to
which I will revert after
dealing with the merits. First it is necessary to deal with the
circumstances in which our law permits
a class action to be brought
and the requirements for doing so.
Class actions
[14] South African law is familiar with proceedings in
which a number of potential plaintiffs join together in one action to
pursue
claims against one or more defendants on the basis that the
common issues of fact and law in relation to their claims make such
a
joinder appropriate.
7
It is also familiar with the notion of a representative
plaintiff, as in the case of an action pursued by a guardian or
curator
ad litem on behalf of a minor or person under disability. In
some at least of those instances the court steps in to appoint the

representative because the individual is unable to do so. However, it
has not, until recently, recognised an action in which a

representative brings proceedings on behalf of a group of persons who
have not authorised the representative to act on their behalf.
Such
actions trace their roots back to the principles of equity in
England, were developed in the United States of America and
have
spread to a number of jurisdictions around the world.
8
They are generally referred to as class actions. There
is now express provision for class actions in s 38
(c)
of the Constitution, which provides that:
‘
Anyone listed in this section has the right
to approach a competent court, alleging that a right in the Bill of
Rights has been
infringed or threatened, and the court may grant
appropriate relief, including a declaration of rights. The persons
who may approach
a court are –
…
(c)
anyone acting as a member of, or in the interest of, a
group or class of persons.’
[15] The South African Law Commission, in line with many
other jurisdictions to which we have been referred, proposed that the
procedures
applicable to class actions be prescribed by statute, and
to that end prepared a draft bill.
9
However, Parliament has not yet acted on its
recommendations or those of a judicial commission of enquiry which
made a similar recommendation.
10
Academic voices over many years have likewise not been
heard.
11
The utility of a class action in certain circumstances
is clear.
12
We are thus confronted with a situation where the class
action is given express constitutional recognition, but nothing has
been
done to regulate it. The courts must therefore address the issue
in the exercise of their inherent power to protect and regulate
their
own process and to develop the common law in the interests of
justice.
13
This may on some occasions involve us, and courts that
will follow the guidance we give, in having to devise ad hoc
solutions to
procedural complexities on a case by case basis –
a possibility referred to by the Supreme Court of Canada
14
–
but the failure to pass appropriate legislation
dealing with this topic leaves us little alternative in the face of
the constitutional
endorsement of class actions. In what follows we
will give guidance as to the approach to be adopted in these cases.
But first
it is necessary to have clarity as to the essential nature
of a class action.
[16] In class actions the party bringing the action does
so, on behalf of the entire class, every member of which is bound by
the
outcome of the action, so that a separate action by a member of
the class after judgment can be met with a plea of
res
judicata
.
15
The concept is most fully defined, by Professor
Mulheron,
16
in the following terms:
‘
A class action is a legal procedure which
enables the claims (or parts of the claims) of a number of persons
against the same defendant
to be determined in the one suit. In a
class action, one or more persons (“representative plaintiff”)
may sue on his
or her own behalf and on behalf of a number of other
persons (“the class”) who have a claim to a remedy for
the same
or a similar alleged wrong to that alleged by the
representative plaintiff, and who have claims that share questions of
law or
fact in common with those of the representative plaintiff
(“common issues”). Only the representative plaintiff is a

party to the action. The class members are not usually identified as
individual parties but are merely described. The class members
are
bound by the outcome of the litigation on the common issues, whether
favourable or adverse to the class, although they do not,
for the
most part, take any active part in that litigation.’
[17] The class action serves to bring a number of
separate claims together in one proceeding. In other words it permits
the aggregation
of claims. However, that is not its only function. Of
equal or greater importance, as Professor Silver points out,
17
is the fact that the class action is ‘a
representational device’. It is:
‘…
a procedural device
that expands a court’s jurisdiction, empowering it to enter a
judgment that is binding upon everyone with
covered claims. This
includes claimants who, not being named as parties, would not
ordinarily be bound. A class-wide judgment extinguishes
the claims of
all persons meeting the class definition rather than just those of
named parties and persons in privity with them,
as normally is the
case.
Judges and scholars sometimes treat the
class action as a procedure for joining absent claimants to a lawsuit
rather than as one
that permits a court to treat a named party as
standing in judgment on behalf of them. This is a mistake …
Class members
neither start out as parties nor become parties when a
class is certified.’
18
[18] Recognition of the representative nature of a class
action has important implications for determining the requirements
for
such actions. If the action is representative it is essential to
identify, not necessarily by name but by description, those who
are
being represented. As it is their rights that are to be adjudicated
upon, they must either be given the opportunity to be excluded
from
the class (to opt out) or they must be required to join the class (to
opt in). It is also necessary to identify the representative
and to
determine both their suitability to act as such and the basis upon
which they will do so. The element of aggregation of
claims dictates
that the claims brought together in the action, whilst not
necessarily identical, should raise common issues of
fact or law, the
resolution of which will serve to resolve or enable the resolution of
all claims.
When may class actions be brought in South Africa?
[19] The Constitution, in s 38
(c)
,
recognises a class action specifically in relation to infringements
of or threats to rights guaranteed in the Bill of Rights.
That caused
the appellants in this case to invoke s 27(1)
(b)
of the Constitution. However, that was
unnecessary. The class of people on whose behalf the appellants seek
to pursue claims, (leaving
aside for the present the definition of
that class), is both large and in general poor. Any claims they may
have against the respondents
are not large enough to warrant their
being pursued separately, so that it is improbable that any lawyers
would be willing to act
for them on a contingency fee basis. If those
claims cannot be pursued by way of a class action, they are not
capable of being
pursued at all. The effect of that is to engage the
right of access to courts vested in each of the members of the class
by s 34
of the Constitution. The threatened infringement of that
right may be challenged by way of a class action and the appropriate
remedy
is to permit a class action in respect of the underlying
claims. It was accordingly unnecessary to seek in s 27(1)
(b)
an
alternative peg on which to hang the entitlement to proceed by way of
a class action. The right to proceed in that way was clear,
subject
to satisfying the other requirements for such an action.
19
[20] Similar circumstances will be present in many other
potential class actions and the entitlement to proceed by way of a
class
action will be clear. However, one can envisage circumstances
in which parties may wish to bring a class action, but are unable
to
contend that, if they are not able to do so, their s 34 rights will
be infringed. For example, the families of passengers, killed
when an
aeroplane on an international flight crashes, may be able to pursue
litigation in their own interests and be unable to
point to a right
in terms of the Bill of Rights that is infringed or threatened as a
result of the death of their family members.
Similarly, mortgagors,
who contend that the financial institution from which they have
borrowed money has miscalculated the interest
on their loans and
added impermissible charges, may be in a position to sue in a
suitable court to ventilate their claims. However,
in both instances
a class action may be the most appropriate means for determining
their claims.
[21] In my judgment it would be
irrational for the court to sanction a class action in cases where a
constitutional right is invoked,
20
but to deny it in equally appropriate
circumstances, merely because of the claimants’ inability to
point to the infringement
of a right protected under the Bill of
Rights. The procedural requirements that will be determined in
relation to the one type
of case can equally easily be applied in the
other. Class actions are a particularly appropriate way in which to
vindicate some
types of constitutional rights, but they are equally
useful in the context of mass personal injury cases or consumer
litigation.
I accordingly reject the suggestion advanced in some of
the academic writing, and in some of the heads of argument, that we
should
await legislative action before determining the requirements
for instituting a class action in our law. The legislature will be

free to make its own determination when it turns its attention to
this matter and in doing so it may adopt an approach different
from
ours. In the meantime the courts must prescribe appropriate
procedures to enable litigants to pursue claims by this means.
[22] Having said that, it is right to
enter one caveat. It is that, within the limited ambit of a class
action as described earlier
in this judgment, we are only concerned
to determine the broad parameters within which class actions may be
pursued and to lay
down procedural requirements that must be
satisfied in order to do so. Where necessary we must develop the
common law in order
to achieve this, for example, by expanding the
scope of the
res judicata
principle.
But, as the international literature shows, fundamental issues of
policy may arise in determining the structure of such
actions and
their consequences. The resolution of those issues involves difficult
policy choices that have received differing answers
in different
jurisdictions. It is not for us, in laying down procedural
requirements, to make policy choices that may impinge upon,
or even
remove, existing rights. That would be to trespass upon the domain of
the legislature, which the doctrine of the separation
of powers –
fundamental to our constitutional order – does not permit us to
do. Against that background I turn to address
the question of the
requirements for pursuing a class action in South Africa.
Requirements for a class action
Certification
[23] All of the parties accepted that it is desirable in
class actions for the court to be asked at the outset, and before
issue
of summons, to certify the action as a class action. This
involves the definition of the class; the identification of some
common
claim or issue that can be determined by way of a class
action; some evidence of the existence of a valid cause of action;
the
court being satisfied that the representative is suitable to
represent the members of the class; and the court being satisfied
that a class action is the most appropriate procedure to adopt for
the adjudication of the underlying claims. In my view they were

correct to do so and we should lay it down as a requirement for a
class action that the party seeking to represent the class should

first apply to court for authority to do so.
21
My reasons for adopting that requirement are the following.
[24] Most jurisdictions around the world require
certification either before institution of the class action or at an
early stage
of the proceedings. The exception is Australia. The
justifications are various. First, in the absence of certification,
the representative
has no right to proceed, unlike litigation brought
in a person’s own interests. Second, in view of the potential
impact of
the litigation on the rights of others it is necessary for
the court to ensure at the outset that those interests are properly
protected and represented. Third, certification enables the defendant
to show at an early stage why the action should not proceed.
This is
important in circumstances where the mere threat of lengthy and
costly litigation
22
may be used to induce a settlement even though the case lacks merit.
23
Fourth, certification enables the court to oversee the procedural
aspects of the litigation, such as notice and discovery, from
the
outset. Fifth, the literature on class actions suggests that, if the
issues surrounding class actions, such as the definition
of the
class, the existence of a prima facie case, the commonality of issues
and the appropriateness of the representative are
dealt with and
disposed of at the certification stage, it facilitates the conduct of
the litigation, eliminates the need for interlocutory
procedures and
may hasten settlement. Lastly the Australian experience has not
proved entirely satisfactory, with numerous interlocutory

