SAFLII Note: Certain personal/private details of parties or witnesses have been redacted from this document
in compliance with the law and SAFLII Policy


CONSTITUTIONAL COURT OF SOUTH AFRICA


Case CCT 25/24

In the matter between:


BNP PARIBAS Applicant

and

COMPETITION COMMISSION OF SOUTH AFRICA Respondent


Case CCT 27/24

And in the matter between:


CREDIT SUISSE SECURITIES (USA) LLC Applicant

and

COMPETITION COMMISSION OF SOUTH AFRICA Respondent


Case CCT 30/24

And in the matter between:


COMPETITION COMMISSION OF SOUTH AFRICA Applicant

and

BANK OF AMERICA EUROPE DESIGNATED
ACTIVITY COMPANY First Respondent

BNP PARIBAS Second Respondent

JPMORGAN CHASE & COMPANY Third Respondent

JPMORGAN CHASE BANK N.A. Fourth Respondent

AUSTRALIA AND NEW ZEALAND BANKING
GROUP LIMITED Fifth Respondent

STANDARD NEW YORK SECURITIES
INCORPORATED Sixth Respondent

INVESTEC LIMITED Seventh Respondent

THE STANDARD BANK OF SOUTH AFRICA
LIMITED Eighth Respondent

NOMURA INTERNATIONAL PLC Ninth Respondent

STANDARD CHARTERED BANK Tenth Respondent

CREDIT SUISSE GROUP Eleventh Respondent

COMMERZBANK A.G. Twelfth Respondent

MACQUARIE BANK LIMITED Thirteenth Respondent

HSBC BANK PLC Fourteenth Respondent

CITIBANK N.A. Fifteenth Respondent

ABSA BANK LIMITED Sixteenth Respondent

BARCLAYS CAPITAL INCORPORATED Seventeenth Respondent

BARCLAYS BANK PLC Eighteenth Respondent

HSBC BANK USA, N.A. INCORPORATED Nineteenth Respondent

MERRILL LYNCH PIERCE FENNER AND
SMITH INCORPORATED Twentieth Respondent

BANK OF AMERICA N.A. Twenty-First Respondent

INVESTEC BANK LIMITED Twenty-Second Respondent

CREDIT SUISSE SECURITIES (USA) LLC Twenty-Third Respondent

NEDBANK GROUP LIMITED Twenty-Fourth Respondent

NEDBANK LIMITED Twenty-Fifth Respondent

FIRSTRAND LIMITED Twenty-Sixth Respondent

FIRSTRAND BANK LIMITED Twenty-Seventh Respondent

STANDARD AMERICAS INCORPORATED Twenty-Eighth Respondent



Neutral citation: BNP Paribas v Competition Commission of South Africa; Credit
Suisse Securities (USA) LLC v Comp etition Commission of South
Africa; Competition Commission of South Africa v Bank of
America Europe Designated Activity Company and Others [2026]
ZACC 28

Coram: Mlambo DCJ, Kollapen J, Majiedt J, Mathopo J, Mhlantla J,
Musi AJ, Rogers J, Savage J, Theron J and Tshiqi J


Judgment: Rogers J (unanimous)

Heard on: 19 to 21 August 2025

Decided on: 30 June 2026

Summary: Competition Act 89 of 1998 — Competition Tribunal — standard
of pleading — exception procedure

Res judicata and peremption — overlooking same

Referral to Competition Tribunal — whether joinder of firms after
referral permissible — whether fresh initiation needed

Section 4(1)(b) of Competition Act — single overarching
conspiracy — requirements for and pleading of same — personal
and subject matter jurisdiction

ORDER



In Case CCT 25/24 BNP Paribas v Competition Commission of South Africa:
On application for leave to appeal from the Competition Appeal Court (hearing an
appeal from the Competition Tribunal):
Leave to appeal is refused with costs, including the costs of two counsel.

In Case CCT 27/24 Credit Suisse Securities (USA) LLC v Competition Commission of
South Africa:
On application for leave to appeal from the Competition Appeal Court (hearing an
appeal from the Competition Tribunal):
1. Leave to appeal is granted.
2. The appeal succeeds.
3. Paragraph 2 of the order of the Competition Appeal Court, insofar as it
relates to the applicant in Case CCT 27/24 (the twenty-third respondent
in the Tribunal proceedings), Credit Suisse Securities (USA) LLC, is set
aside and replaced with the following order:
“(a) The appeal against the joinder of Credit Suisse Securities (USA)
LLC (CSS), forming part of paragraph A[1] of the Tribunal’s
order, succeeds.
(b) The Tribunal’s decision in that respect is set aside and replaced
with an order dismissing the Competition Commission’s
application to join CSS.”

In Case CCT 30/24 Competition Commission of South Africa v Bank of America Europe
Designated Activity Company and Others:
On application for leave to appeal from the Competition Appeal Court (hearing an
appeal from the Competition Tribunal):

1. The applicant, the Competition Commission (Commission), is granted
leave to appeal.
2. The fourteenth respondent, HSBC Bank plc (HBEU), is granted leave to
cross-appeal.
3. The Commission’s appeal is upheld in relation to the fourth respondent,
JPMorgan Chase Bank N.A. (JPM Bank), and the twenty -eighth
respondent, Standard Americas Incorporated (SAI).
4. The Competition Appeal Court’s decision in respect of JPM Bank and
SAI is set aside and replaced with orders dismissing those parties’ appeals
against the Tribunal’s decision.
5. Save as aforesaid, the Commission’s appeal is dismissed.
6. HBEU’s cross-appeal is dismissed.



JUDGMENT




ROGERS J (Mlambo DCJ, Kollapen J, Majiedt J, Mathopo J, Mhlantla J, Musi AJ,
Savage J, Theron J and Tshiqi J concurring):


Introduction
[1] Before us are three applications for leave to appeal and an application for leave
to cross-appeal a judgment of the Competition Appeal Court (CAC) , handed down on
8 January 2024.1 The prelude to that appeal was an earlier judgment of the CAC in the
same litigation, delivered on 28 February 202 0.2 I shall refer to the earlier and later
CAC judgments as CAC I and CAC II. The appeal proceedings in CAC I and CAC II
were against decision s of the Competition Tribunal (Tribunal) handed down on

1 Competition Commission of South Africa v Bank of America Merrill Lynch International [2024] ZACAC 1;
[2024] 1 CPLR 1 (CAC).
2 Competition Commission v Bank of America Merrill Lynch International Limited [2020] ZACAC 1; [2020] 1
CPLR 26 (CAC); 2020 (4) SA 105 (CAC).

ROGERS J
6
12 June 20193 and 30 March 20234 respectively. I shall refer to th em as TRIB I and
TRIB II.

Procedural history
Initiation and the February 2017 referral
[2] The main litigation arises from a complaint initiated 5 by the Competition
Commission (Commission) in April 2015, amended in August 2016, and referred to the
Tribunal in February 2017. The Commission alleged that a number of banks – 18 were
identified in the February 2017 referral – were guilty of conduct prohibited by
section 4(1)(b) of the Competition Act6 (Act). The accusation was that over the period
from 2007 until at least September 2013 they colluded to manipulate the exchange rate
between the United States Dollar (USD) and the South African Rand (ZAR).

[3] One of the initial respondents, the sixteenth respondent, Absa Bank Limited
(Absa), and the seventeenth and eighteenth respondents, Barclays Capital Incorporated

3 Competition Commission of South Africa v Bank of America Merrill Lynch International Limited ZACT 50;
[2020] 1 CPLR 205 (CT).
4 Competition Commission of South Africa v Bank of America Merrill Lynch International Designated Activity
Company [2023] ZACT 26.
5 Initiation is a term of art in the Competition Act 89 of 1998. In terms of section 49B(1) “the Commissioner may
‘initiate’ a complaint against an alleged prohibited practice”. Once there has been an initiation, the Commissioner
must direct an inspector to investigate the complaints. After initiating a complaint, the Commissioner may in
terms of section 50(1) refer the complaint to the Tribunal. Upon referral, the chairperson of the Tribunal must
publish in the Government Gazette the name of the respondent and the nature of the conduct that is the subject of
the referral – section 51(3).
6 89 of 1998. Section 4(1)(b) provides:
“(1) An agreement between, or concerted practice by, firms, or a decision by an association
of firms, is prohibited if it is between parties in a horizontal relationship and if—
. . .

of firms, is prohibited if it is between parties in a horizontal relationship and if—
. . .
(b) it involves any of the following restrictive horizontal practices:
(i) directly or indirectly fixing a purchase or selling price or any other
trading conditions;
(ii) dividing markets by allocating customers, suppliers, territories, or
specific types of goods or services; or
(iii) collusive tendering.”

ROGERS J
7
and Barclays Bank plc (collectively Barclays), applied for and were granted leniency7
on the basis that they would give their cooperation to the Commission in prosecuting
the complaint. The fifteenth respondent, Citibank NA8 (Citibank), reached a settlement
with the Commission.

[4] This left 14 of the original 18 respondents as active parties in the referral
proceedings: the first respondent, Bank of America Europe DAC, 9 formerly Bank of
America Merrill Lynch International DAC (BAMLI); the second respondent, BNP
Paribas (BNP); the third and fourth respondents, JPMorgan Chase & Co (JPM Co) and
JPMorgan Chase Bank NA (JPM Bank); the fifth respondent, Australia and New
Zealand Banking Group Limited (ANZ); the sixth respondent, Standard New York
Securities Incorporated (SNY); the seventh respondent, Investec Limited; the eighth
respondent, The Standard Bank of South Africa Limited (SBSA); the ninth respondent,
Nomura International plc (Nomura); the tenth respondent, Standard Chartered Bank
(SCB); the eleventh respondent, Credit Suisse Group (CSG); the twelfth respondent,
Commerzbank AG (Commerzbank ); the thirteenth respondent, Macquarie Bank
Limited (Macquarie); and the fourteenth respondent, HSBC Bank plc (HBEU).

[5] The Commission delivered several supplementary affidavits in an attempt to
meet the respondents’ criticisms of the February 2017 referral affidavit. This did not
satisfy the m, and so they filed exceptions. Another development was that in
January 2018 the Commission served an application to join another f ive banks (first
joinder application): as the nineteenth respondent, HSBC Bank USA NA (HBUS); as
the twentieth respondent, Merrill Lynch Pierce Fenner and Smith Incorporated (MLP);
as the twenty-first respondent, Bank of America NA (BANA); as the twenty-second

7 Although express statutory provision for the granting of leniency, by way of section 49E, was only introduced

into the Act in February 2019, the Commission had in February 2004 published and acted upon a corporate
leniency policy. The object of such a policy is to encourage a participant in prohibited conduct to “blow the
whistle” by being the “first through the door”.
8 NA is an acronym for National Association, a designation for national banks under American law.
9 DAC is an acronym for Designated Activity Company, a form of corporation under Irish law.

ROGERS J
8
respondent, Investec Bank Limited (IBL); and as the twenty-third respondent, Credit
Suisse Securities (USA) LLC10 (CSS).

[6] BAMLI, MLP and BANA are part of the Bank of America (BoA) group. Since
the Commission’s allegations against these three entities are in the alternative, I shall
refer to them collectively as BoA, save where it is necessary to distinguish between
them. The same applies to JPM Co and JPM Bank, and to HBEU and HBUS, to which
I shall refer collectively as JPM and HSBC respectively.

[7] In the referral affidavit , the Commission distinguished between foreign
respondents that were “ pure peregrini” (pure foreigners) and those that were “local
peregrini” (local foreigners). The pure peregrini were foreign firms without any branch
or representative office in South Africa. The active respondents that fell into th at
category were BAMLI, JPM Co, ANZ, SNY, Nomura, Macquarie, HBUS, MLP and
CSS. The remaining foreign respondents – BNP, JPM Bank, SCB, CSG,
Commerzbank, HBEU and BANA – were local peregrini. As at February 2017 the
only active respondent that was an incola (local, i.e. South African) was SBSA. The
other South African bank, Absa, was an inactive respondent by virtue of leniency.

[8] One of the issues in the exceptions was whether the Commission had pleaded
facts to show that the Tribunal had personal jurisdiction over the peregrini respondents.
In civil proceedings in this country there is a distinction between a court’s subject matter
jurisdiction and its personal jurisdiction. For example, if a plaintiff claims damages of
R1 million caused by a delict committed in Cape Town, the Western Cape Division of
the High Court (but not , for example, the Cape Town Magistrate’s Court or the
Competition Tribunal) would have subject matter jurisdiction to try the case. It would
also have personal jurisdiction if the defendant was an incola of South Africa. But if
the defendant was a peregrinus, something more would be needed to give the

the defendant was a peregrinus, something more would be needed to give the
High Court personal jurisdiction. There would need to be, for example, an attachment

10 LLC is an acronym for Limited Liability Company, a form of private limited corporation under American law.

ROGERS J
9
of property belonging to the defendant in South Africa in order to satisfy the
requirement of effectiveness.

[9] The Commission’s primary position was that, in the case of proceedings under
the Act, the common law requirements of personal and subject matter jurisdiction are
wholly displaced by section 3(1) of the Act . That subsection states in relevant part:
“This Act applies to all economic activity within, or having an effect within, the
Republic”. The Commission contended that the respondents’ conduct had an effect
within South Africa, which was enough to subject all of them to the Tribunal’s
jurisdiction, even those respondents without any presence in this country.

TRIB I
[10] The exceptions and joinder application were argued in the Tribunal in mid-2018.
The Tribunal gave its decision in June 2019, TRIB I . The Tribunal rejected the
Commission’s argument that section 3(1) did away with the need for personal
jurisdiction. The Tribunal held, further, that the common law had not evolved to the
point where personal jurisdiction over the pure peregrini could be recognised. Despite
this conclusion, the Tribunal considered that it could issue a declaratory order
pronouncing upon the conduct of the pure peregrini, because this would not
compromise the principle of effectiveness. It would also ensure that the declaration of
offending conduct would not be under -inclusive in the context of any follow -on civil
claims brought against respondents over whom there was personal jurisdiction.11 As to
the local peregrini, the Tribunal considered that it had personal jurisdiction on
conventional grounds.12

[11] As to subject matter jurisdiction, the Tribunal addressed the nature of the effects
that needed to be established to satisfy section 3(1). The Tribunal regarded this as
relevant to the local peregrini. This was presumably for the reason that, b ecause their

11 TRIB I above n 3 at paras 35-66.
12 Id at paras 67-83.

ROGERS J
10
acts of collusion had not taken place in South Africa, subject matter jurisdiction
depended upon their foreign conduct having an “effect” in South Africa. In the case of
incolae (locally domiciled) banks (such as SBSA) , subject matter jurisdiction did not
depend on “effect” in South Africa because the alleged conduct was “economic
activity” that took place in South Africa.

[12] After referring to antitrust jurisprudence on effects-based jurisdiction in the
United States (US),13 the European Union (EU)14 and South Africa,15 the Tribunal held
that “effect” in section 3(1) connoted “qualified effects”, the test being whether it was
foreseeable that the prohibited conduct would have a direct or immediate, and
substantial, effect in South Africa (for the sake of brevity, I shall call this the QE test).16
After analysing the 2017 referral affidavit as supplemented, the Tribunal concluded that
the Commission had to make additional allegations to satisfy the QE test.17

[13] The Tribunal then considered the alleged deficiencies in the pleading of the
alleged conspiracy. It was unclear from the Commission’s affidavits, the Tribunal
thought, whether the Commission was relying on a single conspiracy or multiple mini-
conspiracies. In the Commission’s oral argument, however, it became clear that the
Commission was relying on a single overarching conspiracy (SOC), with instances of
multilateral or bilateral collusion being acts evidencing the implementation of the SOC.
The main and supplementary affidavits in the 2017 referral, in the Tribunal’s view, were
inconsistent and appeared to expose various respondents to alternative cases based on
smaller conspiracies. The Tribunal thus took what it called the “unusual approach” of

13 US Foreign Trade Antitrust Improvement Act of 1982 and US Department of Justice and Federal Trade
Commission Antitrust Guidelines for International Enforcement and Cooperation, 13 January 2017 at 21.

14 Intel Corporation v European Commission Case T-286/09; [2014] 5 CMLR 9 (Intel) at paras 231 -2; Intel
Corporation v European Commission Case C-413/14 P; [2017] 5 CMLR 18 at para 49; and Gencor v Commission
Case T-102/96 [1999] ECR II-753 at para 90.
15 American Natural Soda Ash Corporation v Competition Commission of South Africa [2002] ZACAC 5; 2003
(5) SA 663 (CAC); [2005] 1 CPLR 18 (CAC) (Ansac I) at para 18 and American Natural Soda Ash Corporation
v Competition Commission of South Africa [2005] ZASCA 42; [2005] 1 CPLR 1 (SCA); [2005] 3 All SA 1 (SCA);
2005 (6) SA 158 (SCA); 2005 (9) BCLR 862 (SCA) at para 29.
16 TRIB I above n 3 at paras 84-100.
17 Id at paras 101-13.

ROGERS J
11
requiring the Commission, in a superseding referral affidavit (superseding affidavit), to
confine its case to the SOC described by its counsel in argument.18

[14] In regard to the joinder application, I BL, an incola, was joined as the
twenty-second respondent without opposition. The Tribunal deferred its decision on
the joinder of the other four parties until after the filing of the superseding affidavit. In
respect of BANA, a local peregrinus, t he affidavit would need to allege qualified
effects. In relation to HBUS, MLP and CSS, pure peregrini, the affidavit would need
to confine the claim to a limited declaratory order. The Tribunal considered various
other exceptions, upholding some, rejecting others. Its order required the Commission
to file its superseding affidavit within 40 days.

CAC I
[15] The Commission did not file a superseding affidavit , because TRIB I was
appealed to the CAC. The pure peregrini appealed the Tribunal’s decision that it could
grant a declaratory order in respect of their conduct. Three of the respondents appealed
the Tribunal’s order deferring a decision on the joinder application. The Commission
cross-appealed the Tribunal’s decision on the meaning of section 3(1), both in relation
to personal jurisdiction and subject matter jurisdiction. The outcome of the appeals and
cross-appeals was CAC I, delivered in February 2020.

[16] The CAC agreed with the Tribunal that section 3(1) did not do away with th e
requirement of personal jurisdiction . The CAC , though, regarded as illogical the
Tribunal’s view that it could grant a declaratory order in respect of pure peregrini
despite the absence of personal jurisdiction.19 Unlike the Tribunal, however, the CAC
considered that an appropriate development of the common law on personal jurisdiction
would leave open the possibility that the Commission could, by making further
allegations, establish personal jurisdiction over the pure peregrini.

18 Id at paras 113-84.

18 Id at paras 113-84.
19 CAC I above n 2 at para 61.

ROGERS J
12

[17] Citing domestic20 and foreign21 cases, the CAC said there was overwhelming
authority in favour of the extraterritorial application of competition legislation. 22
Support for a limited development of our common law on personal jurisdiction so as to
render it congruent with section 3(1) and the purposes of the Act could , in the CAC’s
view, draw inspiration from Bid Industrial Holdings 23 and Multi-Links.24 In a
competition case, personal jurisdiction would exist if there were “adequate connecting
factors” between the Commission’s complaint and the Tribunal as a forum (for the sake
of brevity, I shall call this the ACF test). Appropriateness and convenience were
relevant considerations. The question was whether the Tribunal had a real and
substantial connection with the suit and whether there were relevant connecting factors
tying the suit to the Tribunal as a forum.25

[18] The CAC said that if the Commission could establish an SOC to manipulate the
USD/ZAR exchange rate and that the SOC involved pure peregrini, local peregrini and
incolae, this could indicate that there were adequate connecting factors sufficient to
found the Tribunal’s jurisdiction.26 From this statement it is apparent that the CAC did
not think it sufficient that the alleged conspiracy targeted the USD/ZAR exchange rate.
It was required, in addition, that the conspiracy should have South African participants.
The CAC also accepted that the test for subject matter jurisdiction was the QE test .27

20 Ansac I above n 15 at paras 16-17.
21 United States v Aluminum Company of America 148 F 2d 416 (2nd Cir 1945); F Hoffmann-La Roche Ltd v
Empagran SA [2004] USSC 2381; 542 US 155; A. Ahlstr öm Osakeyhtiö v Commission of the European
Communities [1988] 4 CMLR 901; [1988] EUECJ C-129/85; [1988] ECR 5193 – opinion of the Advocate General
at para 57; Intel above n 14 – opinion of the Advocate General at para 283-5.
22 CAC I above n 2 at paras 36-44.

22 CAC I above n 2 at paras 36-44.
23 Bid Industrial Holdings (Pty) Ltd v Strang [2007] ZASCA 144; [2008] 2 All SA 373 (SCA); 2008 (3) SA 355
(SCA) at paras 54-6.
24 Multi-Links Telecommunications Ltd v Africa Prepaid Services Nigeria Ltd [2013] ZAGPPHC 261; [2013] 4
All SA 346 (GNP); 2014 (3) SA 265 (GP) at para 23.
25 CAC I above n 2 at paras 45-57.
26 Id at para 56.
27 Id at paras 54 and 58.

ROGERS J
13
Like the Tribunal, the CAC considered it sensible to defer a decision on the joinder
application (except for IBL) until after the Commission filed the superseding affidavit.

[19] Overall, the result was that both the banks and the Commission achieved some
success. The CAC adapted the Tribunal’s order so as to accommodate the CAC’s
conclusion on personal jurisdiction. In view of some of the arguments, it is necessary
to quote the relevant parts of the order (I substitute the banks’ abbreviated names for
those used in the order; the numbers that appear in brackets were inserted by the CAC
to indicate the respondent number of the bank in question):

“3.1 The applications for the dismissal of the complaint referral brought by the pure
peregrini, BAMLI (1); JPM Co (3); ANZ (5); SNY (6); Nomura (9), Macquarie
(13); HBUS (19); MLP (20) and C SS (23) are dismissed subject to the
following.
3.1.1 The Commission must file a new referral affidavit to substitute for and
replace all the complaint referral affidavits. This affidavit must be
filed within 40 business days of this order.
3.1.2 The respondents will only be required to file their answers to the new
referral affidavit. The answers must be filed within 20 days of the
service of the new referral affidavit.
3.2 The new referral affidavit applies to the parties referred to in paragraph 1 of
the substitute order.28 It must:
3.2.1 In the case of all the named respondents set out the facts the
Commission relies on to allege that it was foreseeable that the
impugned conduct would have a direct or immediate, and substantial
effect in the Republic;
3.2.2 Confine the case to a single overall conspiracy (SOC), provided,
subject to 3.4.3 below, that the Commission is not restricted from
alleging that this may be founded on an agreement, arrangement or
concerted practice;

28 I understand the CAC to have meant that the new referral affidavit, in addition to applying to the other

respondents, was also to apply to the pure peregrini listed in paragraph 3.1 of the CAC’s order.

ROGERS J
14
3.2.3 Indicate whether the same facts are relied on for proof of the concerted
practice or allege any different fact if they are not;
3.2.4 Allege whether its case for an SOC relies on proof of an express
agreement or arrangement or whether this is an inference based on
facts, if the latter, allege in general terms what those facts are;
3.2.5 Provide each respondent with a date, or period, in which they are
alleged to have joined the SOC or deemed to have joined the SOC;
3.2.6 Provide the facts that are relied on to prove that the particular
respondent joined or had joined the SOC;
3.2.7 If the SOC ceased;
3.2.7.1 provide what dates the SOC is alleged to have ceased;
3.2.7.2 what facts are relied on for establishing that the conduct had
then ceased; and
3.2.7.3 whether all the respondents remained participants in the SOC
on that date; and, if not, when the respective respondent/s
exited.
3.2.8 If the SOC is still alleged to be ongoing;
3.2.8.1 what facts this is based on; and
3.2.8.2 whether all the respondents are still part of it, if not when the
respective respondent/s exited;
3.2.8.3 in relation to the relationship between the respondent bank s29
and their respective traders;
3.2.8.3.1 is it alleged that some traders acted for more than
one respondent at the same time? If so, details
should be provided;
3.2.8.3.2 if a trader ceased to act for a respondent’s bank, did
this end the respondent’s participation in the SOC or
if not, on what basis is it alleged that the
respondent’s participation continued?

29 The order reads “bank”, but “banks” must have been intended.

ROGERS J
15
3.2.8.3.3 is it alleged that all the traders named as participants
in paragraph 40 of the December affidavit were
so-called active participants or were some so-called
passive participants.
3.3 The new referral affidavit must in addition:
3.3.1 in the case of all of the named respondents set out the facts on which
the Commission relies to allege that there are adequate connecting
factors between the parties and the jurisdiction of the Competition
Tribunal; sufficient to establish personal jurisdiction against all named
respondents.”

[20] Paragraphs 3.1.1 and 3.1.2 and the whole of paragraph 3.2 of the CAC’s order
followed the Tribunal’s order , with one qualification . Paragraph 3.2.1 of the CAC’s
order required the Commission to plead qualified effects (for subject matter jurisdiction)
in relation to each respondent, whereas the Tribunal’s corresponding order had been
confined to the local peregrini. Paragraph 3.3 of the CAC’s order did not have a
counterpart in the Tribunal’s order ; it required the Commission to plead adequate
connecting factors (for personal jurisdiction) in relation to each respondent.

The superseding referral affidavit of June 2020
[21] None of the parties appealed CAC I to this Court. In June 2020 the Commission
served its superseding affidavit in purported compliance with CAC I. The Commission
added another five respondents to the referral: as twenty-fourth respondent, Nedbank
Group Limited (Ned Group); as twenty-fifth respondent, Nedbank Limited (Nedbank);
as twenty-sixth respondent, FirstRand Limited (FirstRand); as twenty-seventh
respondent, FirstRand Bank Limited (FRB); and as twenty-eighth respondent, Standard
Americas Incorporated (SAI). Save for SAI, which is a pure peregrinus, these new
respondents are South African banks.

[22] It is clear from the superseding affidavit that the Commission was aiming to
comply with CAC I. On personal jurisdiction and the ACF test, i t pleaded that the

comply with CAC I. On personal jurisdiction and the ACF test, i t pleaded that the
Tribunal had jurisdiction over the pure peregrini, because they participated in an SOC

ROGERS J
16
that involved the exchange rate of the ZAR and which included South African
respondents. In the case of the local peregrini, the Commission added, as additional
connecting factors, that some of them had branches in South Africa , were authorised
under our law to deal in foreign currency and carried on business in this country; and
that others had representative offices in South Africa and carried on business here.

[23] On subject matter jurisdiction and the QE test, the Commission pleaded in
relation to each respondent that, in implementing the SOC, they engaged in conduct that
had “direct/immediate, substantial and foreseeable consequences upon the economy of
South Africa and the welfare of South African consumers and an effect within
South Africa for the purposes of section 3(1) of the Act”. Later in the superseding
affidavit the Commission devoted many paragraphs to the effect of the conduct, starting
with a paragraph alleging that the SOC “had a direct or immediate, and substantial effect
in the Republic and it was foreseeable that the impugned conduct would, or had the
potential to, have such an effect”.

[24] Save for the two Investec entities, the active respondents again filed exceptions,
some by way of dismissal or objection applications on motion supported by affidavit.
In regard to personal jurisdiction over the pure peregrini, the adequacy of the
Commission’s pleading of personal jurisdiction was closely tied to the adequacy of its
pleading of the pure peregrini’s participation in an SOC that included South African
banks. This was because participation in an SOC that included South African banks
was a crucial “connecting factor” alleged by the Commission for purposes of personal
jurisdiction. In the case of other respondents, there were also contentions that the
superseding affidavit did not adequately plead facts to show that they participated in the
SOC. The pleaded facts, they contended, were either benign or, at best for the

SOC. The pleaded facts, they contended, were either benign or, at best for the
Commission, showed limited collusion not forming part of the alleged SOC. Some of
the respondents also contende d that the Commission had not adequately pleaded
qualified effects to establish subject matter jurisdiction.

ROGERS J
17
[25] In September 2020 the Commission served its second joinder application , this
time for the joinder of the twenty-fourth to twenty-eighth respondents. This joinder
application was conditional upon it being found that the Commission was not entitled
to add additional respondents when delivering the superseding affidavit . The two
FirstRand entities did not oppose their joinder, but the two Nedbank entities and SAI
did.

