Vodacom And Business Venture Investments No 2213

The judgment

The court's own file

1 COMPETITION TRIBUNAL OF SOUTH AFRICA Case No: LM148Dec 21 In the matter between: Vodacom Proprietary Limited Primary Acquiring Firm And Business Venture Investments No 2213 Proprietary Limited Primary Target Firm Panel: A Wessels (Presiding Member) T Vilakazi (Tribunal Member) A Kessery (Tribunal Member) Heard on: 20  24 May 2024; 27  28 May 2024; 30 May 2024; 13  14 June 2024; 18  19 June 2024; 15  16 July 2024; 18  19 July 2024; 29  31 July 2024; 1  2 August 2024; 7 August 2024; 28 August 2024; 9 September 2024; 26  27 September 2024 Last submission on: 16 October 2024 Decided on: 29 October 2024 REASONS FOR DECISION PROHIBITION 1. On 29 October 2024, the Competition Tribunal (Tribunal) prohibited the proposed transaction in terms of which    Vodacomintends to acquire 30%, and potentially 40%, of the issued share capital of Maziv (Pty)  Maziv, previously called Business Venture Investments No 2213 (Pty) Ltd 1 the proposed transactionVodacom and Maziv collectively are referred      the merger parties 1             Non-Confidential 2 2. The proposed transaction involves, inter alia, the largest Mobile Network  MNO    , and the largest dark fibre provider in South Africa,      DFA, as well as the largest fibre    FTTH    FNO   Vumatel   Vumatel. 3.          4. We note that it is common cause that the proposed transaction given that it relates to access to the internet/data and its (future) pricing, is of great significance to millions of South African consumers. The Commission   that the proposed transaction raises both horizontal and vertical competition concerns and ultimately negatively affects South African consumers. The merger parties disagree but nevertheless tender mostly behavioural conditions mainly for the vertical concerns (not for the horizontal concerns) and tender fibre roll-out and other public interest commitments, that we assess under the public interest. 5. We shall consider the effects of the proposed transaction particularly on low- income consumers, including future access to products and services through the roll-out of fibre to support the provision of internet into lower income areas and the effects of the proposed transaction on the future costs of those products and services. Various relevant counterfactual(s) will be a key consideration in this assessment. MERGER PARTIES 6. The primary acquiring firm is Vodacom. Vodacom is ultimately controlled by    Vodacom Group       Johannesburg Stock Exchange. Vodacom Group is the sole shareholder of Vodacom. The issued share capital in Vodacom Group is held as follows: •        Vodafone Investments  52.68%; Non-Confidential 3 • the public  18.98%; • Government Employees Pension Fund  14.30%; •     Vodafone International  7.81%; and •       Yebo Yethu  6.23%. 7. The 52.68% shareholding held by Vodafone Investments is a controlling interest. Vodafone Investments and Vodafone International are controlled by Vodafone   Vodafone          8. The primary target firm is Business Venture Investments No 2213 (Pty) Ltd, renamed Maziv. Maziv is a wholly owned subsidiary of Community Investment Ventures Holdings (Pty) Ltd (CIVH). The main operating subsidiaries of CIVH are DFA and Vumatel. 9. Of relevance is that Vumatel currently has a % shareholding in Hero    Herotel        Herotel from the current % interest to a greater than 50% interest. A merger was filed     Commissionon 30 June 2022 in terms of which Vumatel intends to own % of the issued share capital of Herotel. At the time of the hearing, the Commission was still investigating this proposed transaction. 10.           IEI (57%), ( %) and New GX Fund I ( %). IEI is controlled by VenFin (Pty) Ltd VenFin(%), and VenFin        Remgro     controlled by any specific firm. 11. In addition to CIVH, Remgro has an interest in various firms involved in several industries. In the telecommunications industry, Remgro has a non-controlling interest in Seacom South Africa (Pty) Ltd. Remgro has recently sold its indirect interest in fibre operators Octotel (Pty) Ltd and RSAWeb (Pty) Ltd. Non-Confidential 4 PROPOSED TRANSACTION 12. In terms of the proposed transaction Vodacom will acquire a shareholding in Maziv by subscribing for shares, acquiring shares and selling assets to Maziv. Vodacom will first achieve a 30% shareholding. Vodacom will then have the option to subscribe for additional shares for cash in terms of the top-up mechanism set out in the Shareholders Agreement, which would increase its shareholding in Maziv to 40%. 13. In a prior step to the proposed transaction, and in the form of an internal restructuring transaction, the shares in, and claims against, the main operating subsidiaries of CIVH, i.e., DFA and Vumatel, together with their respective subsidiaries, will be transferred by CIVH into Maziv. 2 14. In terms of the proposed transaction Vodacom would retain its MNO business, long-haul fibre assets and retail Internet Service Provider (ISP) business. 15. The following payments and transfers would take place in terms of the agreement: 15.1. Vodacom will (i) pay approximately R6 billion in cash into Maziv; (ii) pay approximately R4.2 billion to acquire shares in Maziv from CIVH; 3 and (iii) transfer fibre to the business (FTTB) and FTTH wholesale assets and metrofibre transmission links (Transfer Assets) 4 valued at approximately R4.2 billion to Maziv. 5 15.2. If the abovementioned top-up to 40% shareholding is triggered, Vodacom will pay a further R4 billion in cash. 6 2 Merger Filing Part A of the Record p 42 para 2.3. 3     FWB     4 See Sale of Transfer Asset Agreement, clause 2.1.64 (Bundle M part 1 p 712). 5 Joosub FWB p 321 para 7. 6 Joosub FWB p 322 para 8. Non-Confidential 5 16. The proposed transaction will create the following structure: MERGER PARTIES’ ACTIVITIES Vodacom 17. As indicated above, Vodacom is an MNO. It is active in the provision of mobile wholesale and retail voice, messaging and data services to residential and business customers. 18. We note that mobile connectivity gives millions of South Africans access to information, by providing extensive geographic and population coverage. As of 2023, South Africa enjoys more than 90 million active SIMs, and 41.6 million mobile data users, representing a 69% population penetration rate. 7 7             https://www.icasa.org.za/legislation-and-regulations/state-of-the-ict-sector-in-south-africa-2024-report;            https://www.statssa.gov.za/?p=15601 (accessed 25 March 2025). Non-Confidential 6 19. Vodacom is the largest MNO in South Africa, followed by Mobile Telephone Networks Proprietary Limited (MTN). The other major MNOs include Cell C, Telkom, Rain Proprietary Limited (Rain) and Liquid Telecom. As of 2023, Vodacom, MTN, Telkom, and Cell C have around million, million, million, and million subscribers, respectively. 8 This shows the relative size of Vodacom in comparison to the other MNOs. Other than Vodacom, only MTN has more than 20 million subscribers. 20.              telecommunications value chain, using various business models to build, acquire and lease infrastructure and sell services using that infrastructure. It owns national long-haul, metro backhaul and last mile fibre, including FTTH and FTTB. 21.   comprises its FTTH network infrastructure and related assets, including contracts, fixed assets and software. It also has an FTTB network, which is used to self-provide retail services to enterprises that are located in business parks and malls. It does not offer access to this infrastructure to third parties. 22. Vodacom leases and builds last mile fibre to connect end consumers (i.e., FTTH and FTTB) and uses this infrastructure to sell directly to consumers and enterprise customers (i.e., downstream retail services). Vodacom also has Vodacom long-haul (referred to as Vodacom Core) and a metro fibre network (referred to as Vodacom Access). 23. Vodacom is also active in the provision of fixed wholesale and retail services to           supported by its own fibre infrastructure and are also provided over third party infrastructure. Vodacom provides limited fixed wholesale services to other telecommunications providers. These include wholesale fixed managed 8 Nunes FWB p 165 paras 6.49  6.50; Smith    EWB p 256 para 181 Figure 7. Non-Confidential 7 services, wholesale bitstream access to ISPs, wholesale fibre and microwave backhaul and managed satellite internet services. Vodacom also provides wholesale Access Point Name services, leases access to its towers and provides wholesale network management services to Rain and Liquid Telecom.        DSL   to business customers, using wholesale DSL access from Telkom and business satellite internet services (particularly for customers without alternative access). Vodacom also provides more specialised services to enterprise customers, including unified communications solutions based on interlinking products and services; cloud and hosting services; connectivity services such as leased lines, microwave links, dedicated internet access, and best effort internet access; security solutions for IT systems and networks; and managed services, such as managed VoIP/telephony/video, managed VPN, and managed LAN services. 24. Vodacom holds several licences for the use of mobile and microwave spectrum. It holds licences for 14% of assigned mobile spectrum, including the following: 11 MHz FDD in the 900 MHz band; 12 MHz FDD in the 1800 MHz band; and 15 MHz FDD and 5 MHz FDD in the 2100 MHz band. It also provides wholesale national roaming services to Telkom and Cell C. 25.   s, and the conditions of those licences as well as submissions that it has made to the sector regulator, ICASA, are of relevance to both the competition and public interest assessments. We shall in the reasons  submissions made to ICASA, as well as its obligations in terms of its licences. Maziv 26. As indicated above, the main operating subsidiaries of CIVH are DFA and Vumatel. Non-Confidential 8 DFA 27. DFA is both a fibre infrastructure provider and an FNO. 28. DFA is a provider of wholesale, open access passive fibre infrastructure (so-  Layer 1 ducts and fibre cables) and managed network connectivity (Layer 2) services in both the metropolitan and long-haul telecommunications markets. DFA provides no Layer 3      primary activity is deploying metro fibre networks and providing wholesale access to this infrastructure to fixed and mobile network operators and service providers. 29. DFA owns fibre networks in Johannesburg, Cape Town, Durban, Midrand, Centurion and Pretoria, as well as in 25 smaller metros, such as East London, Polokwane, Tlokwe, Emalahleni and George. DFA also has spectrum licences in the 26 GHz band, which is allocated for point-to-multipoint microwave services, and which cannot be used for MNO mobile services. 30. DFA was originally established as a wholesale open access provider of passive or dark fibre infrastructure, but it has expanded its activities into managed or lit services provided over its infrastructure at each level of the value chain. 31. The DFA network comprises about km of fibre. DFA has achieved wide           32. In terms of national long-haul and metro fibre, DFA offers Titan, Peregrine and Calypte fibre products. In terms of access or last mile products, DFA offers dark fibre products in the form of Helios, Tachyon, Lumic and Arc and lit products consisting of Magellan and Business Broadband. Non-Confidential 9 Vumatel 33. Vumatel is an FNO that provides open access FTTH and FTTB infrastructure at the last mile level to ISPs. Vumatel installs fibre in a suburb and also installs a fibre spur up to the home wall. After the fibre has been installed, Vumatel provides capacity on the network to ISPs who then provide retail services to end- customers over the last mile infrastructure. In other words, it provides an active network over the fibre infrastructure, allowing ISPs to connect to the network on an open access basis. Vumatel carries the costs of fibre deployment in the suburb and the ISP is responsible for providing the wi-fi router and internet      9 34. Vumatel is active in three segments of the FTTH market, divided according to the monthly income of its customers, Core, Reach and Key: Core 34.1. The Core segment comprises approximately 2.2 million household customers who earn above R30,000 per month. 34.2. It is common cause that almost all homes in the Core segment have been passed (but not connected). Vumatel submits that it is now focused on the Reach and Key segments. Reach 34.3. The Reach segment comprises approximately 4.8 million customers and is aimed at suburbs where the average monthly household income is between R5,000 and R30,000 a month. 10 34.4. Vumatel launched Vumatel Reach in 2019 and states that it developed the roll-out model which made the Reach segment accessible. With Vuma Reach it         house, providing a Wi-Fi connection that can be shared by everyone in 9 Mare FWB p 439 para 38. 10 Mare FWB p 440 para 41. Non-Confidential 10 the house. Vuma Reach prepaid internet products are made available on a wholesale open-access basis to ISPs. 11 34.5.        more formal previously disadvantaged areas, like your Mitchells Plain, Soweto, Vosloorus, Soshanguve, places     . 12 34.6. Competitors are currently also rolling-out fibre in the Reach segment. 34.7.            Vumatel on Vuma Reach to date is approximately R . 13 Key 34.8. Vumatel launched the Key product in 2021. The Key segment refers to customers who earn under R5 000 per month, comprising between 9 and 11 million customers. 14 34.9. Mr Mare explains that these are people living in informal dwellings and include places essentially like Khayelitsha, Alexandra, and Kayamandi. 15 34.10. Vumatel has rolled-out fibre to Kayamandi and, as of February 2024, had passed approximately Key segment homes and the number of subscribers connected amounted to . 16 This shows a relatively low uptake by customers, also referred to in the industry as the penetration rate. 34.11. Mr Mare indicates that Key requires aerial deployment with GPS on the poles to manage uncertainty around addresses. Since Reach is aimed at the low-        mechanism, 17 an efficient distribution model, consistent customer service, and maintenance so that there is no disruption of service at a cost that is affordable to consumers in this segment. 18 11 Mare FWB p 440 para 43. 12 Mare FWB p 440 para 42. 13 Mare FWB p 443 para 51. 14 Inter alia Mare FWB p 441  442 para 48. 15 Transcript p 2577 lines 5  17. 16 Mare FWB p 441 para 48. 17 Transcript p 2580 line 15 to p 2582 line 22. 18 Transcript p 2580 line 15 to p 2582 line 22. Non-Confidential 11 34.12.             Key to date is approximately R . 19 35. We note that in terms of conditions imposed by the Tribunal in the merger involving CIVH and Vumatel, the parties to that transaction had to execute pilot projects in Alexandra and Mitchells Plain with a certain total anticipated capital cost. 20 36. Vumatel offers FTTB Best Effort to connect enterprises that it passes incidentally. These services are provided more to Small Medium and Micro Enterprises SMMEs     , for example, malls like DFA. BACKGROUND Commission’s investigation and concerns raised 37. On 4 August 2023, the Commission recommended to the Tribunal that the proposed transaction should be prohibited. 38. During its investigation, the Commission received submissions, data and other information and documents from inter alia the merger parties, customers of the merger parties, competitors, the Department of Trade, Industry and Competition (dtic), employee representatives and trade unions. 39. Concerns were raised by various third parties about the proposed transaction. For brevity we do not repeat those concerns here; the Commission deals with them in its Report from paragraphs 389 to 456. The concerns raised by these parties include concerns about vertical input and customer foreclosure, market consolidation, horizontal concerns, tying and bundling, durable first mover advantage concerns, 5G based concerns, removal of a competitor, information 19 Mare FWB p 442 and 443 para 51. 20 Community Investment Ventures Holdings Proprietary Limited and Vumatel Proprietary Limited (CT Case no.: LM109Jul18 CIVH/VumatelCT Case no.: LM109Jul18; see Condition 7.1. of the imposed conditions. Non-Confidential 12 exchange concerns, suitability of open access conditions, and AEX driven concerns. The Commission notes in its Report that most third parties are of the view that the proposed transaction should be prohibited and that no remedies would suffice to address those concerns. However, certain third parties made remedy suggestions. 40. After the hearing of evidence before the Tribunal, the Commission persists with its view that the proposed transaction ought to be prohibited. It submitted that    (mostly behavioural) conditions, in their different iterations, do not address the competition concerns that in the Commiss view outweigh the  public interest commitments. Dtic and union participation 41. The dtic 21 and the Communication Workers Union (CWU) participated in the hearing. 42. In terms of section 18(1) of the Competition Act, 89 of 1998, as amended the Actin order to make representations on any public interest ground referred to in section 12A(3), the Minister may participate as a party in any merger proceedings before the Commission, Tribunal or the Competition Appeal Court CAC, in the prescribed manner. 43. The dtic representing the Minister participated in this matter in relation to the public interest in terms of the abovementioned section. Its legal representatives questioned both the factual witnesses and the economic experts on public interest issues and made closing arguments in relation to the public interest. We note that the dtic did not apply to intervene in relation to the competition issues and therefore did not participate in the proceedings in relation to any of the competition issues. It submits that it will abide by the Tribunal's findings in regard to the latter. 21               Non-Confidential 13 44. During the course of 2022 and 2023, the dtic and the merger parties engaged with each other regarding the public interest effects of the proposed merger. This resulted in a draft framework agreement between them circulated in June 2023, which recorded a number of commitments made by the merger parties. 22 45. The dtic proposes that should the Tribunal be minded to approve the proposed merger, it should do so subject to the  ultimate set of proposed public interest conditions. 46. The CWU participated in relation to the effects of the proposed transaction on employment and ownership. It made written and oral submissions seeking the imposition of certain conditions it proposed in relation to employment and ownership if the Tribunal were minded to approve the proposed transaction. 23 Intervenors 47. The Tribunal gave intervention rights to two third parties (i) Rain; and (ii) MTN, both are customers and competitors of the merger parties. The Tribunal did not provide reasons at the time for allowing these interventions. The level of participation of these parties in the proceedings is self-evident from these reasons. Rain 48. On 16 November 2023, the Tribunal granted Rain, an MNO, leave to intervene in the merger proceedings on a limited basis. It is a customer of DFA and both a competitor and a supplier of Vodacom. 49. In 2023 Rain launched RainOne which is a bundled product that offers 5G fixed wireless access (FWA) and two 4G SIMs for mobile services. 22 These commitments are contained in the set of tendered conditions which was marked Exhibit CB during the hearing. 23 See Transcript inter alia p 42 line 12 to p 46 line 9. Non-Confidential 14 50. As a customer of DFA, Rain leases the following from DFA: (i) dark-fibre circuits           (ii)             circuits aggregate; and (iii) backhaul dark-fibre circuits from these aggregation        it is heavily dependent on DFA for these transmission services and that it has a strong preference to use dark fibre only (as supplied by DFA). 24 51. As a competitor to Vodacom, Rain competes as a provider of retail mobile and FWA products. 25 FWA is a home broadband product supplied by MNOs using              FWA router is nomadic insofar as it can be moved to another location where the service is supplied. LTE/4G FWA was the first generation of FWA provided over licenced spectrum in South Africa initially led by Telkom and Rain, with Rain as the first to introduce 5G FWA. 52. Rain offers (i) retail 4G mobile products i.e., through the sale of data and airtime on Rain simcards for use in mobile phones; 26 and (ii) retail 5G FWA products for home broadband and small businesses (through the provision to the client of a router that receives data over 5G signals and which re-transmits that data over a Wi-Fi              device). 27 53. Rain is a supplier of Vodacom in that it provides roaming services to Vodacom on the 4G layer of its network. 28 54. Rain submits that its customers are predominantly people in the middle- to lower-              24 Schoeman FWB p 521 para 8.6; Conrad Leigh, Rain Intervention Application Founding Affidavit, intervention bundle p 19 to 20 paras 35 to 39; Schoeman Transcript p 941 line 17 to p 942 line 10. 25 Schoeman FWB p 521 para 8.7. 26 See https://www.rain.co.za/mobile (accessed 25 March 2025). 27 Schoeman FWB p 519  520 para 8.1; Schoeman Transcript p 929 line 22 to p 930 line 5. 28 Founding affidavit intervention bundle p 19 para 33. This does not mean, however, that Vodacom               2189 line 20 to p 2190 line 6, p 2264 line 10 to p 2266 line 8, p 2266 line 9 to p 2267 line 18. Non-Confidential 15              townships. 55. Before the hearing started, Rain cited satisfaction with    tendered version of the (mostly behavioural) conditions filed on 14 March 2024, 29 and on 28 March 2024 withdrew its objection to the proposed transaction. 30 It thus would no longer intervene in the proceedings. 56. Given the above, the Tribunal requested Rain to explain its position and address the Tribunal on the following issues at the hearing: (i) its theories of harm/concerns regarding the proposed transaction; (ii) why and how the merger                 dered remedies. Following the  , Rain filed a factual witness statement. 57. Rain  after hearing the evidence is that it is not in favour of approving the proposed transaction without conditions. It submits that without appropriate conditions the proposed transaction is likely to substantially lessen competition in the affected markets and Rain would oppose the merger, given that the competitive concerns associated with the merger would outweigh its likely benefits. This is because: 57.1. the merged entity would have an incentive to foreclose competitors of Vodacom, including Rain, which would be paired with an existing ability on the part of DFA to foreclose primarily through the provision of services to Vodacom on preferential terms; and 57.2. the merger         information would leak from DFA to Vodacom as a competitor. Rain submits that this would be to its competitive detriment. It is heavily dependent on DFA for the provision of backhaul dark-fibre circuits. 29 T     iled on 14 March 2024. 30 Rain communicated in its letter to the Tribunal dated 28 March 2024 that it was satisfied that the revised version of the proposed conditions filed on 14 March 2024 addressed the concerns Rain had raised in its intervention application. Also see Transcript p 936 lines 1  7. Non-Confidential 16 MTN 58. On 16 November 2023, the Tribunal granted MTN leave to intervene in the hearing proceedings in terms of a defined scope. MTN is both a customer and a competitor of the merger parties. 59.             behavioural conditions filed on 14 March 2024. MTN submits that, absent effective conditions, the proposed merger would give rise to concerns of substantial anti-competitive harm. It however does not propose a prohibition of the proposed transaction. 60. During the hearing of the factual evidence, MTN on 19 July 2024 informed the Tribunal that it negotiated an agreement with the merger parties on behavioural conditions and therefore would no longer actively participate in the hearing other than closing argument. 31 The above development had certain practical implications since certain factual witnesses, including Mr Van Zyl Botha Mr Bothaof Herotel, were yet to testify and would have been crossed-    . 32 61. Recall that CIVH through Vumatel currently has a shareholding in Herotel (see paragraph 9 above). We note that in relation to Herotel, MTN on 02 June 2024 requested the Tribunal to issue a subpoena duces tecum requiring the CEO of Herotel to testify within the period allocated for the testimony of the factual witnesses nominated by the parties to these proceedings. The Tribunal on 11 June 2024 issued a subpoena requiring Mr Botha, the CEO of Herotel, to appear and testify under oath before the Tribunal on certain issues falling within his personal knowledge. He was further required to provide certain documents to the registrar of the Tribunal for inspection by the independent legal and 31 Transcript p 2683 line 14 to p 2685 line 19. 32 Transcript p 2686 line 3 to p 2694 line 4. Non-Confidential 17 economic advisors of all the parties to the proceedings against the provision of appropriate confidentiality undertakings. 33 62. MTN during the hearing placed its focus on improving    tendered behavioural conditions aimed at dealing with vertical concerns. As it articulates already in its opening statement, it is not a proponent of the prohibition of the proposed Maziv/Vodacom deal because Investment in South        consolidation of the industry is an inevitable and even desirable feature of the national and international landscape. 34 (Own emphasis) The reason why MTN does not favour that the deal be prohibited becomes clear during the hearing  it itself wants to make acquisitions to acquire a fibre footprint and grow fibre assets in order to effectively compete after this proposed merger. Its strategic documents reveal that it   target Openserve as a possible acquisition or JV partner to acquire a fibre footprint & wholesale business that can effectively compete with Vodacom CIVH      Multiple ISP/FNO acquisition & consolidation to grow fibre assets and customer base 35 The need for this to     to mitigate the risk of Openserve acquisition not being approved by regulators 36 Mr James Hodge Mr Hodgemakes the point that certainly, if you want to pursue other deals, then if this was prohibited pursuing those over the deals depending on the structure, is likely to not be feasible. 37 We shall deal with this further under the public interest analysis, i.e., the effects on the particular sector. 33                       inter se and in respect of Herotel; (iii) any document(s) reflecting the terms and conditions on which CIVH extended and/or facilitated funding to Vumatel, whether directly or indirectly, to enable it (or an associated entity) to acquire a shareholding in Herotel; (iv) any document(s) reflecting the terms and conditions on which CIVH extended and/or facilitated funding to the Trust, whether directly or indirectly, to enable it (or an associated entity) to acquire a shareholding in Herotel; (v) the four most recent annual budgets and/or business plans of Herotel; and (vi) the four most recent audited annual financial statements of Herotel. 34 Transcript p 27 line 16 to p 28 line 4. 35 See M        MTN SA  FTTX Way Forward: Role of Fibre and Options to Consider       36 Bundle O p 254. 37 Transcript p 3926 lines 13  20. Non-Confidential 18 Witnesses 63. The following factual witnesses testified at the hearing on behalf of the Commission: • Mr Abraham Van d  Mr Van der Merwe        FrogfootFrogfoot is an FNO that provides services (FTTH, FTTB and FTTT/S) on an open access, wholesale-only basis. It does not provide any services to retail customers; 38 •    Mr Masalesa      Consumer and Small Business; and •    Mr Motlekar       Consumer and Small Business. 64.       Mr Johnson   partner at Aetha Consulting Limited Aethawho testified on issues relating to    efficiencies and proposed behavioural remedies. 65.      Mr Nunes, the General Manager, Network Implementation at MTN, as a factual witness. 66.              Botha (see paragraph 61 above). 67. The merger parties called the following factual witnesses: •    Mr Uys         and the Chairperson of the board of directors of CIVH; •      Mr Joosub  of the Vodacom Group Limited; •    Mr Otty         a non-executive director on the Vodacom Group Limited board; 38 Van der Merwe FWB p 36 para 24. Non-Confidential 19 • Dr Ryan Van d  Dr Van den Bergh     Group Technology Strategy, Architecture, Spectrum and Assurance at Vodacom; and •     Mr Mare    ; and • Dr Marten Scheffer Dr Scheffer, a managing executive at Vodacom. 68. Factual witness statements were filed by the merger parties for the following individuals; however, they were not called to testify at the hearing: • Mr Sitho Mdlalose Mr Mdlalose, the CEO at Vodacom; and • Mr Robin Maduray Mr Maduray, a managing executive for transmission engineering at Vodacom. 69.       Mr Schoeman                 paragraph 56). 70. The following economic experts gave evidence in so- hot tub  • for the Commission, Mr Hodge, the Chief Economist at the Commission; • f     Mr Smith       • f       Prof Theron     Economic and Financial Consulting practice at FTI Consulting; and Mr Paul  Mr Reynolds        trading name of FTI Consulting LL. Hearing and ultimate conclusion 71. The hearing took place over 26 days in the period 20 May 2024 to 27 September 2024, with the last written submission received on 16 October 2024. 72. We note that both the customers and competitors of the merger parties have raised competition concerns with the proposed transaction and do not support the outright approval of the proposed transaction. Other than the firms Non-Confidential 20 associated with the above factual witnesses, several other third parties raised concerns with the proposed transaction     73. Although the merger parties tendered several iterations of behavioural remedies in an attempt to deal with the vertical concerns, the proposed remedies do not, other than a divestiture remedy in relation to FTTH infrastructure, address the horizontal competition concerns. Furthermore, the tendered behavioural remedies for the vertical competition concerns, as raised by customers and competitors alike, are technical in nature, cumbersome, and will not be effective and furthermore cannot be adequately monitored and enforced by the competition authorities. In addition, the tendered behavioural remedies, which are extremely complex and technical in nature, and affect many customers/competitors of the merger parties, will place a huge regulatory burden on the Commission and Tribunal and both institutions cannot take on this (sector) regulatory burden of indefinite duration. 74. Having considered the factual and economic evidence, numerous strategic and other documents and argument, the Tribunal on 29 October 2024 issued its order prohibiting the proposed transaction. CONTEXT Sector background and terminology used 75. To provide context to the assessment that will follow, we explain certain general features of the sector and the terminology used, including the different fibre infrastructure layers, our history of high mobile data costs in South Africa, and alleged past cartel behaviour involving Vodacom and MTN. Layers of fibre infrastructure 76. There is broad consensus between the parties on the general features of the four layers that make up the fibre infrastructure industry. These layers comprise the Non-Confidential 21 (i) international; (ii) national long distance; (iii) metropolitan (including backhaul); and (iv) last mile levels. 39 We explain each layer below. International and national long distance layer 77. We note that t        NLD infrastructure layer are not particularly relevant to the assessment of the competition issues in this merger. For completeness, we note that at the international layer, connectivity occurs through high-capacity undersea fibre cables which connect different countries and continents through multiple landing sites. 40 This infrastructure has high capacity given that it carries country aggregated traffic. 41 Consortia typically built these connections given the high cost to lay this infrastructure. 78. The NLD fibre/infrastructure connects major cities and towns across South Africa, transporting traffic between them and connecting them to the international connectivity infrastructure. 42 Metropolitan 79. The metropolitan layer is relevant to the competition issues raised in this merger. 80. At the metropolitan layer, metropolitan fibre rings and backhaul (or metro fibre) aggregate traffic from last mile access networks (both fixed and mobile) or aggregation nodes (including major datacentres), and transport aggregated traffic between these networks and nodes, and connect them to the infrastructure NLD layer. 43 Backhaul may also be provided by microwave links, but typically only where fibre is not in place or is too expensive to lay, and is 39 Van der Merwe FWB p 33  34 para 16; Nunes FWB p 129  131 para 2.4; Van den Bergh FWB p 188  189 para 5; Hodge EWB p 49 and 50; Smith EWB p 211 to 213; Reynolds EWB p 437 and 438. 40 Transcript inter alia p 74 line 3 to p 75 line 6. 41 Transcript p 74 lines 9  18. 42 Transcript p 75 line 11 to p 78 line 8. 43 Transcript p 78 lines 11  22; p 79 line 13 to p 80 line 12. Non-Confidential 22 typically self-provided rather than wholesaled. There are also some legacy copper connections, but these are currently being replaced by fibre. Last mile 81. Relevant to the competition issues raised in this merger, is the final layer known as the last mile fibre infrastructure. This infrastructure provides the connection between the metropolitan backhaul network and the final customer premises. 44 FTTB and FTTH 82. It is common cause that a distinction can be drawn between the markets for (i) residential customers using FTTH infrastructure; and (ii) commercial customers or enterprises using FTTB infrastructure. This distinction is due to the difference in service levels required by residential consumers and businesses. We note however that many SMMEs in South Africa also make use of FTTH infrastructure. 83. Both the FTTH and FTTB networks are built by FNOs and are typically offered on a wholesale basis to ISPs. The ISPs provide internet connectivity services and package the completed product to final consumers or businesses. However, there are some operators that offer the ISP and retail service for FTTB themselves and do offer their FTTB networks on a wholesale basis. Mobile and fixed retail services to consumers, SMMEs and businesses 84. MNOs use the fibre backbone infrastructure as the building blocks to form their core mobile networks and connect their core network to their last mile radio access network (RAN) infrastructure. The RAN makes use of base stations (or towers) and antennas to provide a wireless last mile connection to the end user devices using privately held spectrum. It connects directly to the metropolitan network through fibre to the site (FTTS) infrastructure. 44 Transcript p 81 line 14 to p 82 line 15. Non-Confidential 23 Technologies 85. Another view of the market is to segment it across different technologies. The market is segmented across the following technologies in South Africa: 85.1. FTTx which includes FTTH, FTTB, and others such as fibre to the site or tower (FTTT/S). Examples of service providers include DFA, Vodacom, Liquid Telecom, Openserve, Frogfoot, Octotel, Metrofibre, WIOCC, and Seacom. 85.2. Mobile Wireless Access which includes the 3G, 4G/LTE and 5G technologies (mobile) and some of this also is delivered as FWA products (e.g. Rain, and Fixed LTE). Examples of service providers here include Vodacom, MTN, Telkom, Rain, and Cell C. 85.3. Point-to-point FWA would usually use unlicensed or licensed spectrum (not 4G/5G). Examples of service providers include Comsol, Herotel, Sentech, MTN, and Liquid Telecom. 85.4. Satellite refers to technologies such as Very Small Aperture Terminal, and Low Earth Orbit. Examples of service providers include Yahsat, Eutelsat OneWeb, Vodacom, Liquid, Vox, Paratus, CMC, and Q-KON. 45 Data costs in South Africa 86. As context, we note that South Africa has a history of high data pricing that over recent years have reduced following intervention inter alia by the Commission. Mobile broadband (MBB) has long been flagged as a costly alternative source of internet access for consumers, which resulted in the initiation in 2017 by the Commission of its     DSMI          MBSI 45 Van der Merwe FWB p 35  36 para 18. Non-Confidential 24 87. The Commission during its market inquiry received a number of submissions arguing that access to data has become akin to a human right or an essential utility on the level of electricity and water. Affordable data was seen as critical for social and economic inclusion. Mr Hodge testifies that there is still a constant gripe about data prices in South Africa and that organisations like            three concerns, every month. 46 88. The DSMI found inter alia the structure of data pricing in South Africa to be anti- poor, meaning the poor who typically purchase small pre-paid data bundles, pay more for data than richer consumers who purchase larger bundles and on a contract basis. 47 Whilst the DSMI did reach agreement for short-term price reductions, it recognised that the mobile market remained uncompetitive which required regulatory interventions in the medium-term. The DSMI also included recommendations to extend fibre alternatives, including public Wi-Fi and FTTH, precisely to bring competitive pressure on persistent high mobile data prices. 89. Mr Hodge notes that there were two settlement agreements involving Vodacom and MTN after the Commission finding of a prima facie case of excessive pricing. 48 MNOs now offer a range of MBB packages with higher usage levels at lower prices. 90. Mr Reynolds provides a useful analysis of the significant decrease in the price per gigabyte for mobile services for MNOs in South Africa over the period 2018 and 2022, as illustrated by his Figure 13 as replicated below. 49 46 Transcript p 3663. 47 DSMI Final Report 2019. 48 Transcript p 3635. 49 Reynolds EWB p 546 para 7.34, Figure 13. Non-Confidential 25 Figure 13: Average effective mobile price per gigabyte by MNO, 2018-2022 Notes: Prices are exclusive of VAT. Source: Compass Lexecon analysis based on Africa Analysis data, PR0005 Effective mobile prices. 91. The above graph illustrates the historic high average mobile price per gigabyte per MNO in South Africa and the significant drop in the five-year period considered. 92. Mr Smith also notes significant changes in the price of mobile connectivity over time in South Africa. His Figure 8 50         customers for mobile data connectivity fell dramatically in the eight years from 2011 to 2019. This dramatic fall again highlights the historic high mobile data connectivity prices of MTN. 50 Which was submitted by MTN to ICASA during the MBSI. Non-Confidential 26 Economies of scale at MNO level 93. The history of the South African telecommunications sector has other            them a larger customer base and a higher share of customer revenue. This along with wholesale revenues make both players highly profitable and able to make vast investments in their network from retained earnings. The ability to invest at the level that Vodacom and MTN can ensures that they have a network quality advantage, including network speeds as they rollout new generation faster than rivals to more areas. 51 94. National network coverage as an MNO is important to benefit from economies of    national coverage provide them with significant economies of scale, as the largest cost components of an MNO are rental of space on towers/high sites for radios, backhauls, as well as equipment. Each of Vodacom and MTN have an advantage in that they can share tower space with their GSM radios and reduce a significant cost component and can also negotiate with tower companies to get lower rates due to scale. 52 In contrast, other (smaller) MNOs wanting to advertise national presence are forced to roam on one of the two networks at a much higher variable cost. 53 Alleged history of collusion between Vodacom and MTN 95. According to a 1996 Financial Mail report, Vodacom and MTN executives met in London in 1994 to discuss pricing. The result of the meeting became known as  London Agreement - a memorandum where cellular tariffs for South Africa allegedly were set. The memorandum allegedly outlined agreements on tariff structures, airtime discounts, and connection bonuses. At the time, Vodacom and MTN said that the agreement reached in London w legal and not anti- competitive The erstwhile Competition Board referred the London Agreement 51 Hodge EWB p 93 para 128.3. 52 Hodge EWB p 92 para 128.1 referring to Liquid Telecom letter dated 25 March 2022, Bundle M 6857. 53 Hodge EWB p 92 para 128.2. Non-Confidential 27 to the criminal authorities, but not much progress was made. My Broadband reports that according to Gordon Institute of Business Science Professor and former Tribunal Chairperson, Mr David Lewis,    concocted legal stratagems designed to keep the issue out of court 54 Issues in dispute between the parties in this matter 96. We note that many issues are in dispute between the parties in this case which in very broad terms include: 96.1. the (true) rationale of the proposed transaction (including the issue of so- called co-control) and   post-merger incentives; 96.2. the competitive dynamics and effects of the proposed transaction relating to the following markets: 96.2.1. whether or not dark vs lit fibre are in separate relevant markets and the market position of DFA; and 96.2.2. market delineation in regard to home broadband and (the degree of) competitive interaction between FWA and FTTH; 96.3. relevant counterfactuals including a competition, investment and fibre roll- out counterfactual; 96.4. the horizontal effects of the proposed transaction relating to the following markets: 96.4.1. metro fibre and FTTB actual and potential future competition; and 96.4.2. competitive interaction between FWA and FTTH, including in the future; 96.5. portfolio effects, specifically post-merger bundling concerns; 96.6. vertical foreclosure in relation to: 96.6.1. metro/FTTS used by MNOs; 96.6.2. wholesale metro and FTTB used by FNOs; and 96.6.3. wholesale FTTH/FTTB used by ISPs; 54 See My Broadband article of 26 November 2014 titled Secret Vodacom, MTN Pricing Agreement Warning available at: https://mybroadband.co.za/news/cellular/114623-secret-vodacom-mtn-pricing- agreement-warning.html (accessed 25 March 2025). Non-Confidential 28 96.7. the adequacy and effectiveness of the  proposed mostly behavioural (and one structural divestiture) conditions; and 96.8. the public interest effects of the proposed transaction including the merger-specificity of the effects. Legal framework 97. In terms of section 12A(1) of the Act, whenever required to consider a merger, the Tribunal must initially determine whether or not the merger is likely to substantially prevent or lessen competition, by assessing the factors set out in subsection (2), and if it appears that the merger is likely to substantially prevent or lessen competition, then determine (a) whether or not the merger is likely to result in any technological, efficiency or other pro-competitive gain which will be greater than, and offset, the effects of any prevention or lessening of competition, that may result or is likely to result from the merger, and would not likely be obtained if the merger is prevented; and (b) whether the merger can or cannot be justified on substantial public interest grounds by assessing the factors set out in subsection (3). 98. The assessment of a substantial prevention or lessening of competition under section 12A is determined on a case-by-case basis.   substantially     materially or considerably in amount or duration 55 99. In Imerys         [t]here is thus much to be               more probable than not. One is concerned with a predictive exercise where future outcomes cannot be measured with fine callipers 56 (Own emphasis) Indeed, this is a case where reasonably probable future outcomes are important 55 Medicross Healthcare Group (Pty) Ltd and Another v Competition Commission [2006] 1 CPLR 1 (CAC) (Medicrossat para 19. 56 Imerys South Africa (Pty) Ltd and Another v Competition Commission (147/CAC/Oct16, IM013May15) [2017] ZACAC 1; [2017] 1 CPLR 33 (CAC) (2 March 2017) Imerysat para 53. Non-Confidential 29 for the assessment of the competitive effects of the proposed transaction, with        . 