applications and significant costs and delays being experienced.
24
[25] Accordingly in my judgment an application for
certification of the proposed action as a class action was necessary
in the present
case. As the dismissal of the application finally
disposed of the question whether the appellants could institute such
an action
on behalf of the proposed class, it was appealable.
25
It is unnecessary for present purposes to determine whether the grant
of certification would be subject to an appeal or whether
it is a
decision capable of alteration by the court of first instance and
therefore not appealable. I turn next to consider the
requirements
for a certification application.
[26] In the course of argument the presiding judge put
to counsel the following list of the elements that should guide a
court in
making a certification decision. They were:
the existence of a class identifiable by objective
criteria;
a cause of action raising a triable issue;
that the right to relief depends upon the determination
of issues of fact, or law, or both, common to all members of the
class;
that the relief sought, or damages claimed, flow from
the cause of action and are ascertainable and capable of
determination;
that where the claim is for damages there is an
appropriate procedure for allocating the damages to the members of
the class;
that the proposed representative is suitable to be
permitted to conduct the action and represent the class;
whether given the composition of the class and the
nature of the proposed action a class action is the most appropriate
means
of determining the claims of class members.
There is an element of overlapping in these
requirements. For example, the composition of the class cannot be
determined without
considering the nature of the claim. The fact that
there are issues common to a number of potential claimants may
dictate that
a class action is the most appropriate manner in which
to proceed, but that is not necessarily the case. A class action may
be
certified in respect of limited issues, for example, negligence in
a mass personal injuries claim, leaving issues personal to the

members of the class, such as damages, to be resolved separately.
[27] This list corresponds substantially with the
factors identified by the Law Commission as the requirements for
certification.
26
It also overlaps with what Cameron JA said were ‘the
quintessential elements of a class action’, in dealing with a

contention that a class had been inadequately described, namely
‘…
that (1) the class is
so numerous that joinder of all its members is impracticable; (2)
there are questions of law and fact common
to the class; (3) the
claims of the applicants representing the class are typical of the
claims of the rest; and (4) the applicants
through their legal
representatives, the Legal Resources Centre, will fairly and
adequately protect the interests of the class.’
27
Similar requirements are prescribed in Federal Rule
23(a) of the Federal Rules of Civil Procedure in the United States of
America,
namely that the class is so numerous that joinder of all its
members is impracticable; that there are questions of law or fact
that are common to the class; that the claims of the representative
parties are typical of the claims of the class; and that the

representative parties will fairly and adequately represent the
interests of the class. These requirements are referred to as
numerosity, commonality, typicality and adequate representation.
28
Similar requirements are to be found in other jurisdictions.
29
[28] Counsel did not dispute that these were the issues
that require to be addressed in an application for the certification
of
a class action and directed their arguments at certain of them,
primarily the definition of the class; the nature and existence
of
the cause of action being advanced; the commonality of the claims of
members of the class; and the appropriateness of the relief
being
sought. Without excluding the possibility of there being other issues
that require consideration, it suffices for our purposes
to say that
a court faced with an application for certification of a class action
must consider the factors set out in the list
in para 26 and be
satisfied that they are present before granting certification. I now
address in greater detail some of these
factors that are of
particular relevance for this case.
Class definition
[29] In defining the class it is not necessary to
identify all the members of the class. Indeed, if that were possible,
there would
be a question whether a class action was necessary, as
joinder under Uniform Rule 10 would be permissible. It is, however,
necessary
that the class be defined with sufficient precision that a
particular individual’s membership can be objectively
determined
by examining their situation in the light of the class
definition. It is important to be able to do this for three reasons.
First,
it affects the manner in which notice is given to the members
of the class. In the conventional situation of an ‘opt out’

class the entitlement to opt out is negated if people cannot
ascertain with reasonable certainty whether they are members of the

class in the first place. Second, it is necessary for people to know
whether they can commence their own litigation against the
defendant
or defendants in the class action. Third, it is essential to the
identification of those who are bound by the judgment.
30
[30] Two problems with class definition that arise in
this case are over-inclusive definition and definition by subjective
criteria.
Where the class suffers from these problems it impacts upon
other elements relevant to the certification decision. Thus if the
class is too wide, as in an Australian case
31
where the original pleaded case included ‘every man, woman and
child who has been in this country between 1992 and 1999’,
the
litigation will be unmanageable because of the need to take the
personal circumstances of every person in the class into account.

That indicates that a class action is inappropriate.
[31] An over-inclusive class also raises the question
whether there are common issues of fact or law that can conveniently
be resolved
in the class action in the interests of all members of
the class. The broader the class the less likely it will be that
there is
the requisite commonality. This was one of the reasons for
the majority refusing certification in
Wal-Mart Stores, Inc,
Petitioner v Betty Dukes et al
, the most recent decision of the
United States Supreme Court on class actions. The claim for
certification was denied because the
statistical evidence of systemic
discrimination against the 1.5 million members of the class,
consisting of women employees from
1998 until the commencement of the
litigation, scattered across 3 400 stores employing 1 million
people in 50 states (in respect
of 14 of which there was no anecdotal
evidence of discrimination at all), was held to be inadequate. The
sheer size of the class,
the disparate circumstances of its members
and the fact that policy for dealing with staffing issues was
formulated by individual
store managers, excluded the existence of a
common issue and precluded certification.
32
That did not mean that some, even many, of the members of the class
may not have been the victims of discrimination. It meant only
that
the claim of systemic discrimination, essential to the existence of a
common issue for determination in the class action,
was
insufficiently demonstrated.
[32] The problem of certifying a class on the basis of
subjective factors is particularly manifest when the definition makes
membership
of the class dependent upon the outcome of the litigation,
for example, because it is dependent upon the class member having
suffered
loss, or dependent upon the ability of the defendant to
raise defences to some claims and not others. The point is
illustrated
by the case of
Emerald Supplies Ltd & another v
British Airways PLC
.
33
The claimant claimed that British Airways participated in illegal
price fixing cartels operating in the area of air freight charges.
It
sought redress for itself and all consumers of international air
freight services. The members of the class were those persons
who
were direct or indirect purchasers of such services at prices
inflated artificially by the activities of the price fixing cartels.

Three different types of loss were identified as having potentially
been suffered by members of the class. Whether they had in
fact
suffered such loss depended on the individual user and it was
accepted that loss would have to be proved individually. The
claim
was for a declaration that British Airways was liable ‘in
principle’ to class members for the three different
forms of
loss.
[33] The problem was that, until it had been established
that the price fixing cartels existed; that British Airways
participated
in them; and that prices were artificially inflated
thereby, it was impossible to identify any member of the class. In
addition,
as the class included both direct and indirect purchasers
of such services, whether any particular purchaser had suffered loss
depended on whether the overcharge had been absorbed by the direct
purchaser in fixing its prices or passed on to an indirect purchaser

by way of a price enhancement. In a judgment by Mummery LJ the
entitlement to pursue the action as a representative action was

rejected. He said the following:
‘
Judgment in the action for a
declaration would have to be obtained before it could be said of any
person that they would qualify
as someone entitled to damages against
BA. The proceedings could not accurately be described or regarded as
a representative action
until the question of liability had been
tried and a judgment on liability given. It defies logic and common
sense to treat as
representative an action, if the issue of liability
to the claimants sought to be represented would have to be decided
before it
could be known whether or not a person was a member of the
represented class bound by the judgment.’
34
[34] Any attempt to define a class must take account of
these potential pitfalls. The essential question will always be
whether
the class is sufficiently identified that it is possible to
determine at all stages of the proceedings whether a particular
person
is a member of the class.
A cause of action raising a triable issue
[35] The appellants accepted that a class action should
not be certified if the case is ‘hopeless’. I am not sure
that
this constitutes a sufficiently clear standard to be applied on
a case by case basis. Whether a case is hopeless has two aspects.
It
is hopeless if it is advanced on a basis that is legally untenable.
It is also hopeless if it is advanced in the absence of
any credible
evidence to support it. These are categories that have long been
recognised in our law and practice. A case is legally
hopeless if it
could be the subject of a successful exception. It is factually
hopeless if the evidence available and potentially
available after
discovery and other steps directed at procuring evidence will not
sustain the cause of action on which the claim
is based. In other
words if there is no prima facie case then it is factually hopeless.
[36] In my judgment these are the standards that should
be applied in assessing whether a proposed class action reflects a
cause
of action raising a triable issue. I will deal with each in
turn. Causes of action are not in the first instance dependent on
questions
of law. They require the application of legal principle to
a particular factual matrix. The test on exception is whether on all

possible readings of the facts no cause of action is made out. It is
for the defendant to satisfy the Court that the conclusion
of law for
which the plaintiff contends cannot be supported upon every
interpretation that can be put upon the facts.
[37] An issue raised in argument was how this test could
be applied in the context of a novel claim. The answer is that,
provided
the novel claim is legally plausible, the standard is met
and the claim survives scrutiny and must be determined in the course
of the action. Take a delictual claim based on a novel legal duty not
to act negligently.
35
The existence of such a duty depends on the facts of the case and a
range of policy issues. The need for the court to be fully
informed
in regard to the policy elements of the enquiry militate against that
decision being taken without evidence. As Hefer
JA said:
36
‘
As the judgments in the cases
referred to earlier demonstrate, conclusions as to the existence of a
legal duty in cases for which
there is no precedent entail policy
decisions and value judgments which ‘shape and, at times,
refashion the common law [and]
must reflect the wishes, often
unspoken, and the perceptions, often dimly discerned, of the people”
(
per
M M Corbett in a
lecture reported
sub
nom
“
Aspects
of the Role of Policy in the Evolution of the Common Law” in
(1987)
SALJ
104
at 67). What is in
effect required is that, not merely the interests of the parties
inter se
,
but also the conflicting interests of the community, be carefully
weighed and that a balance be struck in accordance with what
the
Court conceives to be society's notions of what justice demands.
(
Corbett
(
op
cit
at
68); J C van der Walt “Duty of care: Tendense in die
Suid-Afrikaanse en Engelse regspraak”1993 (56)
THRHR
at 563-4.)
Decisions like these can seldom be taken on a mere handful of
allegations in a pleading which only reflects the facts
on which one
of the contending parties relies. In the passage cited earlier
Fleming
rightly stressed the
interplay of many factors which have to be considered. It is
impossible to arrive at a conclusion except upon
a consideration of
all
the circumstances of
the case and of every other relevant factor.’
37
[38] That does not mean that certification should not be
refused in an appropriate case, where there is no prospect of a trial
court,
with the benefit of all the evidence that the plaintiff can
muster or suggest will be available to it, holding that the claim is

