---

 

COMPETITION TRIBUNAL OF SOUTH AFRICA

 

Case No: LM183Sep18

 

 

In the matter between

 

MIH eCommerce Holdings Pty Ltd t/a OLX South Africa  Primary Acquiring Firm

 

and

 

WeBuyCars Pty Ltd                                                          Primary Target Firm

 

Panel:            AW Wessels (Presiding Member)

M Mazwai (Tribunal Member)

I Valodia (Tribunal Member)

 

Heard on:       14-18; 21; 23-25; 28 October 2019;

5;12; 15 November 2019;

2; 3; 13 March 2020

 

Order Issued on:     26 March 2020

 

Reasons Issued on: 7 December 2020

 

 

REASONS FOR DECISION (PUBLIC VERSION)

 

 

INTRODUCTION

 

[1.] The proposed transaction entails the subscription by MIH e-Commerce Holdings (Pty) Ltd (MIHE) t/a OLX South Africa ( OLX ) for 60% of the shares in We Buy Cars (Pty) Ltd ( WeBuyCars ). OLX is a wholly owned subsidiary of Naspers Limited ( Naspers ), a public company listed on the Johannesburg Stock Exchange.

 

[2.] OLX is a generalist (horizontal) online classified advertising platform which carries advertisements for a broad range of goods and services for purchase and sale, including used cars. Naspers also operates an online specialised (vertical) classified advertising platform for the purchase and sale of used and new vehicles called The Car Trader (Pty), t/a AutoTrader.

 

[3.] WeBuyCars operates a guaranteed-purchase used vehicle business, also referred to in these reasons as a car-buying service.[1] It operates on both sides of the market in that it is the largest used vehicle buying business in South Africa, but also a large seller of used vehicles.

 

[4.] The proposed transaction between Naspers and WeBuyCars would combine the horizontal online classified advertising platform of OLX with vertical online classified platform, with vehicle stock for sale and purchase through these platforms.

 

[5.] The Competition Commission ( Commission ) referred the proposed merger to the Tribunal ( Tribunal ) in terms of section 14A of the Competition Act, 1998, as amended ( Act ) for determination.

 

[6.] The Commission recommended the prohibition of the merger on grounds that it would firstly lead to significant unilateral effects as it would remove a potential competitor in the car-buying services market in South Africa; and secondly that it would give rise to conglomerate effects.

 

[7.] In regard to unilateral effects, the Commission took into account recent acquisition of an online car-buying and selling platform, which can also be categorised as a guaranteed car-buying service, Frontier Car Group, Inc ( FCG ), a Berlin based company which was allegedly poised to enter the South African market in competition with WeBuyCars. The Commission contended that FCG (through Naspers) would have entered the car-buying services market in South Africa in competition with WeBuyCars were it not for proposed acquisition of WeBuyCars.

 

[8.] The merging parties contended that no unilateral effects would arise from the proposed transaction since there are many other alternatives in the used car- buying and selling markets which would constrain the merged behaviour post-merger. They submitted that the contestable relevant product market was wider than the Commission sought to argue.

 

[9.] Regarding the potential entry of FCG, the merging parties disputed that FCG had the intention to enter the South African market. Furthermore, they submitted that whether FCG would have entered the market should be assessed not as a counterfactual at the time of the assessment of the proposed merger by the Commission, but on whether FCG would enter the market if this transaction were to be prohibited.

 

[10.] They argued further that, regardless of whether FCG would have entered the market or not, WeBuyCars would continue to face effective competition in the future. They submitted that even if FCG was to enter, it was unlikely to materially enhance the intensity of competition in the relevant markets.

 

[11.] With regard to conglomerate effects, the merging parties submitted that given the complementary nature of their post-merger service offering, the proposed transaction would result in so-called i.e. WeBuyCars may benefit from accessing OLX and platforms on the buy-side which can be leveraged on the sell-side. They however argued that there was no evidence that these benefits would be harmful to competition. The merging parties nevertheless tendered behavioural remedies which they argued addressed any concerns relating to conglomerate or portfolio effects.

 

[12.] It bears mention that the Commission and the merging parties were in agreement that any potential significant unilateral effects arising from the removal of a potential competitor, i.e., FCG, could not be cured through a remedy.

 

[13.] The Tribunal heard the matter over 15 days in the period May 2019 to March 2020. The hearing experienced a number of procedural disputes which protracted the proceedings.

 

[14.] We found that the proposed transaction would result in the removal of an effective potential competitor, FCG, from the guaranteed car-buying services market in South Africa.

 

[15.] Although the merging parties proposed behavioural remedies, which were amended during the course of the hearing in an attempt to address the competition concerns resulting from the proposed transaction, the remedies proposed could not address the significant unilateral effects resulting from the removal of a potential competitor.

 

[16.] In addition, we found that the evidence before us sustained the second theory of harm namely, that the transaction would give rise to conglomerate effects. We are of the view that these effects could be leveraged by the merging parties to: (i) entrench existing dominance; (ii) raise barriers to entry for other actual or potential competitors of WeBuyCars; and (iii) also provide reciprocal benefits to the Naspers businesses which would have been instrumental in entrenching dominance, and which no competitor could replicate.

 

[17.] Accordingly, we prohibited the transaction for the reasons that follow.

 

[18.] Prior to dealing with our reasons, we note that during the hearing, the Tribunal was required to decide two interlocutory applications, an Intervention Application by Weelee (Pty) Ltd ( Weelee ), and a Discovery Application brought by the Commission against the merging parties during the proceedings.

 

[19.] Our reasons are therefore set out as follows: Part A deals with the Intervention Application by Weelee, in Part B we deal with our reasons for the prohibition (the Merits), and in Part C, the reasons for dismissing the Discovery Application.

 

PART A

 

Intervention Application Reasons

 

INTRODUCTION

 

[20.] On 26 September 2019, some 11 business days before the hearing of this matter, Weelee filed an application in terms of section 53(c)(v) to intervene in the merger proceedings before us.

 

[21.] Weelee acts as an intermediary between private used vehicle sellers and car dealers by providing an online platform on which vehicles are auctioned. In providing a platform for the sale of used vehicles, Weelee competes with WeBuyCars.

 

[22.] During its investigation, the Commission engaged with Weelee, and intended to call a Weelee representative as one of its witnesses. To this end the Commission filed a witness statement by Marc Friedman (Mr Friedman) of Weelee.

 

[23.] The merging parties opposed intervention application. The Commission did not oppose the application, choosing to abide our decision.

 

[24.] We heard the intervention application on 4 October 2019 and granted it for the following reasons.

 

THE LAW

 

[25.] Section 53(c)(v) of the Act empowers the Tribunal with the discretion to recognise, as a participant in merger proceedings, a person that is not party to the merger provided they can show material interest in the matter. This power must be exercised judicially, according to the rules of reason and justice, having regard to the particular facts and circumstances of each case.[2]

 

[26.] The position as enunciated by the CAC on intervenors in merger proceedings is as follows:

 

26.1. A party who can prove a material and substantial interest would fall within the class of parties who may be admitted upon the exercise of judicial discretion by the Tribunal.[3]

 

26.2. A party who is unable to show a material substantial interest in the matter may well be admitted if it is able to provide evidence of its ability to assist the Tribunal in the consideration of the application of the various purposes of the Act as contained in section 1 thereof to the relevant merger transaction.[4]

 

[27.] An intervention application may thus succeed if the applicant demonstrates a genuine ability to assist the Tribunal in discharging its statutory mandate, i.e. conducting its mandate in terms of section 12A.

 

[28.] In this regard, a potential probable contribution to merger proceedings must be weighed up against the foreseeable consequence of intervention in terms of the expedition and resolution of the proceedings.[5]

 

The Tribunal also held in Caxton and CTP Publishers and Printers Ltd v Media24 (Pty) Ltd and others [2017] 1 CPLR 432 (CT) (Caxton) para 45:

 

 assessment is designed to be expeditious so that merger considerations are not unavoidably delayed. Competition authorities for this reason adopt a pragmatic approach. It is not necessary to pontificate on issues for the sake of them if they bear no relevance to the outcome of whether the merger should be cleared or

 

[29.] It has also been ruled that commercial interest or strategic aspirations are insufficient to warrant participation in merger proceedings,[6] and that inference or speculation will not suffice, the founding papers must detail the unique contribution that the applicant is able to make.[7]

 

[30.] It is also noted that a showing of (potential) material inadequacies in the recommendation may be favourable to the outcome of an application.[8]

 

THE ARGUMENTS

 

[31.] In its application, Weelee submitted that it both had a material and substantial interest in the merger and would be able to assist the Tribunal in its consideration of the proposed merger.

 

[32.] Weelee argued that as a used vehicle buying platform it was a relatively new entrant in the used vehicle market in South Africa and thus had a material and substantial interest in the merger.[9]

 

[33.] Weelee also submitted that it would be able to assist the Tribunal in its determination of the merger. To this effect, it was submitted that although its CEO, Mr Friedman, had already submitted a witness statement on behalf of the Commission, the case for prohibition was, in one respect, different to the concerns raised by Weelee.

 

[34.] Weelee submitted that its case for a prohibition of the proposed transaction would include argument advancing a direct exclusion/ foreclosure theory of harm which, unlike the theories, was not dependent upon proving harmful portfolio effects. Weelee submitted that it would argue that in a digital economy, the post-merger access to data gained by the merged entity through this merger would allow the merged firm the ability to exclude competitors. argument was that access to data was an important element in this case, relying on the increasing realisation that, in the modern digital economy, access to consumer behavioural data (in particular, their behaviour on the internet) can be a key competitive tool which can be used, with ruthless efficiency, to exclude competitors from the market.

 

[35.] Where the Commission and cases aligned, as it did on the issue of market definition, Weelee submitted that the Tribunal would benefit from the presence of a market participant in the hearing.

 

[36.] Weelee furthermore submitted that its presence for the duration of the hearing as an active participant would significantly bolster the proceedings as it would be in a position to participate throughout the proceedings and not only for the limited duration when providing evidence as a witness for the Commission.

 

[37.] The merging parties opposed the application and argued firstly, that Weelee did not have a material and substantial interest in the merger which is not sufficiently represented by the case; and secondly that Weelee would not be able to assist the Tribunal in such a way as to warrant the prejudice arising in the form of further delays in the matter triggered by     late intervention application.

 

OUR ASSESSMENT

 

[38.] As a relatively new entrant providing an online platform on which used vehicles are auctioned in South Africa, Weelee would have first-hand knowledge of the features and characteristics of the used vehicle market in South Africa, and furthermore could speak to entry requirements and barriers to entry into the market.

 

[39.] Given that this merger involves evolving markets where access to and the use of data may be an important factor to consider in the competitive assessment, we were persuaded that Weelee would be of assistance to the Tribunal in understanding the role of data in the competitive process and the competitive intricacies and complexities of the markets concerned. The proposed merger would combine assets including data in the hands of WeBuyCars with online classified advertising platform, data in the hands of AutoTrader, and expertise in data analysis. Understanding the dynamics of the market(s) from the perspective of a relatively new participant who was providing services in South Africa, using a car-buying model that was data intensive, similar to the potential entrant FCG, would have utility in our assessment of competitive dynamics in these markets.

 

[40.] However, we noted that granting an intervention in which Weelee could enjoy full intervenor rights, including leading new evidence, requesting documentation, and calling new witnesses at such a late stage would undoubtedly lead to severe prejudice and delay the matter further. To prevent this, we granted the intervention on circumscribed grounds.

