IN THE HIGH COURT OF SOUTH AFRICA
GAUTENG LOCAL DIVISION, PRETORIA
Case No: 2022/026554
In the matter between:
MOBILE TELEPHONE NETWORK (PTY) LTD Applicant
and
INDEPENDENT COMMUNICATIONS
AUTHORITY OF SOUTH AFRICA First Respondent
VODACOM (PTY) LTD Second Respondent
TELKOM SOC LTD Third Respondent
CELL C (PTY) LTD Fourth Respondent
LIQUID TELECOMMUNICATIONS SOUTH
REPORTABLE: YES
OF INTEREST TO OTHER JUDGES: NO/YES
19 September 2025
SIWENDU J DATE:
2
AFRICA (PTY) LTD Fifth Respondent
RAIN (PTY) LTD Sixth Respondent
FREE MARKET FOUNDATION Seventh Respondent
ICT SMME CHAMBER Eighth Respondent
INTERNET SERVICE PROVIDERS’
ASSOCIATION NPC Ninth Respondent
MEDIA MONITORING AFRICA Tenth Respondent
SOUTH AFRICAN COMMUNICATIONS FORUM Eleventh Respondent
AFRIHOST SP (PTY) LTD Twelfth Respondent
JUDGMENT
Siwendu J
Introduction
[1] This is an application to review, set aside, and declare invalid, certain
provisions of the Mobile Broadband Services Regulations, 2021 (the
Regulations),1 published by the Independent Communication Authority of South
Africa (ICASA). The Regulations purport to regulate and remedy competition in
1 Mobile Broadband Services Regulations, 2021, GN 1960 in GG 46155 of 31 March 2022.
3
the mobile telecommunication services market and impose pro -competitive
conditions on market operators with significant market power.
[2] The applicant, Mobile Telephone Networks (Pty) Ltd (MTN), is a mobile
network operator ( an MNO alternatively, an operator) and offers mobile and
fixed-line telecommunication services to approximately 30 million subscribers in
South Africa. MTN has an individual electronic communications network service
licence issued by ICASA. ICASA is the first respondent and an organ of state
established by s 3(1) of the Independent Communications Authority of South
Africa Act 13 of 2000 ( the ICASA Act). Its objects include the regulation of
electronic communications in the public interest. 2 MTN is subject to regulation
by ICASA.
[3] The second respondent is V odacom (Pty) Ltd (V odacom) and is regarded
as the largest mobile telecommunications services provider in the South African
market, after MTN. V odacom opposes the application on limited grounds
pertaining to the relief sought by MTN . I n the event the review application
succeeds, it abides by the outcome of the review application.
[4] The third and fourth respondents are Telkom SOC Ltd (Telkom) and Cell
C (Pty) Ltd (Cell C) and are MNOs. The fifth and sixth respondents are Liquid
Telecommunications South Africa (Pty) Ltd (Liquid) and Rain (Pty) Ltd (Rain) ,
who are mobile virtual network operators (MVNO). The seventh to eleventh
respondents are interest ed organisations, associations and not for profit
companies involved in free trade, Information and Communications Technology
sector (ICT sector), the provision of internet services, and th e monitoring of the
media and communications. The third to twelfth were cited for their interest in
2 Section 2(b) of the Independent Communications Authority of South Africa Act 13 of 2000 (ICASA Act).
4
the subject matter of the review. No relief is sought against them, and they do not
participate in the review.
[5] ICASA’s regulatory activities are performed under the auspices of the
Council established terms of s 5 of the ICASA Act. Section 4B (1) of the ICASA
Act empowers ICASA to conduct an inquiry into any matter concerning the
achievement of the objects, regulations and guidelines, compliance with
regulations, guidelines and licences made in terms of the of ICASA Act and
underlying statutes falling under its jurisdiction. The relevant underlying statutes
are the E lectronic Communications Act 36 of 2005 (the ECA) and the
Competition Act 89 of 1998 (the Competition Act).
[6] In August 2018, ICASA embarked on an inquiry into the electronic
communications sector to determine markets to be prioritised for a market review
(priority markets). That review identified two market categories for mobile
broadband services, namely (i) the retail and (ii) wholesale markets. In November
2018, it decided to conduct an inquiry into ‘mobile broadband services’, which
are also referred to as ‘mobile data services’ to promote competition in the ICT
sector.3 Following findings about the functioning of the identified markets, it
promulgated the regulation subject to this review.
[7] MTN challenges these Regulations on several grounds, but before I deal
with the basis for the review, a n overview of the mobile telecommunication
service market and value chain is necessary to give context to the impugned
Regulation.
3 Notice of intention to conduct market inquiry into mobile broadband services, GN 713 of 2018, GG 42044 of
16 November 2018.
5
Mobile telecommunications services market and value chain
[8] MTN and V odacom entered the mobile telecommunications market early.
It is common cause that they enjoy almost 100% national population coverage. It
is not disputed that both operators have invested substantially in their
infrastructure. MTN calculates its capital expenditure and investment on
infrastructure at approximately R100 billion over more than 20 years.
[9] Mobile telecommunication services are delivered to consumers in the
following manner: by providing international bandwidth, through submarine
cables that connects South African MNOs with the rest of the world. Certain
MNOs have installed alternative fixed networks at national transit level and have
high-capacity fixed connections between each of their core networks and the
national transit network. The MNOs connect the South African population from
the termination points of the submarine cables by a high capacity fibre optic
transmission links between c ities and towns and the service providers’ point of
presence.
[10] End users in South Africa connect to the network using their mobile
devices through wireless transmission technology infrastructure, provided by the
MNO. The infrastructure includes radio frequency spectrum (to transmit the radio
signal) and base transceiver station , Radio Access Network ( RAN) electronics,
fibre or microwave backhaul and core transmission, controllers and core/network
management systems. In addition, RAN sites, usually located at municipal and
metro level are utilised to host RAN towers, antennae,4 shelters and backup power
systems. MNOs or third parties operate and maintain the network infrastructure,
purchase and manage internet protocol connectivity to deliver end to end data
services to enable end users a connection.
4 Antennae facilitate the ability to connect a subscriber’s device to a network.
6
[11] Access to suitable sites is an important part of the value chain . They are
utilised to instal l transmission infrastructure to provide network coverage. It is
common cause that the se sites range from ‘macro-solution’ like rooftops and
indoor sites to ‘micro-solutions’ like lampposts and billboards , referred to as
‘micro-sites.’
[12] MNOs like Telkom Mobile (linked to Telkom, a well-known fixed landline
operator) and Cell C, entered the mobile telecommunications market after MTN
and V odacom. Cell C and Telkom Mobile do not have national coverage on their
own networks, but achieve it by roaming on the network of either V odacom or
MTN.
[13] MNOs with limited geographic footprint enter the mobile
telecommunications market by: first, acquiring access to the use of radio
frequency spectrum and base transceiver station (known as infrastructure-based
entry); or second by concluding agreements with other MNOs or MVNO s for
roaming and through access point name (APN) services (known as services-based
entry). These are categorised as either: (a) passive infrastructure sharing, which
involves ‘passive’ network elements such as the site (such as the piece of land or
rooftop on which the mast is located) or the mast (such as the tower or other forms
of base transceiver station ), or (b) ‘active’ infrastructure sharing, which entails
sharing ‘active’ network elements such as antennae or RAN.
[14] MNO entrants often enter into agreements with MTN and V odacom for use
of their infrastructure for connectivity or for roaming 5 outside of their area
network to either gain capacity or national coverage for their customers.
5 Roaming means that the traffic of an MNO's customer is carried and routed on another MNO's network.
7
[15] The review is confined to MNO’s operating within the South African
telecommunications market.
Regulation making process
[16] ICASA’s market regulation power and task are common cause. Sections
4(3)(b) and (j)6 together with s 4B(1) (a) and (b) of ICASA Act7 confer it the
power to monitor, inquire into and regulate sectors falling under its jurisdiction.
[17] On 16 November 2018, ICASA published a Notice of Intention to Conduct
Mobile Broadband Services Market Inquiry (MBSMI) 8 into services offered by
MNOs to assess the state of competition and determine whether there are markets
or segments of the markets which warrant regulation. The notice states that
ICASA would conduct an inquiry in six phases by, namely:
(a) The announcement of the market inquiry; solicitation of views, information,
opinions and representations from market participants and stakeholders through
a questionnaire requiring data between 2015 to 2018. Participants were given an
election to make oral submissions if ICASA decided to hold public hearings.
(b) Publish a Discussion Document informed by the information submitted by
stakeholders and any other research or benchmarking exercises which it may
conduct.
6 Sections 4(3)(b) and (j) of the ICASA Act read:
‘Without derogating from the generality of subsections (1) and (2), the Authority-
(a) . . .
(b) must monitor the broadcasting, postal and electronic communications sectors to ensure compliance with
this Act and the underlying statutes;
. . .
(j) may make regulations on any matter consistent with the objects of this Act and the underlying statutes
or that are incidental or necessary for the performance of the functions of the Authority’.
7 Section 4B(1)(a) and (b) read:
‘(1) The Authority may conduct an inquiry into any matter with regard to-
(a) the achievement of the objects of this Act or the underlying statutes;
(b) regulations and guidelines made in terms of this Act or the underlying statutes’.
(b) regulations and guidelines made in terms of this Act or the underlying statutes’.
8 See Notice of intention to conduct market inquiry into mobile broadband services, GN 713 of 2018, GG 42044
of 16 November 2018.
8
(c) Public Hearings on the Discussion Document if deemed necessary
(d) Publish in the Gazette, a summary of Findings Document and draft
Regulations for comment (if necessary)
(e) Hold Public hearings, if necessary,
(f) Final Regulations and the reasons document and will publish in the
Government Gazette the final regulations and the reasons document.
[18] A Mobile Broadband Inquiry Questionnaire, as mentioned in the notice,
and available on ICASA’s website, solicited information about the retail and
wholesale of mobile services markets for the period from 2015 to 2018 . Where
detailed subscriber information was required, ICASA considered it adequate for
MNOs to provide the data based on a sample of 30 000 subscribers.
[19] MTN made its submission to ICASA on 29 March 2019, based on a sample
of 1 million subscribers. It reasoned a larger sample from which to draw
conclusions would be a more representative.
[20] On 29 November 2019, ICASA issued a discussion document as
preliminary findings based on information solicited from MNOs, international
data and its internal market research. 9 MTN submitted its response to the
discussion document on 27 February 2020. Pursuant to the public hearings held
in October 2020, MTN addressed supplementary submissions, these dated 11
November 2020 and 24 February 2021 , to ICASA. The submissions detail ed
several misgivings about ICASA’s approach and findings to market regulations.
9 See ‘Discussion document on mobile broadband services inquiry for public comments’ ICASA, 29 November
2019. https://www.icasa.org.za/news/2019/discussion-document-on-mobile-broadband-services-inquiry-for-
public-comments (accessed 19 August 2025). Published in GN 1560 in GG 42878 of 2 December 2019.
9
[21] On 26 March 2021, ICASA published the draft regulations together with
its final findings document .10 On 28 May 2021 , MTN responded to the draft
regulations and the findings document. ICASA held public hearings in August
2021, where MTN made a presentation and recommendations on ICASA’s
findings. According to MTN, ICASA invited comments on the draft regulations
but not on the findings document. As will be seen later in the judgment, MTN’s
persistent complaint is about procedural unfairness . MTN alleges that ICASA
considered the findings document as ‘final’ and failed to properly consider
representations and information placed before it.
[22] On 31 March 2022, ICASA published the impugned Mobile Broadband
Services Regulations, 2021, together with a reasons document . The reasons
document11 largely mirrors the findings document . ICASA’s market inquiry
findings are linked inextricably with each of the impugned Regulations; it is
convenient to deal with those findings in conjunction with the Regulation
challenged.
[23] It is noteworthy that on 5 July 2024, after the launch of the review, ICASA
amended the regulations. It deleted and substituted the provisions of regulation
7(e) and (g) and amended regulation 7 (h)(iii) and (v). This amendment is
published as the Mobile Broadband Services Amendment Regulations, 2024.12
[24] All parties agree, consistent with settled authority that the making of
regulations is reviewable under Promotion of Administrative Justice Act 3 of
2000 (PAJA).13 However, at the hearing, the debate centred on the boarders of the
10 Draft Mobile Broadband Services Regulations pursuant to section 67(4) of the Electronic Communications Act
No. 36 of 2005, GN 272 in GG 44337 of 26 March 2021, which includes the findings document.
11 The reasons document forms a part of the Mobile Broadband Services Amendment Regulations, 2024, GN 2617
in GG 50910 of 5 July 2024.
in GG 50910 of 5 July 2024.
12 Mobile Broadband Services Amendment Regulations, 2024, GN 2617 in GG 50910 of 5 July 2024.
13 City of Tshwane Metropolitan Municipality v Cable City (Pty) Ltd [2009] ZACC 34; 2010 (3) SA 589 (SCA)
para 10.
10
court’s power to interfere with ICASA’s findings of fact forming the basis for the
Regulations. I must point out on this score that s 3(5) of the ICASA Act states
that:
‘A person affected by any action, finding or decision of the Authority may apply to a court with
competent jurisdiction for review of that action, finding or decision.’ (My emphasis.)
From a plain reading of s 3(5), it appears it would have been legally permissible
for MTN to review ICASA’s findings under this section of the ICASA Act. The
findings constitute the reasons for the Regulations.
[25] MTN elected to bring the review of the regulations in terms of PAJA . Its
approach is consistent with the decision in Esau and Others v Minister of
Cooperative Governance and Traditional Affairs and Other s14 that policies and
decisions, which are often formulated based on findings , will not be ripe for
review until implemented and given external legal effect through the ir
publication, in this case, the promulgation of the regulation.
Non Joinder of the Competition Commission.
[26] This review intersects administrative law with competition law by virtue
of the applicable regulation framework to the subject of the review . I will deal
with this in due course.
[27] ICASA first contended that the non -joinder of the Competition
Commission (the Commission) to the proceedings was an impediment to the
14 Esau and Others v Minister of Co-Operative Governance and Traditional Affairs and Others [2021] ZASCA 9;
2021 (3) SA 593 (SCA) para 45, the court held that
‘As a general rule, policies that have been formulated and adopted by the executive will not be ripe for review
until they are implemented, usually after having been given legal effect by some or other legislative instrument.
Two principles come into play in this regard: first, that in order for an exercise of public power to be ripe for
review, it should ordinarily be final in effect; and secondly, that the decision must have some adverse effect for
the person who wishes to review it, because otherwise its setting-aside would be an academic exercise which
courts generally eschew.’
11
review. It is trite that a non-joinder is a question of law, which can be raised mero
motu by a court, but it need not detain the court in this instance.15
[28] The role of the Commission , whose relevance emerges later in the
judgment, is set out in section 4B (8) of the ICASA Act, where in the relevant
part it states that:
‘Before the exercise and performance of any of its powers and duties in terms of this section,
the Authority must –
(a) . . .
(b) subject to section 67 of the Electronic Communications Act and the terms and
conditions of any concurrent jurisdiction agreement concluded between the Authority and the
Competition Commission, bear in mind that the Competition Commission has primary
authority to detect and investigate past or current commissions of alleged prohibited practices
within any industry or sector and to review mergers within any industry or sector in terms of
the Competition Act.’
[29] It was common cause that the Commission participated in the market
inquiry and made submissions to ICASA. In my view, the provision serves to
prompt ICASA (as the telecommunications sector regulator) to co -operate and
consider the views of the Commission (as the Competition regulator) . The
provision does not render the Commission a necessary party to proceedings
brought against ICASA.
[30] Although the Commission as the competition regulator, may have an
interest in the outcome of the review, the nature of the orders sought have no
direct effect on the Commission if granted. 16 ICASA correctly abandoned the
non-joinder point. I now turn to the review grounds.
15 Snowy Owl Properties 284 (Pty) Ltd and Others v Mziki Share Block (Pty) Ltd [2024] ZASCA 79 para 25.
16 Judicial Service Commission and Another v Cape Bar Council and Another [2012] ZASCA 115; 2013 (1) SA
170 (SCA) para 12.
12
Review grounds
[31] MTN has launched extensive grounds for review premised on the findings
and PAJA. They can be categorised into the following themes. The first challenge
is directed at the Regulations 3(a), 3(b) and 3(c) defining the retail, site access
and roaming markets respectively. MTN contends that the market definitions are
irregular, unlawful, flawed and arbitrary:
(a) ICASA failed to properly apply the SSNIP Test and consider competitive
dynamics in the markets defined. The f indings document and the r easons
document show that ICASA disregarded MTN’s representations and it
impermissibly sought to rationalise the determination ex post facto . ICASA
departed from its own Guidelines in making the determinations.
(b) In so far as the site access market, the Regulation resulted from a flawed
definition of the site access market. ICASA excluded sites owned by non -
licensees, micro sites and did not apply a SSNIP test to define the market. It is
alleged that it failed to take account of relevant considerations and took account
of irrelevant ones. Similarly, with regards to the Regulation of the roaming
product market, the flawed definition resulted from confining the product market
to roaming for coverage and the exclusion of roaming for capacity.
[32] The second category of the challenge concerns ICASA’s finding that the
defined markets are ineffectively competitive. Although the factual basis for the
finding differs based on the market identified, the common complaint is that the
determinations are unlawful because ICASA failed to have regard to mandatory
requirements in s 67(4A) of the ECA, and failed to consider the dynamic character
and functioning of each of the retail market , site access and roaming product
markets defined, in particular the market shares and relative market power of the
participants. In making the determinations, ICASA failed to consider relevant
considerations and had regard to irrelevant considerations.
13
[33] The third theme concerns the finding that MTN has significant market
power in the markets defined. In terms of section 67(5)(a) of the ECA, a licensee
has significant market power if it is dominant within the meaning of section 7 of
the Competition Act. It submits that ICASA relied on outdated data to determine
MTN’s dominance in the retail market. It is alleged that ICASA’s determination
was not only influenced by a material error of law, but to the extent that it sought
to regulate the wholesale site access and roaming markets, it conflated the
meaning o f a vertical relationship with vertical integration in th e w holesale
markets identified. ICASA also failed to take account of relevant information and
relevant considerations in determining MTN’s market power in the site access
markets.
[34] As the f ourth basis, MTN submits that the pro-competitive conditions
imposed by ICASA are ultra vires and not for the purpose for which the power to
regulate the market was conferred.
[35] The fifth ground is premised on several provisions of s 6(2) of PAJA:
(a) The material aspect of the procedural complaint is that ICASA treated the
findings document as ‘final,’ even though the findings document made new
findings, which were not canvassed in the discussion document furnished to
MNOs.