legally tenable. Sometimes the test on exception can appropriately be
applied in respect of novel legal claims. Harms JA pointed
that out
in the following passage from his judgment in
Telematrix
:
38
‘
[2] The plaintiff's
particulars of claim, with annexures, runs to 158 pages and contains
a full exposition of the events surrounding
the Directorate's
decision. In addition we were provided with the ASA's Code of
Advertising Practice and Procedural Guide and the
parties, prudently,
were content that regard could be had to it even though it does not
form part of the pleadings. The case does
not, therefore, have to be
decided on bare allegations only, but on allegations that were
fleshed out by means of annexures that
tell a story. This assists in
assessing whether or not there may be other relevant evidence that
can throw light on the issue of
wrongfulness. I mention this because,
relying on the majority decision in
Axiam
Holdings Ltd v Deloitte & Touche
,
the plaintiff argued that it is inappropriate to decide the issue of
wrongfulness on exception because the issue is fact-bound.
That is
not true in all cases. This Court, for one, has on many occasions
decided matters of this sort on exception. Three important
judgments
that spring to mind are
Lillicrap
,
Indac
and
Kadir
.
Some public policy considerations can be decided without a detailed
factual matrix, which by contrast is essential for deciding

negligence and causation.
[3] Exceptions should be dealt with sensibly. They
provide a useful mechanism to weed out cases without legal merit. An
over-technical
approach destroys their utility. To borrow the imagery
employed by Miller J, the response to an exception should be like a
sword
that 'cuts through the tissue of which the exception is
compounded and exposes its vulnerability'. Dealing with an
interpretation
issue, he added:
“
Nor do I think that the mere
notional possibility that evidence of surrounding circumstances may
influence the issue should necessarily
operate to debar the Court
from deciding such issue on exception. There must, I think, be
something more than a notional or remote
possibility. Usually that
something more can be gathered from the pleadings and the facts
alleged or admitted therein. There may
be a specific allegation in
the pleadings showing the relevance of extraneous facts, or there may
be allegations from which it
may be inferred that further facts
affecting interpretation may reasonably possibly exist. A measure of
conjecture is undoubtedly
both permissible and proper, but the shield
should not be allowed to protect the respondent where it is composed
entirely of conjectural
and speculative hypotheses, lacking any real
foundation in the pleadings or in the obvious facts.”’
[39] It must be borne in mind that, as a result of the
procedure we now lay down, the party seeking certification will have
set
out in a draft pleading and in affidavits the basis for the
proposed action. In so doing the court will probably have more
material
available to it in regard to the cause of action than would
be the case with a normal exception. That will enable the court to
make a proper assessment of the legal merits of the claim and,
sensitively applied in this new area of law and procedure there
should not be a difficulty. Unless it is plain that the claim is not
legally tenable, certification should not be refused. The court

considering certification must always bear in mind that once
certification is granted the representative will have to deliver a

summons and particulars of claim and that it will be open to the
defendant to take an exception to those particulars of claim.
The
grant of certification does not in any way foreclose that or answer
the question of the claim’s legal merit in the affirmative.
[40] Establishing a prima facie case on the evidence is
not a difficult hurdle to cross. In the context of an attachment to
found
jurisdiction Scott JA set out the test as follows:
‘
[12] The requirement of a
prima facie case in relation to attachments to found or confirm
jurisdiction has over the years been said
to be satisfied if an
applicant shows that there is evidence which, if accepted, will
establish a cause of action and that the
mere fact that such evidence
is contradicted will not disentitle the applicant to relief —
not even if the probabilities
are against him; it is only where it is
quite clear that the applicant has no action, or cannot succeed, that
an attachment should
be refused. . Nestadt JA, in the
Weissglass
case … warned
that a court must be careful not to enter into the merits of the case
or at this stage to attempt to adjudicate
on credibility,
probabilities or the prospects of success.
[13] …
[14] What is clear is that the evidence on which an
applicant relies, save in exceptional cases, must consist of
allegations of
fact as opposed to mere assertions. It is only when
the assertion amounts to an inference which may reasonably be drawn
from the
facts alleged that it can have any relevance. In other
words, although some latitude may be allowed, the ordinary principles
involved
in reasoning by inference cannot simply be ignored. The
inquiry in civil cases is, of course, whether the inference sought to
be
drawn from the facts proved is one which by balancing
probabilities is the one which seems to be the more natural or
acceptable
from several conceivable ones … While there need
not be rigid compliance with this standard, the inference sought to
be
drawn, as I have said, must at least be one which may reasonably
be drawn from the facts alleged. If the position were otherwise
the
requirement of a prima facie case would be rendered all but nugatory
…’
[41] A similar standard is applied in other instances
such as the test for the existence of a defence in summary judgment
proceedings.
There is no reason why it cannot be applied to determine
whether the applicant for certification has shown the existence of a
cause
of action. I would add only this to Scott JA’s
exposition. The test does not preclude the court from looking at the
evidence
on behalf of the person resisting certification, where that
evidence is undisputed or indisputable or where it demonstrates that

the factual allegations on behalf of the applicant are false or
incapable of being established.
39
That is not an invitation to weigh the probabilities at the
certification stage. It is merely a recognition that the court should

not shut its eyes to unchallenged evidence in deciding a
certification application. Properly applied the test for a prima
facie
case should not pose any insuperable difficulties for an
applicant for certification.
[42] The appellants accepted in their heads of argument
that to obtain certification a prima facie case had to be
established. They
submitted that the existence of such a case did not
involve any enquiry into the merits. In doing so they relied on two
cases,
Eisen v Carlisle & Jacquelin et al
, from the United
States of America,
40
and
Hollick v Toronto (City)
, from Canada.
41
Neither case supports this contention. The passage from
Eisen
on which reliance was placed was explained in
Wal-Mart
42
as not excluding the necessity for evidence to show that the
requirements of Federal Rule 23(a) were satisfied and this would
necessarily involve evidence on the merits. In
Hollick
the
question was posed to what extent the class representative ‘should
be allowed or
required
to introduce evidence in support of a
certification motion.’
43
The answer in the light of the recommendations of the Ontario Law
Reform Commission was that:
‘
In my view, the Advisory
Committee’s report appropriately requires the class
representative to come forward with sufficient
evidence to support
certification, and appropriately allows the opposing party an
opportunity to respond with evidence of its own.’
Evidence is therefore required to identify the class,
identify the common issue or issues and show that a class action is
appropriate.
That necessarily means that there must be evidence
showing a prima facie cause of action, because the existence of a
cause of action
underpins the existence of a class and serves to
identify the issues common to that class that require determination.
[43] It is desirable to say something about the
procedure to be adopted in certification applications. The appeal was
complicated
by the absence of a clear statement by Mr Solomon of the
cause of action that the appellants intended to advance. It was
unclear
whether the claim was based on the breach of the Act’s
provisions or was a constitutional claim seeking constitutional
damages.
44
In the appellants’ heads of argument it was said to be a
delictual claim, with an alternative claim based upon a breach of
the
constitutional right to sufficient food. This confusion would have
been avoided if the application had been accompanied by
a draft set
of particulars of claim in which the cause of action was pleaded, the
class defined and the relief set out. The affidavit
or affidavits
filed in support of the application would then have set out the
evidence available to the appellants in support of
that cause of
action and the further evidence that they anticipated would become
available to them to sustain the pleaded case
and the means by which
that evidence would be procured. That procedure should be followed in
future applications. That will enable
those opposing certification to
respond meaningfully and the court to decide the application with a
clear understanding of the
nature of the case. That is not to say
that the court must treat the draft pleading as if it were the law of
the Medes and Persians,
insofar as the content of the applicant’s
case is concerned. Manifestly it can be amended and the need for
amendment may
emerge in the course of the certification application.
But it should at least serve to clarify the issues arising in the
certification
application.
Common issues of fact or law
[44] This does not require that every claim advanced in
the class action, save possibly in relation to quantum, be identical.
It
requires that there be issues of fact, or law, or both fact and
law, that are common to all members of the class and can
appropriately
be determined in one action. Dealing with the issue of
commonality in
Wal-Mart
Scalia J said
45
that the claims:
‘…
must depend upon a
common contention …That common contention, moreover, must be
of such a nature that it must be capable
of classwide resolution –
which means that determination of its truth or falsity will resolve
an issue that is central to
the validity of each one of the claims in
one stroke.’
In my view that is correct. The simplest example of such
a common issue would be the issue of negligence in a case involving
the
derailment of a train. That could give rise to different claims,
such as damages for personal injuries by passengers, dependents’

claims for loss of support in respect of those killed, claims for
loss of or damage to goods being carried on the train and damage
to
other property arising as a result of the derailment, but there would
be sufficient commonality on the issue of negligence to
sustain a
class action.
[45] That highlights the point that the class action
does not have to dispose of every aspect of the claim in order to
obtain certification.
It might in an appropriate case be restricted
to the primary issue of liability, leaving quantum to be dealt with
by individual
claimants. Certain common issues could be certified for
the entire class, and other subsidiary issues certified in respect of
defined
sub-classes. But the question in respect of any class or
sub-class is always whether there are common issues that can be
determined
that will dispose of all or a significant part of the
claims by the members of the class or sub-class.
The representative
[46] In some jurisdictions, such as the United States,
it is an express requirement that the representative plaintiff has a
claim
that is typical of the claims of the class. In Canada and
Australia, whilst there is no express requirement of typicality,
Professor
Mulheron suggests that the jurisprudence of those countries
in regard to commonality, makes that a requirement.
46
That question does not arise in South Africa, because s 38
(c)
of the Constitution expressly contemplates a class action being
pursued by ‘anyone acting as a member of,
or in the interest
of
, a class’. Accordingly, while the appellants include
five individuals who may be typical of the class they are seeking to

represent, the other four appellants may permissibly act in the
interest of the class. As already indicated there is no reason
to
differentiate in that regard between class actions based on
infringement of rights protected under the Bill of Rights and other