100. When determining this the Tribunal must assess the strength of competition in the relevant market(s), and the probability that the firms in the market after the merger will behave competitively or co-operatively, taking into account any factor that is relevant to competition in the relevant market(s), which may include issues such as an increase in price, 57 a reduction in quantity, quality, consumer choice or a loss of innovation. Concerning the substantiality of any likely price increase, we note that the Tribunal in Draslovska   any negative price effect as a result of a merger arguably represents a substantial effect for purposes of section 12(1) of the Act 58 101.         n Imerys explains:              that the Commission bears the burden of proving likely SLC. It is nevertheless so that, if on all the evidence before the Tribunal, a likely SLC cannot be found, the Tribunal must approve the merger unless the public interest override is operative 59 102. In relation to remedies, as we have (mostly behavioural) remedies put up in this case, the CAC in Imerys held I think it is permissible for the Tribunal to reason              conditions are more likely than not to remedy the likely SLC, there is a reasonable possibility that they will fail to do so. Therefore we prohibit the merger 60 (Own emphasis) 103. The Act furthermore in terms of section 12A(1A) requires us to, despite our determination in subsection (1), also determine whether the merger can or 57 Competition Commission of South Africa v Mediclinic Southern Africa (Pty) Ltd and another 2022 (4) SA 323 (Constitutional Court) Mediclinic) at para 54. 58 Draslovska Holdings A.S v Competition Commission of South Africa and others IM139Dec21 (11   Draslovskaat para 194. 59 Imerys (CAC) para 38. 60 Imerys (CAC) para 40. Non-Confidential 30 cannot be justified on substantial public interest grounds by assessing the factors set out in subsection (3). These factors are the effect that the merger will have on- (a) a particular industrial sector or region; (b) employment; (c) the ability of small and medium businesses, or firms controlled or owned by historically disadvantaged persons, to effectively enter into, participate in or expand within the market; (d) the ability of national industries to compete in international markets; and (e) the promotion of a greater spread of ownership, in particular to increase the levels of ownership by historically disadvantaged persons and workers in firms in the market. 104. In these reasons we first assess the competition effects, including efficiencies, and then assess the public interest factors. We then perform a balancing exercise of the anti-competitive effects and the merger-specific public interest benefits. TRANSACTION RATIONALE 105. We next consider the evidence regarding the rationale for the proposed transaction. Determining the true rationale of the proposed transaction is essential since it gives context and informs the competition and public interest analyses that will follow.    106. The merger parties submit that Vodacom wishes to invest in Maziv because Maziv has the unique capability, know-how and expertise to roll out fibre at scale and with speed. 61                    61 Joosub FWB p 329  330 para 24. Non-Confidential 31 that it is off-        businesses from the outset and is likely to realise a future return derived from       -up of fibre. 62 107.            required to focus on connecting homes that are already passed, rather than passing new homes, because of its capital constraints. 63 Mr Uys submits that the proposed transaction will provide Maziv with the necessary capital to continue with the roll-out of fibre, supporting the provision of internet into lower income areas 64 In this way, the proposed transaction will contribute     digital divide 65 Maziv also sees an opportunity through the proposed transaction to expand its fibre footprint through the             access offering to customers. 66 108. The merger parties rely on Mr Uys oral evidence during the hearing that until Maziv reduces its ratio it cannot invest further in expanding its network. 67 The capital injection from Vodacom will improve its ratio. 109. They also argue that Vodacom (alone or in a FibreCo JV) will not build fibre in competition with Vumatel and DFA.   110. The Commission submits that both Vodacom and Maziv have defensive reasons for concluding the transaction. 62 Joosub FWB p 330  331 para 26  27. 63 Uys Supplementary Witness Statement FWB p 551 para 23. 64 Uys FWB p 486 para 58. 65 Uys FWB p 486 para 58. 66 Transcript p 1116 lines 14  22. 67 Transcript p 1125 lines 3  4. . Non-Confidential 32 111. It submits that CIVH initially rejected the transaction before later agreeing to it,                  112.    customer, and the proposed transaction presents an            customer. The Commission further submits that pre-merger DFA must price aggressively in order to retain its customer and, post-merger, this pricing constraint is lessened or lost as the risk of losing Vodacom is eliminated by this proposed merger. 113. According to the Commission, Vodacom has sought a fibre deal given the risk posed to its mobile revenues as a result of mobile consumers embracing fibre. Our assessment 114. Throughout our assessment we         strategic and other documents since these documents, unlike the merger            business and not prepared specifically for this merger hearing. This weight is consistent with international practice. The USA 2023 Merger Guidelines explain this weight as follows: The Agencies often obtain substantial information from the merging parties, including documents, testimony, and data. Across all of these categories, evidence created in the normal course of business is more probative than evidence created after the company began anticipating a merger review. Similarly, the Agencies give less weight to predictions by the parties or their employees, whether in the ordinary course of business or in anticipation of litigation, offered to allay competition concerns. Where the testimony of outcome-interested merging party employees contradicts ordinary course business records, the Agencies typically give greater weight to the business records  Similarly, other suppliers, indirect customers, distributors, consultants, and industry analysts can also provide information helpful to a merger inquiry. As with other interested parties, the Agencies give less weight Non-Confidential 33 to evidence created in anticipation of a merger investigation and more weight to evidence developed in the ordinary course of business 68 (Own emphasis) 115.            considering the deal reveal the true rationale for the proposed transaction rather than what is submitted later to the competition authorities. These documents              of substantial competition from Vodacom at both the DFA and Vumatel levels, as shown below. 116.            than concluding the current deal with CIVH as made clear during the negotiations to Mr Uys and reflected in the board minutes. 117.                                            69 was not concluded in December 2020 and a shareholder subcommittee was formed to continue discussions with             an opportunity for them to invest in CIVH. Vodacom is still interested in investing in CIVH, however, there is limited time to conclude the transaction as Vodacom is considering other options. It was noted that should the parties fail to reach an agreement on the transaction by the end of March 2021 Vodacom would consider other options. Own emphasis) Mr Uys confirms that these are his words as contained in the signed minutes. 70 118. Mr Uys further indicates that Mr Joosub, the Chief Executive Officer and Executive Director of Vodacom Group Limited, had reached the end of his tether at that point and was applying pressure for Maziv to confirm the conclusion of 68 Merger Guidelines of the U.S. Department of Justice and the Federal Trade Commission, 2023 Section 4.1 page 35. 69 Project Lindt is the internal nomenclature used by CIVH denoting the proposed transaction. 70 Minutes of the meeting of the CIVH Board on 17 March 2021.Bundle M p 1174 para 2.1.1.10; Transcript p 1194 lines 6  21. Non-Confidential 34            do other things 71                things 72 119. He further confirms that CIVH delayed the transaction for a considerable time: We then asked Mr Joosub, Mr Joosub, we are sorry that CIVH did not approve the transaction in December. Mr Joosub, I cannot use the words he used to us, but they were not friendly and he said, you guys have messed me around for the whole of 2020 and then at the end you did not approve the deal 73 120.              businesses and specifically what Vodacom would do absent this proposed transaction. These documents show that among the threats to its businesses   aggressive land-grab and overbuild 74 strategies of competitor metro and backbone networks in secondary cities price erosion due to large ISPs and network operators looking to self provide rapidly reducing ARPUs 75 and overbuild  competition from other technologies such as satellite broadband, tapping into rural and other low return areas. 76 (Own emphasis) 121. The documents furthermore identify certain high probability scenarios which presented a threat to CIVH. Mr Uys under cross examination sought to downplay          hypothetical scenarios that we discussed 77      thinking back  identified scenarios were not a real credible outcome. 78 We do not find this attempt by Mr Uys to downplay the strategic documents credible. The internal                     -ante view when questioned 71 Transcript p 1195 lines 12  22; p 1196 line 12 to p 1197 line 6; p 1202 lines 20  21. 72 Transcript page 1197 lines 1  6. 73 Transcript page 1195 lines 12  17. 74 Overbuild refers to the duplication of fibre infrastructure by two or more fibre infrastructure providers that have laid their own fibre optic cables in the same area or even in the same roadside trench. 75 Average Revenue Per User. 76 Bundle M p 1221; Transcript p 1209 lines 2  20. 77 Transcript p 1226 lines 6  7. 78 Transcript p 1217 line 22 to p 1218 line 1. Non-Confidential 35 about any document before the Tribunal. Furthermore, the scenarios were prepared by management and advisors, 79 considered and discussed by the CIVH board and, importantly, informed the decision to do the deal with Vodacom. 80 Notably, no other scenarios were presented to the CIVH board. 122. When the strategic documents are put before Mr Uys    counsel, he acknowledges that they were put together by management and advisors, but   the Chairman of the Board  we debated each and every of these had no idea where the information came from: ADV           information from? MR UYS: No idea 81 We do not find this credible. 123. We consider the threats identified in the CIVH internal documents in relation to DFA and Vumatel respectively. Threats to DFA 124. In relation to DFA the identified threats relate inter alia to Vodacom establishing (i) a FibreCo; and (ii) a TowerCo. We deal with these in turn indicating what the anticipated effect will be on the competition that CIVH will face absent the proposed deal, the fear being significant increased competition for DFA. The FibreCo threat 125.       has identified  establishment of a FibreCo by pursuing deals with other FNOs, making a significant investment, and expanding as a rival to DFA: 82 Vodacom has identified the establishment of a Fibre Co as a key strategic priority (with or without Lindt). Without Lindt, to deliver its strategic ambitions, Vodacom will partner with another FNO (e.g; ) or MNO (e.g; ) to 79 Transcript p 1213 line 22 to p 1214 line 1. 80 Transcript p 1225 lines 21  22. 81 Transcript p 1214 lines 1  12. 82 Bundle M p 1223. Non-Confidential 36 establish a Fibre Co which will now be a rival to DFA and in which it will have invested R6 billion cash which can be used as CAPEX to fund its expansion. (Own         Vodacom has identified   strategic document. 126. The implications of this FibreCo threat are recorded in the internal documents and will play out at three levels of effects on DFA: 126.1.       tenant and the largest MNO in South Africa, Vodacom - together with major revenue churn in a competitor FibreCo in which Vodacom would have a shareholding; 126.2. the risk of losing R in revenue per annum with average contract tenure of years; and 126.3. in addition to the above, all future new business from Vodacom which would shift to its own FibreCo (new links for existing base stations and new 5G densification). 83 127. The above concern          would move to the FibreCo - being a rival to DFA. The implications of that are that DFA would be unable to expand its network footprint as planned to match that rivalry and would be unable to quickly and efficiently capture new FTTB market share whilst mitigating downward pressures on pricing. 84 128. A further issue identified in this CIVH       rival FibreCo could expand its existing routes to rooftop and indoor mobile sites in buildings, to also include , as well as the expansion of the FibreCo model to footprint in other markets on the continent. The implication of this is that the FibreCo FTTB offers, resulting in further downward             83 Bundle M p 1223; Transcript p 1215 line 3 to p 1216 line 3. 84 Bundle M p 1223. Non-Confidential 37 Vodacom on expansion opportunities to enter new markets outside of South Africa. 85 129. The further scenario contained in this strategic document is of competitors concluding similar deals to establish FibreCos with with the implication that  86 The TowerCo threat 130. The further threat identified in the CIVH internal document relates to Vodacom       Vodacom has also identified the establishment of a separate TowerCo as an additional key strategic priority (with or without Lindt). Own       Vodacom is looking to partner with , to establish the TowerCo with  87 131. From a DFA perspective, the implications of this scenario were inter alia that 88 Vodacom will also not DFA products and services resulting in for DFA. DFA will not only previously mentioned, but also stands to lose . Own emphasis) 132. A further negative consequence of the above highlighted in the strategic document is that CIVH would lose due to       89 85 Bundle M p 1223. 86 Bundle M p 1223. 87 Bundle M p 1223. 88 Bundle M p 1223. 89 Bundle M p 1223. Non-Confidential 38 Vumatel 133. In relation to Vumatel, the threats identified by   Vodacom      result in 90 We note that this clearly suggests that Vodacom has lower Average Revenue Per User ARPUs. The impact would be      ARPUs and  ARPUs. 91 This confirms that Vodacom has ARPUs than Vumatel. 134.                         assets at a in the present transaction to commercially its core business to future competition. The CIVH Board pack explains this as follows: Although CIVH is acquiring the assets at a , the overall transaction is set to commercially the core business of CIVH significantly into the future. The strategic benefits in a rapidly evolving and uncertain market and economic environment potentially outweigh the short-  92 (Own emphasis) 135. The CIVH Board pack further records one of the main strategic benefits of the     93 136. Mr Uys in his oral evidence concedes that the proposed transaction is a means by which Maziv de-risks itself from losing Vodacom as a client and gaining it as a formidable competitor. 94 90 Bundle M p 1223 and 1224. 91 Bundle M p 1224. 92 Bundle M p 1231 and 1236. 93 Bundle M p 1236. 94 Transcript p 1229 lines 6  13. Non-Confidential 39 137. As regards Vodacom, the transaction allows it to have a stake in fibre revenues given that fibre roll-out reduces FNO revenue.     Value at Risk to its mobile business (also see paragraph 285 below). The Value at Risk that Vodacom sees for its mobile business, is estimated at half of R11.8      95 Mr Joosub confirms a loss of up to 30% of mobile data spend within a house as they move to fibre, although he sought to claim it was less than   Consumer VaR estimated to be R11.8bn (Euro570mn) for the total period up to FY24 96 He explains the     essentially                data spend within the household and you could lose as much as 30% of the              97 138. The merger parties concede that a segment of mobile revenues has moved to fibre, but argue that the demand for mobile data outside the home continues to grow rapidly. 98 This however is not responsive to the fact that fibre is taking away revenue from the FNOs    both identify it as a concern and quantify this. As Mr Van der Merwe of Frogfoot testifies, As fibre roll-outs into new low-income areas continue and new products are developed to target consumers at a lower price points than what has traditionally been targeted by FTTH providers, the MNOs will lose customers and revenue. The MNOs will need to respond by rolling out better technologies (like 5G) on a more widespread basis and/or dropping prices for their mobile and FWA products 99 Dr Van den Bergh    the installation of fibre in these               100 139. In conclusion, CIVH true rationale for the proposed transaction is its defensive reasons for the proposed transaction in relation to both DFA and Vumatel, as 95 The figure of R11.8 billion was presented to the Vodacom Board. See, for example, Transcript p 2003 lines 2  12. 96 Bundle M p 3498. 97 Joosub Transcript p 1639 lines 6  14. 98 Transcript p 1953 line 14 to p 1954 line 5. 99 Van der Merwe FWB p 40  41 para 34. 100 Van den Bergh FWB p 215 para 80. Non-Confidential 40 explained above.     s to have a significant stake in the future fibre revenues of the largest dark fibre and FTTH provider in South Africa, given its Value at Risk, as explained above. VODACOM’S RIGHTS IN TERMS OF THE TRANSACTION, INFLUENCE AND INCENTIVES 140. In order to contextualise the competition assessment, one must have regard to the rights that Vodacom will enjoy in terms of the proposed transaction post- merger, and how the proposed transaction changes the   incentives. 141. The Commission submits that after the merger there are shared incentives between Vodacom and Maziv which cannot be separated from the co-control ownership structure of the proposed deal. Inter alia    economic interests will be mutually aligned, and various incentives give Vodacom reason to limit the degree to which it competes against Vumatel post- merger, thereby substantially lessening competition in the relevant markets. 142. The merger parties submit that the acquisition of negative control rights by            in assessing the effects of the limited number of veto rights it will be granted. It does not mean that M         of Vodacom. They say that         remaining MNO and retail operations. Therefore, the only aligned incentive between Vodacom and CIVH that the merger will create is the incentive to maximise the profitability of Maziv. The merger will not give either Vodacom or CIVH an incentive to promote the interests of Vodacom over the profit maximising interests of Maziv itself. 143.     -merger rights as 30%-40% shareholder in Maziv. Non-Confidential 41 Merger parties submissions 144. When the merger parties on 10 December 2021 notified the proposed merger to the Commission, they disclosed The Proposed Transaction will result, inter alia, in Vodacom acquiring at least 30% of the issued share capital of Newco [Maziv]. This will, in turn, result in Vodacom qualifying for various rights in terms of the Newco Memorandum of Incorporation ... that will give Vodacom joint control of Newco. The Proposed Transaction will accordingly result in Vodacom acquiring control over Newco in terms of section 12(2)(g) of the Competition Act 89 of 1998. 101 (Own emphasis) 145. During the hearing, the merger parties alleged that the term joint control  simply a useful label to describe the nature of the rights exercised by Vodacom  rights that will enable Vodacom to prevent a limited set of decisions being made unilaterally by CIVH. The merger parties adopt the position that, post- merger, Vodacom would not have an unfettered ability to control Maziv, and to             . 146. The merger parties also submit that Vodacom would have no influence over the operations of subsidiaries such as DFA and Vumatel. 102 Mr Joosub testifies that as far as he knows, the Reserved Matters (explained below) relate only at the level of Maziv itself. 103 147. In his first witness statement Mr Uys, the chair of the boards of directors of CIVH, Maziv, DFA and Vumatel, 104 contends that post-merger Vodacom would have (only) a minority shareholding in Maziv, by virtue of which Vodacom will qualify for certain shareholder reserved matter veto rights that will provide it with joint control of Maziv, without compromising the operational independence of Maziv 142 or without provid[ing      -to-day management 101 DLA Piper letter to Commission dated 10 December 2021 para 2.3, Bundle M p 35; Bundle M p 160  170, specifically p 161 par 1.1.4; Transcript p 1860 line 5 to p 1862 line 20. 102 Joosub FWB p 320  321 para 6; p 322  324 paras 10  14; Uys FWB p 483  486 paras 50, 53  57. 103 Transcript p 1869 line 20 to p 1870 line 4. 104 Transcript p 1399 lines 4  9. Non-Confidential 42    105 In substantiation of this assertion, Mr Uys mentions only   post-merger rights and confirms this at the start of his evidence in chief without qualifying the evidence. 148. Mr Joosub, the CEO of Vodacom Group Limited and a director of Vodacom, in his witness statement states that post- Maziv remains a business controlled by CIVH as a majority shareholder, completely separate from the Vodacom business        protect          majority to make material changes to the business of Maziv 106 149. Relying on the evidence of Messrs Joosub and Uys, the merger parties contend that the package of   normal minority protection rights- given to a shareholder, i.e., Vodacom investing R14 to R18 billion to acquire a minority interest in a company. i.e., Maziv. 107 Mr Uys submits that the veto rights are not unusual for an investment of this nature. He states that these minority protections are less than what current CIVH shareholders get at a 30% shareholder level and that they are good rights for a R15 billion investment 108   submissions 150.           obtaining substantial rights in terms of the Memorandum of Incorporation of Business Venture Investments No 2213 (Pty) Ltd, renamed Maziv MOI 109 151. MTN submits that post-merger, Vodacom would have joint control of Maziv and material influence over the operations of its subsidiaries and even of its controlled investee companies. It alleges that the merged entity could use the business operations of group companies (at least partially) to foreclose MNOs, FNOs and ISPs in the industry. 105 Uys FWB p 483 para 50. 106 Joosub FWB p 320  322 para 6 and 10. 107 Transcript p 1675 line 20 to p 1676 line 8. 108 Transcript p 1236 lines 9  13. 109 Bundle M p 801  855. Non-Confidential 43 Our assessment 152. Section 12(2)(g) of the Act provides that a person controls a firm if it has the ability to materially influence the policy of the firm in a manner comparable to a person who, in ordinary commercial practice, can exercise an element of control referred to in paragraphs (a) to (f). 153. In Caxton and CTP the CAC held that the concept policy of a firm should be viewed in a wide sense and within the context of each case. 110 Indeed, context is important in this case when one considers both the ability to influence policy and post-merger incentives. 154.        While it should be accepted that influence on one aspect of a firm may not be sufficient to constitute material influence over the policy of that firm, context is very important. There may be matters whose nature is so material to the strategic direction of the firm (even if numerically few) such that influence on them may be reasonably extensive in a manner that qualifies to control contemplated by paras 12 (2) (a) to (d) of the Act. That qualification, we would suggest, was ma                 ) 111 As we shall show below, Vodacom will post-merger have the ability to influence numerous aspects of Maziv. 155. It is common cause that Vodacom, as a 30%-40% shareholder, will be able, through the veto rights it will         control on specific elements of the Maziv business, requiring consensus on those aspects before they are implemented. The veto rights relate to a number of matters as set out in Schedule 2 of the MOI. The parties describe this as   negative control      110 Caxton and CTP Publishers and Printers Limited and Others v MultiChoice Proprietary Limited and Others        Caxton and CTP) at para 79. 111 Caxton and CTP at para 79. Non-Confidential 44 156. The merger parties concede that the proposed transaction was notified as a merger because certain of the veto rights granted to Vodacom relating inter alia              112 are typically regarded in competition law as a form of de facto (or negative) control in terms of section 12(2)(g) of the Act. 157. The first to note is that Mr Joosub indicates that Vodafone, in order to approve this acquisition as one of three conditions, required that Vodacom secure that it could obtain co-control. 113 Mr Otty of Vodafone confirms that Vodafone made the transaction conditional upon co-control. 114 Mr Uys also confirms during his testimony that Vodacom wanted joint control with CIVH when negotiating the proposed transaction. 115 158. The factual evidence further confirms that Vodacom will be an active strategic investor in Maziv rather than a passive financial investor. As we shall show below, the rights granted to Vodacom go well beyond the rights ordinarily granted in terms of the Companies Act for the protection of minority shareholders. 116 As Mr Joosub testifies, we [Vodacom] want a seat around the table 117  given the size of the cheque, it needs to be a strategic investment as opposed to a financial investment. 118 159. We further note that Maziv does not have any assets beyond its investments. 119 Maziv also does not have any operations beyond those of its subsidiaries. 120 Indeed the basis for seeking approval of the merger was that Vodacom would gain the ability to influence the policy of subsidiaries such as DFA and Vumatel or operations beyond those of such subsidiaries. 121 A driver for such influence 112 Schedule 2: Reserved Matters, Bundle M p 853. 113 Transcript p 1793 lines 4  17. 114 Transcript p 2022 lines 3  10. 115 Transcript p 1205 lines 6  16. 116 Companies Act 71 of 2008: sections 26; 31(1); 39(2); 61(3); 61(4); 61(5); 62; 66(4); 65; 163; 165. 117 Transcript p 1776 lines 13  15. 118 Transcript p 1777 lines 11  12. 119 Transcript p 1865 lines 5  9. 120 Transcript p 1865 lines 12  18. 121 Transcript p 1864 line 20 to p 1866 line 5. Non-Confidential 45             -billion rand investment in Maziv and its transfer of valuable fibre assets to Maziv. 122 160. Regarding decision-making at shareholder level, the MOI provides that the Company [Maziv] shall not perform, permit, conclude or implement any Reserved Matters in respect of any Group Member [including DFA and Vumatel 123 ] (or in respect of any investee company, to the extent within its control 124 ) unless every Controlling Shareholder [CIVH and Vodacom 125 ] has confirmed its support for the relevant Reserved Matter in writing. 126 This means that Maziv requires the written consent of Vodacom to engage in any conduct specified in Schedule 2 of the MOI. 127 161. Schedule 2 lists 23 Reserved Matters in respect of which any valid decision would require the concurrence of Vodacom. We note that contrary to the evidence of Messrs Uys and Joosub, most of such matters relate to decisions not only of Maziv but also of subsidiaries such as DFA and Vumatel and controlled investee companies (potentially including Herotel). 162. The Reserved Matters 128 include that Vodacom has the ability to veto the              financing of debt and the adoption or amendment of the dividend policy. 129 For completeness, we list certain of the Reserved Matters: 162.1. the approval of the annual budget and business plan of Maziv and its subsidiaries and/or any deviation therefrom. Mr Joosub clarifies that DFA and Vumatel fit within the Maziv business plan and that in terms of the  tendered behavioural remedies Vodacom would not by virtue of this reserved matter claim an entitlement to veto the underlying 122 Transcript p 1866 line 6 to p 1867 line 14. 123 Transcript p 1872 line 14 to p 1873 line 15. 124 This potentially includes Herotel. 125 Transcript p 1871 line 16 to p 1872 line 13. 126 Bundle M p 837 MOI clause 5.5. 127 Bundle M p 806 MOI clause 1.1. 128 Bundle M p 853  855. 129 Transcript p 1805 line 18 to p 1807 line 19; p 1808 lines 11  17. Non-Confidential 46            impairment at the Maziv level; 130 162.2. the appointment or dismissal of the (group) CEO and CFO. 131 This means that the approval of both Vodacom and CIVH is required for the appointment or dismissal of the CEO and CFO. 132 Vodacom would have              proposed transaction. 133 The CEO and CFO of Maziv are ex officio            NomComwill suggest a preferred candidate in respect of which the Board will vote and that vote will be subject to approval by the Controlling Shareholders, CIVH and Vodacom; 134 162.3. the cessation or discontinuation of any material business of the Maziv Group, where the business shall be regarded as material if: (i) it contributes more than % of the annual revenue of the Group: or (ii) it is reasonably anticipated that it will start to contribute more than % of the annual revenue of the Maziv Group within the next financial years; 135 162.4. any material changes to the nature of the business and/or strategic direction of any Maziv subsidiaries or of the Maziv Group as a whole (unless already approved as part of the budget and business plan) that is not in the ordinary course of business; 136 162.5. the acquisition and disposal of assets if the aggregate purchase price payable in respect of the transaction, or the aggregate value of the assets disposed of, exceeds R500 million; 137 130 Bundle M p 853 MOI Schedule 2 clause 1 - 2; Transcript p 1875 line 8 to p 1876 line 22. 131 Bundle M p 854 MOI Schedule 2 clause 14; see also Exhibit AA; Transcript p 1884 line 8 to p 1885 line 10. 132 Clause 6.2.4.2 at Bundle M p 838 and clause 14 of Schedule 2 at Bundle M p 854. 133 Transcript p 1886 line 19 to p 1889 line 2. 134 Transcript p 1897 line 12 to p 1898 line 12; clause 6.2.4 at Bundle M p 838. 135 Bundle M p 853 MOI Schedule 2 clause 11; Transcript p 1883 lines 18  21. 136 Bundle M p 853 MOI Schedule 2 clause 3; Transcript p 1877 line 1 to p 1879 line 7. 137 Bundle M p 853 MOI Schedule 2 clause 6; Transcript p 1881 lines 16  22. Non-Confidential 47 162.6. any transaction that is not in the ordinary course of business and which: (a) exceeds R or which (b) binds the relevant Group member to obligations for longer than ; 138 162.7. the incurral of debt finance above a certain amount; 139 162.8. the provision of security above a certain amount; 140 162.9. the establishment or implementation of an employee profit or share incentive scheme and/or B-BBEE participation scheme; 141 162.10. any Group Member making any loan or otherwise extending any credit to a third party in excess of a certain amount; 142 162.11. the commencement of a litigation or arbitration process; 143 162.12. a related party transaction in excess of R or which binds the Group for more than years; 144 and 162.13. the conclusion of a profit- or revenue-sharing agreement with a person that is not a member of the Maziv group above a certain amount in any financial year. 145 163. Mr Joosub concedes that all these rights remain, except for any specific carve- outs in the merger partiesproposed conditions. 146 138 Bundle M p 854 MOI Schedule 2 clause 15; Transcript p 1891 line 18 to p 1893 line 6. 139 Bundle M p 853 MOI Schedule 2 clause 4; Transcript p 1880 lines 2  16. 140 Bundle M p 853 MOI Schedule 2 clause 5; Transcript p 1880 line 17 to p 1881 line 15. 141 Bundle M p 853 MOI Schedule 2 clause 7; Transcript p 1882 lines 1  15. 142 Bundle M p 853 MOI Schedule 2 clause 9; Transcript p 1882 line 16 to p 1883 line 17. 143 Bundle M p 853 MOI Schedule 2 clause 10; Transcript p 1883 lines 18  21. 144 Bundle M p 854 MOI Schedule 2 clause 16; Transcript p 1891 line 18 to p 1893 line 6. 145 Bundle M p 854 MOI Schedule 2 clause 17; Transcript p 1891 line 18 to p 1893 line 6. 146 Transcript p 1806 lines 5  10. Non-Confidential 48 164. The Maziv board, whose decisions are made by majority vote, shall consist of the following tree types of director: (i) two ex officio board members; (ii) up to three independent board members; and (iii) up to 14 directors appointed by CIVH and Vodacom. 147 165. Despite Vodacom only acquiring up to 40% of Maziv shares, CIVH and Vodacom will have equal representation on the Maziv board and the same voting rights. 148 CIVH can nominate up to seven directors (as a 70% shareholder) and Vodacom can match this number of directors. 149 We note that CIVH and Vodacom need not nominate their full quotient of directors and if either of them elects to nominate less directors than provided in the MOI, the nominated directors would vote as if the full quotient were elected. 150 166. Vodacom SA would appoint two directors from the Mergers and Acquisition and Business Development Teams in Vodacom Group Limited and two further appointees from Vodacom International. The four Vodacom appointees together      rry 61% of the votes on the Maziv Board, and would be able to overrule the other directors; alternatively, if Vodacom is able to persuade three independent directors to vote with it on a particular issue, it would have more than 50% of the Board votes. 151 167. As indicated above, the Maziv group CEO and CFO would be the ex officio directors of Maziv, subject to certain veto rights. Any successor CEO or CFO would be recommended by NomCom and appointed by the board, subject to the approval of each of CIVH and Vodacom. 152 NomCom would comprise five members, two of whom would be appointed by each of CIVH and Vodacom. NomCom would make recommendations in respect not only of members of the     all members of senior executive management in the Group, 147 Clause 6.2.2 Bundle M p 838; Transcript p 1896 line 8 to p 1897 line 11. 148 Clause 6.3.14 Bundle M p 845. 149 Clause 6.2.6.1(a) Bundle M p 840. 150 Clause 6.2.6.8 Bundle M p 842. 151 Bundle M p 1082; Transcript p 1802 line 5 to p 1803 line 6. 152 Clause 6.2.4 Bundle M p 838; Transcript p 1897 line 12 to p 1898 line 12. Non-Confidential 49 as well as independent non-executive Directors and candidates to serve as the chairperson of the Board 153 168. The approval of both Vodacom and CIVH is required for the appointment of independent non-executive directors of the board. 154 The chairperson of the Board must be an independent non-executive director. 155 169. The shareholder representatives nominated by CIVH and Vodacom must always exceed in number the aggregate number of the ex officio and independent directors. 156 170. If an independent director is to be removed or replaced, the approval of both CIVH and Vodacom is required. 157 Corporate dynamics and incentives 171. As indicated above, the merger parties argue that the above rights were              minority interests against CIVH exercising its majority shareholding to make material changes to the business of Maziv which would undermine the     in Maziv. 158 The merger parties further             -to-              preference for its own businesses or have any insight into the business of any other DFA or Vumatel customer or any planned DFA or Vumatel roll out. 159 172.             realities of corporate power dynamics and economic incentives. It relies inter alia 153 Clause 6.3.17 Bundle M p 846; Transcript p 1893 line 18 to p 1896 line 7. 154 Clause 6.2.5 Bundle M p 838 - 840; Transcript p 1898 line 22 to p 1899 line 16. 155 Clause 6.2.9.2 Bundle M p 842 and 843. 156 Clause 6.2.5.5 Bundle M p 839.Transcript p 1899 line 17 to p 1900 line 14. 157 Clause 6.2.5.6 Bundle M p 840; Transcript p 1900 line 15 to p 1901 line 5. 158 Transcript p 52 lines 6  13. 159 Transcript p 52 lines 13  18. Non-Confidential 50 on the evidence of Mr Van der Merwe of Frogfoot who submits that  hey [Vodacom] can veto just about anything. They have to approve the annual budget and the business plan. They have to agree to the CEO and CFO that ever gets appointed. Any sort of funding they can really kind of decline or so  . even if they acquire 30% of the         10% potentially being invested  I mean I do                    sides to           disappear whether they have control or not. I think there is economic incentives to engage in harmful      that. And then I think as well, I mean the letter of the law means one thing, but what happens in practice is different if [Vodacom]      do it. 160 173. MTN, through its legal representative, observes: 161 We pause to dispel a point that, with respect, is divorced from the reality of corporate life. The merger        -merger, Vodacom would have no influence over the operations of DFA and Vumatel. That it is a kind of benevolent investor seeking to release a promising business from the strictures of its debt burden. MTN is unconvinced. A corporate colossus of Vodacom standing would not entrust the arteries of its lifeblood to or invest up to 14 billion in targets at liberty to act as they see fit, including by granting all comers open access to Vodacom's Metro and Last Mile fibre assets in the hope that the targets would do a decent job with these resources. In the real world, in every matter of strategic significance, Vodacom would materially influence the decision of Maziv, which would materially influence the related decision of DFA or Vumatel 174. W           would align post-merger from a competition perspective. 160 Transcript p 129 line 8 to p 131 line 5. 161 Transcript p 29 lines 3  16. Non-Confidential 51 175. First, one can accept that     -merger would be different to their incentives pre-merger     and post-merger significant shareholding in Maziv. 176. Second, the shareholders agreement prohibits any shareholder (whether through its representatives on the Board or on any committee of the Board, or through its representatives on any subsidiary boards or committees, or directly in its capacity as a shareholder) from requiring or having regular meetings, reviews or reports from any operational members of the Group without the         allowed the same opportunity to participate in such process at operating subsidiary level 162 177. Third, both CIVH and Vodacom as shareholders may have representatives on                      shareholder) could require regular meetings, reviews or reports from any operational member of the group. 163 178. Fourth, an important issue           customer. Given that Vodacom will have a 30%-40% economic interest in Maziv,           Moreover, since     , if Vodacom grows, Maziv will have a larger anchor customer in Vodacom, in the context where there is a     ROFR        Vodacom always gives Maziv the right to match to supply Vodacom, which           thus has an economic incentive to give Vodacom preference and to grow   179. Mr Van der Merwe of Frogfoot submits that in terms of market dynamics    would be aligned post-merger and that 162 Clause 11.4 Bundle M p 783. 163 Transcript p 1889 line 6 to p 1890 line 21. Non-Confidential 52 Vodacom    could substitute any business lost by Maziv with more business from itself. 164 Furthermore, Maziv would have a real incentive to appease        s requirements at board level. 165 180. The abovementioned incentives could give Vodacom reason to limit the degree to which it competes against Maziv in a number of ways: 180.1. the more aggressively Vodacom competes with Maziv, the lower               result, Vodacom will be incentivised to, where possible, not compete or compete less aggressively with Maziv. This would be of particular concern in the low-income areas where both firms are set to roll out absent the proposed merger; 180.2. the incentive to lessen competition could be observed in rollout decisions where the quality (for example the density of towers) or extent of rollout may be affected; 180.3. Vodacom would have less incentive to develop, promote, and competitively price products in ways that would compete more directly against any Maziv product or service. 181. Further opportunities for partnership to the benefit of Maziv that emerged during the hearing include a potential FWA bundle for FTTB services 166 and the potential to use their very strong retail presence in the form of Vodashops for the distribution of prepaid vouchers for FTTH products. 167 182. Combined post-merger strategies may include that the merger parties could use             164 Transcript p 127 lines 1  9. 165 Transcript p 127 lines 7  9; p 130 line 19 to p 131 line 5. 166 Transcript p 2493 lines 2  8. 167 Transcript p 2424 lines 7  12; p 2718 lines 2  21; p 3307 lines 4  10. Non-Confidential 53 to offer customers a bundled combination of immediate FWA access at a discount to be followed in due course by fibre access which rivals may not be able to replicate. 168 Mr Masalesa of Telkom CSB echoes this concern. 169 183. Although the merger parties tender behavioural conditions in an attempt to deal with strategic roll-out decisions and information sharing concerns, MTN, Rain and others point out that if the transaction is implemented, strategic information on issues such as the future rollout of fibre could permeate corporate walls before finding its way onto websites. 170 Mr Van der Merwe of Frogfoot   So, I think first thing is, I mean, how do you even monitor that there is no whispers in the corridors, which you know I think is very real probability. 171 (Own emphasis) 184.      :                     discussions are being had 172 Maziv may in future if the transaction is implemented require additional capital investment from Vodacom. This means     should not be viewed as a one-off event,       185. Prof Theron argues that even if Vodacom has the incentive to benefit its own          cannot instruct   Maziv into conduct that does not benefit Maziv. 173 We disagree with this characterisation in the full context of the documentary and factual evidence. Vodacom has the ability, as co-controlling shareholder with a significant shareholding              broader interests because of their joint commercial and economic incentives. Mr Otty confirms        EBITDA- 168 Transcript p 34 lines 17  22; Mr Nunes p 632 line 5 to p 633 line 10. 169 Transcript p 329 line 13 to p 332 line 12. 170 Transcript p 34 lines 16  17. Transcript p 935 lines 13  16. Rain Intervention Application Founding Affidavit p 27  28 paras 77  82. 171 Transcript p 130 lines 11  13. 172 Transcript p 1674 line 14 to p 1675 line 14. 173 Transcript p 3896. Non-Confidential 54 AL, 174 operating profit and cash flow, and that it would seek to extract as much short- and medium-term profit as possible. 175 186. One must further consider       CIVH Board presentation clearly identifies areas of strategic rationale/benefits from the partnership opportunities with Vodacom to the benefit of Maziv and its operating entities, DFA and Vumatel and    Key Transaction Considerations       ; etc); l i i.e., Vodacom. 176 187. Furthermore, the post-merger joint incentives between Vodacom and Maziv would not be restricted to their activities in South Africa. The factual evidence that we shall deal with later in these reasons indicates that Vodacom has a fibre strategy absent the proposed transaction, including through pursuing off- balance sheet fibre   JVs. 188.     contained in its internal document titled Infrastructure Sharing  FibreCo and Rural Coverage. 177 Mr Otty testifies that It's possible that this was presented to a strategy meeting of Vodacom. We have an annual strategy meeting in every year. It looks like the sort of document that will be presented there. 178 The document sets o  174 EBITDA refers to earnings before interest, tax, depreciation and amortisation. 'AL' means 'after leases'. 175 Transcript p 2008 line 9 to p 2009 line 17. 176 Bundle M p 1269 and 1397. 177 The author is indicated as Mr Sean Bennett; the document   within the last two years according to Mr Joosub. 178 Transcript p 1987 lines 5  11. Non-Confidential 55 strategy on how it intends to approach for the setting up of a in different markets. 179 The document records: Investment structures  We are developing partnerships for off-balance sheet 3rd party funded investments. It includes and has the bullet points: (i) (ii) ; (iii) Up to will be held by Vodacom ; and (iv) n. Mr Joosub confirms that this is still the strategy. 