 

[41.] The first was to grant Weelee the opportunity to lead its witness, Mr Friedman, only on the witness statement already provided (as a Commission witness) and only on the issues which had not been led by the Commission.

 

[42.] The second was the ability to cross-examine the merging witnesses only to the extent that such did not result in a duplication of issues of the cross examination. To facilitate this, Weelee was afforded procedural rights of access to the record and any additional documents filed in the lead up to the hearings under the conventional confidentiality regime.

 

[43.] The intervention rights granted to Weelee were furthermore subject to there not being any disruption to the       existing timetable for the hearing. Further, that Weelee would adhere to the time allocation set for the hearing and any limitations imposed from time to time. In our view, these limitations would cure any potential prejudice that could have arisen from the intervention.

 

PART B

 

THE MERITS

 

PARTIES TO THE PROPOSED TRANSACTION

 

Primary Acquiring firm

 

[45.] OLX is an operating division of MIH eCommerce Holdings (Pty) Ltd,[10] which is an investment holding firm ultimately controlled by Naspers, a public company listed on the Johannesburg Stock Exchange.

 

[46.] OLX is a generalist (horizontal) online classified advertising platform which carries advertisements for a broad range of goods and services, including electronics and computers, fashion and beauty, sports and outdoors, pets, and various kinds of used vehicles.

 

[47.] Naspers controls in excess of 79 firms in South Africa and an even greater number abroad. Three firms which are of particular relevance to the assessment of the proposed transaction are:

 

47.1.  AutoTrader: AutoTrader is a specialised (vertical) online advertising portal, specialising in the advertising of motor vehicles.

 

47.2.  Media24 (Pty) Ltd (Media24): Media24 is a print media division of Naspers, with interests in various magazines, newspapers and digital platforms in South Africa.

 

47.3.  FCG: FCG is a car-buying and selling platform in which the Naspers Group had recently acquired a 34.92% shareholding and by November 2019 when we heard the matter, it had increased its shareholding to becoming the controlling shareholder in FCG.

 

[48.] Naspers has a multinational presence. Of relevance is that it owns and operates OLX advertising platforms in various other countries (under different legal entities).

 

[49.] In addition, Naspers controls FCG which operates used car platforms in various emerging market economies, pursuant to which used cars are auctioned online on behalf of clients. The Naspers Group acquired a minority interest in FCG in May 2018 and in November 2019 announced a further investment in terms of which it became controlling shareholder.

 

[50.] FCG, is a Berlin-based firm operating in various countries, buys used vehicles from private individuals and from car rental companies. FCG then sells the used vehicles to dealers (it does not sell to individual consumers). When FCG purchases a used vehicle from a consumer, it has either already arranged to re-sell the vehicle to a dealer immediately or, through the use of analytical software and data collated over time, will purchase the vehicle knowing what the demand and price points for the vehicle are so that the vehicle can be resold immediately.  model operates on the basis that whilst ownership of the used vehicle may transfer briefly to FCG, it does not take vehicles into stock, which would require real estate space. FCG describes itself as a middle-man between a private seller and a dealer and is able to negotiate competitive prices for sellers through the aggregation of demand.

 

[51.] Whilst the operations of FCG will be discussed in greater detail below, we note here that Naspers considered FCG to operate as [CONFIDENTIAL][11] and the firm, founded and run by Mr Sujay Tyle Mr Tyle is operational in several emerging market economies including Nigeria, Mexico, Argentina, Colombia, Chile, Pakistan and Indonesia. Of importance to the competitive assessment of the proposed transaction is alleged planned entry into South Africa in competition with WeBuyCars.

 

Primary Target Firm

 

[52.] WeBuyCars is a South African company controlled by [CONFIDENTIAL].

 

[53.] WeBuyCars has since 2001 been active in the business of purchasing and selling second-hand vehicles in South Africa. WeBuyCars conducts its business under two brands, namely, WeBuyCars and      Car previously WeSellCars

 

RATIONALE FOR THE TRANSACTION

 

[54.] Sjoerd Nikkelen Mr Nikkelen the CEO of OLX Europe, Middle East and Africa EMEA , broadly explained that, from perspective, the proposed transaction was motivated by two main factors: a strategic factor and a financial one.[12]

 

Strategic Rationale

 

[55.] strategic rationale consisted of the following:

 

[55.1  55.4 CONFIDENTIAL].

 

[56.] OLX summarised its investment rationale for the transaction as four-fold:

 

[ 56.1  56.4 CONFIDENTIAL].

 

[57.] [CONFIDENTIAL]

 

[58.] In its recommendation, the Commission situated the current transaction in the context of Naspers having acquired AutoTrader in 2017[13] to strengthen position in the market as a horizontal classified advertising platform through the use of the AutoTrader knowledge and experience. The Commission argued that in this transaction, the purchase of WeBuyCars by Naspers is an attempt to provide users of OLX with an instant cash offering. The merger in view [CONFIDENTIAL]. While the Commission did not dispute the rationale put up by the merging parties, it highlighted that portion of the rationale relating to the unlocking of marketing and other synergies between the Naspers online platforms and the WeBuyCars business. The Commission argued that it was most likely the intent of Naspers to grow its online platforms, and at the same time increase its presence in the car-buying services market.

 

 

financial rationale

 

[59.] Mr Nikkelen explained that there was also a purely financial rationale for the transaction.

 

[60.] When WeBuyCars was brought to   attention, it was impressed by size, the quality and efficiency of its operations, and its profitability. WeBuyCars was a successful business which has shown substantial growth since it was established and has good growth prospects. OLX was also motivated by [CONDFIDENTIAL].[14]

 

WeBuyCars rationale

 

[61.] Adriaan van der Walt [CONFIDENTIAL].[15]

 

[62.] Mr van der Walt also believed [CONFIDENTIAL].[16]

 

THE HEARING

 

[63.] The Commission filed witness statements for:

 

63.1.  Marc Friedman ( Mr Friedman ) of Weelee;

 

63.2.  Shaun Harmse ( Mr Harmse ) of LSM Distributors (Pty) Ltd, trading as Porsche Centre South Africa (Porsche);

 

63.3.  Colin Morgan ( Mr Morgan ) of GetWorth Pty Ltd ( GetWorth ); and

 

63.4.  Its economic expert, James Hodge (Mr Hodge), the Chief Economist.

 

[64.] The merging parties filed witness statements for:

 

64.1.  Adriaan van der Walt of WeBuyCars (also known and referred to in the documentation as    van der Walt or

 

64.2.  George Mienie ( Mr Mienie ) of AutoTrader;

 

64.3.  Mr Nikkelen of the OLX Group;

 

64.4.  Mr Tyle of FCG; and

 

64.5.  its economic expert, Richard Murgatroyd ( Mr Murgatroyd ), of RBB Economics.

 

[65.] A significant issue during the course of the hearing related to the availability, or otherwise, of Mr Tyle of FCG as a factual witness.

 

[66.] On 10 October 2019 (2 business days before the hearing), the merging representatives notified the parties of the order in which their factual witnesses were to be called. In this email, it was indicated that Mr Tyle was not available as initially understood, but the legal representatives were working to resolve the matter. In a subsequent email, the representatives indicated that Mr Tyle would only be available to present his factual witness statements after the hearing of the experts. The Commission objected to this suggestion, saying that it would be manifestly prejudicial and unfair if this witness was led only after the economic experts were heard.

 

[67.] On 14 October 2019, being the first day of the hearing, the Commission approached the Tribunal for direction on Mr Tyle availability. Mr Wilson, for the merging parties, submitted that the merging parties would endeavour to confirm Mr Tyle availability and respond. No more was said on the matter as the hearing progressed.

 

[68.] At the end of the day on 17 October 2019, the merging parties indicated that Mr Tyle would not be available on the hearing dates. The Commission submitted that alternative dates would need to be allocated to facilitate his presence and testimony. The merging parties were unwilling to delay the hearing of the matter for the purpose of calling Mr Tyle and therefore elected not to call him as a witness. The Commission argued that if Mr Tyle would not be testifying, his witness statement should be withdrawn from the record. The merging parties submitted that they were not required to do so, but agreed to withdraw it nonetheless. Counsel for the merging parties confirmed that his witness statement was accordingly pro non scripto, i.e., as if it had never been prepared. [17]

 

[69.] Given that Mr Tyle would have been able to respond to the first theory of harm relating to alleged planned entry into South Africa, we found his failure to make adequate arrangements for coming before us to give version uncooperative. We draw a negative inference from this.

 

[70.] As discussed above, Weelee brought an intervention application which was opposed by the merging parties. We granted application to intervene.[18]

 

[71.] The Commission elected not to call Mr Harmse and only called Mr Morgan of GetWorth and Mr Friedman of Weelee.

 

[72.] GetWorth, a competitor of WeBuyCars, was founded in 2016. It is active in what it styles the car platform buying used cars from the public and on- selling them to private purchasers.

 

[73.] Mr Friedman testified on behalf of Weelee. Weelee, which started trading in 2017, acts as an intermediary between private sellers of used vehicles and car dealers by providing a platform on which the vehicles are auctioned. It operates as an online intermediary, facilitating the sale by private individuals of used vehicles (for cash) to second-hand vehicle dealers through its online platform.

 

[74.] [….] online platform attracts private sellers to advertise and offer their vehicles for sale on its platform and allows pre-approved cash buyers (motor dealers) to compete through cash offers for the vehicles available on its platform. The motor dealers on online platform bid for the vehicle in question. The winning dealer inspects the vehicle, and if the vehicle is as described, pays the winning bid amount to the customer.

 

[75.] Weelee is not involved in the on-selling of the used cars to end consumers, whereas GetWorth, and until recently WeBuyCars do. Weelee is a facilitator and in this sense operates a similar model to FCG.

 

[76.] The merging parties called three factual witnesses: Mr van der Walt, Mr Mienie and Mr Nikkelen.

 

[77.] In closing argument, the Commission and Weelee sought findings against the credibility of the merging factual witnesses. We shall assess the credibility of each testimony, where appropriate, in the proper context.

 

[78.] Before considering the delineation of the relevant markets for purposes of the proposed transaction, we set out the industry background.

 

INDUSTRY BACKGROUND

 

[79.] WeBuyCars is in the business of purchasing used vehicles from private sellers for cash and selling those used cars both to dealers (i.e., wholesale) and private buyers (i.e., retail). Its business model is referred to in the industry as an cash business model and it has been active since 2001.

 

[80.] Prior to the advent of car-buying services such as WeBuyCars, a customer wishing to sell a used vehicle had three options (i) to sell privately through his / her network of family and friends or a     sign on the car, or through the classifieds; or (ii) by trading-in the used car for a new or used car; or (iii) by selling to a used car dealer.

 

[81.] A used car dealer, on the other hand, had the option to purchase vehicles (i) through trade-ins; or (ii) wholesale, from other dealers or auction houses; or (iii) from the public through -

 

[82.] These existing options created inefficiencies for both sellers and buyers of used vehicles.

 

[83.] Selling a used vehicle through a private sale is a time-consuming process and can be a risky exercise:

 

83.1.  the seller is tied up in the logistics of selling, including meeting potential buyers often with no certainty of a sale in the end;

 

83.2.  the seller is exposed to the security risks associated with dealing with unknown or bogus buyers, as well as the risks associated with receiving payment;

 

83.3.  the seller cannot secure or facilitate the granting of financing for purposes of the purchase of the vehicle; and

 

83.4.  information asymmetries between the seller and buyer makes accurate pricing difficult.