TRIB II
[26] The Tribunal heard argument over a number of days in November and
December 2021. It delivered its decision, TRIB II, in March 2023. The Tribunal’s view
was that extraneous facts alleged by respondents in dismissal and objection applications
should be left out of account, and that such applications should be treated as pure
exceptions.30 The facts alleged in the superseding affidavit were to be assumed, for
purposes of the exceptions, to be true. The Commission was entitled to the benefit of
the most favourable reading of which the affidavit was reasonably capable. Although
in general the Tribunal follows High Court exception principles, its guiding principle,
the Tribunal emphasised, is fairness. This is because of the unique nature of referral
proceedings: a mix of motion and trial procee dings; subject matter involving the
intersection of law and economics; and the fact that the Tribunal has inquisitorial
powers.31

[27] The Tribunal said that it had a wide discretion in the conduct of its proceedings.
Its approach was not overly technical. It had to conduct cases expeditiously and
informally. The Commission was accusing the respondents of conduct regarded as the
most egregious in competition law, relating to manipulating the ZAR exchange rate
which had a central and crucial role in South Africa’s economy.32


30 TRIB II above n 4 at paras 235-6.
31 Id at paras 49-60.
32 Id at paras 106-21 and 229-31.

ROGERS J
18
[28] As to the adequacy of the pleading of the SOC, the Tribunal said that an SOC
did not require that all members of the conspiracy meet at the same time or that each
member must have met every other member. What was required was contact between
firms, directly or through an intermediary, and a common anti-competitive objective by
which the participants considered themselves bound.33 The Tribunal concluded that the
superseding affidavit pleaded sufficient facts to make out a prima facie case of an SOC
between foreign and local banks. 34 The Tribunal also was satisfied that the affidavit
pleaded sufficient facts to meet the QE test for subject matter jurisdiction.35

[29] The Tribunal referred to rule 15(2) of the Tribunal Rules (TR), which states that
a referral must be supported by an affidavit setting out in numbered paragraphs “a
concise statement of the grounds of the complaint” and “the material facts or the points
of law relevant to the complaint and relied on by the Commission”. This standard, in
the Tribunal’s view, had been met.36 The Tribunal undertook a detailed examination of
the respondents’ exceptions, finding in each case that the superseding affidavit passed
muster.37

[30] Some of the respondents raised, as a ground of exception, that on the pleaded
facts the prohibited conduct had ceased more than three years before the initiation of
the complaint in April 2015. For this reason, so they argued, proceedings against them
were time-barred under section 67(1) of the Act. 38 The Tribunal rejected this ground.
The Commission had pleaded that it did not know the cessation date because this was
within the knowledge of the banks. Whether the conduct had ceased more than three

33 Id at paras 118-19, citing Whish and Bailey Competition Law 8 ed (OUP, Oxford 2015) at 107-10.
34 TRIB II id at paras 123-43 and
35 Id at paras 146-62.
36 Id at paras 205-9.
37 Id at paras 281-355.
38 Section 67(1) reads:

36 Id at paras 205-9.
37 Id at paras 281-355.
38 Section 67(1) reads:
“A complaint in respect of a prohibited practice that ceased more than three years before the
complaint was initiated may not be referred to the Competition Tribunal.”

ROGERS J
19
years before the initiation had to be pleaded by the respondents and determined after
the leading of evidence.39

[31] Some of the banks that had not been named in the February 2017 referral
contended that there could be no referral in respect of them in the absence of an initiation
specifically identifying them. The superseding affidavit did not disclose that this had
happened. The banks in question invoked the judgment of the Supreme Court of Appeal
(SCA) in Woodlands.40 Some of the banks also argued that it was not permissible to
join a respondent once a complaint had been referred to the Tribunal. For this
proposition they relied on Woodlands and the CAC’s judgment in Loungefoam.41

[32] The Tribunal considered that the SCA’s judgment in Yara42 and this Court’s
judgment in Pickfords43 were a complete answer to these contentions. An initiation did
not require formalities, it could even be done tacitly. An initiation had to identify the
prohibited practice. An initiation was valid even if the Commission did not then know
who all the p articipants were. A new initiation was not needed just to add further
respondents. What Woodlands had found impermissible was the initiation of a
complaint against a whole industry in the absence of material evidence to support a
belief that the conduct was being perpetrated by the whole industry.44

[33] Another ground of exception addressed by the Tribunal was a complaint that the
superseding affidavit did not in all respects meet the pleading requirements set out in
the CAC I order. On the interpretation of court orders, the Tribunal took guidance from

39 TRIB II above n 4 at paras 167-73.
40 Woodlands Dairy (Pty) Ltd v Competition Commission [2010] ZASCA 104; 2010 (6) SA 108 (SCA); [2011] 3
All SA 192 (SCA).
41 Loungefoam (Pty) Ltd v Competition Commission South Africa ; Feltex Holdings (Pty) Ltd v Competition
Commission South Africa [2011] ZACAC 4; [2011] 1 CPLR 19 (CAC).

Commission South Africa [2011] ZACAC 4; [2011] 1 CPLR 19 (CAC).
42 Competition Commission v Yara (SA) (Pty) Ltd [2013] ZASCA 107; [2013] 2 CPLR 351 (SCA); [2013] 4 All
SA 302 (SCA); 2013 (6) SA 404 (SCA).
43 Competition Commission of South Africa v Pickfords Removals SA (Pty) Ltd [2020] ZACC 14; [2020] 1 CPLR
1 (CC); 2020 (10) BCLR 1204 (CC); 2021 (3) SA 1 (CC).
44 TRIB II above n 4 at paras 174-89.

ROGERS J
20
this Court’s decision in S.O.S.45 The context of TRIB I and CAC I, the Tribunal said,
had been a matter of pleading. The CAC I order should not be minutely interpreted as
if it were a legislative checklist. It was enough that the superseding affidavit
demonstrated what the case was about and enabled the banks to answer. That test was
met. Even if the superseding affidavit did not contain all the details required by the
CAC I order, this did not fetter the Tribunal’s procedural discretion in terms of
sections 5546 and 58(1)(c)47 of the Act read with TR 55.48 Non-compliance with CAC I
was not, the Tribunal concluded, a self-standing ground of objection.49

[34] On the joinder applications, the Tribunal said that its joinder powers under
TR 45(1) were discretionary. As a general proposition, it was disinclined to dismiss the
joinder applications, since a dismissal at an early stage of the proceedings might result
in the unintended consequence of letting an alleged cartelist off the hook. 50 The
Tribunal dismissed each of the joinder respondents’ grounds of opposition, which
included issues of jurisdiction and initiation dealt with earlier in the Tribunal’s decision.

[35] The Tribunal made an order granting all the joinders and dismissing all the
exceptions. The respondents were ordered to file their answering affidavits within
40 days. If they persisted with any grounds of objection or exception not dealt with in
the Tribunal’s reasons, they had to do so in their answering affidavits and plead over on
the merits.


45 SOS Support Public Broadcasting Coalition v South African Broadcasting Corporation (Soc) Ltd [2018] ZACC
37; [2018] 2 CPLR 411 (CC); 2018 (12) BCLR 1553 (CC); 2019 (1) SA 370 (CC).
46 Section 55 confers on the presiding Tribunal member the power to determine any matter of procedure, with due
regard to the circumstances of the case. In terms of the same section, the Tribunal is granted the power to condone
technical irregularities.

technical irregularities.
47 The Tribunal referred to section 59, but that seems to have been an error. Section 58(1)(c) gives the Tribunal
the power to condone non-compliance on good cause shown.
48 TR 55 gives effect to sections 55 and 58(1)(c).
49 TRIB II above n 4 at paras 190-204.
50 Id at paras 234-5.

ROGERS J
21
CAC II
[36] All the banks that lost in the Tribunal appealed to the CAC. Some of the banks
also launched review applications to set aside TRIB II. The appeals and review s were
argued over four days in November 2023. During the hearing SCB reached a settlement
with the Commission and thus ceased to be an active respondent. The CAC delivered
judgment, CAC II, in January 2024.

[37] The CAC rejected a preliminary argument by the Commission that TRIB II was
not appealable. It did so on the basis that the dismissal of an exception is appealable
where the exception contests the jurisdiction of the forum.51

[38] The CAC accepted the Commission’s argument on the three requirements for an
SOC, derived from EU case law, 52 namely (1) a common anti -competitive objective,
being the existence of an overall plan pursuing a common economic objective;
(2) participation, being each firm’s intentional contribution by its own conduct to the
common objective pursued by all the participants; and (3) that the firm was either aware
of the actual conduct planned or put into effect by other firms in pursuit of the same
objective or could reasonably have foreseen it and was prepared to take the risk.53 (As
to the third requirement, I shall henceforth refer to knowledge which a firm had or could
reasonably have had as actual or constructive knowledge.)

[39] The CAC pointed out that, according to Team Relocations,54 a series of acts or
continuous conduct may constitute an infringement even where one or several elements
of that series of acts or continuous conduct could in themselves and in isolation
constitute an infringement. The responsibility of a firm that parti cipates in a “single
and continuous infringement” (i.e. an SOC) is—

51 CAC II above n 1 at paras 52-60.
52 Commission v Anic Partecipazioni [1999] ECR I-4125; [1999] EUECJ C-49/92 and Team Relocations NV v
European Commission [2013] EUECJ C-444/11 (Team Relocations).
53 CAC II above n 1 at para 28.
54 Above n 52.

ROGERS J
22

“limited, with regard to the conduct planned or put into effect by the other participants
and of which it was or should have been aware, to the actions which occurred during
the period of participation of that undertaking in that infringement.”55

This formulation acknowledged that a firm “may be held responsible for a single and
continuous infringement even if it does not participate in all of the offending conduct
of which it is made up”. Also, the firm need not have begun its intentional contribution
at the very start of the infringement nor need it have pursued the common objective in
ways identical to those put into effect when the infringement began.56

[40] The CAC said that pleading an SOC is an “onerous exercise” which required, in
this case, that the Commission provide—

“plausible evidence that, even if not a participant in all of the events which made up
the SOC, a respondent bank’s conduct, based essentially on that of its traders,
constitutes sufficient evidence to justify that there was an intentional contribution to
the common objectives of the conspiracy and that the respondent bank was aware or
should have been aware of the offending conduct of the initial participants and hence
was party to the infringement by way of an SOC.”57

[41] The CAC reminded itself that in CAC I , it had rejected the Commission’s
argument concerning section 3(1). In regard to personal jurisdiction, CAC I had laid
down the ACF test. CAC I had foreshadowed the possibility that an SOC targeting the
USD/ZAR exchange rate and that had local banks as participants could satisfy the ACF
test in regard to pure peregrini. But it was then necessary, the Court emphasised in
CAC II, that the pleaded case should show that all of the banks were connected to the
SOC, having regard to the principles set out in Team Relocations.58


55 CAC II id at para 33.
56 Id at paras 29-34, quoting Team Relocations above n 52 at paras 49, 52, 55 and 56.
57 CAC II id at para 36.

57 CAC II id at para 36.
58 Id at paras 38-9.

ROGERS J
23
[42] An enquiry, therefore, into whether TRIB II was right to find that the
Commission had sufficiently pleaded personal jurisdiction in respect of the pure
peregrini required, said the Court in CAC II—

“an examination as to whether there was an SOC between the respondent banks of a
kind that reveals not simply that the Rand was the subject of the trades but that the
conspiracy was sufficiently connected to South Africa because, by virtue of it being an
overall conspiracy, foreign banks had effectively entered into ‘business with South
African banks’. This provided sufficient connection between these respondent banks
and South African jurisdiction to sustain a finding of personal jurisdiction in the
circumstances of the case. This concept of personal jurisdiction as developed by this
Court was predicated on an economy where business is no longer based on bricks and
mortar but rather on modern technology, which has created the conditions for a global
economy where national borders are transcended by virtue of technological
development. It follows that to sustain an argument that there were sufficient
connecting factors, the need is to provide clear evidence of linkages to the South
African banks as part of an overall conspiracy and which thus linked the incolae banks
with the peregrini banks.”59

[43] The Court then embarked on a bank-by-bank analysis to determine whether the
superseding affidavit contained sufficient allegations to sustain the conclusion that the
bank in question had participated in the alleged SOC. This analysis was done for all
the banks, not only the pure peregrini. In part, therefore, it was an assessment as to
whether the Commission had made out a cause of action against each bank in respect
of the subject matter of the case, namely an SOC. But because the alleged existence of
an SOC with South African participants was an important connecting factor for personal

an SOC with South African participants was an important connecting factor for personal
jurisdiction over the pure peregrini, the analysis was also aimed at establishing whether
the Commission had sufficiently pleaded personal jurisdiction over the pure peregrini.

[44] The CAC’s detailed analysis60 led it to conclude that the Commission had made
sufficient allegations in respect of BNP and HBEU (local peregrini) and JPM Co and

59 Id at para 85.
60 Id at paras 87-173.

ROGERS J
24
CSS (pure peregrini). In respect of all the other active respondents, the CAC concluded
that the superseding affidavit was excipiable and that those banks’ appeals should
succeed. Earlier in its judgment, the CAC had also concluded that the Commission
could not properly include a company as a respondent merely because it was the holding
company of an implicated bank. On this basis, so the CAC concluded, the case had to
fail against BAMLI, BANA, Ned Group and FirstRand.61

[45] Although the CAC repeatedly used the expression “subject matter jurisdiction”
in assessing the case pleaded against the various banks, it was plainly not using that
expression in the technical sense of conduct having effects in South Africa meeting the
QE test. It seems to have used that expression loosely as a synonym for a “cause of
action”, more particularly whether there were sufficient allegations to show that the
bank in question was a participant in the alleged SOC. 62 In regard to subject matter
jurisdiction in the sense of effects meeting the QE test, the CAC dealt with this in only
three paragraphs in the concluding part of its judgment,63 the last of which reads:

“It may well be that the effect that various trades documented by the Commission had
on the Rand was so insignificant as to have had no material effect thereon. But that is
a matter which is better dealt with at trial where the respondent banks, which hav e a
case to answer, can provide evidence to gainsay the case made out by the Commission.”

[46] In regard to the second joinder application, the CAC held that CAC I had required
the Commission to “revisit its significantly imperfect referral affidavit” and to
“reconfigure” it so that it passed the test laid down in CAC I. The Court had not intended
that banks that were not part of the initial referral could be added. What had been
intended was the redrafting of the initial referral affidavit, not the generation of a new

intended was the redrafting of the initial referral affidavit, not the generation of a new
one.64 The CAC rejected the Commission’s reliance on Pickfords65 to justify the

61 Id at paras 64-6.
62 See, for example, CAC II above n 1 at paras 126-8 (JPM Co), 135 (HBEU), 139 (CSS) and 173 (SBSA).
63 Id at paras 180-2.
64 Id at paras 37 and 184.
65 Above n 43.

ROGERS J
25
addition of further respondents post-referral. What Pickfords made clear, according to
the CAC, was that a complaint is initiated against a practice, not specific parties. The
Commission can thus add further parties after the initiation. This did not contradict
CAC I with regard to the “cut off” point for the addition of further parties, “that is, after
the referral”.66

[47] On the strength of this reasoning, the CAC held that the case for the joinder of
SAI was fatally flawed. The CAC also rejected the notion that there had been a tacit
initiation against SAI. That bank had at no stage been the subject of an investigation
prior to the referral of the complaint to the Tribunal.67 In the case of FRB, to which the
same reasoning would also have applied, the CAC instead examined whether the
superseding affidavit sufficiently alleged its participation in the SOC and found that it
did not. The CAC did not do the same for Nedbank.

[48] In the concluding part of its judgment, the CAC returned to the question of the
second joinder application. It stated that because CAC I had not permitted the joinder
of further respondents, the granting of the application to join Nedbank, FRB and SAI
had to be set aside.68

[49] The CAC thus made an order upholding the appeals of BAMLI, JPM Bank,
ANZ, SNY, SBSA, Nomura, CSG, Commerzbank, Macquarie, HBUS, MLP, BANA,

66 CAC II above n 1 at para 67. In CAC II, the Court referred at this point to para 67 of CAC I on the “cut off”
date. The latter paragraph in CAC I above n 2 reads:
“While it is correct that the judgment in [Power Construction (West Cape) (Pty) Ltd v
Competition Commission of South Africa [2017] ZACAC 6; [2017] 2 CPLR 589 (CAC)] makes
it clear that it is necessary for a complaint to be initiated against each firm that is alleged to be
a party to the cartel before that complaint is referred to the Tribunal, there is nothing in this

judgment which deals with the facts of this present case; namely, in the event that this Court
does not disturb the relevant component of the Tribunal’s order, namely that the Commission
has been granted permission to file a new referral affidavit to substitute fo r and replace all the
complaint referral affidavits so long as the complaint referral deals with the individual parties
and shows their connection to the alleged overall conspiracy to form a cartel. In short, the
existing jurisprudence presents no fatal obstacle to the granting of a similar order against the
pure peregrini and thus deferring the joinder application until such time as that referral affidavit
has been filed and can appropriately be considered.”
67 CAC II id at paras 143-6.
68 Id at para 184.

ROGERS J
26
the Nedbank respondents, the FirstRand respondents and SAI. It dismissed the appeals
of BNP, JPM Co, HBEU and CSS. The effect of this was to leave those four banks as
active respondents together with the two Investec entities which had played no part in
the second round of appeals.

The applications for leave to appeal to this Court
[50] In CCT 30/24 the Commission has applied for leave to appeal CAC II in respect
of the banks who succeeded in that Court, save for SNY, CSG, Ned Group and
FirstRand. HBEU has applied for leave to cross-appeal its failure in CAC II. BNP and
CSS have applied for leave to appeal their failures in CAC II (in CCT 25/24 and
CCT 27/24 respectively). JPM Co, the remaining unsuccessful bank in CAC II, has not
sought leave to appeal.

[51] In its founding affidavit in CCT 30/24, the Commission pleads eight grounds of
appeal:
(a) First, as to personal and subject matter jurisdiction, the CAC erred by not
following the interpretation of section 3(1) which the Commission had
advanced in the first round of exceptions.
(b) Second, as to exception procedure, the CAC erred by taking into account
facts alleged by the respondents in their affidavits in support of their
exception applications. This was contrary to well -established principles
governing the adjudication of exceptions.
(c) Third, as to the standard of pleading, the CAC ignored the provisions of
TR 15, instead imposing a higher standard.
(d) Fourth, the CAC erred in holding that the superseding affidavit was
excipiable in relation to the banks that succeeded in CAC II.
(e) Fifth, the CAC erred in holding that the order in CAC I precluded the
joinder of further banks. In so doing, it misconstrued Pickfords.
(f) Sixth, the CAC erred in finding that the Commission needed to issue a
fresh complaint initiation in respect of banks not identified as respondents
at the time of the complaint referral in February 2017.

ROGERS J
27
(g) Seventh, the CAC failed to address the Commission’s contention that the
simultaneous reviews brought by some of the banks should be dismissed.
Parallel reviews are open to abuse and impose a significant burden on the
Commission.
(h) Eighth, and finally, the Commission contends that the order in CAC II is
incomplete and confusing in various respects.

[52] In its application for leave to cross-appeal in CCT 30/24, HBEU pleads that the
CAC erred in not having regard to affidavits stating that the traders identified in the
superseding affidavit as acting for HBEU were in fact employed by and representing
HBUS – the so -called “wrong guy” point. In so doing, the CAC treated HBEU
differently from other banks that succeeded in the CAC on the basis of not having
employed or been represented by the implicated traders.

[53] In CCT 25/24, BNP pleads that the CAC erred in failing to uphold its exception.
The exception contended that the superseding affidavit contained contradictory
averments as to the duration of the SOC and as to the continued participation of the
various respondents, and in these respects failed to comply with the pleading
requirements of CAC I.

[54] In CCT 27/24, CSS pleads two grounds of appeal:
(a) Although the CAC found against CSS on its exception application, the
CAC failed to deal with CSS’ appeal against the Tribunal’s joinder
decision. The joinder appeal should have succeeded, because the
Commission failed to establish that it had initiated a complaint against
CSS. This was necessary in view of the fact that CSS had not been named
in the initiation statements of April 2015 and August 2016, and that the
Commission only sought to join CSS after the referral to the Tribunal in
February 2017.
(b) The CAC erred in finding against CSS on its exception application. In so
doing, the CAC wrongly treated CSS as a local peregrinus for purposes

ROGERS J
28
of personal jurisdiction, whereas it was common cause between the
parties that CSS was a pure peregrinus. The allegations in the
superseding affidavit against CSS as a pure peregrinus stood on the same
footing as the affidavit’s allegations against MLP and HBUS. The latter
respondents succeeded in the CAC on the basis that the affidavit failed to
establish that they were implicated in a conspiracy with South African
banks.

Jurisdiction and leave to appeal
[55] This Court’s jurisdiction is set out in section 167(3)(b) of the Constitution. 69
These applications raise several arguable points of law of general public importance
which this Court ought to consider. These include (a) questions relating to joinder and
complaint initiation, particularly when the Commission wishes to add further firms after
a complaint has already been referred to the Tribunal; (b) alleged legal errors committed
by the CAC in relation to the requirements for and pleading of an SOC; and (c) the
proper interpretation of section 3(1), though this last question may not be reached if we
are against the Commission on res judicata (a case finally decided ) and peremption.
We thus have general jurisdiction in terms of section 167(3)(b)(ii).

[56] Res judicata and peremption implicate the rule of law and the right of access to
courts guaranteed in section 34 of the Bill of Rights,70 so our constitutional jurisdiction
is also engaged. The questions of joinder and initiation concern the powers and
jurisdiction of two statutory bodies, the Commission and Tribunal . The proper
delineation of the powers and jurisdiction of such statutory bodies is a constitutional
question. The same is true in relation to section 3(1), since its proper interpretation
affects the jurisdiction of the Tribunal , both as to persons and subject matter . The

69 Section 167(3)(b) provides that the Constitutional Court may decide—
“(i) constitutional matters; and

“(i) constitutional matters; and
(ii) any other matter, if the Constitutional Court grants leave to appeal on the grounds that
the matter raises an arguable point of law of general public importance which ought to
be considered by that Court.”
70 See the authorities discussed later in this judgment at [101] to [126].

ROGERS J
29
interpretation of section 3(1) must also, argues the Commission, be interpreted in
accordance with the injunction in section 39(2) of the Constitution , namely in a way
that best promotes the spirit, purport and objects of the Bill of Rights. In that regard,
the Commission invokes the right of access to courts in section 34 of the Bill of Rights
as an important interpretive influence. We thus also have constitutional jurisdiction.

[57] This does not necessarily mean that what will remain for decision after the above
issues have been d etermined will be matters engaging our jurisdiction. Most of the
banks contend that, unless we find that the CAC went awry on the law, what will remain
are mundane questions as to whether, in the particular circumstances of this case, the
superseding affidavit pleaded sufficient facts to withstand the objections taken against
it. These latter questions, they submit, are not within our jurisdiction.

[58] As shall appear in due course, the Commission enjoys prospects of success on at
least some of the questions which we have jurisdiction to decide. Quite apart from
prospects of success, however, the importance and magnitude of the case make it
desirable to pronounce on the questions I have identified . Although some of th ose
questions are more prominent in relation to some banks than others, it would not be
profitable to draw distinctions. With one exception, the parties seeking leave to appeal
should be granted leave, and the various appeal s should be considered on their merits.
The exception is BNP’s application for leave to appeal, for reasons I shall explain later.

The issues and further structure of this judgment
[59] I shall deal with the issues that arise in this case in the following order:
(a) First, I shall address the issues relating to the standard of pleading and
exception procedure in Tribunal proceedings.
(b) I shall then turn to the Commission’s argument on the interpretation of

(b) I shall then turn to the Commission’s argument on the interpretation of
section 3(1) and the important related questions of res judicata and
peremption.
(c) Thereafter I shall address issues concerning joinder procedure in the
Tribunal, and the related questions whether it is permissible to join

ROGERS J
30
respondents after a complaint has been referred to the Tribunal, whether
further initiations were needed against the joinder respondents and
whether CAC I precluded the joinder of further parties. That part of the
judgment will be concerned with and will determine the issues raised in
this respect by CSS, Nedbank and SAI.
(d) Then I shall deal with the requisites for pleading an SOC and various
related questions. In that part of the judgment I shall, by way of
introduction, summarise how the Commission went about pleading the
SOC, before dealing, under separate subheadings, wi th 13 alleged legal
errors which, according to the Commission, the CAC made in its
assessment of the pleaded case against the various banks. Those alleged
errors were identified by the Commission in a note submitted at the
Court’s request during the cours e of the hearing and to which the banks
were given an opportunity to respond.
(e) With a view to analysing the 13 alleged errors, it will be convenient in
that part of the judgment to summarise the Commission’s pleaded case
against various banks and the CAC’s reasoning in favour of those banks:
JPM Bank (third error); Nomura and Macquar ie (sixth error); FRB
(seventh error); HSBC (eighth error); BoA (ninth error); ANZ (ninth and
twelfth errors); and Commerzbank and SBSA (eleventh error).
(f) I shall then draw the threads together in a part of the judgment dealing
separately with each of the banks. I shall there address any additional
matters not covered in earlier parts of the judgment.
(g) In the concluding section, I shall summarise the overall result and deal
with costs.

Pleadings and exceptions in Tribunal proceedings
[60] Although the Tribunal Rules do not explicitly provide for exceptions, the
Tribunal has over many years held that it can entertain exceptions. In terms of
section 52(2) of the Act, the Tribunal must conduct its hearings in public, expeditiously

section 52(2) of the Act, the Tribunal must conduct its hearings in public, expeditiously
and in accordance with the principles of natural justice; and it may do so informally or

ROGERS J
31
in an inquisitorial way. In terms of section 55(1), the member presiding at a hearing
may determine any matter of procedure with due regard to section 52(2). In terms of
TR 55(1), the presiding member may give directions on procedure and may, for that
purpose, have regard to the High Court Rules.

[61] As to the form for exceptions, the Tribunal has often permitted this to be done
on notice of motion supported by an affidavit setting out the litigant’s objections. This
does not mean that the objecting litigant may, without more, rely on additional facts
contained in its affidavit. In terms of High Court procedure, an excipient must confine
its attack to the terms of the impugned pleading and may not introduce additional facts.
The allegations in the impugned pleading must, moreover, be assumed to be true.

[62] The Tribunal has emphasised, however, that although it may consider
High Court procedure, it has a wide discretion in the conduct of its proceedings.
Although its proceedings are adversarial in form, it has inquisitorial powers to arrive at
the truth. The Tribunal must conduct its proceedings with fairness and guard against
elevating form over substance. Fairness is context -driven, having regard to the
circumstances of each case. Fairness is not a one-way street – all parties have the right
to fairness in conducting their cases.71

[63] The Tribunal has also said that, unlike the approach in the High Court, its general
approach to exception applications has been to decide each one on its own merits and
circumstances, and not to adopt an overly technical approach. Its approach, the Tribunal
has stated, is informed by three considerations. First, complaint proceedings are sui
generis (unique), consisting of elements of both motion and trial procedure. Second,
the subject of the Tribunal’s proceedings involves the intersection of law and
economics. What may appear to be a pure point of law may in truth require an entire

economics. What may appear to be a pure point of law may in truth require an entire

71 Rooibos Ltd v South African Competition Commission [2009] ZACT 58; [2009] 2 CPLR 572 (CT) at paras 5
and 8; BMW South Africa (Pty) Ltd t/a BMW Motorrad v Fourier Holdings (Pty) Ltd t/a Bryanston Motocycles
[2011] ZACT 3; [2011] 1 CPLR 181 (CT) at para 22; and South African Medical Association v Council for
Medical Schemes [2016] ZACT 71; [2016] 2 CPLR 1027 (CT) at para 52 and fn 35.

ROGERS J
32
factual matrix. Third, the Tribunal has inquisitorial powers and a wide discretion in the
conduct of its proceedings. The guiding principle is fairness.72

[64] TR 15(2), as previously noted, provides that a complaint referral must be
supported by an affidavit setting out in numbered paragraphs “a concise statement of
the grounds of the complaint” and “the material facts or the points of law relevant to the
complaint and relied on by the Commission or complainant”. The language of TR 15(2)
is similar but not identical to rule 18(4) of the Uniform Rules of Court (URC). In the
case of URC 18(4), the pleading must contain “a clear and concise statement of the
material facts”. In the case of TR 15(2), it is the “grounds of the complaint” that must
be the subject of a “concise statement”. Additionally, the supporting affidavit must then
set out the “material facts or the points of law” relevant to the complaint. The rule does
not state that the “material facts” must merely be the subject of a “concise statement”.
In High Court motion proceedings, UR C 6(1) requires the notice of motion to be
supported by an affidavit “as to the facts upon which the applicant relies for relief”.

[65] Because referral proceedings are a blend of motion and trial proceedings, and
because the referral affidavit must set out the “material facts” on which the Commission
relies, a skeletal statement of the cause of action, such as might suffice for a High Court
pleading, is unlikely to meet the standards set by TR 15(2). As Wallis JA stated in
Loungefoam:73

“[12] Assuming this reflects the general stance of the Commission it is labouring
under a fundamental misconception as to the nature of the affidavit required by
rule 15(2). It treats it as if it is a type of pleading, subject to amendment from time to
time as the case develops. That is incorrect. An affidavit in competition proceedings
has precisely the same character as it has in any other circumstances. It is a s worn

has precisely the same character as it has in any other circumstances. It is a s worn
statement on oath by a witness that is required by rule 15(2) to set out a concise
statement of the grounds of the complaint and the material facts and points of law

72 Invensys plc v Protea Automation Solutions (Pty) Limited; In re: Protea Automation Solutions (Pty) Limited v
Invensys plc [2014] 2 CPLR 505 (CT) at paras 13-16.
73 Above n 41.

ROGERS J
33
relevant to the complaint and relied on by the Commission. It serves the same purpose
as an affidavit in application proceedings, which contains both the allegations
necessary in a pleading, including any relevant propositions of law, and the essential
evidence in support of those allegations.
[13] It was suggested in argument before us that the affidavit delivered in support
of a referral to the Tribunal is sui generis and does not stand on the same footing as a
conventional affidavit. Counsel made the point that the deponent to the affidavit is
usually an investigator in the employ of the Commission and much of the contents
thereof constitute hearsay. It is unusual fo r the investigator to be a witness in the
proceedings before the Tribunal and in practice the Tribunal determines the cases tha t
come before it on the basis of oral and documentary evidence.
[14] Whilst this may accurately describe what happens in practice it is unclear why
it is thought to alter the fundamental nature of an affidavit. There is no legal prohibition
against an affidavit containing hearsay evidence. In certain circumstances and bef ore
certain tribunals such evidence is inadmissible, but that does not mean that an affidavit
in support of a referral to the Tribunal cannot contain hearsay. It may be convenient
for the Commission to cause the affidavit to be deposed to by the investiga tor who
investigated the complaint. That is likely to be a sensible course, as the investigator
will have the relevant facts and documents at her or his fingertips. However, it is
inevitable in those circumstances that the affidavit will largely be an af fidavit of
information and belief rather than direct evidence. That is immaterial bearing in mind
the practice of the Tribunal to conduct a hearing at which witnesses with direct
knowledge of the facts testify under oath and are cross-examined. No doubt if it sought
to rely only on the investigator’s affidavit that would provoke protest from other parties

to rely only on the investigator’s affidavit that would provoke protest from other parties
but that is a different matter.”