180 189. Mr Joosub also confirms         talk about while we and the context of that is that you can off the investment 181 and                      to them about doing it with us in Africa  182 On the Maziv side, Mr Uys testifies that Ja, we would like to partner with MTN and Vodacom, and whoever, in                    our own. 183 190. In relation to strategic partners in Tanzania, Mr Joosub confirms that Vodacom wants to do a three-way deal including CIVH, with CIVH bringing their capacity, strength and know-how. 184       Strategic Intention    has committed to expanding its fibre footprint via FibreCo JVs in several markets 185 This document also reflects in relation to Tanzania that integrated teams have been formed with and in relation to the Democratic Republic of Congo and Mozambique that there is preliminary interest from . 186 179 Bundle M p 12455: Infrastructure Sharing  FibreCo and Rural coverage. 180 Transcript p 1720 line 16 to p 1723 line 5. 181 Transcript p 1649 lines 4  7. 182 Transcript p 1657 lines 1  7. 183 Transcript p 1276 lines 3  5. 184 Transcript p 1717 lines 10  20. 185               186 Bundle M p 12879. Non-Confidential 56 191. Notably CIVH, in setting out the reasons for doing this transaction to the board, highlights  unique potential to explore international opportunities; as well as the Growing demand across the continent: Meet the growing demand for connectivity infrastructure into the rest of the Africa continent focussed partnerships to market entry in where partners have a market presence. 187 192. Given the above it         in South Africa, even if at a cost to Maziv, may result in more investment by Vodacom with CIVH in JVs in other countries, creating additional income for Maziv. 193. The above serves          realise other benefits from the partnership. Realising these benefits may require preferencing Vodacom as a customer and partner over others, and may offset any loss, assuming there is, from any preferencing. 194. In light of the above strategic and other evidence,    of up to R14 billion in the targets, the merger parties witnesses version rings hollow that post-merger Vodacom would have no influence over the operations of DFA and Vumatel; that it is an investor seeking to release a promising business from the strictures of its debt burden. 15 Conclusion 195. The evidence shows that Vodacom will have extensive decision-making rights and powers at shareholder, director and even committee levels in relation to Maziv as well as its subsidiaries and controlled investee companies. 196. We conclude  30%-40% shareholding in Maziv would give Vodacom joint control of and extensive influence in, as well as strategic alignment with, the Maziv business by virtue of inter alia the suite of contractual 187 Bundle M p 1222. Non-Confidential 57 rights to direct the strategic affairs of the Maziv group. Vodacom would materially influence the decisions of Maziv, which would materially influence the related decisions of DFA or Vumatel and potentially even Herotel (discussed below). 197.                 -merger. One further has to         -merger through   the ROFR granted to Maziv on all Vodacom (future) fibre purchases, and the ability of Vodacom to offer side-payments through other       , as well as potential future capital provided by Vodacom. 198.                         MTN in      artificial         perspectives would remain discrete rather than intertwined 188 Herotel 199. As indicated above, Vumatel currently has a % shareholding in Herotel and intends to increase this to a greater than 50% interest (see paragraph 9 above). As further indicated above, the Tribunal issued a subpoena duces tecum requiring the CEO of Herotel, Mr Botha, to testify at the hearing and provide certain documents. After MTN negotiated behavioural remedies with the merger parties, it fell on the Commission to cross-examine Mr Botha on inter alia the current control structure of Herotel, the Herotel confidential, strategic information that Vumatel has access to and alleged existing co-ordination of roll-out activities between Herotel and Vumatel. One of the issues that arose during the hearing           or is budgeted to be used            in South Africa. 188 Transcript p 33 lines 5  7. Non-Confidential 58 200. Herotel is of relevance in this matter inter alia for the determination of concentration (market shares),      more recent roll-out decisions, and how Herotel affects any potential remedies, as will become apparent below. 201. The merger parties argue that given that Vumatel in their view does not currently control Herotel  and will not do so unless and until it obtains approval from the competition authorities for the recently notified transaction  Maziv is not in a position to use Herotel in the manner suggested by the Commission. The Commission and MTN however, after considering the factual and documentary evidence, specifically the cross-examination of Mr Botha, argue that Vumatel already controls Herotel. 202. In terms of the  tendered conditions and how they are affected by Herotel, the tendered conditions are aimed at preventing preferencing prior to any acquisition of control of Herotel. 189 Further, if Maziv obtains approval to acquire control of Herotel in due course, the merger parties proposed conditions require Maziv to fully integrate Herotel to the Maziv business model and make the Herotel network open access. 190 203.            evidence, submits that the circumstances surrounding the acquisition by Vumatel of its % shareholding in Herotel appears to be tantamount to the acquisition of control of Herotel, as well as prior implementation of the proposed merger with Community SPV. The issue of alleged prior implementation is, however, the subject of a separate investigation by the Commission and it has made no finding in regard to that. Likewise, we do not deal with the issue of alleged prior implementation in these reasons. 189 Clauses 3.2 to 3.5, 5.1 and 5.3 of the         Conditions. 190 Clauses 1.41 read with 4 and 5 of the Conditions. Non-Confidential 59 204. MTN after considering the evidence submits that since at least 2022, CIVH and/or Vumatel has enjoyed sole control of Herotel. Of importance to the             business operations of Herotel effectively to circumvent the tendered conditions and (at least partially) to foreclose MNOs, FNOs and ISPs. Our assessment 205. Relevant to our assessment is inter alia the extent of any economic interest and influence that Vumatel already has over Herotel (regardless of whether it amounts to control or not). Furthermore, the strategic documents of Herotel and CIVH/Vumatel shed light on if Herotel and CIVH/Vumatel have coordinated or potentially could coordinate their activities in the market(s) that they are active in, and the extent of that (potential) coordination. 206. Mr Botha confirms that Herotel and Vumatel compete in the same industry in terms of building, owning and operating fibre networks.    provided on a closed access basis. 191 Although Mr Botha alleges in his witness statement that Herotel and Vumatel are not close competitors, when questioned about it by the panel he concedes that Look, at the end of the day the product is connecting home to the internet. 192 207. In terms of who has control over Herotel, Mr Botha confirms in his witness statement           as follows: • Vumatel  % (since 24 August 2023 according to Mr Mare); 193 •         49.93%. According to Mr Botha, Herotel SPV was established to facilitate the funding for the acquisition of shares in Herotel, by Herotel SPV, on behalf of Community NPC; 194 and 191 Transcript p 1590 lines 18  19. 192 Botha Transcript p 3031 lines 9  20. 193 Mare Transcript p 2787 lines 9  11. 194 FWB p 561 para 17. Non-Confidential 60 • Non-management minorities  0.11%. 195 208. Despite the large shareholdings in Herotel by the abovementioned two parties, Mr Botha contends that Herotel is not currently controlled, jointly or otherwise, by any person or firm. 196 Mr Uys also denies that Vumatel currently has any control over Herotel 197     an associate investment. 198 209. As background, we note that Vumatel acquired its shareholding in Herotel in a series of transactions wherein it bought out numerous minority shareholders and i the Herotel management to establish the Community SPV. CIVH the establishment of Herotel SPV and Community NPC. Mr Botha confirms that he is aware of the arrangement. 199 The Commission contends that SPV may be a sham vehicle used by CIVH to acquire control of Herotel, without having to notify a merger with the competition authorities. 210. The evidence heard confirms that the Community SPV: (i) does not have any powers in terms of its MOI other than to acquire shares from Herotel, hold the shares, and sell them to CIVH; and (ii) has ceded all its rights to CIVH, including its bank accounts, its claims and shares. 200 Mr Botha testifies that SPV is               purpose for setting it up is to have a limited function 201 Furthermore, the directors of the Community SPV are all former and present executives of Herotel, Vumatel, and DFA. 202 Interestingly, none of them is from any group or community whose interests the Community SPV is purported to advance. 203 211. Mr Uys concedes that there has been no benefit to date to any community in which Herotel rolls out fibre. 204 195 FWB p 559 para 10. 196 FWB p 559 para 7. Transcript p 3034 line 10 to p 3036 line 6. 197 Transcript p 1324 line 20. 198 Transcript p 1589 lines 17  19. 199 Transcript p 2956 lines 10  16. 200 Transcript p 2959 lines 5  15. 201 Transcript p 2966 lines 5  14. 202 Exhibit S; Transcript p 2948 line 21 to p 2955 line 20. 203 Transcript p 2957 lines 6  10. 204 Transcript p 455 lines 10  15. Non-Confidential 61 212. Mr Uys could during his testimony not explain a number of issues relating to Herotel: 212.1. First, both Mr Mare and Mr Uys could not explain why CIVH in strategic documents maintains that CIVH retains % economies of secondary citiesif the notified merger with Community SPV were ultimately prohibited by the competition authorities: 205 ADV BERGER SC: No, but why will CIVH get %?       ADV BERGER SC: When its shareholding is at %.          ADV BERGER SC: Mr Mare?      206 212.2. Second, the inclusion of unredacted, detailed     board documents.        mistake 207 212.3.          performance, approval of matters concerning Herotel, a board pack          packs. 208 212.4. Mr Botha also testifies that Vumatel has access to   management accounts as a shareholder holding more than 10% of the shares. 209 He in response to questions from the panel indicates that these 205 Bundle R p 55  Overview of Project Legend decision points and outcomes; Mare Transcript p 2835 line 12 to p 2838 line 9; Uys Transcript p 1346 lines 3  6. 206 Transcript p 2836 lines 14  19. 207 Transcript p 1255 lines 4  14; p 1259 lines 1  21; Bundle M p 11432 to 11440. 208 Transcript p 1553 lines 8  15; p 1553 line 19 to p 1555 line 19; See also Exhibit Z  Minutes of meeting of Maziv, Dark Fibre and Vumatel Board of 16 March 2023 paras 4.1.5, 4.2.1.5, 5.1.1, 5.1.12, 6.1.1.1, 6.1.2.1 and 6.1.2.3. 209 Botha Transcript p 2995 line 16 to p 2996 line 2. Non-Confidential 62      number of stands it builds, referred to as Tier 1 and Tier 2 stands. 210 213. The documents relevant to Herotel in this matter include an internal CIVH management presentation dated July 2021 regarding Project Legend 211 relating to                    % of the shares in Herotel. This is defined in the presentation as the Proposed Transaction. 212  strategic rationale for this transaction as set out in the presentation includes: 213.1. in the third part: 213 Herotel as , and indicates that Herotel has a   s across South Africa, through its wireless and fibre networks rolled out to date which provides an attractive    214 213.2. that                              215 213.3. the presentation further sets out the [a]ncillary benefits of the Proposed Transactionincluding: (i)       216 and thus (ii) investing in Herotel                to 210 Botha Transcript p 3028 line 17 to p 3031 line 8. 211 Bundle R p 12  38 CIVH presentation of July 2021 regarding Project Legend (CIVH 1). 212 Bundle R p 13 CIVH 1 Slide 2 bullet 1, 5; p 31 Slide 20 block 1. 213       is set out in four parts. 214 Bundle R p 18 CIVH 1 Slide 7 block 3 bullet 1; Uys Transcript p 1465 line 4 to p 1466 line 7. 215 Bundle R p 18 CIVH 1 Slide 7 block 3 bullet 3; Uys Transcript p 1466 line 8 to p 1467 line 1. 216 Bundle R p 18 CIVH 1 Slide 7 block 4 bullet 1; Uys Transcript p 1467 line 2 to p 1468 line 7. Non-Confidential 63 . 217 214.       Proposed Transaction is similarly set out in four steps: 214.1. first, a          in in which Herotel already has a presence     which can  218 214.2. second, as regards the stated strategy  , competitor behaviour: in each secondary area where competitor activity and risk of overbuild is limited,        offering, rather than convert the existing footprint to a model   until such time as r offering is required due to other FNO activity in the area 219 while    220 214.3. third, as regards the stated strategy involving  t  214.3.1.       offering be rolled out in areas where Herotel and would seek to  221 (Own emphasis) and 217 Bundle R p 18 CIVH 1 Slide 7 block 4 bullets 3 and 4. 218 Bundle R p 19 CIVH 1 Slide 8 block 1 bullets 1 and 2. 219 Bundle R p 19 CIVH 1 Slide 8 block 2 bullet 1. 220 Bundle R p 19 CIVH 1 Slide 8 block 2 bullet 2. 221 Bundle R p 19 CIVH 1 Slide 8 block 3 bullet 1. Non-Confidential 64 214.3.2.               Herotel is able to effectively . 222 (Own emphasis) 214.4. fourth, a key priority post-transaction would be to       ensure that as much of the existing customer base can be   network to . 223 (Own emphasis) 215. As indicated above, we need not ultimately decide which firm(s) control(s) Herotel for competition law purposes, but note that the presentation records a legal     CIVH [would] gain control of Herotel pursuant to      224 216. Further Herotel-related documentary evidence includes a business plan         , 225 which reflects an aggregated budget per operating business within the Maziv group, including Herotel; 226 as well as an individualised budget per entity within the Maziv group, including Herotel. 227 Significantly, in      document contains forecasts of revenue and other metrics associated with the   driv[ing] penetration on FTTH network  specifically Core, Reach and Key             228 217. Mr Uys accepts that the presentation  numbers products      why it envisages t Vumatel products on . 229 222 Bundle R p 19 CIVH 1 Slide 8 block 3 bullet 2. 223 Bundle R p 19 CIVH 1 Slide 8 block 4 bullet 2. 224 Bundle R p 35 CIVH 1 Slide 24 bullet 4. 225 Bundle M p 12323  12396; Transcript p 1255 line 15 to p 1256 line 16. 226 Bundle M p 12332. 227 Bundle M p 12389  12396; Transcript p 1483 line 5 to p 1485 line 14. 228 Bundle M p 12330. 229 Transcript p 1479 line 7 to p 1483 line 4. Non-Confidential 65 218. This suggests that, as CIVH had strategised in July 2021, Maziv budgeted in FY2024 to use as well as products and/or services in . 219. A budget         , 230 again reflects an aggregated budget per operating business within the Maziv group, including Herotel; 231 as well as an individualised budget per entity within the Maziv group, including Herotel. 232               subscriber growth for each of Vumatel and Herotel 233 and forecasts revenue figures and other metrics associated with the strategy of    , products and/or services. 234 Mr Uys accepts that the   Herotel rolling out those products or at least                  235 220. The above evidence suggests that, as CIVH had strategised in July 2021, Maziv                         secondary areas. 221. Mr Uys concedes that Vumatel Herotel. 236 Mr Mare confirms that Maziv (through Vumatel) is 237 and concedes that there is little overlap in the footprint of Herotel and Vumatel with only homes 230 Bundle M p 11354  11442. 231 Bundle M p 11359. 232 Bundle M p 11432  11440; Transcript p 1487 line 16 to p 1488 line 20. 233 Bundle M p 11357; Transcript p 1486 line 1 to p 1487 line 15. 234 Bundle M p 11434  11435. 235 Transcript p 1489 line 22 to p 1495 line 11. 236 Transcript p 1323 lines 3  22. 237 Transcript p 2821 line 15 to p 2825 line 3; Bundle R p 13 and 17 to 19. Non-Confidential 66 overbuilt. This evidence would be consistent with the two firms coordinating on a complementary rollout strategy. 238 222. Dr Scheffer of Vodacom also confirms       infrastructure in unserved secondary areas where competition for the market still remains in terms of the land grab (discussed below). He agrees that post- transaction, Herotel, being an associate company of Maziv, could approach an unserved secondary area and offer a discounted bundle of FWA, FTTH and ISP services. 239 223. We conclude, based on the documentary and oral evidence before us, that Vumatel has been expanding through Herotel and is pursuing some of its strategies through Herotel. 224. Given the above, and that Vumatel has [80-90]% of the economics of Herotel as recorded in the strategic documents, for the purposes of market structure analysis we shall combine part of    an approximate [10- 20]% share with homes passed) with that of Vumatel. MARKET FEATURES AND CHARACTERISTICS 225. It is important that we first set out certain key characteristics of the South African fibre sector, specifically in relation to the FTTH market. FTTB and metro fibre shall be dealt with in a latter section. 226. These characteristics of FTTH are highly relevant to the assessment of the relevant counterfactuals, relevant markets and competitive effects. They are furthermore also highly relevant to assessing the roll-out commitments under the public interest analysis. 227. The key characteristics of the fibre sector in South Africa, specifically FTTH, include: (i) the so-    the first mover advantages 238 Transcript p 2896 line 15 to p 2897 line 6. 239 Transcript p 2505 line 9 to p 2506 line 13. Non-Confidential 67 enjoyed by the first FNO to provide fibre to a specific area; (iii) relatively low average   d; (iv) relatively low uptake of or penetration rates for FTTH; and (v) the growing demand for fibre/data in South Africa and South   income used for the internet. We consider each characteristic in turn. The land grab phenomenon 228. From the factual evidence it is clear that FTTH rollout in South Africa is characterised by the so-called land grab phenomenon, which has moved to the              South Africa (relating to the higher income areas) is saturated. 229. Mr Van der Merwe of Frogfoot confirms the land grab characteristic in fibre and           He explains that having covered almost all remaining homes in the Living Standard Measure LSM -10 categories (i.e., the wealthiest households) in South Africa, FNOs such as Vumatel, Frogfoot and Herotel have been expanding into the next category of areas  low-income areas, secondary towns and rural areas. FNOs and ISPs use different business models, products and payment plans to drive penetration and make these investments ultimately profitable. This is referred to         240 230. Dr Scheffer confirms that (i) the first land grab in South Africa related to the high value customers in the big metros; 241 (ii) all the FNOs focused their rollout in the first land grab on the higher income areas; 242 (iii) Vodacom has accepted that most of the metro areas in South Africa have been covered by fibre; 243 and (iv) in effect the first land-grab has come and gone. 244 He confirms that (i) Vodacom is now targeting high value customers in secondary cities and towns and that is 240 Van der Merwe FWB p 32 para 14. 241 Transcript p 2430 lines 12  16. 242 Transcript p 2441 lines 1  4. 243 Transcript p 2 440 lines 1  5. 244 Transcript p 2491 lines 3  8. Non-Confidential 68          245   the second land grab is still very much up for grabs. 246 231. He also confirms that in the unserved secondary areas there is land to be gabbed, after it being put to him that the primary areas are saturated. 247 Mr Mare similarly indicates that the Core (higher LSM) homes are penetrated. 248 232. Mr Nunes of MTN also confirms the second land grab and that it relates to the lower LSM areas  the so-called Vuma Reach area(s). 249 233. Mr Mare similarly confirms that the Core market segment in South Africa is  we moved away from the Core because the Core was penetrated  So, the Core, the 2.2 million homes in South Africa effectively covered   moving to Reach. 250 (Own                      251 234. Mr Joosub also confirms that fibre has been installed in most high-income areas in South Africa, where the majority of profits can be extracted. Vodacom sees no benefit in overbuilding the FTTH networks in these areas as the economic returns to be generated will not justify such overbuild. 252 235.         the Reach in secondary cities. So,                grab, yes 253       on the second land grab for Herotel it would be going into the Reach markets, meaning they go into the lower LSM areas between 30 and over R5 000.00 a month household income 254 245 Transcript p 2440 lines 6  11. 246 Transcript p 2475 lines 13  16. 247 Transcript p 2505 line 21 to p 2506 line 7. 248 Transcript p 2648 lines 19  21. 249 Transcript p 788 lines 15  18. 250 Mare Transcript p 2882 line 21 to p 2883 line 4. 251 Mare Transcript p 2928 lines 12  14. 252 Joosub FWB p 333  334 para 31.2. 253 Transcript p 2900 lines 13  16. 254 Transcript p 2900 lines 1  5. Non-Confidential 69 236. W      2023, Mr Mare confirms that the actions   Continue in areas to . 255 (Own emphasis) 237. Mr Uys agrees that aggressive land grab and overbuild strategies of competitor metro and backbone networks in secondary cities is still a threat today and it has always been a threat. 256 First mover advantages 238. A further market characteristic confirmed by the factual evidence is the so-called              geographic area will enjoy. 239. Mr Van der Merwe states that there are benefits to being the first FNO to invest in an area and clear disadvantages to being the second or third FNO to invest in a network. Where there is more than one fibre network in an area, the network         257 240. He notes that what drives the speed and breadth of the land grab is competition. More competition means more investment, and a faster deployment of fibre to these areas than what would happen with less competition - this is because of the first mover advantage. 258 241. Mr Van der Merwe further explains that the significant advantage to being the first to roll out fibre to an area is because it is less likely that another FNO will roll out fibre to the same area, at least in the short term and penetration rates will be higher for the first FNO. As markets mature and demand continues to grow, the investment in rolling out fibre infrastructure pays off. The FNO competes with the MNOs offering services in the area, but the lack of another 255 Transcript p 2857 line 4 to p 2858 line 5. 256 Transcript p 1209 lines 5  9. 257 Van der Merwe FWB p 38 para 27. 258 Van der Merwe FWB p 38 para 27. Non-Confidential 70 FNO ensures the investment is more likely to be successful. Given the long-term benefits of being the first FNO to invest infrastructure in an area, and the difficulty in making a profitable investment as a second FNO in the same area, there are strong incentives to deploy more quickly, hen     259 242. Mr Otty also explains why one wants to be the first fibre operator in a specific area and how that affects both uptake/    And you really need to get to something like a 40% penetration of your households passed. That means, houses connected over households passed in order to generate a return on your investment. Also, if you are the second, even if you are the second player in the market, the first player tends to get first player advantage and take up the customers that really want fibre first. So, as a second player the returns are less. You still have to make the 40% penetration level to                 . 260 (Own emphasis) 243. Mr Nunes states that it is accepted that there is a competitive advantage to the provider that is the first to deploy fibre infrastructure in an area. 261 244. Dr Scheffer states that Vodacom was loath to overbuild and it never made sense to Vodacom, frankly, to the entire market, to overbuild     the first mover into an area is absolutely important. 262 245.                  town where we identified an opportunity. If it gets published to all our competitors, you know, if we identified it,         mover there to get the opportunity to get some business 263  So, for us it gives us a competitive advantage           an underserved area, you know, we get a bit of a head start above our 259 Van der Merwe FWB p 39 para 30. 260 Otty Transcript p 1951 lines 12  20. 261 Nunes FWB p 156 para 6.14. 262 Scheffer Transcript p 2441 lines 5  11. 263 Schoeman Transcript p 933 line 20 to p 934 line 2. Non-Confidential 71              than to acquire them. 264 (Own emphasis) 246.     Vuma was always good to go in the market, we we go in with the , we cover that town super quick and  265 (Own emphasis) 247. In cross-examination Mr Mare makes it clear that strategy is to be the in an area and . Thus, there is a contest to be first mover to achieve the associated benefits: ADV BERGER SC: But the reason why you are so insistent on as as possible is because and then you .                MR MARE: Listen, you want     because        . . You want . 266 (Own emphasis) Relatively low average FTTH penetration rates actually achieved in South Africa 248. The factual evidence confirms relatively low average rates of FTTH penetration in South Africa. In other words, in the areas where FTTH has been made available to consumers through homes being passed, relatively low numbers of South African consumers actually take up the fibre alternative, although it differs per area/type of residence. 249. Mr Otty explains that fibre businesses incur significant capital costs upfront and take a long time, 7+ years, to generate returns because of the time required to achieve high enough penetration levels. 267 264 Schoeman Transcript p 934 lines 14  17. 265 Mare Transcript p 2590 lines 3  5. 266 Mare Transcript p 2929 lines 12  20. 267 Otty FWB p 358 para 15. Non-Confidential 72 250. Mr Van Der Merwe submits that most fibre operators, including Frogfoot, rely on terminal penetration of at least 50% to make their business cases work. The business case for a second or third FNO to invest in fibre infrastructure in a specific geographical area is therefore less likely to be compelling as it will be harder to achieve the necessary penetration levels. This is particularly true where the first operator is well established, with significant market penetration, as the second or third FNO will find it harder to achieve the necessary market penetration to justify the investment. 268 251. Where an FNO has access to existing infrastructure, and demand is particularly strong, the investment for a second or third FNO can potentially be possible. One example of this is Openserve which already has duct and pole infrastructure for its legacy copper network and thus rolling out fibre infrastructure by replacing its copper infrastructure is far less costly. 269 252. The above is consistent with the evidence of Mr Otty who says when you are building a fibre network, homes passed is the first metric, but you have got to achieve approximately 40% penetration fairly quickly otherwise you will make a loss, because of the high capital costs and consequent interest bill. 270 253. Dr Scheffer states that as a rule of thumb, an FNO requires more than 40% uptake (homes connected) of homes passed in an area to achieve a return on the investment involved in rolling out in the area (and potentially higher outside the higher income areas where Vodacom deployed to if customers take cheaper, lower speed products). If a second FNO with a similar cost structure also rolls out in the same area, this means that more than 80% of the households in the area need to be connected for both FNOs to break even. This is far less likely to happen as uptake rates of 80% are seldom achieved. 271 268 Van der Merwe FWB p 38 para 28. Transcript p 85 lines 8  15. 269 Van der Merwe FWB p 39 para 29. 270 Otty FWB p 359 para 17. 271 Scheffer FWB p 240 para 41. Non-Confidential 73 254. Dr Sheffer indicates that the average connection rate for Vodacom is %. 272 255. Importantly, he also accepts that 50% of South African homes passed by fibre are accessing the internet through means other than fibre. 273 256. Mr Mare explains that the initial first uptake of fibre will be relatively high but then    when I spoke about the uptake, what you normally see is your               you get your first uptake up to % you get in the first , and then it trickles out 274 257. Mr Mare indicates that Vumatel considers that it would generally need to achieve FTTH take-up rates of % in Vuma Core and Reach areas and % in Vuma Key areas for the business model to be successful. 275 258. Mr Nunes explains that penetration levels depend on the area/type of residence            depends on                 their HOA 276  277 He then indicates that the situation is different in the suburbs     door-to-     . 278 Penetration in        We normally see that in and around the  between the and little bit higher, % depending on the suburb and the amounts of communications 279 and concedes that in the suburbs only 50% of the homes passed are actually   I agree with you, at face value             of homes passed are connected] and 50% not. Okay 280 272 Scheffer Transcript p 2443 line 21 to p 2444 line 5. 273 Scheffer Transcript p 2454 lines 1  3. 274 Mare Transcript p 2928 lines 8  12. 275 Mare FWB p 439 para 36. 276 Home Owner Association. 277 Transcript p 912 lines 2  14. 278 Transcript p 912 lines 15  19. 279 Transcript p 912 line 20 to p 913 line 1. 280 Transcript p 915 line 17 to p 916 line 9. Non-Confidential 74 259. Mr Masalesa says that the main challenge with FTTH is its uptake or penetration. In terms of penetration numbers,    Vumatel has got a home connection rate of 31%. I think the highest connection rate that you have is with Openserve at %. 281 He also makes the point that the technology must serve the customers and that FWA has a much higher connectivity rate than FTTH:               technology, we do technology to service customers, right. And purely looking at the numbers you have much higher connectivity rate on FWA as opposed to FTTH               reality on the ground is the uptake is a different story. The uptake paints a different story altogether 282 (Own emphasis) 260.    fixed wireless technology because of its ubiquitous                                 after interpreting the signs, take decisive action to put up infrastructure and more                that infrastructure. So, if I may call it the hit rate on fixed wireless access is much higher  283 261. Mr Van Der Merwe indicates that as of end February 2024 where Frogfoot is not overbuilt (i.e., where it is has no fibre-to-fibre competition), its median penetration rate is 6%, while this penetration rate falls to % where there are one to three other FNOs overbuilt. 284 262. Mr Mare testifies that the average penetration rate for fibre in South Africa                              roughly the uptake is about 48%, if you look at it 285 281 Transcript p 345 lines 8  11; p 353 lines 12  22. 282 Transcript p 345 lines 11  18. 283 Transcript p 468 line 18 to p 469 line 5. 284 FWB p 39 para 29.1. 285 Transcript p 2891 lines 7  10. Non-Confidential 75 263. At the end of December 2023, the number of active premises serviced by ISPs using the Vuma Reach product amounted to , with over homes passed. 286 This indicates that a relative low percentage of homes passed by Vumatel are actually connected to the fibre that has been rolled out. 264. As at 18 September 2023, Africa Analysis reports that a total of 5.5 million homes have been passed with FTTH in South Africa. Of this total number of homes passed with fibre only 1.95 million homes are connected. The total number of overbuilt homes is only 1.51 million. 287 Low fibre overbuild in practice 265. Overbuild refers to the duplication of fibre infrastructure by two or more fibre infrastructure providers that have laid their own fibre optic cables in the same area or even in the same roadside trench. Where there is more than one fibre network in an are           266. Where a fibre provider is the first to lay infrastructure in a particular area, it typically achieves a monopolistic position in that area unless another provider is willing to overbuild on its network. 288 267. Mr Nunes indicates that the willingness of providers to overbuild is dependent on several factors, including customer take-up rates or demonstrated interest, poor performance from another FNO, single-trench policies in certain precincts, and Home Owner Association preferences regarding multiple fibre providers within an area. 289 268.        are reluctant to overbuild on the networks of others       investment; and (ii) duplicating infrastructure reduces the return on investment 286 Mare FWB p 441 para 47. 287 FWB p 253. Africa Analysis Report, FTTH Market Tracking Programme, Quarter Ending June 2023 (Updated 18 September 2023). 288 Nunes FWB p 155 para 6.7. 289 Nunes FWB p 155  156 para 6.10. Non-Confidential 76 for fibre providers since revenues are dispersed amongst the providers in an area. As a result, providers often prioritise areas that are underserved to maximise investment returns. 290 269. Mr Mare indicates that Vumatel does not consider it economically feasible to overbuild, in light thereof that where there is overbuild, there is not the required take-up per FNO to show an acceptable return on investment. 291 270. Dr Scheffer testifies that Vodacom does not overbuild anyone. 292 271.       It's not our strategy to overbuild. We specifically do not intend to overbuild              immediately        293 272. Mr Nunes indicates with reference to FTTH, that only 18% of the total number of reported homes passed by FNOs has been overbuilt. 294 He is also referred to a statement of his attorneys reflecting that only 23% of FTTH is overbuilt and indicates that in that case overbuild will primarily be by Openserve because they use their old copper network  blow fibre  295 Growth in demand for data and consumers’ disposable income 273. All the fibre markets relevant to this transaction are poised for substantial growth,          /towns and lower income areas, FTTB through business broadband extension to outlying business areas and secondary cities/towns, FTTS to support the rollout of 5G on mobile networks and metro fibre backhaul to support all of these initiatives. 290 Nunes FWB p 156 para 6.13. 291 Mare FWB p 439 para 36. Transcript p 2592 lines 11  14. 292 Scheffer Transcript p 2443 lines 13  20. 293 Botha Transcript p 2944 lines 3  8. 294 Nunes FWB p 16 para 6.11. 295 Transcript p 756 lines 5  15. Non-Confidential 77 274.              increasing and accelerating. The importance of data in an increasingly digital economy is recognised by all parties. In that context both data/internet access and future prices are important. 275. Consumers in higher income communities in South African have access to the internet through fibre connections. There however is currently a significant deficit in the ability of South Africans in low-income communities to access similar opportunities. As indicated above, the factual evidence is that the high-income fibre areas in South Africa are saturated and that FNO competition has now moved to the lower income areas. It is important to note that this transaction is proposed at a time when FNOs are looking to expand into the lower income areas. These lower income areas are currently mostly supplied by MNOs supplying mobile broadband and FWA home internet products. Further context to note is that MNOs have received spectrum to roll out the latest 5G technology, with FWA as the only use case currently to get a return on their spectrum investment. 276. The merger parties argue that the proposed transaction will contribute to bridging the digital divide by reducing data costs and bringing fibre coverage to areas previously not connected. The Commission, on the other hand, argues that the proposed transaction will harm competition, and that it is competition in the relevant markets that will ultimately lead to the roll-out of infrastructure and cheaper prices for consumers. 277. Mr Motlekar submits that the digital divide is not based on technology or tech. His view is that We create a gateway. Allow you to move up the price points.                                           by the customers 296 We shall assess the FWA and FTTH competitive interaction below. 296 Transcript p 559 line 17 to p 560 line 1. Non-Confidential 78 278. The factual testimony confirms the market growth and opportunities. We give some examples: Mr Van der Merwe confirms that the fibre sector continues to develop rapidly. 297 Mr Masalesa confirms that the demand for data will continue to grow in South Africa and indeed accelerate in the years to come. 298 Dr Van den Bergh  data is growing so that demand in general is everywhere 299            We [Vodacom] have a right to play in this market. We can become the preferred provider to meet the growing demand 300 Mr Uys agrees that there is a growing FTTH market with potential: you can see on both the quarterly and also the annual that there is definitely still growth in the market and the            and versus the Vumatel quarter of . So, there is potential in the market  301 279. In terms of what South African consumers spend monthly on the internet, a study conducted by BMIT, a technology industry research and advisory firm, shows that 75% of South Africans have a spend of R500 or less for internet services. Of that 75%, 50% only have the ability to spend R300 or below. 302 280.        The numbers are tight and my view based on that is you know the two technologies; both fibre and fixed wireless access technology are competing for a share of that wallet 303 We shall assess FWA and FTTH competitive interaction below. 297 FWB p 32 para 14. 298 Transcript p 344 line 19 to p 345 line 1. 299 Transcript p 2281 lines 7  8. 300 Transcript p 1836 lines 1  5. 301 Transcript p 1139 lines 4  12. 302 Masalesa Transcript p 298 line 18 to p 299 line 4; p 349 line 11 to p 350 line 11; p 462 lines 7  9. 303 Transcript p 299 lines 4  6. Non-Confidential 79 COUNTERFACTUALS 281. The disputed issues in this case include the relevant counterfactuals. The counterfactuals to the proposed transaction are important to both the competition and public interest assessments that follow. The counterfactuals          and   counterfactuals. We deal with each in turn. Competition counterfactual 282. Based on the documentary and factual evidence, the Commission argues that the counterfactual to the proposed transaction is a world where Vodacom increasingly puts itself in competition to both DFA and Vumatel. The merger parties disagree. 283. Highly relevant to this assessment is the (true) rationale for the proposed transaction that we have dealt with in paragraphs 105 to 139 above, which must be read together with this section. 284.         in these reasons, has a strategic imperative to look for an expansion in fibre (including FTTS), to expand its network for 5G purposes and to densify and fiberise existing 4G sites 304 and FTTH/B to share in the profits in these markets. 285.             to the Value at Risk to its mobile business, estimated at R11.8 billion in the           305 where Mr Joosub estimates that there could be a loss of up to 30% of mobile data spend within a household if they move to fibre, although he sought to claim it was less than the  Consumer VaR estimated to be R11.8bn (Euro570mn) for the total 304 Bundle M p 3498. 305 The figure of R11.8 billion was presented to the Vodacom Board. See, for example, Transcript p 2003 lines 2  12. Non-Confidential 80 period up to FY24 306 Vodacom wants to compete in fibre to compensate for this VaR, for revenue generation and financial gain. 307 286. Rec                            116 and 12019 above).         Vodacom establishing both a TowerCo and a FibreCo   Vodacom    Focus on and cover where there is no fibre. 308 287. Vodacom sees an opportunity to participate in the fibre market rather than being a customer to infrastructure players such as where it does ,      model). 309 288. Project 310 , albeit that certain of its assumptions are disputed      ses in the hearing, still indicates that Vodacom can , either with or f or through . 311     contemplated investment and expansion shown in Project indicates that Vodacom wanted to create a Maziv in both FTTH and FTTB. 312 Vodacom had : the plan was to become the FTTB player and the t FTTH player, and to operate model. 313 Mr Joosub confirms that Project T projected to the proposed transaction, although he contends that the assumptions . 314 We note that other than factual witnesses disputing the assumptions made in Project in oral 306 Bundle M p 3498. 307 Otty FWB p 360 par 18 to 20; p 362 and 363 para 25; Transcript p 2012 line 16 to p 2014 line 4. 308 Bundle M p 2702. 309 Bundle M p 12854. 310 This is based on a self-build plan with 100% Vodacom ownership. 311 Bundle M p 12483: Project  VSA Fibre rollout acceleration, 21 May 2021. 312 Bundle M p 12475 and 12476. 313 Bundle M p 12475. 314 Transcript p 1760 lines 4  11; Bundle M p 12482. Non-Confidential 81 testimony at the hearing, it produces no contemporaneous documents (e.g. a minute of any meeting or discussion or any other document) indicating which of the assumptions made in Project were wrong (and which not) and, if wrong, how they were wrong, and if corrected, what the position would be. 289. We further note that   strategy document of 2021  containing Mr Joosub  identifies opportunities which will deliver significant and        315 290.       d that Vodacom would not be willing to invest in its own fibre infrastructure absent the merger, as its shareholder, Vodafone, would n   -build fibre projects. They argue that this is because its main business is that of an MNO and it does not have the capability to build, especially in lower income areas. Absent the merger, as they argue, Vodacom would not scale-up its fibre network and it would continue to lease. 291. We do not accept the above assertion on a thorough consideration of all the evidence, including the       evidence t       . What the merger             network infrastructure to reduce its costs. In this strategy, it is looking to enter into JVs and partnerships so that it can participate in fibre, and not just lease. 316 Mr Joosub confirms this strategy, and that the strategy remains absent the proposed merger. He confirms that    s but that  . 317 315 Part A of the Record p 3807 Project   Strategy, September 2021, Shameel Joosub. 316 Transcript p 3993 lines 10  17. 317 Transcript p 1765 lines 12  15. Non-Confidential 82 292. Mr Joosub furthermore confirms that in the counterfactual, to the extent that DFA fails to build enough FTTS, Vodacom would find alternatives to DFA, whether through self-            continue to use DFA and so on? As long as it make sense for us we will continue to                   to then either self-build it or go and build it with somebody else. But Fibre to the Site needs to happen. So, that essentially will        is a need that will be there  318 and         to the Home versus Fibre to the Site. Fibre to the Site is creating a shared infrastructure path and to the extent that the CIVH has the capability to deliver that we will continue to do business with them. And to the extent    we  and we need the path, we would then have to seek alternatives. 319 293.            R23.7 billion in cash and a nett debt to EBITDA ratio which improved from 1.1 to 0.9. 320 Mr Joosub further testifies that there is room to take on more debt since              321 Mr Otty of Vodafone testifies that banks and shareholders consider factors such as the stability of a company, which Vodacom is, when deciding to give a business access to funding. 322 294. Notably, Mr Joosub explains that, as is evident from Vodacom  documents, it can do o  we build and the context of that is that you can the investment. What does that                           calculation and so you can go and look for           at in some of the other entities is 318 Transcript p 1771 lines 2  7. 319 Transcript p 1774 lines 1  6. 320 Exhibit AB1  Vodacom Group Limited Reviewed Annual Results and cash dividend distribution, Slide 18; Exhibit AB2  Vodacom Group Annual Results for the year ended 31 March 2024, Slide 28; Transcript p 1700 lines 7  13. 321 Transcript p 1700 lines 16  20. 