(b) ICASA allegedly made determinations relative to MTN’s market power in
various markets without properly considering MTN’s submissions on the findings
document on the Mobile Broadband Services Inquiry published together with the
draft Regulations. 17 MTN contends it made submissions on several matters
affecting the findings, but ICASA failed to give a proper consideration to those
17 ‘Draft Mobile Broadband Services Regulations’ Pursuant to section 67(4) of the Electronic Communications
Act No. 36 of 2005, GN 272 in GG 44337 of 26 March 2021.
14
submissions.
[36] Lastly, MTN relies on procedural fairness as an over-arching, self-standing
ground to impugn the Regulations as a whole. The complaint i s that despite
several engagements with stakeholders,
(a) ICASA failed to provide MTN with the material information requested, for
MTN to make the necessary meaningful representations.
(b) It deprived MTN of the opportunity to make proper representations on the
findings document and failed to take MTN’s submissions into proper
consideration.
(c) Although ICASA made new findings pertaining to the definition of the
markets in the findings document, and promulgated the Regulations based on the
new findings, it treated the findings document as final, and at the hearing, urged
stakeholders to confine their representations to the draft regulations, but not the
new findings canvased in the findings document forming the basis for the
Regulations.18
Scope of the court’s review powers
[37] The review implicates the findings of facts and determinations made by
ICASA. The parties differ materially in conceptual approach influencing the
reach of the court’s review powers. It is convenient to dispose of this debate first
before deciding whether the regulations are susceptible to review on stated
grounds.
[38] As a starting point, ICASA contended that the court was not required to
‘test the correctness’ of the factual findings made to define and de termine the
issues challenged. Doing so would blur the line between an appeal and a review.
18 Ie the Mobile Broadband Services Regulations, 2021, GN 1960 in GG 46155 of 31 March 2022 (the
Regulations).
15
It relied on the decision in Dumani v Nair and Another 19 and submitted that if
ICASA was mistaken in its evaluation of the facts on the definition of the markets
and the assessment of competitive constraints, that mistake of fact would not be
sufficient grounds for review. A court can only interfere with the decision -
maker’s findings of fact if they are uncontested and verifiable.
[39] ICASA’s second contention is based on the principle of deference. It urged
the court to give due weight to ‘findings of fact and policy decisions made by
those with special expertise and experience in the field’. That submission is based
on MEC for Environmental Affairs and Development Planning v Clairison’ s
CC,20 that once the law entrusted ICASA with a discretion to define the markets,
the weight or lack of weight to be attached to the various considerations making
up its determination, is part of a discretion given to it as the decision -maker. As
such, t he court should pay d ue respect to the decisions and route selected by
ICASA in making the determinations.
[40] It added that this approach would be consistent with the ‘executive role’
because ICASA is ‘part of the executive’. Counsel relied also on the approach in
Justice Alliance of South Africa v Mncube NO and Others and 2 related matters.21
It is contended that ICASA acted rationally and reasonably in making the
regulations. It is settled law that ‘rationality’ is about whether there is a rational
connection between the decision taken and the purpose for which the power
exercised is given. 22 Reasonableness on the other hand posits that a decision is
19 Dumani v Nair and Another [2012] ZASCA 196; [2013] 2 All SA 125 (SCA) (Dumani) paras 22 and 32 and 33.
20 MEC for Environmental Affairs and Development Planning v Clairison's CC [2013] ZASCA 82; 2013 (6) SA
235 (SCA) (Clairisons) para 27-28
21 Justice Alliance of South Africa v Mncube NO and Others and 2 related matters [2015] 1 All SA 181 (WCC)
(Justice Alliance ) para 30. It bears noting however that in Justice Alliance, a review p remised on several
procedural defects, including a failure to grant interested parties a meaningful opportunity to make representations
and consider relevant considerations and considerations of fairness, the court granted the review. There ICASA
committed a material error of law. The licenses challenged were issued based on an incorrect legal assumption
inconsistent with it the prevailing legislation
22 Albutt v Centre for the Study of Violence and Reconciliation, and Others [2010] ZACC 4; 2010 (3) SA 293 (CC)
paras 51 and 72.
16
reviewable if it is one a reasonable decision -maker would not reach.23 MTN’s
complaint can best be resolved through an appeal.
[41] I accept that the findings may not be interfered with simply because a court
considers the decision incorrect.24 The conundrum of demarcating the scope of a
review of a mistake or errors or finding of fact is more complex and nuanced than
argued by ICASA. It must be considered in t he context of the evolution of
administrative law, the complex interplay between those facts and the grounds on
which the review is brought. Although not pertinently referred to during argument,
the complexity is illustrated by Unterhalter J (as he then was), in Airports
Company South Africa v Tswelokgotso Trading Enterprises CC25( ACSA) dealing
with a legality review.
[42] The court’s interpretation of Dumani in ACSA26 suggests an expanded
ambit than advanced on behalf of ICASA when questions of legality are at stake
and as it states that:
‘Dumani, in my view, enlarges the ambit of review on the grounds of mistake of fact. Even
where the functionary enjoys competence as the finder of fact, an error made as to a material
fact that was established, in the requisite sense, is reviewable. The qualifica tion
in Pepcor ousted review for mistake of fact, even if material, where the functionary enjoyed
the power to determine the relevant facts. In Dumani a functionary, though empowered as a
finder of fact, who renders a decision mistaken as to a material fac t which was established as
uncontentious and objectively verifiable, has made a reviewable error.’
[43] I interpose to state that the principle of deference iterated in the Constitutional
Court’s decision in Bato Star Fishing (Pty) Ltd v Minister of Environmental Affairs
23 Bato Star Fishing (Pty) Ltd v Minister of Environmental Affairs and Tourism and Others [2004] ZACC 15; 2004
(4) SA 490 (CC) (Bato Star) para 44.
(4) SA 490 (CC) (Bato Star) para 44.
24 Merafong Demarcation Forum and Others v President of the Republic of South Africa and Others [2008] ZACC
10; 2008 (5) SA 171 (CC) para 63.
25 Airports Company South Africa v Tswelokgotso Trading Enterprises CC 2019 (1) SA 204 (GJ) (ACSA) paras 5
-12.
26 ACSA para 11.
17
and Others27is not unqualified. The court in Bato Star went further than stated during
argument to underscore that the loadstar is the character of the decision challenged,
stating that:28
‘The extent to which a Court should give weight to these considerations will depend upon the
character of the decision itself, as well as on the identity of the decision-maker. A decision that
requires an equilibrium to be struck between a range of competing interests or considerations
and which is to be taken by a person or institution with specific expertise in that area must be
shown respect by the Courts.’ (My emphasis.)
[44] Despite the opinion in ACSA, which must be viewed in the context of the
case before it, the court in Pepcor Retirement Fund and Another v Financial
Services Board and Another29 went further and demarcated two types of mistakes
of fact an ‘ordinary mistakes of fact’ and mistake of facts going to jurisdiction,
and stated that:
‘[46] Nevertheless it is relevant to note in passing that section 6(2)( e)(iii) provides that a
court has the power to review an administrative action inter alia if “relevant considerations
were not considered”. It is possible for that section to be interpreted as restating the existing
common law; it is equally possible for the section to bear the extended meaning that material
mistake of fact renders a decision reviewable.’
. . .
[48] Of course, these limitations upon a reviewing court's power do not extend to what have
come to be known as jurisdictional facts and, in my view, it will continue to be both necessary
and desirable to maintain that particular category of fact.’ (Footnote omitted.)
[45] ICASA is constrained to exercise its power and perform its function within
the borders of its founding legislation and the relevant statutes. It must act in the
27 Bato Star para 46 reads as follows:
‘The use of the word “deference” may give rise to misunderstanding as to the true function of a review Court.
This can be avoided if it is realised that the need for Courts to treat decision -makers with appropriate deference
or respect flows not from judic ial courtesy or etiquette but from the fundamental constitutional principle of the
separation of powers itself.’ (Footnote omitted.)
28 Bato Star para 48.
29 Pepcor Retirement Fund and Another v Financial Services Board and Another 2003 (6) SA 38 (SCA) (Pepcor).
18
manner consistent with the law.30 Its submission is first and foremost tempered by
s 3(5) of the ICASA Act, which permits an unqualified review of any findings made
by it.
[46] The principle of d eference is also tempered by Pharmaceutical
Manufacturers Association of SA and Another: In re Ex parte President of the
Republic of South Africa and Others31 here the Court held that:
‘[85] It is a requirement of the rule of law that the exercise of public power by the Executive
and other functionaries should not be arbitrary . Decisions must be rationally related to the
purpose for which the power was given, otherwise they are in effect arbitrary and inconsistent
with this requirement. It follows that in order to pass constitutional scrutiny the exercise of
public power by the Executive and other functionaries must, at least, comply with this
requirement. If it does not, it falls short of the standards demanded by our Constitution for such
action.
[86] The question whether a decision is rationally related to the purpose for which the power
was given calls for an objective enquiry. Otherwise a decision that, viewed objectively, is in
fact irrational, might pass muster simply because the person who took it mistakenly and in good
faith believed it to be rational. Such a conclusion would place form above substance and
undermine an important constitutional principle.’ (Footnotes omitted).
[47] The following principles emerge: First, a functionary cannot render a
proper decision made in ignorance of material facts. A mistake of fact has been
recognised as a ground of review in terms of PAJA.32 Second, the qualification or
limitation of the ambit of the review power (i.e. the discretionary a point made by
ICASA) is that the functionary must enjoy the power to establish the relevant
facts. Third, the reviewable mistake is confined to uncontentious and objectively
30 Fedsure Life Assurance Ltd and Others v Greater Johannesburg Transitional Metropolitan Council and Others
[1998] ZACC 17; 1999 (1) SA 374 (CC) para 27 states that:
‘Laws are frequently made by functionaries in whom the power to do so has been vested by a
competent legislature. Although the result of the action taken in such circumstances may be “legislation”, the
process by which the legislation is made is in substance “administrative”.’
31 Pharmaceutical Manufacturers Association of SA and Another: In re Ex parte President of the Republic of South
Africa and Others [2000] ZACC 1; 2000(2) SA 674 (CC) (Pharmaceutical Manufacturers).
32 Pepcor paras 47 and 48.
19
verifiable facts. Fourth, that mistake of fact does not apply to the category of
jurisdictional facts which are reviewable. Fifth, an administrative action may be
interfered with if it is arbitrary, irregular and /or where irrelevant factors are
considered, and when relevant considerations are excluded. Sixth, the decisions
must be rationally related to the purpose for which the power was given. I find
that ICASA’s market definition and determinations are reviewable based on any
of the above principles.
[48] It bears emphasising at this stage that the context of the present review does
not flow from an investigation of past or current commission of prohibited or anti-
competitive behaviour or a decision about contraventions by MNOs. It flows
from the statutory power to regulate market and remedy an established market
failure. Against the backdrop of the review principles above, it is appropriate to
consider the applicable framework governing ICASA’s market regulation task,
thereafter, consider whether ICASA has performed the function lawfully within
the powers conferred and for the purpose for which the power was entrusted.
Statutory and competition regulation of the mobile telecommunications
market
Electronic Communications Act
[49] The ECA defines three electronic communications areas with implications
on market participant s subject to ICASA’s regulatory authority , n amely,
‘electronic communications network service ’, an ‘electronic communications
network’ and ‘electronic communication facility’.
[50] ECA defines these three terms as follows:
‘“electronic communications facility” includes but is not limited to any-
(a) wire, including wiring in multi-tenant buildings;
(b) cable (including undersea and land-based fibre optic cables);
20
(c) antenna;
(d) mast;
. . .
or other thing, which can be used for, or in connection with, electronic communications,
including, where applicable-
(i) collocation space;
(ii) . . .
(iii) space on or within poles, ducts, cable trays, manholes, hand holds and conduits
. . .
“electronic communications network ” means any system of electronic communications
facilities (excluding subscriber equipment
. . .
“electronic communications network service ” means a service whereby a person makes
available an electronic communications network’.
The relevance of the above definitions is material to who or what it subject to
licensing and ICASA’s regulatory regime.
[51] The impugned Regulations were published pursuant to s 67(4) of the ECA.
Sections 67(4)(a) to (d) mandates ICASA in the following manner:
‘The Authority must, following an inquiry, prescribe regulations defining the relevant markets
and market segments and impose appropriate and sufficient pro-competitive licence conditions
on licensees where there is ineffective competition, and if any licensee has significant market
power in such markets or market segments. The regulations must, among other things-
(a) define relevant wholesale and retail markets or market segments;
(b) determine whether there is effective competition in those relevant markets and market
segments;
(c) determine which, if any, licensees have significant market power in those markets and
market segments where there is ineffective competition;
(d) impose appropriate pro -competitive licence conditions on those licensees having
significant market power to remedy the market failure’. (My emphasis.)
[52] A definition of what the ‘relevant market s’ are, is imperative and a
precondition to the market regulating power in s 67(4)(d). Although the provision
21
lists discrete determinations to be made, the recurring reference to ‘those markets’
in s 67(4) (b) and (c) dealing with effective competition and significant market
power respectively is significant and is dependent upon a proper definition of the
marker or market se gment. There is a sequential and cumulative connection
between the definition of a market or market segment and the determinations of
their effectiveness and or significant market power.
[53] To determine whether the identified markets have ineffective competition,
s 67(4A) of the ECA stipulates a list of non-exhaustive factors which ICASA must
consider, and states that:
‘When determining whether there is effective competition in markets and market segments, the
Authority must consider, among other things-
(a) the non-transitory (structural, legal, and regulatory) entry barriers to the applicable
markets or market segments; and
(b) the dynamic character and functioning of the markets or market segments, including
an assessment of relative market share of the various licensees or providers of exempt services
in the markets or market segments, and a forward looking assessment of the relative market
power of the licensees in the markets or market segments’ (My emphasis.)
[54] It is mandatory that ICASA first considers the factors listed in the section
to determine competition effectiveness. 33 Any suggestion of a discretion on
factors ICASA can considered in determining the effectiveness of competition
cannot be to the exclusion of the mandatory factors listed in s 67(4A). Hence,
‘amongst others’ signifies that any further factors considered are in addition to
those mandatory factors listed in the section. Even then, those other factors must
be clearly stated.
33 Minister of Environmental Affairs and Tourism and Others v Pepper Bay Fishing (Pty) Ltd; Minister of
Environmental Affairs and Tourism and Others v Smith 2004 (1) SA 308 (SCA) para 32.
22
[55] Whether MTN has the requisite significant market power within the market
in terms of s 67(4)(c), is determined with reference to s 67(5) of the ECA, which
states that:
‘A licensee has significant market power in a market or market segment if that licensee-
(a) is dominant;
(b) has control of an essential facility; or
(c) has a vertical relationship that the Authority determines could harm competition.’
As will be seen below, the question of dominance derives from competition law
jurisprudence. Here, “significant market power” includes dominance and either
of the additional factors and thus presents ICASA a higher bar than mere market
share and or market power.
[56] MTN and other MNOs are obliged to provide ICASA with information
specified by it in carrying out its duties under s 67(4B) of the ECA which states
that:
‘Subject to section 4D of the ICASA Act, licensees must provide to the Authority any
information specified by the Authority in order that the Authority may carry out its duties in
terms of this section.’
[57] The bounds for imposing pro-competitive conditions are set out in s 67(7)
of the ECA, which provides that:
‘Pro-competitive licence terms and conditions may include but are not limited to-
(a) obligations in respect of interconnection and facilities leasing in addition to those
provided for in Chapters 7 and 8 and any regulations made in terms thereof;
(b) penalties for failure to abide by the pro-competitive licence conditions;
(c) obligations to publish any information specified by the Authority in the manner
specified by it;
(d) obligations to maintain separate accounting for any services specified by the Authority;
(e) obligations to maintain structural separation for the provision of any services specified
by the Authority;
23
(f) rate regulation for the provision of specified services, including without limitation price
controls on wholesale and retail rates as determined by the Authority, and matters relating to
the recovery of costs;
(g) obligations relating to accounts, records and other documents to be kept, provided to
the Authority, and published;
(h) obligations concerning the amount and type of premium, sports and South African
programming for broadcasting; and
(i) distribution, access and reselling obligations for broadcasters.’
The Competition Act
[58] As alluded to earlier, ICASA’s competition regulatory remit in the
provisions of the ECA, which regulates the ICT sectors falling under it, intersects
with the Competition Act , which regulates competition across all industries
(together they are referred to as the underlying statutes). Sections 67(4), 67(4A)
and 67(5) of the ECA , as set out above, deal with market definition and
competition regulation. However, the ECA does not define the meaning of the
terms employed in these sections.
[59] It does so through s 1 of the ECA, which serves as a bridge between those
terms employed in s 67(4) and the meaning ascribed to them in the Competition
Act. Section 1 of the ECA states that ‘dominance’, ‘market power’, and ‘vertical
relationship’ bear the same meaning as defined in sections 1 and 7 of the
Competition Act.
[60] Section 1 of the Competition Act provides that:
‘“market power” means the power of a firm to control prices, to exclude competition or to
behave to an appreciable extent independently of its competitors, customers or suppliers;
. . .
‘“vertical relationship” means the relationship between a firm and its suppliers, its customers
or both’.
24
Section 7 sets out the definition of dominance by setting out various percentages
of market share as follows:
‘A firm is dominant in a market if-
(a) it has at least 45% of that market;
(b) it has at least 35%, but less than 45%, of that market, unless it can show that it does not
have market power; or
(c) it has less than 35% of that market, but has market power.’
[61] It is convenient, at this juncture, to also refer to additional terms employed
by ICASA in its answering affidavit, the findings document, and the regulations
themselves, which have a bearing o n the review, namely the meaning to be
ascribed to a ‘vertical relationship’ and ‘vertical integration’. ICASA contended
in its answering affidavit that it employed these terms interchangeably.
[62] The relevance of the phrase ‘vertical relationship’ in s 67(5)(c) of the ECA
is to assess whether a licensee like MTN has significant market power within a
market found to be ineffectively competitive. The only section in the Competition
Act in which the term ‘vertical relationship’ is employed is in s 5(1) ,34 which
refers to ‘an agreement between parties in a vertical relationship’.
[63] On the other hand, ‘vertical integration’ is employed in s 12A(2)(f)35 of the
Competition Act, as a part of various considerations that must be measured before
a merger of two companies may be approved . Hence, the drafters distinguished
between the concepts of a ‘vertical relationship’ and ‘vertical integration’. Had
they intended the former concept to include a vertically integrated firm, they
would have said so.