 

[84.] Selling through a trade-in eliminates most of the hassles of private selling, but often at the cost of an unattractively low price. Testimony from Mr Morgan revealed that this is because the private seller in this instance is often less concerned about the price s/he will receive for the sale of the car as opposed to the purchase of the new vehicle. Often, the used vehicle to be sold is financed; and the private seller is still indebted to a financier. By trading in the used car, the dealer is able to structure a deal, i.e. add the outstanding debt amount on the new car, for example through extended repayment terms on the new vehicle. [19]

 

[85.] The third option, a direct sale to a dealer, means that the private seller will likely receive a wholesale price and, in order to get the best possible price, will have to engage in the time-consuming exercise of ping          for the best deal among dealers.

 

[86.] Mr Morgan and Mr Friedman testified to the fact that many dealers, and large independent dealers in particular (as opposed to large groups such as Bidvest, or large dealers such as Barloworld and Imperial who sell the vehicles of original equipment manufacturers) typically experience difficulties in securing enough stock, as well as the required range of appropriate stock. This is due to insufficient used cars being available for sale in their local catchment area. Sourcing then becomes time-consuming and inefficient, with dealers spending 40% to 90% of their time sourcing second-hand stock. Dealers would also have to spend a significant amount of money on marketing in order to source the level and range of requisite second-hand stock. As the evidence later shows, dealers are typically customers of car-buying services rather than their competitors.

 

WeBuyCars

 

[87.] In his testimony, Mr van der Walt explained that he and his brother Dirk founded WeBuyCars in 2001 in pursuit of their passion for cars, and in order to address the gaps between the private seller of a used vehicle and dealer discussed above. WeBuyCars buys most of its vehicles from private individuals. Potential sellers of used vehicles contact WeBuyCars telephonically ([CONFIDENTIAL]), through their website ([CONFIDENTIAL]) or, in some instances, by visiting a WeBuyCars warehouse or buying pod ([CONFIDENTIAL]).

 

[88.] WeBuyCars does not have any commercial arrangements for the ongoing purchase of vehicles from corporate entities (for example, the purchase of -of- life" fleet or rental vehicles) or from any dealership. In limited instances, WeBuyCars may be approached by a small used vehicle fleet owner and it will consider purchasing those vehicles.

 

[89.] WeBuyCars is headquartered in Midstream, Gauteng. It operates a total of 8 warehouses situated in Gauteng, the Western Cape, KwaZulu-Natal and the Eastern Cape provinces. These warehouses operate as large vehicle showrooms where private sellers can engage with WeBuyCars should they wish to sell their vehicles and prospective purchasers can view the used vehicles that are for sale.

 

[90.] WeBuyCars also hosts a used vehicle auction at its Gauteng branches.

 

[91.] In addition to these warehouses, WeBuyCars has 18 "buying pods" situated across the country. These buying pods are inspection centres where the sellers of used vehicles can take their vehicles for inspection and sale to WeBuyCars.

 

[92.] WeBuyCars employs approximately 147 buyers nationally, whose function is to visit potential private sellers at locations that are convenient to the sellers in order to inspect their vehicles and, if WeBuyCars so elects, to offer to purchase such vehicles.

 

[93.] Most of WeBuyCars stock is listed on its website. Potential purchasers can contact WeBuyCars through its website, and then visit a warehouse or join an auction to purchase a vehicle after viewing the vehicle online.

 

How WeBuyCars purchases a vehicle

 

[94.] WeBuyCars process of purchasing a used vehicle from a potential seller is summarised as follows:

 

[95.] A prospective seller will contact WeBuyCars through one of the channels referred to above.

 

[96.] WeBuyCars will give an indication to the potential seller of the price range that it is likely to offer, subject to the inspection and test drive of the vehicle, and the seller complying with other requirements for the sale of the vehicle. This price indication is provided over e-mail, SMS or telephone.

 

[97.] WeBuyCars then arranges a visit to the prospective seller to inspect and test drive the vehicle.[20] This is done either at a location selected by the prospective seller, at a WeBuyCars warehouse or at a buying pod.

 

[98.] Should WeBuyCars be satisfied with the result of the inspection and test drive, it will make an offer to the prospective seller to purchase the vehicle.

 

[99.] WeBuyCars has extensive experience and expertise in pricing second-hand vehicles for purchase or sale, and relies on this knowledge to determine the best possible price to offer a prospective seller.

 

[100.] The WeBuyCars buyer may negotiate the ultimate selling price with the seller. If the seller accepts offer, WeBuyCars purchases the vehicle and makes immediate payment to the seller. Where necessary, WeBuyCars will first settle any outstanding debt for vehicle financing and then pay the balance to the seller.

 

[101.] WeBuyCars then arranges for the collection of the vehicle from the seller, and completes the administration for the transfer of ownership of the vehicle.

 

[102.] After purchasing the vehicle, WeBuyCars will wash, photograph and price the vehicle. WeBuyCars does not generally do any "value addition" to the vehicles, such as body work or refurbishing of the vehicle.

 

[103.] Although the majority of leads for prospective sellers of used vehicles are generated from its website, no transacting occurs online, and purchases are always subject to a physical inspection, test drive and further required verification.

 

How WeBuyCars sells a vehicle

 

[104.] Once WeBuyCars buys a used vehicle from a private seller, that vehicle will form part of its vehicle stock for sale.

 

[105.] The sales process is summarised as follows:

 

105.1.   WeBuyCars first offers its stock to dealers at a wholesale price.

 

105.2.   WeBuyCars then advertises its stock at a retail price on its website and on various vehicle classifieds portals (such as AutoTrader).

 

105.3.   Prospective purchasers can visit the WeBuyCars warehouses to inspect the stock.

 

105.4.   Sales are concluded on site.

 

105.5.   WeBuyCars also hosts regular auctions at its Gauteng warehouses. The vehicles that are earmarked for sale on auction are listed on the WeBuyCars website prior to the auction. Prospective buyers can inspect the vehicles at the warehouse and elect to either purchase the vehicle prior to the auction, or to attend the auction and bid to purchase the vehicle.

 

105.6.   WeBuyCars assists with the transfer of title to the new owner and also assists in obtaining financing and insurance should the purchaser require this.

 

[106.] WeBuyCars submitted that [CONFIDENTIAL] of its stock is sold to dealers, generally referred to below as [….] , with the remaining sales to private individuals.

 

[107.] WeBuyCars further submitted that most dealers that purchase from WeBuyCars typically do so on auction.

 

[108.] WeBuyCars endeavours to offload stock as quickly as possible otherwise, as Mr Morgan put it, if you have 100 cars at R200 000 each, you are sitting on R20 million of metal [21]

 

[109.] Since WeBuyCars started providing car-buying services in 2001, other firms have sought to follow suit. Weelee, the intervenor in these proceedings, entered in 2017. GetWorth, the    witness, entered in 2016, as did CarZar. At the time of the hearing, all of the parties indicated that CarZar had exited the market between the time of the investigation and the hearing.

 

[110.] A common theme in evidence among the participants offering car-buying services was that they each seek to offer fast, convenient and risk-free transactions, which are more efficient than the cumbersome traditional dealer model.

 

[111.] Car buying services can act as principal, where they acquire the used car for their own stock, or as agents where they facilitate a sale to the ultimate buyer. WeBuyCars acts as a principal in the transaction.

 

111.1. GetWorth is active in what it styles the car platform buying cars as principal from the public and on-selling them to private purchasers.

 

111.2. Weelee operates an online platform that puts private sellers and pre- approved dealers together where the dealers compete by bidding for the cars available on platform. The winning dealer then inspects the car, and if the car is as described, pays the winning bid amount to the customer. In general, Weelee does not take stock.

 

111.3. FCG, the potential entrant into South Africa, was acquired by the Naspers Group (through OLX) shortly before the proposed transaction was concluded. It operates a model whereby it engages with the customer on the buy-side, inspects the vehicle, but does not offer cash or take ownership of the vehicle. Rather, it conducts instant online auctions of the vehicle to dealers signed up on its platform for a commission fee. The dealers have the benefit of the full inspection report and FCG acts as a trusted intermediary to facilitate the transaction. The FCG model also seeks to provide other services, like finance, the absence of which makes car trading complex in developing countries. FCG model is highly automated. In rare instances FCG will make its own cash offers for vehicles where it believes it can secure a sale at a potentially better price.

 

111.4. Carzar.co.za operated a similar model to FCG where it facilitated auctions but also did some instant cash purchases itself to build stock of vehicles.

 

[112.] It is common cause that once a car-buying service has purchased a used vehicle, it will on-sell it. Here, car-buying services may differentiate their business models by targeting a particular type of buyer. In this regard, three broad categories were identified:

 

112.1.   Customer-Dealer (C-D). These firms buy used vehicles from the public and then on-sell primarily to dealers (wholesale). Weelee is an example of such a firm.

 

112.2.   Customer-Customer (C-C). Firms which buy used vehicles from the public and then on-sell to the public (retail). GetWorth is an example of such a firm;

 

112.3.   Firms that use a combination of methods, firms such as WeBuyCars, YouDriveWeSell, Cars2Cash, which sell to both wholesale and retail.

 

COMPETITION ASSESSMENT

 

Theories of harm

 

[113.] The Commission consistently advanced two theories of harm:

 

113.1.   That the merger would result in unilateral effects as it would remove a potential entrant, FCG, from the car-buying services market;

 

113.2.   That the merger would entrench dominance through the portfolio effects arising from the proposed merger.

 

[114.] Both these theories, so the Commission argued, would result in a significant lessening of competition in the relevant market which it argued was the market for the purchase of used , which is also later referred to as the car- buying market.

 

[115.] The merging parties contended that both of the theories were incorrectly premised on an overly narrow definition of the relevant markets in which WeBuyCars competes and were misplaced. In essence, the merging parties contention was that the relevant product markets were those for the purchase and sale of used cars, whether over platforms or not and, most importantly, that used car dealers should be considered to form part of the relevant product markets.

 

Relevant market

 

[116.] It was common cause that the broad used car market can be further delineated into separate product markets for (i) the purchasing of used cars; and (ii) the selling of used cars.

 

[117.] The buying of used cars is a differentiated product market. Assessing the competitive effects of the proposed merger in a differentiated product market requires an examination of the closeness of competition among the various participants involved in the buying of used cars in South Africa. This assessment was a point of argument between the parties.

 

[118.] At the heart of the contested market definition on the buying side, is the question of whether guaranteed used car purchase operators or car-buying services compete with       nal dealers and, if so, to what extent.

 

[119.] It is common cause that the acquiring group currently does not provide a car-buying service similar to WeBuyCars in South Africa. The Commi argument however, was that FCG which is part of the Naspers Group, intended to enter the South African market in competition with WeBuyCars. OLX is a generalist (horizontal) online classifieds advertising platform which connects buyers and sellers of a broad range of products and services, including vehicles. AutoTrader, which also forms part of the acquiring group, is an online advertising portal specialising in the advertising of motor vehicles.

 

The  view

 

[120.] According to the Commission, the relevant product market in which WeBuyCars operates is the market for the guaranteed purchase and sale of used cars or car-buying     market. This is a two-sided market, consisting of i) the purchasing and ii) the selling of used cars.