322 Transcript p 1996 line 19 to p 1997 line 2. Non-Confidential 83 and then and so then basically that stays . So, when you look at the Vodacom , that will not be in and then when you         as well. 323 (Own emphasis) 295. Mr Joosub does not contest that Vodacom can find other fibre partners to roll out fibre in South A absent the proposed transaction), but testifies that it would not suit Vodacom because it will be a very small play and you would  coming late to the party 324 This is inconsistent with the abovementioned       296. Significantly, the current transaction is off-balance sheet. Mr Joosub explains the benefit of this, so your debt          revenue, but you            as we call it. 325 Vodacom has been able to raise billions of Rands off-balance sheet for this transaction. 297. Dr Scheffer also confirms that Vodacom does have capital available for fibre investments in the form of the R14 billion to R19 billion currently set aside for this proposed transaction: ADV BERGER SC: But there is capital available, and that capital is the 14 to R19 billion which is being set aside for this deal? DR SCHEFFER: Yes  326 298. Regarding skills and capacity as a potential barrier, Mr Otty claims that, now that most of the high income areas are covered, Vodacom does not have enough skills or capacity and any model to expand into low-income areas which would require a different way of doing things. 327 Dr Scheffer however concedes that the skills and know-how required to roll out successfully in the lower income 323 Transcript p 1649 lines 4  16. 324 Transcript p 1763 line 18 to p 1764 line 5. 325 Transcript p 1671 lines 7  11. 326 Transcript p 2474 lines 12  18; also see p 2478 lines 10  13. 327 Transcript p 2020 lines 7  11. Non-Confidential 84              yes, we can; 328 Yes, that's correct odacom could acquire the skills and the know-how for rolling out fibre successfully in low-income areas. 329 299.        credible that Maziv is the only entity capable of in the future rolling out fibre with sufficient scale. 330 The Project Discussion Materials dated September 2020 identify , and  good, challenger  331 It is also contrary to the        , as discussed under the true rationale. 300. As also          world without the proposed merger lays bare the counterfactual of real and          where it discusses the threats of not doing this deal articulate this counterfactual. The documents reflect that CIVH was anxious to conclude a deal with Vodacom              becoming a significant competitive force against both DFA and Vumatel, possibly triggering further competition with other FNOs as they conclude similar deals to counter Vodacom. 301. Recall further that            competition and saw the proposed merger as likely to de-   When questioned about whether the transaction de-risks the core business of CIVH by taking away the risk of losing Vodacom as a client and gaining it as a formidable competitor Mr Uys testifies: 332      saying. So, in light of these scenarios that management put to us, they said that this will  this deal will also commercially de-risk the core business of CIVH because then they hopefully will keep their business with us. (Own emphasis) 328 Transcript p 2476 lines 6 to p 2477 line 18. 329 Transcript p 2477 lines 15  18. 330 See Hodge Transcript p 3575. 331 Bundle M p 3544. 332 Transcript p 1229 lines 10  13. Non-Confidential 85 302. The evidence furthermore indicates that Vodacom has in the previous five years. This is largely because, as Vodacom was still pursuing the proposed transaction, it expanded its network by completing the construction of FTTS links and km of metro fibre, seemingly to keep its self-supply alternatives open in case the deal fell through. 333 This infrastructure would compete with Maziv absent the proposed merger. 303. In summary, the documentary and oral evidence confirm that Vodacom is a competitive threat to Maziv, and we conclude that the counterfactual to the proposed transaction is a world where Vodacom increasingly puts itself in competition to Maziv. Investment and fibre roll-out counterfactual 304.             to roll out fibre faster and to speed-up the process of bridging the digital divide. 305.             to funds to roll out fibre as it has done in the past, and even if it does not, the evidence shows that other market players will roll out fibre in any event. 306. We first consider investment and then roll-out. Investment 307. The Commission argues that the relevant counterfactual is that, absent the merger, the rollout into low-income communities will occur anyway. Either Maziv will obtain the necessary funding from other sources to finance further FTTH           in South Africa will do so in its stead. 333 Hodge EWB p 117 para 179. Non-Confidential 86 308. The merger parties argue that currently, Maziv is c , which limits its ability to pursue its F           and enable it to roll out fibre faster, enabling it to narrow the digital divide. 334 They argue that without the merger, as conceded by Mr Hodge, Maziv cannot draw on its . 309. We note that Mr Hodge concedes that Maziv has a at the moment but that it that until it meets certain and is able to then, in future, on it once it does meet those . 335 310. The merger parties argue that the relevant counterfactual is the status quo since as a result of having 336 and having no , Maziv cannot pursue its plans to pass homes in low-income areas. Instead, it will have to focus on connecting homes in middle- and high-income areas (which have already been passed with fibre). 337 311. Regarding further funds from existing shareholders, Mr Uys testifies that the and   would be to put in the money and that's why there would be a delay. We'll have to go and find  338 (Own emphasis). When asked if this was discussed at the , he says that he has only  339 312. The merger parties do not contend that Maziv cannot seek alternative investors. They argue that Maziv will have to try to find new investors. However, they say that will take a number of years, and there is no guarantee that it will be successful. 340 They rely on     I can't say we can find it. I will try my best to find one, but I can't guarantee that we will find one 341 They 334 Uys Supplementary Witness Statement FWB p 548 para 11. 335 Transcript p 4024 lines 4  19. 336 Uys FWB p 486 para 59. Transcript p 1593 line 19 to p 1594 line 10. 337 Uys Supplementary Witness Statement FWB p 551 para 23. 338 Transcript p 1593 lines 19  22. 339 Transcript p 1594 lines 5  10. 340 Merger Parties Heads of Argument (HOA) p 115 para 219.2. 341 Transcript p 1593 lines 14  17. Non-Confidential 87 also rely on       in which it sought to from i and 342 313.     t other firms in the market will roll out fibre, the merger parties submit that no other competitor besides Telkom/Openserve can roll-out fibre at scale and pace  and Telkom/Openserve has never led the market in penetrating low-income areas given the risks involved. Our assessment 314. The evidence shows that Maziv has up to recently always found capital to fund its infrastructure rollout plans. 315.    r 0, the evidence is that they had a plan of action to r and prepared and executed a further R in 2021 to and support the growth of the business. 343 That capital supported the of DFA and the of Vumatel in FY2023 and FY2024. 316. Along with the above was a further R1 to acquire     CIVH provided funding     b sitting there that we have to r , so part of that R to Vumatel to acquire the 49% in Herotel 344 317. Although the merger parties claim that because of     existing shareholders could not provide further funding to the business absent the merger, the evidence is that they did provide significant further funding while the merger was under consideration. This capital injection by shareholders supported R in capex (of the R10 billion proposed in the   342 Uys Transcript p 1108 lines 1  16. 343 Transcript p 1189 line 19 to p 1190 line 12. 344 Transcript p 1331 line 18 to p 1332 line 1. Non-Confidential 88 initial tendered condition) and delivered on almost homes passed since 1 April 2022 (that the parties initially made a condition to the merger). 345 Thus, the previous roll-out commitments (that were claimed to be a benefit of this merger and were to be achieved over 5 years) were met (without this deal) and in a relatively short space of time. 318. Furthermore, CIVH refinanced its debt and renegotiated and extended its debt facility with the banks in December 2023 to R25 billion. This resulted in an additional R6 facility above its current debt levels, accessible on meeting its debt covenants. 346 With regard to meeting the c for September 2024 and March 2025, Mr Uys confirms    has sought . 347 He testifies that the banks have  they are     a good chance that the banks will give Maziv or so 348 319. Furthermore, Mr Uys and Mr Mare indicate that Maziv through concentrating on connecting the homes passed rather than new rollout (as it has been doing), will drive EBITDA and make it easier for Maziv to meet its debt covenant requirements. 349 The presence of the facility means that the R6 billion is not all required for capex. 320. We also note that Maziv has    to take out some of the money that comes in       that was going to be used as a      . 350 321. The worst-case scenario absent the proposed transaction, according to the evidence of Mr Uys, is that      , although he 345 Transcript p 1354 lines 12 to p 1355 line 5; Hodge EWB p 165 para 317. 346 Transcript p 1121 lines 1  15; Bundle M p 10590. 347 Transcript p 1124 lines 2  12 and p 1142 lines 11  16. 348 Transcript p 1124 lines 13  17. 349 Uys Transcript p 1123 lines 9  11; and Mare p 2926 lines 13  22. 350 Transcript p 1118 lines 7  13. Non-Confidential 89 does not explain why it would take that l We refer to the following exchange between panel member, Ms Kessery and Mr Uys: ADV KESSERY: And then to just get to what I call the counterfactual, which is if the deal doesn't go through, if the transaction is not approved. You spoke about your vision and your strategy, and you said that you will find a way to do it and you will find funding elsewhere. So, I'm taking that you're not saying that it won't happen; you're just saying there will be a delay in it happening. MR UYS: Yes. It will take another  351 322. With regard to potential investors in Maziv,     at the time of contemplating an external investor in CIVH was to source investment in the market in return for a minority shareholding in CIVH. 352 Mr Uys confirms that  there was definitely interest from local investors 353 Furthermore that there were discussions with MTN as well, not just about South Africa but the rest of the continent. 354 Notably, Maziv however only engaged with Vodacom post-Covid. 355 323. Even when market conditions had changed post-Covid, and it had raised R b from its shareholders, CIVH did not consider other potential investors. We note that w            the evidence confirms that there was interest from multiple investors of under R1 324. There is no evidence to suggest that absent the merger, Maziv would not continue to look for investors, but only that it would take time. 356 On the question of Mr Uys not being able to guarantee an investor, he puts up no evidence of having made a serious attempt since talks initially started with Vodacom years ago, even during the period when a deal could not be reached that Mr Joosub 351 Transcript p 1592 lines 10  16. 352 Transcript p 1179 lines 1  15. 353 Transcript p 1180 lines 18  21. 354 Transcript p 1181 lines 7  16. 355 Transcript p 183 line 20 to p 1184 line 4. 356 Transcript p 1593 lines 5  18. Non-Confidential 90 was upset about. In the context of a growing sector, with rapidly growing demand for data, and given the skills and expertise of CIVH as the market leader and largest incumbent in fibre, together with attractive margins and limited overbuild generally in FTTH, we see no reason why external investors would not be interested in a stake in Maziv. Vodacom certainly sees its investment in Maziv as one that will reap attractive benefits and profits and Maziv puts up no evidence to the contrary. 325. Furthermore, given the market characteristics (as discussed above) Maziv will be incentivise to seek an investor in order to achieve its rollout goals since FTTH is a market where there is so-called competition for the market and significant first mover advantages, with limited overbuild  as borne out by the factual evidence. If Maziv does not roll out FTTH, it will lose significant market share to its competitors actively involved in the second land grab and these areas would then be permanently sterilized for Maziv given that it would be reluctant to overbuild. This will significantly incentivise it to speed up any external investor seeking process. Fibre rollout 326. We have above dealt with the key market characteristics of FTTH in South Africa, as confirmed by the factual evidence. This includes (i) competition for the market; (ii) the second land grab phenomenon; (iii) significant first mover advantages to the first FNO to roll out fibre in an area; and (iv) limited overbuild by other FNOs (see paragraphs 22765 to 272 above). Given the combination of all these factors, this means that if Maziv fails to roll out fibre into the lower LSM areas absent the proposed transaction, other operators will as part of the second land grab to achieve the first mover advantages. 327.           for the market and what the consequences would be for Maziv if it stops rolling out fibre. Mr Joosub states that, were Maziv to stop rolling out fibre in a particular Non-Confidential 91     fibre will go to those areas and then effectively CIVH has completely lost out 357 (Own emphasis) 328. This was echoed by Mr Otty who stated that without this merger in areas which  ripe for fibre              would roll out very soon thereafter and capture the       358 (Own emphasis) 329. Mr Mare explains the fluidity of the market and what happened in practice in the      basically what we saw in Reach is the competitors then copy the solution and then suddenly they start building at speed as well               this. If you get a workable solution then basically the competitors follow suit. So, basically we started in 2019 with our Reach. In 2021 two of the biggest operators were building Reach at scale. And if you look at it today everybody is building Reach at scale         . 359 (Own emphasis) 330. Mr Mare further indicates that Vumatel is a significant player with homes passed, and that competitors are active and collectively have passed 1.2 million      all the competitors, but I think what I can say is we launched Reach or the prepaid product in 2019. Our competitors started following us only in 2021. We got homes in I think in  we got homes passed at this point. I think our competitors or the whole rest of the market is about 1.2 million roughly the figure I have, yes 360 (Own emphasis) 331.            aggressive in this market at the front 361 357 Transcript p 1829 lines 8  11. 358 Otty FWB p 365 para 34.2; Transcript p 2016 line 19 to p 2017 line 11. 359 Transcript p 2893 line 17 to p 2894 line 4. 360 Transcript p 2580 lines 7  12. 361 Mare Transcript p 2756 lines 1  2. Non-Confidential 92 332.    If you look at today, I mean on par people are building Reach more than us at this point 362 (Own emphasis) 333. He also gives the example of Soweto where O l and are  in Soweto . 363 334. Openserve we have heard is more likely to overbuild because it has exclusive access to an existing duct and pole infrastructure that was previously used for     -reaching copper network. This allows it to install fibre on existing infrastructure and avoid some of the construction costs incurred by other FNOs. 364 Openserve is part of the second land grab and would be able to roll out fibre in the Reach areas. 335. Mr Mare further indicates   you look at Stats SA and you look at all the            seeing a very positive, lucrative economy in the Reach area  365 336. Mr Joosub testifies that the fibre train will happen regardless of the proposed  the fibre train is going to happen anyway    Vuma              ,                over the world           do    . 366 337. Mr Van der Merwe of Frogfoot testifies that in the counterfactual world if the          step up and say we want a slice of the pie        catch on to the opportunity that exists to          362 Transcript p 2894 lines 12  13. 363 Transcript p 2869 lines 4  7. 364 Scheffer FWB p 240 para 42. 365 Mare Transcript p 2596 lines 2  10. 366 Transcript p 1641 lines 6  11. Non-Confidential 93 townships 367 If competition is allowed to play out over the next couple of years, it will be disruptive and it will change that market create jobs    positive things which the merger parties are saying they are going to do in circumstances where this will happen anyway without the merger. 368 338. Given that Maziv is the largest fibre player in South Africa, and the undisputed market characteristics of (i) competition for the market; (ii) the second land grab; (iii) significant first mover advantages; (iv) coupled with the fact that Maziv does not overbuild, as confirmed by the factual testimony, we are unconvinced by the          for a significant period of time just because it did not get the cash injection from Vodacom through this transaction. In any event  s evidence reveals that if this merger does not proceed     rollout as it 339. Mr Otty agrees that it does not matter to the consumer who provides the fibre and that rollout will happen without the proposed transaction, but contends that        fibre rollout would happen more slowly without the Vodacom investment in Maziv 369 Mr Otty however concedes that he does not know the capabilities of other players such as Openserve and other FNOs. 370 Competitors   obviously be influenced by whether or not the proposed Vodacom/Maziv deal proceeds or not. 340. The merger parties also assert that Maziv has a track record of pioneering the rollout of fibre in new areas. Vumatel was the first to roll out FTTH. They claim that Openserve/Telkom was not doing it despite their previous monopoly and the availability of their copper network. However, the factual evidence, as discussed above, is that competitors have caught up and are now seriously competing in the second land grab in relation to the Reach areas. 367 Transcript p 285 lines 6  14. 368 Transcript p 285 line 15 to p 286 line 2. 369 Transcript p 2018 lines 9  12. 370 Transcript p 2076 lines 1  8. Non-Confidential 94 341. The merger parties also claim that none of the other FNOs has managed to develop a commercial model to roll out at scale into the lower income areas  when Vumatel rolled out in 2019, the other operators followed only much later, in 2021. That Vumatel did so first in our view is not material  what is relevant is that the other operators followed and are now actively competing to provide FTTH access in the Reach areas. 342. Mr Reynolds concedes under cross-       out to Reach, other players will start taking some of those addressable  some                customer base of Maziv, ja, but that could take a process 371 343. We conclude that the relevant counterfactual will be a combination of all interested market participants, including Maziv without the proposed deal, competing for the land grab in the lower-income areas. This applies specifically to the Reach market segment. We shall deal with the Key segment under the public interest analysis. 344. In relation to the Key segment there is no player at scale at this stage. Mr Mare indicates that there are different players in this market with different solutions at this point. He does not think that there are big players at scale at this point, and Vumatel is also not at scale. 372 MARKET DELINEATION 345. There is consensus between the parties that the broad markets for wholesale metro fibre and last mile fibre can be further segmented into the following narrower relevant markets: 345.1. within the metro fibre market, there are separate relevant markets for (a) FTTS, wholesaled to MNOs; and (b) fibre backhaul, wholesaled to FNOs; and 371 Reynolds Transcript p 4111 lines 9  17. 372 Transcript p 2755 lines 16  19. Non-Confidential 95 345.2. the broad last mile fibre infrastructure market can be further delineated into separate relevant markets for (a) FTTH; and (b) FTTB, wholesaled to ISPs because of the different needs of household and SMME customers who typically consume FTTH versus enterprise customers who consume FTTB. 346. The first contentious market definition issue is whether dark and lit fibre are part of the same relevant market or distinct markets in the provision of wholesale metropolitan backhaul to MNOs/FNOs and last mile fibre (FTTH and FTTB) to ISPs. 347. The other contested market delineation issue is if retail FTTH services and FWA services are part of a broader market or in totally separate relevant markets. 348. We first consider the issue of dark vs lit fibre. Dark vs lit fibre 349. Dark or passive fibre is network fibre that has been installed but not yet turned on by a network provider. It is only once the dormant cables are lit that data can be transmitted through the cables by pulses of light. 373 Lit fibre, also known as active fibre, refers to fibre that has been turned on by the network provider using specialised equipment; it is already operational and being used to transmit data by pulses of light. 374 350. Dark fibre at long-haul (to an extent), metro and last    the             375 373 Nunes FWB p 132 para 2.7. 374 Nunes FWB p 132 para 2.8. 375 Nunes FWB p 153 para 6.3.1; p 165  167 paras 6.51.1  6.51.4. Non-Confidential 96 351.             dark fibre from DFA are precluded from reselling (i.e., sub-letting) such dark fibre. 352. Vodacom does not currently offer dark services on any of its fibre infrastructure, with the exception of one link to each of MTN and FibreCo. In addition to its self- provided fibre, Vodacom leases dark and lit fibre where self-supply is not viable, mainly from DFA, and , which it uses as an input into its downstream fixed and mobile network services. 353. The Commission defines separate relevant markets for dark and lit fibre. 376 MTN submits that dark and lit fibre are not substitutable products or services, do not place significant competitive constraints on each other and are therefore in separate relevant product markets. 377 Rain submits that its experience supports the finding of a separate relevant product market for dark fibre. 354. The merger parties submit that the Tribunal can leave open whether dark and lit fibre are in the same or separate relevant markets, because it does not alter the competitive assessment. 378 Prof Theron clarifies: we are very happy to work within an assumption that there is a separate dark     impact on our theories of harm. 379 They argue that (i) since Vodacom does not offer dark fibre products, the proposed transaction will have no impact on shares in a putative dark fibre market; 380 (ii) if a combined market is considered for both dark and lit fibre      as a result of the proposed transaction is small and there are many other market players 381 including Openserve which is significantly larger than Maziv. 382 376 Hodge EWB p 51  53 paras 33  37. 377 Nunes FWB p 142  147 paras 5.3  5.18; Smith EWB p 234  238 paras 104  125. 378 Theron EWB p 389  393 paras 346  364; Exhibit BQ Theron Slide 21; Transcript p 3478 and 3479. 379 Transcript p 3478 lines 8  10. 380 Theron EWB p 393 para 363. 381 Theron EWB p 393 and 394 paras 365  368; para 318 and Table 7; para 319 and Table 8, p 383 and 384. 382 Transcript p 59 line 19 to p 60 line 13. Non-Confidential 97 355. We note that DFA provides both dark and lit fibre whereas Openserve only provides lit fibre services. Therefore, market definition is important to the concentration (i.e., market share) assessment and the analysis of effects that will follow. Our assessment 356. The evidence of Mr Uys in relation to the network levels suggests that dark and lit fibre are in separate relevant markets, illustrated in the following diagram: 383 357. The above diagram illustrates how dark and lit fibre operate at different levels of the value chain. Dark fibre is an input into lit fibre since it comprises the underlying infrastructure (i.e. passive fibre optic cables) that transmits data once activated. Lit fibre is what is used by customers to provide the bandwidth capacity for applications (including internet, email, file sharing, web hosting, data backup, video, VOIP and VPN). 384 In industry terminology, dark fibre is provided 383 See Exhibit Q read with Transcript p 1090 Iine 5 to p 1093 Iine 6. 384 Inter alia Nunes FWB p 147 para 5.16.1. Non-Confidential 98 as a layer 1 network product 385 whilst lit fibre is provided as a layer 2 managed service. 358. Mr Van der Merwe of Frogfoot provides a useful analogy illustrating the differences between dark and lit fibre in the context of the telecommunications value chain: 386          the bottom of that value stack we really have layer one physical infrastructure or what we call dark fibre and if I can use an analogy, if we were in the business as a fibre operator of building houses or business parks or so forth, dark fibre would be the bricks that we build the houses with. So, we go and acquire these bits and pieces of dark fibre and we then use it to create a finished product and then if we then go up the value stack we see the layer 2 service providers, so that would be the analogy of that property developer that goes and builds a house and then if we go further down the value stack to layer 3 operators, that would almost be the WeWork or the Workshop17 that goes and leases a premises, furnishes it, provides all sorts of services on top of that 359. In the case of dark fibre, each customer has a dedicated end-to-end line. However, a lit or managed network involves a connection that is shared across        -to-point dedicated solution. 360. Dark fibre (as a layer 1 product)        equipment to create a component of a layer 2 lit service offering. This does not mean that a provider of dark fibre can readily become a supplier of lit fibre. Mr Uys explains that               required substantial capital expenditure. 387 361. In terms of the dark fibre products offered by DFA, Mr Uys indicates that DFA developed the Titan dark fibre product specifically for MNOs and developed the 385            -users with connectivity: layer 1 involves deployment; layer 2 involves lighting of fibre optic cables and managing of lit cables; and layer 3 involves the provision of internet connectivity to end-users via lit fibre. Nunes provides a description of each of these layers at FWB p 133  134 paras 2.11.1  2.11.3. 386 Transcript p 88 Iine 5 to p 89 line 3. 387 Transcript p 1090 Iine 13 to p 1091 line 2; p 1092 line 2 to p 1093 line 6. Non-Confidential 99 Peregrine dark fibre product, which was the initial dark fibre product, for other FNOs. Variants of the dark fibre products, namely Tachyon and Helios were added to target enterprise customers and large ISPs. It is only when DFA sought to target smaller customers that it introduced lit services. 388 362. Notably, lit fibre cannot readily           require providers to disconnect ancillary equipment and switch off the lit fibre so that it could be leased as an asset to other providers 389 and, even then, it would still be necessary to source a separate dedicated end-to-end (dark) fibre line. 390 There is no evidence of this been done in South Africa. 391 363.        strong preference to use dark fibre only i.e., to lease dark fibre from DFA, as opposed to lit fibre services. 392 Mr Van der Merwe of Frogfoot also regards dark and lit fibre to be in separate relevant               the merging parties make clear that they see no distinction between dark fibre and lit fibre and so that lit fibre could very well become a substitute for dark fibre, which means all the cri        well disappear or become more and more unattractive and I can give a lot of examples of how that has actually affected us in the past 393 364. In addition to the above, certain demand-side factors are indicative of separate relevant product markets for dark and lit fibre: (i) they are generally procured for different intended uses by different types of customers; (ii) customer switching between dark and lit fibre cannot occur quickly and/or without significant investment; and (iii) material differences exist in the commercial terms on which dark and lit fibre are procured. We address each of these factors in turn. 388 Transcript p 1094 lines 16  22. 389 Smith EWB p 237 para 119; Hodge EWB p 53 para 36.2. 390 Transcript p 1090 Iine 5 to p 1093 Iine 6. 391 The European Commission, in paragraphs 97 to 102 of its decision in Telefonica/ Liberty Global JV Case No. M. 9871, considers the significant differences between   dark fibre)    (lit) but it does not conclude on the issue of separate relevant markets. Available at: https://ec.europa.eu/competition/mergers/cases1/202115/m9871_543_3.pdf (accessed on 25 March 2025). 392 Schoeman Transcript p 941 line 19 to p 942 line 2. 393 Transcript p 107 Iines 9  15. Non-Confidential 100 Different uses and customers 365. Different customers groups prefer and use either dark or lit products/services due to the nature of and conditions of supply. Dark fibre is typically used by businesses that wish to establish their own connectivity either for own use or for on-supply to third parties. The business can lease or purchase the fibre network and use its own equipment to light the fibre network. 394 The business will assume responsibility for deploying, operating, managing, securing and maintaining the equipment. Since the entire fibre network is leased by, and dedicated to, a particular business, the business benefits from having control of the network as well as flexibility in determining its capacity/bandwidth. 395 366. Due to the scalability and flexibility of dark fibre as articulated by Mr. Uys, 396 infrastructure provider customers have a strong preference for dark fibre to supplement their own self-build in order to offer a competitive lit managed fibre service. 397 367. Lit fibre, on the other hand, is generally operated by an FNO that sells connectivity to an ISP or sometimes directly to enterprises in the case of FTTB. The ISP then supplies it as a retail service to end-users such as businesses and homes. 398 368. The ISP, as a lit fibre customer, is not responsible for the deployment, operation, management, security or maintenance of the fibre infrastructure and outsources these responsibilities to a third party (i.e. the FNO) that has the requisite equipment and expertise. 399 394 Transcript p 736 Iine 15 to p 737 Iine 2. 395 Nunes FWB p 144 para 5.7. Also see Transcript p 1073 Iine 18 to p 1074 Iine 1. 396 Transcript p 1093 lines 13  20. 397 Van der Merwe FWB p 34 para 16.4. 398 Nunes FWB p 145 para 5.11. 399 Nunes FWB p 145 para 5.12.1. Non-Confidential 101 369. Lit fibre also does not offer the same control or flexibility as dark fibre in terms of the capacity/bandwidth of the fibre network. 400 370. Mr Van der Merwe indicates that a further problem with procuring lit services is that one has no control over redundancy / routing of network which means it becomes difficult or even impossible to engineer high levels of resilience. 401 Customer switching and costs 371. In the case of leasing or purchasing dark fibre, the upfront costs involved are substantially higher than those associated with procuring lit fibre. 402 Over and above those upfront costs, the customer is required to cover the costs inter alia of managing and maintaining the infrastructure and is required to purchase the ancillary equipment necessary to light the dark fibre. 403 372. It takes significantly longer for a customer of dark fibre to establish and use connectivity via dark fibre deployment involving extensive configuration and set up processes that contribute to longer lead times 404 - in contrast to the immediate connectivity offered by lit fibre. 405 373. Dark fibre constitutes essential infrastructure for FNOs and MNOs that could not be replicated without significant investment. Mr Van der Merwe testifies: if that                and that would break the business model. 406 Mr Uys confirms all the MNOs               it will harm Vodacom if we [Maziv] just terminate dark fibre. It will harm the mobile operators. 407 400 Nunes FWB p 145 para 5.13; Transcript p 626 Iines 15 - 19. 401 Van der Merwe FWB p 53  54 para 62. 402 Nunes FWB p 145 para 5.12.2. 403 Nunes FWB p 144 para 5.8; Transcript p 1092 Iines 6  7. 404 Nunes FWB p 144 para 5.9. 405 Nunes FWB p 145 para 5.12.1. 406 Transcript p 94 Iine 5 to p 95 Iine 22. This is consistent with the evidence of Mr Nunes, FWB p 162 para 6.39; p 163 para 6.41. 407 Transcript p 1145 Iines 9 14. This is consistent with the evidence of Mr Nunes, FWB p 157 and 158 para 6.19 and 6.20. Non-Confidential 102 374. Prof Theron that        substitutable for ISPs compared to MNOs   ISPs may prefer dark fibre to capitalise on economies of scale, while smaller ISPs may favour lit fibre, if they want to avoid risk of having their own active equipment and large, dedicated core network capacity 408 finds little support in the factual evidence. 375. Customers of lit fibre in our view are unlikely to view dark fibre as a viable alternative to lit fibre because of the significant costs and lead times involved in lighting dark fibre (including the costs associated with securing the necessary infrastructure which is itself a significant barrier to switching). Conversely, customers that purchase dark fibre do so with the intent of lighting it to provide a managed, value-added lit service downstream. This means that it is unlikely that such customers would view lit fibre as a substitute for dark fibre. 409 Different commercial terms of procurement 376. The evidence suggests significant differences in the manner in which dark and lit fibre are priced and contracted: 376.1. Dark fibre is typically contracted as an asset on the basis of a (largely) fixed monthly rental charge on a price per metre or link. 410 There are different procurement models, including lease options (typically for years) and indefeasible right of use options (usually for to years) that offer an exclusive and unrestricted right to the dark fibre infrastructure. Customers may light or operate these assets at various levels of utilisation and capability. 411 376.2. Lit fibre, on the other hand, is typically contracted as a service and is charged on the basis of port speed (e.g. per Mb of upload/download 408 Exhibit BQ Theron Slide 21. 409 Smith EWB p 236 para 112. 410 Nunes FWB p 146 para 5.15.1. 411 Nunes FWB p 146 para 5.15.1. Non-Confidential 103 speed), capacity or usage (e.g. in the case of prepaid options). Where an end-user requires changes to capacity/bandwidth, it will generally need to contact the ISP to alter the terms of the connectivity package. 412 376.3. Furthermore, dark fibre pricing is based on distance (i.e. price per metre) whereas lit fibre pricing is based on speed (i.e. price per Mb). 413 Conclusion 377. Given the above, we conclude that there are separate relevant product markets for dark and lit fibre respectively and that these products/services do not place significant competitive constraints on each other. This is relevant to the wholesale markets for provision of wholesale metro fibre and last mile FTTB and FTTH. Concentration and dependence 378. The above distinction between dark and lit fibre is important since the extent of  /market power in dark fibre has a bearing on whether the proposed merger would give rise to a substantial prevention or lessening of competition. 414 379. DFA is the dominant provider of dark fibre in South Africa to MNOs (FTTS) and to FNOs (metro backhaul) with a national market share of well above %. By revenue share, its national market share of the dark fibre market is estimated at [80-90]%. Furthermore, adding to DFAmarket power, it in many cities provides the only open access dark fibre network, or the only network of acceptable density. 415 412 Nunes FWB p 146 para 5.15.2. 413 Hodge EWB p 52 para 35. 414 If dark and lit fibre were to be regarded as being in the same relevant market, which is not supported                        415 Exhibit C Slide 12. Non-Confidential 104 380. The above is consistent with the evidence of Mr Nunes of MTN who indicates:          the metros and secondary towns covered today from a fibre perspective. A lot of the operators have climbed onto them in terms of leasing fibre to build theirs. 416 381. Vodacom states  DFA dominates the passive metro and long-haul network with c% and c.% of the market share. 417 Vodacom further states that the metro market is expected to show strong growth driven by the deployment of 5G and expansion of the access market. 418 Vodacom also indicates  DFA have the metro as it sells dark fibre as a standard product and is well positioned to future MNO backhaul opportunity associated with densification 419 Also that DFA gets up to c.80% of wallet share. 420 382.              market. She argues, however, as already indicated, that the relevant market could include both dark and lit fibre, in which case DFA would have a much lower market share. 421 We have dealt with this issue and indicated why we disagree. 383. The second-largest player of dark fibre for metro fibre connectivity in South Africa, Liquid Telecoms, is significantly smaller than DFA with an estimated national market share of [0-10]%; and the third-largest player is Link Africa, with an estimated [0-10]% national market share. 422 This illustrates the dominance of DFA. 384. Of importance to the vertical competitive assessment that will follow, is that customers are heavily reliant on DFA for dark fibre. C    a 416 Transcript p 757 lines 13  17. 417 Bundle M p 3592. 418 Bundle M p 3592. 419 Bundle M p 10164. 420 Bundle M p 10164. 421 Exhibit BQ Theron Slides 19 and 22. Openserve, DFA, MTN, Liquid Telecom and MFN provide lit services to the downstream market on a wholesale basis to ISPs. 422 Smith EWB p 254. Non-Confidential 105 critical player in the industry and an unavoidable supplier of dark fibre as a key input. 423 385. As indicated above, mobile operators mainly lease dark fibre from DFA, or may build their own dark fibre, to connect to their base stations. 424   business was premised on the connectivity of mobile operators to their base stations. 425 Mr Nunes confirms         dark fibre assets so that the FNO market could be built, and the MNO market could be built. 426 386. FNOs rely heavily on dark fibre to build their networks. Mr Van der Merwe testifies that approximately        427                 go and buy dark fibre from anyone else                      this network and also there may not be another provider that even has a feasible dark fibre offering that I can buy from them 428 (Own emphasis) These product rules      387. Mr Schoeman explains why Rain is highly dependent on DFA for the provision of dark fibre services: (i) Rain is % dependent on DFA for the provision of fibre backhaul services; 429 and (ii) because DFA is so dominant in the South African              would be exceedingly difficult for Rain to uncouple itself from DFA by finding an alternative service provider or by self-      very difficult for us [to] unplumb ourselves from DFA 430 423 Schoeman Transcript p 1034 lines 4  5, Van der Merwe p 77 lines 1  6; p 88 line 6 to p 89 line 3; p 89 line 22 to p 90 line 10; p 95 lines 13  20; p 108 lines 2  3. 424 Van der Merwe Transcript p 84 lines 6  16; p 107 line 18 to p 108 line 3. 425 Van der Merwe Transcript p 94 line 13 to p 95 line 5; p 232 lines 15  18. 426 Transcript p 909 lines 10  20. 427 Transcript p 108 lines 3  6; p 153 line 18 to p 154 line 2. 428 Transcript p 114 lines 8  14. 429 Rain Founding Affidavit para 38, intervention bundle p 20. 430 Schoeman Transcript p 956 line 9. Non-Confidential 106 388. Mr Nunes confirms that Openserve only provides lit services and indicates that although MTN has some reliance on Openserve, it is marginal. MTN only utilises Openserve where there is no other fibre infrastructure or if it is in sparse rural area. 431 389. Furthermore, overbuilding is uneconomical and rare, particularly in respect of FTTS. Mr Van der Merwe explains DFA volunteered to say on an offtake basis we will go and interconnect all these base stations, they signed very large agreements with DFA that enabled them to finance and deploy this national MetroFibre network that interconnected all the base stations. Once that network existed, the business case disappeared for anyone else to go and build MetroFibre, because all the base stations are already connected 432 FWA and FTTH 390. A major issue of contention during the proceedings was whether, or the extent to which, retail FTTH services compete with FWA services. 391. From a market delineation perspective, Mr Hodge for the Commission submits that at the retail level there is a market for the retail provision of home and SMME broadband services with differentiated products where fibre (FTTH/FTTB) competes with FWA. 433 392. MTN submits that FWA and FTTH are neither demand- nor supply-side substitute technologies or, at most, there is immaterial competitive interaction between FWA and FTTx products and services. It submits that more immediate and intense competition exists within (i.e., fibre competes with fibre) and not between these technologies. Recall however that there is limited overbuild of FTTH (see paragraphs 265 to 272 above). 431 Transcript p 629 lines 1  8. 432 Van der Merwe Transcript p 94 line 13 to p 95 line 8. 433 Hodge EWB p 71 para 74. Non-Confidential 107 393. Mr Reynolds and Prof Theron for the merger parties contend that FWA does not compete in the same market as fibre. Mr Reynolds specifically argues that FWA is inferior to FTTH. For this assertion these economic experts, as well as Mr Smith, rely mainly on arguments around supply-side substitution. 394. The merger parties advance the following reasons for why FWA and FTTH are complementary rather than substitutable: 394.1. they allege that the witnesses confirm that FWA is not an alternative to fibre; 394.2. FWA and FTTH have significant differences in capacity, quality and price; 394.3. FWA and FTTH cater for different use cases; 394.4.            customers for whom FTTH would be a viable substitute for their needs; 394.5. there is no market evidence of churn and pricing pressure between FWA and FTTH; 394.6. the Aetha Report confirms that in countries with greater 5G take-up, FWA services are generally not considered by regulators to be close substitutes for FTTH; and 394.7. even if there was a level of substitutability between FWA and fibre, (current) low FWA market share and its limited ability to offer FWA due to the demands of its mobile customers, show that the merger would not have any material horizontal effect on competition 434 395. The Commission disputes all the above factors relied on by the merger parties.           documents, which are not prepared for contested proceedings, reveal the following: 395.1. from a demand perspective, FWA and FTTH are both home broadband products and target the same demand; 435 434 In a market for the retail provision of home and SMME broadband services, i.e., a combined fibre/FWA market. 435 See Transcript p 3491 lines 11  18. Non-Confidential 108 395.2. CIVH in its Board Strategy Session dated 2 February 2022, explicitly cites and as ; 436 395.3. Hardiman Telecommunications Hardiman   consultant) concurs in a report prepared for Vodacom that by competitors in neighborhoods may impact and          two threats being the neighbourhoods, but it also articulates how this could likely impact on l, namely through a potential impact on and . In essence, if a share of household demand, then this will reduce by FTTH in those areas, making those areas or A reduction in ARPUs may occur either through , and improving or pushing more FTTH connections. Importantly, where there is no fibre overbuild in an area, which FTTH operators seek to avoid, then LTE/5G FWA is the only alternative to the FTTH provider for consumers and hence the only remaining direct constraint; 437 395.4. FWA and fibre both deliver the speeds that are consumed by most consumers, i.e., 100Mbps and lower; 438 395.5. FWA and fibre are both marketed by operators as substitutes to one another, for example, MTN and Vodacom have considered FWA as a fibre-like service; 439 395.6. their pricing and packaging have evolved to align with one another; 440 395.7. there is a shift by both to towards playing increasingly in the middle to low- income segments; 441 395.8. FWA access is growing in coverage and subscription numbers, and Vodacom markets speeds up to 100Mbps on its 4G and 5G FWA. The Commission refers to Dr Van den Bergh conceding that the coverage 436 Exhibit BA p 67; Hodge FWB p 66 Figure 16. 437 Bundle M p 3113. EWB p 67 para 64 and Figure 17. 438 Transcript p 3491 lines 11  18. 439 Transcript p 3497 line 22 to p 3499 line 9. 