34 Section 5(1) of the Competition Act reads:
‘An agreement between parties in a vertical relationship is prohibited if it has the effect of substantially preventing
or lessening competition in a market, unless a party to the agreement can prove that any technological, efficiency
or other procompetitive, gain resulting from that agreement outweighs that effect.’
or other procompetitive, gain resulting from that agreement outweighs that effect.’
35 Section 12A(2)(f) provides that ‘the nature and extent of vertical integration in the market’.
25
[64] In sum, h aving regard to the review principles and the regulatory
framework above, ICASA is institutionally entrusted with the function to regulate
the telecommunications sector, textually, the provisions of the ECA reinforce that
the use of the language and the evaluation of ICASA’s determinations are to be
sourced from the Competition Act. Section 67(9) of the ECA makes it clear that
subject to its provisions, the Competition Act applies to competition matters in
the electronic communications industry. The inescapable conclusion is that the
determinations must be congruent with accepted competition law principles.
[65] It is necessary to refer to the ICASA Guidelines published in March 201036
before evaluating the regulations . The guidelines explain how ICASA would
implement the provisions of the ECA:
(a) By entrenching the participation of stakeholders and operators when
making regulations.
(b) The guidelines provide that ICASA
‘may use the Hypothetical Monopolist Test, including the Small (but) Significant Non-
transitory Increase in Price (SSNIP) test, as well as other alternatives, including the
examination of ‘practical indicia’, during the process of defining a market.’
(c) The guidelines confirm the interplay between the (i) demand-side and (ii)
supply-side, which, when considering market definition , includes geographic
availability.
(d) It looks at how easily it is to substitute a product (ie demand and supply),
the barriers of entry (both the structural and legal) all of which form part of the
assessment of the market’s dynamic character and how it functions.
36 A guideline for conducting market reviews, ICASA, 8 March 2010. https://www.icasa.org.za/uploads/files/
Guideline-for-Conducting-Market-Reviews.pdf (accessed 21 August 2025).
26
Did ICASA define the relevant markets lawfully within the bounds of the
applicable principles and legislation?
[66] I now turn to the review of the market definition s in the Regulations,
whether the defined markets are ineffectively competitive, and whether MTN has
significant market power in respect of those markets, commencing with the retail
market.
Retail market
[67] MTN contends that regulation 3(a) of the Regulation, which defines the
retail markets is arbitrary within the meaning of s 6(2) (e)(vi) of PAJA, ie that
ICASA has breached s 6(2)(e)(iii) of PAJA.37 ICASA failed to consider materially
relevant considerations and took account of irrelevant considerations when
drafting the Regulation 3(a). MTN submits that t he Regulation is not rationally
connected to the purpose for which it was made and the empowering provision in
the ECA or the information placed before ICASA within the meaning of s
6(2)(f)(ii)38 of PAJA.
[68] MTN’s grievance is that sub -national markets (provincial, segmented to
rural and urban) are wholly inconsistent with the competitive dynamics between
MNOs as they compete on a national basis. Its subscribers access mobile services
across the country ( in the form of national coverage). Many of the elements
necessary to provide mobile services, for example RAN, are shared across the
37 The relevant provisions of s 62(e) reads that
‘the action was taken—
. . .
(iii) because irrelevant considerations were taken into account or relevant considerations were not
considered;
. . .
(vi) arbitrarily or capriciously’.
38 “The action itself is not (ii) is not rationally connected to—
(aa)the purpose for which it was taken;
(bb)The purpose of the empowering provision;
(cc)the information before the administrator; or
(dd)the reasons given for it by the administrator”
27
provision of mobile services to other consumers, in other areas. Competitive
constraints manifest at a national level. Consumers emulate the mobile nature of
the service and move from one region to another. It will be recalled that parties
agreed that the product market combines SMS, voice and data.
[69] ICASA determined that there was as a single retail product market for
telecommunication mobile services encompassing (a) voice, (b) SMS and (c) data
services, made of subscribers for mobile services . The parties agree to the
identified product market. ICASA identified the geographic retail market as
follows: t he discussion document issued to MTN, and other stakeholders,
narrowly defined the geographic markets as pertaining only to local and metro
municipalities. It thus regarded each of the 234 municipalities in South Africa as
a separate geographical retail market for mobile telecommunication services .
However, in the findings document and the reasons document it broadened the
definition of retail market for mobile services to include regional and provincial
geographic areas segmented into rural and urban areas.
[70] The Hypothetical Monopolist Test, also referred to as a SSNIP Test, is used
by competition authorities to define geographic or product markets. ICASA
referred to the same test in Regulation 4 , which s tates the methodology to be
applied. The test presupposes a hypothetical monopolist who controls all supplies
and products in a particular candidate market but has no such control outside that
market. Generally, competition authorities will pose the question along the
following lines:39 ‘whether a hypothetical monopolist will be able to profitably
impose a small but significant non-transitory increase in price above competitive
levels (a SSNIP), typically in the order of 5-10%?’ Put differently, in ‘contrast, if
a 5-10% price increase would be unprofitable, then the candidate market should
a 5-10% price increase would be unprofitable, then the candidate market should
39 S Bishop and M Walker The Economics of EC Competition Law: Concepts, Application and Measurement
(2010) at 111-115, and 505 as cited in MTN’s submissions, dated 27 February 2020.
28
be widened to include additional products and/or geographies that were previous
excluded.’40
[71] Correctly applied to the defined markets , the SSNIP test is repeated
iteratively until a price increase would be profitable, at which point the product
and geographic dimensions of the candidate market over which the price increase
has been evaluated would be a relevant market for anti-trust purposes.
[72] The SSNIP test means that ICASA would have evaluated and explained
why a monopolist in rural Free State or rural Eastern Cape ( being one of the
provincial retail markets in which MTN was found to have significant market
power) would profitably increase prices in the region by 5 to 10%, but it would
not be profitable to increase prices in a larger area.
[73] ICASA’s stance was that it was not obliged to apply the SSNIP test. It
claims it determined the geographic regional market by ‘aggregating geographies’
with similar competitive conditions, considered regional pricing behaviour and
localised marketing strategies, based on submission s from market MNOs and
stakeholders. ICASA emphasised regional pricing by MTN and V odacom to
justify the regional geographic market definition. The justification for the finding
was that ‘stakeholders not only indicated differences in costs between rural and
urban areas, but stakeholders also indicated that there are regional differences in
mobile operator management pricing and investment decisions’.
[74] Accordingly, it defined the geographic market s after considering internal
documents submitted by V odacom and MTN that demonstrated:
(a) cost and price differences between rural and urban areas;
40 S Bishop and M Walker The Economics of EC Competition Law: Concepts, Application and Measurement
(2010) at 111-115, and 505 as cited in MTN’s submissions, dated 27 February 2020.
29
(b) regional differences in mobile operator management pricing; and
(c) investment decisions.
MTN’s submission supported that there was regional competition based on the
quality of the network albeit on a limited basis. V odacom, MTN, Telkom and Cell
C offered regional pricing. In addition, V odacom’s monthly net promoter score
(NPS)41 supported the regional market definition and contend ed that V odacom
supported the geographic market as defined. In its answ ering affidavit , it
explained that MTN informed ICASA that:42
‘more recently, it responded to V odacom's region-specific pricing by introducing “My Town”
offers allowing MTN customers to purchase specific data bundles offered in specific towns,
which provided direct and concrete evidence of operators swiftly responding to changes in
prices in different areas.’
MTN had put additional price pressure on the market with its ‘My Town’ offer
and offered discounted data bundles based on the location of the customer.
[75] ICASA did not deny that it did not consider the supply-side substitution at
market definition stage but did so when determin ing competitive effectiveness.
In its view, there is ‘no binding rule of law that obliged it to do so at the market
definition stage’. It contends, the practice of the competition authorities in South
Africa is to consider supply -side substitution at effectiveness of competition
assessment phase rather than the market definition stage, or to determine whether
any firms have market power, to avoid defining overly broad markets, (referred
to as overinclusion) which was reasonable in the circumstances. The approach
can be altered on a case-by-case basis.43
41 It is based on consumer survey and data which indicated customer willingness to switch or recommend products
of a company.
42 Findings document.
43 Caxton and CTP Publishes and Printers Ltd v Competition Commission and Others [2011] ZACT 54; [2011] 2
CPLR 304 (CT) (Caxton).
30
[76] ICASA claims it relied on the European Commission’s recent practice and
contended ‘a mere omission to apply the SSNIP test in aggregating municipalities
into provinces, segmented by rural and urban on its own does not render the
determination irregular’. It may use the SSNIP test including an examination of
other ‘practical indicia’ to define a market. It claims th e information before it
provided a rational basis for its approach.
[77] MTN contended that ICASA’s reliance on regional pricing to support its
regional market definition conflates the inquiry expected of it. That view is based
on leading authors Bishop and Walker44 who state that:
‘The use of differences in absolute price levels stems from a confusion between defining a
market as an area in which the “law of one price ” holds — the so-called “economic” market
— and defining the relevant market, which takes into account the competitive constraints
between products and regions. The relevant market can coincide with, be wider than or be
narrower than the economic market.’
[78] It will be recalled that ICASA and the parties accepted that the product
market includes voice, data and SMS. ICASA did not gainsay MTN’s contention
that different pricing would be subject to different usage by the customers and
will in most instances differ even within the same geographic region. That the
MNOs offered different regional pricing is an unreliable means of determining
the geographic retail market.
[79] ICASA’s stance about the failure to apply the SSNIP test conflicts with the
Competition Court of Appeal’s decision in Medicross Healthcare Group (Pty) Ltd
v Competition Commission (Medicross),45 which endorsed the requirement for a
proper definition of the relevant market. It held that conclusions on what exerts
44 S Bishop and M Walker The Economics of EC Competition Law: Concepts, Application and Measurement
(2010) at at 139.
(2010) at at 139.
45 Medicross Healthcare Group (Pty) Ltd and Another v Prime Cure Holdings (Pty) Ltd [2006] ZACAC 3; [2006]
1 CPLR 1 (CAC) (Medicross) para 25.
31
competitive constraints in any market ‘requires a full market analysis’. Contrary
to ICASA’s submission, Medicross46 endorsed the SSNIP test as a mechanism to
define the market and had this to say:
‘In short, a market does not define itself. It is determined by the demand side substitution and
the supply side substitution of customers and producers respectively. In so assessing the scope
of the referral market, there is a need to enquire as to the extent to which customers may switch
to alternate products in the event of a price increase and the ability of competing producers to
other alternate product offerings employed by the Tribunal in arriving at its finding. Hence
there is a need to enquire into the evidence employed by the Tribunal in arriving at its finding.’
[80] ICASA’s view that it was not obliged to supply -side substitution and /or
apply SSNIP test at market definition level also conflicts with Medicross, which
endorsed that the supply -side substitution is integral to market determination to
gauge a market response to the hypothetical monopolist price increase, product
substitutability and the geographic boundary beyond which the hypothetical
monopolist may or ma y not act profitably. It is an inherent consideration of a
market definition and application of the SSNIP test.
[81] On the information before ICASA , it demonstrated that capacity and
coverage are linked inextricably to availability of RAN and form a part of
competitive dynamics of the national mobile telecommunications retail market.
MTN’s uncontested submission was that the underlying infrastructure for
delivery of the mobile services is contiguous , with many of its elements shared.
It contended that the adjacent regions may be linked by a ‘chain of substitution’
resulting in an overlap in areas served by RAN towers. A RAN tower located near
the border of one region would be a substitute for a RAN tower in the adjacent
region. Adjacent RAN towers impose (direct) competitive constraints on each
region. Adjacent RAN towers impose (direct) competitive constraints on each
46 Medicross para 30.
32
other and, in turn, impose indirect competitive constraints on non -adjacent
towers.
[82] The thrust of this information, which was not in dispute is two-fold, namely
that first, subscribers who contract with MNOs do so based on capacity and
availability of national coverage. Second, the competitive constraints posed by
the underlying infrastructure, like RAN, are materially relevant considerations for
determining the geographic market. The above factors distinguish the judgment
in Competition Commission of South Africa v Mediclinic Southern Africa (Pty)
Ltd and Another (Mediclinic)47 from the present case, which ICASA suggested I
should consider since it related to regional geographic market.
[83] ICASA’s last word and justification for its market definition is the findings
document and reasons document. They are intended to provide its reasons for the
decision. The requirement is that they must be based on correct facts, be rationally
connected to the information before ICASA and take account all relevant factors
before it.48
[84] Although in the discussion document ICASA identified narrow municipal
geographic markets, ICASA accepted it did not apply the SSNIP test to the regional
markets and or the provinces split by urban and rural it had identified. While the
failure to apply the SSNIP test in the discussion document can be understood in the
context of the discussions with stakeholders, the difficulty is that the findings
document and the reasons document does not provide a justification for concluding
that the retail markets are regional/provincial split to urban and rural.
47 Competition Commission of South Africa v Mediclinic Southern Africa (Pty) Ltd and Another [2021] ZACC
35; 2022 (4) SA 323 (CC) (Mediclinic)
48 Pepcor paras 47 and 48.
33
[85] ICASA’s findings document or reasons document do not provide reasons
or explanation s for its assumption that retail activities in one province exert
competitive constraints in another province , but that retail activities beyond
provincial boundaries do not. The findings document, the reasons document and
answering affidavit do not provide an objective explanation to the critical
question why a monopolist would profitably increase prices in the regions
identified by 5 to 10%, but it would not be profitable to increase prices in a larger
area. The geographic dimensions of the market over which the price increase has
been evaluated is not defined. ICASA’s findings document and reasons document
is silent on the consequence of the overlap in RAN services to justify its regional
market determination.
[86] The court cannot overlook that in answer to the inherent mobile nature and
use of mobile services, ICASA relied on a survey by the Council for Scientific
and Industrial Research Gauteng Household Travel as ‘evidence’ which
‘suggests’ that most travel (90%) undertaken in the Gauteng Province is within
municipalities. Other than this limited geographic region, the findings document
is silent on whether the same data applies to the other regional markets to justify
the provincial/regional definition.
[87] As the court in National Lotteries Board v South African Education and
Environment Project49 held, reasons for the decision ‘cannot be remedied by
giving different reasons after the fact’. The court proceeded to state that:50
‘The duty to give reasons for an administrative decision is a central element of the
constitutional duty to act fairly. And the failure to give reasons, which includes proper or
adequate reasons, should ordinarily render the disputed decision reviewable. In England the
courts have said that such a decision would ordinarily be void and cannot be validated by
courts have said that such a decision would ordinarily be void and cannot be validated by
49 National Lotteries Board and Others v South African Education and Environment Project [2011] ZASCA 154;
2012 (4) SA 504 (SCA) (National Lotteries Board) para28.
50 National Lotteries Board para 27.
34
different reasons given afterwards — even if they show that the original decision may have
been justified. For in truth the later reasons are not the true reasons for the decision, but rather
an ex post facto rationalisation of a bad decision.’
[88] It is not immaterial that ICASA’s retail market definition was at odds with
the submissions made by the Commission, stating that:
‘The Commission disagrees with the definition of local geographic markets rather than
national, as competition dynamics are clearly national. Furthermore, defining geographic
markets using municipal boundaries is arbitrary and not properly justified by the discussion
document. This raises concerns for the analysis and the conclusions reached.’
The Commission also observed that the regulatory remedy employed by ICASA,
requiring disclosure of pricing information across the country was at odds with
ICASA’s finding of sub-national markets, a matter that has relevance later in the
judgment.
[89] While acknowledging the differences in approach, ICASA misconstrued
the legal importance of the Commission’s submission, contending that it did not
mean that ICASA’s determination of a local market ‘is unreasonable’. Tellingly,
it sought to provide ‘a rational basis’ for the differences on the grounds that the
Commission may not have had the extensive data and documentary evidence on
geographic markets that ICASA had, nor did it carry out the detailed analysis on
prices, usage, and costs that ICASA had carried out.
[90] Although the issue fell within ICASA’s jurisdiction, the disagreement by
the Commission, which is well versed in competition economics and competition
law, warranted a proper analysis of the data and information and a cogent
justification, none of which ICASA provides.
[91] Significantly, ICASA’s submissions contradicts its own methodology when
it drafted regulation 4. The relevant part of the regulation states that:
35
‘In determining the effectiveness of competition in the markets defined in regulation 3 above,
the Authority applied the following methodology:
(a) the identification of relevant markets and their definition according to the principles of
the Hypothetical Monopolist Test, taking into account the non -transitory (structural, legal, or
regulatory) entry barriers to the relevant markets and the dynamic character and functioning of
the relevant markets.’
[92] Although the above Regulation deals with the effectiveness of competition
in the defined markets, it refers to the methodology applied to define th ose
markets. It states that ICASA applied the SSNIP test when determining what the
definition of the market to be used whilst drafting the Regulations, when the facts
before the court, and ICASA’s version, point to the contrary.
[93] Lastly, ICASA developed the Guidelines to give guidance on how it would
interpret and apply the ECA. Instead, it elected to follow recommendations by
the European Commission. ICASA’s stance was that it is not bound by ‘a mere
guideline’ which may be applied flexibly or that it had a discretion on how to
decide the markets , is not sustainable . The importance of guidelines has been
considered by the court s. The guidelines provide operators with regulatory
certainty, and it cannot ignore them at will. As the Supreme Court of Appeal stated
in CTP Limited and Others v Director-General Department of Basic Education
and Others:51
‘Regarding the deviation from the implementation guide, it has to be noted that this guide is
not legislation but a policy. The objects of a policy are to achieve reasonable and consistent
decision making, to provide a guide and a measure of certainty to the public. It is trite that when
the government makes a policy, its officials are not entitled to simply ignore it, but must act in
accordance with it. They can only deviate from it if there is a reasonable basis for such deviation
in which case that basis should be clearly articulated.’
51 CTP Limited and Others v Director -General Department of Basic Education and Others [2018] ZASCA 156
para 30.
36
[94] As I have endeavoured to show, the exercise of ICASA’s regulatory power
is circumscribed by the provision of the ECA. The nature of the fact s sought to
be established are jurisdictional requirements , the absence of which will be
considered irregular and open to review.
[95] The justification for determining that the retail markets are provincial and
‘split by urban and rural ’ is lacking in the ‘full market analysis’ as propounded
in Medicross.52 The ‘set of products or services which exert competitive
constraints on one another ’ in the regional mobile retail markets which ICASA
identified are not discernible from the findings document or the reasons
document. The real-world dynamics of the mobile retail market that ICASA was
tasked to define and the prospect that MNOs would respond to a price increase in
a particular geographic area by offering competitive services was not undertaken.
An approach which fails to consider both the demand -side and supply -side
substitution in defining the market for retail mobile services is flawed and not one
contemplated by the ECA and Medicross.