 

[121.] The Commission identified features of car-buying service operators which, it submitted, distinguish them from traditional dealerships. The primary feature is that car-buying services focus on the buying rather than the selling of cars, unlike traditional dealers. Consequently, the marketing and advertising focus, capabilities required, working capital requirements, scale economies, network effects (resulting from two-sided markets), and first mover advantages are unique features of the car-buying service market when compared to the traditional dealer model.

 

[122.] The difference in focus between car-buying service operators and traditional dealers is illustrated, on the buying side, by the source of stock for each. Car- buying services primarily source their stock from private sellers, whereas d primary source of used vehicle stock is from trade-ins. In a trade-in scenario, the private seller is drawn to the purchase of a new vehicle, and the dealer is primarily interested in selling a new vehicle in other words, the purchase and sale of the used vehicle is ancillary to the purchase and sale of a new vehicle.

 

[123.] Since car-buying services cannot compete for these trade-ins, the Commission contends, they should not be taken into account when defining the relevant market in which WeBuyCars competes: car-buying services and traditional dealers do not exert a competitive constraint vis-à-vis one other.

 

[124.] Using the hypothetical monopolist test, the Commission argued that on the buying side, traditional used car dealers are unlikely to place a significant competitive constraint upon car-buying services (particularly once a car-buying service reaches scale, as it alleges is the case in respect of WeBuyCars) and are therefore not in the same relevant product market.

 

[125.] On the selling side, the Commission found that, since car-buying services seek to purchase at scale and offload stock quickly (or, in the case of certain platforms, such as FCG or Weelee, immediately without taking ownership of the stock themselves), the initial target for the selling side is the wholesale used car dealers who require constant volumes of stock to on-sell to individuals. The selling side may accordingly be segmented into two narrower markets: (i) the market for the sale of used vehicles to dealers (wholesale); and (ii) the market for the sale of used cars to individuals (retail).

 

[126.] The Commission regards car-buying services as to traditional used car dealerships. Traditional car dealers are set up to sell cars. This is their primary focus. In the past, in order to retrieve used car stock, dealers would primarily rely on passive channels to source stock such as trade-ins, cash trades, walk-ins, the local network of dealers, bank auctions of repossessed vehicles and, for the larger groups, ex-rental car stock. However, as explained by Mr Hodge, outside of these channels, traditional dealers were often left pursuing cash trade used car purchases from the public which were inefficient and costly. In addition, the dealer would often be limited to its local area, which could result in there being a mismatch between what was available from the public and what the dealer might require.

 

[127.] According to the Commission, dealers themselves saw the value of car-buying services as a means to obtain stock at a lower cost per lead, as well as allowing them to buy stock at scale.

 

[128.] On the consumer side, customers were also limited to selling their cars privately, or trading in their vehicles. However, car sellers would face the inconvenience of driving from dealer to dealer to seek out a sale and would also typically be limited to dealers within their local area. Given the narrow, localised market, there was a risk that there would be limited demand for the vehicle and hence the seller would have to deal with what Mr van der Walt in his witness statement, called low ball offers.

 

[129.] It was in this context that car-buying services entered and disrupted the market. It enabled owners to sell their vehicles at competitive prices, while also providing dealers with inventory which would be readily available.

 

The view of the merging parties

 

[130.] The merging parties contended for broad product markets in the two-sided market for the purchase and sale of used cars, with particular focus on the existence, and extent of competition, between traditional dealers and car-buying services.

 

[131.] In addition, they argued that nothing precludes car-buying services from contesting trade-in sales and accordingly that the market was significantly broader than the narrow market definition adopted by the Commission.

 

[132.] Mr Murgatroyd, the merging economic expert, was of the view that although WeBuyCars offers a differentiated service from traditional used car dealers, this differentiation does not insulate WeBuyCars from competition by traditional used car dealers. According to Mr Murgatroyd, if one were to look at the used vehicles purchased by WeBuyCars annually in South Africa as a function of the total number of used vehicles sold annually, using eNaTIS data, WeBuyCars accounts for a small portion of such vehicles. On Mr Murgatroyd analysis this means that WeBuyCars is losing sales to other operators in the market, including traditional dealers. Mr Murgatroyd accordingly submitted that the contestable market was wider than that which was contended for by the Commission.

 

[133.] This argument was also put forward by Mr Nikkelen who said that the universe of used cars sold per annum in South Africa is 125 000 as recorded by eNaTIS, while WeBuyCars sells approximately 60 000 used cars annually.

 

[134.] We focus first on the buying side before turning to the selling side.

 

Buy-Side: Do Traditional Dealers compete with Guaranteed Used Car Purchase Operators?

 

The evidence

 

[135.] In his witness statement, Mr Morgan stated that the actual addressable market open for used car purchasing is the so-called as opposed to the entire used car market as the merging parties contended. According to him, the excludes trade-ins, ex-rental fleets and private sales to dealers.[22]

 

[136.] This is because trade-ins of used vehicles, as mentioned above, are driven by the purchase of a new car. The sale of the vehicle is an ancillary means to an end. Insofar as ex-rental fleets are concerned, these are often kept within the relevant group and generally are not available to car-buying services.

 

[137.] Regarding private sales to dealers; the evidence was that the private seller who typically sells to a dealer is     i.e. s/he financed the purchase of his/ her existing vehicle and owes more on the used vehicle to the financier than s/he will receive from a potential wholesale purchaser of the vehicle. These kinds of sellers would likely only sell their vehicles to dealers, who would try to structure a deal: for example, by adding the difference between the financed amount owing on the vehicle to the amount being financed on the new vehicle being purchased so that the seller would settle the debt. These sources of stock should accordingly be excluded from the available market for car-buying services, according to Mr Morgan.

 

[138.] Both Mr Morgan and Mr Friedman testified that, by and large, when they lose sales, the lost sales (excluding those relating to sellers who are are to other car-buying services rather than to traditional dealers.[23]

 

Merging response

 

[139.] Mr van der Walt contended that a significant majority of the offers that WeBuyCars makes to prospective private sellers are rejected in favour of, inter alia, dealers:

 

[CONFIDENTIAL][24]

 

[140.] In support of the claim that most of the leads generated were lost to dealers, (and thus dealers were strong competitors to car-buying services), Mr van der Walt relied on a document containing an analysis of used vehicle leads on the buying side.[25] His evidence was that of the applications received on the WeBuyCars website (individuals reaching out to sell their cars), between [CONFIDENTIAL] of those leads were sent to WeBuyCars buyers. During his cross-examination the note on which he relied upon for this evidence was put up as exhibit E, reproduced below. exhibit E is entitled WeBuyCars buying for the period January 2018 to May 2019.

 

[EXTRACT CONFIDENTIAL]

 

WeBuyCars buying exhibit E in the Tribunal Record, handed up 18 October 2019.[26]

 

[141.] Having regard to the actual percentages reflected in exhibit E for the full period January 2018 to May 2019, Mr Murgatroyd observed that from the majority of by WeBuyCars, between [CONFIDENTIAL], do not result in sales, and that this conclusion still holds at the stage, where between [CONFIDENTIAL] of leads assigned do not result in sales. Even at the see stage, between [CONFIDENTIAL] of leads seen are lost by WeBuyCars.

 

[142.] Mr Murgatroyd pointed out further that even these percentages are likely to understate the true proportion of sellers who would consider switching to other options if WeBuyCars were to seek to degrade its buying offers, because they only reflect those sellers that reject WeBuyCars offers. As Mr Murgatroyd explained, it is likely that a proportion of those sellers who, in fact, accept WeBuyCars offers at their current level would also be prepared to switch if WeBuyCars were to seek to degrade its offers to them by even a small amount.

 

[143.] However, as the Commission pointed out, exhibit E did not accord with exhibit F which was a spreadsheet produced by the WeBuyCars buying department and which had formed part of the discovered documents. According to exhibit F, the total leads assigned at least for January 2018 was [CONFIDENTIAL]. This therefore brought into question the accuracy of the data being put up by Mr van der Walt in his evidence, as well as the reliance on this by Mr Murgatroyd. A point which the Tribunal debated with Mr van der Walt:

 

[CONFIDENTIAL].[27]

 

[144.] Mr Murgatroyd submitted that an assessment of WeBuyCars leads indicates that, even after leads have reached the assignment stage (i.e. from inspection of the used vehicle to an offer being made on the vehicle), between [CONFIDENTIAL] of the leads are rejected. His argument was that WeBuyCars is losing sales to other participants in the market, including traditional dealers.[28]

 

[145.] Mr Murgatroyd submitted further that the number of private sellers who chose to trade-in to a dealer, or sell to a dealer for cash, was materially higher than the number of sellers who sold used vehicles to a car-buying service.[29]

 

[146.] He acknowledged that WeBuyCars might be a relatively closer competitor to firms like Weelee and GetWorth but argued that such firms are unlikely to be the primary competitive constraint upon WeBuyCars.

 

[147.] We found that two uncertainties plagued the reliance on the merging leads analysis. The first was that there was a discrepancy between the number of leads assigned which may in turn result in an inaccuracy about the percentage of leads ultimately lost by WeBuyCars. In short, exhibit E shows WeBuyCars losing more leads than exhibit F does, thus downplaying performance for the purpose of showing that dealers were strong competitors.

 

[148.] The second, and perhaps more fundamental uncertainty was that there was no reliable indicator that the leads lost were lost to traditional dealers. Mr van der Walt could not confirm whether the leads assigned but ultimately lost were going to dealers or going to other car-buying services.

 

[149.] These uncertainties existed in the context of other car-buying services testifying that a significant number of sales lost by GetWorth are typically to other car-buying services and not to traditional dealers. He said:

 

I say, there is overlap between the various segments of the market. I can only really speak to what we see and what we perceive as the competition and the best way to gauge that is to look at, when a deal close for a customer asked for a leave and you ask the question, what happened to that deal? And again, transverse this. in impressionistic terms, but we the bulk of the deals that we lose to competition. And here excluding deals that we lose to privates, to someone selling privately. And excluding deals that we lose to trade-in, where we ask the price and we find that the price is kind of up here and in the retail range. So, of the customers who are selling in our universe, when we lose a deal to competition in that arena, typically to WeBuyCars or a Weelee of the world. And in fact, those two names and quite rare that another car-buying service that comes up. It is some it does sometimes occur that we lose a deal to a traditional dealer, but it is very uncommon. And our perception of that is that the traditional dealers compete on price to the same extent we [30]

 

[150.] Mr Friedman held the same view. He indicated that:

 

MR MAJENGE: And the purchases that you lose or the customers that you lose, those lost customers, to which competitors do those purchases go to? That is if you lose a customer or a seller to whom do you lose those to?

 

MR FRIEDMAN: Look the vast majority of the time it is to WeBuyCars. It is very clear, and its you know I am sure if I I during this process checked my emails for the last three days, so I am assuming in the 2 000/3 000 emails that are sitting in my inbox at the moment I can find a various emails to say sorry I have sold it to WeBuyCars or WeBuyCars have placed a different offer. It happens all the time. There are, to be fair, other small competitors that we do lose cars to and a customer may at stages sell a car privately or not necessarily be in the market or ultimately trade in their car. That is the case, they do have options there, but the vast majority of the time we lose our sales to [….] [31]

 

Documentary Evidence

 

[151.] Against the backdrop of the testimony of messrs van der Walt, Morgan, and Friedman dealt with above, we had particular regard to what the merging own strategic documents reveal about the relevant market(s) in which WeBuyCars competes and who they identify as WeBuyCars closest competitors.