440      441 Bundle M p 12949. Non-Confidential 109 maps, showing that a large portion of the country is covered with            deliver up to 100Mbps of FWA broadband; 442 and 395.9. market analysts, Africa Analysis and BMIT, include FWA and FTTH together in analysing home broadband services. 443 396. We note that it is common cause that FWA and FTTH are both home broadband products, the dispute is around the extent of any competitive interaction. Our assessment 397. In Corruseal, 444 the Tribunal cautioned that the boundaries of market definition do not determine the outcome of an analysis of the competitive effects of a merger in a mechanistic way. The purpose of market definition is to provide a framework for an analysis of the competitive effects of the merger. 445 The Tribunal further explained that in practice, the analysis underpinning the identification of the market(s) and the assessment of the competitive effects of a merger may overlap, with many of the factors affecting market definition also being relevant to the assessment of competitive effects, and vice versa. Therefore, market definition and the assessment of competitive effects should not be viewed as distinct analyses. 446 398. In this matter the evidence that we rely on in relation to market delineation should be read alongside our competitive assessment. 399. Those that seek to contend that there is a lack of competitive interaction between FWA and FTTH would typically argue that FTTH is a superior product in terms of speed and is uncapped, which means that consumers/households prefer 442 Transcript p 2279 lines 1  4. 443 Transcript p 3495 lines 4  15; Bundle M p 12756  12824. 444 Corruseal Group (Pty) Ltd and another v Competition Commission of South Africa and others        Corrusealat paras 36 - 38 and 47 with reference to the UK CMA Merger Assessment Guidelines (2021). 445 Corruseal para 36. 446 Corruseal para 38. Non-Confidential 110 FTTH absolutely where it is available. These proponents argue that FWA is complementary, not substitutable, offering home broadband services where FTTH is yet to roll out and once FTTH is available then households will switch to it. We shall test whether this applies to South African consumers, noting the relative low levels of uptake or penetration in South Africa where FTTH has been rolled out. 400. The factual witnesses of the merger parties and MTN, and their respective economic experts, largely take a technology-based or supply-side perspective to market delineation when arguing that FWA and fibre do not compete. The panel asked what evidence is available from a demand-side perspective. As Mr Masalesa points out, which we concur with, customers consume data not technologies                  . 447 As we shall explain below, once fibre has been rolled out in a specific area, the marginal consumer is relevant in the assessment of the competitive interaction between FTTH and FWA. 401. Mr Van der Merwe of Frogfoot explains how FWA and fibre compete. He testifies that FNOs target a terminal take-up (also referred to as penetration rate) of 50% (also see penetration discussion above in paragraphs 248 to 264) and that once they achieve the 50% terminal take-up, they compete with LTE and 5G FWA to        So, predominantly those mobile services are either packaged as fixed wireless access products, so LTE or 5G service              to. [...]. They do compete with the fibre operators and what is important to highlight there is really how they compete. [...]. But what is important to understand is that once the fibre operator reach terminal penetration, so we                 get to around 50% of the subscribers signing up onto our network, that would be a good business case. So, once we get to the 50% then we have to ask ourselves the question of why is the other 50% of subscribers not signing up 447 Transcript p 293 lines 11  22; p 460 line 15 to p 61 line 1. Non-Confidential 111 with the fibre networks         subscribers are typically subscribing to entry level fixed wireless or mobile wireless services. [...]. so once the fibre operator really gets terminal penetration, it will then start creating packages that directly compete with fixed wireless access services 448 (Own emphasis) 402. Dr Scheffer concedes that for FNOs to increase their fibre/FTTH penetration rates over time, they compete with FWA, for (in economic terms) the marginal customer, as is evident from the following exchange with counsel: ADV BERGER SC: And what the FNOs have to do, in order to win the trust of people above the 35% and to get them up to the 90%, is to convince them that fibre is better for them than, for example, fixed wireless or other means of connecting to the internet. And so you have to compete and  against the fixed wireless operators to attract those customers that, initially, are not being attracted to fibre. Not so? DR SCHEFFER: Yes, customers will connect to the appropriate solution out of their free choice and, typically, ISPs will offer beneficial products and so forth that allow the customer to use that advantage. ADV BERGER SC: Yes, and so once you get  it's easy for you to get to the 35% and then as  in order to increase it over 35%, connection to 45% and 55 and all the way up to 90, you've got to compete harder and harder to get those customers to come from whatever other means they've connected to the internet, to come onto fibre. Would you agree with that? DR SCHEFFER: Yes, the  one has to keep on providing products that customers would prefer and that would become more and more competitive products as one progresses during this time. 449 403. Based on the evidence before us, for purposes of our assessment FWA includes (i) LTE/4G FWA; and (ii) 5G FWA. This is because, from a consumer perspective, the two are considered as home broadband technologies, and 448 Transcript p 84 line 6 to p 85 line 22. 449 Transcript p 2455 lines 3  22. Non-Confidential 112 MNOs offer FWA services using both these technologies.    So, the 5G market will be driven strongly by migration of customers. Ja, I mean to be honest               still have an FWA product and the customer receives a broadband product. They           . 450 (Own emphasis) 404. Mr Motlekar of Telkom CSB submits that in experience FTTH and FWA are substitutable products for home-connectivity depending on the primary use requirement. He says that market participants continue to develop and introduce innovative fixed wireless and mobile products into the market which provide alternative last mile options for consumers. The degree to which these newer technologies will compete with FTTH and FWA will only be known in the future. 451 405. Similarly, Mr Masalesa       FWA and FTTH are substitutable technologies for the majority of consumers in terms of a home broadband solution. From a customer perspective, he indicates that customers consider a variety of factors in making their purchasing decisions. He states that g            their individual connectivity needs, affordability and the pricing of available products both within, as well as across, different technologies. 452 406. Rain submits to the Commission that competition exists between 5G and FTTH services as both services offer internet connectivity at similar speeds and price to the end consumer. It adds that 5G provides the customer with flexibility to move with their router to different locations where there is coverage, whereas FTTH does not provide such flexibility. It further submits that wireless access technologies such as LTE, LTE-Advance or 5G can be used effectively as an alternative to FTTH. This it says is evidenced in South Africa by the large percentage of customers using wireless access technologies as their primary 450 Transcript p 2036 lines 9  13. 451 Motlekar FWB p 9  10 paras 7  11. 452 Masalesa FWB p 21 and 22 paras 6  9. Non-Confidential 113 broadband connection. In regard to price, it submits that in the South African market wireless services typically offer similar value to FTTH services. 453 407. Mr Otty agrees that in relation to the lower income segment of South Africa substitutability between FWA and FTTH offerings makes sense in the short run       454 408. Dr Scheffer concedes that 5G FWA and fibre both offer customers access to broadband services and that Vodacom in August 2021 was of the view that its 5G FWA deployment must be aligned with its fibre offering in that it would not provide the same rollout in the same area. 455 409. As indicated above, the Tribunal asked questions around the consideration of the demand side, i.e., evidence from the perspective of consumers. In our view a vital portion of the demand-side evidence is the take-up or penetration levels that typically are achieved by FNOs in South Africa where they have rolled out FTTH. The Tribunal wanted to know if fibre is as technologically superior as the merger parties argue, why the average penetration levels are relatively low when fibre has been made available in an area, especially in the Core segment areas where high LSM households live. We have dealt with the relatively low average penetration levels actually achieved in the factual evidence in paragraphs 248 to 267 above. This suggests to us that South African consumers have alternatives other than FTTH. The merger parties and their experts could not satisfactorily respond to this. 410. Mr Nunes when asked by the Tribunal what the large portion of customers use that choose not to connect to FTTH where it has been made available (given the relatively low penetration rates) indicates:         including old legacy technologies and fixed wireless access. 456 453 Bundle M p 8575. 454 Transcript p 2032 line 21 to p 2033 line 12. 455 Transcript p 2462 lines 3  22. 456 Transcript p 914 line 7 to p 915 line 16. Non-Confidential 114 411. Responding to the same issue about what the large portion of customers use that do not use FTTH where it is available, Prof Theron  Mr Hodge also said I mean what is the aim of the market definition and I think listening to the debate and the question if the fibre penetration even in higher income areas is only 50%, then what do the rest use          can speculate they use FWA. 457 (Own emphasis) One can from the factual testimony infer that many use FWA. 412. In relation to the competitive dynamics in township areas, which is of particular relevance in this case, Mr Van der Merwe testifies that when fibre enters those areas, MNOs will have to respond to keep customers through lower prices or less restrictive packages. 458 He describes the competitive interaction between fibre operators and mobile operators as follows once fibre enters these areas:  the prime candidates to lose market share there is the mobile operators and so we can expect that once the fibre operators enter those markets, they will start reducing pricing and actively compete with the fibre operators in those markets. 459 413. Mr Motlekar confirms that is one of the factors taken into account by Telkom in its pricing of 460 and that in relation to the lower LSM areas Telkom is seeing switching to fibre and back to what the consumer is using, some other fixed wireless access productby customers in low-income areas. 461 414. A further important issue when considering the dynamics of the market, is the fact that there exists in many local geographic areas in South Africa local monopolies in the provision of FTTH. Recall the evidence regarding the reluctance of FNOs to overbuild and the low percentage of actual overbuild in South Africa (see paragraphs 265 to 272 above). 457 Transcript p 3543 lines 16  20. 458 Transcript p 86 line 13 to p 87 line 7. 459 Transcript p 86 lines 17  21. 460 Transcript p 491 line 17 to p 492 line 10. 461 Transcript p 506 lines 2  19. Non-Confidential 115 415. In all of these local areas that are not overbuilt there is no direct fibre-to-fibre competition and FWA is an alternative. Mr Hodge  I think 80% of                                increases to get to build in overbuild, especially where Telkom never had infrastructure before, which is the majority of the country. So, this idea that they are distinct, I think the evidence does suggest quite the contrary. 462 We concur that this is a strong indication that FTTH and FWA are not in distinct product markets where there is no overbuild (also see Dr Scheffers evidence in paragraph 402 above). 416. We accept that when there is overbuild of FTTH in a specific area (which is overall limited), one or more other FTTH providers would provide competition and that fibre-to-fibre competition would be more homogenous. Mr Hodge  I think in this case and given the framework of how these markets operate, you know, you've got some overbuild where you might think about that, FTTH being closer, but you've got a lot of other        the case 463 (Own emphasis) 417. One has to consider what would constrain the local FTTH monopolies in the many geographic areas where there is no overbuild. In that context it becomes imperative to consider the potential (future) competitive interaction with other technologies, i.e., FWA. 418. All of the above is persuasive to conclude on the market delineation issue, and we            regarding the competitive interaction between FWA and FTTH under the effects analysis. 462 Hodge Transcript p 3299 line 19 to p 3300 line 6. 463 Transcript p 3644 lines 7  17. Non-Confidential 116 Conclusion 419. As we have explained above, the purpose of market definition is to provide a framework for an analysis of the competitive effects of the merger and market definition and the assessment of competitive effects should not be viewed as distinct analyses. We noted that the evidence that we rely on in relation to market delineation should be read alongside our effects assessment. 420. We conclude that, for market delineation purposes, there is sufficient evidence to consider a broader market for the provision of home broadband services including FWA and FTTH. Geographic scope of the markets 421. We focus on metropolitan and last mile fibre given their relevance to the merger. 422. Mr Hodge for the Commission submits that the geographic scope of the metropolitan and last mile markets has three dimensions: (i) local; (ii) regional; and (iii) national, and that each dimension is important for competition. 464 He submits: 422.1. A last mile infrastructure provider, whether FTTH, FTTB or mobile, is ultimately interested in securing backhaul from a specific location. The options available to them depend on local competition for metro fibre and the existence of alternatives within a short distance. Options that are not as close pose less of a competitive constraint as the greater distance makes them relatively more expensive. Metropolitan fibre providers compete to offer coverage in an area, including aggregation nodes, to win customers located in those areas. It is also cheaper for a metro fibre provider to extend its coverage once it has a local presence, connecting adjacent areas at incremental cost. 464 Hodge EWB p 54 para 38. Non-Confidential 117 422.2. A metro fibre provider will benefit from a greater regional presence too, as it enables them to offer build and support services with greater scale and cost efficiencies. This includes across enlarged metro areas, but also in cities and towns across a region. Once more, it is easier to extend their network build and service in regions where they already have a presence, and they benefit from the economics of density. 422.3. Lastly, there is a national dimension to competition insofar as most infrastructure providers operating across regions tend to have a single rate card that applies nationally. However, the pricing will reflect the strength              is active. The ability to raise capital and enter new areas (competition for new markets) is also likely to depend on the national position of a metro fibre provider, as this will determine the extent of EBITDA margin generated from existing business which can be channelled or leveraged to secure funding into new build expansion. 423. Prof Theron does not conduct a geographic market delineation exercise but criticises Mr Hodge for considering local markets; 465 she assesses the effects of the proposed merger at a national level. In oral evidence, Prof Theron does not            the relevant geographic markets. 466 424. In relation to FTTH, Mr Reynolds agrees that the market is highly localised. 467 425. In our view the geographic dimension of the markets, both metro and last mile, is important in understanding the dimensions of competition in that it relates to scale economies and efficiencies. A greater regional presence enables operators to offer build and support services, and also to extend their networks to cover wider areas, benefitting from economics of density. 468 465 Theron EWB p 375 to 380. 466 Transcript p 4171 line 18 to p 4172 line 13. 467 Transcript p 4086 lines 2  5. 468 Hodge EWB p 54 para 38.2. Non-Confidential 118 426. FTTH and FTTB providers compete for local markets and benefit from economics of density if they expand locally. In addition, national players who are present in local markets tend also to compete on localised pricing and discounting to improve churn rate and take-up within specific local geographic markets. 427. Mr Otty in response to a question from the Tribunal regarding how he sees the geographic market for the fibre operators indicates  fibre is naturally covering a geographic area. So, you go into a geographic area and you then become hopefully the first entrant into that geographic area, which means it makes it a less attractive proposition for the next person to come along and that makes it a completely unattractive proposition for the third person to come along. 469 (Own emphasis) 428. The impact of the proposed merger on the local, regional, as well as national markets is relevant. HORIZONTAL EFFECTS 429. On horizontal effects the Commission argues that there will be a loss of future competitive rivalry if the proposed merger is approved. The merger parties argue that there will be no significant accretion and no meaningful loss of rivalry as a result of the proposed merger. 430. We consider the horizontal effects in relation to respectively (i) a lessening of competition in FTTH; and (ii) the removal of a competitor, Vodacom, in metro fibre and FTTB. HORIZONTAL ASSESSMENT: FTTH 431. The Commission argues that FWA and FTTH are in the same relevant product market, that they are home broadband substitutes and that consumers stand to 469 Transcript p 2076 line 9 to p 2077 line 2. Non-Confidential 119 benefit from increasing competitive rivalry between FWA and FTTH that will lower prices. It argues that absent the proposed transaction, Vodacom will compete more aggressively with FWA and that will benefit South African consumers. 432. The merger parties argue that FWA and FTTH are not in the same relevant product market and that FWA and FTTH are not sufficiently close substitutes to place a material competitive constraint on each other. They argue that FWA complements FTTH and is only useful where fibre is not available and is a             other words, FWA competes with FWA, and fibre competes with fibre. 433. The merger parties qualify their assertions regarding the lack of competitive interaction between FWA and FTTH by arguing that to the extent that there is competitive interaction between these technologies, this is not a primary constraint as there are other constraints on each of Maziv and Vodacom. They argue that the c         services of other MNOs and that Vumatel is constrained by rival FNOs, particularly Openserve, which has overbuilt Maziv the most. MTN makes similar arguments. 434. They further submit that Vodacom requires its spectrum to service its mobile customers. More of these mobile customers would be in the densely populated lower-income areas where fibre is to be rolled out. In these areas, Vodacom would not have the spectrum capacity to provide a sustainable FWA service for large numbers of FWA customers (in addition to its mobile services) as its radio network could not carry the load. Our assessment 435. We first deal with the horizontal overlap between Vodacom and Maziv, through Vumatel, in relation to wholesale FTTH. We consider current as well as future dynamic competition. Non-Confidential 120 436. We then deal with the extent of competitive interaction between FWA and FTTH in the context of the market characteristics as previously discussed. FTTH overlap in merger parties’ current and future activities 437. In terms of homes passed at a national level, Vumatel is the largest wholesale FTTH provider in South Africa. It has an estimated national market share of approximately [30-40]% with approximately homes passed. 470       the proposed transaction   Acquisition of 40% equity in the largest enterprise & FTTH FNO in SA 471 438. The next player in FTTH in terms of size is Openserve with an estimated national market share of approximately [20-30]% with million homes passed. We note that           merger parties. Furthermore, in terms of future competition, Vumatel is seen by the merger parties as able to achieve more sustainable investment and growth relative to Openserve. 472 439. We have above dealt with Herotel and indicted that (i) CIVH already has a % shareholding in Herotel; (ii) in terms of a proposed merger notified to the Commission it intends to increase this     CIVH retains % economies of secondary cities  212.1 above). 440. The implications of Herotel for the concentration analysis in this matter are that at the very least Vumatel has a       network and furthermore Maziv does not intend to challenge it in the secondary             have a market share of approximately % in South Africa in homes passed with 470 Hodge EWB p 84 para 104. 471 Part A of the Record p 3804 Project      2021, Shameel Joosub. 472 Hodge EWB p 82 para 101; Bundle M p 3544. Non-Confidential 121 FTTH. (Vumatel has a market share of approximately [30-40]% and its % share in [0-10]% share, translates to approximately [0-10]%, together      441. In terms of homes connected with FTTH, the merger parties have a national market share of approximately [30-40]%. (Vumatel has an approximately [30- 40]% share, it has a %   approximately [0-10]% market share, and Vodacom has an approximately [0-10]% share). 473 442. As we have found under the geographic market, competition in FTTH is highly localised, although there is a national dimension to competition as national pricing can be observed. Mr Reynolds confirms under cross examination that the FTTH market is highly localised: 474 ADV BERGER SC: Ja, but my  again, my question is simply that you recognise local markets. In fact, you say they are highly localised? MR REYNOLDS: Yes, that's right, ja 443. As we have found in the market characteristics section, there is competition for the market and FTTH providers tend not to overbuild each other, i.e., they compete in the land grab and want to obtain the first mover advantages in new rollout areas. It is common cause that the land grab is ongoing, with the high- income areas of South Africa largely covered, and that the FTTH providers are now focusing on secondary towns and cities and low-income areas. 444.             overlap in the provision of FTTH since they have overbuilt each other. 445. The Commission argues that in these local areas the removal of Vodacom as a competitor will result in a loss of choice and loss of competition in both price and non-price factors such as marketing and service. 473 Hodge EWB p 84 para 104. 474 Transcript p 4086 lines 2  5. Non-Confidential 122 446. As of October 2023, Vodacom and Vumatel have overbuilt each other with respect to an estimated homes passed. 475 In respect of approximately of these homes the proposed merger results in a merger to monopoly. For the rest of the affected homes the proposed merger would represent a three-to- two or four-to-three merger. 447. Mr Reynolds concedes that for the areas where the proposed merger reduces the number of competitors from two to one, there will be a reduction in choice for consumers: ADV BERGER SC: Okay, if you look at the 2-1 concentration, those consumers are  have no choice anymore, until perhaps Openserve comes in. MR REYNOLDS: Yes, that's right. ADV BERGER SC: And so, in those localised markets there most certainly would be a substantial lessening of competition.            448. Furthermore, the Commission has found that in local markets that are overbuilt, there is more competition with lower prices and greater discounting to improve take-up. 476 Some FTTH operators will remain focused on the particular local markets in which they entered and will price locally. National FTTH (and FTTB) providers which have national pricing policies will still engage in localised discounting and ISP offers designed to improve take-up and churn from competitors within specific local markets as ISPs in turn pass them onto prospective customers. 477 449. Mr Hodge           there is an overlap with other FNOs. 478 He testifies that Vodacom discounts wholesale rates where they are overbuilt. 475 Reynolds EWB p 507 para 5.52 (a). 476 Commission Report, summary, para 23. 477 Hodge EWB p 54  55 para 39.1. 478 Transcript p 4088 line 11 to p 4089 line 5. Non-Confidential 123 450.            post- merger there will be a push to close that gap. 479 451. Although Dr Scheffer argues that too high fibre prices will attract new entry, he when questioned about this by the Tribunal indicates that in his past experience the price increase would have to be about 20% before competitors would be attracted to enter/overbuild. 480 This means that a SSNIP 481 can be successfully implemented by the merger parties post-merger without attracting new entry. 452.         that it, when assessing the proposed transaction in October 2021, and            proposition for the deal was that post-merger it would be able to  by a . 482 This was not seriously contested       453. Mr Reynolds sought to justify the above difference in ARPUs by referring to his            broadly similar, and therefore argued that any expectation of higher revenues must be customers that will be "trading up" by choosing greater volumes or better quality service. There is however no evidence or reason proffered as to why customers might trade up post-merger (i.e., chose greater volume or better quality services) if they could already do so pre-merger. 483   assertion is also not supported by any evidence to be found in the relevant CIVH internal documents assessing the proposed transaction. A more plausible explanation is that ARPUs (effectively prices) would rise because of a lessening of competition or the exercise of market power. 479 Transcript p 3293 line 19 to p 3294 line 4. 480 Transcript p 2555 lines 4  16. 481 A small but significant non-transitory increase in price. 482 Bundle M p 1352 and 1398. 483 Transcript p 4103 lines 1  5. Non-Confidential 124 454. Mr Reynolds also notes that there is different pricing for the Vuma Reach and Vuma Key areas 484        lower speed packages would only be found in the Reach areas, not the Core          located. 485 Customers purchasing lower speeds and lower priced products are likely to be those with lower disposable incomes. Those customers that prefer cheaper packages with lower speeds should not have to pay for products with higher prices. Thus, the impact of Vu     -called Core areas, would fall disproportionately on consumers at the lower end of the income distribution in these areas. 455. We conclude that where the proposed merger results in a merger to monopoly in respect of approximately homes passed with FTTH (and a reduction of competition in relation to homes), this will reduce the choice of those          are lower than that of Maziv and that means higher ARPUs post-merger which will be to the detriment of consumers. 456. The merger parties offer a divestiture remedy in relation to the FTTH overlap in activities 486 , but reserve the right for Maziv to apply for a waiver of the condition if it is unable to find any potential purchasers for the assets within of the implementation date. The latter in our view will remove the incentive to find a buyer or buyers for these assets. The proposed condition does not make         to ensure that the assets are sold and that the competitive harm caused by the transaction is neutralised. The proposed condition furthermore makes no provision for a 484 Transcript p 4120 lines 6  7. 485 Reynolds EWB p 516 para 5.92. 486 See clause 8.1 of the tendered remedies that reads: In all areas, including suburbs, estates, or business parks, where Vodacom SA Group has rolled out infrastructure which: (i) has overbuilt Maziv Group FTTH infrastructure as at the Implementation Date, (ii) is being transferred to Maziv in terms of the Merger; and (iii) Vodacom SA is using to provide Wholesale FTTH Services as at the Implementation Date, the Merger Parties shall apply good faith and best endeavours to divest the overlapping infrastructure of either Vodacom SA Group or Maziv Group within of the Implementation Date (or as soon as regulatory approvals required for such disposal have been granted) but, if Maziv is unable to find any potential purchasers for the assets within of the Implementation Date, Maziv may, in terms of clause 21 apply to the Tribunal for this condition to be waived. Non-Confidential 125 pricing remedy until the assets are sold. We deal with the shortcomings of this proposed condition further under remedies. 457. As we have further indicated, competition analysis is forward looking, and a non- static analysis is appropriate in this case. Context is crucial, and thus the           proposed transaction must be considered in the horizontal competition assessment. The proposed transaction will also impact competition for the market i.e., the land grab in wholesale FTTH, which leads to irreversible - takes- -like outcomes in future as we have discussed. The merger parties proposed conditions do not address this. 458. As we have indicated, it is common cause that FNOs are engaged in a second FTTH land grab, focusing on the lower-income areas following the near complete coverage of the high-income areas. Based on the strategic evidence, absent the proposed transaction, Vodacom would be looking for opportunities to build FTTH, and importantly Vodacom would be looking to partner in off balance sheet JVs to expand its whole fibre offering in South Africa. 487 The removal of Vodacom as an FTTH player means that there would be less competitors in the land grab and thus lower incentives across the market to build out more quickly. 459. The           loss of other MNO customers to Vodacom in the event that Vodacom enters the            that Vodacom would present either through o its fibre network and resulting in DFA and Vumatel and (see paragraph 133 above). 488 Rather than , CIVH decided that it must conclude the proposed transaction as a 489 487 Bundle M p 12454, 12475 and 12483. 488 Bundle M p 1223 and 1224. 489 See inter alia Bundle M p 1174. Non-Confidential 126 Interaction between FWA and FTTH 460. FWA services were initially offered in South Africa using unlicensed non-GSM spectrum with end-customers having antennas installed in their homes directed at the tower of the FWA provider. These services have been supplemented and largely displaced by FWA services provided over LTE/4G mobile networks. Rain launched 5G FWA services in South Africa in 2019. 490 461. We note that this section must be ready together with the evidence in the market definition section. 462. The two factual witnesses from Telkom CSB, both testified that FWA and FTTH are substitutable access technologies which consumers use to consume data. 491 FWA does not complement fibre but competes with fibre. 492 Rain also submitted to the Commission that FWA and FTTH compete (see paragraph 4066 above). 463. The various factual witness of the merger parties and Mr Nunes from MTN disagree. Much of the evidence lead by them focused on supply-side considerations of the market and the alleged superiority of fibre. 464. As already indicated, the Tribunal was interested in demand-side considerations, in other words, how customers that make the choice between these technologies, FTTH and FWA, see the market. 465. The factual evidence confirms that customers considering whether to opt for FTTH or FWA consider a wide range of factors such as price - specifically relevant for the lower-income households; connectivity speed; connection reliability; and product flexibility. 493 490 Reynolds EWB p 450 para 2.28. 491 Masalesa FWB p 21 para 6; Transcript p 293 line 19 to p 294 line 4, p 460 line 10 to p 467 line 20. 492 Masalesa Transcript p 403 lines 5  14. 493 For example, Masalesa FWB p 21 para 7, also see, for example, Masalesa Transcript p 294 line1 to p 299 line 16. Non-Confidential 127 466. From a customer perspective Mr Masalesa explains that a customer will walk              to prove the point of substitutability, a customer might actually walk into a store with a mindset of buying an LTE product and he gets educated about availability of fibre products and how fibre works and you know, he can walk out of a store with a fibre product instead of an LTE product. And the same could be true with a customer going into a store to buy fibre and only to be told that you know what, you can actually get decent speeds on fixed wireless access 494 467. On affordability Mr Motlekar states: ... affordability but the features of the technology and obviously linked to affordability, is the pricing of the technology of what the customer want. Ultimately, it's the same use that's being transmitted across this technology. The customer wants access to the Internet and the features of those technologies together, where the price will determine which product the customer selects 495                               496  From our experience it starts with affordability and price and then the technology that services that demand and what are the price points within the various categories and this refers to                slide and that talks to what the customer  how the customer consumes, when they want to consume it and where they want to consume it 497 (Own emphasis) 468. Mr Nunes confirms that in lower-income areas, the flexibility of pricing and pre- paid options is important when consumers make decisions about which products to choose. 498 494 Transcript p 465 line 8 to p 466 line 5. 495 Transcript p 481 lines 7  12. 496 Transcript p 557 lines 5  16. 497 Transcript p 606 lines 1  7. 498 Transcript p 923 line 8 to p 924 line 6. Non-Confidential 128 Evidence about consumer conduct in South Africa 469. The best evidence that we have from a demand-side, i.e., customer, perspective is how many South African consumers actually choose FTTH where it has been rolled out. We have provided the average FTTH penetration rates for the Core, Reach and Key market segments. This factual evidence confirms that relatively low average FTTH penetration rates have been achieved in South Africa  meaning that consumers have other options for accessing the internet other than FTTH. Those home broadband alternatives to consumers include LTE 4G and 5G. 470. Furthermore, in the market characterisation section, we have made the point that an important characteristic in FTTH is that FNOs compete against each other in the so-called land grab, i.e., there is competition for the market. What is crucial is that - unless the FTTH FNOs overbuild each other - they do not compete. Overbuilding however is something that they typically would avoid according to the factual evidence. In fact, % of Vumatel FTTH areas are not overbuilt by other FNOs. In the non-overbuild areas, FWA is available as an alternative for            competitor in all of the non-overbuild areas. Since both technologies are growing, the future will continue to see this interaction between FWA and FTTH, specifically competition for the marginal customer as explained above. 471. We have already referred to Dr Scheffer confirming that to drive up connectivity rates in relation to fibre, there is constant and ever-increasing competition to formulate competitive products to attract customers to switch to fibre where those customers use methods other than fibre, such as FWA: ADV BERGER SC: Yes, and so once you get  it's easy for you to get to the 35% and then as  in order to increase it over 35%, connection to 45% and 55 and all the way up to 90, you've got to compete harder and harder to get those customers to come from whatever other means they've connected to the internet, to come onto fibre. Would you agree with that? Non-Confidential 129 DR SCHEFFER: Yes, the  one has to keep on providing products that customers would prefer and that would become more and more competitive products as one progresses during this time. 499 472. The above makes sense from an economics perspective since an FNO has the incentive to improve its FTTH penetration rates to drive revenue returns on the capex that has been spent to roll out the fibre. To improve that, the operator will increase marketing and be incentivised to offer discounts and do promotions, as is happening in practice in South Africa. This is competition for the marginal customer. 473.           submissions in these proceedings and their internal documents, as well as their submissions to the sector regulator, ICASA, regarding the competitive interaction between FWA and FTTH. The same applies to MTN. What Vodacom told ICASA 474.               hearing about a lack of interaction between FWA and FTTH, since it is inconsistent with          ICASA during or about 2019/2020 in which Vodacom submitted: 500 The assessment should also have considered the role of fixed-mobile substitutability. ICASA should also have considered the potential effect that the expansion of fixed broadband services may have on mobile broadband prices. In particular, there is a degree of substitutability between fixed and mobile services which can be expected to grow over time: customers may offload data usage to fixed services relieving capacity constraints currently faced by Vodacom and other operators, and forcing them to compete with fixed broadband services that offer high or unlimited usage caps. Both factors may result in lower prices for mobile broadband services. 499 Transcript p 2455 lines 14  22; p 2456 lines 14  17. 500 Exhibit            Mobile Broadband Services p 13057  13058. Non-Confidential 130 ...Overall, we consider fixed services to represent an important and growing competitive constraint on mobile services. ... • With wider availability of higher quality fixed services, customers are likely to off-load an increasing amount of their mobile data traffic to Wi-Fi, both at home and in public places. • The wider adoption of fixed broadband services may also enhance the use of devices that are reliant on fixed broadband services over mobile services (for example, customers may start using video streaming services on a computer instead of a smartphone). In this context, we note          data is 9 pm, i.e. a time at which people are normally at home and could use Wi-Fi. •          drivers behind mobile prices ... suggest that higher levels of fixed line penetration are significantly correlated with lower mobile data prices. This is consistent with there being a material degree of substitutability between fixed and mobile services               475. We note the above to highlight the inconsistencies in submissions to regulators.              includes bundles that are used through data SIMS, MiFi 501 devices and through routers. 502 Mr Joosub sought to argue that Vodacom distinguishes between mobile broadband (a SIM and data package only) and FWA. 503 However, FWA would clearly be an even closer competitor to FTTH than mobile broadband              501 MiFi stands for Mobile Wi-Fi. A MiFi is a mobile router that, when equipped with a SIM card, connects to a provider of mobile broadband. 502 Transcript p 1852 lines 17  20. 503 Transcript p 1854 lines 17  21. Non-Confidential 131 only rings truer for FWA. However, the factual witnesses and economic experts in this matter agree that normal    and FTTH do not compete, the dispute here is about FWA and FTTH interaction. 476. Furthermore, the merger parties in their respective internal documents identify FWA/fibre as threats. 504 Vodacom has identified a strong fibre constraint on its business, calculated as Value at Risk, in addition to any other MNO competition.  al spectrum 477. It is common cause that Vodacom has acquired significant additional spectrum in the recent spectrum auction. The use of that spectrum will enable Vodacom              ICASA for additional spectrum allocation, it submitted: 505 477.1. There is congestion on its network, and the additional spectrum would allow the effective deployment of additional capacity not just to deal with congestion but also provide significant capacity for other services. 477.2.  business case was largely that there is low fixed penetration in South Africa and the additional spectrum would enable Vodacom an opportunity to provide a fibre-like service to customers, that fibre-like service referred to being 4G and 5G FWA. 477.3. The additional spectrum is required to support the ever-increasing demand for data services. 477.4. By acquiring access to the new 2600 MHz frequency band, Vodacom             capacity. 477.5. Amongst the 5G use cases identified by Vodacom is fibre-like mobile broadband and it is stated that 5G fibre-like mobile services, namely FWA, are anticipated to be a particularly important new product portfolio. 506 504 See, for example, Bundle M p 1212: 27 October 2021 CIVH Board pack and Bundle M p 3102: Project Lindt - Technical Assessment - 21 September 2020. 505 Exhibit AM; Transcript p 2199 line 14 to p 204 line 21. 506 Transcript p 2199 line 7 to p 2204 line 21. Also see Exhibit AM. Non-Confidential 132 478. Dr Van den Bergh confirms that Vodacom told ICASA that it would be able to expand its carrier footprint, improve customer experience, and enable and support a wide variety of internet of things applications. 507 479.              commercial use-                     capacity. 508 Mr Otty also confirms that FWA provides useful incremental     509 480. Even before the additional spectrum was acquired, Vodacom offered 4G through refarming with the spectrum that it had. 510 Dr Van den Bergh confirms that the                       should enable significant 4G coverage. 511 481. Dr Van den Bergh concedes that Vodacom did not tell ICASA that, with the increased spectrum, Vodacom would remain a congested network and barely able to serve the same customers as it had before the spectrum allocation. 512     482. Although Mr Nunes of MTN concedes that fibre and 5G FWA serve the same basic function of providing connectivity to end users, he still argues that they are not interchangeable. 513 Mr Nunes however concedes that FWA at least in the short term offers a tool by which MTN as an MNO is able to compete for market share. 514 507 Transcript p 2204 lines 6  10. Also see Exhibit AM. 508 Transcript p 2030 lines 3  6. 509 Transcript p 2039 lines 15  21. 510 Transcript p 2203 lines 10  14. 511 Transcript p 2204 line 22 to p 2205 line 10. 512 Transcript p 2205 lines 15  20. 513 FWB p 148 para 5.20; Transcript p 638 lines 2  3. 514 Transcript p 650 lines 11  16; p 678 lines 7  10. Non-Confidential 133 483.              consistent with its own internal documents, as we show below. 484. In response to the announcement of this proposed transaction in the market, MTN                     MTN to pursue a  strategy targeting and and operate under until they can be into the . 515 485. MTN (with Supersonic) also sought to reach its growth targets and attack any available opportunities by using both its fixed and FWA products. Its strategy is to obtain organic growth of customers and inorganic growth to add an additional customers. 516 486.                 low latency of 5G offers the opportunity 517 487.           is only possible in a narrow category of intended uses such as voice calls and single device video streaming 518 its FWA products are, in fact, able to perform many more functions, as is evident from its own advertising. 519 488. Importantly,      -                   Home is where                 515 Transcript p 647 line 20 to p 648 line 2; Bundle O p 231  MTN SA Board Plan B Update dated February 2023; also see Bundle O p 253 and following -    MTN SA  FTTX Way Forward, Role of Fibre and Options to Consider    516 Transcript p 648 line 21 to p 649 line 21; Bundle O p 222 and following  MTN SA Board Plan B Update dated February 2023. 517 Transcript p 707 lines 8  21. 518 Nunes FWB p 150 para 5.23.6. 519 Transcript p 683 line 19 to p 684 line 6. Non-Confidential 134  Blistering speeds for fastest gaming and streaming. Reduced latency 520 Mr      -  his response is  targeted at the consumer and what the consumer needs within his house. So, his needs, all right, for a better connected life. 521 489. MTN, in its efforts to compete by using FWA, is incentivised to price aggressively to grow its market share, which results in a benefit for consumers. 