class actions.
[47] The court must be satisfied of two broad matters in
regard to the representatives of the class. The first is that they
have
no interests in conflict with those whom they wish to represent.
A conflict of interest arises if the purpose of the litigation
is to
enrich the representatives, or to serve interests other than those of
the class. The second issue is whether the representative
has the
capacity to conduct the litigation properly on behalf of the class.
That is important because unsuccessful litigation will
have the
effect of destroying the claims.
[48] The capacity to conduct the litigation has a number
of aspects that must be dealt with in the application for
certification.
First, has the representative the time, the
inclination and the means to procure the evidence necessary to
conduct the litigation?
Second, has the representative the financial
means to conduct the litigation and, if not, how is it going to be
financed? This
will involve making some assessment of the likely
costs. Third, does the representative have access to lawyers who have
the capacity
to run the litigation properly? This will require some
consideration of the likely magnitude of the case and the resources
involved
in dealing with it. Fourth, on what basis are those lawyers
going to be funded? Fifth, if the litigation is to be funded on a
contingency
fee basis, details of the funding arrangements must be
disclosed to ensure that they do not give rise to a conflict between
the
lawyers and the members of the class. The court must also be
satisfied that the litigation is not being pursued at the instance
of
the lawyers for their own gain rather than in the genuine interests
of class members, as the risk of conflicts of interest is
inherent in
that situation. It is for this reason that in other jurisdictions the
court’s approval of any settlement is required.
Whilst that
issue does not arise in these proceedings, so that it is unnecessary
for us to be prescriptive, some similar requirement
will need to be
imposed when that situation does arise.
Application of principles in this case
[49] Whilst much of the argument focussed around the
Western Cape and what the appellants now describe as Class 1, it is
convenient
to deal in the first instance with Class 2 and the
application to certify a class action in respect of the issues
flowing from
the national complaint to the Commission.
Class 2
[50] This application did not originally seek
certification in respect of the national complaint and a class
consisting of purchasers
of bread in the four provinces, Gauteng,
Free State, North West and Mpumalanga, where the events giving rise
to that complaint
occurred. Accordingly we lack any evidence in
regard to that class. All we have is the Tribunal’s
determination in the Pioneer
case of the conduct that gave rise to
the Tribunal’s findings that Pioneer, in conjunction with the
other bread producers,
had engaged, directly or indirectly, in price
fixing and the division of markets in those four provinces. However,
there is no
evidence as to the effect of that conduct in the market
place and its impact on the price of bread for consumers. In addition
the
orders by the Tribunal in respect of Tiger and Foodcorp do not
appear to overlap entirely with the findings in the determination
in
respect of Pioneer.
[51] This is relevant because the claim is based upon
anti-competitive conduct in terms of the Act. That claim must be
pursued on
the basis of a determination by the Tribunal. Given the
passage of time and the provisions of s 67(1) of the Act it
seems
improbable that any further determination will be made about
anti-competitive conduct in the areas covered by the national
complaint.
The certification of a class action must then be addressed
on the footing that the case to be advanced will be in accordance
with
the determination by the Tribunal in the Pioneer case.
[52] The Commission did not allege before the Tribunal
that a cartel existed at all times across all regions of the country
or that
the co-ordination or agreement or understanding between the
bread producers related to the same subject matter in all regions.
47
The factual findings of the Tribunal were varied. In regard to the
division of markets in 1999 Pioneer sold its bakery in Welkom
to
Tiger and Tiger agreed ‘to keep out of the wider Free State
area’ in favour of Pioneer. In the North West, also
in 1999,
the agreement was confined to the informal trade, where it was agreed
that the four bread producers would not compete
in supplying that
trade in the areas around Krugersdorp, Orkney, Stilfontein and
Potchefstroom. In Mpumalanga in 2001, Tiger and
Pioneer jointly sold
a bakery in Bushbuckridge, in which they owned equal shares, to
Foodcorp. In return Foodcorp sold its bakery
in Groblersdal to
Pioneer and its share in the bakery in Ermelo to Tiger.
48
[53] It is impossible on this evidence to say that this
division of markets affected all purchasers of the respondents’
bread
in these four provinces, or even that it detrimentally affected
all purchasers of their bread in the particular areas. Nor can it
be
said that its impact on consumers would have been similar. That will
have depended on what occurred as the bread passed down
the value
chain from producer to distributor or retailer and thence to the
consumer. Any impact on the national chains of supermarkets
would
have been different from any impact on smaller retail outlets and the
informal sector. There is no evidence that it actually
resulted in
higher prices in those areas or caused prejudice to consumers. It
may, for all we know have resulted in greater production
efficiencies
and lower prices, particularly if the bakery sales occurred because
the sellers were not operating them profitably.
Even if it had an
adverse effect on consumers, the nature, geographic scope and
duration of that effect is not known and would
have differed from
area to area.
[54] In regard to price fixing, whether direct or
indirect, there was no evidence of this in either the Free State or
Mpumalanga.
The Tribunal heard evidence about meetings in North West
in 2003 or 2004, and another meeting in 2005, where information was
exchanged
about impending price increases and arrangements made to
co-ordinate increases, not to undercut prices and not to poach
customers
when this occurred. These arrangements did not involve
Tiger, but did include Foodcorp. There is no evidence that it had any
effect
elsewhere in the other three provinces. Its impact in the
North West is not dealt with and we do not know the nature,
geographic
extent or duration of any adverse effect.
[55] In relation to Gauteng the Tribunal heard the
evidence of one witness who testified to regular communications about
price increases
during 2003 and 2004 and the efforts made by the
bread producers to ensure that this did not set off a round of
discounting directed
at increasing market share – probably the
most significant element in determining profitability. This was done
by exchanging
information about price increases and, if one producer
broke ranks by discounting in response to an increase, threats to
retaliate.
The same witness also testified to an agreement between
Pioneer and Premier not to compete on price when a new distribution
centre
was opened in Vanderbijlpark and to an agreement on the
co-ordinated implementation of price increases in Gauteng generally
in
December 2006. Once again there is no evidence of the nature,
geographic extent or duration of any adverse effect. This is
particularly
necessary in relation to price increases because, as the
Tribunal’s determination records, a major driver of price
increases
is the cost of the inputs into production, in particular
wheat and flour, but also electricity, labour costs, transport costs
and
the like. Co-ordination of price increases would have had little
impact if an increase by the price leader (the dominant producer
in a
particular region) would, in any event, have resulted in the other
producers following suit.
[56] It is plain from this summary that the class for
which certification is sought in Class 2 is over-broad; that the
potential
claims lack any central common feature; and that any losses
that this anti-competitive conduct may have occasioned (none having

been alleged or shown by evidence), would have varied from place to
place, time to time and person to person. The class includes
a
considerable number of bread consumers who cannot possibly have been
affected by the anti-competitive conduct of the bread producers,
or
would have been affected by some, but not all, of such conduct. I
have considered whether there are common issues that would
warrant
certification in the wide class sought, together with the creation of
sub-classes, to cover particular situations. However,
the diversity
of the Tribunal’s findings in regard to the conduct falling
under the national complaint precludes that. It
cannot readily be
segmented without information identifying the people affected by the
division of markets in the Free State, or
the agreement not to
compete on price in Vanderbijlpark, or the coordinated increase in
price in Gauteng in December 2006.
[57] A class can only be identified with reference to
the conduct of a specific potential defendant or defendants. However,
Premier
was not involved in the anti-competitive conduct in the Free
State and Tiger was not involved in the agreement in respect of
Vanderbijlpark
or those in the North West. Premier was not party to
the market division in Mpumalanga. This diversity precludes the
identification
of a common issue the disposal of which would have a
decisive effect on the claims of all members of the class.
[58] The findings by the Tribunal in the Pioneer case
also preclude the possibility of there being a common issue or issues
that
can apply to a defined class of persons for the purpose of a
class action. They involve six different types of conduct occurring

respectively in 1999, 2003 and 2004, July 2006 and December 2006. The
1999 conduct did not extend to or impact upon Gauteng. The
2003 and
2004 agreements were confined to parts of the North West. The
anti-competitive conduct in 2006 related to Gauteng alone.
Although
the findings and the papers before us repeatedly use the sweeping
term ‘cartel’ to describe this conduct,
it is misleading
to the extent that it suggests widespread conduct beyond that which
the Tribunal held had occurred. There was
no finding of the existence
of a general cartel operating in the bread producing industry in
Gauteng, Free State, North West and
Mpumalanga from 1 September 1999
to 8 May 2008, which is the area and the period for which
certification was sought. Indeed there
could not have been such a
finding because that was not an allegation made by the Commission.
[59] Lastly it is necessary to mention that the absence
of any evidence concerning the impact of this anti-competitive
conduct on
consumers also precludes certification. Without such
evidence the requirement of commonality among the members of the
class is
not satisfied. The claim is said to be one for damages. The
assessment of whether damages have been suffered arising from
anti-competitive
conduct is a complex issue. The following method is
taken from an English case involving a price fixing cartel
49
and adapted to this case:
Determine or estimate the actual prices charged by the
cartel during the relevant period.
Estimate the price (known as the ‘but for’
price) that would have been charged had there been no cartel.
Subtract the ‘but for’ price from the
actual price.
Determine the quantity of bread purchased by each
claimant or by the class of claimants.
Estimate the proportion of the ‘over-charge’
absorbed by intervening distributors and the retailers as the bread
is
passed down the chain of purchases to the end consumer.
[60] Applying this methodology – which was not
challenged - to this case, involves entirely separate exercises being
undertaken
in a number of small regional areas in the four provinces,
in respect of widely divergent conduct taking place at differing
times
and affecting consumers differently depending on the actions of
the retailers and resellers. In addition the exercise is to take

place in relation to three different bread producers who set list
prices separately in relation to the different types of bread
that
they produced
50
and negotiated discounts separately and in different ways for
different groups. The result was that they charged varying prices
to
the retailers and resellers to whom they sold bread. In turn the
retailers and resellers fixed their own prices independently.
The
inevitable result is that there is no commonality within the
suggested class in regard to harm suffered or the nature of any

damage that may have been caused to individual consumers. That would
not matter if there were some overriding common issue that
applied to
all bread consumers in these four provinces during the relevant
period, but none is identified. There is simply the
sweeping
statement (admittedly made in relation to the Western Cape) that
‘every consumer that bought their products during
the period in
question suffered damages’. That cannot be correct in the light
of the Tribunal’s analysis of the offending
conduct.
[61] The inevitable conclusion is that, on the
assumption that there is a delictual action available to consumers
who suffer damage
in the form of higher prices in consequence of
anti-competitive conduct, the proposed Class 2 is too broad because
it includes
people who were not injured by the conduct of the
producers. Furthermore there is no common issue of fact or law shared
by all
the members of the class. Consumers who suffered damages as a
result of any of the anti-competitive conduct constituting the
national
complaint did so for varying reasons arising from different
conduct in different areas at different times. The cause, nature and