 

[152.] Investment Committee described WeBuyCars as a [CONFIDENTIAL][32] [33](emphasis added)

 

[153.] When considering its purchase of a minority interest in FCG, [CONFIDENTIAL] [34](emphasis added)

 

[154.] Finally, in describing WeBuyCars market position, Naspers explained that it has a [CONFIDENTIAL] [35] and that:

 

WeBuyCars is the [CONFIDENTIAL] [36] (emphasis added)

 

[155.] Later in the document, [CONFIDENTIAL]:

 

[156.] Furthermore, two surveys, one conducted by OLX and one by Autotrader are also instructive of the acquiring views of the market.

 

[157.] In its survey, OLX [CONFIDENTIAL].[37]

 

[158.] Secondly, in a survey conducted by AutoTrader that tested views on a possible partnership between AutoTrader and FCG, [38] AutoTrader asks [CONFIDENTIAL]. Mr Mienie testified that this question was posed out of a desire to test customer responses to the WeBuyCars transaction, without mentioning WeBuyCars by name. In Mr Mienie view, dealers would be concerned about a tie up between AutoTrader and WeBuyCars and thus we needed to be really, really careful about how we communicated the transaction [39]

 

[159.] The survey is, in our view, a clear indication that the merging parties wished to test AutoTrader reaction to a tie up with FCG (a car-buying service) as a proxy for a tie up with WeBuyCars (another car-buying service). Implicit then is the impression that, to AutoTrader and the Naspers group broadly, there was some differentiation in the competitive world of car-buying services and traditional dealers. Case in point, the survey did not test reaction vis-à-vis a traditional dealership.

 

[160.] The evidence traversed above points to a car-buying market in which traditional dealers do not feature as competitors in the acquiring      own internal strategic documents, consistent with the evidence of Mr Morgan and Mr Friedman. Mr van der evidence that dealers provide a competitive constraint to car-buying services is not supported by the evidence before us.

 

The sell-side

 

[161.] Mr Hodge and Mr Murgatroyd also differed regarding the selling of used vehicles. Mr Hodge contended that there was a separate wholesale market for the sale of used cars to dealers; and a retail market where used cars are sold to end-users.

 

This is because car-buying services seek to offload the stock they purchase on the buy-side, as quickly as possible to dealers who then sell to end-users.

 

[162.] Mr Murgatroyd was of the view that there was a single market for the selling of used vehicles because dealers that sell to private individuals can and do sell to other dealers. Mr Hodge accepts that the guaranteed purchase model is capable of evolving to a situation where used cars are sold to end-users for quicker offloading.

 

The evidence

 

[163.] All the witnesses appear to accept that on the selling side WeBuyCars competes with dealers in the selling of used vehicles to end consumers.

 

[164.] Mr Morgan stated:

 

Then those get handed across into the sales function and we operate, like I said, very similar to what a traditional dealer would operate. And a traditional dealer is what you see when you drive down the road, the named franchises, like the local Audi, or something like that, or the small guy with 10 cars on the lot on the corner, a broad universe, but generally it is business that are focussed on, primarily on selling vehicles. With buying almost just being an ancillary portion of their business.[40]

 

[165.] Mr Friedman, under cross examination supported this view too:

 

ADV WILSON: And I think, if I understand your evidence correctly, that in that, can I call it selling, a private selling market a market in which WeBuyCars is competing with all dealers big and small, what you say there?

MR MR FRIEDMAN: Hm, but only for the sale of cars. ADV WILSON: Yes.

MR MR FRIEDMAN: The sale of cars.

ADV WILSON: talking on the sales side, ja. But in that selling market WeBuyCars is dealing with, is competing with all dealers big or small?

 

MR MR FRIEDMAN: Yes, in respect of used cars.[41]

 

[166.] Mr Murgatroyd contended that even if there were a narrower wholesale market for the sale of used cars to dealers on the selling side, the competitor set faced by WeBuyCars would logically be no narrower, and potentially materially broader, than on the buying side.[42]

 

[167.] With regards to the selling side, we are of the view that nothing of substance turns on whether the market for the sale of used cars is further segmented on the basis of wholesale to dealers or the sale to private individuals.

 

[168.] What is pertinent however from the evidence on the sell side, is firstly that the activities of WeBuyCars and other car-buying services are complementary to one another. WeBuyCars and other car-buying services are an important source of stock in trade for traditional dealers.

 

[169.] Mr van der Walt testified that WeBuyCars offers dealers the bite of the when it comes to the vehicles which it has purchased. It gives dealers the first opportunity to purchase an array of vehicles: vehicles which they would otherwise not have been able to source, either in the numbers or breadth of range offered by WeBuyCars.

 

[170.] He said:

 

ADV WILSON: But from the dealer perspective, in other words, insofar as existing dealers are concerned, I understand they are your customers on the sales side at least in part.

MR VAN DER WALT: Correct

ADV WILSON: What would be the benefit you would bring those dealers that would cause them to buy from you rather than direct from the public?

 

MR VAN DER WALT: I understand. So, buying from the public, if a dealer wants to source vehicles, you know, and a lot of them also put up advertisements nowadays that say they buy cars. One of our biggest dealers who buys from us also has a billboard on the highway saying, I am not going to mention their name, but they buy cars, so they also they want to attract as much stock as they can in the easiest fashion they can. So, a proposition we have for dealers is, firstly they get first pick on our stock, they get a generous dealer discount. So, if you would see a vehicle advertised on our website discounts varying according to price ranges from up to R12 000 for the higher price range vehicles.

ADV WILSON: And sorry, carry on.

MR VAN DER WALT: Further it is we have dedicated dealer sale staff who would look after dealers and nurture that relationship and contact them and I know you are looking for this type of vehicle, I have it in now or we had this vehicle on the floor, previously, it was too expensive for you but now had price drops, I think this is now right for you.[43]

 

[171.] Secondly, the evidence nn the selling side suggests that traditional dealers compete with car-buying services, in the sale of used cars to the end consumer. Accordingly, traditional dealers and would are a competitive constraint to car- buying services. Therefore, the proposed transaction does not appear to raise concerns on the sell side.

 

Overall Conclusion on the Relevant Market

 

[172.] Section 12A(1), read with 12A(2) of the Act stipulates that the effect of a merger on competition must be assessed within the context of the market (or markets) affected by the merger.

 

[173.] In differentiated markets, such as this one, it is useful to identify those firms and products that place a competitive constraint on the merging products and services and then rank them based on the magnitude of that constraint. In this regard, identifying close competitors allows the Tribunal to assess the likely effects of a merger without having to definitively conclude on the relevant market.

 

[174.] In Primedia,[44] the Tribunal considered that some of the most persuasive evidence on the issue of closeness of competition could be derived from the internal documents of the firms in the market, and in particular, those documents that were not prepared in contemplation of the merger hearing:

 

Absent reliable econometric evidence and inconclusive evidence from the demographics of audience and advertiser profiles, the next best evidence we have of who competes in the market are the opinions of the stations themselves. While Primedia witnesses testified to this in the hearing, some caution must be expressed about this evidence, given their interest in the outcome. One piece of evidence that cannot be regarded as having been drawn up in contemplation of this hearing are the internal documents of Kaya FM which were obtained under subpoena prior to the hearing.[45] (emphasis added)

 

[175.] It is clear from our consideration of the strategic, and other documentation, of the merging parties which we have traversed above, as well as from those which we will discuss below, that much has turned for us on these documents. They reveal the true opinions of the executives concerned, before an awareness about how they would be interpreted by the competition authorities or any other regulatory bodies had caused them to self-edit, or to be much more circumspect about how they expressed their opinions.

 

[176.] Furthermore, it is important to determine the real purpose behind the endeavour of delineating the relevant product market. In this instance, the product market importance is linked to the theories of harm advanced by the Commission. The Commission s first theory of harm is that the proposed merger resulted in the removal of a potential competitor to WeBuyCars in South Africa.

 

[177.] The importance of market definition in this case is therefore first to provide the analytical basis for determining whether or not FCG         alleged planned entry would have significantly impacted the market in which WeBuyCars operates, specifically on the buying side.

 

[178.] As stated above, there are many competitors on the selling side of the market. As such, it does not raise competition concerns. Our focus, for purposes of determining whether the proposed transaction raises competition concerns, is the buying side of the market. In particular, we are required to form a view on whether the relevant product market is the market for the purchase of used cars by car- buying services, or whether it should be more widely cast as the market for the purchase of used cars by car-buying services and traditional dealers.

 

[179.] We have concluded that on the buying side the relevant product market is the narrow one for car-buying services. We say this for the following reasons:

 

179.1. that the merging strategic documents, as discussed in paragraphs [151.] to [158.] above, confirm that traditional dealers do not exert any meaningful constraint upon car-buying services such as WeBuyCars, Weelee and GetWorth;

 

179.2. we took note of the evidence that traditional dealers are customers, rather than competitors of WeBuyCars, something which was borne out by a review of the merging parties own documents, including the Autotrader survey and the AutoTrader analysis;[46] and

 

179.3. the evidence from Mr Friedman that large traditional dealers benefit from having car-buying services as it provides them with an all-important range of stock. They are customers as they are unable to compete effectively with car-buying services for the purchase of used vehicle stock from private individuals.[47] In addition, WeBuyCars largest customers are dealerships.

 

[180.] Furthermore, based on the testimony of Mr Morgan and Mr Friedman and as set out by counsel for Weelee, we conclude that what sets a car-buying service apart from a traditional dealer is the geographic reach and scale of operations as a purchaser of high volumes of used vehicles. A car-buying service provides a centralised online market-place for the buying of used vehicles, where the scale and geographical reach extend beyond that which we believe a traditional dealer could achieve. These features distinguish traditional car dealers from car-buying services, which:

 

180.1.   focus on the purchase, for of a high volume of used vehicles;

 

180.2.   have a good reputation in the minds of the public as a car-buying service (i.e. a purchaser of high volumes of a variety of used vehicles). This good reputation seems to be established through substantial marketing and advertising spend on an ongoing basis through channels that reach substantial numbers of consumers across a wide geographic spread. This is key to the success of a car-buying service because the customer (i.e. private seller) must know that the car-buying service exists (and must be reminded of that fact at the time that he or she decides to sell his or her used vehicle). Without effective marketing, the car-buying service will simply not achieve the scale that it needs to survive;

 

180.3.   offer an online portal, which enables private sellers to approach the car- buying service and initiate the sale of their vehicle by loading the details of their vehicle onto the portal. The online portal significantly increases the geographical reach of the service (relative to the physical limitations which constrain most traditional         ability to secure the quality and variety of used vehicle stock which they seek). In other words, because the first step in the sale of a used vehicle by a private seller is undertaken online, there is no need for the seller to drive from dealership to dealership in the hope of finding a dealer who wishes to purchase his/her vehicle;

 

180.4.   offer speed and convenience. As stated above, WeBuyCars will give an indicative price and conduct a physical inspection of the vehicle (generally at the premises so the customer need not travel at all) and then seek to conclude the agreement of sale at the inspection. The vehicle is then loaded onto WeBuyCars website very soon thereafter. WeBuyCars boasts a 1-hour turnaround time on buying used cars from private sellers. Weelee [CONFIDENTIAL]. GetWorth also recognises the importance of quick turnarounds for quoting and ensuring that sales are concluded on acceptance; and

 

180.5.   the car-buying service must have a large customer base of willing buyers of the used vehicles that it sources from private sellers. Most car-buying services cannot afford to purchase high volumes of used vehicles and keep them in stock. Vehicles have to be offloaded quickly if the model is to work. It is for this reason that Weelee acts as an intermediary between customers and dealers. It also explains why WeBuyCars sold most of its stock to dealers in the first decade, or more, of its existence. It is only in recent years, as its scale and reputation has grown, that it has been able to sell vehicles directly to end customers in significant numbers. Notably, GetWorth has opted to sell its vehicles directly to the public from the outset (so as to enable customers to  for their cars and so that GetWorth can earn at least part of the margin which would otherwise be earned by a dealer selling the same vehicle to a customer), but that model requires significant investment in the selling side of the business, including in marketing and sales.