[66] In practice, the Commission includes a substantial body of evidence in its referral
affidavits, albeit through the mouth of an investigator, and that was done here. How
then should one adjudicate a complaint that the referral affidavit fails to disclose a cause
of action? The mere presence of formal assertions constituting the bare bones of a cause
of action are unlikely to suffice, bearing in mind that the referral affidavit is meant also
to contain evidence.

ROGERS J
34
[67] I do not wish to cast doubt on the general principles which the Tribunal has laid
down for itself in dealing with exceptions or to fetter its procedural discretion, but I
think the following would generally be the appropriate test. Assuming that all the facts
alleged in the referral affidavit are proved and that no other facts are proved, could the
Tribunal, acting reasonably, conclude that the Commission has made out a case for the
relief claimed? Where the case depends on inferences, this would include asking
whether the Tribunal, acting reasonably, could draw the inferences which the
Commission seeks to draw from the primary facts alleged in the referral affidavit. This
is not dissimilar to the test which has been applied by our civil courts when a respondent
takes a preliminary objection that the founding affidavit does not make out a case for
the relief claimed. The question is whether the founding affidavit makes out a “prima
facie case”.74

[68] Ordinarily a respondent raising such an objection is confined to the facts stated
in the founding affidavit, 75 just as an excipient is confined to the facts alleged in the
impugned pleading. However, the principles which the Tribunal has laid down do not
warrant a rigid rule that under no circumstances may additional facts be taken into
account. Considerations o f fairness may, in some circumstances, justify regard being
had to limited additional facts. It is relevant to bear in mind, in this regard, that in a
complex case the trial of a complaint referral may run for many weeks or months. That
would seem to be true of the present case. Moreover, the Tribunal does not have the
power to order the Commission to pay a respondent’s costs if the Commission fails
against that respondent.76

[69] Additional evidence adduced by the respondents may be relevant in the case of
BAMLI, SBSA and HBEU. This evidence relates mainly to matters such as who

BAMLI, SBSA and HBEU. This evidence relates mainly to matters such as who

74 Taylor v Welkom Theatres (Pty) Ltd 1954 (3) SA 339 (O) at 344G-345A; Bader v Weston 1967 (1) SA 134 (C)
at 136B-E; Hart v Pinetown Drive-In Cinema (Pty) Ltd 1972 (1) SA 464 (D) at 465E-G and 469B-F; and Pearson
v Magrep Investments (Pty) Ltd 1975 (1) SA 186 (D) at 187B-188A.
75 See the cases cited in the preceding footnote.
76 Section 57(1) states that in Tribunal proceedings each party must bear its own costs. Where the referring party
is a complainant (i.e. a private entity), the Tribunal may grant costs for or against a respondent, depending on the
outcome, but there is no equivalent power where the referring party is the Commission.

ROGERS J
35
employed the person named as a trader and whether that person was a trader. There
was also evidence from JPM and SBSA about the workings of currency trading. As
respondents in the referral as at February 2017, these firms were, in principle, confined
to raising exceptions.

[70] In TRIB II, the Tribunal set out its guiding principles on exceptions in the way I
have summarised above, but it did not specifically mention the High Court principle
that an excipient may not rely on additional evidence. 77 When dealing with BAMLI,
SBSA and HBEU, the Tribunal did not say that it was precluded from having regard to
the evidence in question, instead observing that it would be reluctant at the exception
stage to reject the Commission’s version or that the matters in question were more
properly a matter for defence on the merits. 78 There was no detailed engagement with
the facts, and this is particularly true for SBSA. The CAC, by contrast, relied freely on
the additional facts, without explicitly considering whether, in the particular
circumstances, it was fair to depart in this respect from usual exception principles.

[71] I shall need to return to these failings by the Tribunal and the CAC in regard to
the additional evidence when considering the Commission’s appeal s in respect of
BAMLI and SBSA, and HBEU’s cross-appeal.

[72] The Tribunal’s modified exception principles apply in principle to the dismissal
applications brought by banks from the initial group of 18 respondents cited in the
February 2017 referral. In the case of further respondents whom the Commission
sought to join, those respondents were entitled to file answering affidavits on the facts.
There can be no question that the Tribunal was bound to have regard to those facts. If,
however, the facts were properly contested by the Commission, the Tribunal could
justifiably have adopted the position that the bank in question should be joined and that
the contested facts be resolved at trial.

the contested facts be resolved at trial.

77 TRIB II above n 4 at paras 49-60.
78 Id at para 286 (BAMLI) and para 313 (SBSA). The additional facts of potential relevance to HBEU were not
mentioned (see paras 345-7).

ROGERS J
36

[73] Although the CAC did not explicitly formulate the test it intended to apply in
assessing the adequacy of the superseding affidavit, its various formulations do not
show it to have been guilty of any material misdirection. The CAC, correctly in light
of Loungefoam, did not hold against the Commission that the investigator who made its
affidavits did not have personal knowledge of the facts. The pleaded facts were
accepted as facts. On several occasions the CAC spoke of the Commission’s need to
establish a “prima facie case”.79 Elsewhere it spoke of a “sufficient case”,80 “sufficient
facts”,81 “sufficient evidence”82 and “plausible evidence”.83 In exonerating some of the
banks, the CAC said that certain allegations were “hardly evidence” 84 of participation
in the SOC or that there was “simply insufficient evidence”85 or an absence of “plausible
evidence”86 or, in some instances, “no evidence”.87

[74] Bearing in mind that the Commission’s case rested on inferences, these various
formulations are all consistent with testing whether the Commission’s facts, if proved,
made out a prima facie case. In those instances where the CAC found for the banks,
the CAC did not consider a reasonable inference of participation in the SOC could be
drawn from the primary facts.

Section 3(1), res judicata and peremption
[75] The Commission submits that the interpretation of section 3(1) in CAC I was
wrong and that we should accept the argument advanced by the Commission in the first
round of exceptions, namely that section 3(1) is an all -encompassing jurisdictional

79 CAC II above n 1 at paras 61, 89, 118, 128, 135 and 139.
80 Id at paras 79 and 151.
81 Id at 61.
82 Id at para 36.
83 Id.
84 Id at paras 104 and 114.
85 Id at paras 115 and 123.
86 Id at para 170.
87 Id at paras 89, 93, 142 and 153.

ROGERS J
37
provision that wholly displaces the ordinary requirements for personal jurisdiction and
subject matter jurisdiction. If this argument were open to the Commission, it would be
an arguable point of law of general public importance engaging this Court’s jurisdiction.
But is the argument open to the Commission?

[76] At the hearing, the questions of res judicata and peremption were raised. Since
these matters were not addressed as fully as perhaps they should have been, post-hearing
directions were issued for the parties to file written submissions. The submissions were
required to address these four questions:
(a) To the extent that the doctrine of res judicata finds application in this
matter, is it procedurally open to this Court to consider relaxing the
doctrine of res judicata?
(b) If it is, what factors weigh in favour of or militate against the relaxation
of the doctrine of res judicata in this case?
(c) Does the doctrine of peremption render moot any relaxation of the
doctrine of res judicata?
(d) If neither the doctrine of res judicata nor the doctrine of peremption
precludes this Court from interfering in the holding in CAC I, on what
basis, if any, may this Court interfere with such holding?

[77] The questions of res judicata and peremption are mainly of relevance to personal
jurisdiction over pure peregrini. Of the pure peregrini, personal jurisdiction was not
argued by SAI, which defended the CAC’s decision in its favour on other grounds.

Was res judicata sufficiently pleaded?
[78] In its submissions , the Commission has contended, preliminarily, that none of
the banks pleaded res judicata in their opposing affidavits in this Court. I should
immediately make the point that the question of res judicata did not arise during the
hearing of TRIB II and CAC II because it was not argued in those fora that CAC I was
not binding. Indeed, the proceedings in TRIB II and CAC II were premised on CAC I

not binding. Indeed, the proceedings in TRIB II and CAC II were premised on CAC I
being binding. What served before the fora in TRIB II and CAC II were exceptions and

ROGERS J
38
objections to the superseding affidavit in which the banks contended that the affidavit
failed to meet the requirements laid down in CAC I on various matters, including
personal and subject matter jurisdiction. CAC I was in essence the standard by which
the exceptions and objections in TRIB II and CAC II were adjudicated.

[79] I accept that usually res judicata should be pleaded. If it is not, the other party
may not know the previous judgment on which its opponent relies and how that
judgment bears on the issues in the later proceedings. A failure to plead res judicata
may, in such circumstances , cause the other party to be taken by surprise if it is first
raised in argument. That concern is absent in a case such as the present, where the
previous judgment was front and centre in the later proceedings. The entire second
round of proceedings in TRIB II and CAC II was concerned with the extent to which the
superseding affidavit complied with CAC I. The banks expressly invoked CAC I as the
basis for their exceptions and objections. The Commission did not, in opposing the
exceptions and objections, contend that CAC I did not apply.

[80] In these circumstances, where res judicata only became relevant because of the
Commission’s belated shift of stance in this Court, there would be no unfairness to the
Commission if the banks were permitted to invoke res judicata for the first time in
argument. This said, res judicata, or at least the facts to sustain that defence in law,
were sufficiently raised in the opposing affidavits of those foreign banks that have made
submissions in response to the Court’s directions:
(a) The three BoA entities in their answering affidavit in this Court did not
use the label “ res judicata ”. They did, however, allege that the
Commission was impermissibly seeking to reopen issues finally decided
in CAC I. This was a defence of res judicata expressed in English rather
than Latin. They also alleged that the Commission had not appealed

than Latin. They also alleged that the Commission had not appealed
CAC I, and had on the contrary accepted that it was bound by CAC I. In
their main written submissions, the BoA entities relied upon res judicata
in terms.

ROGERS J
39
(b) JPM Bank, although not a pure peregrinus, expressly pleaded res judicata
in its answering affidavit in this Court in the context of subject matter
jurisdiction. JPM Co, which is a pure peregrinus, had no opportunity to
plead res judicata because it is not a party to the proceedings in this Court.
In the main joint written submissions by JPM Bank and JPM Co, they
spoke of peremption in relation to CAC I.
(c) ANZ, which saw the Commission’s resurrected argument on section 3(1)
as a disguised attempt to appeal against CAC I , expressly pleaded
peremption, alleging that the Commission had not sought to challenge
CAC I, had been content to abide by it and had demonstrated its
acceptance of CAC I when delivering the superseding affidavit . In
substance, therefore, the Commission knew that ANZ was contending that
it could not reopen the issues decided in CAC I . ANZ also relied on
peremption in its main written submissions.
(d) Like ANZ, Nomura expressly alleged peremption, which in context
suffices. In its main written submissions, Nomura submitted that the
Commission’s resurrected argument on section 3(1) was impermissible
because CAC I, having gone unchallenged, bound the Commission when
drafting the superseding affidavit.
(e) Macquarie, in its answering affidavit in this Court, did not use the
expressions “res judicata ” or “peremption”, but pleaded the facts for
peremption. It alleged that the Commission had not applied to this Court
to appeal CAC I and had instead accepted that it was required to comply
with it and had attempted to do so in the superseding affidavit. In its main
written submissions, Macquarie contended that the Commission was
precluded from resurrecting its previous arguments on personal
jurisdiction and subject matter jurisdiction, because it did not seek leave
to appeal CAC I and had sought to comply with it through the superseding
affidavit, thereby waiving or abandoning those arguments and accepting
the findings in CAC I.

ROGERS J
40
(f) HBUS’ answering affidavit in this Court is similar to Macquarie’s in this
respect. In its main written submissions, HBUS invoked peremption.
(g) CSS did not file an answering affidavit in the Commission’s application
for leave to appeal, since CSS lost in the CAC. When CSS filed its own
application for leave to appeal, the Commission had not yet
communicated its decision to resurrect the section 3 (1) argument. CSS
thus had no reason to anticipate res judicata or peremption in its founding
affidavit. The founding affidavit did, though, refer to the procedural
background, including CAC I and the superseding affidavit filed pursuant
to it. When the Commission raised the section 3(1) argument in its
answering affidavit, CSS did not seek leave to file a replying affidavit. In
its main written submissions, however, it did rely on both res judicata and
peremption. CSS submits that this was the first opportunity it had to raise
these matters subsequent to the filing of the Commission’s answering
affidavit.

[81] All these banks, therefore, expressly put up CAC I and the fact that it had not
been appealed as a basis for the Commission being bound by that judgment. One can
understand why some of the banks saw the Commission’s application in this Court as a
disguised attempt to appeal CAC I, hence the language of peremption. Res judicata was
quite plainly an available defence on the alleged facts if the Commission should clarify,
as it later did in oral argument, that it was not seeking to appeal CAC I.

Two preliminary obstacles for the Commission
[82] As I shall explain presently, both res judicata and peremption should be applied
to bar reconsideration of the section 3(1) arguments if those arguments are otherwise
properly before us. However, there are, in my view, two fundamental obstacles in the
Commission’s way, the effect of which is that we do not even need to reach res judicata

Commission’s way, the effect of which is that we do not even need to reach res judicata
and peremption. Nevertheless, and after dealing with these two obstacles, I shall
address res judicata and peremption to the extent that they may remain relevant.

ROGERS J
41
No appeal against the CAC I orders
[83] The first is the absence of an appeal against CAC I . In oral argument, the
Commission’s counsel confirmed that the Commission was not seeking leave to appeal
CAC I, and that position has been repeated in its latest submissions filed in response to
this Court’s directions. This means that the orders made in CAC I stand.
Paragraph 3.2.1 of the CAC I order required the Commission to file a superseding
affidavit setting out facts to support an allegation that it was foreseeable that the
impugned conduct would have a direct or immediate and substantial effect in South
Africa (the QE test for subject matter jurisdiction); and paragraph 3.3.1 of the CAC I
order required the superseding affidavit to allege adequate connecting factors between
the parties and the Tribunal’s jurisdiction, sufficient to establish personal jurisdiction
(the ACF test for personal jurisdiction). Those orders would in turn need to be
interpreted with reference to the reasoning contained in CAC I. A decision of this Court
on the Commission’s resurrected section 3(1) arguments would thus hang in the air and
be without practical effect, since our decision would not change the Commission’s
obligation to comply with CAC I . This Court should not embark on an academic
exercise.88

This Court’s appellate jurisdiction
[84] The second obstacle goes to this Court’s jurisdiction. The Commission invokes
our appellate jurisdiction. This means that we only have jurisdiction to decide matters
arising from the decision under appeal, in this case CAC II. TRIB II and CAC II were
concerned with exceptions and objections in which the banks contended that the
superseding affidavit failed to comply with the requirements laid down in CAC I. The
correctness of CAC I was not an issue in TRIB II and CAC II. On the contrary, CAC I
was the very standard that formed the basis of the exceptions and objections, and the

was the very standard that formed the basis of the exceptions and objections, and the
litigation was conducted in TRIB II and CAC II on that basis. This being so, the

88 Ferreira v Levin N.O.; Vryenhoek v Powell N.O. [1995] ZACC 13; 1996 (1) BCLR 1 (CC) ; 1996 (1) SA 984
(CC) at para 165.

ROGERS J
42
resurrected section 3(1) arguments, which the CAC rejected in CAC I, are simply not
issues properly arising on an appeal against CAC II.

[85] One can test this by asking what would have happened if, prior to the
adjudication of the second round of exceptions and objections in the Tribunal or the
CAC, the Commission had said that it wished belatedly to apply for leave to appeal
against CAC I. The exceptions and objections would have been held in abeyance
pending the outcome of the belated application for leave to appeal, since a successful
appeal against CAC I would have rendered the premise on which the exceptions and
objections were based largely moot. Yet it is those very exceptions and objections that
are essentially before us in the current application for leave to appeal.

Peremption
[86] Subject to these preliminary obstacles, res judicata and peremption are both
applicable in this case. Peremption can sometimes be overlooked and res judicata can
sometimes be relaxed. I shall first consider whether peremption and res judicata are in
principle applicable. I shall then, under a separate heading, consider whether
peremption should be overlooked or res judicata relaxed.

[87] Peremption occurs most often when a litigant who has initially accepted a
judgment changes course and seeks to appeal it. In that situation, if peremption bars
the appeal, the judgment will stand and become res judicata. The purpose of the appeal
would be to prevent the judgment from standing as res judicata. Where there is no
attempt to appeal a judgment, res judicata would suffice to prevent a litigant from
reopening the case. However, the circumstances may show not only that there was no
appeal but also that the litigant by its conduct accepted the judgment, thereby waiving
its right to impeach it. Peremption, being a species of waiver, can then operate as an
additional basis for barring a challenge to matters decided in the judgment.

additional basis for barring a challenge to matters decided in the judgment.

[88] In regard to peremption, it is clear beyond doubt that the Commission decided
not to challenge CAC I. This is shown not only by the fact that it at no stage sought to

ROGERS J
43
appeal CAC I, but also by the terms of the superseding affidavit which the Commission
filed in purported compliance with the orders in CAC I . Personal jurisdiction and
subject matter jurisdiction were pleaded with reference to the ACF and QE tests laid
down in CAC I. As to personal jurisdiction over the pure peregrini, the superseding
affidavit alleged that there was such jurisdiction by virtue of two “connecting factors”,
namely participation in a conspiracy involving the ZAR exchange rate and such
conspiracy being with South African respondents. Section 3(1) was mentioned only in
relation to subject matter jurisdiction, and in that context the “effect” was formulated
with reference to the QE test.89

[89] A number of the foreign banks expressly allege in their answering affidavits in
this Court that the Commission had chosen not to appeal and had instead accepted
CAC I and tried to comply with it. Although the Commission sought leave to file two
replying affidavits on other matters, there has been no response on oath to the banks’
allegation of a deliberate choice not to appeal CAC I . Factually, therefore, it is
uncontested that peremption occurred. Indeed, we now have it from the Commission’s
own mouth that it still does not intend to appeal CAC I.

[90] In its latest submissions, the Commission contends that its oral argument in
CAC II reflects that it raised the section 3(1) argument, which tells against a finding of
peremption. A transcript of a passage from the oral argument was annexed to one of
the Commission’s replying affidavits to demonstrate this. This replying affidavit was
filed in response to a different point made by some of the banks in their answering

89 In paragraph 51 of the superseding affidavit, the Commission pleaded that the pure peregrini were participants
in an SOC to implement the conspiracy—
“all of which constitute economic activity having a direct/immediate, substantial and

“all of which constitute economic activity having a direct/immediate, substantial and
foreseeable consequence upon the economy of South Africa and the welfare of South African
consumers and an effect within South Africa for the purposes of section 3(1) of the Act.”
The same allegation was made in respect of the local peregrini – see paras 45 and 48 of the affidavit. In the
section of the affidavit headed “The Effect of the Conduct” (paras 259 ff), the very first allegation was formulated
with reference to the QE test: “The Conspiracy had a direct or immediate, and substantial effect in the Republic
and it was foreseeable that the impugned conduct would, or had the potential to, have such an effect”. This section
of the affidavit contained subheadings referencing the QE test: “The direct and/or immediate effect of the conduct”
(paras 268 ff); “The effect of the conduct was cumulatively substantial” (paras 271 ff); and “It was foreseeable to
the cartelists that the conduct would have a direct or immediate and substantial effect” (paras 275 ff).

ROGERS J
44
affidavits, namely that the section 3(1) arguments that the Commission advanced in its
founding affidavit in this Court had not been raised in CAC II.

[91] This extract cannot bear the weight the Commission places on it. The
section 3(1) arguments were not resurrected in TRIB II nor in the heads of argument
filed in CAC II. The subject arose tangentially in the course of the Commission’s oral
submissions on subject matter jurisdiction. Counsel was arguing that a manipulation of
the ZAR among foreign banks could be regarded as “economic activity” within South
Africa for purposes of section 3(1). This argument was raised in case the Commission
could not come home on showing sufficient “effects” in South Africa to meet the QE
test.

[92] One of the Judges in CAC II at one point asked counsel whether, for purposes of
jurisdiction, the economic activity that conferred jurisdiction was the participation of
the banks in an SOC. Counsel replied that what conferred jurisdiction was not the SOC
as such but the effects of the underlying conduct, the manipulation of the ZAR:

“I know, Justice Davis, that I made the point three years ago without success . . . [b]ut
I’m repeating the point today. I’ve already made the first point about foreign banks
engaging with our currency in circumstances where the only institutions that can
protect that currency are local institutions. I think there’s something fundamental about
that question which if it’s time to revisit, it should be revisited. It may be that it’s the
wrong case to revisit it now.”

[93] Counsel then resumed his submissions on subject matter jurisdiction. It is
unclear precisely what point counsel was referring to in the above extract, given that
the context of the debate at this juncture was subject matter jurisdiction. There was no
clear articulation of any argument that, contrary to CAC I, section 3(1) dispensed with
the need for personal jurisdiction or that the effects need not be effects meeting the QE

the need for personal jurisdiction or that the effects need not be effects meeting the QE
test. Counsel appears to have been pressing a different contention, namely that foreign
conduct involving manipulation of the ZAR could be regarded as “economic activity”
within South Africa for purposes of subject matter jurisdiction, thus rendering it

ROGERS J
45
unnecessary to determine whether the resultant effects satisfied the QE test. Moreover,
the fact that the Commission or its counsel might not have agreed with some of the legal
conclusions in CAC I does not show that the Commission did not accept CAC I, for
better or for worse, and it is the latter question rather than the former that is germane to
peremption.

Res judicata
[94] In regard to res judicata , CAC I is a final judgment on the interpretation of
section 3(1) in relation to personal jurisdiction and subject matter jurisdiction. Just as
TRIB I, being a decision on jurisdiction, could be (and was) appealed to the CAC, so
the Commission, and for that matter the banks, could have sought leave to appeal CAC I
to this Court. None of them did.

[95] The Commission maintains that it was wholly successful in the order in CAC I
and could not merely challenge the reasoning in that case, since there is no procedural
mechanism for doing so. I reject that argument. The Commission had partial success
in its cross -appeal against TRIB I , but it lost its primary arguments, namely that
section 3(1) did away with the need for personal jurisdiction and that any effect within
South Africa sufficed for subject matter jurisdiction. It was not only the CAC’s
reasoning that was against the Commission on these matters. The reasoning was
reflected in paragraphs 3.2.1 and 3.3.1 of the CAC’s orders, as noted earlier. The
Commission would have been entitled to seek leave to appeal these orders if it was not
content to accept the rejection of its section 3(1) arguments.

[96] The requirements for personal and subject matter jurisdiction as laid down in
CAC I are thus res judicata. All the elements of res judicata are satisfied.90 It is not
merely a case of issue estoppel. The parties involved in the first round of litigation that
culminated in CAC I are the same as those involved in the second round of litigation

culminated in CAC I are the same as those involved in the second round of litigation

90 For these requirements, see Kommissaris van Binnelandse Inkomste v Absa Bank Bpk [1994] ZASCA 144;
[1995] 1 All SA 517; 1995 (1) SA 653 (A) at 664C-E.

ROGERS J
46
culminating in the present proceedings in this Court. The Commission is pursuing the
same cause of action and seeking the same relief. CAC I decided preliminary issues in
the very same case.

[97] It is true that there are three additional parties whom the Commission wishes to
join – Nedbank, FRB and SAI.91 This does not detract from the operation of res judicata
as between the Commission and all the other parties. Moreover, the section 3(1) issues
that are rendered res judicata by CAC I are not germane to Nedbank and FRB as incolae,
and SAI does not rely on the section 3(1) issues in resisting the Commission’s appeal.
No doubt for that reason, SAI’s attorneys notified the Court that the ir client would not
be filing written submissions in response to the directions, although they did say they
aligned themselves with the supplementary submissions filed by HSBC.

[98] More to the point, perhaps, is that JPM Co, a pure peregrinus, is not a party to
the proceedings in this Court. If we were to reverse CAC I on the questions of personal
and subject matter jurisdiction, JPM Co could be prejudiced in its absence. I recognise
that JPM Co and JPM Bank, in their joint main submissions, dealt with the question of
peremption. However, the Commission specifically ob jected to any submissions on
behalf of JPM Co and requested that it be removed from the submissions in question.
This Court directed that counsel for JPM Bank should not advance oral argument on
any issues relating only to JPM Co. Because JPM Bank is a local peregrinus, and
because the focus of its opposition to the appeal is narrow, its counsel’s oral argument
did not touch on personal and subject matter jurisdiction. The supplementary
submissions in response to our latest directions are in the name of JPM Bank alone.

[99] In Zuma,92 this Court said that “the principles of legal certainty and finality of
judgments are the oxygen without which the rule of law languishes, suffocates and

judgments are the oxygen without which the rule of law languishes, suffocates and

91 I omit Ned Group and FirstRand, because the Commission has not sought to appeal its failure in respect of those
two entities.
92 Zuma v Secretary of the Judicial Commission of Inquiry into Allegations of State Capture, Corruption and
Fraud in the Public Sector Including Organs of State [2021] ZACC 28; 2021 (11) BCLR 1263 (CC).

ROGERS J
47
perishes”.93 A core doctrine giving effect to these principles is res judicata . The
doctrine is a facet of the rule of law. It also draws force from section 34 of the Bill of
Rights.94 That section guarantees access to the courts, and does so for the purpose of
having disputes “resolved” and “decided”. That purpose is frustrated if a final judgment
can be reopened, because then the proceedings that gave rise to the judgment will not
have resolved and decided the dispute.

[100] It has also been held by the United Kingdom’s Supreme Court, and I agree, that
the public interest in finality is “reinforced by the current emphasis on efficiency and
economy in the conduct of litigation, in the interests of the parties and the public as a
whole”.95

Overlooking res judicata and peremption – relevant principles
[101] This Court recognised in Molaudzi96 that the Court could relax the doctrine of
res judicata if the interests of justice demand. 97 The Court emphasised, however, that
there were dangers in eroding the doctrine: “The rule of law and legal certainty will be
compromised if the finality of a court order is in doubt and can be revisited in a
substantive way” .98 Nevertheless, where “significant or manifest injustice” would
result from allowing an order to stand, the doctrine ought to be relaxed so as to allow
the Court to revisit a past decision : “This requires rare and exceptional circumstances,
where there is no alternative effective remedy.”99

93 Id at para 1.
94 Section 34 provides:
“Everyone has the right to have any dispute that can be resolved by the application of law
decided in a fair public hearing before a court or, where appropriate, another independent and
impartial tribunal or forum.”
95 Test Claimants in the Franked Investment Income Group Litigation v Revenue and Customs Commissioners
[2020] UKSC 47; [2021] 1 All ER 1001; [2022] AC 1 ( Test Claimants) at para 59, quoting Lord Bingham’s

statement to this effect in Johnson v Gore Wood & Co [2000] UKHL 65; [2001] 1 All ER 481; [2001] 2 WLR 72;
[2002] 2 AC 1 at 31D.
96 S v Molaudzi [2015] ZACC 20; 2015 (2) SACR 341 (CC); 2015 (8) BCLR 904 (CC).
97 Id at para 32.
98 Id at para 37.
99 Id at para 45.

ROGERS J
48

[102] The relaxation allowed in Molaudzi occurred in the context of a criminal case.
What was rare and exceptional about it was this. Mr Molaudzi, together with several
others, had been convicted of serious crimes and sentenced to life imprisonment. He
brought an application to this Court for leave to appeal without the benefit of legal
representation. This Court dismissed the application in a short judgment without a
hearing, holding in essence that Mr Molaudzi was merely attacking the factual findings
of th e trial court. Afterwards, two of Mr Molaudzi’s co -accused, who raised
constitutional issues about the admissibility of evidence, succeeded in this Court
following a hearing. Since Mr Molaudzi was identically placed, this Court issued
directions inviting him to bring a second application for leave to appeal. It was that
application which this Court granted, thereby relaxing res judicata. There was no other
remedy open to Mr Molaudzi, because no appeal lay from this Court’s earlier decision.

[103] More recently, in Mothulwe,100 this Court granted a second rescission application
where it had previously dismissed the applicant’s application for leave to appeal and his
first rescission application. Such dismissals had occurred without a hearing and without
a substantive judgment. The applicant was not legally represented. In the second
rescission application, this Court was satisfied that it had failed, when making its first
and second orders, to discern the true justice of the applicant’s case: there had been no
adjudication on the counter -application which the applicant had brought in the
Labour Court to challenge a serious finding of corruption against him. Since he had
exhausted his usual legal remedies of appeal and rescission, and was faced with prior
orders from the apex c ourt, his only recourse was to ask this Court to revisit its own
previous orders. This Court was satisfied that the circumstances of the case were “truly
exceptional”.101

exceptional”.101


100 Mothulwe v Labour Court, Johannesburg [2025] ZACC 10; [2025] 8 BLLR 761 (CC); 2025 (8) BCLR 899
(CC); (2025) 46 ILJ 1853 (CC).
101 Id at paras 35, 38 and 48.