extent of those damages are not common to the proposed class. The
claim for certification in respect of Class 2 must therefore
fail.
Class 1
[62] This is the class to which the application related
from the outset. The proposed action arises from the co-ordinated
implementation
of price increases in the Western Cape by the
respondents from 18 December 2006. Although reference is made in the
founding affidavit
to the other practices concerning the distributors
it is not clear how they could have adversely affected the consumers.
If the
distributors’ discounts were reduced, and they absorbed
the reduction by accepting a decline in their profit margins, that

would not have affected consumers at all. If they increased their
prices in order to maintain their profit margins, not only would

their sales have fallen, because consumers refused to buy bread at
those prices or obtained it elsewhere, but any loss to consumers

would have flowed from the independent actions of the distributors.
Similarly the agreement not to poach distributors was supplementary

to the discount agreement. It is, however, unnecessary for present
purposes to decide this, as it suffices for the purpose of
certification for the appellants to show that consumers would have a
cause of action arising from the co-ordinated increase in list

prices.
[63] The appellants have now nailed their colours to the
mast of a delictual action flowing from a breach of statutory duty.
They
persisted with their constitutional claim based on a breach of a
negative obligation not to interfere with the right to sufficient

food in s 27(1)
(b)
of the Constitution, but only in the
alternative. Whilst there was some criticism in argument of the
manner in which their claim
had fluctuated in the course of these
proceedings I do not think that should affect matters. On any basis
the proposed claim is
a novel one and identifying the proper basis
for it is a matter of some complexity. The appellants should not be
penalised in certification
proceedings for variations in the legal
foundation of the claim they seek to advance. On a factual level it
has always been the
same claim, namely that consumers of bread in the
Western Cape were obliged to pay more for bread than they would
otherwise have
done if the bread producers had not engaged in the
prohibited anti-competitive conduct the Tribunal found they had
perpetrated.
[64] The claim is advanced on the following basis. The
appellants say that this type of anti-competitive conduct is
prohibited in
the interests of competition and the interests of
consumers. These prohibitions serve to fulfil the aim of the Act as
set out in
the preamble and in particular the aim of providing
markets in which consumers can freely, that is, without coercion by
anti-competitive
conduct, purchase the quality and variety of goods
they desire. They also serve to provide consumers with competitive
prices as
provided in s 2
(b)
of the Act. Founding upon
cases that say that a breach of a statutory duty can give rise to a
legal duty not to cause financial
loss,
51
the appellants contend that such a legal duty rested on the bread
producers and that they breached it deliberately by their actions
in
agreeing on the co-ordinated increase of list prices. In support of
such a duty they point to s 65(6) of the Act that contemplates

that a person who has suffered loss or damage as a result of a
prohibited practice may have an action to recover that loss or
damage. They contend that the prohibited practice in this case
resulted in consumers in the Western Cape paying more for the
respondents’
bread than they would otherwise have done.
[65] The respondents challenge this case on a variety of
levels. First, they submit that as a matter of interpretation the
Competition
Act was not enacted for the benefit generally of the
consumers on whose behalf it is sought to proceed and therefore does
not create
the alleged legal duty. Second, they contend that there is
no evidence of loss suffered by the consumers arising out of the
anti-competitive
conduct. Third, they submit that there is no
evidence linking the anti-competitive conduct with any increase in
prices or damage
to consumers, and hence the necessary causal element
is missing. In other words, on the first leg they raise a legal
argument and
on the latter two legs they say there is no prima facie
case in relation to these elements of a delictual action.
[66] Premier challenges the delictual claim on two other
bases. Its principal argument is that such a claim is bad in law
because
s 65 of the Act provides for a ‘follow on’
claim founded on a finding of prohibited anti-competitive conduct by

the Tribunal, and that on a proper construction of the Act this
remedy is exclusive and precludes a common law delictual action.
In
the alternative it contends that the proposed claim is, by virtue of
the relief being sought, not a claim for damages and accordingly
not
a valid delictual claim.
[67] The legal arguments about the existence of a legal
duty and the existence of an exclusive statutory claim in terms of
s 65
of the Act are linked. If Premier is correct that the Act
provides an exclusive follow on claim then the legal duty on which
the
appellants rely does not exist. However, if it is incorrect, that
strengthens the appellants’ hand considerably, because,
s 65(6)
recognises the possibility of claims arising from prohibited
anti-competitive conduct, so that the absence of a specific
statutory
claim would suggest that there must be a delictual claim available to
at least some persons injured by such conduct.
[68] I am not convinced that s 65 of the Act
provides for the type of exclusive follow on remedy for which Premier
contends.
The relevant portions of the section read as follows:
‘
(2)  If, in any
action in a civil court, a party raises an issue concerning conduct
that is prohibited in terms of this
Act, that court must not consider
that issue on its merits, and—
(a)
if the issue
raised is one in respect of which the Competition Tribunal or
Competition Appeal Court has made an order, the court
must apply the
determination of the Tribunal or the Competition Appeal Court to the
issue; or
(b)
otherwise,
the court must refer that issue to the Tribunal to be considered on
its merits, if the court is satisfied that—
(i) the issue has not been raised in a frivolous or
vexatious manner; and
(ii) the resolution of that issue is required to
determine the final outcome of the action.
(6)  A person who has suffered loss or damage
as a result of a prohibited practice—
(a)
may not
commence an action in a civil court for the assessment of the amount
or awarding of damages if that person has been awarded
damages in a
consent order confirmed in terms of section 49D(1); or
(b)
if entitled
to commence an action referred to in paragraph
(a)
,
when instituting proceedings, must file with the Registrar or Clerk
of the Court a notice from the Chairperson of the Competition

Tribunal, or the Judge President of the Competition Appeal Court, in
the prescribed form—
(i) certifying that the conduct constituting the basis
for the action has been found to be a prohibited practice in terms of
this
Act;
(ii) stating the date of the Tribunal or Competition
Appeal Court finding; and
(iii) setting out the section of this Act in terms of
which the Tribunal or the Competition Appeal Court made its finding.
(7)  A certificate referred to in
subsection (6)
(b)
is
conclusive proof of its contents, and is binding on a civil court.
(8)  An appeal or application for review
against an order made by the Competition Tribunal in terms of section
58 suspends
any right to commence an action in a civil court with
respect to the same matter.
(9)  A person’s right to bring a claim
for damages arising out of a prohibited practice comes into
existence—
(a)
on the date
that the Competition Tribunal made a determination in respect of a
matter that affects that person; or
(b)
in the case
of an appeal, on the date that the appeal process in respect of that
matter is concluded.
(10)  For the purposes of section
2A(2)
(a)
of the
Prescribed Rate of Interest Act, 1975 (Act No. 55 of 1975),
interest on a debt in relation to a claim for damages in
terms of
this Act will commence on the date of issue of the certificate
referred to in subsection (6).’
[69] Whilst the section contemplates the possibility of
a claim for damages at the instance of a person who suffers loss or
damage
in consequence of a prohibited practice, it does not state
that such a person will have an action, nor does it purport to
determine
the requirements for such an action. Its provisions are
rather directed at the relationship between the Tribunal’s
determination
and the court before which such an action is brought;
the timing of such an action; the running of prescription in relation
to
the claim underlying such action and interest on the claim. As
regards the first of these, it says in s 65(2) that the issue
of
whether prohibited conduct has occurred must be determined by the
Tribunal, or the Competition Appeal Court on appeal from it,
to the
exclusion of the court before which the action is commenced. Proof of
such prohibited conduct is provided by the certificate
of the
Chairperson of the Tribunal or the Judge President of the Competition
Appeal Court (ss 65(6) and (7)). Until the proceedings
before
the competition authorities are complete an action may not be pursued
(s 65(8)). If the Tribunal has already awarded
damages in a
consent order under s 49D(1) a further claim for damages is
excluded. The commencement date for the running of
prescription is
dealt with in s 65(9) and the date upon which interest on a
claim will start to run is dealt with in s 65(10).
All of these
are essentially procedural matters or, in the case of prescription
and interest, matters dealt with in other statutes
that require
special treatment in this particular context.
52
[70] Premier draws attention to the bifurcation of
jurisdiction created by the Act, where the specialist tribunals have
exclusive
jurisdiction over the question whether prohibited practices
have been committed and the civil courts have no say. However, to my

mind that is a two-edged sword. The writ of the specialist
competition tribunals runs only to the extent of determining whether

a prohibited practice has occurred and no further. Beyond that the
civil courts provide the forum to which parties who suffer damages

arising from the prohibited practice must turn in order to recover
those damages. They will clearly apply conventional principles
in
regard to the assessment of damages and causation of loss. However,
the more difficult question is to identify those who are
entitled to
pursue claims to recover such loss. Premier’s argument assumes
that anyone who has suffered loss or damage and
can demonstrate a
causal link between that loss and damage is entitled to be
compensated for that loss or damage. That is consistent
with its
contention that the Act provides for strict liability.
[71] Under the principles of the law of delict it is
insufficient for someone to say that they have suffered loss as a
result of
the conduct of another, even where such conduct was
deliberate or negligent. They must first establish that the person
from whom
they seek to recover that loss owes them a legal duty to
prevent such loss so that their failure to do so is wrongful in
accordance
with the legal convictions of the community.
53
In the context of pure economic loss, into which category this claim
falls, that is a particularly complex enquiry.
54
For example the risk of indeterminate loss would have to be weighed
against the inability of the consumers to protect themselves
against
loss flowing from the prohibited practice. It would be a startling
departure from principles that have been recognised
as compatible
with our new constitutional order,
55
for liability to compensate for loss or damage flowing from a
prohibited practice to exist in the absence of any duty to prevent

such loss. It would eliminate one of the basic principles by which
our law prevents liability for acts causing damage from being

extended beyond acceptable limits.
[72] On the other hand there are textual factors that
lend support to Premier’s argument. Thus s 65(6)
(a)
speaks of a person who has suffered loss or damage as a result of a
prohibited practice commencing an action in a civil court ‘for

the assessment of the amount or awarding of damages’. The first
part of that is consistent with the action concerning only
the
quantum of damages and nothing more. Section 49D(4) refers to a
complainant ‘applying for an award of civil damages’.