 

[181.] We also took into account the different attributes between traditional car dealers and car-buying service, inter alia, that traditional dealers:

 

181.1.   do not spend the money on advertising and marketing necessary to establish themselves as a car-buying service in the minds of consumers;

 

181.2.   do not have the systems and technology which are essential to their operation as such (i.e. an online portal); and

 

181.3.   lack the focus and incentive to make these investments because they generate most of their revenue through other channels.

 

[182.] We therefore conclude that traditional dealers do not fall within the same product market as car-buying services. At best they may provide a very weak constraint to car-buying services on the buying side of the market.

 

WeBuyCars market position in the guaranteed car-buying service

 

[183.] There were two approaches applied to assessing WeBuyCars position in the market.

 

[184.] The first was by assessing WeBuyCars share relative to other car-buying services in South Africa. Given the contrasting approaches to market definition, the Commission and the merging parties arrived at different market share positions for WeBuyCars.

 

184.1. According to the Commission if the market is defined as the market for car- buying services then the market share of WeBuyCars would be [>80%]. This market share was calculated by dividing the number of used cars that WeBuyCars purchases per month by the total number of used cars purchased across a number of competitors, namely Cars2Cash, cash4vehicles.co.za, CarZar, Sellmycar2day.co.za, WeBuyCars, Weelee, and Youdrivewesell.

 

184.2. The merging parties on the other hand sought to define the market broadly i.e. as a market in which all used car dealers compete, including traditional vehicle dealerships and car-buying services. Mr Murgatroyd estimated WeBuyCars market share using eNaTIS data on the number of monthly used car registrations. Based on the merging parties calculation, WeBuyCars would have a market share of [0%-10%] in this broadly defined product market.

 

[185.] The second approach to assessing WeBuyCars position in the market is in terms of the so-called The free float is the segment of used vehicles that is ultimately contestable by car-buying services, or the for car- buying services.

 

185.1.   To get to the free float portion of vehicle sales, the Commission excluded from the total number of used car transactions: (i) private transactions, (ii) trade-ins, and (iii) ex-rentals. Hodge found that WeBuyCars would have [0%- 30%] market share of the free float. However, Mr Hodge maintained that the current share of the held by WeBuyCars is not informative of the market power of WeBuyCars in a car-buying service market. That is better informed by the share of WeBuyCars amongst car-buying services, which the       expert report identifies as [>80%].

 

185.2.   On the other hand, Mr position was that WeBuyCars share of an appropriately calculated - is informative of the competitive position of WeBuyCars. According to Mr Murgatroyd on his estimation WeBuyCars would account for a very small percentage [0%-10%] of used vehicles bought and sold in South Africa. However, it is clear that the data used by Mr Murgatroyd in estimating this figure was not reliable. It double counted vehicles which are bought by dealers and then on-sold, as well as including repossessions and scrapped vehicles.

 

[186.] Given that we find that the relevant product market is one for car-buying services, and is limited to a few new close competitors who have not been able to gain traction in the market, we consider the market share estimate of position to be the most dispositive on the issue of market position. We do not consider the free float market to be a good indicator of WeBuyCars position as it underestimates the true market position of WeBuyCars.

 

[187.] We are therefore of the view that WeBuyCars could be considered to be dominant in the car-buying services market.

 

The removal of FCG as a potential competitor

 

[188.] We note that Mr Murgatroyd conceded that if we were to find a narrower market definition, then there could be a concern regarding the entry of FCG as well as portfolio effects. As Mr Murgatroyd stated:

 

MR MURGATROYD: needs to die in a ditch over precise market definition but ultimately both species of concern rely on WeBuyCars not being subject to effective competition currently. So, if the Tribunal to find that WeBuyCars is not subject to effective competition, then there would be a concern around the entry of FCG, assuming that the Tribunal also found that FCG would have entered. And similarly, if the Tribunal found that WeBuyCars is not subject to effective competition then key, the potential for some of the benefits that are conveyed through the merger to further entrench that position would be greater than if WeBuyCars was already subject to effective competition. And then the threshold one would need to apply to think about the benefits conveyed through the conglomerate facts, theories, would need to be high. I hope that clarifies from my perspective. [48] (emphasis added).

 

[189.] Case law indicates that in order for such a removal to warrant the prohibition of the merger, two conditions need be met:[49] The first is that there is reasonable certainty that the firm would have entered the market absent the merger. The second is that the entrant would have been considered a formidable competitor in the market, or would have significantly altered the competitive dynamics.

 

Was FCG likely to have entered the relevant market?

 

[190.] In March 2018 the Naspers Investment Committee Recommendation for FCG FCG     approved the recommendation to acquire 34.92% equity in FCG.[50]

 

[191.] In the Naspers FCG ICR, FCG is [CONFIDENTIAL]. [51] The recommendation constitutes one of a number of pieces of evidence clearly confirming not only that FCG intended to enter South Africa, but that these plans were extremely advanced. Indeed, [CONFIDENTIAL].

 

[192.] The recommendation established a post-investment operational plan which envisaged that [CONFIDENTIAL].[52]

 

[193.] It is common cause that FCG successfully entered Argentina and Columbia (the two other OLX operating countries specifically targeted for entry) under an FCG entity in October or November 2018 as reflected in in the initial plan.[53]

 

[194.] In terms of cooperation between [CONFIDENTIAL]

 

[CONFIDENTIAL] [54] (Underlining is our emphasis.)

 

Reference is made [CONFIDENTIAL].[55]

 

[195.] After the approval by the Naspers investment committee, a Heads of Agreement was signed on 10 April 2018 between OLX Global and FCG to begin the process of concluding a formal co-operation agreement. The purchase was fully concluded on or about 18 May 2018.

 

[196.] The Heads of Agreement set out that [CONFIDENTIAL].[56]

 

[197.] Mr Hodge advanced the argument that in addition to South Africa being included in the post-investment operational plan, the future proceeds from South Africa were included in the valuation model for FCG, including synergies. Naspers [CONFIDENTIAL].[57]

 

[198.] Public statements made by FCG and OLX indicated that, owing to large presence in South Africa, it was one of the next markets for FCG to enter. In an article published in the Financial Times, posted on website, it was reported that FCG: ady operates in six countries including Nigeria, Mexico, Chile and Pakistan. In Africa FCG had planned to expand to second-tier cities in $12bn second- hand car market, before taking its model to Ghana and Kenya. Because of strong presence in South Africa, it would now start there, the company said, opening car inspection lots in Johannesburg, Cape Town and Pretoria. The FCG model promises car vendors immediate cash and dealers the chance to build up [58] (emphasis added)

 

 

[199.] By the time of the conclusion of the Heads of Agreement, FCG had started taking active steps to implement some of its provisions.

 

199.1. [CONFIDENTIAL].[59]

 

199.2. [CONFIDENTIAL]. [60]

 

[200.] [CONFIDENTIAL].

 

[201.] At the end of May 2018, FCG and Naspers issued press statements that publicly announced their tie-up and         entry into South Africa within 4-6 months.

 

201.1. One article indicates that FCG will launch its still unnamed, new brand in Cape Town and Johannesburg in the next four to eight and that will be able to sell their vehicles at one of the business inspection centers FCG founder Mr Tyle was reported to be expecting to sell several thousand cars in the first year post-launch in South Africa It also anticipates that FCG would come up against others that are already operating this business model including WeBuyCars.[61]

 

201.2. A second article states that FCG aims to become the largest player in the second-hand car market in South Africa within the next two years [62]

 

OLX / FCG and WeBuyCars

 

[202.] Interwoven into the above narrative were a number of interactions between WeBuyCars and OLX / FCG. This must be considered against the backdrop of co-founder Peter statement that [CONFIDENTIAL].[63]

 

[203.] [CONFIDENTIAL].[64]

 

[204.] [CONFIDENTIAL].[65]

 

[205.] [CONFIDENTIAL].

 

[206.] Following the email regarding closing ranks, FCG and OLX proceeded to engage with WeBuyCars. Mr Tyle wrote to the van der Walts on 15 June 2018 indicating FCG s intention to take matters forward through OLX. To that end, he set up a meeting between Robin Voogd (head of M&A at OLX) and the van der Walts at WeBuyCars head office on 19 June.

 

[207.] [CONFIDENTIAL].[66] […].

 

[208.] [CONFIDENTIAL].[67] [….].

 

[209.] [CONFIDENTIAL].[68] [….].

 

[210.] [CONFIDENTIAL]. [….].

 

[211.] The following was agreed in the due diligence engagements

 

211.1.   In an attempt to allay the WeBuyCars very clear concern regarding information to be disclosed to OLX from the e-mail correspondence exchanged between the parties, it was agreed:

 

[CONFIDENTIAL] [69]

 

211.2. the inclusion of the following sentence in the agreements between OLX and WeBuyCars, with reference to the disclosure of the competitively sensitive data, that [CONFIDENTIAL] [70]

 

[212.] [CONFIDENTIAL].[71] [CONFIDENTIAL].[72]

 

[213.] Having considered the conspectus of evidence before us, as well as the testimony we heard, we are of the view that FCG was poised to enter the South African market but for the events discussed above. It is also clear that if the shareholders of WeBuyCars did not agree to sell WeBuyCars to OLX, they would face considerable competition in the South African market, in the form of an OLX/ FCG combination. [CONFIDENTIAL].

 

[214.] We wish at this stage to highlight certain glaring discrepancies between this analysis based on actual strategic and other documents of the interaction amongst OLX, FCG and WeBuyCars and a letter addressed by Baker & McKenzie, the legal representatives of the merging parties, to the Commission, during the course of the investigation and dated 29 April 2019:

 

[CONFIDENTIAL] [73]

 

[215.] This letter was sent to the Commission before the discovery of documents in these proceedings. If one reads the balance of the letter, it appears that Baker & McKenzie took an instruction on the various issues canvassed in the letter from, inter alia, Mr Nikkelen. We cannot but draw an adverse inference from the discrepancy between the real position vis-à-vis FCG and the contents of this letter. This inference affects the credibility of the merging submissions at a general level, and that of Mr Nikkelen in particular. The factual evidence has shown inter alia that [CONFIDENTIAL],[74] resulting in [CONFIDENTIAL].[75]

 

[216.] In this case, had it not been for the discovery process which revealed the true state of affairs, the Commission might have laboured under the serious misapprehension that FCG was completely at liberty to enter the South African market.

 

[217.] Given that one of the main theories of harm relates to whether the proposed merger results in the removal of FCG as a competitor in South Africa, misinformation of the nature contained in the Baker & McKenzie letter could have deleteriously affected the Co process and recommendation and ultimately hampered our ability to properly consider the true factual position in this case.