ROGERS J
49
[104] This Court, in Molaudzi, considered comparative law on the question of a court’s
discretion not to apply res judicata. I wish simply to note, in that regard, that the latest
authoritative judgments from the United Kingdom hold that res judicata in its fullest
sense – cause of action estoppel – is an absolute defence in respect of points that were
raised and decided in the previous case. Some relaxation is, however, possible, in
respect of points that could have been raised in the previous proceedings but were not
and in respect of issue estoppel.102

[105] In the leading judgment of the Supreme Court of Canada, Danyluk,103 the Court
drew a distinction between cases where the previous adjudication was judicial on the
one hand and administrative on the other. The discretion not to apply res judicata in
court-to-court cases was said to be narrower than in tribunal-to-court cases.104 Danyluk
itself was a tribunal-to-court case, as indeed are most of the subsequent Canadian cases
dealing with the discretion not to apply res judicata. So much is this the case that some
Canadian courts have doubted whether there is truly a discretion not to enforce res

102 Virgin Atlantic Airways Limited v Zodiac Seats UK Limited [2013] UKSC 46; [2013] 4 All ER 715; [2014] 1
AC 160 at para 22, where Lord Sumption drew the following conclusions from his analysis of Arnold v National
Westminster Bank plc [1991] 2 AC 93; [1991] 3 All ER 41 (this part of Lord Sumption’s speech was concurred
in by the other members of the Court – see at para 42):
“Arnold is accordingly authority for the following propositions:
(1) Cause of action estoppel is absolute in relation to all points which had to be and were
decided in order to establish the existence or non-existence of a cause of action.
(2) Cause of action estoppel also bars the raising in subsequent proceedings of points
essential to the existence or non-existence of a cause of action which were not decided

essential to the existence or non-existence of a cause of action which were not decided
because they were not raised in the earlier proceedings, if they could with reasonab le
diligence and should in all the circumstances have been raised.
(3) Except in special circumstances where this would cause injustice, issue estoppel bars
the raising in subsequent proceedings of points which (i) were not raised in the earlier
proceedings or (ii) were raised but unsuccessfully. If the relevant point was not raised,
the bar will usually be absolute if it could with reasonable diligence and should in all
the circumstances have been raised.”
See also Test Claimants above n 95 at para 62.
103 Danyluk v Ainsworth Technologies Inc 2001 SCC 44 (CanLII); [2001] 2 SCR 460.
104 Id at para 62. See also Lange The Doctrine of Res Judicata in Canada 3 ed (LexisNexis, Ontario 2010) at 225:
“Courts in Canada have regularly maintained that, in a court -to-court context, issue estoppel
and cause of action estoppel should apply. The basic principle is that only in the rarest of cases
should these doctrines be rendered inoperative in a court-to-court context.”

ROGERS J
50
judicata in court-to-court cases.105 Australian courts have yet to recognise the existence
of a discretion to depart from res judicata.106

[106] I mention this not to cast doubt on what was held in Molaudzi and applied in
Mothulwe, but to emphasise that this Court should be extremely wary of departing from
res judicata in a court-to-court case. To what extent res judicata applies in this country
to administrative adjudications is not something I need consider.

[107] A court may also overlook peremption in the interests of justice. The most recent
authority on the subject in this Court, where earlier decisions were reviewed, is United
Manganese.107 Generally speaking, there seems to be greater latitude to overlook
peremption than to disapply res judicata.

Factors for and against overlooking res judicata and peremption
[108] The differences between Molaudzi and this case are obvious. Apart from the fact
that Mr Molaudzi was at risk of having to serve a life sentence despite a wrong ful
conviction, he was initially unrepresented and had exhausted his ordinary remedies.
The question was whether this Court should revisit its own earlier judgment, where
substantively it had already done so in the context of Mr Molaudzi’s co-accused. In the
present case, by contrast, the Commission is a well -resourced regulator which was
legally represented at all relevant times. There was a remedy open to the Commission
if it disagreed with CAC I, namely seeking leave to appeal to this Court. Instead the
Commission chose to abide by CAC I, and delivered a superseding affidavit that
manifestly sought to give effect to the requirements laid down in CAC I for personal
jurisdiction and subject matter jurisdiction.


105 See Avalon Bookkeeping Services Ltd v Furlong 2004 NLCA 46 (CanLII); 243 DLR (4th) 153 at paras 41-4
and Patrick Street Holdings Ltd v 11368 NL Inc 2024 NLCA 11 (CanLII) (Patrick Street) at para 370.

106 See Charafeddine v Morgan [2014] NSWCA 74 at paras 22-7 and Mayfield Development Corporation Pty Ltd
v NSW Ports Operations Hold Co Pty Ltd (No 4) [2024] FCA 538 at para 112.
107 United Manganese of Kalahari v Commissioner, South African Revenue Service and Four Similar Cases [2025]
ZACC 2; 2025 (5) BCLR 530 (CC); 2026 (2) SA 227 (CC) at paras 288-300.

ROGERS J
51
[109] There is nothing rare or exceptional about the circumstances in which the
Commission now finds itself. It is in exactly the same position as any other litigant who
decides not to pursue an appeal. The fact that this matter concerns the extraterritorial
reach of section 3(1) and the prosecut orial scope of competition authorities, as
submitted by the Commission, is not a sufficient consideration to relax res judicata.
This Court is not being asked to revisit its own earlier judgment. It is being asked to
revisit a substantive judgment of the CAC that could have been appealed but wasn’t.
This Court is, moreover, being asked to interfere with CAC I in circumstances where
there is no formal process to set aside or alter the orders in CAC I, such as an application
for leave to appeal or to rescind CAC I. This stands in contrast with Molaudzi and
Mothulwe. And this Court is being asked to do so in circumsta nces where neither the
Tribunal nor the CAC was invited to reconsider the questions now raised.

[110] The only remarkable feature is that the Commission sought to argue the
section 3(1) issue in this Court as if CAC I did not stand in its way. In their answering
affidavits in this Court, the respondents emphasised that the Commission was bound by
CAC I, variously invoking res judicata and peremption. Although the Commission
sought leave to file two replying affidavits on other aspects, it did not do so in relation
to the finality of CAC I. Nor did it respond to the repeated allegation by the respondents
that the Commission deliberately chose not to appeal CAC I but instead to comply
with it.

[111] We do not have before us an application or request by the Commission to relax
res judicata or overlook peremption. The question of rare and exceptional
circumstances for relaxing res judicata or the existence of sound reasons for
overlooking peremption in the interests of justice have not been ventilated in affidavits.

overlooking peremption in the interests of justice have not been ventilated in affidavits.
Departing from res judicata or overlooking peremption is not for this Court to do of its
own accord. When faced with an objection of res judicata or peremption, a litigant who
would otherwise be bound by a judgment or its litigation choices must seek the aid of
this Court and set out on affidavit the circumstances which justify a relaxation.

ROGERS J
52
[112] The importance of the legal question – the correct interpretation of section 3(1) –
is not an exceptional circumstance for purposes of overlooking res judicata and
peremption. If the Commission thought that the question required this Court’s attention,
an appeal against CAC I was the time-honoured way in which to do it. The Commission
chose not to pursue that option. Many cases come before the courts that involve
important legal questions of statutory interpretation, and cases heard by this Court are
almost invariably of that character. This cannot possibly be a reason for declining to
apply res judicata or for overlooking peremption. The fact that a court applies res
judicata or peremption does not mean that the legal question cannot be raised in another
suitable case. These doctrines do not foreclose future consideration of the legal
question; they simply bring finality as between the particular litigants.

[113] It is not unknown for a litigant who seeks to challenge a binding precedent to
raise the point in the lower fora, acknowledging that it must fail there but reserving the
right to pursue the point on appeal. If in a future case the Commission or a pure
peregrinus wished to contend that CAC I was wrongly decided (either because it
imposed too stringent or too lax a standard for personal jurisdiction), I have no doubt
that the matter could without undue delay be presented to the CAC and ultimately to
this Court for determination.

[114] The Commission has submitted that the banks would suffer no prejudice if the
section 3(1) issue were reopened. However, as some of the banks have pointed out in
their responding submissions, they have not had the opportunity of canvassing this in
affidavits. Moreover, there is some self -evident prejudice. The first round of
exceptions a nd objections, culminating in CAC I , was intended to establish the
principles with which the Commission had to comply to plead a proper case against the

principles with which the Commission had to comply to plead a proper case against the
banks. CAC I gave the Commission a final opportunity to do so in line with the
principles laid down in that judgment.

[115] A great deal of time and irrecoverable costs were expended since February 2020
in determining whether the superseding affidavit complied with CAC I. There were

ROGERS J
53
multiple exceptions and objection applications as well as exchanges of affidavits in the
two joinder applications. The issues were argued over six days in the Tribunal by legal
teams that included 33 advocates and many attorneys. The three members of the
Tribunal wrestled with the issues for 15 months before issuing their decision. Then
followed the appeal to the CAC, which was argued over five days with a similarly large
cast of lawyers. A number of the banks also brought review proceedings to challeng e
TRIB II. It was in February 2024, four years after CAC I was decided, that the
Commission in its application in this Court for the first time signalled that it wanted this
Court to reverse certain key legal conclusions underpinning the orders in CAC I. To
allow it to do so would bring the administration of justice into disrepute.

[116] Another factor militating against overlooking res judicata and peremption is that
this Court is being asked to decide issues in the absence of the full ventilation they
deserve. In regard to personal jurisdiction, the question on the merits is not only
whether section 3(1) dispenses with this requirement. The Commission in this Court
has an alternative contention if its main argument on section 3(1) fails. The alternative
argument is that the common law should be developed so that, in competition cases, a
connection to South Africa by virtue of participation in a scheme where the central
instrument was the South African currency is sufficient for personal jurisdiction. This
argument was not raised in the first or second rounds in the fora below, so we wo uld
have to address it at first instance.

[117] Moreover, if we were to hold that CAC I is not binding in relation to personal
and subject matter jurisdiction, it is not only the Commission that should be entitled to
argue for an outcome more favourable to it than CAC I. In CAC I, the banks argued that

argue for an outcome more favourable to it than CAC I. In CAC I, the banks argued that
the Tribunal had been right in TRIB I to hold that there was no basis on which the
common law could accommodate personal jurisdiction over the pure peregrini. The
banks have not resurrected that argument and have been willing to abide by CAC I. But
if we hold that CAC I should not be treated as binding, we would , in fairness, have to
reopen all the jurisdictional arguments to both sides. And regardless of what the parties

ROGERS J
54
chose to argue, this Court would need to be satisfied that it has come to the right
conclusion on all the possibilities.

[118] I thus conclude that res judicata should not be disapplied and peremption should
not be overlooked. The parties to this litigation are therefore bound by CAC I insofar
as section 3(1) is concerned. This involves no expression of opinion by this Court as to
whether the CAC’s interpretation of section 3(1) in CAC I is right. The whole point of
res judicata and peremption is that the later court cannot revisit the earlier judgment,
even though it might have been wrong.108 An allegation that the earlier court reached a
wrong decision is not an exceptional circumstance justifying the non-application of res
judicata.109

[119] Of course, whether CAC I was wrong in its interpretation of section 3(1) is hotly
contested. Although the Commission’s section 3(1) contentions have previously failed
in the Tribunal and CAC, its case in that respect is at least arguable. Given the important

108 See, for example, Mulkerrins v Pricewaterhouse Coopers [2003] UKHL 41; [2003] 4 All ER 1 (HL) at para 41
per Lord Millett:
“As I observed in Crown Estates Commissioners v Dorset County Council [1990] Ch 297, 305
res judicata (or to give it its full name estoppel per rem judicatam) is a form of estoppel which
gives effect to the policy of the law that the parties to a judicial decision should not afterwards
be allowed to re -litigate the same question, even though the decision may be wrong. If it is
wrong, it must be challenged by appeal or not at all. As between themselves, the parties are
bound by the decision, and may neither re -litigate the same cause of action nor re -open any
issue which was an essential part of the decision. The doctrine comes into its own only when
the decision is wrong; if it is right, it merely serves to save time and costs.”
109 MacDougall v Lake Country (District) 2012 BCCA 408 (CanLII) at paras 35-6:

109 MacDougall v Lake Country (District) 2012 BCCA 408 (CanLII) at paras 35-6:
“[35] The appellants, in my view have failed to raise any reasonable basis upon which a court
could refuse to apply the doctrine of issue estoppel in the case before us. The 1963 decision
was taken by a court of competent jurisdiction after full argument by the appellants’ predecessor
in title. The issue decided by the Court was not tangential to the arguments presented; rather it
was the very question that was argued before the Court. The applicant in the 1963 case had a
right to appeal, but did not exercise it.
[36] As I analyse the appellants’ argument, it is simply that the 1963 decision was wrong,
and that it would therefore be unjust to follow it. To accept that as a basis for refusing to apply
the doctrine of res judicata would be to eliminate the doctrine entirely. The doctrine of res
judicata exists precisely to obviate the re-litigation of issues. At least where a previous decision
is not patently perverse, the question of whether the court, in the end, reached the right
conclusion is not one which should be addressed in exercising discretio n to apply or not apply
the doctrine.”
See also Amgen Inc. v Pfizer Canada ULC 2020 FC 522 (CanLII) at para 160 and Patrick Street above n 105 at
paras 393-8.

ROGERS J
55
nature of the Commission’s complaint against the banks, it is regrettable that it did not
bring an appeal against CAC I. I do not express any view as to whether such an appeal,
if pursued, would have succeeded; that would entail a detailed analysis of a kind
foreclosed by res judicata and peremption. An appeal would, however, have allowed
an important issue to be finally resolved. Its absence is an opportunity missed.

[120] What I have just said should not be confused with the doctrine of precedent. If
in a future case the jurisdictional effect of section 3(1) should be an issue properly
before it, this Court as the highest court in the land could, as a matter of precedent,
overrule CAC I, just as it could, as a matter of precedent, overrule any other decision of
a court beneath it in the judicial hierarchy. In overruling such a precedent, this Court
does not reopen the case as between the litigants in the overruled preceden t.110 The
Commission could in another case ask the CAC itself to depart from its previous
decision by seeking to persuade that Court that CAC I was clearly wrong.

[121] It will thus be necessary to consider whether the superseding affidavit adequately
pleaded personal jurisdiction and subject matter jurisdiction in accordance with the ACF
test and QE test respectively, as well as the anterior question of whether this is a matter
engaging the Court’s jurisdiction.

[122] As I mentioned earlier, the fact that pure peregrini conspired among themselves
to manipulate the USD/ZAR exchange rate was rejected in CAC I for not being a
sufficient connecting factor to South Africa. The Commission has argued, however,
that a sufficient additional connecting factor would be if pure peregrini conspired
together with local peregrini, even if no link between the pure peregrini and South
African banks was properly alleged.

[123] If the superseding affidavit sufficiently alleged that the interactions between the

[123] If the superseding affidavit sufficiently alleged that the interactions between the
pure peregrini and the local peregrini involved the business activities of the local

110 See also Handley Spencer Bower and Handley Res Judicata 4 ed (LexisNexis, London 2009) at para 1.15.

ROGERS J
56
peregrini’s South African branches and offices, the argument might well be sustainable.
However, the affidavit does not allege that any of the traders for the local peregrini
worked at the South African branches or offices of those banks. Moreover, in pleading
jurisdiction, the Commission did not allege, as a relevant connecting factor, that the
pure peregrini were part of an SOC with local peregrini. The affidavit pleaded, as the
relevant connecting factor, that the pure peregrini were part of an SOC “together with
the South African respondents”. The “South African respondents” were earlier defined
as the incolae banks: SBSA, Investec, Absa, Nedbank and FRB.

[124] The ACF test, as formulated by the CAC, is concerned with real connections
between the suit and the local forum. If a pure peregrinus and a local peregrinus
conspire to manipulate the USD/ZAR exchange rate, with all relevant activity taking
place abroad through the banks’ foreign offices and traders , without any involvement
of the local peregrinus’ South African branch and South African-based traders, it would
seem to be an irrelevant coincidence that the local peregrinus has a South African
branch or office. On the other hand, there would arguably be a real connection between
South Africa and the suit if the pure peregrinus conspired with the South African-based
traders of the local peregrinus, since some of the conduct involving the formation and
implementation of the SOC would then have occurred in this country. In Bid Industrial
Holdings, for example, the Court said that appropriateness and convenience, for
purposes of adequate connection, were “elastic concepts which can be developed case
by case”, but that the “strongest connection would be provided by the cause of action
arising within that jurisdiction”. 111 It is considerations of this kind that might have
caused the Commission not to plead a conspiracy with local peregrini as an adequate
connecting factor.

connecting factor.

[125] I must emphasise that what I have just said concerns personal jurisdiction in
respect of pure peregrini, more particularly the adequacy of the connection between the
suit and South Africa for purposes of satisfying the ACF test. It is a different question

111 Bid Industrial Holdings above n 23 at para 56.

ROGERS J
57
whether the Tribunal has personal jurisdiction over the local peregrini themselves
purely by virtue of the existence of a South African branch or office, even though the
impugned conduct is not alleged to have been perpetrated by or through the South
African branch or office.

[126] In regard to personal jurisdiction over local peregrini, there is authority that, in
ordinary civil proceedings, our courts have jurisdiction over a local peregrinus company
in respect of any cause of action arising out of the business carried on by that company
at its South African branch. Such a company is regarded, for purposes of ordinary civil
jurisdiction over that particular suit, as residing in two places, na mely in the foreign
country and at the location of the South African branch.112 In the present case, however,
the Tribunal held that it had personal jurisdiction over the local peregrini merely by
virtue of the fact that they had local branches and offices, without any insistence that
the impugned conduct was associated with those branches and offices. None of the
local peregrini appealed that holding of personal jurisdiction to the CAC. The point
was not argued in this Court and I express no view on it. But as I shall presently explain,
the Tribunal’s reasoning has given rise to some confusion in relation to JPM Bank,
where a similar issue has been raised in relation to subject matter jurisdiction.

Joinder and initiation
[127] CSS, as the applicant for leave to appeal in CCT 27/24, and Nedbank and SAI,
as respondents in CCT 30/24, have raised various legal arguments as to why they could
not be joined.


112 Appleby (Pty) Ltd v Dundas Ltd 1948 (2) SA 905 (E ); [1948] 3 All SA 37 (E) (Appleby) at 909-11, relying
inter alia on TW Beckett & Co Ltd v H Kroomer Ltd 1912 AD 324 at 338-9; Skjelbreds Rederi A/S v Hartless (Pty)
Ltd [1982] 1 All SA 1 (W); 1982 (2) SA 739 (W) at 743D-744H; ISM Inter Ltd v Maraldo 1983 (4) SA 112 (T)

at 117A-H; and Lin v Minister of Home Affairs 2015 (4) SA 197 (GJ) (Lin) at para 88.

ROGERS J
58
Did CAC I preclude the addition of further respondents?
[128] The first argument, applicable to Nedbank and SAI, 113 is that the CAC I order
did not permit the Commission to add further respondents, i.e. , beyond the 18 initial
respondents and the five further respondents who were the subject of the first joinder
application. This argument, which was accepted in CAC II, involves the interpretation
of the CAC I order. Because the interpretation affects the Commission’s right to pursue
its cartel case against Nedbank and SAI, section 34 of the Bill of Rights is implicated
and it is a constitutional matter.

[129] At the time CAC I was argued, the question of the possible addition of further
parties was not before the Court. Unsurprisingly, therefore, CAC I did not deal with
that question. The order was intended to regulate what the Commission had to plead in
respect of the existing respondents. The expression “named respondents” in the CAC I
order merely emphasised that the requisite particularity had to be pleaded in respect of
each of the named respondents rather than in broad brushstrokes.

[130] It is so that CAC I did not grant the Commission leave to add further respondents.
No such leave was sought, and the CAC’s attention was thus not directed to that
question. I thus accept that the Commission was not entitled simply to name the
twenty-fourth to twenty-eighth respondents as additional respondents in the superseding
affidavit. However, if the Commission was otherwise entitled to bring an application
to join them, CAC I did not prohibit this from being done. The Court in CAC II thus
erred in holding that CAC I was a bar to joinder.

Was a fresh initiation needed?
[131] The next question is whether the Commission was entitled to join further
respondents, subsequent to the referral of February 2017, without a separate complaint

113 The argument would also apply to FRB, though it did not raise this as an objection.

ROGERS J
59
initiation against them. This argument is applicable to CSS, Nedbank and SAI. 114 It
raises an arguable point of law of general public importance which this Court should
consider. Our jurisdiction is thus engaged.

[132] In Pickfords, the Commission had initiated a complaint in November 2010
alleging collusive tendering in the furniture removal business and naming various firms.
Pickfords Removals SA (Pty) Limited (Pickfords) was not named.115 In June 2011 the
Commission issued a further initiation which named Pickfords and several others. 116
After the complaint was referred to the Tribunal, Pickfords filed an exception in which
it contended, among other things, that some of the complaints against it were time -
barred under section 67(1). This turned on whether the relevant initiation date, f or
purposes of the three-year period in that section, was November 2010 or June 2011.117

[133] This Court held that the relevant initiation was the complaint initiated in
November 2010. The Court emphasised that an initiation is against an alleged
prohibited practice. Self -evidently, this Court said, the Commission would have in
mind some of the firms potentially involved in the practice:

“But this does not mean that the names of all the firms or parties must be included. The
section emphasises ‘prohibited practices’ over firms or parties. There are no
formalities required, save for a decision by the Commissioner to cause the
commencement of an investigation into the alleged prohibited practice. The initial
omission of a firm or party at the stage when the complaint is first initiated by the
Commission and its subsequent addition to the complaint are not fatal, given the
wording of sectio n 49B(1) and the informality of the procedure.” 118 (Footnotes
omitted.)


114 It would also be applicable to HBUS, MLP, BANA, IBL and FRB, but none of those parties raise this as an
objection.
115 Pickfords above n 43 at para 5.
116 Id at para 6.
117 Id at para 8.

objection.
115 Pickfords above n 43 at para 5.
116 Id at para 6.
117 Id at para 8.
118 Id at para 21.

ROGERS J
60
[134] The Court stated that on the facts the Commission had expressly left open, in
the first initiation, the possible addition of further firms. This was the inference to be
drawn from the fact that the named firms in the first initiation were described as th e
“main companies” implicated in the alleged conduct. 119 Regarding the second
initiation, the Court said that there was no legal impediment to the Commission
amending a complaint initiation. The question was whether in that particular case the
second initiation was an amendment of the first one or a separate initiation. This Court
held that it was an amendment.120

[135] Although in Pickfords there was an amending initiation that named Pickfords, I
do not read this Court’s judgment as holding that such an amendment was essential.
The Court said only that it was permissible. The statement I have quoted from Pickfords
appears to me to convey that an initiation in respect of particular conduct and naming
certain firms is a valid initiation in respect of all firms whom the Commission’s
investigation thereafter identifies as having been guilty of the practice, even though they
were not initially n amed. If, as held in Pickfords, the relevant initiation date was
November 2010, even though Pickfords was not yet named, it follows that even without
an amendment to the initiation, Pickfords could not have complained if it was included
as a respondent in the referral to the Tribunal.

[136] It is important to bear in mind the manifest purpose of requiring an initiation in
respect of alleged prohibited conduct. Section 49B(3) provides that, upon initiating a
complaint, the Commissioner must direct an inspector to investigate the complaint as
quickly as practicable. An inspector has wide powers for purposes of conducting an
investigation. The complaint initiation circumscribes what the inspector may
investigate. The complaint initiation is not a process that commences legal proceedings

investigate. The complaint initiation is not a process that commences legal proceedings
against a firm. If a firm is named in the complaint initiation, it may not know that fact
until the case is referred to the Tribunal.

119 Id at para 23.
120 Id at para 27.

ROGERS J
61

[137] In the present case, the Commission’s first complaint initiation in April 2015
named 11 firms, not all of whom remained part of the complaint referral in
February 2017. The April 2015 initiation identified the prohibited conduct as
price-fixing of the exchange rate between the ZAR and several other currencies
including the USD. In August 2016 the Commission issued a further complaint
initiation. The accompanying initiation statement was styled an “amended” initiation
statement. It named another 12 banks. A further type of prohibited conduct was also
added, namely market division by allocating customers.

[138] The first and amended initiation statements did not expressly say that there might
be further firms implicated in the prohibited practices. They did not expressly exclude
that possibility either. Given the nature of the alleged practices, that 11 banks w ere
already identified in April 2015 and that by August 2016 another 12 had come to light,
the initiation statements should not be viewed as excluding the possibility that
investigation might yield yet further names. On my reading of Pickfords, the
August 2016 amendment would not have been necessary if its only purpose was to
identify 12 further banks. However, the amendment may well have been necessary in
order to allege additional prohibited conduct.

[139] In principle, therefore, it was not necessary, after August 2016, for there to be a
further initiation statement identifying CSS, Nedbank and SAI. The conduct of which
they are alleged to be guilty is the same conduct identified in the April 2015 and
August 2016 initiation statements. They could be added to the referral without the need
for yet another initiation.

Was the post-referral joinder of CSS, Nedbank and SAI permissible?
[140] What must now be considered, however, is whether what I have just said holds
true even though the Commission sought to add CSS, Nedbank and SAI only after the

true even though the Commission sought to add CSS, Nedbank and SAI only after the
complaint was referred to the Tribunal in February 2017. There are two strands to the
argument on this issue: first, that such a joinder is absolutely precluded by the Act;

ROGERS J
62
second, if such joinder is not absolutely precluded, that a fresh initiation against the new
parties is required. Pickfords does not directly address this issue, because there the
addition of the new parties took place before the referral to the Tribunal.

Is there a bar to post-referral joinder?
[141] SAI is the proponent of the argument that the Act prohibits joinder post-referral.
It emphasises the three stages in complaint proceedings: initiation, investigation and
referral. By the time the Commission decides to refer the complaint to the Tribunal, it
will have concluded its investigation and decided on the respondents it intends
prosecuting. The door cannot be left open indefinitely for the addition of further
respondents.

[142] SAI bolsters its argument with reference to the definition of “respondent” in
section 1 of the Act, namely “a firm against whom a complaint of a prohibited practice
has been initiated in terms of this Act”. SAI also relies on section 51(4), in terms of
which, upon referral of a complaint to the Tribunal, its chairperson must publish in the
Government Gazette a notice of the referral that includes “the name of the respondent”
and the “nature of the conduct” that is the subject of the referral.

[143] I am willing to assume, without so deciding, that upon referral to the Tribunal,
the Commission’s statutory powers of investigation in respect of the complaint come to
an end. But the fact that the Commission can no longer use its investigative powers
does not preclude the possibility that it may, during the referral phase, come into
possession of information indicating that yet further firms were involved in the
prohibited practice. Information may be revealed by the respondents’ answering
affidavits, through the process of discovery and further particulars, in consultation with
witnesses and in consequence of information volunteered by firms that have obtained
leniency or settled.

leniency or settled.

[144] Having regard to what I have said about Pickfords, the definition of “respondent”
must be understood as including a firm whose conduct falls within the scope of the

ROGERS J
63
initiated complaint and which has been identified as such, even though it was not named
in the initiation statement. The definition is neutral as to the point in time at which the
firm is so identified. As to section 51(4), the chairperson will naturally only be able to
include the names of those firms that are already respondents, but there is no necessary
implication that the addition of further respondents is precluded. The purpose of the
notice in the Gazette seems merely to give information to the p ublic about the referral
as it then stands.

[145] In my view, the architecture of the Act does not compel the conclusion that the
addition of parties after the referral is prohibited. As I have said, the processes that
occur after referral are of a kind that might reveal further implicated parties. In t erms
of Pickfords, these further parties would have fallen within the scope of the complaint
initiation up to the time of referral, and there is no reason why that should not continue
to be so after the referral. To interpret the Act in the way SAI argues is to reward firms
whose involvement in the prohibited conduct remains hidden the longest. Such an
intention cannot be imputed to the lawmaker.

Must there be a fresh initiation in the case of post-referral joinder?
[146] The argument that there must be a fresh initiation for a post -referral joinder is
based on the following propositions. A firm cannot be the subject of a referral to the
Tribunal unless its conduct has been investigated. If there has already been a referral
to the Tribunal, and the firm in question has not yet been identified as a respondent, that
firm’s conduct will not ha ve yet been investigated. In order for there to be an
investigation, there would need to be a preceding complaint initiation.

[147] This argument is unsound. In the circumstances posited, the additional firm’s
conduct falls within the scope of the complaint that was initiated before the referral to

conduct falls within the scope of the complaint that was initiated before the referral to
the Tribunal. What is investigated is the prohibited conduct. If, post -referral, the
Commission comes into possession of information indicating that an additional firm
was also part of the prohibited conduct, it will be adding the firm to a complaint that
has already been investigated. In the circumstances supposed by the ar gument, the

ROGERS J
64
Commission does not need to conduct a further investigation, because the information
will have come into its hands without the need for further investigation.

[148] The respondents’ argument rests on a high degree of formalism. Although the
Commission does not need to exercise its investigative powers to know that the
additional firm should be added, the argument insists that there should be a fresh
initiation followed by a notional investigation phase, even though such phase may exist
for only a brief moment in time. If these formal steps were taken, the case against the
additional firm could itself then be referred to the Tribunal, and the hearing of that case
could be consolidated with the hearing of the existing referral.

[149] Why should this roundabout procedure be needed? The only purpose served by
this formalism is to improve the additional firm’s chance of raising a successful time-bar
defence in terms of section 67(1): for the additional firm, the three -year period would
be reckoned from the date of the later initiation, not the earlier one. In this way, the
additional firm would be rewarded for the fact that its conduct remained h idden the
longest. It is an interpretation that would tend to frustrate rather than promote the
purposes of the Act. Those purposes include to detect and address behaviour within a
market that tends to impede, restrict or distort competition. 121 As I have said, the
scheme of the Act can comfortably accommodate the post-referral joinder of parties.

The section 67(1) time-bar
[150] The above conclusions mean that, in the present case, the respondents whose
joinder was sought after the February 2017 referral do not gain any advantage with
regard to section 67(1), which bars referrals in respect of prohibited practices that
ceased more than three years before the complaint was initiated.122 For them, as for all
the other respondents, the April 2015 and August 2016 initiations are the relevant

the other respondents, the April 2015 and August 2016 initiations are the relevant
trigger dates for price-fixing and market allocation respectively.