That is also consistent with the making of the award being a
mechanical process in which only the quantum of the claim must be

assessed.
[73] The record as it stands is not apt for determining
questions of this importance and difficulty. The claim by the
appellants
has been advanced on too simple a basis to determine
whether the foundation for the legal duty necessary to support a
delictual
claim exists. That is not surprising. After all, the claim
was only advanced on that basis when their heads of argument were
filed
in this court. However, the contention by Premier is not
clearly correct and the corollary to its rejection appears to me to
be
that a delictual action would lie at the instance of at least some
claimants. In my view it is undesirable to determine these issues
in
an appeal where the arguments being advanced, both as to legal duty
and as to the effect of s 65, are raised for the first
time on
papers that are admittedly incomplete.
[74] The arguments that the factual allegations made by
the appellants were insufficient to make a prima facie case in
respect of
a delictual claim have force in the light of the limited
evidence proffered by the appellants. However, the case is one of
price
fixing by agreeing on a date for implementation of price
increases by all three respondents. That agreement would have served
to
stifle competition that might otherwise have resulted from any one
of the respondents increasing its prices without such an agreement.

As such it would have tended to increase the prices that the bread
manufacturers were able to charge to their customers and, markets

being what they are, it is probable that to some extent at least
those price increases would have been passed on to consumers.
After
all part of the justification given by the bread producers for the
increases was that prices of wheat and flour and other
major input
costs had risen and they needed to pass these on to consumers if
their businesses were to remain profitable. If artificially
high
increases in the price of bread resulted from this prohibited
practice by the respondents, it is a logical inference that
consumers
would have paid more for bread than would otherwise have been the
case. In the absence of rebutting evidence from the
respondents,
demonstrating the falsity of that line of reasoning, the appellants’
case on the facts cannot be rejected at
this stage of proceedings. It
is plain that the claim as now advanced raises both legal and factual
issues that would be common
to all affected consumers.
[75] In summary the claim that the appellants seek to
advance has a potentially plausible basis, but it is premature at the
stage
of this appeal for this court to determine the questions raised
by these arguments in view of their novelty, complexity and the
fact
that they are raised for the first time in this court. The appellants
should not be non-suited on these grounds, which would
be the effect
of dismissing their appeal, but equally the respondents’
arguments cannot be rejected at this stage. That indicates
that it is
desirable to refer the present application back to the high court,
with appropriate directions for the delivery of further
affidavits,
for it to be dealt with on a complete set of papers and in the light
of the principles laid down in this judgment.
[76] There are two possible reasons why this court
should not make that order. The first concerns the definition of the
potential
class and the second the proposed remedy. If we were to
conclude that the class is incapable of satisfactory definition or
that
no proper relief can be claimed in this action then it would be
wrong to remit the matter and cause further expense to be incurred

and valuable court time consumed on a fruitless exercise. However, in
considering that question one must be cautious not to stifle
what may
be litigation properly, if inadequately, commenced when we do not
have all the facts and do not have the advantage of
full affidavits
on both sides of the question and argument that would clarify what is
at present obscure.
[77] In regard to the proposed class it is clearly too
broadly stated. The proposition that it includes virtually all
consumers
of bread in the Western Cape cannot be correct, when the
market share of the respondents in South Africa as a whole was said
by
the Tribunal to be some 50 to 60 per cent of the domestic bread
market. For all we know the respondents’ operations may have

been largely confined to the Cape peninsula and the winelands,
leaving the vast hinterland of the province to smaller operations.

That would suggest that the class should be confined to people
resident in that more limited area. The sales of the respondents
in
the Western Cape will have included sales to institutions, such as
schools, hospitals and prisons, and commercial establishments,
such
as hotels, bed and breakfast establishments, restaurants and other
establishments where food is served. These should be excluded.
It may
be that on further consideration it is appropriate to exclude
consumers whose bread purchases were made through the national
chains
of supermarkets and garage forecourts, where prices were negotiated
nationally and the impact of the price fixing may have
been
significantly different. It may also be appropriate to restrict the
case to the basic types of bread – the standard
loaf of white
and brown bread, sliced and unsliced – and not include all the
varieties that were available from the bread
producers at the time,
which are probably not the staple fare of the poor consumer. The fact
that none of this was done illustrates
that the proposed class is
over-broad and cannot be certified as such.
[78] Having said that, however, can it be said that the
proposed class is incapable of adequate definition? The evident aim
of the
appellants is to represent the interests of poor consumers in
the Western Cape, or the relevant part of it, those who would have

been hardest hit by any artificially sustained increase in the price
of bread. It should be possible to define that group with
greater
specificity perhaps by using the income bands that economists, and
many in the commercial world, use to differentiate differing
economic
groups. No doubt statistical information is available from Statistics
SA and other sources. Information about the sources
of bread for
these consumers may enable the class to be defined with greater
precision. It is probably inevitable that any class
will include some
people who do not consume bread or did not consume the respondents’
bread, but that should not preclude
certification if the class is
otherwise adequately defined and statistical controls are in place to
accommodate that possibility.
[79] If one takes all these factors into consideration I
do not think that it is necessarily impossible for the appellants to
define
the class they wish to represent with the degree of clarity
that is required. I may be wrong in the light of information produced

by the respondents when they have an opportunity to deliver full
affidavits, which may show that the bread market is more fragmented

than it may appear at first sight. However, that is not a reason at
this stage to say that the appellants’ task in seeking

certification is doomed to failure.
[80] That leaves the question of remedy. The remedy that
the appellants propose at this stage suffers from many defects.
Whilst
the appellants say that they wish to represent the consumers
of bread and pursue the claims for damages of that class, the remedy

they seek is that money be paid to unnamed and as yet unconstituted
trusts or similar bodies that they say will use the funds generally

to benefit bread consumers. How this is to be done is not explained.
In the heads of argument it is suggested that the money would
be used
to fund community and school feeding schemes. Commendable though that
may be it is not a means of compensating consumers
of bread, who
suffered loss or damage in consequence of the prohibited practices of
the respondents, for their loss or the damage
they suffered. It is a
means of providing food – not necessarily bread – to
those who lack it. Many of those will be
people who suffered no loss
or damage at all in consequence of the prohibited practices. Some,
such as children, who are frequently
the principal beneficiaries of
such schemes, would not even have been alive at the time. Others may
not have had the means to buy
bread at the time or may not have been
in the Western Cape. This is clearly recognised in the appellants’
heads of argument
where they say that the purpose of the proceedings
is not to identify the members of the class and distribute the
proceeds to them.
In fact the appellants seek to turn this to their
advantage by saying that in view of the intended form of relief ‘an
over-inclusive
class definition would be harmless’.
[81] The litigant who sues in delict sues to recover the
loss suffered in consequence of the wrongful act of the defendant.
56
The appellants propose to prove the claim on this basis, but then not
pay the damages to the members of the class. The justification
for
this approach is said to be that in circumstances where it is
impractical to distribute to the members of the class directly
the
damages should be distributed
cy-près
, that is, in a
manner as near as possible to a direct distribution. That would be a
novel development in our law, but we were referred
to foreign
jurisprudence in relation to class actions that supports some
alternative form of distribution of damages where distribution
to the
members of the class is impractical and urged to permit such a mode
of distribution in the present case.
[82] According to Professor Mulheron,
57
Australia does not permit a distribution other than to the members of
the class, although this has occasionally caused difficulties
in
distributing the balance of a settlement fund when all the
beneficiaries that have come forward have been paid and there is
a
residual amount in the fund. In Canada provision is made by statute
for such distributions in certain cases. Only in the United
States
does it appear that courts have permitted such distributions. Even
there this has usually been in relation to settlement
agreements,
where the court is asked to confirm a settlement the terms of which
have been devised by the parties to the litigation,
or in respect of
residual amounts remaining after distribution to all members of the
class who have come forward with claims. As
a remedy in disputed
litigation, distributions
cy-près
appear to have gained
little purchase and they have by and large not been welcomed by
appellate courts.
[83] In South Africa the Law Commission’s working
paper on class actions
58
proposed the application of an aggregate award of damages or a
settlement amount in a way that could reasonably be expected ‘to

compensate or benefit class members, where actual division and
distribution of the award among the class members is impossible
or
impracticable’. It gave as examples a price reduction for a
certain time, where there had been an illegal overcharge for
goods or
services, and the use of compensation to clean up pollution or
provide health services to those who were injured by the
pollution.
In each case whilst the money would not go directly to the
beneficiaries it would be used for their benefit as a remedy
for the
harm they had suffered. However, when the Law Commission came to
deliver its final report even this relatively modest proposal
did not
find favour. It limited its recommendations to the following:
‘
When an aggregate assessment
is made the court should give directions regarding the distribution
of the award to class members and
may, where appropriate, require the
defendant to distribute the damages directly to the class members.
The Act should contain an
express provision with regard to the
aggregate assessment of monetary awards and the disposal of any
undistributed residue of an
aggregate award.’
59
The draft bill that accompanied the report did not in
fact make provision for the distribution of an undistributed residue
of an
aggregate award, but presumably that was a drafting oversight.
[84] The remedy that is being sought at this stage does
not aim to compensate the members of the proposed class in any way.
It is
unconnected to that goal. Its aim is purely punitive, a matter
already dealt with by the Tribunal in the substantial administrative

penalties that it levied against the respondents.
60
It has the effect of depriving the members of the class of their
claims in order to enable the appellants (and this is likely in

practice to mean the first four appellants, not the individual
appellants) themselves, through a trust they control or set up,
to
oversee the distribution of the damages in a manner they think
appropriate. That is a remedy that in some other jurisdictions
has
been introduced by way of legislation after a careful policy review
and in other jurisdictions is simply impermissible. It
is not a
remedy foreshadowed by the Law Commission in its report. It has been
adopted by some lower courts in the United States
in different
circumstances and has received at best a lukewarm and at worst a
hostile reception from higher courts. As a remedy
it is controversial
as Professor Mulheron demonstrates in her comprehensive works in this
area of the law.
[85] It is at this point that the warning I sounded
earlier, about the court not making policy decisions that are
properly the preserve
of the legislature, comes to the fore. In my
view the suggested remedy is not a permissible one. It departs from
the purpose of
the class action to compensate those who have suffered
loss for that loss, by stripping them of their claims, with the
excuse that
as they are small they are worthless, and vests those
claims in others to pursue in their own interests. Worthy though
those interests
are that cannot justify the court in permitting this.
It would not involve a development of the common law, but rather a
substantial
alteration to it. The new principle that would be created
would be that where a large number of people had relatively small
claims
against a defendant, that it would not be worth their while to
pursue individually, those claims can be confiscated from them by