 

[218.] The circumspection with which we treated the evidence given by Mr Nikkelen extends to evidence from Mr van der Walt who professed not to know about the two-year non-compete clause, something which is entirely implausible when considered in the context of the proposed transaction as a whole. We similarly find it difficult to rely upon Mr van der Walt consistent refrain, when asked about the prospect of FCG entering the market, that he was not concerned about it and did not believe that FCG had any prospect of becoming a serious competitor to WeBuyCars. This does not explain the very deliberate attempts on the part of the WeBuyCars shareholders to prevent an OLX/FCG combination from entering the South African market and the eagerness with which the WeBuyCars shareholders pursued the proposed transaction with Naspers.

 

[219.] In terms of precedent, in determining whether a party was likely to enter a market absent the proposed merger, the Tribunal emphasised in the Du Pont[76] case that there must be proof of significant intentions to enter a market for the potential entry to gain traction. This emphasis was corroborated in Timrite[77] where the Tribunal was not persuaded that the mere presence of an agreement preventing parties from entering a market was sufficient proof of the intent to and ability to enter.

 

[220.] In Telkom/ BCX the Tribunal held that:

 

The future prospects of BCX cannot be assessed on the basis of pure conjecture but necessarily requires us to take into account the intentions of the company as articulated in its projections, its business plans, it internal high level documents, the views of competitors and where available third party industry sources and any unique qualities it may possess. [78]

 

[221.] The European Commission in its assessment of the Google/DoubleClick merger, considered  internal documents and past actions as persuasive enough to define it as a potential entrant.[79]

 

[222.] The Commission argues that in a merger context the relevant potential entry enquiry is a historical one, namely where, if the OLX-WeBuyCars merger negotiations had not occurred, would FCG likely have entered the market at that stage? The merging parties oppose this view and argue that the relevant counterfactual is a likely one if the proposed merger does not proceed.

 

[223.] Relying on the decision in SAB,[80] the merging parties held that the relevant counterfactual is a likely one in the sense of being the scenario that is probable in the event that the merger does not proceed. Ultimately the merging parties sought to advance the argument that     entry was unlikely because it had opted for the proposed merger as a method of entry qua shareholder, as opposed to de novo entry. The merging parties argued that FCG was unlikely to enter South Africa even if the merger were to be prohibited. However, Mr Tyle of FCG, who could speak to this issue, was not called as a factual witness by the merging parties.

 

[224.] Had the merging parties wished to undermine the evidence before us regarding the imminent entry of FCG into the market, they would have taken steps to secure Mr Tyle giving testimony before us. He is the only witness who could have spoken to the question of      likely entry into South Africa. However, as noted above, despite initially filing a witness statement for him, the merging parties did not put Mr Tyle up as a factual witness and withdrew his witness statement.

 

[225.] Nevertheless, an assessment of a theory of harm related to potential entry is a historic exercise, which contemplates the intentions of the potential entrant prior to the proposed transaction. This is the only way in which we can give effect to precedent which requires an examination of the intention of the potential competitor.

 

[226.] Within this framework, we find that the merging own strategic documents, including relevant documents of FCG,    public statements, email exchanges between WeBuyCars and OLX / FCG, and the factual evidence, all indicate that FCG was poised to enter the market in South Africa, and would have done so, but for the present transaction.

 

Was FCG set to become a formidable entrant?

 

[227.] There are several reasons why FCG was set to become a formidable entrant in South Africa as a car-buying service in competition with WeBuyCars. It had achieved success in the other developing markets it was operating in; it had access to capital and support through Naspers and OLX; on the buy side, it would have access to leads from OLX, and to a certain extent AutoTrader; and on the sell side, it would have access to AutoTrader. In addition, FCG, a firm which appears to have significant data analytics capacity, would have had access to both OLX and data on both vehicle specifications and listing prices, as well as on customer behaviour on their websites.

 

[228.] These factors, detailed further below, lead us to conclude that FCG would have become a formidable competitor to WeBuyCars had it entered the South African market. We discuss the facts with reference to the relevant jurisprudence on the issue.

 

[229.] According to the following statement from an OECD chapter on mergers: merger between an important competitor and an important potential competitor, especially if no other firms are similarly poised, could remove the competitive discipline provided by the potential entrant. Experience shows that the loss of potential competition is likely to be of greatest concern when a dominant domestic firm is acquired by an important multinational firm that produces the same product.[81]

 

[230.] FCG had a successful business model and was scaling up quickly in the markets that it was operating in. The Naspers ICR FCG acknowledged [CONFIDENTIAL].

 

[231.] The Naspers ICR FCG notes, [CONFIDENTIAL] [82] The other countries included [CONFIDENTIAL].[83]

 

[232.] More recent actual and forecast figures for FCG confirm its ability to roll out and grow rapidly in the markets in which it enters. The figures from the FCG financial model include actual performance to April 2019, with a May 2019 estimate and forecast for future dates beyond May. This shows a picture of strong growth in the markets in which FCG is operating.[84] In its first full calendar year following launch, i.e. 2017, FCG sold [CONFIDENTIAL] across six territories; that trebled to [CONFIDENTIAL] in 2018, with the trebling occurring in the existing territories (with FCG also then adding three new territories). The 2019 figures [CONFIDENTIAL].[85] The future projections [CONFIDENTIAL].[86] These are [CONFIDENTIAL].

 

[233.] FCG has access to working capital and dealer funding at favourable rates. [87] [CONFIDENTIAL]. [88] Furthermore, strong balance sheet would have been available to FCG for further growth and expansion.

 

[234.] When OLX invested in FCG, the synergies and cooperation between FCG and OLX were expected to, and indeed resulted in FCG incurring lower costs and scaling faster:

 

[235.] The post-deal assessment done by FCG in December 2018, which was only six months after final closing of the FCG/OLX deal, concludes that [CONFIDENTIAL]. The specific benefits achieved are listed as follows:

 

235.1.   Leads to purchases: OLX leads were [CONFIDENTIAL] of all car purchases in the JV and overlap markets. This number rose from [CONFIDENTIAL] in Indonesia.[89]

 

235.2.   Speed of rollout: FCG noted that territories that overlap with OLX had [CONFIDENTIAL].[90]

 

235.3.   [CONFIDENTIAL].[91]

 

[236.] In addition to the synergies that could have been achieved with OLX outlined above, entry into the South African marketplace would also have benefitted from additional synergies because of the position of Naspers in the South African economy. These include Naspers ownership of AutoTrader, an auto online vertical platform used by dealers to sell their stock to individual customers (B2C platform) and its ownership of a vast media footprint, including online news (24.com group), newspapers and magazines (Media24) and subscription television (MultiChoice, subsequently unbundled). Whilst the full extent of these synergies and the impact it would have on supercharging WeBuyCars are the focus of the portfolio theory of harm assessed below - we pause here to note that all the analysis conducted therein could be read to apply to FCG in bolstering the strength of its entry and the potential disruption to the market it could have brought.

 

[237.] The US DOJ Horizontal Merger Guidelines hold that:

 

The Agencies consider whether a merger may lessen competition by eliminating a firm, i.e., a firm that plays a disruptive role in the market to the benefit of customers. For example, if one of the merging firms has a strong incumbency position and the other merging firm threatens to disrupt market conditions with a new technology or business model, their merger can involve the loss of actual or potential competition. Likewise, one of the merging firms may have the incentive to take the lead in price cutting or other competitive conduct or to resist increases in industry prices. A firm that may discipline prices based on its ability and incentive to expand production rapidly using available capacity also can be a maverick, as can a firm that has often resisted otherwise prevailing industry norms to cooperate on price setting or other terms of competition[92]

 

[238.] In this regard the high margins earned by WeBuyCars (both in and of themselves, as well as relative to local and international competitors) are also worthy of mention. WeBuyCars margins average at [CONFIDENTIAL].

 

[239.] While the merging parties and Mr Murgatroyd attempted to downplay them, the Commission highlighted that WeBuyCars margins exceed the margins earned by in international counterparts. Auto1 [CONFIDENTIAL]. While Mr Murgatroyd criticised the comparison, the comparison was done by Naspers which believed it relevant when purchasing WeBuyCars.

 

[240.] Naspers modelled    margins at the lower end of the margin spectrum [CONFIDENTIAL] when compared with those of WeBuyCars, but this is still a very attractive level and significantly higher than that of existing competitors in South Africa. This speaks clearly to the anticipated success which FCG would have enjoyed had it entered the South African market.

 

[241.] This anticipated success would be all the greater when placed in the context of substantial growth even over the period of the merger assessment. Mr van der Walt submitted that, over the period of the merger assessment, WeBuyCars went from selling [CONFIDENTIAL] a month to [CONFIDENTIAL]. A growth rate of over [CONFIDENTIAL] despite comparatively high margins was a strong indication that, if FCG had entered the market, it would have enjoyed both short term success and longer-term growth at the expense of WeBuyCars.[93] In the Expert Joint minute on Free Float (the latest document to be submitted during the hearing) the absolute number of used cars purchased by WeBuyCars in May 2019 was listed as 5,882 vehicles.

 

[242.] Based on the above evidence, it is clear to us that Mr Tyle had good reason to believe that FCG would become the largest player in the second-hand car market in South Africa within the next two years. [94] This growth would have occurred at expense, i.e., FCG would have been a close competitor of WeBuyCars.

 

[243.] Mr Murgatroyd argued that for the Commission to succeed in proving that the proposed transaction would lead to the removal of an effective competitor, we would not only need to conclude that FCG would have entered the relevant market absent the merger, (which he conceded was a factual and not an economic issue)[95] but would also need to conclude that this entry was likely to materially intensify competition within the relevant market, and hence the competitive constraints to which WeBuyCars was already subject.

 

[244.] As set out above, we have found on the buying side that the relevant product market is that for car-buying services, of which WeBuyCars is the most successful example. We do not consider traditional dealers seeking to purchase vehicles as being a significant constraint on WeBuyCars on the buying side. Rather, the evidence before us has shown that traditional dealers are customers of car-buying services such as WeBuyCars. This means that the likely entry of FCG into the market would have had a powerful disrupting effect on WeBuyCars.

 

[245.] Furthermore, we have considered the evidence relating to how WeBuyCars, during the deal negotiations, responded to indications of entering the South African market and we found this evidence compelling.

 

[246.] Armed with the capital, experience from other emerging market economies and valuable know-how and technical capabilities, in our view, FCG would have presented WeBuyCars with very real competition, an experience which would have been largely novel to WeBuyCars.

 

Conclusion

 

[247.] For the reasons set out above, we conclude that:

 

247.1.   the proposed transaction would result in the removal of FCG from the South African car-buying services market as it would have entered the market absent the proposed transaction and that the likely effect of FCG would have had a powerful disrupting effect on WeBuyCars;

 

247.2.   The removal of FCG as a result of the proposed transaction would lead to a substantial prevention or lessening of competition.

 

[248.] In our view this conclusion cannot be cured by conditions tendered by the merging parties.

 

[249.] As the merging parties conceded early on in proceedings, Mr Wilson indicated that:

 

If the Commission were to prevail on both of the elements of the first theory of harm, in other words FCG would have entered and its market definition is upheld well then we would accept that the conditions that have been tendered in relation to non- discrimination would obviously not address that first theory of harm.[96]

 

[250.] The merger is prohibited on this ground.