121 Section 2(g) of the Act.
122 Section 67(1) is quoted in n 38 above.

ROGERS J
65

Joinder procedure
[151] The fact that additional firms can be added post -referral does not mean that the
Commission is at liberty to do so unilaterally. Once the complaint has been referred to
the Tribunal, it is the Tribunal that regulates the further conduct of the proceedings. The
Act gives the Tribunal wide powers in that regard.123 TR 45 deals with the joinder and
substitution of parties. It was suggested in argument that TR 45 did not comfortably
cover the joinder of further respondents post-referral. Even if that were so, TR 55(1)
states that in cases of uncertainty as to practice and procedure, the presiding member of
the Tribunal may “give directions on how to proceed” and may for that purpose “have
regard to the High Court Rules”. In the High Court, it is a common occurrence for
parties to be joined as respondents or defendants subsequent to the institution of an
application or action.

The prerequisites for pleading an SOC and related questions
[152] None of the parties takes issue with the SOC test derived from EU competition
law and summarised in paragraph 28 of CAC II. The Commission contends, however,
that the CAC committed errors of law when it applied the tests for an SOC and
jurisdiction to the pleaded case, thereby departing from or adding unwarranted glosses
to the correct tests. The contention conveys that the CAC applied the wrong legal tests
in assessing the adequacy of the superseding affidavit, not merely that it misapplied the
right tests. If this is what the CAC did, it might raise arguable points of law of general
public importance that this Court should consider, since the CAC’s tests would bind the
Tribunal in future cases.


123 See section 27(1)(d), in terms of which the Tribunal may “make any ruling or order necessary or incidental to
the performance of its functions in terms of this Act”; section 31(5), in terms of which the Chairperson or an

assigned member of the Tribunal may make orders of an interlocutory nature that do not warrant a hearing before
a panel of three members; and section 55(1), in terms of which the presiding member “may determine any matter
of procedure for that hearing, with due regard to the circumstances of the case, and the requirements of
section 52(2)”. Section 52(2) provides that the Tribunal must conduct its proceedings in public, as expeditiously
as possible, and in accordance with the principles of natural justice, and provides further that the Tribunal may
conduct its hearings informally or in an inquisitorial manner.

ROGERS J
66
[153] To address the Commission’s contention about legal errors, it is necessary to
give a brief overview of how the Commission went about pleading the SOC:
(a) The “single anti -competitive economic objective” of the SOC was the
manipulation and distortion of normal competitive conditions in the
trading of the USD/ZAR currency pair through price -fixing and division
of markets.
(b) The “general and consistent terms” of the SOC were that the respondents’
traders would be in frequent and regular communication with each other
when trading in the currency pair. The purpose of such communication
was to coordinate trading activities, prov ide information and reach
understandings on trading strategies.
(c) The SOC involved (i) the fixing of prices in the currency pair in relation
to bid and offer prices, bid -offer spreads, the spot exchange rate and the
terms and margins for executing client orders at the FIX; 124 and (ii) the
division of markets through the allocation of customers in the currency
pair.
(d) The existence of the SOC, and its terms and objective, could be inferred
from conduct that implemented the SOC. The conduct entailed extensive
communication between traders over a lengthy period, continuity in the
mode of communication and the existence of permanent chatrooms on the
Bloomberg messaging platform in which the traders frequently
participated. The conduct also included unusual market behaviour
marked by the absence of random fluctuations, the use of round figures
for quotes and a consistent spread of 0.05 and 0.10 charged by the South
African banks.
(e) The chatrooms were pleaded to have been the primary mode of
communication between the traders. The Commission defined (i) a
“member” of a chatroom as a person who created or administered the
chatroom and anyone else who accepted an invitation to join it; and (ii) a

124 The FIX is the specified time each day at which large currency trades are usually transacted.

ROGERS J
67
“participant” in a chatroom as somebody who logged onto or opened the
platform, entered the chatroom and remained present as an active or
passive participant (an active participant being one who posted an instant
message or received a message in response to a message they previously
posted). Instant messages posted in a chatroom were visible to all
participants, whether or not they actively engaged in the communication.
(f) The Commission identified two chatrooms by name, the Old Gits
chatroom and the ZAR chatroom. Other chatrooms were simply referred
to as “implicated chatrooms”.
(g) The traders representing the various respondents were named. The
Commission alleged that there were other representatives of the
respondents who were authorised to trade in the currency pair and who
implemented the SOC, but whose details were unknown to t he
Commission.
(h) Each respondent was alleged to have joined the SOC at the time its
representatives engaged in any conduct implementing the SOC. In the
alternative, certain dates were pleaded in relation to the various
respondents. The Commission pleaded that it did not know the date on
which each respondent ceased to participate in the SOC.
(i) The Commission pleaded numerous specific incidents of communication
over the period 2007 to 2013 ( some 162 incidents), organised under the
headings “Sharing of information and understandings on bid -offer
spreads”, “Sharing information and arrangements to coordinate trading”,
“Sharing information and arrangements to consolidate and offset trades at
FIX”, “Manipulation of USD/ZAR FX rate” and “Sharing competitively
sensitive information”. Each inciden t specified the date, sometimes the
time (expressed as Universal Time Coordinated ( UTC) time, i.e. two
hours behind South Africa), and the implicated traders. If the incident
involved communication in a chatroom, the chatroom was identified.
Some of the incidents involved market behaviour observed on the

ROGERS J
68
“Reuters trading platform”. 125 The Commission pleaded that these
specific instances were not the only conduct undertaken by the traders to
implement the SOC.
(j) The alleged direct or immediate and substantial effects of the conduct
were also pleaded at length.

[154] At the request of this Court, the Commission’s counsel prepared a note
identifying the errors of law allegedly committed by the CAC. This request was made
because it was unclear from the Commission’s argument whether the true import of its
criticisms was merely that the CAC had misapplied the accepted test and reached factual
conclusions with which the Commission disagreed. Criticisms of that kind would on
the face of it not engage this Court’s jurisdiction.126 The Commission’s note identified
13 alleged errors of law (the numbering is my own), and I propose to deal with each
one in turn. The first two errors were said to relate to personal jurisdiction, the third
and fourth to subject matter jurisdiction, and the remaining errors to attribution i n
respect of the SOC.

[155] The respondents were granted leave to file notes in response to the Commission’s
note. Some of them complain that certain supposed legal errors were not pleaded as
grounds of appeal, either at all or in respect of the bank concerned. ANZ does so in
relation to the ninth error, Commerzbank in respect of the eleventh and thirteenth errors,
Macquarie in respect of the fifth error and Nedbank in respect of the thirteenth error. I
shall nevertheless deal with all the alleged errors on their merits.


125 Here and elsewhere, I place this term in quotation marks in the light of evidence contained in SBSA’s dismissal
applications. SBSA explained that there are two relevant Reuters systems: an anonymous trading platform, FX
Trading, used by most market-makers to transact peer-to-peer trades, with bids and offers being anonymous; and

the Reuters information platform, primarily a news outlet akin to Bloomberg, where trades are not executed but
where many institutions post indicative rates, usually by way of an automated feed that may update many times
in an hour. SBSA states that, in the light of the particularity provided by the Commission in the superseding
affidavit, the only reasonable inference is that what the Commission refers to as the Reuters trading platform was
in truth the Reuters information platform. Whether regard may be had to this evidence is a question I defer for
later consideration.
126 Villa Crop Protection (Pty) Ltd v Bayer Intellectual Property GmbH [2022] ZACC 42; 2023 (4) BCLR 461
(CC); 2024 (1) SA 331 (CC) at paras 37-9.

ROGERS J
69
First error – conflating SOC and personal jurisdiction requirements
[156] The Commission complains that the CAC erred by conflating the requirements
for pleading an SOC and for establishing adequate connecting factors for purposes of
personal jurisdiction. Paragraph 39 of CAC II is said to reflect this error. In that
paragraph, the CAC said the following:

“It is important to emphasise that care must be taken not to conflate subject matter and
personal jurisdiction. The point of [CAC I] was that, if the jurisprudence regarding
personal jurisdiction was developed beyond that of the constraints of the common law,
it was important in this case that, at the very least, the overarching conspiracy pleaded
by the Commission should show that all of the banks were connected to that
overarching conspiracy; that all were participants in an overarching conspiracy
designed to have a detrimental effect on the South African economy by virtue of their
joint conduct. To plead an SOC meant having careful regard to the principles set out
in Team Relocations, thereby ensuring a clear showing that the available evidence fell
within these principles. This was key to establishing personal jurisdiction over the
peregrini respondents.”127

[157] I reject this criticism. In accordance with CAC I, personal jurisdiction over pure
peregrini required the Commission to plead more than that the SOC involved the ZAR.
In CAC I it was foreshadowed that an allegation that the pure peregrini were involved
in an SOC with South African banks might meet the ACF test. The Commission in the
superseding affidavit expressly pleaded this as a connecting factor for purposes of
personal jurisdiction. Accordingly, in the case of pure peregrini, the question was not
merely whether those banks were part of an SOC. The Commission had to plead that
the SOC in q uestion also included South African banks. This was the point that the
CAC was making.


127 CAC II above n 1 at para 39.

ROGERS J
70
Second error – requiring South African linkages
[158] The Commission contends that the CAC erred in requiring an SOC in which
foreign banks had effectively “entered into business with South African banks” and
where there were linkages to South African banks as part of the SOC . The phrase
“entered into business” comes from paragraph 85 of CAC II.

[159] The impugned phrase does not point to a legal error. The CAC, which itself
placed the phrase in quotation marks, was merely expressing in vivid language that, in
order for there to be personal jurisdiction over the pure peregrini, it had to be shown
that the pure peregrini were part of an SOC that included South African banks. If that
were properly alleged, all the members of the SOC would have been in the “business”
of advancing the alleged anti-competitive objective pleaded by the Commission. CAC
II as a whole demonstrates that the Court was concerned with linking the pure peregrini
in an SOC to which South African banks also belonged. This was the key connecting
factor which the Commission had alleged, in line with CAC I.

Third error – subject matter jurisdiction and local branches (JPM Bank)
[160] The third alleged error relates only to JPM Bank. In the superseding affidavit,
the Commission pleaded that JPM Bank conducted the business of a bank in
South Africa through a branch in Johannesburg as defined and authorised by sections 1
and 18A of the Banks Act, 128 and that the Johannesburg branch was authorised to deal
in foreign exchange in South Africa in accordance with the Exchange Control
Regulations.129 The affidavit further alleged that the Tribunal had jurisdiction over
JPM Bank because of the foregoing and also because it participated in an SOC
involving the ZAR, which included South African banks.

[161] The Commission contends that the CAC erred in holding that, to establish subject
matter jurisdiction over JPM Bank, the Commission had to plead facts to show that its

128 94 of 1990.

128 94 of 1990.
129 Exchange Control Regulations, GN R1111 GG 123, 1 December 1961.

ROGERS J
71
cause of action arose out of business carried on at JPM Bank’s Johannesburg branch or
was attributable to its authorised dealers in South Africa. This is indeed what the CAC
held:

[124] “Significantly, these parties [JPM Co and JPM Bank] did not appeal the finding
that the Tribunal had personal jurisdiction over this case. The case [of] these
respondents was therefore based on the failure to establish subject matter jurisdiction
on the part of the Commission. The case of [JPM Bank] was that the Commission was
required to plead facts to establish that the cause of action, being a participant in the
SOC, was based on business carried on by [JPM Bank’s] Johannesburg branch or that
the collusive conduct was attributable to its appointment as an authorised dealer in
South Africa.
[125] In relation to [JPM Bank], the Commission pleaded that Mr Akshay Aiyer and
Mr Paul Simister were the two individuals whose conduct was evidence of
participation by [JPM Bank] in the SOC. However, at no point is there any allegation
that either Mr Aiyer or Mr Simister worked for [JPM Bank’s] South African branch
and exercised the powers of an authorised dealer within South Africa. Nor is there any
evidence which attributes their conduct to [JPM Bank] in respect of participation in the
SOC. In summary, the Commission did not plead facts to establish that the cause of
action arose out of business carried on with [JPM Bank’s] Johannesburg branch or that
the collusive conduct was attributable to its appointments [as authorised dealers] 130 in
South Africa.”

[162] I have difficulty understanding why this point was thought to relate to subject
matter jurisdiction, rather than personal jurisdiction. The difficulty can be traced back
to TRIB I. The Tribunal found that there was personal jurisdiction over JPM Bank and
similarly-placed local peregrini by virtue of their South African branches. When it
turned to subject matter jurisdiction, the Tribunal considered effects -based antitrust

turned to subject matter jurisdiction, the Tribunal considered effects -based antitrust
jurisdiction in foreign law and adopted the QE test. It concluded that, as against the
local peregrini (including JPM Bank), the referral affidavit did not adequately plead
that the foreign conduct of these banks had qualified effects in South Africa.


130 The CAC wrote “and authorised deal” at this point, which is plainly a typographical error.

ROGERS J
72
[163] TRIB I did not require the Commission to plead that the SOC cause of action
against JPM Bank was based on business conducted through its Johannesburg branch.
What TRIB I required was that the Commission plead that JPM Bank’s foreign conduct
in pursuance of the SOC had qualified effects in South Africa. This is clear from the
hypothetical case sketched by the Tribunal in paragraph 111 of its reasons. The
Tribunal referred to two local peregrini who engaged in price-fixing of the USD/ZAR
exchange rate “through an offshore office”. A particular price-fixing transaction might
only affect an offshore customer, in which case the QE test would not be met. But if
the price-fixing transaction affected the price paid by South African consumers, the QE
test would be met, so reasoned the Tribunal.

[164] JPM Bank presumably made a similar argument in CAC II to the one it has made
in this Court. JPM Bank refers to a statement in TRIB I that, under the common law, a
court has jurisdiction over a peregrinus that conducts business in South Africa “in
respect of any cause of action which arose out of its activities here”. This is indeed
what the Tribunal said, 131 and in a footnote it referenced three South African cases in
support of that view: Appleby,132 Bisonboard133 and Lin.134 However, the Tribunal said
this in relation to personal jurisdiction, not subject matter jurisdiction.

[165] If the Tribunal was intending to apply this principle, it should have held that, for
purposes of personal jurisdiction, the Commission could only rely on the existence of a
local peregrinus’ South African branch if the relevant SOC conduct of that bank was
committed through its South African branch. Inconsistently, however, this was not the
Tribunal’s decision. It found that it had personal jurisdiction over JPM Bank and other
local peregrini merely by virtue of the existence of the South African branches, without

local peregrini merely by virtue of the existence of the South African branches, without
any requirement that the SOC conduct of those banks was attributable to the
South African branches.

131 TRIB I above n 3 at paras 71 and 80.
132 Above n 112 at 910 and 912.
133 Bisonboard Ltd v K Braun Woodworking Machinery (Pty) Ltd 1991 (1) SA 482 (A) at 496-7.
134 Above n 112 at paras 88-96.

ROGERS J
73

[166] One may question whether the Tribunal was wrong or right in this respect, but
the fact is that JPM Bank did not appeal the decision in TRIB I that the Tribunal had
personal jurisdiction over it. Because CAC I endorsed the QE test in relation to subject
matter jurisdiction, all that the Commission needed to plead, in respect of JPM Bank,
was that its SOC conduct had qualified effects in South Africa. The Commission did
not need to plead that the qualified effects were brought about by the conduct of
JPM Bank’s South African branch. No such ruling had been made in TRIB I or CAC I.
On the contrary, TRIB I clearly envisaged that it was JPM Bank’s offshore conduct that
would need to have had qualified effects in South Africa. If the conduct in question had
been the conduct of JPM Bank’s Johannesburg branch, qualified effects would have
been irrelevant, because one would then have been dealing, for purposes of section 3(1),
with “economic activity” in South Africa, and JPM Bank would then have been in the
same position as the South African banks.

[167] I am therefore satisfied that the CAC erred in law in holding that, for purposes
of subject matter jurisdiction, the Commission needed to plead that the SOC cause of
action against JPM Bank was based on business carried on by its Johannesburg branch.
It is important that we correct this error, otherwise in future the Commission may be
hamstrung in prosecuting anti -competitive conduct by foreign entities with
South African branches.

[168] This correction pertains only to subject matter jurisdiction. The effect of this
judgment will be to restore the proposition in TRIB I that, in the case of a peregrinus
(whether pure or local), subject matter jurisdiction requires that the anti -competitive
offshore conduct of such a firm needs to have had qualified effects in South Africa.

[169] Whether the mere existence of a South African branch, unconnected to the anti-

[169] Whether the mere existence of a South African branch, unconnected to the anti-
competitive conduct, is sufficient for purposes of personal jurisdiction, as was held in
TRIB I, is not a question that is before us. In this particular case it might be academic.
The superseding affidavit alleged that JPM Bank was party to an SOC that targeted the

ROGERS J
74
ZAR and included South African banks. On the approach adopted in CAC I and CAC II,
this would satisfy the ACF test for a local peregrinus, just as it would for a pure
peregrinus. If the allegations in respect of JPM Co, a pure peregrinus, satisfied the
ACF test (which JPM Co has evidently accepted, since it has not pursued an appeal),
the same would apply to JPM Bank, since the Commission’s allegations were made
against the JPM entities in the alternative.

Fourth error – subject matter jurisdiction in respect of local banks
[170] The Commission contends that the CAC erred in holding that, for purposes of
subject matter jurisdiction over incolae, it was insufficient for the Commission to show
that the incolae engaged in economic activity in South Africa. The Commission points
to paragraph 173 of CAC II, where the Court said this when concluding its discussion
of the case against SBSA:

“In the case against [SBSA] the skeletal nature of [the] allegations reveals that there is
no basis by which its activity fell within the scope of subject matter jurisdiction. To
find, as the Tribunal did, that there are disputes that only can be determined after the
benefit of a full hearing bears little relationship to that which is contained in the referral
affidavit. The case does not get out of the legal starting blocks.”

[171] The CAC’s use of the expression “subject matter jurisdiction” in this passage is
unfortunate. It is clear from the preceding paragraphs of the CAC’s judgment 135 that
what the CAC was assessing was whether the superseding affidavit made out a cause
of action against SBSA, namely a case that SBSA was a participant in the alleged SOC.
SBSA’s dismissal application did not include, as a ground, that the affidavit did not
sufficiently plead subject matter jurisdiction. Its central objection was that the affidavit
did not sustain an inference that SBSA participated in the SOC, and it was this point

did not sustain an inference that SBSA participated in the SOC, and it was this point
that the CAC was addressing. The CAC did not mention section 3(1) or qualified effects
or consider the allegations in the affidavit concerning qualified effects, as it would have
done were it addressing subject matter jurisdiction properly so called.

135 CAC II above n 1 at paras 163-72.

ROGERS J
75

[172] The CAC was thus using “subject matter jurisdiction” in a very loose sense, as
the equivalent of “cause of action”. It did the same thing when discussing the cases
against HBEU136 and CSS.137 In their cases, too, the CAC analysed the allegations in
the superseding affidavit and concluded that they had a “case to answer”. There was no
discussion of qualified effects as pleaded in the affidavit. In fact, in none of the
instances where the CAC found for the respondents did it do so on the basis that the
affidavit failed to plead qualified effects sufficiently. If the CAC, in the cases of SBSA,
HBEU and CSS, had been using “subject matter jurisdiction” in its technical sense, it
could not have found in favour of SBSA but against HBEU and CSS. This is so because
the qualified effects pleaded in the affidavit in support of subject matter jurisdiction
properly so called did not differentiate between the banks. The Commission pleaded
the qualified effects that the SOC as a whole had in South Africa.

[173] There was thus no legal error by the CAC, only an inaccurate use of the
expression “subject matter jurisdiction”.

Fifth error – failure to keep SOC requirements distinct
[174] The Commission argues that the CAC conflated two separate questions: (a) a
firm’s contribution to the overall objective of the SOC; and (b) the attribution of the
SOC to that firm. According to the Commission, the CAC ought to have engaged in
two separate enquiries. The first enquiry was to establish whether the pleaded conduct
of a firm fell within the scope of the SOC. The second was to establish the question of
attribution, namely whether the firm had actual or constructive knowledge of the scope
and general characteristics of the SOC. This error is said to permeate the entire
judgment.


136 Id at para 135.
137 Id at para 139.

ROGERS J
76
[175] I reject this criticism. In Team Relocations , the Court of Justice for the
European Union (CJEU) outlined three discrete but related requirements in relation to
an SOC: first, the existence of an overall plan pursuing a common objective; second,
the intentional contribution of a firm to the overarching plan; and third, actual or
constructive knowled ge of the conduct of the other participants. 138 The European
authorities do not require a court to consider these three basic requirements for an SOC
in hermetically-sealed compartments. Moreover, the second SOC requirement is not
that a firm’s pleaded conduct “fell within the scope of the SOC”. The requirement is
that the firm participated in the SOC through an intentional contribution by its own
conduct to the SOC’s common objective.

[176] For example, there might be evidence that two firms conspired to fix the
USD/ZAR exchange rate on a particular date. Viewing that conduct in a purely
objective way, one might say that the conduct is the same as conduct carried out by
participants of the alleged SOC. This, however, says nothing about whether the two
firms were participants in the SOC, since their conduct is equally consistent with purely
bilateral collusion. There must be evidence that, in acting as they did, the firms were
intentionally contributing to the common objectives of the participants in the SOC.
Intentional contribution requires knowledge of the SOC.

[177] Intentional contribution, in accordance with the second SOC requirement, is not
the same as the actual or constructive knowledge that features in the third requirement.
The third requirement is that the firm was aware of the actual conduct planned or put
into effect by the other firms in pursuit of the same objectives, or that it could reasonably
have foreseen such conduct and was prepared to take the risk as to whether it happened
or not. 139 This third requirement concerns the extent of one firm’s liability for the

or not. 139 This third requirement concerns the extent of one firm’s liability for the
conduct of other participants in the SOC. A firm (X) may intend to contribute to an
SOC in accordance with the second requirement without having actual or constructive

138 Team Relocations above n 52 at para 51.
139 See Fresh Del Monte Produce v Commission [2015] EUECJ C-293/13; EU:C:2015:416; ECLI:EU:C:2015:416
at para 157.

ROGERS J
77
knowledge of everything else that the other firms were doing in pursuit of the same
objectives. X may, for example, be aware that A, B and C are participants in the SOC
but be unaware that A, B and C have co-opted a fifth firm, D, into the SOC. If X could
not reasonably have foreseen this, it will not be liable for D’s conduct in pursuit of the
common objectives. Similarly, if X did not know and could not reasonably have
foreseen that A would perpetrate conduct of a particular kind in pursuit of the common
objectives, it would not be liable for that conduct.

[178] The EU doctrine of the SOC is ultimately a tool for holding each firm liable for
the conduct of all the conspirators, and the third requirement is the element which
moderates the extent of each firm’s liability for the conduct of its co -conspirators.
Before, however, one reaches the question of the extent of each firm’s liability for the
conduct of the others, it must be shown that the firm knowingly joined the conspiracy.
This is the second requirement for an SOC, an intentional contribution to the objectives
of a known SOC.

[179] The CJEU has noted that the mere fact that there is identity of object between an
agreement in which a firm participated and a global cartel does not suffice for a finding
that the firm participated in the global cartel. 140 Article 101(1)141 of the Treaty on the
Functioning of the European Union 142 (TFEU) does not apply unless there is a
“concurrence of wills between the parties concerned”.143


140 Sigma Tecnologie di rivestimento Srl v Commission of the European Communities [2002] EUECJ T-28/99 at
para 45.
141 It reads:
“The following shall be prohibited as incompatible with the internal market: all agreements
between undertakings, decisions by associations of undertakings and concerted practices which
may affect trade between Member States and which have as their object or effect the prevention,
restriction or distortion of competition within the internal market.”

restriction or distortion of competition within the internal market.”
142 Treaty on the Functioning of the European Union, 13 December 2007.
143 Toshiba v Commission [2015] EUECJ T-104/13; EU:T:2015:610; ECLI:EU:T:2015:610; [2015] 5 CMLR 21
at paras 52-6.

ROGERS J
78
[180] In UBS144 the CJEU said this:

“[I]t is settled case -law that an undertaking which has participated in a single and
continuous infringement through its own conduct, which meets the definition of
‘agreements’ or ‘concerted practices’ for purposes of Article 101(1) [of the] TFEU and
was intended to help bring about the infringement as a whole, may also be liable in
respect of the conduct of other undertakings in the context of the same infringement
throughout the period of its participation in the infringement. That is the position where
it is shown that the undertaking intended, through its own conduct, to contribute to the
common objectives pursued by all the participants and that it was aware of the
offending conduct planned or put into effect by other undertakings in pursuit of the
same objectives or that it could reasonably have foreseen it and was prepared to take
the risk.”145

[181] The distinction between the second and third SOC requirements can be
illustrated with reference to the UBS case. The European Commission (EC) found that
there was an SOC implemented through interactions in two chatrooms called the
Cods & Chips chatroom and the DBAC chatroom. One of the firms, Bank of America,
actively and intentionally contributed to the SOC’s objective through its participation
in the Cods & Chips chatroom. Some of the firms that participated through that
chatroom also participated in the DBAC chatroom. The EC found, however, that
Bank of America could not be held liable for the conduct perpetrated through the
DBAC chatroom. This was because it could not be found with sufficient certainty that
Bank of America was aware or ought reasonably to have foreseen and was willing to
take the risk of the existence and functioning of the DBAC chatroom.146

[182] Furthermore, the CJEU observed:


144 UBS Group and UBS v Commission [2025] EUECJ T-441/21; ECLI:EU:T:2025:337; EU:T:2025:337. This

was the case in which various firms sought the annulment of the European Commission’s decision in Case
AT.40324 – European Government Bonds (20 May 2021) ( EG Bonds). The EG Bonds case was the subject of
much discussion in the written submissions in this Court. The UBS case featured only in oral argument.
145 UBS id at para 126.
146 Id at paras 1045-97 and 1131-7.

ROGERS J
79
“[T]he fact that collusive practices relate to the same product and the same activity, the
fact that the same means of communication are used, and the fact that the same
individuals are involved are not, taken in isolation, sufficient to demonstrate the
existence of a common plan in pursuit of a single anti -competitive aim and, therefore,
of a single infringement.
However, assuming that they are established, those objective elements are relevant to
assessing the existence of a single infringement and, taken as a whole, are capable of
confirming that the Cods & Chips and DBAC chatrooms were linked and
complementary in nature and sought to achieve the aims pursued by the common plan
found by the [EC].”147

[183] In analysing the sufficiency of the case pleaded against each bank, the CAC’s
focus was the second SOC requirement, namely an intentional contribution to the SOC.
In regard to the banks it exonerated, the CAC did not reach the third requirement,
because it found in each case that the superseding affidavit failed adequately to plead
that the bank in question had intentionally contributed to a known SOC or, in the case
of the pure peregrini and for purposes of personal jurisdiction, that the bank had
contributed to a known SOC that included South African banks.

[184] It needs to be clearly understood that the acceptance by the CAC of European
SOC principles does not mean that the CAC has created a new form of anti-competitive
conduct outside the framework of the Act. The SOC concept is a particular way of
applying section 4(1)(b) of the Act. This means that it must still be found that the
restrictive horizontal practices (here, the fixing of prices and the dividin g of markets)
were the subject of an “agreement” or “concerted practice” by the firms concerned , as
is required by that section.

Sixth error – insistence on regularity (Nomura and Macquarie)
[185] The Commission contends that the CAC erred in applying the requirement of

[185] The Commission contends that the CAC erred in applying the requirement of
regularity as a decisive factor in determining whether a respondent’s conduct fell within

147 Id at paras 1099-100.

ROGERS J
80
the scope of the SOC. This legal error was said to have manifested itself in the CAC’s
reasoning in respect of Nomura, 148 Macquarie149 and JPM. 150 This is a convenient
juncture at which to consider the Commission’s pleaded case against these three banks.

[186] The superseding affidavit contains the following Nomura-specific allegations:
(a) Nomura’s traders were Messrs Guido Arlan and Darren Dempsey.
(b) Nomura joined the SOC “by at least 18 October 2012”, when
Mr Dempsey was a participant in an implicated chatroom.
(c) On 27 May 2010, between 16h30 and 16h49 and on the Reuters trading
platform, SBSA and Absa, together with Barclays, Nomura and other
banks, “held” the USD/ZAR exchange rate around the “focal point of
7.5760”. SBSA and Absa, together with HSBC and other ba nks,
“pushed” the rate to 7.5600. SBSA and Absa, together with Barclays ,
HSBC and other banks, then “returned” the rate to a focal point of 7.5600.
The “whole shift” was 0.0160.
(d) In regard to this event of 27 May 2010, it can be inferred (alleges the
Commission) that in this period the banks held ZAR in their accounts,
while collecting orders to buy the ZAR from customers: “[w]hen the time
came, they simultaneously bought the ZAR, strengthening it to the focal
point. At this point, they offloaded the ZAR to gain USD profits. The
reverse process was then coordinated, weakening the ZAR, from which
they gained ZAR profits”.
(e) On 2 March 2012 M r Jason Katz (then with BNP) and
Messrs Akshay Aiyer (JPM), Christopher Cummins (Citibank),
Nicholas Williams (Barclays) and Arlan (Nomura) were participants in
an implicated chatroom in which the following communication took
place: “Williams informed the traders that [a customer] was asking three
people for a price and the traders should be aware of this.”

148 CAC II above n 1 at para 185 read with para 102.
149 Id at paras 121-2.
150 Id at para 128.

ROGERS J
81
(f) Also on 2 March 2012 the same traders were participants in an implicated
chatroom in which Mr Aiyer “shared information about customer
quotes”.
(g) On 18 October 2012 Messrs Dempsey (Nomura) and Duncan Howes
(Absa) were participants in an implicated chatroom “in which they
discussed the bid-offer spread for USD/ZAR”.