judicial
fiat
and vested in a person that will be able to use
the proceeds of those claims in a socially useful manner. In my view
that is a
bridge we should decline to cross.
[86] Does that mean that, where the claims are so small
that there is no practical way in which to pursue them and distribute
the
proceeds to the individual claimants, no class action can be
brought? In my view not. The problem can be solved by a small
extension
of our existing principles of the law governing damages
along the lines suggested in the Law Commision’s working paper.
The
action proceeds on the basis that the claim is one to recover the
damages suffered by the members of the class. Where those damages
are
all of the same nature, which is the case here where the complaint is
that consumers were allegedly unlawfully forced to pay
more for bread
than they would otherwise have done, they can be computed on an
aggregate basis using well-established statistical
methods. There is
nothing novel in this. Statistical methods are used in many aspects
of the computation of damages.
[87] Once the aggregate damages have been computed the
next step is for the appellants to identify the mode of distribution
that
will serve as a surrogate for the distribution directly to
individuals of the amount of their loss. That may be by way of a
targeted
price reduction for a period, a remedy that Professor
Mulheron says has been found to be ‘particularly effective for
remedying
overcharges on items which are repeatedly purchased by the
same individuals’.
61
Alternatively it may be by way of distribution in a way that can be
shown to be likely to benefit, directly or indirectly, the
members of
the class. This calls for the type of judicial creativity that Harms
JA said is essential to provide an effective relief
to those affected
by a constitutional breach.
62
As we are extending the availability of a class action in order to
give effect to the right of access to courts it is incumbent
on us to
ensure that the right is rendered effective by ensuring that
appropriate remedies will flow from its exercise.
63
In my judgment that can be done in a class action for damages, where
the damages cannot be distributed practicably to the members
of the
class, by way of alternative methods that will, directly or
indirectly compensate the members of the class for their loss.
The
methods outlined above are examples of how that can be done, but do
not constitute a closed list of possibilities. In each
case the
proposed remedy must be identified at the outset and must be
appropriate to the facts of the particular case. What is

impermissible is the type of remedy proposed in the present case
where the members of the class are not compensated either directly
or
indirectly for the loss they have suffered.
Conclusion in respect of Class1
[88] In my judgment on the record before us the
appellants have shown that there is a potentially viable claim for
delictual damages
vested in a class of consumers. That does not mean
that the claim is a good claim in law or that on these papers there
is sufficient
evidence available to demonstrate the existence of a
prima facie case if it is sound in law. Those matters are not yet
ripe for
determination for the reasons given in para 75. All that can
be said at this stage on these papers is that the claim is not
legally
untenable. The class is defined in terms that are overbroad,
but there are grounds for believing that it is capable of more
precise
definition. The case raises issues common to all members of
the class. It may, however, be open to attack both as a matter of law

and on the grounds that there is insufficient evidence to sustain it
even at a prima facie level. That can only be determined once
a
complete set of affidavits has been filed. In addition the relief
that the appellants have indicated they wish to claim is not

competent. However, there is no reason to think that in the light of
the principles laid down in this judgment an appropriate prayer
for
relief cannot be formulated.
[89] I have considerable sympathy for the learned acting
judge who was confronted with this application on a basis of extreme
urgency
and had to wrestle with these novel and difficult issues
within a short time and against a background where there was no
clarity
as to the requirements for a class action beyond those he
distilled from the Law Commission’s report. In addition the
arguments
presented to us were not presented to him and he did not
have the advantage we have had of full argument and reference to
international
authority by able counsel. His principal concern
appears to have lain with the appellants’ view that the claim
against Tiger
would prescribe and the consequent prayer to be
permitted to issue summons before certification, which he treated as
a prayer for
final relief. That concern may have been misplaced. If,
as we now hold, an application for certification is the first
necessary
step in proceedings to pursue a class action there is much
to be said for the proposition that, for purposes of prescription,
service
of the application for certification would be service of
process claiming payment of the debt for the purposes of s 15(1)
of the Prescription Act.
64
Such an interpretation would be supported by cases, where the
institution of similar necessary preliminary proceedings, have been

held to constitute the bringing or commencement of suit for various
purposes.
65
[90] Even if the prayer seeking authority to issue
summons before finalising the issue of certification was bad, it did
not necessitate
the dismissal of the application at that stage. Had
the learned acting judge had deployed before him the arguments that
we have
heard, and had the benefit of knowing the parameters that we
now lay down within which such applications must be determined, I do