[96] ICASA’s reliance on a survey regarding the movement of consumers in
Gauteng to determine regional markets elsewhere drives home the complaint that
its determination was arbitrary. Findings were made for the first time in the
findings document, having not been reached previously – or even canvassed – in
the discussion document. MTN was not given an opportunity to comment on
ICASA’s findings that the market was regional. The justification does not form a
part of the record of its decision, ie the findings document or reasons document.
It is impermissible for ICASA to ex post facto justify its regional market
determination on ‘My Town offers’ made by MTN in its answering affidavit.
52 Medicross para 25.
37
[97] As MTN contends the irregularity occurred at the following levels:
(a) The findings document and reasons document do not show reliance on
MTN documents and constitute an impermissible ex post facto rationalization of
the findings
(b) ICASA failed to apply the SSNIP test;
(c) In determining the geographic dimension of the retail market, ICASA
expressly declined to have regard to the supply-side substitution in the geographic
market and whether the likelihood of supply -side substitution would be ‘timely,
likely and sufficient to have an impact on the definition of [the] market’, contrary
to its own guidelines on market.
[98] ICASA’s determination in regulation 3 (a) that retail markets are regional
for services consisting of provinces, split by urban and rural, is flawed, arbitrary,
and unlawful . It failed to consider relevant consideration . It s determination
process conflicted with its own guidelines without cogent explanation for the
deviation. The regulation falls to the reviewed and set aside.
Ineffectively competitive retail market
[99] MTN challenges regulation 5, which states that:
‘Pursuant to regulation 4, the Authority has determined that competition in the Retail market,
Upstream market 1, Upstream market 2, and Upstream market 3b, as defined in regulation 3,
are ineffectively competitive.’
It submits that regulation 5 should be set aside because its determination that the
retail market, the site access market and the roaming market are ineffectively
competitive is based on an unlawful the market definitions. In the case of the
retail and site access markets, ICASA did not decide the ineffective competition
in each of the markets it had identified:
38
(a) It did not consider several mandatory considerations listed in s 67(4A) of
the ECA.
(b) The determinations that the retail, site access and roaming markets are
ineffectively competitive was unreasonable and irrational, as ICASA failed to
have regard to various relevant considerations and the determinations are based
on irrelevant considerations in breach of s 6(2)(e)(iii) of PAJA.(c) The
determination that the markets are ineffectively competitive was procedurally
unfair and tainted by the failure to consider MTN’s submissions on the findings
document. It violated the principle of legality in s 1(c) of the Constitution because
the findings of ineffectively competitive markets contravened s 67(4A) of the
ECA and were not rationally connected with the power conferred . ICASA also
failed to allow meaningful representations on the contents of the findings
document.
[100] The challenge implicates Regulation 4, dealing with the SSNIP test, the
market definition methodology already discussed and purported to have been
applied. The rest of it requires that ICASA conducts:
‘(b) the assessment of licensees' market shares in the relevant markets; and
(c) the assessment on a forward-looking basis of the level of competition and market power
in the relevant markets.’
[101] I am of the view that the finding that the retail market definition is flawed,
unlawful and tainted by an irregularity has a consequential effect on the
determination of effectiveness of the competition as provided in regulation 5. As
counsel for MTN submitted, the unlawful determination has a ‘domino effect’ on
the balance of the Regulations. As I have already sought to demonstrate in respect
of the definition of the retail markets, I agree with the submission. This should be
dispositive of the determination that the markets were ineffectively competitive.
It is nonetheless imperative to address the issue squarely on its merits.
39
[102] ICASA’s source of its power to determine the effectiveness of the
competition in the defined markets is s 67(4)(b) of the ECA. The recurring use of
‘in those markets’ and ‘in the relevant markets’ signifies that a sine qua non of
the assessment is properly and lawfully defined markets. Accordingly, i f
regulation 5, in respect of which the effectiveness of the competition is made, and
the finding of ineffective competition is based on flawed regional markets, tainted
by an irregular and a flawed market definition, it yields a flawed result.
[103] Even if I am incorrect in this regard, in determining the ineffectiveness of
the competition, s 67(4A) of the ECA stipulates that ICASA ‘must consider,
amongst other things . . . dynamic character and functioning of the markets or
market segments . . . relative market share . . . and relative market power’ that is
in issue. The way the provision is framed is mandatory. It does not grant ICASA
a discretion to overlook the terms of the provision.53 If there is any discretion, it
seems to me that would apply to any other additional factors ICASA may consider
relevant. These too must be clearly articulated.
[104] The discussion document first referred to the national market shares of
various MNOs. ICASA’s final determination in the findings document sets out:
(a) the combined market shares of MTN and V odacom in the identified regions
at the end of 2018 and the end of 2019 based on their share of 90 -day active
connections over this period; and
(b) concludes that ‘the shares of 90 -day connections accounted for by the
incumbents (MTN and V odacom) has [have] not changed significantly over the
year in most regions’ and ‘there is a lack of dynamism in these markets’.
53 Pepper Bay para 32.
40
[105] On ICASA’s analysis, MTN moved from a market share of more than 45%
in five of the 16 regions in 2018 to a market share above this threshold in two
regions in 2019.
[106] Since ICASA decided that the retail markets were regional, it was required
to assess the relative market shares of the MNOs within each of the regions it had
identified. Instead, ICASA set out the combined market shares of MTN and
V odacom in each of the regions. Thereafter, it referred to the international price
benchmarking and made observations on whether the retail market might be
competitive on a forward -looking basis . The international pricing comparison
was inconclusive. ICASA concluded in the discussion document that:54
‘high relative market shares of individual licensees in many municipalities in South Africa
suggests that there are a number of geographic areas characterised by ineffective competition.
The high levels of concentration and lack of dynamism in market shares nationally over time
suggests that these market shares are unlikely to change significantly over the medium term.’
(My emphasis.)
[107] However, in the findings document, ICASA with regards to market shares
concluded that:55
‘An important question is why there is a lack of dynamism in these markets. It is likely that
this is linked to barriers to entry for challenger networks (including Telkom and Cell C),
discussed above, and to competition problems in the market for voice services, as set out in the
Discussion Document. There were no serious objections to the latter analysis in any of the
stakeholder submissions.’
[108] MTN challenged ICASA’s failure to apply the mandatory factors in
s 67(4A) to determine ineffectiveness of the competition. As demonstrated in
54 Discussion document para 45.
55 Findings document para 85.
41
Pepcor, that justifies a scrutiny of the application of the provisions to the
determination.
[109] First, a ‘suggestion’ is not a definitive finding of fact. ICASA contend ed
that MTN and V odacom’s combined market shares had not changed significantly
over the years in most regions. ICASA’s justification is that it considered the
supply-side substitution when assessing the competitiveness in the markets.
There were high barriers to entry, generally. Thus, there was little reason to expect
that supply-side substitution would be timely and sufficient to defeat any exertion
of market power, particularly in rural markets. In its answering affidavit, it
justified the failure to publish individual market shares by region in the findings
document on grounds of confidentiality. I have not been directed to the
confidential parts of the record in which it did so. On the papers, the conclusion
is based on information assessed of over a period of a year.
[110] Other factors considered were that Cell C, Telkom, Rain and Liquid are not
present in certain areas, and MTN and V odacom are the only incumbents. As a
result, ICASA was statutorily bound to intervene to ‘remedy market failures ’.
Telkom’s growth did not have ‘a substantial dent ’ on V odacom and MTN’s
significant market power. Entrants have not been able to reduce the market shares
of MTN and V odacom much below 90% ‘in many rural markets. ’ Cell C had
declined and entry in various regions is limited. It concluded that ‘this suggests
that markets are ineffectively competitive and will continue to be so on a forward-
looking basis’. (My emphasis.)
[111] In ICASA’s view, even on MTN ’s and the Commission ’s national
geographic market in retail mobile services, the Commission found that the
mobile retail market is ineffectively competitive. ICASA stated that ‘If it is bad
at national level, it is certainly worse at regional level .’ It contended that the
42
outcome would not changes regardless of whether the market was defined
narrowly or broadly.
[112] This makes clear that ICASA did not determine the ineffective competition
in each of the regional markets it had identified in regulation 3(a). The criticism
that ICASA conducted a ‘rudimentary analysis’ based on ‘outdated data’ and did
not consider the matters mandated in s 67(4A) is not unfounded. There is no
indication that ICASA considered the mandatory factors identified in s 67(4A) of
the ECA per region. Such an approach is irregular.
[113] ICASA did not analyse the relative market shares and relative market
power of the various MNO s in each of the regional market identified . I was not
directed to any evidence in the rule 53 record, or during argument to show that
ICASA considered the effectiveness of the competition in each region identified
and had referenced to individual market shares of MTN and V odacom or those of
other MNOs relative to one another. ICASA's allegation that it considered the
individual market shares of the MNOs for the purposes of assessing effectiveness
of the competition in the retail market falls to be rejected on the papers.
[114] MTN made representations and placed further information before ICASA
stating that:
(a) MTN’s subscriber market share moved from roughly 42% in 2011 to
approximately 30% in 2018, reflecting a 28% decrease in market share over a
period of seven years.
(b) The market share of Telkom Mobile had increased by 145% from 2011 to
2018 (and Cell C's market share had increased by 83% over this period). Telkom
Mobile's number of subscribers roughly quadrupled in just four years from 2016
to 2020.
43
(c) Additionally, Telkom's data network continues to carry nearly twice the
volume of MTN's data network as it carries 942PB in comparison to MTN's 524
PB year to date. Further, evidence of Telkom's significant growth has been
submitted by MIN to ICASA in its letter of 24 February 2021,
(d) Telkom's market update for the nine months ended 31 December 2020
showed, amongst other things, that:
(i) Telkom Mobile grew its mobile data revenues by over 46% to more
than R9 billion which was comparable to MTN's mobile data
revenues over this period (R11.45 billion);
(ii) Telkom Mobile served 2.6 million of the more lucrative post -paid
customers (in comparison to the 3.3 million post-paid subscribers of
MTN); and
(iii) Telkom grew its number of sites by more than 10%, adding 500
additional sites year on year.
[115] ICASA did not dispute that MTN's market share declined from roughly
42% in 2011 to approximately 30% in 2018, a decline of 28.6% (or 12 percentage
points). There is no evidence of how or where it engaged or considered the
information relating to the changed national competitive dynamics and a decrease
in its national market share. It failed to have regard to this material consideration
in its assessment of effectiveness of the competition.
[116] A further cause for complaint was that ICASA had regard to high barriers
to entry whilst failing to have sufficient regard to the low barriers to expansion
in the retail market. The point made is that low barriers to expansion may mitigate
against the effect of high barriers to entry. ICASA had acknowledged in the
discussion document that ‘it may be easier for an existing rival to expand capacity
into new product ranges than for a new firm to enter the new market ’. The
lopsided assessment of market dynamic s fails to take account the information
44
before it, that Cell C and Telkom Mobile already had national coverage under
their roaming and other network sharing arrangements and could potentially
respond to any price movement by other operators. Instead, ICASA merely stated
that Telkom’s growth in retail mobile services has been ‘minimal’ and that Telkom
data included fixed-wireless customers.
[117] It will be recalled that the product market included data. In any event, on
information before it, ICASA did not dispute that Telkom Mobile's subscribers
increased to just under 10% by 2018 after entering the market towards the end of
2010. ICASA failed to consider the resurgence of Telkom based on the relevant
information before it.
[118] It was not disputed that V odacom offered regional pricing strategies. MTN
in response offered its customers offerings like ‘MTN Zone’, ‘Made 4 U ’ and
‘PAYG MTN Zone’ to increase competitiveness and retain its customer base,
before it introduced the above -the-line ‘My Town’ offerings in August 2020.
These responses are inconsistent with ineffective competition dynamics.
[119] Accordingly, whether viewed conjunctively with regulation 3(a), or as a
self-standing regulation, the above failures support the conclusion that ICASA
did not lawfully determine the issues mandated by the ECA. The Regulations
were only promulgated in March 2022. It had information which indicated a shift
in market dynamics and an effect on relative market shares. It failed to consider
the information or provide a cogent explanation for rejecting it. Its determination
was not supported by the requisite analysis or justified in the findings document.
[120] ICASA lacked a factual basis for the finding of ineffective competition in
the retail market. The findings document does not indicate in respect of which
region the finding of ineffective competition made. It did not determine the
45
relative market shares of the MNOs. The failure to consider the mandatory factors
in s 67(4A) of the ECA constitutes a legitimate ground for review . I CASA’s
determination cannot be said to have been rationally connected with the power
conferred on it by the section. Its determination is also susceptible to be reviewed
under s 6(2)(e)(iii) of PAJA. ICASA failed to take account of materially relevant
information provided by MTN even though this was provided prior to the
publication of the regulations. For the reasons stated above, ICASA 's
determination of ineffective competition in the retail market in regulation 5, is
irregular, unlawful and is reviewed and set aside.
Significant market power in the retail market regulation 6(a)
[121] This challenge is directed at regulation 6(a) which states that:
‘The Authority has determined that MTN and V odacom are dominant in the following markets:
(a) Retail market: MTN is dominant with a market share of between 49% -55% in two
geographic markets for retail mobile services and therefore has SMP [significant market power]
in those markets. V odacom is dominant with a market share of between 47%-75% in 7
geographic markets for retail mobile services and therefore has SMP in those markets. MTN
and V odacom also have SMP as a result of vertical relationships that could harm
competition.’(My emphasis.)
[122] The contention is that to the extent that ICASA finds that MTN is dominant
and has significant market power in the markets defined in regulations 3 (a),
regulation 6(a) is based on the unlawful definition or based on the unlawful
determination that these markets are ineffectively competitive.
[123] The challenge to the substance of the findings is that:
(a) ICASA’s findings that MTN had significant market power in various
markets were unreasonable, based on a consideration of irrelevant considerations,
tainted by a failure to consider relevant considerations, and were materially
influenced by an error of law.
46
(b) ICASA’s findings that MTN had significant market power in the defined
markets was procedurally unfair and tainted by the failure to properly consider its
submissions on the findings document.
[124] The discussion document and findings document show that ICASA found
MTN had significant market power not only because it was dominant in two
regional markets by virtue of its market share of more than 45%, as contemplated
in s 67(5)(a) of the ECA, but also because it had ‘a vertical relationship’ that could
‘harm competition’, as contemplated in s 67(5) (c) of the ECA. Thus, ICASA
relied on two jurisdictional requirements, namely, ‘dominance’ and the existence
of ‘vertical relationship’ to establish that MTN had the significant market power
mandated in s 67(5)(c).
[125] When ICASA first defined the retail markets narrowly as municipal, it
found that MTN had significant market power in 78 municipalities and reached
the 45% dominance threshold. ICASA alleges it had regard to each of the MNOs
to assess significant market power. On its calculations, MTN's market share in
the Eastern Cape and the Free State exceeded 45%. It found that in 2018, MTN
was dominant and had significant market power in both ‘urban and rural’ markets
of the two regions based on a 90-day active subscriber share. However, the
findings document shows that in 2019, MTN was dominant only in rural markets
in both regions.
[126] It explained its premise for ‘vertical integration’ as follows:56
‘MTN and V odacom are both vertically integrated since they operate downstream in offering
retail services as well as upstream , having been assigned spectrum, operating their own high
sites and offering roaming services. This degree of vertical integration is likely harmful to
56 Discussion document para 72 on page 53.
47
competition and gives rise to both operators having significant market power at the wholesale
and retail levels.’ (My emphasis.)
[127] ICASA stated further that:57
‘The Authority’s consideration of the degree of vertical integration in the markets in the
Discussion Document is not really contradicted, save for the claim that there ought to be
evidence of abuse of market power, such as foreclosure. However, the Authority’s primary role
is to regulate on a forward -looking, “exante” basis, and so finding market power only after
concluding there has been an abuse is not a proper approach.’
[128] First, the consequential flaw already alluded to in respect of retail market
definition and effectiveness of competition impacts the finding of significant
market power. Section 67(4) (c) of the ECA authorises ICASA (as also set out
above) to ‘determine which, if any, licensees have significant market power in
those markets and market segments where there is ineffective competition’. (My
emphasis.)
[129] The provision requiring a determination of significant market power is
sequentially connected with ‘those markets’ lawfully and regularly defined, and
the lawful finding of ineffective competition. An irregular market definition,
similar with the evaluation the effectiveness of competitiveness, will taint
determination of significant market power in those markets, in turn. In essence,
regulation 6(a) and the finding of significant market power is the result of
unlawful and irregularly determined markets.
[130] A second difficulty is that ICASA’s answer departs from the position that
‘market power ’ and ‘significant market power ’ bear the same meaning. MTN
57 Findings document para 98.
48
contends ICASA made a material error of law by conflating ‘market power’ with
‘significant market power’.
[131] As set out above, s 1 of the Competition Act defines ‘market power’ as:
‘the power of a firm to control prices, to exclude competition or to behave to an appreciable
extent independently of its competitors, customers or suppliers’.
It is considered as the ability to raise prices above competitive levels.58
[132] Section 7 of the Competition the Act deems that market power exists where
a party has a market share of at least 45%. At the lower end of the scale, the
legislature recognises that market power is less likely to exist when a firm has
a market share under 35% and therefore requires proof that a party actually
possesses market power.59
[133] Under the ECA, significant market power with in a market is defined in
s 67(5). As set out above, a licensee will have significant market power where it
is proven that they have market dominance, or they are in control of essential
facilities, or has a vertical relationship, which, as determined by ICASA,
potentially harms competition.
[134] Contrary to contention by ICASA, in the context of market regulation as
envisaged by the ECA, significant market power and market power are not the
same. In my view, the bar is higher than in merger regulation. Proof of dominance
must be established together with either a control of an essential facility, or an
existence of a vertical relationship determined to harm competition determines
significant market power. Moreover, facts which establish either a theory of harm
or harm which it seeks to remedy must be articulated.
58 P Sutherland Competition Law of South Africa (November 2024 – SI 27) para 7.7.6.2.
59 P Sutherland Competition Law of South Africa (November 2024 – SI 27) para 7.7.6.1.
49
[135] To establish dominance, ICASA must show , in terms of s 7 of the
Competition Act, that MTN has:
‘(a) it has at least 45% of that market;
(b) it has at least 35%, but less than 45%, of that market, unless it can show that it does not
have market power; or
(c) it has less than 35% of that market, but has market power.’
[136] ICASA claims to have ‘recomputed’ market share by regions. The findings
document states that MTN was dominant in 5 regions in 2018 and in two regions
in 2019. In respect of the two region MTN was dominant in ‘urban and rural’ in
2018, but dominant in the ‘rural’ markets of both regions in 2019 based on a 90-
day subscriber share.