[187] In upholding Nomura’s appeal, the CAC reasoned thus:151
(a) The difficulty for the Commission was that its entire case for Nomura’s
alleged participation in the SOC rested on three unrelated chats over a
six-year period.
(b) Two of the chats, both on 2 March 2012, happened before the date on
which Nomura was alleged to have joined the SOC. It was “regrettable”
that the Tribunal had overlooked this.
(c) The allegation in respect of the chat on 17 October 2012 was “so skeletal
and vague” that it was hardly the evidence needed to sustain a case of
personal jurisdiction (by which the CAC plainly meant evidence to
sustain the conclusion that Nomura was party to an SOC with one or more
South African banks).
(d) This was particularly so in view of the fact that Mr Dempsey’s interaction
was with an Absa trader. If there was evidence linking Nomura to a
South African-based SOC, this would have found its way into the
superseding affidavit , since Absa had been granted leniency and was
cooperating with the Commission. The silence was “instructive”.
(e) The case against Nomura “luminously reveal[ed]” the difficulty of the
Commission’s ambitious case. The Commission had to establish that
Nomura knew or ought to have known that it was participating in conduct
forming part of an SOC, that it knew or ought to have known the essential

151 Id 1 at paras 100-8.

ROGERS J
82
features of the SOC and that it intended to contribute to achieving the
SOC’s overall objective.
(f) The CAC quoted a passage from the CJEU’s decision in HSBC152 where
the latter Court emphasised that a firm that participated in only some of
the forms of anti -competitive conduct comprising the single and
continuous infringement, but which was aware or should have been aware
of all the other conduct planned or put into effect in pursuit of the same
objectives, could be held liable for all the conduct.153
(g) The CAC concluded that the three instances alleged were insufficient to
meet the requirements for Nomura’s participation in the SOC.

[188] The superseding affidavit contains the following Macquarie-specific allegations:
(a) Macquarie’s traders were Messrs Mark Chia, Chris Harkins,
Jason Atkins, Bevan Murray, Luke Fryday and Tim Donnelly.
(b) Macquarie joined the conspiracy “by at least 11 September 2013”, when
its traders were participants in an implicated chatroom.
(c) On 11 September 2013, these traders , together with Mr Thulani Kunene
(Absa) and “Bhana and Naidoo” (their genders, full names and the banks
they represented are not pleaded), were participants in an implicated
chatroom “in which Chia and Kunene discussed bid -offer spread for
USD/ZAR” and “Murray shared competitively sensitive information ”.
(This is the last date of the specific incidents catalogued in the
superseding affidavit and evidently forms the basis of the Commission’s
allegation that the SOC endured until at least September 2013.)

[189] In upholding Macquarie’s appeal, the CAC reasoned as follows:154

152 HSBC Holdings plc v European Commission [2019] EUECJ T-105/17; ECLI:EU:T:2019:675; EU:T:2019:675;
[2019] 5 CMLR 21.
153 Id at para 199.
154 CAC II above n 1 at paras 119-23.

ROGERS J
83
(a) The sole Macquarie -specific allegation in the superseding affidavit was
the chat of 11 September 2013.
(b) On its own, and at best for the Commission, this would establish a
prohibited practice between Macquarie and Absa involving a single
instance of information-sharing on a single day.
(c) The Commission had sought to “salvage” its case against Macquarie by
arguing that anyone who became a participant in a chatroom “somehow
would know or should reasonably have known [that] the chatroom was
used as a primary mode of communication to implement the terms and
further the objectives of the SOC”. The difficulty, however, was that
there was nothing in the affidavit from which those conclusions could be
inferred.
(d) Moreover, the affidavit failed to distinguish between information in
relation to currency trading that was in the public domain and information
to which only participants in the SOC were privy.
(e) It was difficult to see how the case could be made that a bank participated
in the SOC as pleaded on the basis of a single chat with an Absa trader.
Once again, there was the absence of information that Absa as a leniency
applicant would have provided, if it had existed.
(f) In pleading the effects of the SOC, the affidavit contained no reference to
Macquarie. But even leaving aside subject matter jurisdiction, there was
“simply insufficient evidence to connect Macquarie to the overall SOC,
which was an essential requirement of [ CAC I ] to establish personal
jurisdiction”.

[190] The superseding affidavit contains the following JPM-specific allegations:
(a) Its traders were Messrs Aiyer and Paul Simister.
(b) JPM joined the SOC “by at least 19 July 2011” when Mr Aiyer
participated in the ZAR chatroom. (The affidavit does not in fact refer to
any incident of 19 July 2011. Mr Aiyer is alleged to have participated in

ROGERS J
84
an incident on 19 July 2012, but this is not the first instance allegedly
implicating JPM.)
(c) Of the 162 specific incidents catalogued in the affidavit, JPM through one
or other of its traders is alleged to have participated in 71 of them,
spanning the period 26 April 2011 to 26 April 2013. These were all
incidents involving chatrooms, of which eight took place in the ZAR
chatroom.

[191] It is unnecessary to deal with the CAC’s reasoning in respect of the JPM entities.
The CAC found against JPM Co, which has not appealed that decision to this Court.
The CAC found in favour of JPM Bank, but on a basis which, in an earlier part of this
judgment, I have found to be legally unsound. The Commission seems to have referred
to JPM in the context of the sixth error only in order to contrast the CAC’s approach to
JPM on the one hand and to Nomura and Macquarie on the other. In respect of JPM Co,
the CAC said that the “regularity” with which Mr Aiyer participated in implicated
chatrooms was a prima facie basis for concluding that JPM Co participated in the
alleged SOC and had a case to answer.155

[192] In respect of the supposed sixth error, the Commission has also referred to
paragraph 185 of CAC II, which was part of the CAC’s summation after dealing with
all the individual banks. There, the CAC said:

“This Court has emphasised that there are clear separate requirements to establish
personal and subject matter jurisdiction. In the case of the pure peregrini both
requirements must be established in order for the referral to meet the requisite legal
standards. It was made clear in [CAC I] that the Commission was required to allege
that there were adequate connecting facts between the parties and the jurisdiction of
the Tribunal sufficient to establish personal jurisdiction against all of the named
respondents. As repeatedly emphasised that is an onerous requirement. The reference
to occasional participation in a chatroom without any additional evidence and where

to occasional participation in a chatroom without any additional evidence and where
there was no link to any South African bank is inadequate to meet the test as set out in

155 Id at para 128.

ROGERS J
85
the [CAC I] order. Accordingly, the Commission has failed to show the requisite
personal jurisdiction in the case of the fifth respondent (ANZ), ninth respondent
(Nomura), twelfth respondent (Commerzbank), thirteenth respondent (Macquarie) and
nineteenth respondent (HBUS).” (Emphasis added.)

[193] To come back to the supposed error, there is nothing in the CAC’s reasoning to
show that it held, as a matter of law, that a particular bank had to have engaged in
frequent interactions in order to justify an inference that it participated in the SOC. The
CAC was concerned only with the inferences that could be drawn from the facts alleged
in respect of Nomura and Macquarie. The fact that a particular bank was alleged to
have participated in an implicated chatroom on only one, two or three occasions is
plainly relevant in the process of inferential reasoning. The more numerous the chats,
and the wider the array of traders with which a particular bank engaged in those chats
(as with JPM Co), the greater would be the likelihood of drawing an inference again st
the bank. The Commission itself pleaded that frequent and regular communication was
the hallmark of this particular SOC.

[194] The sentence I have emphasised from paragraph 185 of CAC II makes it clear
that the CAC did not reject the possibility that a bank could be found to have participated
in the SOC even though it was only directly implicated in a few chats. But then there
needed to be some additional evidence to justify the inferenc e. This evidence was
lacking in the case of both Nomura and Macquarie.

Seventh error – insistence on chatroom participation (FRB)
[195] The Commission contends that the CAC erred in law by finding that the
superseding affidavit needed to contain allegations that a firm’s employee or
representative was a member of the chatroom in order to justify a finding that the
pleaded conduct fell within the SOC. The CAC’s reasoning in respect of FRB is said
to manifest this error.

ROGERS J
86
[196] The superseding affidavit contains the following FRB-specific allegations:156
(a) No FRB traders are named.
(b) FRB joined the SOC “by at least 28 May 2010 ”, when it and other
South African banks “affirmed a focal point” of 7.5750.
(c) The affidavit alleges eight incidents (or six, if one includes incidents that
occurred on the same day and in close time proximity) over the period
January 2009 to September 2012, in which conduct by various banks that
included FRB was observed on the “Reuters trading platform”.
(d) On 11 January 2009, between 20h59 and 21h48, various banks “engaged
in a pattern of price formation”. SNY/SAI and UBS quoted the same ask
and bid prices of 9.8100 and 9.7800, respectively, resulting in an identical
spot exchange rate and spread of 0.03. FRB and Nedbank thereafter also
quoted identical bid prices, the ask price only differing “by a pip”, leading
to identical spot exchange rates and spreads. FRB matched SNY/SA I at
the bid price of 9.7600 and ask price of 9.8599. (I note that UBS is not
alleged to have been a party to the SOC.)
(e) On 28 May 2010, between 07h23 and 07h24, FRB, SBSA, Absa and
Investec “colluded in posting quotes in succession which had the effect
of reducing volatility in the exchange rate”. They “maintained” a spot
exchange rate of between 7.5750 and 7.5765.
(f) On the same date, between 07h41 and 07h44, banks that included the
same group of four South African banks “affirmed a focal point” of
7.5750.
(g) On 7 March 2012, at around 10h13, “banks engaged in ‘flat -lining’, and
[FRB] was allowed to dominate the market with its quotations while other
traders withheld bids and offers”. (The affidavit does not identify the
other “banks”.)

156 The allegations in the superseding affidavit referring to “RMB” (RMB Holdings Limited), the entity later
identified as FirstRand (the twenty-sixth respondent), were made applicable also to FRB (the twenty-seventh
respondent).

ROGERS J
87
(h) On the same day, between 10h12 and 10h16 (that is, over a period that
included the immediately preceding incident), “local banks withheld
quotes and enabled [FRB] to dominate the market”. FRB thereafter
dropped the price from 7.6425 to 7.6390, thereby creating a new lower
focal point above which the other banks quote[d]”.
(i) On 19 September 2012, between 00h26 and 00h34, banks that included
Investec, FRB, Absa and HBEU/HBUS “alternated in posting quotations
in a manner that enabled them to set the exchange rate at various levels
for nearly eight minutes”.
(j) On 20 September 2012, between 06h45 and 06h47, FRB “paired its bid
and ask prices with various banks in succession”. This occurred at least
24 times, and included pairings with HBEU/HBUS, Commerzbank,
Barclays, Investec and Nedbank.
(k) On 28 September 2012, between 18h32 and 18h39, FRB, Absa, Investec
and Nedbank “through a pattern [of] consecutive quotes coordinated their
bid and ask prices generat[ing] a sequence of the same spot exchange
rates”.

[197] In upholding FRB’s appeal, the CAC reasoned thus:157
(a) In contrast to the case brought against many of the other banks, the
Commission did not provide any examples of FRB traders engaging in
contact in chatrooms. Indeed, the Commission accepted that it did not
even know who the FRB traders were and thus had no evidence that they
participated in chatrooms.
(b) The case against FRB was based on “trading data on the Reuters
platform”, the allegation being that the posting of quotations on the
platform was coordinated to manipulate the currency.
(c) The CAC summarised the incident of 11 January 2009 and identified the
dates of the other incidents.

157 CAC II above n 1 at paras 152-61.

ROGERS J
88
(d) It was “regrettable” that the Tribunal’s decision against FRB “was
justified on hopelessly incorrect information”. The Tribunal had stated
that FRB traded through the chatrooms, despite the fact that no such
allegation was made in the superseding affidavit. The Tribunal had
supported its statement with reference to named traders, none of whom
were ever employed by FRB.
(e) The Commission’s counsel had made much of “consistent prices being
observed across time and banks”. A “fundamental problem” with that
submission, one that “percolate[d]” throughout the affidavit, was the
“inability to distinguish between information which is in the public
domain; that is, information available to all banks who simply have access
to Reuters or a similar screen … and information that could only be the
product of some form of nefarious cartel activity”. The superseding
affidavit should at least have made this distinction clear, thus locating its
case within the latter as opposed to the former category.
(f) There was thus no case made out implicating FRB.

[198] The very way in which the Commission frames the seventh error shows that it is
not an error of law, or at any rate not one of general public importance. It is a supposed
error in a particular case where collusion in chatrooms was an important feature of t he
Commission’s pleaded case. It hardly needs to be said that it is not an inherent feature
of cartels that they involve currency traders and chatrooms. The supposed error is case-
and fact-specific.

[199] In any event, the CAC did not make the alleged error. It correctly identified the
absence of an allegation against FRB that its traders communicated with other traders
through implicated chatrooms. This was a relevant factual circumstance to be taken
into account in deciding whether the superseding affidavit adequately alleged FRB’s
participation in the SOC. The Commission itself pleaded that the SOC could be inferred

participation in the SOC. The Commission itself pleaded that the SOC could be inferred
from, among other things, frequent communication through permanent chatrooms, this
constituting the “primary mode” of communication between the respondents’ traders

ROGERS J
89
for purposes of implementing the SOC. Moreover, the Tribunal had, albeit mistakenly,
found that FRB had participated in chatrooms.

[200] The CAC went on to state that the Commission’s case against FRB was based
on “trading data on the Reuters platform”. The CAC did not state that such information
on its own could not in law justify an inference of participation in the SOC. What the
CAC found was that the allegations in regard to the trading data did not in this case
justify the inference.

Eighth error – insistence that traders be named (HBUS)
[201] The Commission contends that the CAC erred in law by holding that the
superseding affidavit needed to identify a bank’s employee or representative by name
in order to justify a finding that the bank’s conduct fell within the SOC. The
Commission submits that this error was made in relation to HBUS158 and Nedbank.159

[202] The superseding affidavit contains the following HSBC-specific allegations:
(a) Its traders were Messrs Christopher Hatton (until 30 October 2010) and
Mijo Mirkovic.
(b) HSBC joined the SOC “by at least 1 October 2007”.
(c) Mr Hatton features in 27 alleged chatroom incidents over the period
September 2007 to June 2010. All these interactions were with other
foreign banks.
(d) The affidavit mentions Mr Mirkovic only twice. On 29 October 2007, so
the Commission pleads, Mr Hatton and traders from Citibank, BoA and
SCB were participants in the Old Gits chatroom. Mr Mullaney (SCB) and
Mr Cook “informed each other of [the] intended trading directions”. The
traders agreed to meet for drinks. Mr Cummins (Citibank) asked the other
traders “to preserve the agreed spreads as he has to support his family”.

158 Id at paras 131-2.
159 Id at para 164.

ROGERS J
90
The Commission continues: “Hatton makes reference to his conversations
with [Mirkovic]. The traders all share information on their trading
positions”.
(e) On 25 and 26 July 2012, traders from BNP, JPM Co, Citibank and
Barclays are alleged to have been participants in an implicated chatroom.
Mr Katz shared information about customer inquiries. Mr Aiyer said he
“needs to sell 25 million and requires everyone to post offers so as to drive
the price down”. Mr Katz indicated that he is “skewing lower on his ecom
[platform] and if he gets paid Aiyer can have”. Mr Aiyer said “that ‘Mijo’
[Mr Mirkovic] has taken 15”.
(f) HSBC is also alleged to have been part of three incidents of market
conduct observed on the “Reuters trading platform”. These occurred on
27 May 2010, 19 September 2012 and 20 September 2012.
(g) The Reuters incident on 27 May 2010 is also alleged to have involved
Absa, SBSA, Barclays, Nomura and other banks. Various banks held the
exchange rate “around the focal point” of 7.5760. SBSA and Absa
together with HSBC and other banks then “pushed the exchange rate” to
7.5600, before returning the exchange rate to a “focal point” of 7.5600.
(h) The second Reuters incident, on 19 September 2012, occurred between
00h26 and 00h34. Banks, alleged to include Investec, FRB, Absa and
HSBC, “alternated in posting quotations in a manner that enabled them to
set the exchange rate at various levels for nearly eight minutes”.
(i) The third Reuters incident took place on 20 September 2012 between
06h45 and 06h47. FRB allegedly “paired its bid and ask prices with
various banks in succession”. This occurred 24 times, and included two
pairings with HSBC.

[203] In relation to HSBC, the CAC reasoned thus:160

160 Id at paras 129-34.

ROGERS J
91
(a) The Commission’s case mainly centred on HBUS’ trader, Mr Hatton. He
was implicated for HBUS in 28 161 of the incidents catalogued in the
superseding affidavit . These interactions were all with other foreign
banks.
(b) There were only three instances where HBUS was alleged to have
engaged with a South African bank, but in none of those did the affidavit
name HBUS’ representative or the representatives of the local banks.
(c) The affidavit also identified Mr Mirkovic as an HBUS representative, but
he was not alleged to have been an active or passive participant in any of
the conduct implementing the SOC. In the only incident in which he was
mentioned, his involvement was “peripheral” and h ardly took the
Commission’s case further.162
(d) The CAC emphasised that, to show personal jurisdiction over the pure
peregrini such as HBUS, the Commission had to show “sufficient
South African involvement”.
(e) The position was different, said the CAC, in relation to HBEU. Because
it was a local peregrinus, only subject matter jurisdiction had to be
established. The allegations in respect of Mr Hatton were “clear prima
facie evidence” of participation in the SOC. Although HBUS admitted
that Mr Hatton had been its employee (i.e. not HBEU’s employee), the
allegations of participation in trading justified drawing an inference that
HBEU was “correctly connected to the SOC”. There was “no clear
denial” that Mr Hatton had nothing to do with HBEU’s activities.

[204] As noted, the CAC mentioned three instances where HBUS was alleged to have
engaged with South African banks, these being the instances where no traders were
identified. These were the three Reuters incidents. It is relevant to note, in this regard,

161 By my count it was 27 incidents, but nothing turns on this. The CAC may have included an incident that took
place after 30 October 2010, by which time Mr Hatton had moved from HSBC to CSS.

place after 30 October 2010, by which time Mr Hatton had moved from HSBC to CSS.
162 There were two references in the superseding affidavit to Mr Mirkovic, but the CAC’s observation is equally
applicable to both.

ROGERS J
92
that in opposing its joinder, HBUS stated that Mr Hatton was employed by it between
September 2005 and October 2010, a fact not denied by the Commission in its replying
affidavit. It follows that Mr Hatton could not have been the unnamed trader involved
in the two incidents in September 2012, which took place more than 27 months after
the last chatroom incident involving HSBC.

[205] Fairly read, the CAC’s judgment cannot be understood as laying down, as a legal
proposition, that a firm’s representative has to be named in order for that firm to be
susceptible to prosecution for cartel conduct. It would be absurd to suppose that the
CAC held such a view. The CAC, in my view, was merely making the point that in this
particular case the alleged facts were insufficient to justify a conclusion that HBUS was,
in these three instances, colluding with the South African banks as part of the SOC.

[206] As I have previously mentioned, CAC II contains no separate analysis of
Nedbank’s position, seemingly an oversight. In the context of the alleged error which
I am now addressing, the Commission refers to a paragraph forming part of the CAC’s
discussion of SBSA’s appeal. Although the CAC said in that paragraph that the
superseding affidavit did not name any trader who represented Nedbank, this
observation, made in passing, appears to have played no part in the CAC’s decision in
favour of SBSA.163

Ninth error – insistence on “active” chatroom participation (BAMLI and ANZ)
[207] The Commission contends that the CAC erred in law in finding that the
superseding affidavit needed to allege that a trader “actively participated” in chatroom

163 CAC II above n 1 at para 164. That paragraph reads:
“Turning to the allegation that [SBSA] entered the SOC on 1 January 2008 (a public holiday in
South Africa, but presumably on the Commission’s case South African traders were still hard
at work), the Commission was not able to refer to a single contact between Absa, [SBSA] and/or

at work), the Commission was not able to refer to a single contact between Absa, [SBSA] and/or
Nedbank, whether before or after that date. No trader was named who was employed by or who
represented Nedbank.”

ROGERS J
93
conduct in order for such conduct to be within the scope of the SOC. This error, submits
the Commission, was made in respect of BAMLI164 and ANZ.165

[208] The superseding affidavit contains the following BoA -specific allegations (the
BoA entities are BAMLI, MLP and BANA):
(a) BoA’s traders were Messrs Gavin Cook and Mark Sheppard, who were
employed by or representing “[MLP], alternatively [BAMLI],
alternatively [BANA]”.
(b) Mr Cook is alleged to have been involved in 40 of the 162 incidents listed
in the affidavit, spanning the period September 2007 to March 2011.
These were all chatroom interactions, many of them in the Old Gits
chatroom. The only other alleged participants in these incidents were
other foreign banks: Citibank, HSBC, SCB and SAI are the recurring
names; CSS features once.
(c) Curiously, the affidavit does not allege that Mr Sheppard participated in
any of the incidents.

[209] In respect of the BoA entities, the CAC reasoned as follows:
(a) In relation to BAMLI, 166 the Commission had been told under oath in
BAMLI’s exception application that Mr Cook was at all times employed
by MLP, not BAMLI, a fact never denied by the Commission. This might
have explained why the Commission sought to join MLP. Mr Sheppard
was not alleged to have participated in a single instance of discussion in
an implicated chatroom, which was the Commission’s key evidence for
the prima facie case justifying the inference that there was an SOC.
(b) The superseding affidavit thus did not contain facts necessary to justify a
conclusion that BAMLI, a pure peregrinus, was connected in a conspiracy
with South African banks.

164 Id at para 89.
165 Id at paras 94 and 96-7.
166 Id at paras 87-93.

ROGERS J
94
(c) Moreover, CAC I had required the Commission to specify the bank on
behalf of which any particular trader had acted and whether the trader was
alleged to have acted for more than one respondent at a time. The
allegation that Messrs Cook and Sheppard were acting for “[MLP] ,
alternatively [BAMLI], alternatively [BANA]” was “devoid of the kind
of specificity which was required by [ CAC I] or by a referral that could
pass legal muster”.
(d) In relation to BANA, 167 a local peregrinus, the affidavit purported to
justify linking Mr Cook to this entity by virtue of email addresses
containing his name together with the domains @[…] and @[…]. This
did not provide a basis to join BANA.
(e) In relation to MLP, 168 a pure peregrinus, Mr Cook’s impugned
interactions involved only other foreign banks. Mr Sheppard was not
implicated at all. On balance, therefore, there were insufficient
connecting factors to South Africa to justify personal jurisdiction over
MLP.

[210] The superseding affidavit makes the following ANZ-specific allegations:
(a) ANZ’s traders were Mr Murat Tezel and, as from 2013, Mr Katz.
(b) Mr Tezel features once in the affidavit. The Commission alleges that on
18 October 2012 Mr Katz (a BNP trader at the time) and Messrs Aiyer,
Williams and Cummins (JPM, Barclays and Citibank respectively) were
participants in an implicated chatroom. Messrs Katz and Aiyer “matched
each other’s opposite FIX positions in order to offset their respective
exposure at the upcoming FIX”, and “Aiyer ‘copied and pasted’ a portion
of a communication between Madaras and Tezel from another chatroom”.
(Mr Dave Madaras is identified elsewhere in the affidavit as a Citibank
trader.)

167 Id at paras 91-2.
168 Id at para 93.

ROGERS J
95
(c) Although the affidavit alleges that Mr Katz was an ANZ trader as from
2013, the last pleaded incident involving Mr Katz is the one just
mentioned, when he was still with BNP. In other words, the affidavit does
not allege any implicated conduct by Mr Katz in his capacity as an ANZ
trader.

[211] The CAC reasoned thus in respect of ANZ:169
(a) Although Mr Katz was a central player in the case brought by the
Commission, every instance of his participation preceded the year in
which he joined ANZ.
(b) Regarding the single reference to Mr Tezel, the CAC said that it was
unclear precisely what was meant by “copied and pasted”. There was no
suggestion that Mr Tezel had participated in a chatroom or any similar
discussion with other traders.
(c) The CAC concluded that no sufficient case had been made out to find that
ANZ was a participant in the SOC.
(d) The CAC added170 that Mr Katz had been criminally charged in the U S
and had entered into a plea agreement there relating to a conspiracy
among foreign traders targeting emerging markets ’ currency prices,
including the ZAR. (These facts were pleaded in the superseding
affidavit.) Since the Commission would, said the CAC, have had access
to the case brought against Mr Katz, the Court would have expected the
affidavit to include allegations that Mr Katz had exploited his knowledge
when he moved to ANZ in 2013 if this had been the case.

[212] As with the seventh error, the formulation of the ninth error is case - and
fact-specific. It is not by its nature an error of law, or at any rate not one of general
public importance. In relation to the BoA entities, the CAC did not hold that, as a matter

169 Id at paras 94-7.
170 Id at paras 98-9.

ROGERS J
96
of law, Mr Sheppard’s conduct could not have implicated them in the SOC unless he
participated in chatrooms. The CAC was nevertheless justified in pointing out that
collusion in chatrooms was the hallmark of the SOC as pleaded by the Commission.
The important point , insofar as Mr Sheppard is concerned, is that the superseding
affidavit did not identify any conduct by him, in chatrooms or elsewhere.

[213] In regard to ANZ, the CAC was emphasising the obscure nature (in the CAC’s
view) of the Commission’s single reference to Mr Tezel. In the context of an SOC
where chatrooms were said to be the primary means of communication, it was not out
of place for the CAC to observe that the single reference to Mr Tezel did not suggest
that he had participated in a chatroom or had “any similar discussion with other traders”.

[214] Insofar as Mr Katz is concerned, the superseding affidavit did not identify any
conduct by him, whether in chatrooms or otherwise, from the time he became employed
by ANZ. The CAC considered that, in the circumstances, the affidavit did not contain
facts that could support an inference of ANZ’s participation in the alleged SOC.

[215] ANZ’s position in relation to Mr Katz is analogous to that of UniCredit in
UBS.171 A certain trader, actively implicated in the SOC during his former employment
with other implicated banks, switched employment from Portigon to UniCredit with
effect from 1 September 2011. In regard to liability for penalties, it was relevant to fix
the date from which UniCredit became implicated in the SOC. The EC had set the
starting date of participation as 9 September 2011, because the trader had logged into
the Cods & Chips chatroom on that date, albeit under his former employer’s account.
However, there was no evidence of anti -competitive conduct in the chatroom on that
date. The first implicated conduct by the trader took place on 26 September 2011, and

date. The first implicated conduct by the trader took place on 26 September 2011, and
the CJEU held that this was the relevant s tarting date. UniCredit could not have been
required to distance itself from conduct that had not yet taken place. The CJEU said:


171 Above n 144.

ROGERS J
97
“1250. If the [EC’s] line of reasoning were to be followed, it would result in allowing
it to find that UniCredit participated in a single and continuous infringement even if,
thereafter, that undertaking’s trader had not participated in any anticompetitive
discussions or the infringement had ceased before any fresh anticompetitive
discussions took place.
1251. It follows that, in the absence of anticompetitive exchanges that took place on
the day on which the access of the trader concerned to the Cods & Chips chatroom was
renewed on behalf of his new employer, namely UniCredit, that renewed access is not,
per se, sufficient to set the starting point of that bank’s participation in the infringement
at issue at 9 September 2011, even though, owing to his previous employment
relationships with RBS, Natixis and Portigon, that trader had knowledge of the fact that
the exchanges that took place in that chatroom could be anticompetitive in nature.”

Tenth error – test for determining a firm’s knowledge of the SOC
[216] The Commission submits that, in determining whether a bank had actual or
constructive knowledge of the scope and general characteristics of the SOC and was
prepared to take the risk, the CAC erred in applying, as the only relevant factors,
regularity of pleaded conduct, membership of and active participation in chatrooms and
whether the superseding affidavit named the bank’s trader. The CAC ought to have
considered all the factual allegations about the bank’s conduct, together with
surrounding factual allegations, to determine whether the affidavit contained allegations
“from which to draw a reasonable possible inference of knowledge or assumed
knowledge”.

[217] The Commission identifies the other relevant circumstances as including: a
common instrument (the USD/ZAR currency pair ); common strategies or forms of
conduct; common platforms; the same geographies (the US and South Africa); the same

conduct; common platforms; the same geographies (the US and South Africa); the same
period, namely 2007 to 2013; and common individuals. The Commission contends that
the CAC should have found that the presence of an individual in the chatroom together
with competitors could be sufficient to draw a reasonable possible inference of a
contribution to the overall objective of the SOC.

ROGERS J
98
[218] The tenth error is, on analysis and at least in the main, merely an aggregation of
the sixth to ninth errors, with which I have already dealt. Furthermore, it does not in
truth assert an error of law. It is formulated with reference to the unique factual
circumstances of this particular case. It is a complaint about how the CAC assessed the
pleaded facts.

[219] Moreover, the CAC did not hold that, as a matter of law, only certain
circumstances were relevant in considering whether one could reasonably draw the
inference that a bank participated in or had knowledge of the objectives of the SOC.
The CAC identified the facts that were alleged in relation to the bank in question, which
is unobjectionable. The more general circumstances listed by the Commission formed
part of the background that the CAC dealt with earlier in its judgment. There is no
reason to think that the CAC ignored them in the context of analysing the bank-specific
allegations.