not think that he would have disposed of the application in that
summary way. He would (and should) have refused to certify the
claim
in respect of the national complaint now described as the claim in
respect of Class 2. However, viewing the claim in respect
of Class 1
in the light of the requirements of this judgment, he should have
required that the appellants supplement their application
by
presenting a draft set of particulars of claim and afforded them the
opportunity of addressing the issues of a prima facie case,
the
definition of the class, the appropriateness of the relief being
claimed and the suitability of the representative (in the
sense dealt
with in paras 47 and 48) in further affidavits. He should then have
given an opportunity to the respondents to file
full answering
affidavits and to the appellants to reply, after which the
application could have been dealt with in the light of
a full
appreciation of the respective parties’ cases.
[91] As that is the approach that in my judgment should
have been adopted, the appeal must succeed in relation to Class 1 and
the
matter must be referred back to the high court to be dealt with
in accordance with the requirements of this judgment. In view of
the
fact that this is novel litigation, in which at first instance the
parties were in large measure operating in the dark, and
in view of
the fact that such success as the appellants have obtained as a
result of this judgment may in the long run bring them
little
advantage, the fairest order to make in regard to costs is that all
the parties should bear their own costs in this appeal.
[92] The following order is made:
1. The appeal against the refusal to certify a class
action in respect of the national complaint and the class 2 claimants
is dismissed.
2. The appeal against the refusal to certify a class
action in respect of the Western Cape complaint and the class 1
claimants is
upheld and the application is remitted to the high court
for determination in accordance with the principles in this judgment.
3. The order of the high court is set aside and replaced
with the following order:
(a) If the applicants choose to pursue the application
they are granted leave to supplement their papers within two months
of this
order by delivering supplementary affidavits, to which are
annexed a draft set of particulars of claim in respect of their
delictual
claim against the respondents, embodying such further
evidence as they deem meet in amplification of that claim.
(b) The respondents are to deliver such further
answering affidavits as they deem meet within four weeks of the date
for delivery
of the affidavits referred to in para (a) of this order.
(c) The applicants are afforded two weeks thereafter to
deliver their replying affidavits, if any.
(d) The costs of the application are reserved.
4. Each party is ordered to pay his, her or its own
costs of this appeal.
M J D WALLIS
JUDGE OF APPEAL
Appearances
For appellant: W H van der Linde SC (with him Steven
Budlender and Michelle le Roux), the heads of argument having been
prepared
by Wim Trengove SC with Mr Budlender and Ms Le Roux.
Instructed by:
Abrahams Kiewitz Attorneys, Cape Town
Honey Attorneys, Bloemfontein
For first respondent: S F Burger SC (with him J P V
McNally SC and J A Cassette)
Instructed by:
Cliffe Dekker Hofmeyr Inc, Sandton
Symington & De Kok, Bloemfontein.
For second respondent John Dickerson SC (with him
Michelle O’Sullivan and Ross Garland)
Instructed by:
Edward Nathan Sonnenbergs, Cape Town
Matsepes Inc, Bloemfontein
For third respondent David Unterhalter SC (with him Max
du Plessis and Michelle Goodman)
Instructed by:
Nortons Inc, Sandton
McIntyre & Van der Post, Bloemfontein.
For amicus curiae Tembeka Ngcukaitobi, the heads of
argument having been prepared by Geoff Budlender SC with Mr
Ngcukaitobi.
Instructed by:
Bloemfontein Justice Centre, Bloemfontein.
1
Pioneer
Food (Pty) Ltd, Tiger Consumer Brands Ltd and Premier Foods Ltd.
2
Foodcorp
(Pty) Ltd.
3
The
corporate leniency policy was dealt with recently in the judgment of
this court in
Agri Wire (Pty) Ltd v The
Competition Commissioner
[2012] ZASCA
134
(660/2011);
[2012] 4 All SA 365
(SCA).
4
The
one agreement related to the Vanderbijlpark area alone. The bakery
sales related to the Free State and Mpumalanga and did
not involve
Premier. An agreement in relation to price increases was confined to
Gauteng.
5
The
related application, heard at the same time as this case, was
brought by a Mr Mukkadam on behalf of distributors. Judgment
in that
matter will be handed down simultaneously with this judgment.
6
It
is irrelevant whether this view was correct.
7
Uniform
Rule 10. In admiralty proceedings it is expressly permitted to bring
proceedings under a collective title such as ‘the
cargo laden
and lately laden on board the MV …’ (Admiralty Rule
2(3)).
8
World
Class Actions: A Guide to Group and Representative Actions around
the Globe
(ed Paul G Karlsgodt) (Oxford, 2012) lists 38
jurisdictions plus the European Union with some or other form of
class action. The
majority of these regulate class actions by
special legislation or rules of court.
9
South
African Law Commission, Project 88,
The Recognition of Class
Actions and Public Interest Actions in South African Law
, August
1998.
10
Commission
of Enquiry into the Affairs of the Masterbond Group and Investor
Protection in South Africa
, Vol 3, Chapter 16 by Mr Justice H C
Nel.
11
F
R Malan ‘Siviele proses, verbruikersbeskerming en kollektiewe
optrede’
1982
TSAR
1
; Wouter de Vos, ‘Reflections
on the introduction of a class action in South Africa’
1996
(4)
TSAR
639
at 642; Wouter de Vos, ‘Is a class action
a “classy act” to implement outside the ambit of the
Constitution?’
2012
TSAR
737.
12
See
the history and justification proffered by McLachlin CJC in
Western
Canadian Shopping Centres Inc v Dutton
[2001]
SCC 46
; 201 DLR (4
th
)
385 paras 19 – 29.
13
Section
173 of the Constitution. In
Ngxuza & others v Permanent
Secretary, Department of Welfare, Eastern Cape, and Another
2001
(2) SA 609
(E) Froneman J crafted an order to give effect to a class
action that in due course received the approval of this court.
Permanent Secretary, Department of Welfare, Eastern Cape &
another v Ngxuza & others
2001 (4) SA 1184
(SCA). However,
the appeal was argued on narrow grounds so that the general
questions facing us were not canvassed in any detail.
14
Hollick
v Toronto (City)
2001 SCC 68
;
[2001] 3 SCR 158
; 205 DLR (4
th
)
19 (SCC) para 14.
15
A
separate action instituted during the pendency of the class action
could be met with a plea of
lis pendens
.
16
Professor
Rachael Mulheron,
The Class Action in Common Law Legal Systems: A
Comparative Perspective
3.
17
Charles
Silver ‘Class Actions – Representative Proceedings’
5
Encyclopedia of Law and Economics
194.
18
See
also Debra Lyn Bassett ‘Constructing Class Action Reality’
2006
Brigham Young University Law Review
1415
at 1417-8.
19
This
approach largely accords with that advanced by the Legal Resources
Centre, which was admitted as
amicus
curiae
.
20
As
was done in
Permanent Secretary, Department of Welfare, Eastern
Cape & another v Ngxuza and Others
2001 (4) SA 1184
(SCA).
21
In
Ngxuza and Others v Permanent Secretary, Department of Welfare,
Eastern Cape & another
supra 624D-E Froneman J suggested
that ‘the possibility of unjustified litigation can be
curtailed by making it a procedural
requirement that leave must be
sought from the High Court to proceed on a representative basis
prior to actually embarking on
that road’. In
Firstrand
Bank Ltd v Chaucer Publications (Pty) Ltd
2008 (2) SA 592
(C)
para 26 Traverso DJP said that this suggestion should be followed in
the future, which no doubt explains the applications
in the two
cases that served before us.
22
In
Tiemstra v Insurance Corp of BC
(1996) 22 BCLR (3d) 49 (SC)
para 20 Esson CJSC said ‘class actions have the potential for
becoming monsters of complexity
and cost’.
23
Milton
Handler
25 Years of Antitrust
864-5 (1973) wrote: ‘Any
device which is workable only because it utilises the threat of
unmanageable and expensive litigation
to compel settlement is not a
rule of procedure – it is a form of legalised blackmail.’
24
Mulheron,
op cit (fn 13), 23-29. SA Law Commission Report, fn 9 ante, para
5.5.
25
Zweni
v Minister of Law and Order
1993 (1) SA 523
(A) 532I–533B.
26
Law
Commission Report, ante, para 5.6, pp 41-50 and recommendation 11.
27
Permanent
Secretary, Department of Welfare, Eastern Cape & another v
Ngxuza & others
2001 (4) SA 1184
(SCA) para 16.
28
Wal-Mart
Stores, Inc, Petitioner v Betty Dukes et al
131 S Ct 2541
at
2550.
29
See
Mulheron ante and Karlsgodt ante.
30
Bywater
v Toronto Transit Commission
(1998) 27 CPC (4
th
) 172
(Ont. Gen. Div.); [1998] OJ No 4913 (Ont Gen Div);
2038724
Ontario Ltd v Quizno’s Canadian Restaurant Corp
(2008) 89
OR (3d) 252 (SCJ).
31
Bray
v Hoffman-La Roche Ltd
[2003] FCA 1505.
32
The
minority dissented on the commonality issue but concurred on another
issue leading to the same result. In Canada the class
was held to be
over inclusive in a claim for compensation for wrongful dismissal
where it was defined as including persons dismissed
for just cause
(
Webb v K-Mart Canada Ltd
(1999), 45 O.R. (3d) 389 (Ont
S.C.J.)). An amendment to the class definition remedied the problem.
Certification of a class action,
in a claim against a school that it
misrepresented the quality of its education and thereby prejudiced
its graduates in seeking
employment, was refused on the basis that
the class included students who had found work after graduation
without problems. (
Mouhteros v DeVry Canada Inc
(1998), 41
O.R. (3d) 63 (Ont. Gen. Div.).
33
Emerald
Supplies Ltd & another v British Airways PLC
[2010] EWCA Civ
1284.
34
Para
63. In para 65 he added:
‘In brief, the
essential point is that the requirement of identity of interest of
the members of the represented class for
the proper constitution of
the action means that it must be representative at every stage, not
just at the end point of judgment.
If represented persons are to be
bound by a judgment that judgment must have been obtained in
proceedings that were properly
constituted as a representative
action
before
the
judgment was obtained. In this case a judgment on liability has to
be obtained before it is known whether the interests
of the persons
whom the claimants seek to represent are the same. It cannot be
right in principle that the case on liability
has to be tried and
decided before it can be known who is bound by the judgment. Nor can
it be right that, with Micawberish optimism,
Emerald can embark on
and continue proceedings in the hope that in due course it may turn
out that its claims are representative
of persons with the same
interest.’ Mulheron, ante, 329, fn 39 refers to a body of
Canadian authority that highlights the
same point.
35
Minister
of Safety and Security v Van Duivenboden
2002 (6) SA 431
(SCA)
para 22;
Trustees, Two Oceans Aquarium Trust v Kantey &
Templer (Pty) Ltd
2006 (3) SA 138
(SCA) paras 10 and 11.
36
Minister
of Law and Order v Kadir
[1994] ZASCA 138
;
1995 (1) SA 303
(A) 318E-I.
37
See
also
Axiam Holdings Ltd v Deloitte & Touche
2006 (1) SA
237
(SCA) para 25.
38
Telematrix
,
supra, paras 2 and 3. Footnote references omitted.
39
Dabelstein
& others v Lane and Fey
NNO
[2000] ZASCA 156
;
2001 (1) SA 1222
(SCA) at
1227H–1228A;
MV Pasquale Della Gatta; MV Filippo Lembo;
Imperial Marine Co v Deiulemar Compagnia Di Navigazione Spa
2012
(1) SA 58
(SCA) paras 22 to 24.
40
Eisen
v Carlisle & Jacquelin et al
[1974] USSC 108
;
417
US 156
(1974).
41
Hollick
v Toronto (City)
supra para 16.
42
Page
2552, fn 6.
43
Para
22.
44
Fose
v Minister of Safety and Security
[1997] ZACC 6
;
1997 (3) SA 786
(CC).
45
At
p 2551. Similarly in
Hollick v Toronto (City),
supra,
McLachlin CJC said ‘an issue will be common “only where
its resolution is necessary to the resolution of each
class member’s
claim” … Further, an issue will not be “common”
in the requisite sense unless the
issue is a “substantial …
ingredient” of each of the class members’ claims.’
(para 18)
46
Mulheron,
supra, 309–313.
47
Tribunal
determination in Pioneer para 81.
48
Curiously
the Tribunal found anti-competitive intent in the fact that in two
instances the seller
paid
the
purchaser to take over the bakery. One would have thought that more
indicative of the bakeries being unprofitable to the sellers.
49
Devenish
Nutrition Ltd & others v Sanofi-Aventis SA & others
[2007]
EWHC 2394
(Ch) para 19
50
The
Tribunal determination reflects that Pioneer produced 32 different
types of bread, not all of which are available throughout
the
country.
51
Steenkamp
NO v Provincial Tender Board, Eastern Cape
2006
(3) SA 151
(SCA) paras 19-21 and
Olitzki
Property Holdings v State Tender Board & another
2001
(3) SA 1247
(SCA) paras 12-14.
52
There
is nothing unusual in this.
Section 8(3)
of the
Extension of
Security of Tenure Act 62 of 1997
contains a similar provision.
53
Minister
of Safety and Security v Van Duivenboden
, supra, para
21;
Telematrix (Pty) Ltd t/a Matrix Vehicle Tracking v Advertising
Standards Authority SA
2006 (1) SA 461
(SCA) paras 12-16;
Trustees, Two Oceans Aquarium Trust v Kantey & Templer (Pty)
Ltd
supra, paras 10-12.
54
Fourway
Haulage SA (Pty) Ltd v South African National Roads Agency Ltd
[2008] ZASCA 134
;
2009
(2) SA 150
(SCA).
55
Rail
Commuters Action Group & others v Transnet Ltd t/a Metrorail &
others
[2004] ZACC 20
;
2005 (2) SA 359
(CC) para 15;
Steenkamp NO v
Provincial Tender Board, Eastern Cape
2007 (3) SA 121
(CC) paras
37-47 and 69-70.
56
Trotman
& another v Edwick
1951 (1) SA 443
(A) at 449B-C;
Ranger v Wykerd &
another
1977 (2) SA 976
(A) at 986D-E.
57
Mulheron
ante 426-434 and
The Modern C
y-près
Doctrine
Chapter 7, pp 215-252.
58
South
African Law Commission
The Recognition
of a Class Action in South African Law
(Working
Paper 57, 1995) para 5.38.
59
Recommendation
18, para 5.13.5, p 66.
60
Para
3 ante.
61
Mulheron,
supra 427. See the potential pitfalls with this remedy in her
The
Modern C
y-près Doctrine
218-222.
62
Modderfontein
Squatters, Greater Benoni City Council v Modderklip Boerdery (Pty)
Ltd (Agri SA and Legal Resources Centre, Amici
Curiae): President of
the Republic of South Africa & others v Modderklip Boerdery
(Pty) Ltd (Agri SA and Legal Resources
Centre, Amici Curiae)
2004
(6) SA 40
(SCA) para 42
.
63
Fose
v Minister of Safety and Security
[1997] ZACC 6
;
1997
(3) SA 786
(CC) paras 19 and 69.
64
Act
68 of 1969.
65
The
Merak: T B & S Batchelor & Co Ltd (Owners of Cargo on the
Merak) v Owners of SS Merak
[1965] 1 All ER 230
(CA) at 238
‘'(t)o bring suit, it is said, means to pursue the appropriate
remedy by the appropriate procedure'.
Dave Zick Timbers (Pty) Ltd
v Progress Steamship Co Ltd
1974 (4) SA 381
(D) at 384A-D;
IGI
Insurance Co Ltd v Madasa
1995 (1) SA 144
(TkA) at 147B-C.