[137] ICASA’s analysis demonstrated a decreased market share by 9% in rural
Eastern Cape and an increased market share of 19% in rural Northern Cape over
the one -year period. The discussion document reflects that Telkom Mobile ’s
subscribers increased to approximately 10% by 2018 , after entering the market
towards the end of 2010. It was shown that MTN’s market share declined from
roughly 42% in 2011 to approximately 30% in 2018 a decline of 28.6% (or 12
percentage points).
[138] MTN submitted that its subscriber market share had been declining over
time while the market shares of Cell C and Telkom Mobile had increased
dramatically over time. It is not clear how ICASA accounted for MTN’s
decreased market share and the resurgence of Telkom in its determination of
significant market power. Notably, on the facts before it, ICASA accepted that
Telkom Mobile had grown over the past years. It nevertheless concluded that
50
Telkom’s growth did not have ‘a substantial dent ’ on V odacom’s significant
market power. There is no justification for this in the findings document.
[139] As MTN contended, if the mobile retail market w ere defined as national,
MTN’s market share would likely be below the 45% threshold for dominance.
ICASA did not dispute that MTN ’s subscriber market share moved from
approximately 42 % in 2011 to 30% in 2018, reflecting a 28 % decrease in market
share over the period. If national subscriber numbers were properly considered,
they reflect that MTN’s market share dropped from 40% to less than 30% between
2011 and 2020.
[140] Once more, there is no indication that ICASA established the individual
market shares of the various MNOs in the regions that it had identified, other than
to consider, for determining significant market power, whether the market shares
of MTN and V odacom were under or over 45%. ICASA did not consider the
relative market power of the licensees in the regional retail market s. MTN’s
complains, justifiably, that this likely ‘masked material fluctuations ’ of the
individual market shares.
[141] A material aspect of MTN’s complaint is that ICASA conflated ‘vertical
integration’ with ‘vertical relationship’. MTN disputes that a vertical relationship
involves a relationship between an MNO and third parties, and claimed ICASA
used the terms interchangeably, which constitutes a material error of law. When
the content of this regulation is juxtaposed with the conclusions of fact in the
findings document, regulation 6(a) refers to ‘vertical relationships and states that:
‘MTN and V odacom also have SMP [significant market power ] as a result of vertical
relationships that could harm competition.’
51
[142] ICASA simultaneously stated in its answering affidavit that:
‘I deny that ICASA found that vertical integration is determinative of significant market power.
The extent of vertical integration in the sector is an indication of significant market power .
Vertical integration can be good for competition. It can also be a source of ineffective
competition. It may result in input or customer foreclosure concerns. Here, MTN and V odacom
may engage input foreclosure. i.e., the refusal to grant a firm competitor or customer access to
an input which is vital for them to compete effectively in the mark et. They have the incentive
and ability to engage in this conduct.’ (My emphasis.)
[143] According to ICASA, it was immaterial whether one firm provided the
services in -house, and the others are outsourced these services . It described
‘vertical integration’ as a species of a ‘vertical relationship’ and argued that there
is no authority to suggest that a ‘vertical relationship ’ excludes ‘vertical
integration’. It states that a vertically integrated firm may engage in ‘a margin
squeeze’, which could harm competition, and significant market power was
compounded by the fact the firms are vertically integrated. In its answer, ICASA
claims it used the terms interchangeably; and contended in argument that MTN
must have ‘understood and accepted ’ that vertical integration is a specie s of a
vertical relationship, albeit internal to the firm.
[144] However, a s stated earlier, section 1 of the Competition Act defines a
‘vertical relationship’ as a ‘relationship between a firm and its suppliers, or its
customers or both’. On the other hand, ‘vertical integration’ is internal to the firm,
and is a business arrangement in which a firm participates at different levels of
the supply chain. On this score, a ‘margin squeeze’ is an exclusionary conduct by
a firm and an abuse of its dominance. 60 As submitted by MTN, this ‘clearly
a firm and an abuse of its dominance. 60 As submitted by MTN, this ‘clearly
indicates that a vertically integrated firm is one firm for purposes of that Act’.
60 Section 8(1)(c) of the Competition Act. See also Telkom SA Ltd v Competition Commission South Africa [2011]
ZACT 4 para 40, referring to the Constitutional Court in Competition Commission of South Africa v Senwes Ltd
[2012] ZACC 6; 2012 (7) BCLR 667 (CC).
52
[145] It is clear from the preceding paragrap hs that ICASA conflated ‘vertical
integration’ and a ‘vertical relationship’. The submission that one is a species of
the other lacks merit. It is inconsiste nt with the meaning employed in the
Competition Act. It is a fundamental and material error of law to treat vertical
integration and a vertical relationship interchangeably , as ICASA did. The
determination of significant market power on this basis is flawed and influenced
by a material error of law within the meaning of s 6(2)(d) of PAJA. Furthermore,
it was not supported by an analysis of the true economic nature of the
relationships involved.
[146] As submitted by MTN, when assessing dominance and significant market
power, a vertical integration does not in itself indicate dominance or significant
market power since it can have economic efficiency rationale. This approach is
consistent with considerations a Tribunal or a Court would consider in the context
of merger regulation. As pointed out, although that jurisprudence is relevant, the
exercise of the regulation power arises from statute to remedy market failures. It
is not necessary for the court to reach a conclusion on whether ICASA had to
investigate the economic efficiency rationale. What is evident is that the irregular
determination is patent and established from the conflation of the two terms. The
error is material and sufficient for the purpose of the review.
[147] The Regulation is based on outdated market share data. ICASA failed to
meaningfully engage with the information placed before and after the circulation
of the findings document . ICASA failed to consider relevant considerations,
which is influenced by a flawed geographic market definition. That includes, but
is not limited to, the decrease in MTN’s national market share and the resurgence
of Telkom. ICASA's determination in regulation 6 (a) that MTN has significant
market power in the retail markets is irregular and falls to be set aside.
53
Review of site infrastructure access
Infrastructure access market definition
[148] Regulation 3(b) defined the site infrastructure market as:
‘Upstream market 1: wholesale site infrastructure access in local and metropolitan
municipalities.’
It purports to regulate the ‘wholesale site infrastructure access’.
[149] The challenge is that regulation 3(b), defining the wholesale site
infrastructure markets is not rationally connected to the purpose for which the
empowering provision in the ECA they were made and or the information placed
before ICASA within the meaning of s 6(2)(f)(ii) of PAJA. In respect of regulation
3(b), I CASA irregularly excluded infrastructure sites made available by non -
licensees and micro solutions, lampposts and billboards (‘micro sites’) from the
applicable market.
[150] MTN contends that there is no separate market for wholesale site access
provided by licensees, as opposed to access to property provided by non-licensees
to construct passive infrastructure for the delivery of their mobile services. Many
sites are controlled by non -licensees, like American Tower South Africa (ATC),
landlords, other private businesses or organs of State, and these sites form part of
the product market.
[151] The criticism is that the market definition is ‘artificial’ and ‘overly
narrowed’ and is legally flawed. On a proper application of the SSNIP test, the
geographic scope of the market is broader than a local municipal. ICASA failed
to adequately apply the SSNIP test framework to define the site access market.
54
[152] MTN’s second complaint concerns the product market definition of the site
access market. It contends that the definition is flawed because it excludes micro-
sites and micro-site solutions, like lampposts and billboards , from the product
market. ICASA’s accounting for roof top sites was flawed.
[153] ICASA disputed that there is an irregularity in its geographic site market
definition, and contended if found to be irregular , that the irregularity is
immaterial. It submitted, based on the discussion document, that non-licensees
account for approximately 14% of the site access market as confirmed by ATC
South Africa on 7 August 2020 . It disputed that many sites are owned or
controlled by non-licensees.
[154] ICASA claims that ‘all suitable sites have been included’ in its analysis of
the site access market. It claims that it considered competitive constraints exerted
by non -licensees in determining the relevant market and when assessing
competitive dynamics and constraints. As the regulator concerned with ensuring
competitive outcomes and the efficient operation of the markets, it accepted that
the impugned Regulation is targeted at licensees only. ICASA was advised to
include ‘rooftops used but not owned by mobile operators in its market analysis’.
[155] With regards to the application of the SSNIP test, it contended that the
absence of substitutability between micro - and macro-sites engages the SSNIP
test. ICASA justified the exclusion of micro-solutions on the grounds that they
‘were unlikely to provide the equivalent coverage and capacity’ offered by macro-
sites, roof tops and indoor sites. However, this argument relates to the product
market rather than the geographic site access market that ICASA sought to
identify.
55
[156] Although ICASA, based on an earlier submission, stated in the findings
document that ‘all suitable rooftop have already been taken up by MTN and
V odacom’ it is common cause on the papers (evident from ICASA’s answering
affidavit attaching a letter dated 7 August 2020), that ATC held more than 500
rooftop sites for use by communication networks at the time of ICASA’s market
inquiry and the time when the Regulation was promulgated. There was no factual
basis to claim that all suitable rooftop sites have been taken MTN and V odacom.
[157] Importantly, ICASA equally stated that ‘not all rooftops across South
Africa have been included in the market ’. It did so without specifying which
rooftops were included, nor did ICASA identify the municipal site access markets
in which they fell.
[158] The difficulty with ICASA’s submissions is compounded by s 67(4)(a) of
the ECA, which requires for ICASA to define the relevant wholesale and retail
markets or market segments. The ECA defines ‘wholesale’ as:
‘the sale, lease or otherwise making available an electronic communications network service
or an electronic communications service by an electronic communications network service
licensee or an electronic communications service licensee, to another licensee or person
providing a service pursuant to a licence exemption.’
[159] However, the ECA envisages a sale or lease of a wholesale network or
communication service in instances where they are provided by an electronic
communications network service licensee or an electronic communications
service licensee, to another licensee or person providing a service pursuant to a
licence exemption. It appears that sites in themselves, as infrastructure do not fit
the definition of wholesale electronic communication serv ice, but the
infrastructure installed on the sites.
56
[160] Assuming for the moment that ICASA is correct and there was a separate
wholesale site access market, as already stated above, the question for the SSNIP
test would be – if either MTN or V odacom raised prices for site infrastructure
access, would the other participants likely respond to render t he increase
unprofitable? How ICASA applied the SSNIP test to define the site access market
is not clear.
[161] In opposition , ICASA counter ed that the absence of substitutability
between micro- and macro-sites engages the SSNIP test. However, that relates to
the product market and not the geographic site access market it set out to define.
I have not been able to discern how ICASA engaged the essential inquiry in
respect of the identified site access market. Importantly, t he ECA excludes
infrastructure sites that are not made available by licensees from the definition of
‘wholesale’. The inclusion of sites by non -licensees to determine the relevant
market does not accord with the ECA.
[162] Turning to the exclusion of micro -sites, MTN’s complaint is that ICASA
irregularly excluded micro-sites and micro- solutions by non-licensee. Given the
view I take on the meaning of ‘wholesale’ in the ECA, I am unable to conclude
the specific exclusion was irregular per se. However, ICASA accepted broadly in
its answering affidavit that micro sites and solutions may play an important role
in G5 networks in the future, but found that currently, they were not ‘sufficiently
substitutable’. It contended that micro-solutions, such as lampposts and
billboards, were unlikely to provide the equivalent coverage and capacity offered
by micro-sites, such as rooftops and indoor sites. The exclusion of micro-sites
meant there will be no effective substitution in the event of a SSNIP.
57
[163] The difficulty is that ICASA based its decision on a different test, namely
‘equivalent coverage and capacity’, which does not engage the SSNIP test. The
exclusion of micro-sites, even those owned by licensees from the product market,
on the grounds that they are unlikely to provide equivalent coverage, or capacity
is not supported by detailed analysis envisaged in Medicross.
[164] How the different forms of sites, which MTN alleges are linked by chains
of demand -side and supply -side substitution , exert competitive constraints on
each other, particularly when it is alleged that individual sites are likely to overlap
substantially with the areas that can be covered by other sites is not addressed.
The findings document merely stated that ‘it would seem unreasonable that a
price increase in Johannesburg would be constrained by a site in Cape Town by
means of a chain of substitution ’. This is wholly insufficient for the purpose of
the inquiry required.
[165] ICASA did not dispute that MTN made representations that MTN and
V odacom use the same underlying and contiguous site infrastructure networks to
serve customers throughout the country. ICASA was informed that different
forms of sites are linked by chains of demand -side and supply-side substitution,
and there are typically several options for any network requirement. It was
submitted there was a high level of demand -side switching between different
forms of sites. MTN alleges that ICASA failed to consider rel evant information
placed before it in its market assessment.
[166] The reasons document states that the comments from MTN related to
matters already considered during the inquiry and dealt with in the findings
document. But the findings document had not considered any submissions ,
specifically regarding the inclusion of micro-sites in the market definition,
because ICASA had reached that finding for the first time in the findings
58
document, had not raised it as a possible finding in the discussion document.
MTN had made those submissions following the findings document . It
demonstrates, contrary to ICASA ’s assertions, it did not consider (or did not
properly consider) MTN’s representations on the findings document.
[167] The disagreement and challenge to ICASA’s wholesale site access market
definition is once more compounded by the concerns raised by the Commission,
and referred to in the findings document . ICASA acknowledged that the
Commission:
‘raised a concern about the Authority’s focus on sites rather than facilities included in the
findings document on priority markets (i.e. upstream infrastructure markets) stating that the
Authority should have considered additional markets for access to other types of facilities.’
[168] What emerges from the papers and the issues raised is the complexity
associated with the type of each site, the value derived from the infrastructure, its
functionality and operational deployment. How ICASA accounted for these
variances, at least in respect of the sites owned by licensees is not clear. Whether
the micro -sites are intermediate products that pose competitive constraints
directly or indirectly was not considered. Instead, ICASA sought to rely on
representations made by MTN to support its exclusion of micro-sites. As already
seen in respect of the retail market definition, t he ex post facto reliance on
representations by MTN constitutes an impermissible justification of its
determination, as similarly found in National Lotteries.
[169] Having regard to the above, I find that ICASA did not identify the relevant
markets and the definition according to the principles of the SSNIP test envisaged
in the methodology in regulation 4(a) of the impugned regulations. ICASA did
not established ‘the set of products or services which exert competitive
constraints on one another ’, in respect of the wholesale site access market. The
59
wholesale site access market definition is arbitrary and flawed. The determination
was not rationally related to the purpose of determining an appropriate market for
site access. It falls to be reviewed and set aside.
Ineffective competition in the wholesale site infrastructure access market
[170] This challenge is directed at regulation 5 and the determination that the
wholesale site infrastructure market in regulation 3(b) defined as ‘Upstream
market 1’ is ineffectively competitive.
[171] ICASA found that there were ‘high levels of [site] concentration in many
municipalities’. Possible sites are limited in urban areas, and often, were already
occupied, making it difficult to roll out new sites. It claims the questionnaire
elicited ‘considerable challenges to establishing new sites ’. MTN and V odacom
had the first mover advantage , which meant they were able to build out sites in
the most advantageous locations first. Their downstream market power is in part
a reflection of this.
[172] ICASA acknowledged that infrastructure sharing was common and did not
dispute MTN and V odacom’s submission that there is an incentive to share sites
to reduce costs. However, it found that barriers to entry were ‘considerable’ since
wholesale services are not supplied competitively in both facility and service -
based entry level. The high cost and complexity of building sites, the regulatory
municipal approvals, environmental impact assessments and wayleaves lead to
substantial costs. These factors and delays posed further barriers. Much of the site
sharing occurs between MTN and V odacom.
[173] ICASA claims that MTN's historical prices with its customers, like Cell C
and Telkom, drives this point home. Incumbent operators acted as a further barrier
and take a long time to consider and approve co -location requests by smaller
60
operators, or reserve space for expansion of their coverage or capacity. In
contrast, tower companies allocate space on a first-come-first-serve basis. Since
V odacom and MTN are two mobile network operators with national networks, it
considered them a ‘duopoly’ and found that they do not constrain each other.
[174] The findings document concludes there are ‘indications’ that markets for
access site s are ineffectively competitive. These competitive dynamics are
unlikely to change in the next three years. On this basis, ICASA found that there
is ineffective competition in the site infrastructure access market in
municipalities. ICASA persisted in opposition that it reasonably and rationally
found that there was ineffective competition in upstream market 1
(site infrastructure access market). It justifie d the impugned regulations on the
grounds that they are intended to solve the ‘MTN-V odacom only rivalry
problem’.
[175] MTN challenges ICASA’s determination on the grounds that it is flawed
and irregular in the following respects:
(a) Since each municipality constituted a separate geographical market for site
infrastructure access, ICASA did not decide on ineffective competition in each of
the municipal site markets it had defined.
(b) ICASA did not consider all the mandatory factors in s 67(4A) to determine
the effectiveness of competition in the identified market.
(c) There is no indication that it evaluated the relative market shares, the
dynamic character and functioning of the market and the relative market power
of licensees in the municipal markets, as determined by ICASA.
(d) ICASA treated MTN and V odacom market shares on a combined basis, and
ICASA relied on unexplained cost range prices, unsubstantiated complaints and
made contradictory findings on MTN’s dominance.
61
[176] The claim is that ICASA relied on the number of sites as the metric to
calculate market share. There is no indication that the individual market shares of
MTN and V odacom, or any of the other licensees , were considered for the
purposes of assessing the effectiveness of competition in the site access market.
[177] ICASA in its answering affidavit claims that it was advised to consider ‘a
list of municipalities with high levels of concentration and had regard to
competitive conditions in each municipality ’. ICASA claims that its analysis of
market power in each of the municipal markets and the calculation of Herfindahl
Hirschman Index (the HHI) as published ‘ requires the calculation of market
shares by mobile network operator as a prior step’, therefore there is no substance
to MTN's complaint in this regard. ICASA contended that MTN offered no
countervailing evidence showing that it did not consider MTN and V odacom's
high market shares in the provision of site access, nationally or by municipality.
[178] MTN accepted that the HHI measures market concentration based on
market shares. MTN submitted however that ‘market share concentration
measures are generally not a sufficient condition to draw conclusions about
whether firms possess market power’.
[179] The findings document sets out in a table the municipalities in which
ICASA determined that MTN and V odacom had market shares of 45% or more
(although this table did not state so). 61 The findings document , found the
combined market shares of MTN and V odacom exceeded 60% without an
indication that it had any regard to the competitive conditions prevailing in each
municipality.62 The findings document and reasons document merely rely on the
61 Findings document at 27.
62 Findings document para 147.
62
percentage of the market. Yet , it concluded that V odacom and MTN do not
constrain each other.