[220] I must confess that I struggle to see how the more general circumstances
mentioned by the Commission could add much weight to the bank-specific allegations.
In regard to the common instrument (the USD/ZAR currency pair) and common
geographies (the US and South Africa), it is perfectly plausible that, if the currency pair
were susceptible to manipulation, different groups of traders in the US and South Africa
could independently collude to manipulate the price. As to common platforms, the
superseding affid avit does not disclose a pattern of all or most of the banks being
implicated in the same chatrooms. 172 Regarding a common period of time (2007 to

172 Of the 162 incidents, 29 took place in the Old Gits chatroom. This was from September 2007 to April 2011.
The hard-core members were Citibank, BoA, HSBC, SCB and SAI, later also Barclays and CSS (this followed
the moves by Mr Katz to Barclays in July 2010 and Mr Hatton to CSS in November 2010). Nine incidents took

place in the ZAR chatroom. This was from March 2012 to October 2012. The only participants were BNP,
Citibank, JPM and Barclays. Of the 16 2 incidents, 104 took place in otherwise unidentified “implicated
chatrooms”. Since these appear to have been random chatrooms, it is not possible to infer any pattern of
participation in any one particular chatroom. What can be said is that Absa and Investec are alleged to have
colluded on four occasions between May 2008 and November 2011 in implicated chatrooms, with no other banks
alleged to have been present. SBSA and Barclays are alleged to have colluded twice in implicated chatrooms, in
January 2012 and October 2012, again with no other banks being present. There are no allegations of participation
in chatrooms by Nedbank and FRB. One incident took place via telephone, 18 were matters of observed market
conduct on the “Reuters trading platform” and one was of observed market conduct on an unidentified platform.
That currency traders should be active on the Reuters trading platform is, of course, unremarkable. These

ROGERS J
99
2013), on the Commission’s own version , not all the banks are alleged to have been
involved throughout that period. Based on the 162 specific instances catalogued in the
affidavit, the periods of the banks’ alleged participation vary widely.173

Eleventh error – disregarding conduct evidence (SBSA, Nedbank, Nomura,
Commerzbank, HBUS and FRB)
[221] The Commission contends that the CAC erred in law by not considering the
admissibility and probative value of “conduct evidence in its assessment of attribution”.
This evidence ought to have been regarded as admissible, relevant and potentially of
high probative value. The Commission submits that this error was committed in relation
to SBSA,174 Nedbank, Nomura,175 Commerzbank,176 HBUS177 and FRB.178

[222] I have already summarised the bank -specific allegations made in respect of
Nomura, FRB and HBUS, and the CAC’s reasoning in respect of those banks.179

[223] The superseding affidavit contains the following Commerzbank -specific
allegations:

incidents mainly involved South African banks, with foreign banks from time to time alleged also to have been
involved in the observed activity.
173 Citibank – September 2007 to October 2012; BoA entities – September 2007 to March 2011; HSBC –
September 2007 to September 2012; SCB – October 2007 to March 2012; SAI – October 2007 to December 2010;
Absa – January 2008 to September 2013; SBSA – January 2008 to October 2012; Nedbank – January 2008 to
September 2012; Investec – May 2008 to September 2012; FRB – January 2009 to September 2012; Nomura –
May 2010 to October 2012; Barclays – May 2010 to April 2013; Commerzbank – July 2010 to September 2012;
CSS – November 2010 to May 2011; JPM – April 2011 to April 2013; BNP – November 2011 to October 2012;
Macquarie – 11 September 2013 (a single date); ANZ – 18 October 2012 (a single date). I recognise, of course,
that if a bank is properly alleged to have joined the SOC, its termination date would not necessarily be the last

known date of its participation. A bank might in those circumstances need to show that it disassociated itself from
the SOC. However, for present purposes we are concerned with the inferences that could reasonably be drawn
from the pleaded facts, and those pleaded facts do not positively show that all the banks were involved throughout
the whole period of 2007 to 2013.
174 CAC II above n 1 at paras 165-71.
175 Id at paras 100-6.
176 Id at paras 109-18.
177 Id at para 131.
178 Id at paras 152-62.
179 See [186] to [187] (Nomura); [196] to [197] (FRB); and [202] to [203] (HBUS).

ROGERS J
100
(a) Commerzbank’s traders were Messrs Nigel Dousie and Kevin Wilson.
(However, only Mr Dousie features in subsequent allegations. The
affidavit does not implicate Mr Wilson in any specific incidents.)
(b) The specific allegations against Commerzbank involved one chatroom
interaction and two instances of observed conduct on the “Reuters trading
platform”.
(c) The chatroom interaction occurred on 28 July 2010, when Mr Katz (who
had just moved from SAI to Barclays) and Mr Dousie were participants
in an implicated chatroom in which Mr Dousie “shared information with
Katz about a potential customer in the market”.
(d) As to observed conduct on the Reuters platform, on 29 March 2012,
between 21h33 and 23h18, “a focal point was initiated by Commerzbank
at 7.7170”. (No other banks are mentioned.)
(e) On 20 September 2012, between 02h46 and 04h17, “a number of foreign
banks (not referred to in this present referral) posted the same bid and ask
prices”. Investec and Tatra Banka both posted the same bid and ask prices
of 8.3027 and 8.3527, respectively, each resulting in an average of 8.3277.
Commerzbank posted “different bid and ask prices (8.3227 and 8.3327
respectively) [but] these prices nonetheless result[ed] in Commerzbank
matching the average of Investec at 8.3277”. (Tatra Banka is not cited as
a respondent. It is apparently a Slovakian bank.)
(f) About two hours later, between 06h45 and 06h47, there occurred the
incident I have already summarised in relation to FRB, which allegedly
“paired its bid and ask prices” with various banks. It allegedly did so
twice in relation to Commerzbank.

[224] The CAC reasoned thus in relation to Commerzbank:180

180 CAC II above n 1 at paras 109-18.

ROGERS J
101
(a) It referred to and summarised the incidents that took place on
28 July 2010, 29 March 2012 (the CAC mistakenly said 29 July 2012)
and 20 September 2012.
(b) As to the incident of 28 July 2010, there were no details of the information
shared, a significant omission in the CAC’s view, given that the
Commission would have had access to the information in the case against
Mr Katz in the US. The allegation in respect of this incident did not
suggest that Commerzbank was acting in concert with any other banks,
nor was there an allegation that the information shared was commercially
sensitive.
(c) As to the Reuters incidents, the CAC regarded as relevant the evidence
given by the deponent to SBSA’s dismissal application about the
distinction between the Reuters trading platform and the Reuters
information platform. The SBSA deponent inferred, from the level of
detail given in the superseding affidavit, that the supposed conduct on the
“Reuters trading platform” in fact refer red to automated indicative
exchange rates posted on the Reuters information platform.181
(d) Regarding the Reuters incident of 29 March 2012, the fact that no other
banks were mentioned stood in contrast to other allegations in the
affidavit of focal points involving multiple banks: “ [a]n allegation of
unilateral conduct without more is hardly evidence of participation in an
SOC”.
(e) There was thus insufficient evidence to show that Commerzbank was a
participant in the alleged SOC.

[225] The superseding affidavit contains the following SBSA-specific allegations:
(a) Two interactions in an implicated chatroom are alleged. First, on
10 January 2012 Messrs Peter Taylor (allegedly a Barclays Capital trader)
and Bryan Brownrigg (allegedly an SBSA trader) were participants in an

181 As to which, see above n 125.

ROGERS J
102
implicated chatroom “in which they discussed [the] bid-offer spread for
USD/ZAR”.
(b) Second, on 1 October 2012 the same two individuals were participants in
an implicated chatroom, on which occasion “Taylor enquired what quote
Brownrigg would show a top customer in 25 and 50” . Mr Brownrigg
responded “by telling ‘50 100’”.
(c) Additionally, conduct involving SBSA together with other banks was
observed on eight occasions on the Reuters platform over the period
1 January 2008 to 19 September 2012. On five of these occasions the
other banks were South African banks. On the remaining three occasions
the banks included foreign banks: SAI (on 12 January 2009) ; Barclays,
Nomura and HSBC (on 27 May 20 10); and Barclays
(19 September 2012). The alleged conduct included withholding quotes
in favour of one of the banks, setting focal points and the posting of
identical bid and ask prices.
(d) The first of these Reuters incidents took place on 1 January 2008 between
20h14 and 05h13. Absa and SBSA were said to have withheld quotes to
enable Nedbank to fix the exchange rate at 7.5850.
(e) The last of these Reuters incidents took place on 19 September 2012
between 18h57 and 02h37 . SBSA, allegedly represented by
Mr Brownrigg, provided “an unusually high spread in the market, in line
with the conversation he held on the Bloomberg chatroom with Taylor”.
The unusually high spread of 0.1001 “was in force for quite some time,
so that the chat between Brownrigg and Taylor confirmed what was
already long agreed between the participants in the conspiracy”. The data
showed that banks normally did not charge a spread exceeding 0.06. The
spread of 0.1000 “is associated with [SBSA], Absa and Nedbank”.

[226] In its dismissal application, SBSA stated that : Mr Brownrigg was employed by
it as a salesperson, not a trader; that he played no role in determining SBSA’s bid-offer
spreads and could not bind SBSA to foreign currency transactions; that Mr Taylor had

ROGERS J
103
not worked for Barclays Capital, but for Citigroup until February 2012 and Barclays
Investment Bank thereafter, neither of which were cited as respondents; that, unlike
SBSA, Barclays Bank was not authorised by the South African Reserve Bank to trade
in the ZAR; and that the communications between Mr Brownrigg and Mr Taylor could
thus only have been of a vertical nature, between a forex dealer (SBSA) and potential
customer (the bank for which Mr Taylor was acting). I shall return to the admissibility
of this evidence at a later stage, when I give separate consideration to the Commission’s
case against SBSA.

[227] The CAC reasoned thus in upholding SBSA’s appeal:182
(a) Regarding the allegation that SBSA joined the SOC on 1 January 2008 –
a public holiday in South Africa – the Commission had not referred to a
single contact between SBSA, Absa and Nedbank, the three banks
involved on the New Year’s Day incident on the Reuters platform,
whether before or after that date.
(b) Although on that occasion the banks were said to have withheld quotes to
enable Nedbank to fix the exchange rate at 7.5850, no further details had
been provided. It was not unreasonable to speculate that the “withholding
of quotes” was because the South African forex markets were closed due
to the public holiday, rather than because of any “nefarious activity”.
(c) The CAC also referred to the Reuters incident on 19 September 2012.
With reference to the distinction drawn by SBSA’s deponent between the
Reuters trading platform and the Reuters information platform, the CAC
said that information gleaned from the latter was “hardly indicative of
participation in the SOC”.
(d) The CAC referred to SBSA’s evidence about the roles of Mr Brownrigg
and Mr Taylor. The Commission had not explained whether the
communication between the two of them was a “normal conversation that
took place in the ordinary course of forex dealing” or whether “there was

182 CAC II above n 1 at paras 163-73.

ROGERS J
104
something nefarious in this particular conduct”. It was inexplicable that
the Commission persisted in claiming plausible evidence to link SBSA to
the SOC.
(e) The “skeletal nature” of the allegations in respect of SBSA revealed that
there was “no basis by which its activity fell within the scope of subject
matter jurisdiction”. The case did not get out of the “legal starting
blocks”. (I have already addressed the CAC’s unfortunate ly inaccurate
use of the term “subject matter jurisdiction” in this context.)

[228] As is apparent from my summary of the CAC’s reasoning in respect of
Commerzbank and SBSA, the CAC did not regard the observed conduct on the
Reuters platform as inadmissible. It made no such legal ruling. Nor did the CAC
overlook such evidence. The CAC referred to that evidence (both of the Commerzbank
incidents and two of the eight SBSA incidents ) as examples, and explained why such
evidence could not reasonably support a conclusion that Commerzbank and SBSA were
implicated in the alleged SOC. I should add that even if the CAC had overlooked the
evidence, that would merely have been an alleged error in the assessment of the
evidence, not an arguable point of law of general public importance.

[229] It is unclear why the Commission claims that the legal error now under
discussion was made in relation to Nomura. There were only four Nomura -specific
allegations in the superseding affidavit , three chatroom incidents (of which two
occurred on the same date) and one Reuters incident. The CAC referred to all these in
its reasoning.

[230] Similarly, in relation to HBUS, the CAC referred to the three Reuters incidents
as being the only ones where South African banks were allegedly involved. The CAC
was critical of the fact that the Commission had not identified a single employee of any
of the South African banks with whom HBUS could have reached an agreement to
participate in the SOC. The CAC did not say that these three incidents did not constitute

participate in the SOC. The CAC did not say that these three incidents did not constitute
legally admissible evidence. The CAC found that the evidence was lacking in detail

ROGERS J
105
and was insufficient to link HBUS to an SOC involving South African banks. If that
assessment of the allegations was erroneous, it was not in consequence of a legal error,
let alone one raising an arguable point of law of general public importance.

[231] The same is true of FRB. The only FRB-specific allegations in the superseding
affidavit were of conduct observed on the Reuters platform. The CAC referred to this
evidence. It regarded the “fundamental problem” with the evidence as being the
affidavit’s “inability to distinguish between information which is in the public domain;
that is, information available to all banks who simply have access to Reuters or a similar
screen available to all banks and information that could only be the product of some
form of nefarious cartel activity”. 183 At least, the affidavit “should have made this
distinction patently clear and thus located the case . . . within the latter as opposed to
the former category”.184

Twelfth error – awareness tied to specific acts by employees (ANZ)
[232] The Commission submits that the CAC erred in finding that the actual or
constructive knowledge of a bank could not be established or inferred unless the
superseding affidavit contained allegations of conduct by its employees during their
period of employment. The Commission says that this error was made in relation to
ANZ.185

[233] It will be recalled that Mr Katz, who had previously been employed by SAI,
Barclays and BNP, moved to ANZ in 2013. The CAC pointed out that the superseding
affidavit contained no allegations of anti -competitive conduct by Mr Katz after 2012.
The Commission submits that the correct legal position is that the knowledge of a trader,
acquired before their arrival in the service of a new employer, can be attributed to the
new employer where it was made available to such employer. Based on a benevolent

183 Id at para 159.
184 Id at para 160.
185 Id at para 95.

ROGERS J
106
and generous reading of the affidavit, as is required at the exception stage, so argues the
Commission, it is a reasonable inference that a trader with a history of collusive
participation during prior employment continued the collusive activity with the new
employer.

[234] Regarding the attribution of a trader’s knowledge to their new employer, the
Commission refers to UBS186 at paragraph 1295. There the CJEU was dealing with the
liability of UniCredit for the conduct of a trader who had been involved in the
conspiracy during his previous employment with other banks. I have already dealt, in
an earlier part of this judgment, with the position of ANZ in relation to Mr Katz,187 inter
alia quoting what the CJEU said in paragraphs 1250-1.

[235] In the passage from UBS cited by the Commission, the CJEU was not saying that
the knowledge of a trader could be imputed to the new employer in the absence of
evidence of actual ant i-competitive conduct by the trader after entering the service of
the new employer. The CJEU had already held that UniCredit could not be held liable
for any anti-competitive conduct before 26 September 2011, the first instance of such
conduct by the trader after joining UniCredit. It was from that date that the trader, as a
UniCredit employee, intended to contribute to the SOC.188

[236] In the part of the judgment referenced by the Commission, the CJE U was
explaining that, in regard to UniCredit’s participation as from 26 September 2011, the
EC could properly have found that UniCredit had actual or constructive knowledge of
all the infringing conduct put into effect by the other banks. According to settled EU
case law, the EC had been entitled to rely on contacts before and after the period of the
infringement “in order to construct an overall picture and to show the preparatory stages

186 Above n 144.
187 See [214] to [215].
188 UBS above n 144 at paras 1275 and 1288-9.

ROGERS J
107
of the cartel as well as to corroborate the interpretation of certain evidence”. 189 Then
follows the paragraph to which the Commission has referred us:

“The [EC] was therefore entitled to use the knowledge acquired by an employee prior
to his or her arrival in the service of a new company and which he or she in fact makes
available to that new employer when that knowledge corroborates other elements
available to the [EC].”190

[237] In the absence of any allegation that Mr Katz acted anti-competitively in pursuit
of the SOC after joining ANZ, the above reasoning of the CJE U cannot apply. The
absence of the required allegation cannot be filled through a generous and benevolent
interpretation of the superseding affidavit.

Thirteenth error – unwarranted speculation by the CAC
[238] The final error identified by the Commission in its note is that the CAC engaged
in speculation about the knowledge of the Commission flowing from the fact that certain
traders had been employed by or represented one or more banks that obtained leniency.
The CAC, so contends the Commission, erred in law when relying on these assumptions
in its assessment of the sufficiency of the superseding affidavit. The CAC ought to have
applied exception principles. This error is said to have been made in relation to ANZ,191
Nomura,192 Commerzbank193 and Macquarie.194

[239] I have already dealt with the CAC’s reasoning in relation to these four banks. In
the cases of ANZ and Commerzbank, the CAC did not refer to information which the
Commission would have got from leniency applicants but to information relating to
Mr Katz’s prosecution in the US. That Mr Katz had been so prosecuted was pleaded in

189 Id at paras 310 and 1249.
190 Id at para 1295.
191 CAC II above n 1 at para 99.
192 Id at para 108.
193 Id at para 112.
194 Id at para 122.

ROGERS J
108
the superseding affidavit . As to information that could be sourced from leniency
applicants, the fact that Absa and Barclays had sought leniency was known factual
background – it was mentioned in TRIB I195 and TRIB II.196

[240] The CAC’s so-called reliance on assumptions associated with prosecutions and
leniency was no more than a minor makeweight point. The CAC’s primary conclusion
in each case was that the superseding affidavit did not plead sufficient facts from which
reasonable inferences could be drawn against the banks in question. It was merely a
matter of surprise that, if such facts in truth existed, they would not have been pleaded,
having regard to the sources of infor mation available to the Commission, in particular
information from banks that had settled or were granted leniency. I should add that the
supposed error is really a criticism of the way the CAC assessed the pleaded facts. If it
is an error of law at all, it is not one of general public importance.

The Commission’s application for leave to appeal in CCT 30/24
[241] Having addressed various issues of principle, I now deal individually with each
bank, starting with the Commission’s application for leave to appeal in CCT 30/24.

The BoA entities: BAMLI, MLP and BANA (first, twentieth and twenty -first
respondents)
[242] I have already dealt with the superseding affidavit’s BoA-specific allegations
and the CAC’s reasoning in the context of the ninth alleged error. There were no legal
errors affecting the CAC’s assessment of the Commission’s case against MLP and
BANA. Evidentially, those entities were entitled to plead facts in opposition to their
proposed joinder. The Commission in reply raised no plausible dispute about
Mr Cook’s employer having been MLP, not BANA. The Commission’s attack on the
CAC’s judgment is ultimately that the CAC’s factual assessment of the case for joinder
was wrong. This does not engage our jurisdiction.

195 TRIB I above n 3 at para 12.

was wrong. This does not engage our jurisdiction.

195 TRIB I above n 3 at para 12.
196 TRIB II above n 4 at para 12 and fn 16 and 23.

ROGERS J
109

[243] The CAC’s reasoning in respect of BAMLI attached significance to the evidence
that Mr Cook was employed by and represented MLP, not BAMLI or BANA. BAMLI,
unlike its associated companies, was in principle confined to an exception. However, I
have already explained that the Tribunal’s approach to exceptions is more flexible than
in ordinary civil litigation.

[244] The Tribunal does not appear to have held, in relation to BAMLI, that no regard
could be had to the evidence that MLP was Mr Cook’s employer. The Tribunal
summarised the Commission’s argument as being that its case against BAMLI related
to its membership of the BoA group, in that BAMLI was, like MLP, a subsidiary of
BANA. The Tribunal recorded that the Commission had not abandoned its case against
BAMLI just because the attorney who made the affidavit on behalf of the BoA entities
stated that Mr Cook had been employed by MLP. The Commission argued that the
superseding affidavit contained factual allegations from which a “reasonable possible
inference” might be drawn “that BAMLI or one of the BoA entities” participated in the
SOC with the requisite knowledge. After referring to one of Mr Cook’s email
addresses – which ended with the domain @[…] – the Tribunal said it would be
reluctant to dismiss the referral against BAMLI without hearing evidence as to whom
Mr Cook was representing at the time.197

[245] The CAC by contrast did not consider that significance could be attached to the
email addresses, and that the version contained in the BoA affidavits had not been
disputed. This seems to be a purely factual difference between the Tribunal and the
CAC, not raising any issue engaging this Court’s jurisdiction.

[246] In any event, it does not ultimately matter. BAMLI, like MLP, is a pure
peregrinus. The CAC concluded that the superseding affidavit failed to adequately
allege that MLP was a party to an SOC that included South African banks. I have found

197 Id at paras 281-7.

ROGERS J
110
that this conclusion did not involve any legal error. If that is so in relation to MLP, it
must also hold true for BAMLI, since the Commission’s case against BAMLI does not
rely on allegations different from those made against MLP.

[247] It follows that the Commission’s appeal in respect of BAMLI, MLP and BANA
must fail.

JPM Bank (fourth respondent)
[248] I held earlier, in relation to JPM Bank, that the CAC committed the third legal
error alleged by the Commission. Since that disposes of the only basis on which
JPM Bank defends the CAC’s order, the Commission’s appeal in respect of JPM Bank
must succeed.

ANZ (fifth respondent)
[249] I have already concluded, in the context of the alleged ninth and twelfth errors,
that the CAC’s reasoning in respect of ANZ does not suffer from those or any other
alleged legal errors. The Commission’s criticism essentially involves an attack on the
CAC’s factual assessment of the superseding affidavit . This does not engage our
jurisdiction, from which it follows that the Commission’s appeal against ANZ must fail.

SBSA (eighth respondent)
[250] In discussing the eleventh alleged error, I have set out the SBSA -specific
allegations contained in the superseding affidavit as well as the CAC’s reasoning.
Subject to one caveat, the Commission’s attack on the CAC’s reasoning concerns that
Court’s factual assessment of the pleaded case, something that does not engage our
jurisdiction.

[251] What must be considered, however, is whether the CAC erred in law in having
regard to the evidence in SBSA’s dismissal application about the roles of
Messrs Brownrigg and Taylor and about the two Reuters platforms. The Tribunal did

ROGERS J
111
not say that no regard could be had to the evidence of the roles of Messrs Brownrigg
and Taylor. What it said was that the evidence was “more in the nature of a defence
and cannot be determined without recourse to a factual enquiry”. 198 The CAC, by
contrast, relied on this evidence without addressing whether this was permissible in the
context of what was, at least according to the Commission, essentially an exception.
Counsel for the Commission in this Court assured us that this argum ent was made in
the CAC.

[252] SBSA does not accept that its dismissal application was no more than an
exception. It claims to have been entitled to bring a substantive dismissal application
because of the circumstances in which it came to be included in the referral and because
of the Commission’s alleged non -compliance with CAC I. As to the former, SBSA
stated in its dismissal application that its business reputation had been badly tarnished
by the Commission’s conduct. At no stage between the initiation of the complaint in
April 2015 and the referral in February 2017 did the Commission seek information from
SBSA or notify SBSA that its conduct was being investigated. It did not ask to
interview Mr Brownrigg or ask him to explain his conduct. SBSA was alarmed to find
itself named in the February 2017 referral, an event attended by much publicity.
Mr Brownrigg was publicly named without having been given any opportunity to
defend himself. SBSA regarded the Commission’s conduct as reckless and inconsistent
with the obligations of a public regulator.

[253] SBSA immediately sought through its attorneys to engage with the Commission
to understand the case against it, because it believed it had been erroneously included
in the referral. The Commission refused to engage. If the Commission had conducted
a prope r investigation, it would, says SBSA, have found out the facts alleged in its
dismissal application. In earlier proceedings, the Commission had been told under oath

dismissal application. In earlier proceedings, the Commission had been told under oath
that Mr Brownrigg was never employed by SBSA as a trader or authorised to trade.
This was confirmed in an affidavit from Mr Brownrigg himself.

198 Id at para 313.

ROGERS J
112

[254] Although the CAC did not pertinently address the admissibility of the evidence
contained in SBSA’s dismissal application, it did not necessarily commit an error of
law by having regard to it. I said, earlier in this judgment, that the flexible principles
which the Tribunal has adopted in relation to exceptions do not warrant a rigid rule that
under no circumstances may additional facts be taken into account; and that
considerations of fairness may in some circumstances justify regard being had to limited
additional facts. 199 Assuming in favour of the Commission that SBSA’s dismissal
application was in substance no more than an exception, it would not follow as a matter
of law that regard could not be had to the additional facts put up by SBSA.

[255] It would have been better for the CAC to explain why, in fairness, it thought that
regard could properly be had to SBSA’s allegations. However, and taking into account
the circumstances under which SBSA was joined and the nature of the additional
evidence, the absence of such an explanation in this particular instance d oes not point
to a legal error. The Commission failed, before adding SBSA to the referral, to direct
any enquiries to the bank. SBSA was thus deprived of the opportunity of presenting
facts to the Commission to show that its thesis in relation to the bank was factually
flawed.

[256] The evidence of Mr Brownrigg’s position within SBSA, and for that matter
Mr Taylor’s position at Barclays, is not evidence of a kind that one would expect to be
contentious. It is difficult to imagine that the Commission would have alleged that
Mr Brownrigg was an SBSA trader if, before including SBSA in the referral, it had
made basic enquiries with SBSA and been told the true position. It is fanciful to imagine
that a bank would lie on a matter of this kind (assuming it were inclined to lie at all),
since the lie would immediately be exposed by the most rudimentary process of
documentary discovery.

documentary discovery.


199 See [67] to [68].

ROGERS J
113
[257] If the Commission’s case against SBSA hinged on the identification of
Mr Brownrigg as a trader and of Mr Taylor as a competing fellow trader, one might
well ask why SBSA should sit through a long trial in circumstances where the
Commission has not raised any genuine factual dispute about the roles played by these
two gentlemen or, for that matter, about the two Reuters platforms. SBSA would not
even be able to look forward to the soothing balm of a costs order against the
Commission.

[258] I need only add that the two chatroom interactions between Mr Brownrigg and
Mr Taylor, as pleaded in the superseding affidavit , appear to lack any self -evident
sinister content. For the rest, and as I have said, the attack mounted by the Commission
on the CAC’s reasoning goes to the factual adequacy of the affidavit, which does not
engage our jurisdiction. For these reasons, I conclude that the Commission’s appeal in
respect of SBSA must fail.

Nomura (ninth respondent)
[259] The Nomura -specific allegations in the superseding affidavit and the CAC’s
related reasoning have been set out in my analysis of the sixth alleged error. There were
no legal errors tainting the CAC’s reasoning. What remain are criticisms of a factual
nature that do not engage this Court’s jurisdiction. The Commission’s appeal in respect
of Nomura must fail.

Commerzbank (twelfth respondent)
[260] I have dealt with the case against Commerzbank and the CAC’s reasoning in the
context of the eleventh alleged error. There were no legal errors. There is nothing else
that engages this Court’s jurisdiction. The Commission’s appeal in respect of
Commerzbank must fail.

ROGERS J
114
Macquarie (thirteenth respondent)
[261] What I have said in relation to Nomura applies also to Macquarie. The
Commission’s appeal in respect of Macquarie must fail.

HBUS (nineteenth respondent)
[262] The case in respect of HBUS has been addressed in the context of the eighth
alleged error. Once again, in the absence of any relevant legal error, the criticism of the
CAC’s adjudication of the adequacy of the allegations against HBUS does not engage
this Court’s jurisdiction. The Commission’s appeal in respect of HBUS must fail.

FRB (twenty-seventh respondent)
[263] It is convenient to deal with FRB before Nedbank. I have addressed the pleaded
case against FRB and the CAC’s related reasoning in the context of the seventh alleged
error. Since there were no relevant legal errors, the Commission’s appeal in respect of
FRB must fail because the attacks on the CAC judgment do not engage our jurisdiction.

Nedbank (twenty-fifth respondent)
[264] Earlier in this judgment I rejected Nedbank’s arguments on the question of
initiation and whether CAC I precluded its joinder. I have already noted that CAC II
does not contain any separate reasoning on the merits of the case against Nedbank,
despite the fact that Nedbank, by way of an exception, asserted the superseding
affidavit’s failure to make out a cause of action against it.

[265] The superseding affidavit contains the following Nedbank-specific allegations:
(a) No Nedbank traders are named.
(b) Nedbank is alleged to have joined the SOC “by at least 1 January 2008”.
(c) There are no allegations of Nedbank ’s participation in an y implicated
chatrooms.
(d) Ten incidents are pleaded of observed conduct on the “Reuters trading
platform”. They occurred on 1 January 2008, 7 July 2008,

ROGERS J
115
11-12 January 2009 (divided into three incidents which occurred over a
10-hour period between 20h59 and 07h07), 28 May 2010,
16 August 2011, 19 September 2012, 20 September 2012 and
28 September 2012.
(e) I have already dealt with the incidents on 1 January 2008 and
19 September 2012 in the context of the case against SBSA. I have also
dealt with the first incident of 11 January 2009, and the incidents of
28 May 2010, 20 September 2012 and 28 September 2012 in the context
of the case against FRB.
(f) The second incident of 11 -12 January 2009, which took place between
23h20 and 00h07, was that Nedbank and SNY/SAI put up “matching bid
and ask quotes in succession”, resulting in a consistent average of 9.8100.
This “pattern of price formation” took place over 47 minutes.
(g) The third incident of 11 -12 January 2009, which took place between
00h16 and 07h07 on 12 January (starting about nine minutes after the
second incident), was that Absa and SAI/SNY “posted the same ask and
bid quotes in succession”, starting at 9.9000 /9.8000 and rising in four
stages to 9.9300/9.8300, “before reducing in turn in matched amounts”.
In addition, and over the same period, SAI/SNY, Nedbank and Absa
“posted the same ask and bid prices” of 9.9100/9.8100 and of
9.9200/9.8200.
(h) The incident of 16 August 2011 took place between 04h10 and 05h54.
Nedbank “increased the spot exchange to create a focal point affirmed by
Investec, Absa and Nedbank at 7.1100”. Another “jump” then occurred
“to create a focal point” for SBSA, Absa and Investec at 7.1000. Finally,
“Investec and Nedbank join[ed] other banks to create a focal point at
7.1050”.
(i) In all, SBSA and FRB together were involved in one of the 10 Nedbank
incidents, while separately SBSA was involved in another five and FRB
in another three.