522 490. During cross examination, Mr Nunes sought to distance himself from any claim of substitutability between FWA and fibre and alleges that FTTH and FWA are not            product which saw litt          523 Airfibre however is a different product. It is a product developed by Supersonic      Wi-Fi frequencies to provide a point- to-point link. 524 Furthermore, the        distinguishable in that the former had a slow uptake whilst the latter, which is the source of comparison with FTTH, and as conceded by Mr Nunes, 525  We note that both Mr Smith and Mr Reynolds rely on this Airfibre example in their price comparisons without any adjustments based on the cross examination of Mr Nunes. 491. Recall also Mr Nunes evidence that        [consumers] using including old legacy technologies and fixed wireless access when asked by the Tribunal what the high number of South African consumers - in the high-income areas - use that have chosen not to use the rolled out FTTH in these areas (see paragraph 4100 above). 520 Transcript p 681 line 12 to p 682 line 22; p 688 lines 5  13; Exhibit J: MTN 5G Integrated Campaign;   Fibre now has competition 521 Transcript p 684 lines 7  15. 522 Transcript p 677 line 16 to p 678 line 13; also see Exhibits H and I. 523 Transcript p 638 lines 4  15; p 640 lines 15  21; FWB p 152 para 5.27. 524 Transcript p 664 line 20 to p 665 line 11. 525 Transcript p 669 lines 1  8. Non-Confidential 135 492.               for 4G and 5G FWA: between April 2023 and April 2024, its 4G FWA subscribers grew by % from approximately 9 to and its 5G FWA subscribers grew by % from approximately to 526 5G FWA competition with FTTH 493. It is common cause that 5G usage is growing globally, and that the devices and equipment associated with 5G will in future reduce in cost. 527 494.      in areas where other players have FTTH networks, suggesting that Vodacom does not see FWA as an inferior product that would not survive where fibre is available. This is contradictory                switch to fibre. As we have noted, this is further contradicted by the factual evidence regarding relatively low average penetration rates where FTTH has been rolled out in South Africa. 495.            of which he and his team were involved in, that in 2021 Vodacom recommended that going forward, it would in those areas where Vodacom has no fibre. 528 This suggests that Vodacom assessed that both products serve the same purpose from an end-customer perspective. Vodacom itself thus sees substitutability and competitive interaction between 5G FWA and FTTH. 496. Rain indicates that it is not seeing high churn rates of customers from its 5G in areas that received fibre, 529 suggesting most of Rain customers remain satisfied with their FWA service. This contradicts the argument that FTTH is superior to FWA. 526 Transcript p 666 line 16; p 667 lines 10  22. 527 Transcript p 2460 line 22 to p 2461 line 2. 528 Transcript p 2468 lines 3  17. 529 Rain letter dated 23 August 2022, para 4.1. Part B(2) of the Record p 6252. Non-Confidential 136 497.                market and a 4G market since you still have an FWA product and the customer receives a broadband product. He noted that the customers probably do not care what G it is for fixed wireless access 530 (see paragraph 4033 above). Responses in the market and price competition 498. We have referred to a study conducted by BMIT 531 showing that 75% of South Africans have a spend of R500 or less for internet services, and of that 75%, 50% only have the ability to spend R300 or below. 532 It is in this part of the latter market segment that Telkom sees its FWA LTE products compete with fibre- based products. Furthermore, Telkom indicates that it does consider FTTH               and fixed wireless access technologies are competing for a share of that wallet. 533 499. We have referred to Dr Scheffer confirming that to drive up connectivity rates in relation to fibre, there is constant and ever-increasing competition to formulate competitive products to attract customers to switch to fibre where those customers use solutions other than fibre, such as FWA (see paragraphs 402 and 4721 above). 500. Consumers have different needs, some require less usage at a cheaper price,              market and trade-offs are being made. Mr Hodge provides examples of the closeness of pricing by FWA and FTTH providers currently in the market. 534 His evidence on current Vodacom FWA pricing shows that it is lower than the 530 Transcript p 2036 lines 9  13. 531 A technology industry research and advisory firm. 532 Transcript p 298 line 18 to p 299 line 4; p 349 line 11 to p 350 line 11; p 462 line 6  10. 533 Transcript p 299 lines 2  6. 534 Exhibit BO Hodge Slides 6 and 16. Non-Confidential 137 Vodacom ISP advertised pricing for fibre, and substantially so as against Vodacom ISP services on the Vumatel network. 535 501.               problematic (price drop comparison) 536 or could not be used to make inferences on the competitive interaction question (correlation analysis). 537 There are two issues with  analysis: (i) it does not deal with FWA penetration but                                     and we know these markets are              538 Mr Reynolds had no convincing response to these criticisms. We note that Mr Reynolds himself finds that LTE FWA pricing is similar to low-cost FTTH products. 539            comparison of usage levels using data from 2021 was unreliable and ignored other evidence. 540 502. What we have found to be persuasive evidence of the competitive interaction between FWA and FTTH, is that FWA has over time responded in the way it is marketed and sold to more closely mimic fibre. FWA used to be marketed on a Gigabits basis with unlimited speeds, and now is throttled and offered at different speeds and marketed as uncapped. 541 For example, MTN reduced its pricing since 2021 by 46% for 10Mbps, 39% for 20Mbps and 38% for 35Mbps. 542 MTN currently (when the analysis was done) prices 10Mbps at R269, 20Mbps at R429, and 35Mbps (5G) at R499, all with a free router. MTN has also introduced faster 5G FWA packages priced at R699 for 60Mbps and R999 for unlimited 535 Exhibit BO Slide 6. The Vodacom leaflet on the left shows that Vodacom offers FWA packages from R299 with a 1TB 50Mbps service at R399 and a 2TB 100Mbps service at R599 whereas the cheapest fibre package offered was R629 on Octotel, but on Vumatel it is R749. If this is a 50Mbps service, then it is almost double the lower range FWA price. 536 Transcript p 3498 line 9 to p 3503 line 9. 537 Transcript p 3506 lines 7 to 16. 538 Transcript p 3506 lines 11  16. 539 Reynolds EWB p 475 para 4.63. 540 Transcript p 3506 line 16 to p 3507 line 15. 541 Hodge Transcript p 3299 lines 1  21. 542 Exhibit BO Hodge lide 17. Non-Confidential 138 speed. These are competitive against Vumatel FTTH priced at R449 for 25Mbps, R757 for 50Mbps and R927 for 100Mbps, and Vuma Reach at R399 for 20/10Mbps, R529 for 40/10Mbps and R987 for 100/50Mbps. 543 503. All of the above evidence, together with the evidence covered in the market delineation section, in our view suffices to conclude that there indeed is competitive interaction between FWA and FTTH, as we have noted, specifically regarding the marginal customer. 504. We nevertheless deal with some of the more technical, supply-side arguments of the merger parties. They elevate the speed of fibre and assert that where fibre is available, customers will invariably opt for fibre. 544 As we have noted, what is important is not only what engineers assess are the technical advantages of fibre, but what consumers who buy these products do in practice. To quote Mr Masalesa, consumers buy data, not technology. 505. Although fibre, from a technical perspective, offers better performance in certain respects, this does not mean that FTTH and FWA do not compete for customers, specifically for the marginal customer (as we have explained in the market definition section). The merger parties and MTN could not, from a demand-side perspective, explain away the large number of South African consumers that in making their decisions about accessing the internet  in the more affluent areas of our country  do not opt for fibre despite FTTH being available to them. This means that FTTH to a very large extent does not displace the alternatives, 4G and 5G FWA, upon entry, since a high proportion of households do not switch to FTTH and are using alternatives and consider them to be viable. How competitive the FTTH is with the FWA will influence its penetration levels and ARPUs, and how competitive FWA is when fibre enters an area will determine how many customers churn. 543 Exhibit BO Hodge lides 6 and 16. 544 See for example Mare Transcript p 2664 line 12 to p 2669 line 8. Non-Confidential 139 506. The other common cause key issue as we have highlighted is that there is no FTTH-to-FTTH competition in many local areas since FTTH FNOs tend not to overbuild one another. 507. There is no need for us to deal in detail with the other technical arguments about the alleged superiority of fibre. One can accept that FTTH and FWA technologies have their pros and cons. None of these technical differences change the fundamental facts that we have traversed above. We nevertheless below summarise some of the issues. 508. On reliability FWA may be at risk of unstable electricity supply and transmission line breaks, fibre is at risk of fibre breaks to individual homes. 509. Loadshedding was debated as part of assessing the comparative reliability of the technologies. T           factors such as loadshedding, it is less reliable than fibre. However, the industry has taken measures to ameliorate issues caused by loadshedding. 545 In the case of Telkom CSB, for example, if service is affected at a particular physical           between different sites, providing backup service from neighbouring sites. 546 Where fibre services are affected due to loadshedding, the customer may be affected for longer periods. 510.             reliability in that the connectivity of fibre will remain stable as long as there is secure cable connection, whilst that of wireless will remain stable in areas where there is network coverage. The main cause for outages in Rain's network is unreliable electricity supply and frequent fibre breaks in the transmission network connecting its towers to its core infrastructure. FTTH operators are equally impacted by fibre breaks in the transmission networks, but also by breaks in the fibre connecting individual subscriber homes. 547 In terms of fibre 545 Masalesa Transcript p 296 lines 1 to 8. 546 Transcript p 296 lines 8  18; FWB p 21 para 7.3. 547 Bundle M p 8575. Non-Confidential 140 breaks (for example contractors digging up cables and damaging the fibre infrastructure or vandalism), there is no backup for fibre, and therefore it can take time to resolve these issues from a fibre point of view. 548 511. Fibre lacks redundancy whilst this may be provided in wireless through other towers. Rain submits that FTTH does not offer any redundancy as once the cable breaks there is no second option for connection. Often restoring a fibre break to a user's premises may involve civil work and consequently it may take significantly longer to restore than a wireless service. Wireless operators are often able to offer redundancy by allowing subscribers to connect to the internet using a different technology within its coverage area, i.e. 5G to 4G or a different tower using the same technology, provided there is coverage in the area. 549 512. Mr Van der Merwe testifies that many homes are aggregated to the same single fibre strand that goes to the head end equipment and so share that resource. 550 He explains that latency and buffering are challenges, but that fibre can have similar limitations: depending on how it is deployed, FTTH and FWA share common characteristics; fibre can also be a shared resource and face similar limitations as FWA, particularly where a single fibre strand is split between multiple customers, which can cause higher latency and degraded bandwidth performance; it is about how the technology is deployed. 551 513. Whilst FTTH may have an advantage with subscribers that want high-throughput packages, wireless offers the flexibility to move with the router to different locations. FWA routers are nomadic, i.e., mobile, in general without requiring             the router can be moved to an optimum position. 552 Fibre on the other hand is a physical cable that terminates at a point in the building or house and there is equipment that is placed next to that. 548 Transcript p 296 line 18 to p 297 line 5; FWB p 21 para 7.4. The evidence on Mean Time to Repair is dealt with under the analysis of vertical theories of harm further below. 549 Bundle M p 8575. 550 Transcript p 145 line 11 to p 146 line 2. 551 Transcript p 144 line 18 to p 145 line 10. 552 Transcript p 297 line 21 to p 298 line 12; Masalesa FWB p 22 para 7.6. Non-Confidential 141 514. FWA, because it is a wireless offering without the need for significant infrastructure            553 515. Both are able to offer value-  VAS      VAS is more suited to mobility and fibre where VAS is more suited to large uplink. Fibre is better placed to offer guaranteed speeds but ISPs are often congested, resulting in lower effective speeds. Speed 516. Customers using the internet have different demands. Some consumers will require maximum speeds and others not, depending on what they use the service for. Products with different speeds are offered in the market. As we have indicated above, FWA has over time responded in the way it is marketed and sold to more closely mimic fibre. 517. Mr Masalesa indicates that for 5G FWA the connection speeds are comparable to those provided by fibre, but not for 4G. 554 5G reaches speeds of over 200Mbps. 518. Rain submits that both fibre and wireless services are able to offer consumer level broadband speeds. In many cases, South African customers achieve higher download speeds from their 5G services than the fibre operator in the area is able to offer. Wireless services are however impacted by signal quality and congestion. Operators manage congestion by increasing capacity in their networks, but a challenge for wireless operators is that it is not always that easy to determine exactly what signal quality and consequently throughput a customer will achieve. 555 553 Masalesa Transcript p 347 line 19 to p 348 line 2; Masalesa FWB p 22 para 7.5. 554 Transcript p 294 line 8 to p 295 line 1; p 335 lines 7  15; FWB p 21 para 7.1. 555 Bundle M p 8575. Non-Confidential 142 519. Whilst FTTH offers the highest speeds, depending on use cases, some customers may opt for entry level products with entry level speeds and are satisfied with speeds of 10Mbps and 20Mbps, with the most popular fibre speeds being 50Mbps 556          evidence is that 20Mbps is a sufficient speed to consume broadband services like streaming without issues with buffering on the video. 557    experience, the majority of its FTTH customers use 50Mbps with a very small number above 100Mbps. 558 LTE FWA is marketed by Vodacom with speeds up to 100Mbps. 520. Mr Nunes confirms that products with a speed of 10Mbps are sufficient for many consumers that do not stream and states that in the last two years the shift to 50Mbps has become the norm. 559 This is consistent with the evidence given by Mr Masalesa and Mr Motlekar. For fibre, average speeds are 50Mbps and the majority at 100Mbps or less, speeds that lie within the range for FWA services, even on 4G. Congestion 521. Recall that Vodacom argues that it would not have the spectrum capacity to provide a sustainable FWA service for large numbers of FWA customers (in addition to its mobile services) as its radio network could not carry the load. 522. The evidence suggests that players running mobile networks all have challenges with congestion and they manage it. They have an incentive to manage congestion because it affects their revenue, and once a certain threshold is reached, additional investments will be made by them to relieve congestion. This is the common industry practice. 556              557 Transcript p 467 lines 7  11. 558 Transcript p 342 lines 6  11; p 344 lines 8  14. 559 Transcript p 693 lines 8  18. Non-Confidential 143 523.            operators use congestion thresholds as prompts to expand capacity. Mr               congesting. It needs to gradually get to that point where it congests and generally speaking we have enough time to deal with that congestion 560 Capacity is added because there is opportunity to grow revenue given there is demand in the area. 561 524. The mobile operators use various methods to monitor and manage data demand daily as network utilisation fluctuates, depending on factors such as marketing campaigns which may cause a spike in data traffic. 562 It is managed through a             also ways of deploying additional technologies, additional bands in terms of spectrum, and also moving up to Massive MIMO 563 . 564 525. Telkom further indicates that one of the many avenues available to it to deal with the issue of congestion is that it can manipulate the angle of antennas and direct customers who hog any radio coverage to other more resourced sites to relieve capacity where there is high utilisation. 565 526. Telkom does not dedicate any spectrum for any particular service but instead steers FWA devices away from spectrum which encourages FWA devices to   -capacity spectrum as opposed to its mobility spectrum (or FDD spectrum). 566 527. The industry has matured such that there is a proliferation of tower companies whose sole function is to build towers with the benefit that densification no longer 560 Transcript p 375 lines 10  14. 561 Transcript p 304 lines 3  8. 562 Transcript p 229 line 21 to p 300 line 6. 563 Mr Masalesa describes Massive MIMO as a technology where an operator can take an existing site, on which it has used up all the spectrum that it has, and it conducts an upgrade to Massive MIMO which can double the capacity that the operator has at the site level. Transcript p 302 line 8  13. 564 Transcript p 561 lines 13  18. 565 Transcript p 378 lines 2  13, p 378 line 18 to p 379 line 12. 566 Transcript p 309 lines 7  19. Non-Confidential 144 requires actually building a site but instead refers to negotiating with the particular tower company for access to a particular site to allow the operator to                 basis. 567 528. Dr Van den Bergh of Vodacom testifies that it has certain benchmark measures for congestion and it continues to invest capex in relieving the observed congestion. Data presented by Dr Van den Bergh shows the percentage of 4G cells congested in FY2024 was % reducing in the April 2024 to June 2024 quarter to %. 568 Dr Van den Bergh testifies that Vodacom invests capex to relieve        % congestion. 569 529. Notably Dr Van den Bergh indicates that Vodacom currently does not prioritise between data for and 570 Importantly, despite the stated concerns about congestion, Dr Van den Bergh confirms in response to questions from the Tribunal that Vodacom does not presently at all as congestion levels in reality have not reached a . 571 Furthermore, Vodacom is able to make adjustments in their annual financial planning cycle to respond to network information that the network in a particular area may be this level 572 , consistent with Mr M            which MNOs are not able to respond effectively, and that they will address constraints where required. This is not surprising because the FWA revenue is incremental which Vodacom would want to keep. 530. Vodacom argues that it must prioritise its mobile businesses over its other businesses, and that its network is congested. The testimony of Dr Van den Bergh however shows that the level of congestion experienced by Vodacom is 567 Transcript p 304 line 18 to p 305 line 12. 568 Exhibit AL:    Presentation slide  Data Traffic growth is exceeding           569 Transcript p 2346 lines 5  7. 570 Transcript p 2282 lines 7  18. 571 Transcript p 2286 lines 2  13. 572 Transcript p 2283 to p 2285. Non-Confidential 145 All MNOs manage and continuously invest to relieve congestion and grow their businesses including FWA. Importantly, Vodacom continues to market and grow its FWA business with data volume growth for FWA greater than mobile. This shows that congestion is not a factor currently and Vodacom continues to invest capex to relieve congestion in any event. FWA according to its own documents is a strategic focus. 531.             auction and what its submissions were to ICASA. 532. Furthermore, as per government policy, 2G and 3G technologies will be phased out around 2027. 573 Dr Van den Bergh agrees that this transition from lower technologies will enhance actual capacity on the network. 574 533. Furthermore, more spectrum is likely to become available in terms of the ICASA              of its allocated spectrum, and there is more to come: 575 ADV MUVANGUA: Am I understanding correctly from what you just said, that with reference to ICASA that more spectrum is on the way? DR VAN DEN BERGH: Certainly there's going to be more spectrum. There's spectrum on the [ICASA] roadmap 534.         most of us are aware of the fact that the regulator had intended to go on a second phase of spectrum                             is still a wash off spectrum 576 He said that this spectrum auction possibly will take place next year. 577 573 Transcript p 2208 lines 7  9; p 2220 lines 4  14. 574 Transcript p 2210 lines 5  8. 575 Transcript p 2268 line 22 to p 2269 line 4. 576 Transcript p 318 lines 8  20. 577 Transcript p 318 line 22 to p 319 line 2. Non-Confidential 146 535. Dr Van den Bergh testifies that even without the additional spectrum to come, Vodacom has s for about for . 578 536. Notably, data provided to the Commission in July 2024 shows that there is currently n     permitting room for its growth. Dr Van den Bergh   site 579 There is also no congestion that cannot be managed in the vast majority of areas offering 4G FWA as is evident from the Vodacom coverage map. 537. Moreover, Vodacom is currently under-utilising the spectrum it procured in the ICASA auction. This is reflected in data that the Commission received from Vodacom on 12 July 2024. 580      configuration for each of its sites (as per the data at exhibit AN). The analysis was presented to the Tribunal as exhibit AO. Dr Van den Bergh argues that this analysis is not of value as it does not identify where the different sites are located, as remote sites may use limited spectrum. 581 However, in the context where % of sites use less than 90MHz, or available to Vodacom, it is apparent, that the majority of sites in urban areas have by 539. Furthermore, more capacity is released as 5G demand grows and displaces 4G, and as 2G and 3G are phased out by 2027. Dr Van den Bergh confirms: ADV MUVANGUA: So, as 5G grows, 4G declines, and this frees up capacity on the spectrum. Do you agree? DR VAN DEN BERGH: I do 582 540. Vodacom further has access to substantial additional spectrum, through its roaming agreements with Liquid and Rain, which provides it with more capacity. 578 Transcript p 2319 lines 10  20. 579 Transcript p 2310 lines 20  22. 580 Exhibit AN. 581 Transcript p 2227 lines 3  14. 582 Transcript p 2296 lines 17  19. Non-Confidential 147 Vodacom allocates MHz to in the 3500MHz band by roaming on Liquid spectrum. 583 This arrangement with Liquid gives it access to MHz, which can be used for 584 Vodacom also uses wholesale 4G FWA capacity on Rain. 585 541. The final point is that, and as conceded by Dr Van den Bergh, as more homes get connected with FTTH, then this too will create more capacity on the networks. 586                              587 It follows that as fibre is rolled out in an area, constraints are lifted in precisely those areas as FTTH is taken up. And as the Vodacom submission to ICASA identifies, in relieving the capacity constraints, this results in lower prices to not only meet fibre pricing but also to use the excess capacity that the network now has. Mr             unused capacity             in spectrum. 588 542. The merger parties argue that the above confuses issues since the location of the alleged offload is (by definition) in areas where fibre is already installed. They say that this assumes that customers in the relevant areas would be connecting to fibre, and so there would be no reason to want FWA. However, as we have indicated not all customers will choose fibre. The factual evidence is that the average FTTH penetration rates are low and therefore customers will continue using LTE/4G and 5G for internet access. 543. Vodacom has the most sites in South Africa. When assessing overall capacity to provide FWA on a network, one must consider the total number of sites and the number of users that can be accommodated per site. Mr Masalesa indicates that 583 Transcript p 2264 lines 4  21. 584 Transcript p 2265 lines 11  22. 585 Transcript p 2266 lines 9  11. 586 Transcript p 2210 lines 9  12. 587 Exhibit AG p 13057  p 13058        Market Inquiry into Mobile Broadband Services 588 Transcript p 2030 lines 3  11. Non-Confidential 148 in terms of capacity for FWA per site, Telkom provisions for users on each site for FWA 589 but that he can double this with Massive MIMO technology, and he furthermore has the option of site densification and more spectrum. 590 544. Dr Van den Bergh states that Vodacom can only do roughly that amount, or around users per site using its 2MHz spectrum for MHz 591 ) 592 or half what Telkom can achieve. Considering 5G, Dr Van den Bergh indicated that one can get a t people. 593 In addition, Vodacom can access additional spectrum from roaming agreements, namely the MHz from Liquid in the MHz band 594 and roaming on 595 Thus, considering all the spectrum available to Vodacom, and if it were to use its actual capacity per site is likely greater than 190 users. 545. As indicated, site density must also be considered. Vodacom has the sites in South Africa, currently sites 596 and at users per site this theoretically would provide overall capacity for FWA users. We acknowledge that as some of the sites are not in , the actual capacity would be This figure however does not consider 5G technology     . 546. In             the argument that congestion is a future primary constraint convincing. It is not consistent with its own submissions to ICASA, and the fact that it is currently not constrained to provide FWA. Furthermore, Vodacom has acquired significant additional spectrum, with a ICASA roadmap for further spectrum to come, and 2G and 3G technologies to be phased out around 2027. What is further significant is that as 589 Transcript p 338 lines 1  9. 590 Transcript p 381 line 8 to p 382 line 10. 591 Van den Bergh FWB p 195 Figure 4. 592 Transcript p 2168 lines 1  9. 593 Transcript p 2168 line 18 to p 2169 line 4. 594 Transcript p 2264 line 16 to p 2265 line 12. 595 Transcript p 2267 lines 1  22. 596 Exhibit AO                            2024- 07- Non-Confidential 149 FTTH is rolled out to the lower-income areas, that will further lift congestion constraints.         547. The merger parties argue that Rain and Telkom are better positioned than Vodacom to offer FWA services as they currently at a national level account for most FWA connections and this competitive interaction will remain post-merger, together with other primary constraints. 597 548. The merger parties further argue that Vodacom needs its spectrum resources to service its mobile customers and will not have capacity for FWA customers. They add that Vodacom needs the new spectrum obtained in the recent auction to meet mobile demand on existing sites. However, as we have indicated, this is not consistent with what Vodacom told ICASA in its spectrum application. 549. Having regard to all FWA technologies (including non-mobile fixed wireless access), there are approximately 1.3 million connections in South Africa as of June 2023. 598 550. There are five main 4G/5G FWA providers in South Africa: As of June 2023, Rain is the largest FWA network operator with consumer FWA connections ([30-40]% of the market as at June 2023). Telkom is the second largest FWA network operator with consumer FWA connections ([30-40]% of the market as at June 2023). Telkom offers a wide range of FWA products at various price points, with options including router or SIM-only offers and regular           across South Africa but is mostly distributed in the more densely populated and high-demand urban and peri-urban areas of South Africa. MTN is the third largest FWA network operator with consumer FWA connections ([10- 20]% of the market as at June 2023). Vodacom is the fourth largest FWA network 597 Transcript p 57 lines 11  17. 598 Reynolds EWB p 451 para 2.33. Non-Confidential 150 operator with consumer FWA connections ([10-20]% of the market as at June 2023). Cell C is the smallest FWA network operator with 25,000 consumer FWA connections (2% of the market as at June 2023). 599 551. FWA services are also sold by ISPs at the retail level. 552. We again note that one should not take a static approach to the market and must consider evidence related to company strategy     growth in 5G, as well as the additional spectrum that it acquired in the latest auction,             fibre is rolled out in the lower-income areas that we are concerned with, that will        553. As we have noted, competition analysis is forward looking and in our assessment Vodacom would offer FWA services on a larger scale in future given its significant spectrum holdings, spectrum roaming and far greater site density than Rain and Telkom. Furthermore,    strategic documents, FWA, as a business case, for it in order to and particularly as household demand and use cases for data grow over the foreseeable future. 554. We note the following from        spectrum: (i) it has been able to achieve % population coverage; (ii) data traffic is significantly growing year-on-year; (iii) due to the low fixed penetration compared to mobile penetration in South Africa there is also an opportunity to provide a fibre-like service to customers; (iv) by acquiring access to the new MHz frequency band, Vodacom will be able to network capacity; (v) given Vodacom's already extensive network, customers will , in terms of once Vodacom begins to ; 599 Reynolds EWB p 450  451 paras 2.30  2.32, Table 6. Non-Confidential 151 (vi) GHz and in the GHz band to deploy 5G on will allow Vodacom to provide services in and M . The type of services that 5G could enable are l (vii) 5G-enabled fibre-like mobile services (fixed wireless access - FWA) are anticipated to be a product portfolio; (viii) access to additional spectrum will on the Vodacom network and, 600 implementation of new technologies like 5G, Vodacom will be able to from both perspective and . 601 (Own emphasis) 555. As %. 602 As we shall discuss under the public interest assessment, it plans to significantly increase the number of 5G sites in the (see paragraphs 1168 to 1170). 556.            subscribers grew very significantly. The spectrum auction provided Vodacom with significant capacity in LTE and Vodacom provided statistics of its 4G FWA subscribers. 603            5G FWA subscribers grew by 2% to c. in the short period from September 2023 to April 2024, or at a rate of c. per month. Dr Van der Berg accepts these figures provided by Vodacom. 604 This illustrates how Vodacom can, absent this transaction, grow and compete against FTTH with its 4G and 5G offerings. 601 See Exhibit AM: Fundamental assumptions for the business plan with financial forecasts & economic efficiency. 602 Reynolds EWB p 538 para 7.7. 603 Exhibit AP1: Number of subscriptions and data usage for Vodacom consumer FWA subscriptions by tariff, April 2024. 604 Transcript p 2245 line 15 to p 2246 line 10. Non-Confidential 152 557. As we have indicated, the land grab is moving to the lower-income areas. Mr Van d              the two major mobile operators, Vodacom and MTN, are dominant in the   there are many townships where Vodacom is absolutely                 they might have better coverage of that township, yes 605 Mr Hodge echoes this                      (referring to Vodacom and MTN). 606 558.                    607 This shows that Vodacom, MTN, Rain and Telkom all cover these 20 largest townships with LTE, and some of these players cover some townships with 5G. 608 There are however more than 500 townships in South Africa, and therefore the 20 largest townships are a very small sample and not representative of the whole of South Africa including the many smaller townships. 559. Regarding FNO fibre-to-         to not overbuild each other; and there are many local areas where there are monopolies in FTTH that are not constrained by other FNOs. As mentioned, in % of Vumatel areas there is no fibre overbuild and so FWA is an alternative home broadband option for consumers in those areas. Here FTTH and FWA compete for the marginal customer as we have shown. 560. Furthermore, Vodacom and MTN have higher quality networks and pervasive coverage. 609 605 Transcript p 241 lines 5  20. 606 Hodge Transcript p 3303 lines 8  10. 607 According to the 2011 Census. 608 EWB Appendix D p 595 Table 25. 609 Otty FWB p 357 para 11. Non-Confidential 153 561. From a costs perspective, Mr Hodge submits that Rain and Telkom are unlikely               high. 610 562. Vodacom, because of its extensive network, also has other advantages over                for someone who has a site large as Vodacom already is very low cost to put up                  whereas an ISP has their input cost 611 563. Vodacom and MTN both have acquired significant additional spectrum. Relative to the other players they were the two players that bought the most spectrum. Vodacom now has as much spectrum as Telkom. As indicated it further has roaming agreements with Rain and Liquid. 564. As the GSMA identified, FWA pricing has to be competitive against FTTH to woo customers. 612 As Vodacom submitted to ICASA, fibre offload from mobile results in lower prices as not only does it release capacity but also it is forced to respond on price. 613 565. Mr Van der Merwe explains what the proposed merger means for future pricing in the townships. FNOs, including Frogfoot, have begun deploying fibre in the townships. Post-merger, once Maziv deploys fibre in townships there will be no incentive for other FNOs including Frogfoot to enter those townships with low FTTH pricing because they tend not to overbuild because of the economics. 614 This will sterilize a large portion of the market for competitor FNOs and entrench    Importantly, it will shape pricing, innovation and consumer choice in these markets for the medium- to long-term, most likely irreversibly. 610 Hodge Transcript p 3303 lines 5  8. 611 Hodge Transcript p 3301 lines 18  21. 612 Hodge EWB p 98 para 142. Bundle M 10967: The 5G Guide April 2019. 613 Exhibit AG p 13057  p 13058. 614 Van der Merwe Transcript p 106 lines 6  11. Non-Confidential 154 566. Price is not the only factor on which competition takes place in FWA, as MNOs have also sought to offer a range of speed and price plans, improve the data       FUPs       free routers even for month-to-month deals. These are all to improve the value        will play price versus FUP versus speed, and offer differences. And even the different mobile operators might take different positions 615 Indeed, the growth in data demand (including during the COVID-19 pandemic), interacting with the expansion of fibre as an alternative in South Africa, appears to have stimulated significant innovation amongst the MNOs, and most importantly the two smallest MNOs, to provide customer-centric data solutions. 567. More aggressive pricing or other value propositions by Vodacom absent the proposed merger would force Vumatel to respond with lower prices to raise uptake rates in the areas where it operates. Hardiman and Mr Van der Merwe indicate that uptake or penetration rates are impacted by FWA competition, resulting in an FTTH response through lower prices to drive uptake 616 given the incentive to improve uptake due to the sunk cost of fibre deployment in an area. A better range of speed and price packages of FWA would also force Vumatel to continue to offer its own range of speed plans, including lower speed plans that are often retired so that consumers are forced to upgrade to higher speed plans with higher prices. 617        (have to) upgrade post-merger, which will not be in their interest if their choices are thereby being limited (see paragraph 453 above). 568. Post-merger Vodacom has less incentive to price FWA as aggressively as pre- merger, or improve the value proposition of the packages in other ways such as better FUPs or lower router costs, given that it will recoup a large portion of lost spend with it through its up to 40% share of the profits of Maziv. Aggressive price            615 Hodge EWB p 60  97. Exhibit BO Hodge Slide 6; Hodge Transcript p 3501 lines 10  21. 616 Hardiman identifies that targeted FWA in Reach areas may impact penetration and ARPU. Mr Van der Merwe identifies that FTTH needs to then offer more affordable i.e. cheaper products to raise penetration levels. Transcript p 85 line 8 to p 86 line 2. 617 Transcript p 3579 lines 1  9. Non-Confidential 155 mobile and post-merger FTTH interests. Less aggressive pricing by Vodacom will place less competitive pressure on Vumatel to itself price more aggressively to increase FTTH uptake rates in all the areas where it operates. 569.         -merger Vodacom will have less incentive to prioritise 5G rollout in the Vumatel areas and to market both LTE and 5G FWA aggressively in those areas. We have indicated that         deployment of 5G FWA be n Recall that        internal documents, the creation of which he and his team were involved in, that in 2021 Vodacom recommended that going forward, it would in those areas where Vodacom has no fibre 618 (see paragraph 4955 above). 570. The merger parties i           Vodacom were to slow down the rollout of 5G, that would be the same as it writ[ing] its own death warrant 619  [y]ou have to keep pace with     620 He went on to say that: So, you have to keep                             and so on and so on, customers will move, so you do have to keep pace with        just spent 5.4 billion buying spectrum, so if               spectrum 621 . This confirms how imperative it is for Vodacom to significantly grow its 5G offering in South Africa to keep pace with technology. It does not have an incentive to slow down its 5G rollout but can post-merger target and align this 5G FWA to avoid competition where Maziv has fibre. It will know     feasibly coordinate their strategies. 618 Transcript p 2468 lines 3  17. 619 Transcript p 1659 lines 8  10. 620 Transcript p 1659 lines 8  10. 621 Transcript p 1659 line 17 to p 1660 line 13. Non-Confidential 156 571. The Vodacom Group        5G Strategy and footprint  622 The strategy is this: to 5G FWA, Vodacom will utilise 5G capacity to provide a experience. 623 By February 2024, Vodacom had rolled out 5G FWA sites. It intends to reach sites by 2029. 624          underplayed when it comes to its tendered public interest commitments, this will be dealt with in the public interest section.) 572. The merger parties,      submit that Vodacom has committed to spending R60 billion in South Africa in the next five years of which a large portion will be used to deploy 5G:      we will invest 60 billion in South Africa post the transaction in the next five years,                  625 This implies significant increased future competition from Vodacom in 5G that can be        FTTH. Conclusion 573. The competitive interaction between FWA and fibre is borne out by the evidence. This is important considering that FTTH is shifting to the lower-income segment of the market where there will be competition between FWA and FTTH for the marginal consumer. 574. For a very large part of the Vumatel FTTH areas, there is no overbuild and hence the only competition can come from FWA for home broadband services. The factual evidence is that average FTTH penetration levels in South Africa are relatively          to its shareholding in Maziv, the proposed transaction will chill competition in those areas resulting in harm to consumers, in a growing market. Absent the proposed transaction, Vodacom will likely compete more aggressively with its 622 Bundle M p 5626. 623 Transcript p 2307 to p 2310. 624 Transcript p 2309 lines 4  14. 625 Transcript p 1659 line 21 to p 1660 line 10. Non-Confidential 157 FWA, and in the areas where Vumatel has rolled out FTTH. That will increase absent the proposed transaction, forcing Vumatel to respond on price, and on overall value-proposition to consumers, including a mix of speed, FUP and router packages. Price levels influence affordability and usage, both of which are harmed by higher pricing. 575. Furthermore, as the proposed merger would be permanent, it will likely entrench Maziv as the leading FTTH provider going forward, and the harm to competition (together with the foreclosure effects that cannot effectively be remedied) will grow over time. The proposed transaction enables both the merger parties to strengthen their market positions and reinforce and grow existing concentration in the telecommunications sector as a whole. 576. As we find under the effects of the proposed transaction on the sector in the public interest section, this transaction between the largest MNO in South Africa, and the largest dark fibre provider and wholesale FTTH player, will likely trigger further consolidation in the sector since other players would need to compete with the Maziv/Vodacom combination. This is borne out in the strategic documents of MTN as the second largest MNO that is very concerned about its ability to compete post-merger. What this means is that the abovementioned lessening of competition between FWA and FTTH will spread further as other players pursue deals involving FNOs and MNOs. This would result in a structural change that would substantially lessen competition. As we have indicated, the evidence shows that MNOs with either no, or with a small, FTTH footprint are more aggressive on FWA home broadband substitution. 577. Apart from a divestiture condition in relation to the overlapping FTTH infrastructure between Maziv and Vodacom (assessed further below under remedies), the merger parties proposed behavioural remedies attempt to address the vertical effects and not the horizontal (i.e., price and non-price competition parameters such as FUPs) effects. Non-Confidential 158 578. We note that the above horizontal effects must be considered together with the foreclosure effects of the proposed transaction on competitors including Telkom, MTN and Rain. All of these players have raised substantial vertical foreclosure concerns, as we shall discuss next under the vertical effects analysis. We also note that many more third parties that did not testify in these proceedings raised          METRO FIBRE AND FTTB 579. The Commission argues that one must not take a static view of the effects associated with the proposed transaction in the metro fibre and FTTB markets when assessing dynamic future competition, and must have regard to the competition counterfactual, specif       strategic documents. The merger parties rely inter alia   - merger market positions and argue that Vodacom currently is too small in both metro fibre and FTTB for the proposed transaction to result in a substantial prevention or lessening of competition. 580. Metropolitan fibre infrastructure comprises tower connectivity, as well as fibre infrastructure that connects core sites, branch sites and aggregation nodes (or transmission), which is used as backhaul infrastructure for the provision of FTTx. 626 581. Metropolitan fibre infrastructure (and particularly backhaul infrastructure) is a critical input for the provision of wholesale FTTB and FTTH. 627 582. Metro fibre is used by network operators, both FNOs and MNOs, to aggregate traffic from their last mile access networks (both fixed and mobile) or aggregation nodes (including major datacentres) and transports aggregated traffic between these networks and nodes, connecting them to the national long-distance infrastructure. Therefore, FNOs and MNOs are the primary drivers of demand 626 Nunes FWB p 162 para 6.37. 