[180] I accept that whether a market is characterised by ineffective competition,
requires more than the tallying or simple calculation of market shares. As
demonstrated earlier, ‘market power’, as quoted above, refers to the ability to
‘control prices, to exclude competition or to behave to an appreciable extent
independently of their competitors, customers or suppliers ’. I understand, the
argument is also influenced by the appropriateness of the tool ICASA utilised to
measure site market shares. It is not necessary for this Court to enter the economic
debate about the appropriateness of the tools used.
[181] ICASA’s answer to the criticism of its assessment and application of these
provisions is that it took into account the market shares, pricing over time,
complaints against licensees, as well as how agreements for site sharing have
changed over time to inform a forward-looking and predictive assessment
[182] On the facts, there is no indication in the record that ICASA considered the
MNOs market shares either individually or relative to one another in its
assessment of the effectiveness of the competition in the site access market.
ICASA claims that Dr Hawthorne's evidence refutes this claim and explains that
the site access market shares for MTN were based on data and submissions from
MTN and other stakeholders. MTN's site share per municipality was determined.
[183] This is clear from confidential documents filed, Dr Hawthorne, in a table,
lists municipalities in which the combined market shares of MTN and V odacom
exceeded 60%. There is no indication that ICASA had any regard to the
competitive conditions prevailing in each municipality , and whether i ndividual
63
market shares of each MNO featured in ICASA's analysis determination of
significant market power.
[184] Moreover, MTN recently disposed of most of its towers to IHS Holdings
Limited (‘HIS’), a company listed on the New York Stock Exchange, and one of
the largest independent owners, operators and developers of shared
communications infrastructure in the world. The tran saction, which became
unconditional on 31 May 2022, and operationally came into effect on 1 August
2022, involved the disposal of 5 701 towers. ICASA did not dispute this
transaction, yet its findings document is silent on its effects on both the market
shares and the determination of MTN’s significant market power.
[185] Importantly, when evaluating barriers to entry, market shares and levels of
concentration in the site access market, it observed in the discussion document
that there were signs that the market may becoming more competitive. Two major
extensive site sharing agreements were recently signed. Telkom had grown its site
footprint rapidly in recent years, while Cell C has added a small number of sites.
The discussion document did not draw any conclusions about the effectiveness of
competition.
[186] When ICASA made a comparison of the costs and an analysis of the prices
charged by the operators for site rental, it observed that in many cases the prices
charged for new rentals were not at a substantial premium to costing in 2018,
although this may have been the case in the past. Yet, in the findings document,
it stated that ‘many access seekers pay considerably more than reasonable costs
ranges for sites’. In its answer, it averred that ‘in respect of the complaint received
by ICASA against the incumbent site access operators, ICASA has prima facie
evidence of high prices ’. In addition, it found that MTN and V odacom ‘do not
appear to be willing to functionally separate or divest sites to tower companies in
64
South Africa’ and that these competitive dynamics are unlikely to change in the
next three years despite MTN and V odacom's public statement to the contrary.
[187] The question is whether in making the determination, ICASA took into
account the factors listed in s 67(4A). Section 67(4A) (b) calls for an assessment
of the dynamic character and functioning of the markets or market segments ,
including an assessment of relative market share of the various licensees or
providers of exempt services in the markets or market segments , and a
forward- looking assessment of the relative market power of the licensees in the
markets or market segments
[188] There is no logical connection between an ‘admission’ of competition
between MTN and V odacom and the justification for ICASA's findings of
ineffective competition. ICASA’s defence must be viewed against the available
information at the time of considering and drafting the Regulation. The site
market share and the determination were based on outdated information. The
reasons document does not disclose that it considered MTN’s submissions.
[189] The disposal of MTN’s site portfolio contradicts the finding in the findings
document which was not altered in the reasons document. The number of site
access facilities owned by non -licensees at the time that the regulations were
promulgated was material to the determination of the effectiveness of
competition. The disposal of MTN’s portfolio changed the landscape of the site
access market. This was a relevant consideration to ICASA's findings of the
effectiveness of competition on a forward-looking basis.
[190] ICASA submitted that there was a dispute of fact on the papers, about
whether it considered competitive constraints by non -licensees. The dispute is
inconsequential, and it relies on Dr Hawthorne’s analysis of competitive
65
constraints, which means that the Plascon Evans rule should be applied. Since,
the disputed issue engages the proof a jurisdictional fact to trigger regulation, the
onus is on ICASA to establish the existence of that fact . The application of
Plascon Evans to such a dispute is inappropriate in my view.
[191] In conclusion, ICASA's finding that the site access market is ineffectively
competitive is flawed. ICASA did not consider the relative market power of the
MNOs in each of the municipal site access markets. It failed to have regard to
relevant considerations and made contradictory statements and findings in respect
of the market. Materially, together with the failure to consider the factors listed
in s 67(4A) of the ECA renders the determination irregular.
Significant market power in site access infrastructure market
[192] Regulation 6(b) determines that MTN has significant market power in the
site infrastructure access markets identified and states that:
‘Upstream market 1: MTN is dominant with a market share of between 45% and 52% in 8
geographic markets for site infrastructure access and therefore has SMP [significant market
power] in those markets. V odacom is dominant with a market share of between 45% and 65%
in 39 geographic markets for site infrastructure access and therefore has SMP in those markets.
MTN and V odacom also have SMP as a result of vertical relationships that could harm
competition.’ (My emphasis.)
[193] The rational e for the regulation is based on MTN’s participation in the
wholesale upstream site infrastructure and downstream activities, and the finding
that it has significant market power in the site access. ICASA justified the finding
of significant market power on the ground that MNOs were vertically integrated
with regards to site access market since most sites are owned by operators who
also provide wholesale and retail mobile network services. The high levels of
also provide wholesale and retail mobile network services. The high levels of
concentration at site level are correlated with high concentration at retail level.
66
[194] The s ame legal requirements for determination of ‘dominance’ and
‘significant market power’ dealt with in respect of the retail market apply. I need
not belabour the legal requirements, already dealt with in respect significant
market power in the retail market. Those requirements apply with equal measure
in this assessment.
[195] ICASA predicated MTN’s dominance and significant market power on the
grounds that the computation of MTN's market share exceeded 45% in these
municipal site access markets. The flaw with the finding that MTN was dominant
and had significant market power in the eight municipal geographic markets
identified is that the discussion document refers to 234 municipalities. The
findings document lists only the 47 municipalities in which either MTN (in 8
municipalities), or V odacom (in 39 municipalities) is dominant. ICASA excluded
district municipalities from its definition of municipalities and made this finding
without reference to all 234 municipalities in which it defined the site
infrastructure access market.
[196] MTN’s further grievance is that ICASA excluded all sites in which, Cell C
and Telkom could be dominant from its assessments. I cannot discern a cogent
answer from the papers for the exclusion of district municipalities, or of Telkom
and Cell C from the computation of site access market shares.
[197] As already alluded to, s 67(4)(c) of the ECA makes it clear that ICASA is
only empowered to decide significant market power in lawfully defined markets
where there is a lawful finding of ineffective competition. It follows that ICASA
could not have lawfully and reasonably made the findings of significant market
power within those markets. First, the consequential flaw in respect of site access
market definition and effectiveness of competition taints the finding of significant
67
market power and the lawfulness of the regulation. This alone should be
dispositive of the determination of significant market power.
[198] To amplify the flaw in computing the market shares, MTN averred that
ICASA ‘simply tallied up the number of sites in each area without regard to the
differences as between the various types of sites ’. Whether an MNOs’ market
shares are more than or less than 45% within a municipality is a far cry from
considering their relative market shares within each municipality , a mandatory
requirement under the provision. There is no indication that the relative market
shares and relative market power of licensees featured in its analysis.
[199] The second factor apparent from the regulation is ICASA’s premise that
MTN and V odacom have significant market power ‘ as a result of vertical
relationships that could harm competition’. The vertical relationship is premised
on site access pairing agreements between MTN and V odacom. In its answer,
ICASA ‘in part ’ attributed the non -availability of space to third parties to the
vertical relationship between V odacom and MTN in respect of site sharing. It
asserted that ‘the site sharing arrangement between V odacom and MTN has both
pro-competitive and anti-competitive effects’. To reinforce the anti -competitive
effects of site pairing arrangements, the deponent claimed that third parties
complained that the location(s) they obtain on masts is “disadvantageous”
because MTN and V odacom have already taken the best locations.
[200] ICASA’s reliance on complaints that access to MTN's site is problematic
were not investigated nor was relevant information provided to MTN. As the
court held in Bertie van Zyl (Pty) Ltd and Others v Minister of Agriculture,
Forestry and Fisheries and Others,63 a failure to provide an affected person with
63 Bertie Van Zyl (Pty) Ltd t/a ZZ2 and Others v Minister of Agriculture, Forestry and Fisheries and Others [2021]
ZASCA 101; [2021]4 All SA 1 (SCA) para 28.
68
the necessary information to place it in such a position that it responds, engages
or considers said information it will ‘irredeemably compromise’ the fairness of a
consultative process.
[201] An important consideration is the notification of the divesture of MTN’s
site portfolio, and its disposal, which altered the market shares of the participants.
ICASA had no regard for the change even though the information was presented
to it. The assertion in the findings document that MTN is unwilling to divest its
portfolio and the conclusion that the competitive dynamics were unlikely to
change in the next three years is not supported by the facts placed before it. There
is no rational basis on which it could be contended that MTN h as significant
market power and was dominant in any site access markets following the
disposal. There is no evidence that ICASA recomputed the significant market
power after the transaction. I agree with MTN’s assertion that ICASA’s
computation is inaccurate, irregular and unlawful.
[202] What c ompounds the difficulty with ICASA’s determination is the
undisputed fact that various sites provide coverage, capacity and /or both. As I
understand the contention, the weighting of the different sites is relevant to the
determination of the market share of each licensee and its significant market
power in the site market. An example of the type of sites is rooftops. ICASA
accepted in its answer that they form part of the macro-site access market. There
is no indication that ICASA weighted different types of sites even within the
municipalities it had identified for the purpose of market share calculation to
determine significant market power.
[203] On the facts above, ICASA’s finding that MTN has significant market
power in eight of the municipal site access markets is flawed and based on an
unlawful determination of the relevant market and competition in those markets.
69
ICASA’s determination was not only unreasonable, but it was also influenced by
a material error of law resulting from the conflation of vertical relationships and
vertical integration to justify the finding of significant market power in the site
access market. ICASA failed to take account of relevant consideration.
[204] For the reasons stated above, ICASA's finding of significant market power
in the site access market is irregular and falls to be reviewed and set aside.
Review of roaming market
Wholesale roaming market definition
[205] ICASA found that national roaming is separate from other wholesale
infrastructure and was in a separate market to other wholesale services such as
MVNO and wholesale APN services. There is no dispute about the geographic
roaming market. Regulation 3(c) identifies the market as:
‘Upstream market 2: wholesale national roaming services for coverage purposes .’ (My
emphasis.)
[206] MTN takes issue with the product market definition , which is limited to
coverage. It contends that roaming for coverage is not a distinct separate market
from roaming for capacity. It claims ICASA failed to apply the SSNIP test to
determine the product market and to properly consider the role of supply -side
substitutes. There was no ‘robust evidence ’ to support the different product
markets.
[207] ICASA erroneously departs from the view that the challenge to its roaming
market determination does not constitute a ground for review but is a ground for
appeal. As has already been shown in respect of the other markets, the market
determination is not an opinion of fact, but a jurisdictional requirement, required
by the relevant provisions.
70
[208] ICASA claims the delineation ‘was informed by the engagement between
ICASA and stakeholders, and comments from them’. It found, based on changes
in the market, that a single market for all roaming services would be ‘simplistic’.
It explained the distinction between markets for coverage and capacity as follows:
‘The provision of national roaming that provides supplementary coverage across the country
is distinct from the provision of roaming for additional capacity as given an increase in price,
customers requiring supplementary coverage would not switch to roaming provided by an
operator with capacity but without a national network that can provide coverage . As such, the
Authority defines a market for national roaming that allows for supplementary coverage.’
[209] ICASA found that national roaming was driven by a need for coverage. In
its assessment, there were two types of roaming agreements . The first type, the
‘traditional roaming agreement’, is driven by a need for coverage and is used by
‘smaller operators’ such as Cell C and Telkom to supplement their coverage in
areas in which they had not yet built a network. The second type were more recent
multi operator core networks agreement, which is driven by a need ‘for additional
capacity’ used by operators with a limited geographic footprint. For the latter, it
relied on agreements between MTN and Liquid, MTN and Cell C, V odacom and
Rain, V odacom and Liquid.
[210] The fact that there were only two operators that could provide additional
coverage, namely MTN and V odacom, appears to have held sway . ICASA
claimed that other operators do not have the same footprint to provide coverage.
Unlike traditional providers for roaming coverage services who can also provide
additional capacity, Rain, Cell C and Liquid have no capacity to provide national
roaming coverage. Therefore, they do not pose competitive constraints on MTN
and V odacom on the national roaming services market for coverage. ICASA
and V odacom on the national roaming services market for coverage. ICASA
found there was no supply substitution because ‘these providers of only additional
71
capacity do not exert any competitive constraints on V odacom and MTN who can
also provide coverage because of their national footprint’. ICASA submitted this
demonstrated the application of a ‘classic SSNIP test’.
[211] The justification is that w hen v iewed from a demand substitution,
customers requiring supplementary coverage would not switch to a roaming
provided by an operator with capacity but without a national network that can
provide coverage. According to ICASA, it was ‘quite clear’ that customers of
roaming services for coverage will not be able to switch to capacity suppliers ,
hence it considered them as a separate market in the wholesale market , because
‘they are not interchangeable’. During argument it was submitted this reflects that
ICASA considered and applied the SSNIPP Test.
[212] MTN contends that competition is based on customer demands and service
requirements, specifically voice and/or data, period of service, service terms and
rates. MNOs take decisions on where to roll out or augment network coverage
and capacity based on the attractiveness of the customer base in different areas.
There is no distinction between coverage and capacity. There can be no coverage
without capacity, or vice versa. For this reason, there are no limitations to specific
frequency spectrum bands, and no roaming agreements are limited to high level
definitions of coverage and/or capacity.
[213] An MNO that contracts for roaming services might either require these
services to supplement its coverage in a particular area (ie for additional capacity)
or it might require roaming services to extend its coverage to an area in which it
currently has no coverage. The mobile roaming agreements are not distinguished
by coverage or capacity requirements , but rather by service requirements,
specifically voice and/or data service requirements.
72
[214] Regulation 5 indicates that the market determination was made pursuant to
the methodology set out in regulation 4, which entrenches the SSNIP test. During
argument, the court was directed to the confidential findings document, in support
of the contention that ICASA applied the SSNIP test in determining the two
separate markets. There is no explanation or analysis on how SSNIP test was
applied other than merely stating: ‘[n]o plausible to switch one for the other using
SSNIP Test ’. This is wholly inadequate and does not constitute a proper
application of SSNIP test, as already explained in respect of the other ma rket
forms challenged and discussed above.
[215] ICASA’s justification of its market definition based on existing agreements
between MTN, V odacom and other operator customers is not the basis for a
proper determination of the product market. Other than to state that, MTN and
V odacom are the only providers, there is no reference to the role of supply-side
responses as between the se two primary infrastructure players in South Africa,
and the competitive constraints they pose on each other.
[216] What is more, roaming for coverage purposes as a distinct product market
was introduced for the first time in the findings document. The findings document
differed from the discussion document in material respects in relation to the
roaming market. The complaint about procedural unfairness is premised on the
fact that MTN ’s submissions on the findings document were not considered or
not properly considered . The principle in Earthlife Africa (Cape Town) v
Director-General: Department of Environmental Affairs and Tourism and
Another64 (as discussed below) applied to those submissions as representations.
64 Earthlife Africa (Cape Town) v Director-General: Department of Environmental Affairs and Tourism and
Another 2005 (3) SA 156 (C) (Earthlife) para 52 onwards.
73
[217] I find that ICASA’s product market definition is flawed. By failing to
properly consider the information before it and to correctly apply the SSNIP test
to both product markets , it irregularly excluded roaming for capacity . The
‘practical indicia’ relied on are not fully explained nor do they constitute a proper
application of the SSNIP test. Accordingly, ICASA did not determine the roaming
product market as mandated . Its determination is not connected with the
information placed before it. It thus falls to be reviewed and set aside.
National roaming market coverage is ineffectively competitive
[218] Regulation 5 states that ICASA determined that competition in the national
roaming market (ie Upstream market 2) is ineffectively competitive. As stated in
relation to the other markets above, the regulation confirms that the determination
was made ‘pursuant to Regulation 4’, which engages the SSNIP test.
[219] Here too, on the strength of s 67(4) (b), which envisages that the
determination of ineffectiveness of competition i s made in respect of properly
defined markets, the irregular market definition taints the determination of the
ineffectiveness of the competition in the roaming market.
[220] MTN contends that the finding was unreasonable , irrational based on
irrelevant considerations as ICASA failed to have regard of the three mandatory
factors as provided in s 67(4A) namely, the relevant market shares, the dynamic
character and functioning of the markets , and the relative market power of
licensees.
[221] The findings document states that prices had been high in the past. There
were complaints about poor quality and the average effective price paid for
roaming was often higher than the average retail price for the roaming providers
indicating ineffective competition. Nevertheless, ICASA acknowledged that the
74
national roaming market was ‘fairly dynamic’ and that new roaming agreements
that had recently been concluded were likely to produce pro -competitive
outcomes, improved technology arrangements and better coverage. This
dynamism is reflected in the decline of roaming prices over the last three years.
[222] With regards to m arket share ICASA observed in its findings document
that:65
‘The only parties that are able to offer national roaming for the purpose of coverage are MTN
and V odacom with population coverage of close to 99% for 2G and 3G, and 4G coverage is not
far behind. As such, market shares in terms of national roaming are very high. While it is true
that Liquid and Rain are providing roaming services, the view of the Authority is that this is
for the purpose of capacity and not national coverage and as such is different.’
[223] The past analysis and justification were confirmed in the findings
document, but without a comparative analysis on pricing for the roaming market.
They were based on a combined market share of MTN and V odacom. Neither the
discussion document nor the findings document provides any indication that
ICASA considered the:
(a) National roaming market shares of MTN or V odacom in relation to either
effectiveness of competition or significant market power;
(b) Individual market shares of MTN and V odacom in the national market. The
findings document merely states that ‘market shares in terms of national roaming
are very high’.