ROGERS J
116
[266] In Part A of its decision in TRIB II, the Tribunal dealt in only broad strokes with
the adequacy of the pleaded case against the various banks. The Tribunal did not give
individualised attention to the adequacy of the bank-specific allegations, and this is true
also of Nedbank, which is not even named in Part A of the decision. In Part B of its
decision, where the Tribunal dealt with the banks under individual headings, the
Tribunal said, of the complaint that the superseding affidavit failed to make out a cause
of action against Nedbank, t hat this fell to be dismissed “for the reasons provided in
Part A”. In short, the Tribunal appears not to have applied its mind to whether the
specific allegations against Nedbank made out a cause of action against that bank.

[267] The CAC held, for reasons I have rejected, that Nedbank, FRB and SAI could
not permissibly be joined. Despite this finding, the CAC analysed the merits of the case
against FRB and found it to be wanting, but it did not do the same for Nedbank. The
reason may be that, unlike Nedbank, FRB did not object to its joinder; its appeal in
CAC II went to the merits of its exception. Having found for Nedbank on its joinder
point, the CAC might have thought it unnecessary to consider the merits of the case
pleaded against that bank. Another possibility is that, amidst the welter of banks and
issues with which the Court had to deal, the CAC simply overlooked that it had not dealt
with the merits of the pleaded case against Nedbank.

[268] Whatever the explanation, the absence of a separate analysis by the CAC on the
merits of the pleaded case against Nedbank puts this Court at a disadvantage. A
non-apex court must always bear in mind that this Court may reverse it on an issue
which the other court regards as dispositive. It is for this reason that this Court has
stressed that it is in general desirable for non -apex courts to decide all issues. 200

stressed that it is in general desirable for non -apex courts to decide all issues. 200
Nevertheless, in this instance it would not be in the interests of justice to remit the matter
to the CAC for a decision on the remaining issues against Nedbank. We have the CAC’s
reasoning in relation to multiple other banks. It is clear enough that the CAC would not

200 Spilhaus Property Holdings (Pty) Ltd v Mobile Telephone Networks (Pty) Ltd [2019] ZACC 16; 2019 (4) SA
406 (CC); 2019 (6) BCLR 772 (CC) at para 45; Casino Association of South Africa v Member of the Executive
Council for Economic Development Environment Conservation and Tourism [2023] ZACC 39; 2024 (5) BCLR
611 (CC) at para 33; and United Manganese above n 107 at para 378.

ROGERS J
117
have found the case against Nedbank to pass muster, bearing in mind that, on
substantially the same allegations, the CAC held that the superseding affidavit failed to
plead an adequate cause of action against SBSA and FRB. I say so because Nedbank
was not alleged to have been involved in any incidents in which SBSA and FRB were
not also involved (in saying this, I treat the second and third incidents of
11-12 January 2009 as one continuing incident).

[269] The Commission’s application for leave to appeal does not raise, as a ground of
appeal, that the CAC failed to address the merits of the case against Nedbank. The
Commission’s written submissions and Nedbank’s counter -submissions squarely
address whether the superseding affidavit contains sufficient allegations to justify a
conclusion that Nedbank was party to the alleged SOC. However, these arguments are
of a kind that do not engage our jurisdiction. In the circumstances, the CAC’s decision
in respect of Nedbank must be allowed to stand.

SAI (twenty-eighth respondent)
[270] I have already rejected SAI’s arguments about initiation and about the
impermissibility of its post-referral joinder. What must now be considered is whether
there is any other basis on which SAI seeks to defend the CAC’s order in its favour.

[271] In its affidavit opposing joinder, SAI criticised the Commission for the delay in
bringing the joinder application and claimed that it was prejudiced by the delay. The
facts in this regard are as follows:201
(a) In the February 2017 referral , the Commission alleged that Messrs Katz
and Louis Friedman were traders representing SNY.
(b) In March 2017 SNY’s attorneys notified the Commission that SNY was
not a forex trader, that it had never employed Messrs Katz and Friedman
and that they had been employed by SAI.

201 This includes material from its affidavit in the dismissal application which was incorporated by reference into
its affidavit opposing its joinder.

ROGERS J
118
(c) The Commission ignored this information when supplementing the 2017
referral in March and April 2017.
(d) In May 2017 SNY filed an exception application. Attached to its
founding affidavit was the letter its attorneys had written to the
Commission in March 2017.
(e) In December 2017 and January 2018 the Commission filed a further
supplementary affidavit and its first joinder application. The Commission
did not seek to join SAI and made no reference to the information received
from SNY’s attorneys.
(f) In March 2018 SNY filed a further exception application, reiterating the
information previously given.
(g) Yet it was only in June 2020, in the superseding affidavit delivered
pursuant to CAC I, that the Commission first sought to include SAI in the
referral, now alleging that Messrs Katz and Friedman had represented
SAI.
(h) Contrary to the Commission’s assertion, SAI was prejudiced by the
delayed joinder. The first to twenty-third respondents had the opportunity
to participate and potentially influence the outcomes in TRIB I and CAC I.
SAI was denied that opportunity. Moreover, Mr Katz left SAI’s employ
in June 2010, and Mr Friedman in October 2014 when SAI closed its
forex desk. SAI was thus prejudiced in its ability to respond to the
complaint on its merits.

[272] In its replying affidavit the Commission did not deny the procedural history but
disputed that SAI was prejudiced. SAI did not contend that it would have adopted a
unique position on any of the issues litigated in TRIB I and CAC I, and it was unlikely
to have done so since it was represented by the same attorneys as SNY and SBSA. The
Commission did not address the alleged prejudice arising from Messrs Katz and
Friedman’s departure from SAI.

ROGERS J
119
[273] In TRIB II , the Tribunal correctly rejected SAI’s opposition based on the
supposed absence of an initiation and the related time -bar defence. It also rejected
SAI’s opposition based on the supposed lack of personal jurisdiction, a ground that SAI
has not pressed in this Court. The Tribunal considered that, in view of SAI’s
employment of Mr Katz until June 2010 and Mr Friedman until October 2014, it clearly
had a case to answer. The Tribunal, however, made no reference to delay.

[274] In CAC II, the CAC referred to some of the procedural background, including
SNY’s attorneys’ letter to the Commission in March 2017, which identified SAI as the
entity that had employed Messrs Katz and Friedman. 202 After dealing with other
grounds which I have found to be legally unsound, the CAC concluded:

“On the Commission’s own version it was only when it prepared its amended referral
on 1 June 2020 that the Commission considered including [SAI] in its referral. This
was not the kind of case where the Commission could have argued that as the cartel
operated in secret it had difficulty in detection because it had obtained clear information
as at 1 April 2015 which, had it wished, it could have also included [SAI]. It failed to
act prudently. There was, in short, no basis for the Tribunal to have upheld t he
Commission’s application to join [SAI] as the 28th respondent in the complaint
referral.”203

[275] It is unclear what “clear information” the Commission had as at 1 April 2015.
The information that SAI had been the employer of Messrs Katz and Friedman was first
communicated to the Commission in March 2017, shortly after SNY received the
February 2017 referral. Even so, the criticism that the Commission failed to act
prudently is not unjust. The Commission has not alleged that it undertook any
investigations of its own. It seems to have joined SAI on the strength of the information
supplied by SNY, yet it took more than three years to act on it.

supplied by SNY, yet it took more than three years to act on it.


202 CAC II above n 1 at paras 140-1.
203 Id at para 146.

ROGERS J
120
[276] If the paragraph I have quoted from the CAC’s judgment was intended to uphold
an independent objection by SAI based on prejudicial delay, it would be a decision on
an ordinary matter of practice and would not engage our jurisdiction. I am far from
satisfied, however, that this is the true import of the paragraph. The CAC did not
identify prejudicial delay as an independent ground of objection. If it was intending to
address such a ground, I would have expected it to deal with the question of prejudice.

[277] Although SAI’s counsel emphasised the Commission’s unexplained delay in her
oral submissions, SAI’s answering affidavit and written submissions in this Court do
not advance prejudicial delay as an independent basis for defending the CAC’s order or
even that SAI understood the CAC to have upheld an independent defence of prejudicial
delay. SAI identified the critical questions before the Tribunal and CAC in respect of
SAI as being : (a) whether CAC I permitted the addition of further respondents;
(b) whether the Commission had initiated a complaint in respect of SAI; and (c) if such
a complaint was initiated, whether it was timeous, having regard to the time -bar in
section 67(1). SAI’s written submissions in this Court follow the same pattern.

[278] SAI said in its affidavit in this Court that the CAC found in its favour on the first
two questions and thus did not have to deal with the third. SAI continued that, on the
undisputed facts, the time -bar would have presented an insurmountable hurdle to the
Commission. Although Pickfords held that the Commission’s failure to comply with
the three -year time -limit in section 67(1) could in principle be condoned, the
Commission’s imprudent delay in the present case would have been fatal to
condonation. This perhaps explains the emphasis on delay; it was viewed as being
relevant to the time-bar defence.

[279] However, on the assumption that prejudicial delay needs to be addressed as an

[279] However, on the assumption that prejudicial delay needs to be addressed as an
independent objection to SAI’s joinder, the issue was not properly considered either in
TRIB II or CAC II. Since the legal grounds on the strength of which the CAC found for
SAI are matters within our jurisdiction, and since we have decided them against SAI, it

ROGERS J
121
is within our power to determine what should happen in respect of the independent and
as yet undecided objection based on prejudicial delay.

[280] In my view, we should not find for SAI on this ground. Although the
Commission was guilty of unreasonable delay, the case has still not got underway on
the merits.204 Real prejudice has not been shown. It is fanciful to think that a separate
argument by SAI would have led to a different outcome in CAC I. SAI has not stated
that it has any new insights justifying a more favourable outcome on personal
jurisdiction than the one laid down in CAC I. In regard to the departure of Messrs Katz
and Friedman, SAI has not stated that it does not have access to them or that any
difficulties that might exist in getting their versions were increased by the delay from
March 2017 to June 2020.

[281] It follows that, in respect of SAI, the Commission’s appeal must succeed.

HBEU’s (fourteenth respondent’s) application for leave to cross-appeal in CCT 30/24
[282] I dealt with the Commission’s case against HBEU and the CAC’s reasoning
when discussing the eighth alleged error. HBEU relies inter alia on evidence put up by
HBUS, in opposition to its joinder, for the contention that Mr Hatton was employed by
HBUS, not HBEU, and that the Commission “has the wrong guy”. HBEU complains
that the CAC relied on similar “wrong guy” evidence when exonerating BAMLI.

[283] The problem for HBEU, as it has been for the Commission in a number of
instances above, is that this Court’s jurisdiction is not engaged by a complaint that the
CAC should have assessed the facts differently or even that the CAC got the facts
plainly wrong. It is contrived for HBEU to argue, as it did, that the supposedly
differential treatment implicates its constitutional right to equality. That a factual line
of reasoning applied by the CAC in relation to BAMLI could arguably also have been
adopted (or adapted) in relation to HBEU does not in itself implicate section 9.

adopted (or adapted) in relation to HBEU does not in itself implicate section 9.

204 Compare SA Steel Equipment Co (Pty) Ltd v Lurelk (Pty) Ltd 1951 (4) SA 167 (T) at 174G-175F.

ROGERS J
122

[284] Different facts were pleaded as against BAMLI on the one hand and HBEU on
the other. In BAMLI’s case, the CAC said that the Commission had been informed on
oath that Mr Cook was at all material times employed by MLP, “a fact never denied by
the Commission”. In HBEU’s case, by contrast, the CAC said that there was “no clear
denial” by HBEU that Mr Hatton had nothing to do with its activities. HBEU argues
that this statement is wrong and that there was such a denial. The Commission submits
that there was no such denial in the record that served before the CAC in relation to
HBEU’s exception. It is unnecessary to delve into this, because even if the CAC’s
statement was wrong, it is an error in assessing the pleaded facts. There is nothing in it
that engages this Court’s jurisdiction.

[285] HBEU’s associated company, HBUS, resisted the Commission’s appeal inter
alia on the basis that the superseding affidavit did not, in relation to subject matter
jurisdiction, contain allegations to show that the effects of the SOC’s conduct meet the
QE test. Because I have found for HBUS on other grounds, I have not needed to address
those submissions. HBEU, in its application for leave to cross -appeal, did not raise as
a ground of appeal that the CAC should, quite apart from the “wrong guy” point, have
found for HBEU because subject matter jurisdiction was not sufficiently pleaded.
HBEU’s written and oral submissions likewise did not include such a contention.

[286] It follows that HBEU’s cross-appeal must fail.

BNP’s application for leave to appeal in CCT 25/24
[287] BNP, the second respondent in CCT 30/24, has applied for leave to appeal in
CCT 25/24. BNP, a local peregrinus, filed a pure exception. BNP complained, in the
first place, that the superseding affidavit was vague and embarrassing on the duration
of the SOC. BNP’s exception pointed to the following allegedly conflicting statements
in the affidavit:

in the affidavit:
(a) that the banks conspired to manipulate the ZAR “between 2007 and
2013”;

ROGERS J
123
(b) that “with effect from September 2007 and until at least September 2013”
the respondents reached an agreement to participate in the SOC;
(c) that the Commission “does not know the date on which the [SOC]
between the respondent banks ceased to operate, or if it has indeed ceased
to operate”;
(d) that the SOC “continued to operate using different modes of
communication after September 2013, alternatively after the implicated
chatrooms were closed, alternatively after the respondents’ traders no
longer had access to the Bloomberg electronic messaging system”; and
(e) that the Commission “invites” the respondents to explain “if and when the
[SOC] ceased to operate” and “the date on which the respondent s’
participation in the conspiracy ceased”.

[288] BNP contended that these allegations were contradictory and could not all be
correct. The period of BNP’s participation in the SOC was important , because it was
“one of the determinants of the quantum of the penalty” and BNP was “entitled to be in
a position to assess its potential liability”. BNP also complained that it was unable to
assess the application of the time-bar in section 67(1).

[289] BNP’s exception contended, in the second place, that the superseding affidavit
was vague and embarrassing as to the individual banks’ continued participation in the
SOC. In that regard, the exception pointed to the following further allegations in the
affidavit:
(a) that the Commission “does not know the date on which each respondent
ceased to participate in the conspiracy, or if they have ceased to
participate in the conspiracy”; and
(b) that “the respondents continued as participants in the conspiracy,
alternatively are deemed to have continued as participants in the
conspiracy, until such time as they expressly dissociated themselves with
the conduct, terms and objective of the conspiracy”.

ROGERS J
124
[290] BNP complained that these allegations, pertaining to each bank’s participation
in the SOC, contradicted the allegations about the duration of the SOC itself. The
allegations could not all be correct. BNP was thus embarrassed in pleading.

[291] The Tribunal gave no individualised attention to BNP’s complaints. The CAC’s
reasoning is not quite clear. The Court indicated that the dismissal of an exception
complaining of vagueness and embarrassment is not ordinarily appealable. In the light
of this Court’s judgment in Lebashe,205 appealability would depend on an assessment
of the interests of justice.206

[292] The CAC went on to say that BNP employed Mr Katz from September 2011 to
2013. (The superseding affidavit alleged that in 2013 Mr Katz moved from BNP to
ANZ.) Mr Katz had been the “glue” which bound a number of the other traders together
in sharing confidential information. This placed a burden on BNP to explain its
conduct. Bearing in mind that BNP’s complaint wa s confined to vagueness and
embarrassment, the Commission had, in the CAC’s view, made out a sufficient case
calling for an answer from BNP.

[293] Insofar as non-compliance with CAC I was concerned, in particular relating to
the SOC’s terminal date, the superseding affidavit alleged “that the SOC ended in 2013
and that Mr Katz was employed until that date”, so there was, in the CAC’s view, “very
little merit” in BNP’s argument that it was adversely affected by the alleged
non-compliance.207 (The CAC’s statement that the affidavit alleged that the SOC
“ended in 2013” fails to take account of BNP’s complaint that there were other
allegations made by the Commission in relation to the SOC which were inconsistent
with this terminal date.)


205 United Democratic Movement v Lebashe Investment Group (Pty) Ltd [2022] ZACC 34; 2022 (12) BCLR 1521
(CC); 2023 (1) SA 353 (CC).
206 CAC II above n 1 at paras 147-9.
207 Id at para 151.

ROGERS J
125
[294] It is unclear whether the CAC intended, by these latter remarks, to find that it
was not in the interests of justice to treat the dismissal of BNP’s exception as appealable
or whether it intended to reject the appeal on its merits. The CAC’s order suggests the
latter, since BNP’s appeal was dismissed, not struck from the roll.

[295] In my view, the dismissal of BNP’s exception, grounded in alleged vagueness
and embarrassment, is not appealable. That would usually be the case when such an
exception is dismissed. Applying Lebashe, there are no overriding considerations of
justice which require us to treat this interlocutory order as nevertheless constituting an
appealable order. No overriding constitutional interests are at stake , nor does the
exception raise any arguable point of law of general public importance. A litigant’s
rights under section 34 of the Bill of Rights are not automatically implicated where a
court requires the litigant to plead a case which, in the litigant’s opinion, is vague and
embarrassing.

[296] The primary supposed inconsistency , which was the focus of BNP’s counsel ’s
oral argument in this Court , was between the allegation that the SOC lasted “until at
least September 2013” and the allegation that the Commission does not know the date
on which the SOC ceased to operate, if indeed it had. This, in my opinion, is an
unobjectionable way of pleading. Those two allegations convey that the Commission
knows that the SOC lasted at least until September 2013 but does not know when after
that date, if at all, it ceased. This occasions no embarrassment.

[297] The allegation that the respondents continued as participants in the SOC, or are
deemed to have done so until they expressly dis associated themselves from it, appears
to me to be an unobjectionable legal assertion. It conveys the Commission’s stance that
cessation in respect of any particular bank will require evidence from such bank of a

cessation in respect of any particular bank will require evidence from such bank of a
positive act of disassociation. The correctness of the legal assertion need not detain us.
If it is sound, it can stand alongside the two allegations discussed in the i mmediately
preceding paragraph, on the basis that the Commission does not know whether after
September 2013 there were such acts of disassociation.

ROGERS J
126

[298] By contrast, the allegation that the SOC “continued to operate using different
modes of communication after September 2013” is a positive assertion seemingly at
odds with those I have just been discussing. But even if this creates some vagueness
and embarrassment, BNP should have no difficulty in filing an answer on the merits. If
BNP’s internal investigations reveal that it must admit its participation in the SOC to
some extent, it will plead when the admitted conduct ceased, and deny any allegations
in the superseding affidavit inconsistent with its version. BNP can assess the viability
of a time -bar defence and its potential liability for penalties on the basis of its own
version of the facts.

[299] Allegations in the superseding affidavit that are too vague to allow BNP to
undertake targeted internal investigations (the allegation of the continuation of the SOC
after September 2013 may fall into this category) can, if necessary, be placed in issue
with an appropriate explanation as to why BNP cannot be more specific. BNP in due
course will be entitled to seek directions from the Tribunal entitling it to request further
particulars of any post -September 2013 conduct on which the Commission intends to
rely.208

[300] It follows that BNP’s application for leave to appeal must be refused.

CSS’ application for leave to appeal in CCT 27/24
[301] CSS, the twenty-third respondent in CCT 30/24, has applied for leave to appeal
in CCT 27/24. As with Nedbank and SAI, I have dealt with issues relating to initiation
and post-referral joinder. CSS’ arguments in that regard have been rejected.

[302] The superseding affidavit contains the following CSS-specific allegations:

208 The Tribunal Rules do not make specific provision for further particulars, but the Tribunal often orders such
particulars to be given, either for pleading or trial preparation, the guiding principle being fairness : Competition

Commission v Tiger Brands Ltd t/a Albany; Competition Commission v Pioneer Foods (Pty) Ltd t/a Sasko [2009]
ZACT 34 at para 48. See also Paramount Mills (Pty) Limited v Competition Commission [2012] ZACAC 4;
[2012] 2 CPLR 215 (CAC) at para 63.

ROGERS J
127
(a) Its traders were Messrs Hatton (who moved from HBUS to CSS with
effect from November 2010) and Heinrich Putter.
(b) CSS joined the SOC “by at least 3 November 2010”.
(c) The affidavit alleges 11 chatroom incidents in which Mr Hatton was
involved for CSS on eight dates 209 between November 2010 and
April 2011. The only other banks in these chatroom incidents were
foreign banks.
(d) The affidavit alleges one chatroom incident involving Mr Putter, the other
trader being Mr Katz (at that time with Barclays). This was on
17 May 2011. Commercially sensitive information was allegedly shared.
(e) CSS is not alleged to have been involved in any specific incident of
market conduct observed on the Reuters platform.

[303] The CAC reasoned thus in respect of CSS:210
(a) The superseding affidavit pleaded eight chatroom incidents 211 involving
its two traders. Most of these involved Mr Hatton, “a particularly active
member of the Old Gits chatroom, the longest running and most prolific
of all the implicated chatrooms”.
(b) CSS’ counsel had argued that CSS was predominantly in communication
with foreign banks – BoA, Citibank and SCB. However, so the CAC
observed, Barclays was also in the Old Gits chatroom. The evidence
suggested that other participants in these chatrooms in cluded Citibank,
which conducted a banking business in South Africa and had settled with
the Commission, and SCB, which also had an office in South Africa and
had also settled with the Commission. It was thus incorrect to say that the
case against CSS was that it was a party to chats only with “American
banks”.

209 There were two incidents on each of 3 November 2010, 2 December 2010 and 6 January 2011.
210 CAC II above n 1 at paras 137-9.
211 I have stated that there were 11 incidents, which occurred on eight dates. The CAC may have been referring
to the eight dates on which Messrs Hatton and Putter were involved for CSS.

ROGERS J
128
(c) Moreover, CSS was a local peregrinus, so that only subject matter
jurisdiction was required.
(d) CSS’ counsel had stressed non-compliance with CAC I, complaining that
CSS could not investigate whether the complaints against it were
time-barred. This was because the superseding affidavit had not provided
the date on which the SOC ended or when CSS’ participation ceased. In
substance, however, the affidavit prima facie implicated CSS. The
“intensity” of Mr Hatton’s involvement “on behalf of a local peregrinus”
justified the conclusion that CSS had a case to answer.

[304] It is common cause that, contrary to what the CAC stated, CSS is not a local
peregrinus. It is a pure peregrinus. Personal jurisdiction thus indeed had to be
established. This is not a mere error in the assessment of the adequacy of the pleaded
facts. It implicates the statutory power of the Commission and the Tribunal to prosecute
and adjudicate the case against CSS. The exercise of public power by these institutions
is a constitutional matter.

[305] The question is thus whether the Tribunal has personal jurisdiction in respect of
CSS. Based on the law applicable in this case (as laid down in CAC I), this depends on
compliance with the ACF test. The CAC observed that CSS was alleged to have been
involved in implicated chatrooms inter alia with SCB and Citibank, both of which were
local peregrini with South African branches. The CAC did not, however, explicitly
state that this sufficed for personal jurisdiction, perhaps because it went on to
say – mistakenly – that CSS was a local peregrinus. In respect of other pure peregrini,
such as MLP and HBUS, the CAC plainly did not regard interactions with Citibank and
SCB as satisfying the ACF test.

[306] As I mentioned earlier, the superseding affidavit does not allege that
participation by pure peregrini in an SOC including only local peregrini as the other

participation by pure peregrini in an SOC including only local peregrini as the other
participants is an adequate connecting factor for purposes of personal jurisdiction. I
have also said that, while impugned communications with the South African branch of

ROGERS J
129
a local peregrinus might suffice to satisfy the ACF test, the same cannot be said if the
relevant interactions had nothing to do with the local peregrinus’ South African branch.
The affidavit does not allege that any of the local peregrini’s implicated traders worked
at the South African branches of those banks.

[307] In the circumstances, CSS’ appeal must succeed.

Conclusion and costs
[308] The overall result is as follows. In CCT 25/24, BNP’s application for leave to
appeal must be refused. The other applications for leave to appeal, as I said earlier in
this judgment, should be granted. In CCT 27/24, CSS’ appeal must succeed . Since
CSS opposed its joinder inter alia on the basis that the Tribunal lacked personal
jurisdiction, the correct substitute order is one having the effect of reversing the
Tribunal’s decision joining CSS, thus rendering a decision on CSS’ exception
application moot. In CCT 30/24, the Commission’s appeal succeeds against JPM Bank
and SAI, but fails against BAMLI, ANZ, SBSA, Nomura, Commerzbank, Macquarie,
HBUS, MLP, BANA, Nedbank and FRB. In the same case, HBEU’s cross-appeal fails.

[309] This means that, at the level of pleading, the SOC case may go ahead against the
following active respondents: BNP, JPM Co, JPM Bank, Investec, HBEU and SAI.
Additionally, although the relevant banks have settled or been granted leniency, the
pleaded SOC will also include SCB, Citibank, Absa and Barclays.

[310] Costs should not be awarded against the Commission where it has acted in good
faith, even if zealously. 212 On this basis, the CAC did not award costs against the
Commission (to the extent it lost) or in favour of the Commission (to the extent it won).
In my view, we should follow the same course. I would, however, make an exception

212 Competition Commission of South Africa v Pioneer Hi-Bred International Inc [2013] ZACC 50; 2014 (2) SA

480 (CC); 2014 (3) BCLR 251 (CC); [2015] 1 CPLR 1 (CC) at paras 19-28 and Coca-Cola Beverages Africa (Pty)
Ltd v Competition Commission [2024] ZACC 3; 2024 (4) SA 391 (CC); 2024 (6) BCLR 771 (CC); [2024] 7 BLLR
665 (CC); (2024) 45 ILJ 1507 (CC) at para 94.

ROGERS J
130
in relation to BNP. In my view, its application for leave to appeal was sufficiently
lacking in substance that it should be dismissed with costs, including the costs of two
counsel. For the guidance of the Taxing Master, the case as a whole was argued ov er
three days, and the argument in respect of BNP’s application did not take up more than
one hour.

Order
[311] The following orders are made:

Case CCT 25/24
Leave to appeal is refused with costs, including the costs of two counsel.

Case CCT 27/24
1. Leave to appeal is granted.
2. The appeal succeeds.
3. Paragraph 2 of the order of the Competition Appeal Court, insofar as it
relates to the applicant in Case CCT 27/24 (the twenty -third respondent
in the Tribunal proceedings), Credit Suisse Securities (USA) LLC, is set
aside and replaced with the following order:
“(a) The appeal against the joinder of Credit Suisse Securities (USA)
LLC (CSS), forming part of paragraph A[1] of the Tribunal’s order,
succeeds.
(b) The Tribunal’s decision in that respect is set aside and replaced
with an order dismissing the Competition Commission’s
application to join CSS.”

Case CCT 30/24
1. The applicant, the Competition Commission (Commission), is granted
leave to appeal.

ROGERS J
131
2. The fourteenth respondent, HSBC Bank plc (HBEU), is granted leave to
cross-appeal.
3. The Commission’s appeal is upheld in relation to the fourth respondent,
JPMorgan Chase Bank N.A. (JPM Bank), and the twenty -eighth
respondent, Standard Americas Incorporated (SAI).
4. The Competition Appeal Court’s decision in respect of JPM Bank and
SAI is set aside and replaced with orders dismissing those parties’ appeals
against the Tribunal’s decision.
5. Save as aforesaid, the Commission’s appeal is dismissed.
6. HBEU’s cross-appeal is dismissed.

Case CCT 25/24

For the Applicant: J Campbell SC and S Mohammed
instructed by Bowman Gilfillan
Incorporated

For the Respondent: T Ngcukaitobi SC, F Hobden, L Zikalala,
H Drake, M Nxumalo and K Monareng
instructed by Ndzabandzaba Attorneys
Incorporated

Case CCT 27/24

For the Applicant: M Norton SC, G Marriott and N Nyembe
instructed by Werksmans Attorneys

For the Respondent: T Ngcukaitobi SC, F Hobden, L Zikalala,
H Drake, M Nxumalo and K Monareng
instructed by Ndzabandzaba Attorneys
Incorporated

Case CCT 30/24

For the Applicant: T Ngcukaitobi SC, F Hobden, L Zikalala,
H Drake, M Nxumalo and K Monareng
instructed by Ndzabandzaba Attorneys
Incorporated

For the First, Twentieth and Twenty-First
Respondents:
P Farlam SC, P Ngcongo and M Mbikiwa
instructed by Webber Wentzel

For the Third and Fourth Respondents: D A Turner SC instructed by Webber
Wentzel

For the Fifth Respondent: C D A Loxton SC and P Maharaj-Pillay
instructed by Cliffe Dekker Hofmeyr
Incorporated

For the Eighth Respondent: K Hofmeyr SC and L Mokwena instructed
by Herbert Smith Freehills Kramer South
Africa Attorneys Incorporated

For the Ninth Respondent: M M Le Roux SC instructed by Mackenzie
Granville Incorporated

For the Twelfth Respondent: A R Bhana SC and K L Williams
instructed by Webber Wentzel

For the Thirteenth Respondent: J Wilson SC and P Bosman i nstructed by
Bowman Gilfillan Incorporated

For the Fourteenth and Nineteenth
Respondents:
A Cockrell SC and C Avidon instructed by
Herbert Smith Freehills Kramer South
Africa Attorneys Incorporated

For the Twenty-Fifth Respondent: A Gotz SC and T Marolen instructed by
Werksmans Attorneys

For the Twenty-Seventh Respondent: M Wesley SC, N Muvangua and
N Mahlangu instructed by Cliffe Dekker
Hofmeyr Incorporated

For the Twenty-Eighth Respondent: M Engelbrecht SC and L Mokwena
instructed by Herbert Smith Freehills
Kramer South Africa Attorneys
Incorporated