627 Nunes FWB p 162 para 6.39. Non-Confidential 159 for metro fibre connectivity. These operators acquire fibre links from infrastructure providers such as DFA, Telkom, and Link Africa in the form of metro backhaul links in the case of FNOs, and fibre to the site/tower (FTTS/T) in the case of MNOs. 583. Metro fibre networks typically offer wholesale FTTB as their metro network brings them past businesses to whom services can then be sold through a break-out link. The FTTB business ensures demand for the metro fibre network too. As metro networks are extended to support the FTTB, there is a growing densification of the network and a metro network with a large FTTB portfolio is likely to be denser than rivals. 584. MNOs have provided the demand foundation for companies such as DFA but also have self-built metro networks given they have substantial demand from their own network requirements which also require a densification to provide coverage. 585. Operators with a large metro fibre network are also likely to hold a sizeable position in wholesale FTTB because the metro network rings are deployed using     628 namely they are designed to pass as many customer opportunities as possible and therefore have denser networks in business districts within metro areas. As metro fibre rolls out on the back of MNO demand, that provides an opportunity to pass businesses and develop FTTB demand.    Expansion of mobile networks / Titan 629 links increases the size of the addressable market for enterprise services (i.e. greater densification) 630 586. FNOs leverage off the metro backhaul infrastructure to supply FTTB services, as they need to be close to metro fibre to support FTTB. 628 Part B of the Record p 1305 para 61.3.3.2. Frogfoot letter to the Commission dated 30 March 2022. 629 Titan and Peregrine are categorised as wholesale metropolitan fibre backhaul. These are dark fibre                  own equipment to enable data transmission. 630 Bundle M p 1061: Vodacom Project (20 October 2021). Non-Confidential 160 587. In terms of the proposed deal, Vodacom will transfer its metro fibre transmission links which connect Vodacom base station sites to the access and aggregation layers of its transmission network. After the transaction, those ducts and links that connect to the Vodacom base station sites will be owned by DFA, and Vodacom will use the fibre in those ducts for its own use and to provide managed services. 631 588. The rollout of last mile FTTB has historically been driven by customer orders as business/enterprise connections are typically built to order. Pricing for FTTB products is based on individual negotiations and may vary by customer depending on service type, quantity and bargaining dynamics. 632 589. With regard to the metro fibre market, Prof Theron argues that (i) there is no overlap between Vodacom and DFA regarding the provision of dark fibre access; and (ii) the merged entity will have a combined national market share (dark and lit) of less than 30%  although it varies across regions and provinces. Thus, if one considers current market shares, there is no fundamental change in the structure of the market as a result of the proposed transaction and there is no major market share accretion warranting concerns over a substantial prevention or lessening of competition. 633 Prof Theron also argues that Maziv has a current competitive constraint in the form of Telkom/Openserve. 634 590. In respect of the FTTB market, Prof Theron argues that the merger does not substantially prevent or lessen competition and that there is no change in the market structure as a result of the proposed transaction for three main reasons, namely (i) pre-merger          ISPs do not have access to the Vodacom FTTB infrastructure; (ii) there is a             and (iii) in terms of market share accretion as a result of the proposed            631 Joosub FWB p 321  322 para 7.3. 632 Hodge EWB p 82 para 98. 633 Transcript p 3343 line 9 to p 3345 line 6. 634 Transcript p 3348 lines 1  10. Non-Confidential 161 therefore there is no change in the market structure of FTTB as a result of the proposed transaction. 635 591. Mr Hodge, on the other hand, considers dynamic future competition from Vodacom absent the proposed transaction. He paints a picture of the future            on the evidence it removes a potential alternative. Vodacom, certainly it was the                actually provide an alternative open access network because it is the largest MNO with the largest volumes and can commit, and even the story of DFA, as              reach and density. So, we are at that point that critical inflection point, where we have to be careful about how we set this path going forward 636 592.                                         counterfactual. And we know that Vodacom very publicly has said it has a FibreCo and Telco strate         be done off balance sheet with JV partners which could include other mobile operators, other fibre operators, financial institutions, and other partners. And,                              637 593. The merger parties argue that the Commission has not come close to discharging its onus to prove its dynamic competition argument. 594. This section must be read with our analysis of the competition counterfactual (see paragraphs 2822 to 3033 above). 635 Transcript p 3342 line 11 to p 3343 line 6. 636 Transcript p 3287 lines 10  18 637 Transcript p 3294 lines 3  17. Non-Confidential 162 595. The economic experts prior to the hearing agree on the following regarding FTTB: (i) Vodacom does not provide wholesale FTTB services to third party ISPs (neither dark nor lit); and (ii) the wholesale FTTB market has greater degrees of overbuild relative to FTTH, but they do not indicate what that degree is. We deal with our assessment of the FTTB overbuild evidence below. Our assessment 596. The assessment of a substantial prevention or lessening of competition under section 12A is measured on a case-by-case basis, taking into consideration the strategic evidence from documents, factual and economic evidence. Merger analysis is forward looking and involves an assessment of actual current as well as future competition, including any changes brought about by the merger to the market dynamics and (future) market structure. 597.              that cautions against a static view of the markets based only on current market shares, considering inter alia the evidence as we have traversed in relation to the true rationale for the proposed transaction, including that the proposed             documents plainly show, and the competition counterfactual (see paragraphs 282 to 3033). We consider the current market positions and how the proposed transaction will change or impact the market dynamics. Metro fibre market shares 598. Mr Hodge, using data on the record, estimates the market shares of total metro fibre backhaul network in kilometres for 2021. 638 Openserve is the largest with a market share of [30-40]%, followed by the merger parties at [20-30]%, MTN at 17% and Liquid Telecom at [10-20]%. We note that MTN is largely self-provision whereas Liquid Telecom is wholesale. 638 This excludes NLD but includes self-build by MNOs for their FTTS requirements along with own metro rings. As such, it is a perspective on capabilities not wholesale performance. Non-Confidential 163 599. A distinction between dark and lit fibre is appropriate. As noted in the market definition section, MNOs and FNOs prefer dark fibre as this is both cheaper but also enables them to have more control and flexibility over their network design. For these operators that require unmanaged services or dark fibre, the options are generally limited to DFA, Liquid Telecom and Link Africa. 600. Openserve does not sell or lease dark fibre access to FNOs, which removes it as an option for infrastructure players that seek access to dark fibre. As we have indicated, DFA is the largest provider of dark fibre connectivity and backhaul in South Africa with a market share of more than 80%. 601. DFA and Vumatel have the largest networks in terms of wholesale FTTB/FTTH and the largest dark metropolitan fibre backhaul incl. FTTS, ensuring the most densified networks for dark fibre and lit FTTB/FTTH nationally. Although Openserve has a larger total network, it does not supply dark fibre which is preferred by other infrastructure providers such as MNOs and FNOs, and it has fewer FTTH/FTTB premises passed. Other dark fibre providers have a considerably smaller metropolitan fibre footprint, especially outside of the major metros. This means that DFA may be the only option in many areas for MNOs/FNOs seeking new dark fibre backhaul and is most likely to have the closest proximity and lowest build distances for new connections. Vodacom is the largest MNO and has a sizeable position as both a retail ISP for FTTH and a retailer of FTTB through Vodacom Business. 602. Mr Nunes submits that the merged entity would have control over scarce upstream infrastructure that is essential to the provision of both wholesale FTTB and FTTH; 639          provide wholesale FTTB or FTTH services, even if they self-build metropolitan or last mile fibre infrastructure. 640 639 Nunes FWB p 163 para 6.41. 640 Nunes FWB p 166  167 para 6.51.3. Non-Confidential 164 603.          the market because of its dark fibre offering and its high-quality network,            presence in smaller towns and rural areas is where it has an advantage. 641 FTTB market shares 604. FTTB infrastructure is largely provided by metro fibre networks given these networks pass businesses as they seek to find backhaul customers which may be either MNOs or businesses. This includes the MNOs that have self-built metro fibre backhaul networks. 605. DFA launched as a dark-fibre-only infrastructure provider. However, it has started providing lit FTTB (in the form of its Business Broadband product 642 ) in addition to dark FTTB (in the form of its Helios and Lumic products). DFA has a growing lit FTTB presence and Vodacom has an existing and growing lit fibre presence. 643 606. Recall that the assets that are to be transferred as part of the purchase          wholesale business, including contracts (with ISPs) fixed assets and software. 607. Vodacom does not only sell FTTB on its own infrastructure but also sells it on the infrastructure of other parties. 644 608. Mr Van der Merwe of Frogfoot notes that data for FTTB (and FTTT/S) is not as readily available and the breakdown of market shares between these different technologies is less clear. 645 There was a dispute between Mr Hodge and Prof 641 Bundle M p 3498:  Discussion Materials dated September 2020. 642                 access basis to ISPs. A best effort FTTB product is a contended service as no specific bandwidth is guaranteed. 643 Nunes FWB p 169 para 6.55. 644 Theron Transcript p 3793 lines 15  17. 645 Van der Merwe FWB p 36 para 19. Non-Confidential 165 Theron specifically about the number of businesses passed and connected by Vodacom with FTTB. 609. We start by considering revenue-based market shares that gives a useful big picture of the relative size of some of the different operators in the FTTB market, although businesses passed (discussed below) provide a better indication of future competition since it is not based on sales performance but on the potential customer base. We concur with Mr Hodge that [businesses] passed must have weight because it shows the reach of your network and the potential to connect those businesses. 646 610. The revenue estimates are based on  total market size estimate and share estimates for FTTB fixed access services. BMIT indicates that revenue estimates are at wholesale prices. It is not clear whether any dark FTTB services of DFA are included in this revenue estimate and therefore the figures may understate the DFA position. However, BMIT estimate that FTTB services in terms of actual revenues are led by Telkom/Openserve ([20-30]%) and DFA ([20-30]%), who also lead metro fibre. The other metro fibre operators, Link Africa ([10-20]%) and Liquid Telecom ([0-10]%), each have less than half the share of either Openserve or DFA. MFN, which is more of an FNO, has a similar share to the smaller metro fibre operators ([10-20]%). Vodacom has a 5% share in terms of 2022 revenue. 647 611.             the record (submissions from the merger parties and third parties, excluding Liquid Telecom), is the single largest FTTB provider in South Africa with the merger parties having [60-70]% of businesses passed in wholesale FTTB (2021). 648 Vumatel provides limited wholesale FTTB services to ISPs with 646 Transcript p 3802 lines 7  9. 647 Hodge EWB p 77  78 para 93 Figure 26. 648 This is based on data provided by DFA to FTI for the merger filing, along with third party submissions. See Hodge EWB p 78 Figure 27 showing the estimated share of total business passed in wholesale FTTB (2021). Non-Confidential 166 approximately business customers connected via the Vumatel FTTB network. 649 612. According to the Commission, Vodacom represents [0-10]% of the abovementioned combined figure of [60-70]%, and Herotel has a [0-10]% market share (not included in the [60-70]%, and if included adds up to a [60-70]% share for the merger parties). We note that these estimates exclude data for Liquid Telecom, and therefore the actual shares will be overstated here. We further note that there is a difficulty with the data submitted by Liquid Telecom in that they were asked by the Commission to indicate the total number of business buildings passed in metropolitan areas, but responded that in metropolitan areas     (calculated using a 30m network line buffer) - thus not businesses passed. 650 613. According to the Commission,       basis, had passed over businesses 651 with 652 businesses connected, and it had km of metro backhaul in 2021 to support its operations and to resale to third parties on a lit basis. Mr Joosub indicates that Vo   relate to fibre access routes to approximately business office connections. 653 (Own emphasis) He does not provide a date for this figure; his witness statement is also not dated but was filed in April 2024. 614. The Commission finds that Openserve is a much smaller player in FTTB than Maziv with a market share of approximately [20-30]% (again, excluding Liquid Telecom figures and thus its market share is overstated here), and the rest of the market players, such as Link Africa, Octotel and Metrofibre, have small 649 Mare FWB p 436 para 25. 650 Bundle M p 6866. 651 Vodacom did not provide the number of buildings passed, but provide premises passed. However, data on overlaps provided by the merger        businesses in 2021. Hodge EWB p 79 para 94 Figure 27 footnotes. 652 Hodge EWB p 108 para 158. 653 Joosub FWB p 321 para 7.2. Non-Confidential 167 estimated market shares of [0-10]% and less. 654 Frogfoot and MTN have very small market shares. 655 615. Importantly,          in South Africa in terms of businesses passed. Vodacom in its Board Pack                     both businesses passed and connected, with Openserve as the next largest player with of businesses connected as DFA, and with as the only other key player listed with half the kilometres of FTTB fibre and a fraction of the businesses connected. Of the smaller players, Vodacom rate but no others  largely operates on dark fibre       % of South   656 616. Therefore, the South African FTTB market is already highly concentrated with two significant players, DFA and Openserve, (and Liquid Telecom as the third sizeable player) with Maziv being the largest player and that market position will increase as a result of the proposed transaction. 617. We note that the original FTTB market share data for both Vodacom and DFA, as relied on by Mr Hodge, 657 was done by DFA. 658 Prof Theron disputes the figures provided for Vodacom and the associated 8% market share for Vodacom, saying that the figures provided by DFA    659 She argues that a later submission by Vodacom on its own number of FTTB           combined market share in FTTB. She argues that the lower number of businesses passed for Vodacom accords with what is being transferred in the 654 Hodge EWB p 79 Figure 28. 655 This                will include duplicates as the same businesses may be passed by more than one metro and FTTB network. 656 Bundle M p 1061: Vodacom Board Presentation, Project (20 October 2021). 657        658 Transcript p 3794 lines 1  14. 659 Transcript p 3554 lines 4  17. Non-Confidential 168 transfer asset agreement and what Mr Joosub indicated. 660 Mr Hodge disagrees arguing that the original figures submitted   seem to be reliable because it was submitted and it seems extremely detailed and assessed in that way. 661 It is true that the original figures submitted are detailed. Mr Hodge notes if you look at figure 48, incredibly detailed. This is not some rough estimate, this is down to so many in two buildings passed. And, as I understand through this transaction, because of the transfer of assets DFA would, presumably, have information on the Vodacom FTTB assets if they were to agree on a valuation as part of the transaction 662 618. As indicated above, Mr Joosub (in his April 2024 witness statement) indicates that Vo    fibre access routes to approximately business office connections 663 Prof Theron says that Vodacom has less than wholesale FTTB infrastructure businesses connected in South Africa in 2021. 664 We cannot reconcile these figures. This suggests to us that there is no reason not to accept the original figures for 2021 submitted to the Commission of    ving passed ove buildings with businesses connected. 619. As noted above and indicated by Prof Theron, she has found figures for Liquid Telecom          market share calculations of businesses passed. 665 This is correct, but as indicated the data of Liquid Telecom do not relate to businesses passed and therefore is also unreliable. This means that Mr Hod        include Liquid Telecom, but we have indicated the difficulty with the data provided by Liquid Telecom         is understated. 660 Transcript p 3558 lines 12  14. 661 Transcript p 3557 lines 1  3. 662 Transcript p 3792 line 18 to p 3793 line 5. 663 Joosub FWB p 321 para 7.2. 664 EWB p 412 Appendix Table 15. 665 Transcript p 3554 lines 17  19. Non-Confidential 169 620. Prof Theron testifies that because there are different metrics used i.e. distances used when reporting on business passed, 666 it is quite difficult to put together the data for FTTB. 667 She indicates that the market share results will depend on how one measures businesses passed i.e., based on what distance from the network is used as the metric. 621. There are clearly some issues with the data that we have been provided with, but what does seem apparent is that DFA and Openserve are the two most substantial operators in wholesale FTTB, with Liquid Telecom as a third but smaller operator. This is consistent with the metro fibre network size and revenue estimates from BMIT. 622. We further note that the above market shares are for FTTB wholesale (both dark and lit), without any distinction between dark and lit. Recall that Openserve, included in the above figures, does not provide lit FTTB. Prof Theron indicates          FTTB market, that question has not been answered sufficiently in these RFIs for us to have market shares on FTTB for dark. So, for the           both dark and lit and the only issue is I hear what Mr Hodge is saying that Openserve is not providing dark. 668 623. It is common cause that Vodacom at the FTTB wholesale level does not allow access to its wholesale FTTB infrastructure. 669 It is also common cause that if Vodacom's FTTB infrastructure is transferred to Maziv, it will be available on an open access basis. We deal with this aspect under efficiencies, where Mr Reynolds concedes that this aspect will not pass muster under the efficiency test applied in merger cases. What is however important to note is that, although the Transfer Assets will become open access, it will enhance the market position of the largest incumbent in FTTB, Maziv. 666 One has to determine where fibre passes any business within a certain distance. Prof Theron indicates that sometimes it would be 500 metres; sometimes 50 metres; sometimes one kilometre. Transcript p 3794 lines 4  11. 667 Theron Transcript p 3793 lines 7  9. 668 Transcript p 3561 lines 2  15. 669 Theron Transcript p 3799 lines 10  12. Non-Confidential 170 624. That DFA is the largest player in FTTB is also confirmed by Mr Van der Merwe.            network is likely the largest. Mr Nunes indicates that DFA has near total control over upstream dark infrastructure at the national, metropolitan and FTTB levels. 670 625. Mr Van der Merwe also notes that an important differentiator here is lit and dark services. DFA provides dark and lit fibre whereas Openserve only provides lit fibre services. This has implications for the FNOs providing FTTB or wholesale FTTH, which require metropolitan connectivity, and who may require dark infrastructure rather than lit. 671 Dark is preferred over lit because with lit services the connection speeds are limited and the price of the connectivity increases with speed. This becomes cost prohibitive when lit is used as backhaul. Also, the lit backhaul customer then becomes reliant       network reliability and ability (or lack thereof) to upgrade their own network as and when additional capacity is required. 672 626.               competes for FTTB customers and can self-supply instead of using DFA. 673 627. It is common cause that compared to households, i.e., FTTH customers, businesses require a higher degree of flexibility and customisation. 674 FTTB customers are also of a larger scale than individual household customers and have a more extensive set of requirements (particularly large businesses) compared to households. National FTTB customers often require services in multiple locations and are unwilling to contract with different service providers in each location. For example, a large retailer with numerous different stores and branches across the country would require an FNO to be able to service at least most of those stores or branches. 675 Since a wide geographic footprint is a 670 Nunes FWB p 153 para 6.3.1. 671 Van der Merwe FWB p 34 para 16.4. 672 Van der Merwe FWB p 34 para 16.4. 673 Van der Merwe FWB p 52 para 56. 674 Nunes FWB p 168 para 6.51.5.3. 675 Hodge EWB p 81 para 97. Non-Confidential 171             higher for an FNO that wishes to have a prospective supplier of wholesale FTTB. 676 W            provide wholesale FTTB services. 677 628. Regarding FTTB overbuild, the merger parties rely on comments made by the small players in FTTB regarding the extent to which they overbuild other players (mainly Maziv and Openserve). 678 Interestingly, Prof Theron does not provide          leading position in the FTTB market it is likely that the small players will have to overbuild Maziv in order to compete in the market. These smaller players are at a distinct disadvantage given their market positions. For example, Link Africa notes that approximately % of its network is in areas where DFA also has a presence. 679 As we have indicated, the next player in FTTB in South Africa in terms of market share is Openserve, which itself is significantly smaller than Maziv. Openserve notes that only about % (based on info as at 2019) of its FTTH/FTTB network overbuilds another FNO. 680 This shows that overbuild even in FTTB is relatively limited from the perspective of one of the larger players. 629.                           exists. He submits that market shares on their own do not yield insights into how providers compete, which is at a more localised level. For instance, due to the              no (or limited) alternatives to DFA in such areas. 681 630. MTN submits to the Commission during its investigation that it is not aware that it has any FTTB infrastructure that currently lies alongside (i.e., overbuild) that 676 Nunes FWB p 167 para 6.51.4. 677 Nunes FWB p 166  167 para 5.51.3. 678 These players have estimated market shares of 5% and below. Frogfoot and MTN have very small market shares in FTTB. 679 Link Africaletter to the Commission dated 25 February 2022, Answer to Questions 129.1  129.2 and 131, Confidential Record of Third Parties, Part B of the Record p 2287. 680 Theron EWB p 377 para 289f. 681 Nunes FWB p 159 para 6.26. Non-Confidential 172 of a competitor. It notes that it builds fibre for FTTB based on a sell and build model. If overbuild occurred, it is because the customer ordered a service from MTN. 682                              683 However, as indicated, MTN is a small player in the FTTB market. 631. Mr Nunes further testifies in relation to the DFA footprint in general So, we need to look at it in terms of potential out there. Would you see major fall off? No,                               price and reduce or take the cost and in terms of being able to provide and             684       And if possible, do I have an alternative provider that can provide it to[o]? All right, that would be small and in between. 685 Non-static analysis 632.             inappropriate in the context of assessing the horizontal effects in this merger. A static competitive assessment is deeply unsatisfactory when a merger has a defensive motive (as we have noted Maziv has) and furthermore takes place in markets enjoying substantial and dynamic growth. An assessment of dynamic competitive effects is appropriate given inter alia that the FTTH/FTTB space is set for a second land grab with follow-through growth for the backhaul infrastructure, driven by 5G rollout also. Recall further that Vodacom states that the metro market is expected to show strong growth driven by the deployment of 5G and expansion of the access market. 686     682 Part B of the Record p 3557 para 89: MTN submission to the Commission dated 23 March 2022. 683 Transcript p 757 lines 3  6. 684 Transcript p 775 line 17 to p 776 line 5. 685 Nunes Transcript p 902 lines 15  16. 686 Bundle M p 3592. Non-Confidential 173 planned 5G expansion as per its submissions to ICASA when applying for additional spectrum (and further deal with it below under the public interest). 633. It is the dynamic competition that impacts on investment, pricing and innovation in the immediate future, and which will shape the future market structure and competitive dynamics more than the current structure. 634. Prof Theron asserts that the evidence relating to future increased competition from Vodacom in these markets is not robust enough for the Tribunal to consider. 687          documents created in the normal course of business is robust and in fact more probative than documents, such as factual witness statements, prepared specifically for this merger hearing. 635.                      other options, i.e., alternatives to the proposed merger. It will grow its fibre footprint in both metro and FTTB and growing to be an effective and even a           Therefore, we submit that assessing the impact of this merger based on the current, static market shares, is flawed and that the strategic evidence contained             the counterfactual and future competition dynamics absent the present deal. 636. The factual evidence has confirmed that Vodacom has a strategic imperative for infrastructure sharing and, in the counterfactual, it would look at JVs and partnerships to expand in fibre and compete with Maziv. The pro-competitive effects of Vodacom expanding without Maziv are acknowledged by CIVH in its assessment of the threats of not doing the deal, where CIVH identified the following in respect of Vodacom expanding in fibre absent the proposed transaction. This can be done or with strategic and 687 Transcript p 3345 line 6 to p 3347 line 11. Non-Confidential 174 636.1. CIVH faced        tenant, Vodacom, who is the largest MNO in the country, together with major revenue churn to a competitor FibreCo in which Vodacom will have a shareholding; (ii) the risk of losing R million in revenue per annum; 688 and (iii) in addition to the above, all future new business from Vodacom which would shift to its own FibreCo (see paragraph 126.3 above). 636.2. Vodacom will also not products and services, resulting in than currently forecasted for DFA. DFA will not only the previously mentioned, but also stands to lose future and from other . 689 636.3. CIVH would lose insights from having no     636.4.           implications that DFA would be as planned to match that rivalry and would be to quickly and efficiently capture new market share whilst mitigating downward pressures on pricing. 636.5.            mobile sites in buildings to also include FTTB (see paragraph 128 above). The clear implication was that FibreCo would ,         636.6. It would expand the FibreCo model to other markets on the continent (see paragraph 128 above). 688 Bundle M p 1223. 689 Bundle M p 1223. Non-Confidential 175 636.7. DFA would lose the opportunity to partner with Vodacom South Africa in order to enter (see paragraph 128 above). 636.8. The final scenario was competitors concluding similar deals to establish their FibreCo with to counter Lindt with the implication   . 690 637. The above illustrates the dynamic competitive landscape that will exist absent the proposed deal, and the competition lost from Vodacom if the proposed transaction were to be implemented. If the merger were to be implemented, all the above-mentioned pro-competitive benefits of future competition from Vodacom (and potentially others), seen by Maziv as a massive threat to its businesses with impacts on its pricing, will be lost. 638. It is apparent from this DFA/Vumatel assessment that a counterfactual where Vodacom establishes a TowerCo and FibreCo (with partners other than CIVH), is expected to result in a competitor to CIVH with the effect of not only lost business from Vodacom (a private loss), but also downward pricing pressure for all its services as FibreCo competes for share (a market wide social gain). That downward pressure resulting from lower pricing by FibreCo in an effort to secure market share will put downward pressure on DFA/Vumatel pricing, and presumably other market participants too. It is also likely to result in a more rapid rollout of infrastructure as competition intensifies for the land grab in FTTB/FTTH. 639. The above evidence is corroborated by the evidence of Mr Van der Merwe who confirms that Frogfoot would jump at the chance of rolling out behind Vodacom:                            690 Bundle M p 1223. Non-Confidential 176 all across the country. Now, were such an opportunity to arise, I think lots of FNOs would jump at the opportunity to create that. So, I think Vodacom here hold  is the gatekeeper. They hold the key 691 (Own emphasis). 640. Mr Van der Merwe explains that because Vodacom has so many base stations (as the largest MNO) that need to be connected, a business case truly exists to overbuild DFA or to give an offtake agreement to another FNO that results in an overbuild. 692 641. The above shows that the threat from increased competition from or facilitated by Vodacom to Maziv is not merely academic and can be implemented in practice. Having a large anchor tenant in the form of Vodacom, as the largest MNO in South Africa, with a large demand, will derisk associated investments for other competitor FNOs. This means that Vodacom is uniquely placed, given it size, to facilitate entry and/or expansion. The proposed merger eliminates this,        substantially lessens future competition. 642. Mr Van der Merwe further explains what the current dynamics are and how the proposed transaction will change those dynamics: Vodacom at present is a large counterweight to DFA both as being its largest customer and as a player which poses the threat of overbuilding DFA  importantly, whether Vodacom elects to do so or not. 693            pricing and service decisions. He indicates that pre-merger DFA is disciplined by Vodacom in that if DFA substantially increased its prices or deteriorates its services to Vodacom it would risk alienating its largest customer, Vodacom. 694 He points out that with the proposed merger this continuous threat disappears: In fact, the opposite then exists, because I [Maziv] have a ROFR with Vodacom, so basically as long as it suits Vodacom they will always give me [Maziv] 691 Transcript p 235 lines 11  16. 692 Transcript p 233 lines 9  15. 693 Transcript p 102 line 16 to p 103 line 14; p 235 lines 14  18. 694 Transcript p 103 lines 3  6. Non-Confidential 177 business and they would basically offset the business that I [Maziv] lose from other third party operators. 695 643. The ROFR, which derives from the Transfer Assets Agreement, is included in        . 696 644.             and FTTB compounds the merger-specific competition concerns and has important implications for future competition and market concentration. This ROFR is likely to give Maziv, as the largest incumbent in FTTB and a large player in metro fibre, a distinct advantage in expanding its market positions further in terms of all new fibre network build  importantly, during a future period of infrastructure growth. Maziv also identifies as part of the pros to the transaction, that the Vodacom products. 697 645. We note that the merger parties were not willing to divest the metro and FTTB Transfer Assets to remove any overlap between them. Prof Theron is questioned       why should you offer a divestment if             1% and in metro also very small percentages. So, why should you offer a divestment? 698         FTTB market share estimates for Vodacom appear understated. We do not find her argument persuasive, as we have shown, the Transfer Assets are significant in size and collectively are valued at approximately R4.2 billion. The more plausible               695 Transcript p 103 lines 10  14. 696    The requirements in clause 5.1.2. shall not preclude the Maziv Group from offering the Maziv Group, the Herotel Group or Vodacom SA Group prices that are lower than the standard rate card prices in instances where discounts are given in order to match a legitimate alternative competitor quote received by the Maziv Group, the Herotel Group or Vodacom SA Group for the same product or service, provided that Maziv Group notifies the Commission and the Monitoring Trustee of such discounted price and provides the Commission and the Monitoring Trustee with evidence of the alternative competitor quote that the discounted price seeks to meet  697 Part A of the Record, p 1267  1268: CIVH Board Pack Meeting dated 27 October 2021, Slide 45     698 Transcript p 4168 lines 3  21. Non-Confidential 178 did not want to see these assets move to a competitor(s) and strengthen their market position(s) and ability to compete with Maziv post-merger. The implications for Maziv not getting hold of the Transfer Assets are that DFA would be unable to expand its network footprint as planned to match rivalry and would be unable to quickly and efficiently capture new FTTB market share (see paragraph 127 above, and below        FTTB. 646. Whilst Vodacom does not wholesale its FTTB, it still competes for overlap                       demand and pricing. The merger will reduce this competition and provide scope for higher pricing by DFA in the future for the buildings passed where the parties overlap ( buildings according to Hodge, based on Vodacom total buildings passed and DFA total businesses passed ). 699 As we have             metro and FTTB market positions. 647. As we have further indicated, there is correlation between metro fibre and FTTB rollout. FTTB infrastructure is largely provided by metro fibre networks given these networks pass businesses as they seek to find backhaul customers which may be either MNOs or businesses. In the counterfactual absent the proposed transaction, Vodacom will grow both its metro fibre and FTTB offering which will enhance future competition 700 . 648. Mr Van der Merwe speaks to the implications for Maziv of future competition from Vodacom. Future competition from or facilitated by Vodacom would result in a            701             result of increased competition from a is 699 Hodge EWB p 110  111 para 165 Figure 48. 700 See Hodge EWB p 48 para 26. Bundle M p 1256  1257: CIVH Board Pack of 27 October 2021 (Part A of the Record p.1267/8). 701 Transcript p 102 line 21 to p 103 line 6; p 235 line 18 to p 236 line 9. Non-Confidential 179  one of the providers of choice to the industry, due to the emergence of preferred new competitors that now have a shareholding 702 649. As indicated, the evidence is that the FTTB market is dominated by DFA, with more than         FTTB network will add to its already dominant size giving it even more scale. We have also dealt with the implications of the ROFR for future competition and concentration. 650.             DFA would be unable to quickly capture new market share without causing downward pricing pressure. As Vodacom extends its fibre to support rooftop and in-building mobile sites, selling FTTB in those buildings at incremental cost, this           651. In summary, the proposed transaction changes the market dynamics: (i) there no longer is the pre-merger looming and continuous threat which disciplines DFA (in its pricing and other decisions) as there is no fear of alienating and losing its biggest customer, Vodacom, compounded by the ROFR (included in the tendered conditions); and (ii) if the Vodacom constraint disappears, DFA will be able to profitably increase prices. Increased prices will mean that costs for FNOs (and other purchasers of metro connectivity like MNOs) will rise, which will in turn feed through to ISPs and end-consumers. 652. Vodacom, due to its size and its commercial significance, is uniquely placed to credibly cause or facilitate increased competition for Maziv. Indeed, Vodacom sees the current market conditions as an important point for it to accelerate its fibre presence through both a TowerCo and FibreCo strategy. It using its R6bn cash from the CIVH proposed transaction to invest in competition, along with its significant fibre assets, (i) represents a significant threat that disciplines Maziv; and (ii) would change the market dynamics - DFA would lose its largest anchor 702 Bundle M p 1223. Non-Confidential 180              business. 653. Indeed, Vodacom pursuing other options and the prospect of more competitive pressure on Maziv in the alternative played a large role in it agreeing a transaction, and on terms that favour Vodacom from a co-control and asset valuation basis. The very real threat of Vodacom expanding and competing more aggressively with Maziv in metro fibre and FTTB is confirmed by the strategic documents. Based on these strategic documents absent the proposed transaction, Vodacom would have expanded independently of Maziv, prompting Maziv to ultimately agree to this transaction to avoid dynamic future competition. 654. The proposed deal changes the future dynamics, since Vodacom would establish a TowerCo, as it has already done, but that TowerCo would source            resulting in a loss to Maziv of existing and future Vodacom revenue. Vodacom TowerCo will then be positioned to provide an integrated wholesale mobile backhaul offer to other MNOs, resulting in DFA losing existing and future revenue from other MNOs. This would clearly increase competition for mobile backhaul and bring competitive pressure on DFA pricing for FTTS. 655. The land grab nature of fibre competition means that stronger dynamic competition results in inter alia more innovation and lower pricing. In contrast, the proposed transaction will reinforce and strengthen market concentration, with negative implications for the long run structure and competition to the detriment of consumers. Conclusion 656. The markets concerned, which are expected to grow rapidly, compel us to take a non-static approach and consider future dynamic competition. For all the above reasons, we conclude that the proposed transaction, which eliminates Vodacom as a future competitor, will substantially lessen future dynamic Non-Confidential 181 competition in metro fibre and FTTB to the ultimate detriment of South African consumers. PORTFOLIO EFFECTS 657. The Commission and MTN contend that the proposed transaction is likely to result in a substantial prevention or lessening of competition through portfolio effects, specifically bundling. 658. The Commission submits that bundling can take the form of offering fixed and                           complementary services that make use of connectivity (e.g., streaming and security) and equipment to support the connectivity. 703     considered an attractive proposition for operators to increase ARPUs and helps reduce churn rates to competitors as they are able to offer the customer all the        February 2023 strategy documents where the benefits of fixed-     owing the to of services, and through  704 659. MTNs strategy documents of December 2021 show that it is concerned about the Vodacom/Maziv deal that business. VC selling and into an existing customer base will   and  705 MTN convened a board meeting to respond to the threats posed by the Vodacom/Maziv deal and a resolution was taken that MTN should consider or to avert the threat of Vodacom/Maziv bundles. 706 703 Hodge EWB p 124 para 192. 704 Bundle O p 231. 705 Bundle O p 2  MTN SA - FTTX Way Forward Role of Fibre and Options to Consider dated December 2021. 706 Commission HOA p 126 para 303. Non-Confidential 182 660. The merger parties through Mr Scheffer of Vodacom argues that pre-merger, Vodacom can offer bundling and has tried to market fixed and mobile bundles in the past but has been unsuccessful. 707 In particular, Mr Scheffer states that Vodacom started offering a bundle consisting of a Vodacom fibre ISP product and a mobile product in 2018 and later tried a mobile and FWA bundle    able to achieve a take-up of sales on the fibre-mobile bundle and sales on the FWA-mobile bundles as at October 2023. 708 Mr Scheffer indicates that after investigating the reasons for the failure, Vodacom decided to discontinue bundles. 709 661. Mr Reynolds rejects the bunding theory of harm on the basis that (i) the majority of customers purchase the products individually; (ii) bundles are already offered in the market without market impact; (iii) international experience show low take-up of bundled offers; and (iv) customers in low-income areas where fibre deployment will be happening are unlikely to sign long-term contracts. 710 Our assessment 662. We first consider what bundles the merger parties could offer after the proposed merger since V                   operations, offering: 662.1. Maziv (as an FNO) will post merger have      subscriber base, which is the largest in the country, as well as its FWA capability; and 662.2. Vodacom (as an FNO/ISP) with a substantial customer base in retail connectivity  will have                     711 Given our above discussion of Herotel, the 707 Scheffer FWB p 246 para 59. 708 Scheffer FWB p 246 para 59. 709 Scheffer FWB p 246  248 paras 60 to 65. 710 Transcript p 3672 line 20 to p 3676 line 20. 711 Nunes FWB p 165  168 para 6.49 to 6.51; Exhibit BP Slide 56; Smith Transcript p 3679 line 1 to p 3681 line 22. Non-Confidential 183 same           strategic information. 663. We next consider the post-merger    contends is irreplicable. 664. MTN submits that post-merger the merger parties could, before rolling out fibre        (largest) mobile retail subscriber base and data 712 as a springboard off which to capture fibre customers by offering residents a bundled combination of mobile and fibre products and services. 713 665. MTN gives the following example of how in practice the merged entity will market a bundle            Vumatel would share with Vodacom its rollout plan for the area to afford the merged entity a decisive first-mover advantage, whilst Vodacom would share with Vumatel its area-specific subscriber information to enable focused and effective marketing. Vodacom ISP would then offer to Vodacom subscribers and other residents of the area a bundled internet access product/service comprising: (i) Vodacom will immediately provide FWA connectivity at discretionary (or potentially non-discriminatory) pricing; 714 (ii) Vumatel will in due course provide FTTH connectivity at discretionary (or potentially non-discriminatory) pricing; 715 and (iii) in relation to both FWA and FTTH connectivity, ISP services to be provided by 712 Smith EWB p 256 and 257. 713 Motlekar FWB p 12 para 25; Van der Merwe FWB p 51  52 paras 54 to 55; p 56 para 73; Nunes FWB p 165 paras 6.49 and 6.50; p 169  171 paras 6.56 to 6.64; Smith EWB p 278 para 251; Van der Merwe Transcript p 103 line 14  p 104 line 8; Masalesa Transcript p 329 line 15  p 332 line 12; Exhibit BP Smith Slide 55. 714 Absent any FWA pricing condition, Vodacom could discount its ordinary price for FWA to entice the retail customer to buy the bundled product or service, bearing in mind that Vodacom would share in 30-      subject to a non-discrimination condition, Vodacom may have to offer the same FWA price to both wholesale and retail customers but would still enjoy its share of the FTTH profits. 