[224] An important consideration to determine effectiveness of competition is
how the market shares of participants might have changed over time . ICASA’s
determination and finding of an ineffectively competitive roaming market
appears to have been primarily influenced by the fact that only MTN and
65 Findings documents para 186.
75
V odacom are able to provide roaming for coverage purposes. ICASA did not
consider the relative market power of the license national roaming market.
[225] ICASA’s analysis of the effectiveness of the competition in the roaming
market considered MTN and V odacom on a combined basis. During argument,
ICASA relied on the Commission’s view that since MTN and V odacom have
universal and/or national coverage, this meant that ‘the market shares are more
accurately described as 50-50’. ICASA concluded that based on the market shares
of 50-50 out of 99% population coverage translates to more than 45% of market
share for MTN and V odacom without more.
[226] First, the reliance on the historical market performance means ICASA did
not engage sufficiently with how the national roaming market operated at the time
of making the regulations or at the time that the findings document was drafted
or compiled. In fact, it made findings that conflict with the determination in three
respects: first, it found that the competition has resulted in several pro -
competitive outcomes ; second, that there was a fall in roaming prices with
improved quality resolved by newer technologies; and third, that customer switch
from one roaming provider to another , as is evident from Cell C ’s switch from
V odacom to MTN and Telkom’s switch from MTN to V odacom. There is no
justification for disregarding the undisputed facts. Notwithstanding, ICASA
stated that it did not attribute the changes in the roaming market to effective
competitive constraints between MTN and V odacom but rather to ‘changes in
technology and countervailing power’.
76
[227] How ICASA considered the countervailing power apparent from the new
roaming agreements and on a ‘look forward’ basis is not clear. It merely stated
that:66
‘It is possible though that this countervailing power will be limited once again as spectrum
constraints are lifted, and so the market may change. The Authority notes that countervailing
power has not been sufficient in the past to constrain MTN and V odacom’s pricing and terms
and conditions.’
[228] Second, e ven if MTN and V odacom’s market share was 50 -50, the
determination is not one that could be made on the combined market share basis.
ICASA was obliged to consider the extent of competition between MTN and
V odacom, and whether they pose competitive constraints on each other in the
roaming market. The determination lacks a proper analysis of the national
roaming market shares of MNOs relative to one another and the competitive
dynamics and constraints in the roaming market. ICASA's reasoning reveals that
it did not consider all the mandatory factors in s 67(4A) of the ECA in determining
the effectiveness of competition in the roaming market.
[229] Although s 67(4A) of the ECA permits ICASA to consider other relevant
factors, it could only do so after considering the mandatory factors. Even then,
those other factors must be clearly articulated, objectively verifiable and justified.
[230] Based on the facts above , the determination is not only unreasonable, but
irrational. It is not one that could have been made if ICASA had proper regard to
the relevant considerations and information before it. The latter conclusion is
fortified by ICASA’s acknowledgement that there had been competitive
outcomes, which contradicts the conclusion that there is ineffective competition.
66 Findings document para 189.
77
The finding is irregular. It is not one made within the contemplated provisions
and falls to be set aside.
MTN has significant market power in the national roaming market.
[231] The regulation 6(c) states:
‘Upstream market 2: MTN and V odacom are dominant and have SMP in the market for
wholesale national roaming since there are only two operators that provide this service for
coverage purposes in South Africa. MTN and V odacom also have SMP as a result of vertical
relationships that could harm competition.’
[232] ICASA’s first linked the determination of significant market power with
the market for site access, dealt with above. It stated in the discussion document
that with regards to roaming, from a network capacity perspective, measured by
number of network sites, MTN was dominant ( ie has a market share of 45% or
more) in 34 local and metropolitan municipalities. V odacom is dominant in 86,
and MTN and V odacom both have a market share exceeding 45% in 15
municipalities. ICASA appears to have abandoned this line of justification of the
finding of MTN’s significant market power. The reason does not feature in the
findings document.
[233] Instead, ICASA based the determination of significant market power on
three considerations, namely : first, the 99% coverage by MTN and V odacom,
second, the finding that MTN and V odacom have more than 45% market share and
third, the existence of a vertical relationship which could harm competition.
[234] The determination of significant market power in the roaming market need
not detain the court. It is perforated by similar legal flaws found in respect of the
determination of significant market power in the site access infrastructure market.
78
It bears emphasis that the determination of significant market power is statutorily
prescribed in s 67(5) of the ECA. It requires:
(a) An establishment of ‘ dominance’ as defined in the Competition Act as
prerequisite.
(b) ‘Dominance’ is determined with reference to the market share.
(c) The market share of 45% attributable to MTN , relied on to find it had
significant market power (dealt with in respect of the effectiveness of the
competition) is irregular for the reasons already stated.
(d) The flaw, evident from the failure to determine MTN’s roaming market
share separately from that of V odacom’s market share , affects the inquiry and
finding whether in fact MTN is ‘dominant’ in the roaming market.
[235] The second reason for the finding of significant market power is that MTN
and V odacom have ‘vertical relationships that could harm competition ’. It will
be recalled that ICASA made an overarching finding that
‘MTN and V odacom are both vertically integrated since they operate downstream in offering
retail services as well as upstream, having been assigned spectrum, operating their own high
sites and offering roaming services.’
[236] ICASA found a strong correlation between the level of concentration of
ownership of mobile sites and retail customers and between site market shares
and customer market shares which, ‘suggests that there is likely a strong link
between market power at the wholesale and retail levels ’. It noted that the
evidence of the extent of the vertical integration is harmful to competition as a
result of the limited sharing of the infrastructure in South Africa and the very
high costs of roaming. (My emphasis.)
[237] It is clear ex facie from the regulation that ICASA conflated the legal
meaning of a ‘vertical relationship’ and ‘vertical integration’. I need not repeat
79
the legal meaning of the terms, already referred to in the context of the site access
market as discussed above. The determination of significant market power is not
only unreasonable but also influenced by the same error of law. It is arrived at in
a manner that is inconsistent with the ECA , the Competition Act and a s
contemplated within the meaning of s 6(2)(d) of PAJA. For this reason alone, the
finding of significant market power in the national roaming market is unlawful
and irregular and falls to be set aside.
Is regulation 7 and pro-competitive conditions imposed ultra vires?
[238] Following ICASA’s findings that there is ineffective market competition,
it published regulation 767 and imposed, pro-competitive conditions on MNO’s
with significant market power in the markets discussed above. A contravention of
the pro-competitive conditions attracts a fine not exceeding R 5 million.
[239] MTN challenges regulation 7 on several grounds in terms of PAJA and
alleges that they are in violation of the requirements of just administrative action
(as guaranteed by section 33 of the Constitution and given effect to by PAJA) and
s 14 of the Constitutio n. Although MTN listed numerous grounds in opposition
to regulation 7, the primary complaint was that the imposition of the regulations
was ultra vires of the ECA.
[240] Regulation 7(a), (b), (c), and (d), which is in respect of the retail market ,
requires a publication of quarterly reports on MNO websites. These reports must
provide the following:
‘(a) A report and supporting data on effective retail prices paid by end user customers for
data services overall, calculated by dividing total revenue for data with total volume of data
used (in Gigabytes) over the quarter.
67 As set out above, regulation 7 was amended by Mobile Broadband Services Amendment Regulations, 2024,
GN 2617 in GG 50910 of 5 July 2024, wherein sub -regulations 7 (e) and (f), were deleted, and parts of sub -
regulation 7(h) were substituted/amended.
80
(b) A report and supporting data on effective retail prices paid by end user customer
category calculated by dividing total revenue for data with total volume of data used (in
Gigabytes) over the quarter for each of the following categories:
(i) By prepaid, hybrid and postpaid customer segments;
(ii) By consumer and business customer segments;
(iii) Data used between 5am and 12am midnight and data used from 12am midnight
to 5am; and
(iv) By province, and within provinces, and by urban and rural areas, as defined by
the Authority.
(c) Data revenue should exclude fixed-wireless data traffic, wholesale data traffic, mobile
virtual network operator data traffic, and enterprise business traffic.
(d) All retail tariffs available to customers over the quarter’.
[241]Regulation 7(f) deals with the disclosure of retail pricing, and states that:
‘Furthermore, if any category of retail price is below any wholesale price the operator with
SMP is required to submit an explanation for the differential and fully auditable evidence to
the Authority, with all assumptions clearly specified, showing that this differential is cost based
or temporary or is economically or technically justifiable on other grounds.’
Regulation 7(f) was design to monitor ‘margin squeeze’ in retail and wholesale
prices.
[242] It claims regulation 7 (f) is concerned with ‘margin squeeze ’, namely
whether a retail price is below a wholesale price. It is directed at remedying
market ‘information asymmetry’. The amended regulation retains the obligation
to publish the information on websites but limits the publication to ‘non -
confidential versions’ of the records listed. MNOs must submit the confidential
information to ICASA. The objective to make prices less ‘opaque’ is not the
purpose for which the Regulations may be imposed.
81
[243] Regulation 7(h) regulates the wholesale site infrastructure access market
and applies where an MNO owns a site or controls access to it. An MNO must,
in terms of the amended regulation, provide:
‘(i) A list of sites approved for access within twenty (20) business days of the initial request
during the previous quarter, together with the access seeker's name, date of request, date of
approval, and all charges, whether recurring or non-recurring, for access to the site;
(ii) A list of sites not approved for access within twenty (20) business days of the initial
request during the previous quarter, together with the access seeker's name, date of request, and
reason for not approving it;
(iii) A report on the previous quarter's site access requests summarising the information in
regulations 7(h)(i) and 7(h)(ii) above, including a summary of time to approve the requests, a
summary of reasons for not approving site access requests, and average effective charges for
the sites shared.
(iv) An updated list of all sites used by the SMP operator, and all charges for sharing any
site infrastructure owned or controlled by the SMP operator.
(v) In respect of information provided per site, the licensee must also provide the operator's
identification code for the site, its longitude and latitude, and Statistics South Africa census
2011 main place code, and site category including macro >15m, macro <15m, rooftop, indoor
(including distributed antennae systems).’
[244] In respect of the wholesale roaming services market, regulation 7 (i)
requires that MNOs are to provide:
‘1.2. A report and supporting data on effective prices paid for wholesale roaming services by
each roaming customer calculated by dividing the total roaming revenue and data roaming
volumes, split by:
1.2.1. Each roaming contract; and
1.2.2. Any contractual price variations used (e.g. metro and non-metro).
1.3. A report and supporting data on wholesale national roaming data volumes used by site,
1.3. A report and supporting data on wholesale national roaming data volumes used by site,
together with details of that site including at least the operator’s identification code for the site,
longitude and latitude, and Statistics South Africa census 2011 main place code.’
82
[245] The complaint here is that regulation 7(a) to 7 (h), which imposes pro -
competitive conditions, should be set aside because:
(a) The markets they purport to regulate or the determination that the markets
are ineffectively competitive is unlawful.
(b) The requirement to publish the listed information on the applicant ’s
website is inappropriate and disproportionate, not aimed at remedying any market
failure identified; no r is i t rationally related to the reasons given for it , and
constitutes an unconstitutional infringement of the rights to privacy of MTN and
its wholesale customers.
(c) It is over-broad and not rationally connected with the regional/provincial
markets which ICASA had defined. On ICASA’s market determination, the
question that ar ises is: If retail markets are regional w hy would a subscriber in
Gauteng be interested in pricing information that pertains to a subscriber in rural
Eastern Cape? What would be the harm that ICASA seeks to remedy in relation
to those subscribers? This much is not clear.
[246] Notwithstanding the amendment, R egulation 7(h)(iii) is ultra vires. The
purpose of regulation 7 (h)(iii) is not connected with the information before
ICASA, particularly the effect of the disposal of the site portfolios by MNOs on
their significant market power (dealt with in paragraph 184 above), a prerequisite
to market regulation and the evidence that most sites are owned by the non -
licences.
[247] The regulation of macro-sites in regulation 7(h)(v) is similarly tainted the
irrational exclusion of micro sites from the product market for site access ,
discussed in paras 162 to 168 above.
83
[248] As I have endeavoured to show throughout the judgment, the source of
ICASA’s power to monitor MNO’s and to impose pro-competitive conditions is
in s 67(4)(d) and (f) of the ECA.
[249] As set out above, and in summary here: The power to impose pro -
competitive condition flows from s 67(4)(d) of the ECA. It is linked inextricably
with the determinations ICASA made, which I have found irregular and falls to
be set aside. It is evident from s 67(4)(d) that ICASA’s powers to remedy market
failures through an imposition of pro -competitive conditions in the identified
markets must have been based on a finding that there is ineffective competition
within these markets, and in respect of an MNOs with significant market power.
[250] Section 67(4) enjoins ICASA to impose conditions that a re ‘appropriate
and sufficient’. ICASA relied on the range of non -exhaustive remedies it may
impose in s 67(7). ICASA submitted that I should view the conditions in the
context of ICASA’s duty to monitor MNOs in s 67(4B) of the ECA, but also in
relation to its statutory object to promote competition. It has the statutory power
to seek ‘any information specified’ by it from MNOs in terms of s 67(4B) of the
ECA.68 Section 67(7) (c) of the ECA, permits it to impose conditions , which
include ‘obligations to publish any information specified by the Authority in the
manner specified by it’. ICASA contended that the publication of information of
MNOs on websites is justified and relates to non -confidential information . It
seeks to monitor participants and to remedy market ‘information asymmetry’.
[251] Accordingly, the pro-competitive conditions must be aimed at remedying
a market failure . There must be a rational connection between the remedy
employed and the ability to remedy a market failure. The remedies imposed were
68 Subject to s 4D of the ICASA Act, licensees must provide to the Authority any information specified by the
Authority in order that the Authority may carry out its duties in terms of this section.
84
proportionate to the harm, which is the extent of the ineffectiveness of the
competition.
[252] I have dealt with the irregular market definition and the consequential flaw
on the determinations ICASA was required to make under the ECA. I have in
addition pointed out the irregularities in determining the ineffectiveness of the
competition and significant market power, including the error of law occasioned
by the conflation of a ‘vertical relationship ’ with ‘vertical integration ’ in the
identified wholesale markets.
[253] Even if it is found that I have reached an incorrect conclusion in finding
that ICASA’s market definition is flawed in respect of regulations 3(a), (b) and
(c), s 67(7) of the ECA leaves no room for doubt that a finding of (a) ineffective
competition and ( b) significant market power is statutory prerequisite for
imposing pro-competitive conditions.
[254] ICASA accepted the criticism of its failure to compute the relative market
shares, although this was in relation to the computation of the market shares in
the roaming market. It contended this could be rectified by re -computing the
market shares. As already shown, this shortcoming is not confined to the roaming
market.
[255] The obvious disconnect is that the conditions would apply in markets
where it has been shown MTN does not have significant market power or the
significant market power is irregularly determined across the defined markets.
Although ICASA amended the regulation, the amendments do not rectify the
irregularities identified. The statutory trigger for imposing the pro -competitive
conditions under s 67(7) was simply not met. The finding is dispositive and
supports why the pro-competitive conditions cannot stand. MTN’s complaint that
85
the regulations are ultra vires and unlawful succeeds. Regulation 7 falls to be
reviewed and set aside in its entirety.
[256] Much debate ensued about the appropriateness of the requirement to
publish pricing practices and information on websites, which permeates all the
identified markets, and whether the means employed (i .e. publication of
information on websites) justify the end, (i.e. remedying market failure) and is
proportionate to the mischief sought to be remedied. ICASA agreed that its power
is not unfettered. It accepted, based on the court’s decision in Economic Freedom
Fighters v Speaker, National Assembly and Others, 69 that ‘proportionate’ means:
‘“Appropriate” means nothing less than effective, suitable, proper or fitting to redress or undo
the prejudice, impropriety, unlawful enrichment or corruption, in a particular case.’
[257] I accept that the requirement is in respect of non-confidential information.
ICASA already has the power to seek a publication of information specified by it
in terms of s 67(5)(g) of the ECA to monitor the market. However, in the present
instance, the information required is not connected to remedying lawfully
identified market failures.
Procedural unfairness
[258] The procedural unfairness complaints already alluded to above are on all
fours with the facts in Earthlife. There the court held that:
‘[52] Fairness ordinarily requires that an interested party be given access to relevant material
and information in order to make meaningful representations. De Smith Woolf & Jowell
summarise the principle as follows:
“If relevant evidential material is not disclosed at all to a party who is potentially prejudiced
by it, there is prima facie unfairness, irrespective of whether the material in question arose
before, during or after the hearing.”
69 Economic Freedom Fighters v Speaker, National Assembly and Others [2016] ZACC 11; 2016 (3) SA 580 (CC)
para 71.
86
[53] On the other hand, however, it has been emphasised repeatedly that an interested party's
right to disclosure of “relevant evidential material” is not equivalent to a right to com plete
discovery, as this could “over-judicialise” the administrative process. “The right to know is not
to be equated to the right to be given ‘chapter and verse’.” What is required in order to give
effect to the right to a fair hearing is that the interested party must be placed in a position to
present and controvert evidence in a meaningful way. In order to do so, the aggrieved party
should know the “gist” or substance of the case that it has to meet.’
In Minister of Education, Western Cape and Another v Beauvallon Secondary
School and Others,70 it was held:
‘[I]t must be remembered that although the fairness of any procedure followed will depend on
the circumstances of each particular case, a person affected by a decision usually cannot make
meaningful representations without knowing what factors are likely to be taken into account.
Accordingly, in a test regularly approved by this court, “fairness will very often require that he
is informed of the gist of the case which he has to answer”.’
[259] It bears noting that the rule 53 record alone spans 3 783 pages (this
excludes the affidavits, which are over 500 pages, and other documents filed) and
excludes 18 volumes of confidential material. The record contains no evidence of
the minutes and the transcripts of ICASA's deliberations about the regulations.
Material documents relevant to its decision making process are reported to have
been deleted. ICASA’s approach was that the decision document, the findings
document and the reasons document are the only evidence constituting proof of
its evidence and reasoning. Given the change, it meant MTN had no ‘knowledge
in advance of the considerations’ on which a decision will be based.71
[260] ICASA’s reasons document treats the representations made to it in a
[260] ICASA’s reasons document treats the representations made to it in a
perfunctory manner. It states that the comments from MTN and V odacom relate
70 Minister of Education, Western Cape and Another v Beauvallon Secondary School and Others [2014] ZASCA
218; 2015 (2) SA 154 (SCA) para 19, the court quoted the House of Lords decision in Doody v Secretary of State
for the Home Department and Other Appeals [1994] 1 AC 531 (HL) ([1993] 3 All ER 92) at 106 b – h (All ER)
with approval.