715 Absent any FTTH pricing condition, Vumatel could discount its ordinary price for FTTH to entice the retail customer to buy the bundled product or service, bearing in mind that a controlling shareholder, Vodacom, would share in 30-     subject to a non-discrimination condition, Vumatel may have to offer the same FTTH price to both wholesale and retail customers while Vodacom would still enjoy its share of the FTTH profits. Non-Confidential 184 Vodacom ISP at discretionary pricing. 716 Once the fibre infrastructure was rolled out in the area (which would be open access in the case of Vumatel) the merged entity could increase its FTTH and ISP pricing  to FNOs, ISPs and/or retail customers  without there being any real risk of over-building by a third party. 717 666. MTN argues that it and Supersonic would not be able to replicate or rival the Vumatel bundle for the following reasons: 718 (i) MTN/Supersonic would not have      (ii) MTN/Supersonic would not have access to      (iii) at best MTN/Supersonic could offer to MTN subscribers and other residents of the area a bundled internet access product/service comprising: (a) FWA connectivity to be provided immediately by MTN at discretionary pricing. MTN submits that although it could theoretically discount its ordinary price for FWA to entice the retail customer to buy the bundled product or service, the rationality of doing so would be undermined by              ; (b) FTTH connectivity to be procured in due course from Vumatel or other FNO with network infrastructure in the area (if any), at discretionary (or potentially non- discriminatory) pricing, the cost of which would have to be borne by MTN qua reseller or passed on to the customer. MTN again submits that the rationality of doing so would be undermined by the fact that MTN would not share in any percentage of the FTTH profits; and (c) in relation to both FWA and FTTH connectivity, ISP services to be provided by Supersonic at heavily discounted pricing to offset the cost referred to in (b) above. MTN submits that the rationality of doing so would be undermined by the fact that Supersonic  already subject to low margins characteristic of the ISP market segment  would not share in any percentage of the FTTH profits. Once the merged entity rolled out fibre infrastructure in the area, MTN/Supersonic would have to bear or pass on any 716 Absent any ISP pricing condition, Vodacom ISP could discount its ordinary price for ISP to entice the retail customer to buy the bundled product or service, bearing in mind that Vodacom would share in 30-     717 Transcript p 630 line 20 to p 633 line 17; Exhibit BP Slide 56; Smith Transcript p 3332 lines 7  9. 718 Nunes FWB p 169 to 171 paras 6.56 to 6.64. Non-Confidential 185 increased FTTH pricing without there being any reasonable prospect of over- building by a third party. 719 667. MTN further argues that no other FNO or ISP (e.g. Frogfoot and Vox) would be able to replicate or rival the Vumatel bundle because: 720 (i) the FNO/ISP would not       (ii) the FNO/ISP would not have access to      (iii) at best for the FNO/ISP, it could offer to residents of the area a bundled internet access product/service comprising: (a) FWA connectivity to be procured immediately from Vodacom or other MNO with network coverage in the area (if any), at discretionary (or potentially non-discriminatory) pricing, the cost of which would have to be borne by the FNO qua reseller or passed on to the customer. MTN submits that the rationality of doing so would be undermined by the fact that the FNO would not share in any percentage of the FTTH profits; (b) FTTH connectivity to be procured in due course from Vumatel or other FNO with network infrastructure in the area (if any), at discretionary (or potentially non-discriminatory) pricing, the cost of which would have to be borne by the FNO qua reseller or passed on to the customer. MTN submits that the rationality of doing so would be undermined by the fact that the FNO would not share in any percentage of the FTTH profits; and (c) in relation to both FWA and FTTH connectivity, ISP services to be provided by the ISP at heavily discounted pricing to offset the costs referred to in (a) and (b) above. MTN submits that the rationality of doing so would be undermined by the fact that the ISP  already subject to low margins characteristic of the ISP market segment  would not share in any percentage of the FTTH profits. Once the merged entity rolled out fibre infrastructure in the area, the FNO/ISP would have to bear or pass on any increased FTTH pricing without there being any reasonable prospect of over-building by a third party. 721 668. As indicated above, the merger parties argue that past bunding efforts have not been successful. Curiously however, despite  in the past with bundling, Vodacom has this strategy, and it 719 Transcript p 633 line 18 to p 635 line 6; Exhibit BP Slide 56. 720 Nunes FWB p 169  171 para 6.56  6.64. 721 Transcript p 635 line 7 to p 636 line 10; Exhibit BP Slide 56. Non-Confidential 186 as detailed in a recent document dated August 2023. 722 Importantly however past bundling does not take into account the merger-specific advantages and bundling opportunities as articulated above by MTN. Bundling remains a potential strategy in the long-term plans for Vodacom and CIVH, but also MTN. 669. As explained above, the proposed merger brings about opportunities for consolidation, access by Maziv  large subscriber base, and unique opportunities to discount            proposed transaction. An important aspect of any successful bundling strategy would be scale, which the proposed merger provides to (i) Maziv in terms of FWA/mobile; and (ii) Vodacom in terms of FTTH and FTTB. In other words, Vodacom will have access to the scaled FTTH and FTTB network of Maziv, and         a conducive partnership for bundling in the context where Maziv has already considered for its business to improve its overall offer. 723 670. Therefore, although past bundling may not have been as unsuccessful, and that customers tend to buy the relevant products separately pre-merger, the scale advantages that this merger affords the merged entity will enable such a bundling strategy. Mr Hodge notes you've got a which is August 2023, now a year old. And over on the next page an MTN plan B update. This is in relation to responding to this transaction, but this is February 2023. So, despite the apparent lack of success, this of Vodacom, and it remains part of the ambition of MTN as well. And I would probably say there's a good reason for this is because if you look at the bundle, this is going to be targeted at your wealthier consumer because your starting price, if you have bundled the many products together is going to be a more expensive monthly offering. And if you look at the Vodacom approach, it's also looking to cross sell, upsell a whole range of other value-added offerings in the market where it's bound to be a wealthier customer. [...]. if you get it right and your competitor doesn't, your main 722 Bundle M p 12863. 723 Exhibit BW p 24: CIVH Group Strategy Session dated 2 February 2022 Non-Confidential 187 competitor here being MTN who has high value customers, well then you could weaken them and that is precisely the concern that MTN has when it hears about this transaction . 724 Conclusion 671.      reveal that bundling post-merger and the proposed merger would, due to its size and other advantages from the combination, enable them to execute this strategy. Successful bundling as a result of the merger would further entrench the dominance of Maziv in fibre and     mobile. Therefore, post-merger bundling is a merger-specific concern. VERTICAL EFFECTS Control and the incentive to foreclose 672. Prof Theron argues that the structure of the proposed transaction does not create an incentive to foreclose, and that Maziv will not have an incentive to preference Vodacom. 725 She premises this argument on the assertion that Maziv will not own a share of Vodacom and will thus have no profit incentive to align its interests with those of Vodacom at its expense. 673. In our view this is to take a relatively binary view of the theory regarding incentives, how they change through major transactions such as this one, and the strategic and commercial weight of the affiliation and alignment that a merger of this nature creates between firms. 674. We have dealt with post-merger incentives above considering inter alia the            documents in which Maziv and its shareholders asserted that they did not see 724 Transcript p 3669 lines 11  22, p 3670 line 1, p 3670 lines 21  22, p 3671 lines 1  2. 725 Theron EWB p 366  368. Non-Confidential 188 an alignment with the wider Vodacom business. To the contrary, the evidence we have been presented shows that Maziv perceives significant opportunities through an alignment with Vodacom, no doubt evidenced and shaped by ongoing and previous interactions be     and the Vodacom group at large including in relation to partnerships in South Africa as well as other countries in Africa. 675. If the above holds such that Mr Smith, Mr Hodge and Mr Johnson are correct regarding the association and incentives created structurally by the merger, then               majority of her arguments in relation to vertical foreclosure must fail. It seems to us that the theory does not posit a binary lens in which a firm either has absolute control at 50% and beyond with fully aligned incentives, or none at all. This merger is also not one in which anyone could credibly argue or accept that the significant minority shareholding being acquired by Vodacom, which we note is associated with significant, relatively uncommon rights beyond those typical of a minority shareholder, cannot attribute some forms of preference, alignment or indeed economic incentives to favour Vodacom and its interest in Maziv board decision-making and strategy over those of others. One only needs to consider a scenario in which a proposition is being considered at board level in which a particular innovation or investment by Maziv would benefit it, but could in all likelihood be expected to, in some way, erode the competitive position of Vodacom. As a Maziv board member, Vodacom would surely resist such a position. An example mig          license or spectrum in a manner that might place it in competition with Vodacom or sought to invest in a firm that competes with Vodacom. 676. As we have indicated in the section above dealing with incentives, in this transaction Vodacom is far from being a passive shareholder in the general sense and indeed has rights which would allow it to at the very least shape key decisions of Maziv over time. In a dynamic and evolving market this is a significant risk that we perceive to the extent that it might lead to various anti- competitive outcomes or anti-innovation outcomes over time. Prof Theron Non-Confidential 189 acknowledges in her report that there is at least minority controlling interest being acquired. MTN argues that indeed significant influence is likely and it would be contrary to the facts of this case to ignore the possibility and import of such influence. 677. While we accept the argument that the transaction does not present a full integration of the mobile and downstream business of Vodacom with that of Maziv, we find that the position of minority shareholding, and the strategic documents of CIVH and Vodacom support a view that there is likely to be an alignment at a strategic level between Maziv and Vodacom and their commercial interests. This is precisely at the heart of the concerns raised by MTN and Rain because of the likely alignment of interest of the large fibre infrastructure provider capabilities of DFA and FNO position of Vumatel, to the detriment of      Maziv may not hold an interest in Vodacom, but it has a vested interest in its success and Vodacom exerts co-control over Maziv 726 678. In the context of this case one should do a comprehensive assessment of the ways in which the firms might choose to associate or align their strategies, noting incentives are not binary and certainly cannot only be assessed as being about differences, in this case, in upstream wholesale margins which are said to be larger than downstream retail margins. This would be to obscure the full picture that has emerged from our inquisitorial process and in particular our assessment of submissions by witnesses as well as particularly the strategic documents of the merger parties. 679. Ultimately, the documentary and other evidence suggests that Maziv does not see itself as simply receiving a financial injection from Vodacom akin to that which would be made by an investment bank with unrelated business activities, a general shareholder in a large, listed firm, or a minority rights shareholder with limited interest or powers. Our analysis of the vertical theories of harm departs from this understanding. 726 Hodge EWB p 134 para 211. Non-Confidential 190 680. In what follows, we evaluate the three vertical theories of harm based on the vertical overlaps arising as put forward by the Commission, namely: i) foreclosure of metro backhaul to MNOs; ii) foreclosure of metro backhaul and dark FTTB used by FNOs; and iii) foreclosure of access to wholesale FTTH/B used by retail ISPs and for business. VERTICAL OVERLAP 1: FTTS AND FORECLOSURE OF METRO BACKHAUL TO MNOS 681. We begin with an assessment of the theory of harm put forward by the Commission (and Mr Hodge), MTN and Rain that the merged entity would have the ability and incentive to foreclose MNO rivals of Vodacom of access to FTTS infrastructure. This theory of harm arises due to the vertical overlap that the merger would create in the provision (upstream) of metro dark fibre on the one hand, and retail mobile products and services typically offered by MNOs downstream on the other. 682. The             upstream market for the supply of dark fibre access to various MNOs including              infrastructure to facilitate interconnections between cell towers to support the provision of primary mobile communications services (mobile voice and data, and FWA etc.) to their customers. In particular, DFA provides its dark fibre products, primarily Titan, to MNOs such as Vodacom, Rain and MTN, that procure Titan links to provide mobile telephony tower connections referred to as FTTS. It is the alleged dependence of MNOs on DFA for Titan access primarily that gives rise to the claim that the merged entity would have the ability to foreclose. 683. Titan is categorised as wholesale metropolitan fibre backhaul. It is a dark fibre product, meaning that the product does not transmit data. Customers need to            727 727 Mare FWB p 434 para 16. Non-Confidential 191 684. In terms of the incentive to foreclose, the concerns of rival MNOs and the Commission can be distilled as follows. Pre-merger, DFA has no incentive to favour Vodacom over other MNOs in the provision of access to Titan and other dark fibre infrastructure. The proposed merger alters this position. Post-merger an MNO will have a significant shareholding and co-control in Maziv, and DFA will have an incentive to advantage Vodacom over its rivals through price and non-price mechanisms. 685. The concern therefore is that the merger leads to a change in incentives such that Vodacom may cause DFA to engage in input foreclosure of MNO competitors primarily through the provision of services to Vodacom on preferential terms. 728 The alignment of interests and operations of DFA as a fibre infrastructure provider and Vodacom as an MNO creates a risk of substantial price and non-price discrimination or self-preferencing in favour of Vodacom. 729 686. In terms of non-price factors, the concern relates to preferencing such as on the timing of fibre network rollout, coverage, capacity, speed and quality, or inferior terms in the repair of faults. 730 It is also alleged that DFA would have the incentive to prioritise Vodacom in terms of the rolling out of fibre infrastructure to areas preferred or required by Vodacom rather than where its competitors need infrastructure. 731           are important for rivals in favour of those relied upon by the merger parties or lit products, for example. 687. In terms of prices, Mr Smith and the Commission raise that the proposed merger          (Vodacom) switching away. In terms of prices charged by Maziv to MNOs post-merger, Mr              728 Rain HOA p 5 para 8.1. 729 MTN HOA p 65 para 82. 730 MTN HOA p 65 para 82; Rain HOA p 7 para 14. 731 Rain HOA p 8 para 14.4. Non-Confidential 192 switching away is likely to incentivise Maziv to increase the prices it charges for Titan to MNOs. While Maziv may lose some Titan profits in the event that rival MNOs reduce their demand for Titan, the proposed transaction, and the resulting             wide) price increases more attractive and more likely. This is for two key reasons: 687.1. Maziv is likely to perceive a reduced risk of Vodacom  its largest customer and substantial shareholder  switching away. Vodacom is the    product. A reduced risk of its largest customer switching away (or self-building) is likely to result in Maziv becoming more confident to increase prices. His Figure 14 shows the           According to the figure, Vodacom accounts for     revenues, with Cell C, Rain and MTN accounting for 3%, % and % of     687.2. Moreover  -built mobile backhaul (which currently constitutes % of its mobile backhaul) would be transferred to DFA post-            around %                     to increase prices to MNOs (even on a market wide basis). 688. Rain 732 and MTN argue that a partial rather than full input foreclosure strategy may arise as a result of the merger, with significant anti-competitive effects on (competition between) MNOs. It is unlikely that the merger parties would seek to totally foreclose third-party MNO customers due to the revenues derived from               follows regarding the theories of harm presented by the Commission: As Maziv operates open access networks, earning revenue from a broad set of customers, foreclosure is less likely to take the form of a 732 Rain HOA p 8 para 15.3. Non-Confidential 193 complete refusal to supply as against all customers, and more likely to take the form of preferential treatment of Vodacom as MNO / ISP and its own expanded network and selective refusal to supply. One means would be through prioritisation of Vodacom (MNO and ISP) requirements, enabling them to secure new opportunities first, and delaying services requested by rivals where they overlap. Preferential pricing is another means, using discounts notionally justified by volumes or bespoke arrangements where rivals would not even be aware. Selective refusal to supply could be targeted at a single customer, such as MTN as the closest rival, or at locations where the parties seek a first- mover advantage, such as the land grab or 5G coverage. Lastly there is a foreclosure to dark fibre involving a shift to lit fibre. 733 689. Mr Hodge accepts 734 that differential pricing is unlikely to affect competition downstream given that the costs of FTTS metro fibre access constitute a small proportion of total MNO costs of provision. 690. In relation to this theory of harm, both MTN and Rain are unequivocal in their assertions that if the merger was approved without effective remedies it is likely to result in a substantial prevention or lessening of competition. These parties believe that the concerns can be remedied through the conditions proposed by the merger parties, while the Commission believes that the concerns cannot be addressed by the conditions. 691. Lastly, an additional concern arises in relation to information sharing and the business plans of Maziv, and the ability of Vodacom to view information of its competitors and to shape investment and strategy decisions of Maziv. This is related to the information exchange concern expressed by inter alia Rain, in arguing that Vodacom would have access to its commercially sensitive information. 735 This is on the basis that Rain and other MNOs share highly sensitive information with DFA in the ordinary course, including network rollout 733 Hodge EWB p 133  134 para 211. 734 Transcript p 4047 lines 4  5. 735 Rain HOA p 9 para 16. Non-Confidential 194 plans six to eight months in advance, so that DFA can set up requisite infrastructure. In the hands of a competitor, Vodacom, such information could be used by it to gain a first-mover advantage by offering services in those areas first. 692. Relatedly, Mr Hodge contends that Vodacom would need to share information            business plan. The nature of the information would be such that Vodacom could     plan, as a member of its Board, the requirements              accordingly. 693. The merger parties do not strongly contest the concern regarding the potential flow of competitively sensitive information post-merger. They hold that the proposed conditions would be sufficient to address the concern. We agree that such a concern exists and that it is significant from a competition perspective. As it is not contested, we do not deal with it in the analysis to follow although we will consider the matter in our evaluation of the proposed conditions further below. Ability to foreclose rival MNOs Change in market structure 694. As we have indicated, DFA is dominant in the upstream market for the supply of dark fibre, with approximately [80-90]% national dark fibre market share (by 2021 revenues) in the provision of metro dark fibre. 736 Only three players offer dark fibre on an open access basis (DFA, Link Africa with approximately [0-10]% market share, and Liquid Telecom with approximately [0-10]% share). We have above found that dark and lit fibre constitute separate relevant product markets and note that MNOs rely predominantly on dark fibre infrastructure access for FTTS connectivity. 736 MTN HOA p 64 para 81; Rain HOA p 6 para 11.1; Commission Report p 431. Non-Confidential 195 695. Prof Theron states, based on an internal document of Maziv, that the shares of Maziv for the provision of mobile backhaul/FTTS are only approximately [30- 40]              approximately [0-10]% based on the number of links. 737 In this estimation, Telkom is said to have the largest share of the market at around 59%. 738 However, we note that Prof Theron recognises that in the narrower, dark fibre-            of [80-90]%, with her main argument being that the merger does not lead to a significant change in this market and that regard must be had to any change in incentive post-merger. 739          throughout. 696. Prof Theron explained in the hearing that Openserve (affiliated with Telkom) has a larger network but does not provide dark fibre access. 740 However, her            access, given there are no technical constraints as confirmed by Telkom/Openserve 741 . We identified concerns with this approach. 697. We understand that while Openserve has a large dark fibre network measured in links or kilometres, in reality its network is not presently available to other MNOs for mobile backhaul because commercially    business model to offer lit rather than dark fibre access to different third-party              698. Openserve does not extensively provide dark fibre for mobile backhaul access, at least not to some other MNOs competing with Telkom downstream (for example, MTN does not procure a significant proportion of dark fibre Titan or 737 Theron EWB p 393 para 367. 738 BMIT estimate, Theron EWB p 394 para 367. 739 Transcript p 3774 lines 1  9. 740      741 Part B of the Record p 5324 para 14.2. Non-Confidential 196 Titan-like products from Openserve 742 ). 743 The fact that Vodacom has some dark fibre infrastructure but has chosen for its own commercial reasons not to provide access to this infrastructure to third parties (other than in limited ) 744 ,           745 for this access as cited by Prof Theron, is analogous. 699.         of [80-90]% in the provision of metro fibre used by MNOs. 746 This is because Vodacom does not provide dark fibre access in the market, as noted above. However, we note, as accepted by the various experts, that a merger need not give rise to a change in both the ability and incentive to foreclose, and according to the theory it is sufficient for assessment that a merger may only change the incentive of the merger parties to foreclose rivals if it would be profitable for them to do so. We return to the issue of profitability and incentive below. 700.                 of the merged entity to foreclose downstream rivals of Vodacom. In this regard, we consider various factors in drawing our conclusion that DFA has the ability to foreclose MNO rivals of Vodacom. Overbuilding or self-build not viable 701. The evidence before us points to the fact that this infrastructure is relatively difficult to replicate, and that MNOs have few alternatives that provide FTTS particularly in critical high density metro markets. Mr Otty for Vodacom explained that leasing fibre from third-party providers where it is available is affordable because the infrastructure is shared by multiple parties. 747 In this regard, witnesses for both MTN and Vodacom confirmed that for MNOs, self-building 742 Hodge EWB p 135 Figure 61. See also Nunes Transcript p 636 lines 5  8. 743 EWB p 53 para 36.3. 744 Exhibit BQ  lide 14. 745 Exhibit BQ  lide 14, citing Van der Merwe Transcript p 77 line 21. 746 Theron and Smith agree in the Joint Expert Minute (para 3.4.1) that they are not aware of Vodacom routinely providing dark fibre services for metropolitan connectivity or mobile backhaul to third parties. 747 Otty FWB p 360 para 18. Non-Confidential 197 the infrastructure is approximately [%] 748     749 the cost of leasing from DFA. It is thus generally accepted to be uneconomical to self-build when the option to lease the infrastructure through long-term agreements is available, which Mr Reynolds admits in the case of Vodacom. 750 MNOs have varying but significant degrees of dependence on DFA 702. To           infrastructure either in terms of reliance on DFA altogether 751 , in key areas where alternative infrastructure is not available (where DFA is a monopolist) 752 , and given the fact that overbuild or self-provision is considered to be largely not commercially viable relative to the significant investments that DFA has already made in dark fibre infrastructure throughout South Africa 753 . 703.             %) on DFA for the provision of dark fibre services. 754 It argues that DFA is so dominant                 an alternative service provider or to self-      network has been built with a high level of integration and    infrastructure, and the evidence is that it would be difficult for Rain to uncouple itself from DFA. This evidence is not seriously contested, and it is acknowledged            the only MNO truly reliant on DFA for its mobile backhaul needs 755 No further evidence was led to refute that Rain is likely to be harmed were it to be foreclosed access to dark fibre for FTTS and this concern alone weighs heavily in our assessment. 748 Mdlalose FWB p 392 para 29. 749 Smith EWB p 270 para 224; Hodge EWB p 137 Figure 62. 750 Exhibit BR1: s Slide 6; Transcript p 3566 lines 1  12. 751 Rain HOA p 6. 752 MTN HOA p 64 para 81. 753 MTN HOA p 64 para 81. 754 Rain HOA p 6 para 12. 755 Theron EWB p 396 para 378. Non-Confidential 198 704.                It relies on DFA for approximately % of its sites, whereas approximately % of its sites rely on self-built connections. We understand that MTN has a higher proportion of self-built infrastructure due to legacy infrastructure it deployed before DFA was well-established. 756 Presently, it uses a combination of leasing and self-build depending on costs and availability of existing infrastructure. 705. It is worth noting that Cell C until 2023 was also 100% dependent on DFA for Titan links, that is until its strategic decision to sell its site infrastructure and to transition to providing mobile services solely through roaming agreements in future and thus to cease procuring links directly from DFA. 757 That is, at least at the time that the proposed transaction was contemplated in 2021, Cell C could have been understood to be dependent on DFA just as with Rain. It is not clear              be reliant on the existing DFA links and arrangements for a large proportion of its network, although this appears to be likely at least in the short- to medium-                As such, we cannot dismiss this dependency linked to the existing network infrastructure that was being used by Cell C. 706.           -like products for mobile backhaul is significant. Only % of its mobile backhaul infrastructure is self- built, compared to % that is leased from DFA and % that is leased from other providers. 758 707. There can be no dispute that % dependence on DFA Titan links (as with Rain, and previously with Cell C) establishes the existence of an ability to foreclose these firms. Vodacom is dependent on DFA pre-merger for most of the links that it uses. 756 Smith EWB p 271 para 227; Nunes FWB p 158 para 6.22. 757 Hodge EWB p 138 para 225. 758 Smith EWB p 271 Table 1. Maduray FWB p 417 Table 4. Nunes FWB p 182 annexure RN1. Non-Confidential 199 708. MTN has no alternative to DFA in numerous areas. 759 In instances where those Titan links relate to a particular geographic area where no other providers are present, then some interruption, decrease in quality or partial foreclosure of access can be significantly detrimental to MTN, and perhaps most importantly its retail consumers in those areas. 709. Importantly           -built, its annual operational expenditure on third-party providers for FTTS is dominated by DFA at approximately % which is significant. Notably, MTN relies on third- party provision of these links to a non-negligible degree in all main regions of the country (for example, approximately % in Johannesburg and Tshwane in Gauteng, and approximately % in KwaZulu-Natal and the Western Cape). 760 Indeed, MTN believes that it is very reliant on DFA 761 , and no evidence was presented by the merging parties to dismiss the substantive dependence of MTN on DFA even as it has a high proportion of self-built infrastructure. 710. Importantly, any future new MNO entrant in the market is likely to be reliant on DFA given the fact that overbuild or self-provision is considered to be largely not commercially viable. Access to DFA aggregation nodes is restricted 711. High dependency on DFA infrastructure is exacerbated by the manner in which DFA governs access to its aggregation nodes. Aggregation nodes are the physical connection points or joints at which different fibre lines connect with the core DFA network. 762 759 Smith EWB p 271 para 227. Nunes FWB p 158 para 6.22. 760 Exhibit BQ Theron lide 33. Theron EWB p 85 Table 9. 761 Transcript p 634 line 21. 762 Hodge EWB p 135 para 214: Aggregation nodes are central points in the network where multiple connections converge. Aggregation nodes aggregate or collect traffic from multiple sources and distributes it efficiently across the network. They also serve as critical hubs for routing data to and from different destinations. Non-Confidential 200 712. DFA does not allow third parties, including the MNOs, to aggregate connections at its aggregation nodes. As such, any party that chooses to self-build site connections needs to build fibre infrastructure that would terminate at its own or the nearest non-DFA aggregation node 763 particularly in geographic areas where DFA is the primary or only provider of dark fibre infrastructure. Again, the implication of this is that it is generally more cost effective to lease access to links from DFA (or another provider) where it is present 764 , relative to the costs of self-build or the use of third-party infrastructure, which enhances the reliance            713. The large footprint of DFA and restricted access to its aggregation nodes goes to the limitations of any argument that DFA does not have the ability to foreclose. In reality, there appear to be few alternatives available to MNOs. On a static basis, those alternatives only account for 14% of the available dark fibre infrastructure used to connect sites. Any competing network is unlikely to have connections and nodes in all areas where these are required by MNOs in the short- to medium-term. 765 714. In a dynamic sense, switching to a self-build strategy is costly and would take considerable time to replicate the required infrastructure, and switching to the network of another dark fibre infrastructure provider would take time 766 and significant resources even if only in areas where alternative providers are present, and more so where they are not at present 767               which creates   768 763 Or alternatively to use microwave links as Rain would have had to do in some instances where the costs of building its own fibre lines to connect with a third-party node were considered to be prohibitive due to the significant distances. 764 Smith EWB p 250 para 50. 765 For example, only provides dark fibre in Gauteng and KwaZulu-Natal (see Smith EWB p 250 para 163). The only other significant provider of fibre for mobile backhaul is Liquid Telecom. 766 Smith EWB p 255 para 175. 767 Mdlalose FWB p 392 paras 28  29. 768 Smith EWB p 255 para 175 footnote 260. Van der Merwe FWB Annexure 3 p 55 (or p 9 of subsection    Non-Confidential 201 715. Prof Theron notes that in areas where the merger parties are the only providers of FTTS infrastructure, there would be a loss of a competitor. Alternatives and switching 716. A key consideration is that MNOs such as MTN would also have to incur significant costs in un-      reconnecting sites that already have fibre connections were they to switch away from procuring from DFA. 769 Exiting current contractual arrangements would imply certain costs and time, let alone the costs of building or connecting with new infrastructure if at all possible. 717. The significance of DFA in this market is not dissimilar to the role played by other critical infrastructure, such as electricity, relied upon by businesses. That is, while electricity in some cases accounts for a relatively small proportion of total operating costs of a firm, the absence of electricity or interruptions in service even for a small period of time can result in significant disruption in the operations of a firm (in the absence of alternative sources of power). In our understanding, the same appears to be true of core fibre infrastructure connectivity for MNOs  that is, in key areas where a user such as MTN relies on DFA links, a degradation or interruption of service will be harmful to its operations and to its customers. This risk would be even more stark in the case of Vodacom, Rain and Cell C given their relatively higher level of exposure to DFA as discussed above. 718. Our assessment is that there is an all-or-nothing aspect to metro dark fibre access networks  if DFA is present, it makes sense to connect with its network and aggregation nodes. Indeed, this is precisely the nature of the service that DFA set out to provide to MNOs since its establishment so that they could focus on their core activities  that is, to connect their sites with fibre infrastructure and 769 Hodge EWB p 138 para 223           2022, Part B of the Record, Part 1  Competitors ) (2201  6118) p 3850 para 13.2. Non-Confidential 202 generate scale efficiencies in these networks by aggregating connections and creating a shared infrastructure platform. 719. We therefore conclude that Maziv through DFA has an ability to foreclose MNO rivals of Vodacom of access to critical dark fibre inputs for FTTS connectivity or mobile backhaul. Incentive to foreclose rival MNOs 720. In competition theory and practice, the incentive to foreclose relates to whether it would be profitable for a firm to foreclose a customer or supplier. In the case of input foreclosure, it relates to the ability to deny downstream users, buyers or customers access to a good or service or to raise their costs or degrade their terms of access to an important input. The issue at hand is to assess, if an ability to foreclose exists, why a profit maximising upstream firm would choose to foreclose the downstream customer, particularly in this case, where there are elements of vertical integration through the upstream and downstream. Key concerns arising from a change of incentives 721. By way of context, the Commission and Mr Hodge raise several factors that point to the fact that the interests of Vodacom and Maziv will be aligned in foreclosing rival MNOs access to a critical input. 722. First, Vodacom has an interest in ensuring that its need to fiberise its network for 4G, FWA and 5G rollout is met and to seek any advantage over its rivals in the process where it can. 770 To ensure that its requirements are prioritised, so goes the argument, it sought co-control over Maziv otherwise its post-merger fulfilment of its fibre needs was to be no better off than pre-merger wherein it contests with others for DFA capex and priority. 770 Hodge EWB p 139 para 228. Non-Confidential 203 723. Second             expansion strategy through its post-merger partnership with Vodacom so as not to lose its            and expansion secured through a ROFR commitment as part of the proposed transaction 771 . That is, it has made sure that it secured the custom of Vodacom going forward (in a growing market), unlike with other MNOs, consistent with           772 . 724. Third, Mr Hodge adds that it is unlikely that Maziv would not prioritise the requirements of its largest shareholder that relies on it for a key input, particularly in circumstances where that shareholder has veto rights to hold it to account or at least affect its decisions at the Board level, to remove the CEO and CFO, and to veto budgets an         performance. 725.              expansion opportunities in other African countries, which could be withheld by Vodacom if it thought Maziv was not prioritising its needs. 773 As such, it would not matter whether Vodacom had control of the day-to-day operations of Maziv for its needs to be prioritised. 726. Fifth, Mr Hodge argues that DFA could further advantage Vodacom by providing discounts against its standard pricing parameters (rate card, which includes term, volume, upfront payment or monthly components, and build distance parameters). It is able to agree non-standard pricing terms with clients where some of the fixed parameters are altered; or where customer specific or bespoke product sets are provided allowing for technical deviations from standard rate card pricing; or applying its discretion in offering customer specific discounts in response to competitor pricing (secured under the ROFR terms) or discounts for volume and monthly revenue commitments. 774 The above all goes to suggest 771 Part A of the Record p 785  789 clause 19. 772 Part A of the Record p 1223 CIVH Board Pack 27 October 2021. 773 Hodge EWB p 139 para 230. 774 Commission Report p 293  300. Non-Confidential 204 that DFA, post-merger, would be in a position to ostensibly justify providing Vodacom with better, differentiated pricing compared to its MNO rivals, with DFA effectively negotiating with its largest shareholder that has sight of its costs, margins and plans. 727. The merger parties argue that it would not make sense for Maziv to preference Vodacom, particularly in overbuilt areas, because MNO customers could switch to alternatives or self-build infrastructure if Maziv raised prices or sought to preference Vodacom. In areas where there is no overbuild and Maziv is already present, the merger parties argue that the proposed merger does not change   Alignment with Vodacom, preferencing and profitability 728. The economic experts agree that Maziv can be assumed to be profit maximising, pre-merger. Mr Hodge (and Mr Smith 775 ) nonetheless argue that there are partial foreclosure strategies that could be feasible for Maziv to pursue post-merger, pertaining to the partial foreclosure of rivals of Vodacom. 776 That is, even though Maziv operates an open access business model which is premised on multi- tenancy to share infrastructure costs, a selective refusal to supply (such as through prioritising Vodacom-linked rollout to give it first-mover advantage) is sti -             a material impact in the market. 729. Ultimately, because Vodacom would have a significant economic interest in Maziv post-merger, it can, it is argued by the Commission and MTN 777 , treat its DFA purchases differently internally in that it will receive a return from Maziv             rivals. That is, even if Maziv were to put through increases in prices across all 775 Hodge EWB p 141 para 236. 776 Hodge EWB p 140  141 paras 234  236. 777 Smith EWB p 272 para 229. Non-Confidential 205 MNO customers, the effective cost to Vodacom is less because it benefits from the offsetting effect of its (30-40%) share of profits in Maziv. 778 730. Mr Smith argues that in the event of a targeted or market-wide price increase, Maziv is likely to perceive a reduced risk of its largest customer 779 of Titan and substantial shareholder, Vodacom, switching away or self-building, whereas Vodacom has some bargaining power vis-à-vis Maziv pre-merger. 780 731. Prof Theron contends that post-transaction, if Maziv cannot provide services on competitive terms to Vodacom, it will move its business to other suppliers,             ignores all of the corporate dynamics and incentives, as per our analysis in paragraphs 140 to 198 above. We concur with Mr Smith that Vodacom, as a           a significant proportion of increased Titan prices through its shareholding. Competitors of Vodacom will be subject to the same higher price but will however not be able to recoup anything and will have to pass on the price increases to their customers. 732. The benefit from future investment by Vodacom in a dynamic market, as well as the significance of the Transfer Assets, and also the impact on concentration/competition is already evident from the fact that Vodacom will add to the Transfer Assets the investments it has made in fibre since concluding this deal. This in itself is significant. 733. The dependence of Vodacom on DFA for mobile backhaul, and thus the strategic alignment of Vodacom and Maziv, would be enhanced post-merger as we show below: 733.1. Pre-merger Vodacom accounts for       Cell C, Rain and MTN account for %, % and %, respectively.           - 778 Smith EWB p 273 para 229. 779 See Smith EWB p 272 Figure 14. 780 Smith EWB p 272 para 229. Non-Confidential 206 merger,         procured from . Vodacom currently has sites, so this would mean that Vodacom would % of its mobile backhaul 781 As such, the increment in the number of links is significant in the FTTS landscape as they represent            signific           733.2.  metro fibre network is estimated at around 14,000  15,000 km. 782 The addition of km 783 of metro fibre translates to about a %             733.3. In             by DFA post-merger as the transaction agreements give DFA the ROFR                  as part of its plans to fulfil the spectrum obligations, Vodacom plans to deploy a minimum of 5G sites per year for the five years leading to 2027/8. This will result in the addition of new sites, requiring an additional new FTTS links. Absent the merger, DFA would have to compete for this increased demand from Vodacom, but with the merger, this demand will support the expansion of DFA. 734. The above marks a significant increase in links  by comparison and to illustrate the significance of this change, we note that MTN procures links from DFA (of Titan and Peregrine) 784 . Mr Smith also notes that as this dependence increases over time, say to a level where Vodacom procured 100% of its requirements from DFA, other providers of mobile backhaul currently serving DFA could be foreclosed. 735.               a customer. In turn, we consider that Vodacom, acting in the Board of Maziv, 781 See Smith EWB p 273, footnote 325; Maduray FWB p 417 Table 4. 782 Bundle M p 12981: BMIT SA Wholesale Telecoms Report 2023, Slide 81. 783 1,110 Titan links plus 1,326 km Peregrine links. 784 Commission Report p 278 Table 46. Non-Confidential 207 would have a significant incentive to shape the decisions of Maziv in a manner that suits its interests as an MNO and important player in the affected market. Vodacom would be in a position to inter alia influence and have the right to approve the budget and expansion plans linked with the business plans of Maziv. Furthermore, Vodacom would be able to influence decisions to prioritise rollout that suited its own plans such as in relation to site builds for 5G given new spectrum that needs to be utilised, either in terms of: 735.1. rollout          over those of other MNOs which could still be calculated to be profitable or non-loss making for Maziv or DFA; and 735.2. rollout (timing, order and quantum) which prioritised mobile backhaul activities over other investments such as in FTTB or FT