71 Chairman, Board on Tariffs and Trade v Brenco Inc 2001 (4) SA 511 (SCA) at para 13.
87
to matters already considered during the inquiry and dealt with in the findings
document. Earthlife stresses that proper compliance with procedural fairness
requires ‘that the interested party must be placed in a position to present and
controvert evidence in a meaningful way’.72
[261] Although MTN makes a cogent case on this ground, given the finding that
the Regulations are jus ticiable and fall to be set aside, it is not necessary to
determine this issue on this basis. Nevertheless, it has a material bearing on the
appropriate remedy the Court may grant.
Declaration of invalidity
[262] For the reasons stated in the judgment, it is declared that:
(a) In respect of the retail markets that –
(i) ICASA's determination in regulation 3(a), that retail markets are
regional and/or provincial , split by urban and rural, is flawed,
arbitrary, and unlawful. ICASA did not determine the market in
terms of the methodology it had prescribed in the regulations and
failed to follow its own guidelines. It failed to provide reasons for
the geographic market determination. The regulation is reviewed and
set aside.
(ii) The determination in regulation 5 that the retail market defined in
regulation 3(a) is ineffectively competitive is unlawful and irregular.
It is based on a flawed and irregular market definition. In addition,
ICASA failed to consider the mandatory factors in s 67(4A) of the
ECA and failed to consider relevant considerations . The
determination is reviewed and set aside.
72 Earthlife para 53.
88
(iii) The finding in Regulation 6(a) that MTN is dominant and has
significant market power in the retail market is unreasonable and
unlawful. It is tainted by the flawed determination of the ineffective
competition within the retail market and is based on outdated market
data, and irrelevant consideration. ICASA failed to take into account
relevant considerations. The Regulation is influenced by a material
error of law ; ICASA conflated a vertical relationship with vertical
integration. The determination is reviewed and set aside.
(b) In respect of the wholesale site access market that:
(i) Regulation 3 (b) defining the upstream ma rket as wholesale site
infrastructure geographic market as local and municipal is flawed
and is not based on the methodology ICASA had prescribed.
(ii) The product market irregularly excludes other sites from the product
market based on ‘equivalent coverage ’ a different test than that
applied to market definition. ICASA failed to take account of
representations made to it. The Regulation 3(b) is reviewed and set
aside.
(iii) Regulation 5 determining the ineffectiveness of competition in the
site access market is reviewed and set aside . ICASA failed to take
into account the mandatory factors prescribed in s 67(4A) of the
ECA, being the individual market shares and the relative market
shares of operators , including information placed before it , the
disposal of site portfolios being such relevant information. The
regulation is not rationally connected to the purpose of the
empowering provision in the ECA, or the information placed before
ICASA within the meaning of s 6(2)(f)(ii) of PAJA. The Regulation
is reviewed and set aside.
(iv) Regulation 6 determining significant market power is irregular, it is
influenced by unlawfully defined markets and a material error of law.
89
There is no rational basis for the finding of dominance and
significant market power. The Regulation is reviewed and set aside.
(c) In respect of the wholesale national roaming market for coverage that:
(i) Regulations 3(c) defining the roaming market is flawed . It
irregularly excluded roaming for capacity. ICASA failed to apply the
methodology it had prescribe d to define the market, and definition
was not rationally connected to the information before ICASA
within the meaning of s 6(2) (f)(ii) of PAJA. The Regulation is
reviewed and set aside.
(ii) Regulation 5 in respect of upstream market 2 that is defined as
‘wholesale national roaming market for coverage purposes’ is
ineffectively competitive, and it is tainted by the flawed irregularly
determined product market. It is unreasonable and based on
irrelevant considerations. ICASA failed to take account the
mandatory factors prescribed in s 67(4A) of the ECA . The
Regulation is reviewed and set aside.
(iii) Regulation 6 (c) determining significant market power is
unreasonable and is permeated by a material error of law. There is
no logical connection between an ‘admission’ of competition
between MTN and V odacom and the justification for ICASA ’s
findings of ineffective competition and significant market power. It
is reviewed and set aside.
(d) With regards to the pro-competitive licence c onditions imposed in
regulation 7, it is declared in respect of Regulation 7(a), (b), (c) and (d) that:
(i) ICASA could only impose pro -competitive conditions in lawfully
defined m arkets. ICASA has no power in law to impose pro-
competitive conditions on MTN in relation to the geographic
markets in which MTN does not have significant market power.
90
(ii) ICASA was required to show that it met the mandatory jurisdictional
requirements and lawfully determined ineffective competition and
significant market power to impose the conditions contemplated in
terms of s 67(4)(d) of the ECA. It failed to do so. For these reasons
alone, the Regulation is reviewed and set aside.
(iii) The determination of ineffective competition and significant market
power in the retail, site access and roaming markets were influenced
by a material error of law. ICASA’s regulatory power was not
lawfully triggered. The pro-competitive conditions are ultra vires the
ECA. The amendments and substitution do not cure the irregularities
found. The Regulation is reviewed and set aside.
(e) Regulation 7(f) was concerned with ‘margin squeeze’, namely whether a
retail price is below a wholesale price.
(i) it is not competent for ICASA to impose conditions where there is
no findings of ineffective competition and significant market power
in the retail market.
(ii) there was no evidence of exclusionary conduct by either MTN or
other MNOs.
(iii) Its intervention is not justified, and the pro -competitive conditions
are ultra vires the ECA.
(f) In the result, regulations 7(a) to (d) and (f) are irregular and ultra vires and
they are not rationally related to the purpose of the empowering provision, or the
reasons given by ICASA.
(g) Regulation 7(h)(i) to (v) were amended to clarify that the conditions apply
only to markets where an operator has significant market power as said, th e
amendments do not alter the dispute about their lawfulness, or the finding that the
material error of law in determining significant market power renders the
regulation unlawful and ultra vires. The Regulations are reviewed and set aside.
91
The remedy
[263] As already alluded to, V odacom opposed the review on a limited basis in
respect of the qualified relief limiting the outcome of the review to MTN, if MTN
succeeded to set aside the Regulations . In prayers 1.3 and 1.4 of the notice of
motion sought an order to set aside the regulations ‘to the extent’ that they applied
to it. V odacom plays no part in the merits of the review. V odacom contends the
relief sought in prayer 1.3 of the notice of motion is unduly narrow. The excision
of MTN from the application of Regulations, should the grounds for the review
be upheld, do not permit an excision of V odacom or other MNOs from the relief.73
[264] It was evident during argument that an appropriate remedy hinged on the
basis on which the court would determine the review. MTN and V odacom
accepted that if the court found the regulations were unlawful and irregular, and
fall to be set aside, there can be no legal basis to excise V odacom or other MNOs
from the finding as it vindicates the rule of law. MTN abandoned the qualification
to its relief. ICASA accepted too, that in the event of such a finding, it would be
in the court’s sphere to declare the regulations invalid and set them aside. The
debate turned to the consequences of a declaration of invalidity.
[265] Section 8(1) of PAJA gives effect to the wide remedial discretion conferred
on the courts by s 172 of the Constitution.74 It confers the courts with wide powers
to grant ‘any order that is just and equitable’. In refining the nature of the powers
conferred on the courts , t he Constitutional Court in Allpay Consolidated
Investment Holdings (Pty) Ltd and Others v Chief Executive Officer, South
73 Section 67(8)(b) of the ECA, provides that,
‘Where, on the basis of a review under this subsection, the Authority determines that a licensee to whom any pro-
competitive conditions apply is no longer a licensee possessing significant market power in that market or market
segment, the Authority must revoke the applicable pro-competitive conditions applied to that licensee by reference
to the previous market determination based on earlier analysis’.
74 Bengwenyama Minerals (Pty) Ltd and Others v Genorah Resources (Pty) Ltd and Others [2010] ZACC 26;
2011 (4) SA 113 (CC) paras 82-83.
92
African Social Security Agency and Others 75 clarified that the extent of the
powers are wide flexible remedial powers , which are without self -censor, to
fashion a remedy which is appropriate to the circumstances of the case . It
highlighted the multi -dimensional approach necessary to formulate a ‘just and
equitable’ remedy. The court in Central Energy Fund SOC Ltd and Another v
Venus Rays Trade (Pty) Ltd and Others, 76 with reference to the decisions of the
Constitutional Court referred to above, confirmed the remit of the court’s powers
and that the range of remedies provided is not intended to be a closed list.
[266] ICASA advocates for a suspension of the order declaring the Regulation
invalid for a period of 12 months to grant it an opportunity to cure the defects.
The consequence of the suspension of the declaration of invalidity means a
retention of the status quo ante and the Regulation will be in force. By implication
that means an order remitting the Regulations to ICASA to do so. A remittal is
one of the orders the court may grant with or without directions in s 8(1) (c)(i).
V odacom joins issue with ICASA and supports an order suspending the order of
invalidity and a remittal of the Regulations to ICASA. MTN will be required to
comply with the Regulations for the period of suspension.
[267] ICASA’s approach was advanced by the same parties in Mobile Telephone
Networks (Pty) Ltd v Chairperson of the Independent Communications Authority
of South Africa and Others; Vodacom (Pty) Ltd v Chairperson of the Independent
Communications Authority of South Africa and Others ,77 in a dispute about the
validity of Call Termination Regulations. ICASA contended, as it did then that a
failure to suspend the declaration of invalidity would create ‘an unwarranted
75 Allpay Consolidated Investment Holdings (Pty) Ltd and Others v Chief Executive Officer, South African Social
Security Agency and Others [2014] ZACC 12; 2014 (4) SA 179 (CC) (Allpay 2) para 30.
Security Agency and Others [2014] ZACC 12; 2014 (4) SA 179 (CC) (Allpay 2) para 30.
76 Central Energy Fund SOC Ltd and Another v Venus Rays Trade (Pty) Ltd and Others [2022] ZASCA 54; [2022]
2 All SA 626 (SCA) para 37-38.
77 Mobile Telephone Networks (Pty) Ltd v Chairperson of the Independent Communications Authority of South
Africa and Others; Vodacom (Pty) Ltd v Chairperson of the Independent Communications Authority of South
Africa and Others [2014] 3 All SA 171 (GJ) para 109.
93
regulatory lacuna’ if not granted. If the Regulations are set aside with immediate
effect ‘public interest will suffer’.
[268] The court in Mobile Telephone Networks (Pty) Ltd v Chairperson of the
Independent Communications Authority of South Africa and Others; Vodacom
(Pty) Ltd v Chairperson of the Independent Communications Authority of South
Africa and Others with reference to the judgment of Cameron J in Estate Agency
Affairs Board v Auction Alliance (Pty) Ltd and Others 78 dealt with the need to
balance a range of interests and consider, the interests of the successful litigant in
obtaining immediate constitutional relief, on the one hand , and the potential
disruption of the administration of justice that would be caused by the lacuna on
the other. That court confirmed that it is thus enjoined to consider all the interests
implicated and the impact of the irregularity. However, of importance, the remedy
must be appropriate and equitable in the circumstances of the case, fair to all
parties affected by it, and yet effectively vindicate the rights violated.79
[269] These principles call for a pragmatic approach which must be informed by
the facts, the circumstances of each case, where relevant, and the conduct of the
parties. The Regulations were promulgated on 31 March 2022. It was not disputed
that MTN has not complied with the conditions. ICASA suggested MTN took the
law in its hands . There is no evidence that ICASA had sought to enforce the
Regulations. On the contrary, MTN launched this review during September 2022,
within the prescribed period of 180 days.
78 Estate Agency Affairs Board v Auction Alliance (Pty) Ltd and Others [2014] ZACC 3; 2014 (3) SA 106 (CC).
79 Allpay 2 para 71; see also Steenkamp NO v Provincial Tender Board, Eastern Cape [2006] ZACC 16; 2007 (3)
SA 121 (CC) para 29.
94
[270] ICASA should only regulate a market when a regulation thereof is justified
by properly substantiated analysis, and it has assessed the evidence of a market
failure. The difficulties with the suspension and remittal order are that:
(a) Two years and three months after the promulgation, and approximately a
year and some 10 months after the launch of the review, ICASA published the
amended Regulations on 5 July 2024 , which it claims, ‘were necessitated by
regulatory oversight ie clerical errors, incorrect referencing, etc.) identified in the
Regulations.’
(b) A confirmatory af fidavit by ICASA’s Senior Manager Legal (Legal,
Contracts, and General Legal) indicates that the record of deliberations
comprising of the minutes of the meetings of the committee and the transcript of
those minutes was deleted erroneously. ICASA is of the view that the discussion
document, findings document and reasons document stand as prima facie proof
of the deliberations. There is conflicting evidence on whether in fact they exist.
(c) At the hearing, ICASA conceded there may be difficulties with the finding
of ineffective competition, albeit the concession was confi ned to the finding of
ineffective competition in the roaming market. This concession permeates the
finding of significant market power it sought to correct through the amendment.
The error is endemic in all the identified markets.
(d) Even when considering the amended Regulation, ICASA failed to seriously
consider material data presented by MTN in its analysis, or consider, engage with
or seeks input from the Commission, a specialist body well versed with
competition matters.
(e) The impression made on the court, is that the amendments were a mere
‘tweak’, thus superficial and wholly fail to address the fundamental flaws and
errors of law in the analysis of the market data, the premise and the findings on
which the Regulations are based. This leaves little to salvage by way of a
correction which would justify a remittal.
correction which would justify a remittal.
(f) Section 67(8)(b) of the ECA states that if pursuant to a review it is found
95
that a licensee no longer has the market power in that market or market segment,
then ICASA must revoke the conditions in so far as they apply to that licensee.
There is no indication on the papers that it availed to this provision.
[271] Often, the courts have evaluated the State ’s blameworthiness at various
stages of the procurement process: from when the impugned administrative
decision-making takes place , at the time of initiating review proceedings;
and, during litigation proceedings.80 Even though considered in the context of
procurement, those factors have a bearing in determining a remedy, especially
where an administ rator is entrusted with an important duty to exercise a
Regulatory function. On the facts of the present case, public interest is thwarted
by an improper and unlawful exercise of Regulatory power of a far reaching
important sector.
[272] For the reasons set out above, the fast-paced technology driven nature of the
mobile telecommunications market and sector, the evidence of its dynamic nature
and the shifts that have already occurred at the time of the promulgation of the
Regulations in the market shares in the identified markets, a remittal order will not
only be inappropriate, it will be an unjust and inequitable remedy given the extent
of the deficiencies, unlawful premise and irregularities found.
Costs
[273] What remains is the question of costs. MTN seeks a cost order against
ICASA should the review succeed. Given the complexity of the issue raised, it is
of the view that costs at scale C is warranted. However, it does not seek a cost
order against V odacom.
80 Allpay 2.
96
[274] On the other hand, ICASA contends even if MTN and V odacom succeeded
on one ground of review, it is not appropriate in the circumstances of this case for
this court to make an adverse costs award against ICASA, as a statutory body
vested with the responsibility of carrying out certain functions in the public
interest. It proposes that each party pays its own costs, on the basis that it is just
and equitable to do so.
[275] With regards to V odacom’s costs, they are limited, and consistent with the
limited grounds of its opposition. Even though a declaration of invalidity of the
Regulations would affect all MNOs, I accept that it would not have been able to
predict how argument on those issues would unfold. It is fair that V odacom should
bear its own costs.
[276] What persuades the court to grant a cost order against ICASA in this
instance is the way it approached its market regulation task . Its purported
amendments leave an impression that it did not wish to reconsider its stance with
an open mind, despite the time that elapsed since the original Regulations.
[277] ICASA failed to consider the representations by MTN , the c hange in
market information and the disjuncture between its findings and those of the
Commission. T he public interest, time and resources would have been better
served had it done so. In the result, costs must follow the result. There is no
justifiable reason to deprive MTN its costs, which including costs consequent on
employment of two counsel.
Order
[278] In the result, I make the following order:
1 The Regulations are declared irregular, unlawful, and invalid as follows:
1.1 In respect of Market Definition:
97
(a) Regulation 3 (a), defining mobile retail services in regional
geographic areas (provincial, split by urban and rural),
(b) Regulation 3(b) defining Upstream market 1 : wholesale
infrastructure access in local and metropolitan municipalities,
(c) Regulation 3 (c) defining Upstream market 2 : wholesale
national roaming market for coverage purposes,
are reviewed and set aside.
1.2 In respect of Effectiveness of Competition, regulation 5, to the extent
that it determines ‘ the retail market ’, ‘Upstream market 1 ’ and
‘Upstream market 2 ’ defined in regulation 3 are ineffectively
competitive is reviewed and set aside.
1.3 In respect of Significant Market Power Determination:
(a) Regulation 6 (a) determining that MTN and V odacom are
dominant with the market shares stated thus have significant
market power and are in a vertical relationship that could harm
competition is reviewed and set aside.
(b) Regulation 6 (b) providing for the Upstream market 1
determining MTN and V odacom are dominant the market
shares stated thus have significant market power as a result of
a vertical relationship that could harm competition is reviewed
and set aside.
(c) Regulation 6 (c) providing for the Upstream market 2
determining MTN and V odacom are dominant the market
shares stated thus have significant market power as a result of
a vertical relationship that could harm competition is reviewed
and set aside.
1.4 In respect of Pro-Competitive Terms and Conditions:
(a) Regulations 7(a), (b), (c), (d) and (f) are reviewed and set
aside;
98
(b) Regulations 7(h)(i), (ii) and (iii ) as substituted together with
regulation 7(h)(iv) and (v) are reviewed and set aside;
(c) Regulations 7(i) and (j) are reviewed and set aside.
2 The orders in paragraphs 1.1 to 1.4 operate with immediate effect.
3 The first respondent is ordered to pay cost of the Applicant at scale C.
4 The costs include the costs occasioned by employment of two Counsel.
5 The Second Respondent shall bear its own costs.
_____________ ___
NTY SIWENDU J
JUDGE OF THE HIGH COURT
GAUTENG DIVISION, PRETORIA
Delivered: This judgment was handed down electronically by circulation to the parties’ legal
representatives by e -mail. The date and time for hand -down is deemed to be on the 19
September 2025.
Hearing Dates: 18 March 2025 to 20 March 2025
Request for Limited Submissions: 29 May 2025
Appearances:
For the Applicant: A. Cockrell SC
With him: M. Mbikiwa
Instructed by: Webber Wentzel
C/O Hills Incorporated
For the First Respondent N. Maenetje SC
With him: Katlego Monareng
99
Instructed by: Mashiane Moodley & Monana Inc
For the Second Respondent F Snyckers SC
With him: Lerato Zikalala
Instructed by: Cliffe Dekker Hofmeyr