 

CONGLOMERATE/PORTFOLIO EFFECTS

 

[251.] The second theory of harm being advanced by the Commission is based on portfolio effects. The theory analyses whether the complementary businesses of Naspers, including OLX and AutoTrader, but also the broader portfolio of Naspers media assets and technology assets, can be leveraged to entrench the existing dominance of WeBuyCars in the car-buying services market, while also providing reciprocal benefits to those Naspers businesses too.

 

[252.] Portfolio effects occur across different product or service markets and arise from the expansion of a       product or service offerings across complementary markets. The combining of different complementary products within a single firm can change the incentives of the merged entity to engage in exclusionary behaviour post-merger, for example by tying, bundling, full line forcing, exclusive dealings, cross-subsidisation, predatory pricing, and control of information.

 

[253.] The reason for this is that a market power exceeds the market power that it would have in each of the markets individually, so even if leveraging opportunities existed pre-merger, they would increase post-merger. This results in an increase in barriers to entry into the relevant market as a result of scale and scope economies, thereby reducing the likelihood that the merged entity would be significantly constrained post-merger.

 

[254.] With respect to mergers in a digital market, where big data can be a competitive advantage, the products and services would not even necessarily have to be close complements for this to occur, because the markets may be intimately related because of data.

 

[255.] Before proceeding to discuss the portfolio effects arising from this merger, it is necessary to first set out the context in which this concern was raised.

 

[256.] According to the Commission, WeBuyCars is already dominant in the car-buying services market. Furthermore, it is a first mover in this market and has experienced exponential growth in the market over the last 18 years. As WeBuyCars continues

 

to grow its retail presence, there is also likely to be a concentration of used car dealers, as small, independent dealers are driven out of the market. As submitted by Mr Hodge, there would be two avenues by which the retail side would likely tip in favour: starving the dealers of stock by engaging in exclusionary behaviour, or by advantaging WeBuyCars to accelerate growth, which would result in scale and network effects resulting in an unassailable lead. Our finding on conglomerate effects is based on the latter.

 

[257.] One of the central features of the car-buying service model is purchasing vehicles from the public. This largely depends on converting sales leads to purchases when buying vehicles from the public. According to the Naspers WBC ICR, [CONFIDENTIAL]. The number of purchases recorded is therefore positively correlated to the number of leads generated. Similar businesses in the market for car-buying services, such as Weelee and GetWorth, agree that attracting leads is a key function of their business and that this is often related to their overall marketing spend.

 

[258.] Marketing spend is critical to drive leads, which in turn drives purchases. This was apparent to the Naspers deal team developing the valuation model for WeBuyCars, which included marketing as a key expense item [CONFIDENTIAL]. Marketing effectively drives the growth model.

 

[259.] The Commission identified the following synergies which it submitted would arise from the proposed transaction (i) individual private car sales leads emanating from the OLX advertising platform to WeBuyCars; (ii) dealer sales leads emanating from website to WeBuyCars; (iii)       technology which, in combination with the data sourced from WeBuyCars, could be leveraged in favour of WeBuyCars; and (iv) Naspers other media platforms which could similarly be leveraged in favour of WeBuyCars.

 

[260.] The merging parties submitted that there were two fundamental failings with the conglomerate theory overall. The first was an incorrect finding that OLX, AutoTrader and WeBuyCars have significant market power in their respective markets. Secondly, that the alleged benefits are speculative and, to the extent that they do arise, are in fact pro-competitive.

 

[261.] Their main contention was that most of the conglomerate effects relied upon by the Commission for this theory of harm involved efficiencies and other pro-competitive effects such as lower prices, access to data, technology and capital integration. Mr Murgatroyd argued that theories of harm based on benefits, i.e., those that will enable the merged entity to outcompete its rivals, are commonly referred to as efficiency offences.

 

[262.] He submitted that concerns based on efficiency offences should be treated with a high level of caution precisely because they are presumptively prima facie pro- competitive. Further, that while economics recognises that the competitive advantage might potentially give rise to anti-competitive harm this, in his view, was only in exceptional circumstances where the competitive advantages are substantial and would give rise to unassailable competitive advantages.

 

[263.] We have already set out dominance in the car-buying market above. In what follows we turn to consider the market positions of OLX and AutoTrader and the potential synergies which may arise post-merger.

 

OLX and WeBuyCars tie-up

 

 

[264.] The concern raised by the Commission in relation to the integration of OLX and WeBuyCars was with respect to the potential for lead generation that the merger would create. In particular, the Commission was concerned that through the proposed integration of the two firms, private sellers would automatically be afforded an option to obtain a quote from WeBuyCars in the event that they were unsuccessful in attracting a private buyer through advertising on the OLX platform. Post-merger, WeBuyCars would therefore be able to scale its business even faster than it would have been able to pre-merger, denying opportunities to rivals. Further, that there are simply no other competitors who would be able to replicate an equivalent partnership to achieve the same benefits as WeBuyCars through the tie-up with OLX.

 

[265.] As mentioned above, OLX is an online horizontal classified advertising platform connecting buyers and sellers of goods and services through the publication of individual product advertisements for a broad range of goods and services listed on its website, including used cars. While user browsing and ad placement is free, OLX charges a fee to sellers for a premium placement of their listings on its website so as to distinguish it from other content on its website. A fee is also charged for banner advertising.

 

[266.] Where the merging parties and the Commission diverged was with regards to market share and how it should be estimated. According to Mr Nikkelen, the vehicle category of OLX is relatively small and does not account for a substantial volume of the traffic. In addition, he does not consider OLX to be a relatively large participant in the generalist online horizontal classified advertising market in South Africa and, in the vehicle category, submits that they compete against other automotive classifieds.

 

[267.] However, the Commission contended that OLX South Africa holds value for Naspers, particularly with respect to the car ecosystem that it is building. Given that the guaranteed purchase used car marketplace purchases used cars primarily from private sellers, the Co expert considered what proportion of private sellers list their vehicles on the OLX website, as compared to its competitors. For this he relied on the following excerpt from strategic plan.

 

[EXTRACT CONFIDENTIAL]

 

[268.] According to the extract above, of approximately [….][CONFIDENTIAL], [CONFIDENTIAL]. Its closest competitor has only [CONFIDENTIAL] out of a total of [CONFIDENTIAL] on its website. Across all the different online vehicle marketplaces, the Commission found that OLX has [CONFIDENTIAL] of the total private listings.

 

[269.] There were several reasons why Mr Murgatroyd disagreed with the above approach.[97] The first was that the Commission appeared to deviate from how it had defined the relevant market in its merger report. According to the merger report, the Commission had defined the relevant market as the market for the provision of automotive classified services. In its expert report, the Commission had made a further delineation, defining the market as the market for the provision of automotive classified services to private sellers. That being said, Mr Murgatroyd main issue was with the reliability of private listings as a probative indicator of a market power. This is because there can be significant overlap across platforms i.e. individuals are able to list on generalist platforms at no cost, and therefore barriers to switching between advertising platforms are virtually non- existent.

 

[270.] Further that the estimations of market position (and AutoTrader) were significantly overstated given their reliance on site visit information provided by third parties to the Commission. Using Similarweb data provided by the merging parties instead, and including both generalist and specialist online platform visits, Mr Murgatroyd estimated market share to be [CONFIDENTIAL].[98]

 

[271.] Based on this estimation, Mr Murgatroyd was of the view that this was a good indication of    competitive weakness in the online classifieds market in South Africa. Further, while he acknowledged that there would be some synergies that would arise through the integration of the two firms, he disagreed that WeBuyCars would attain an unassailable advantage in the market.

 

[272.] What then is the nature of this benefit? It is common cause that private sellers on OLX (and other classified portals) would prefer to obtain a retail price for their vehicle notwithstanding the associated disadvantages with the process such as inconvenience, safety and trust. This is because they are generally unwilling to sell their vehicle to a dealer at a wholesale price.

 

[273.] As Mr Nikkelen put it:

 

 prospective seller faces a trade-off in this regard, between (i) on the one hand, the increased inconvenience and certainty (as well as reduced risk) of selling to a dealer or trading-in, but at a potentially lower price, and (ii) the greater administrative burden and risks of selling their vehicle privately, but at a potentially higher[99]

 

[274.] Mr van der Walt also confirmed in his evidence the unwillingness of private sellers to sell their vehicles to dealers. He also suggested pricing would not be attractive to these private sellers:

 

my experience, potential sellers of second-hand vehicles that elect to pursue private sales do so because they are aiming to extract a higher purchase price than they would be able to get from dealers or when trading-in a vehicle at the time of purchasing a new vehicle, despite the additional time and effort to do so. Although WeBuyCars prides itself on offering a competitive and reasonable purchase price and convenience to potential sellers, I do not believe that our pricing would be attractive to these private[100]

 

[275.] The purpose of the OLX and WeBuyCars tie-up, however, would be to appeal to private sellers who have grown tired on the OLX platform and might be willing, after a particular period of time of advertising privately, to accept a wholesale price for their vehicle from WeBuyCars. As Mr Nikkelen explained:

 

believed that it might be attractive to potential sellers to be able to link from OLX to [WeBuyCars] for a quote if they did not succeed in selling their vehicles privately on OLX after a certain period of time. As noted above, individual sellers are prepared to incur the administrative hassle of trying to sell their vehicles privately in the hope of a higher price, but might be willing to sell to a dealer if their private efforts are[101] [….].

 

[276.] According to the merging parties, the transaction would also allow them to compete more effectively against Facebook Marketplace and, ultimately, improve the user experience on the OLX platform, rather than drive incremental business to OLX. The incremental business to OLX would be a secondary benefit. Further, there would be no benefit to WeBuyCars. Mr van der Walt stated:

 

WeBuyCars is willing to explore this strategy, but it does not believe that it will bring any particular benefit to WeBuyCars, and it is not one that WeBuyCars actively wishes to pursue. WeBuyCars already generates significant leads to purchase second-hand vehicles from its own systems (i.e. its website, call centres and physical locations) and it does not believe that OLX would be able to add materially to those[102] […].

 

[277.] However, as discussed above, the purchaser, OLX, clearly saw competitive benefits resulting from a tie-up, as explained by Mr Nikkelen.

 

[278.] Aside from the issue of generating leads, a further issue which was raised by the Commission related to the marketing spend. There was consensus from Mr Morgan, Mr Friedman and Mr van der Walt that a significant amount has to be spent on marketing to attract potential private car sellers to a car-buying service. According to the Commission, own management forecast, produced for Naspers, included an assumption of marketing cost savings of [CONFIDENTIAL].

 

[279.] As regards the quantification, Mr Nikkelen pointed out, with reference to the section in the Naspers WBC ICR, [CONFIDENTIAL].

 

[280.] Mr Nikkelen explained that OLX subsequently realised that even that modest synergy had been materially overstated as a result of two factors. The first [CONFIDENTIAL]. The second [CONFIDENTIAL].

 

[281.] Mr Nikkelen explained that, the total number of listings on the OLX website are not reflective of private sellers. This is because, as he suggested, dealers often pose as private sellers so as to avoid being charged on the OLX website for their listings. Further that because of this the OLX strategic plan above was not an accurate reflection of the private listings and that this amount in reality was much lower. In order to substantiate this claim Mr Nikkelen put up exhibit which calculates how many sellers have had more than one live listing in the cars and bakkies category on OLX over the last three months. Exhibit        is the subject of the Discovery Application brought by the Commission during the proceedings and is addressed in Part C of these reasons.

 

[282.] According to Mr Nikkelen, this exhibit shows that private sellers constitute a small percentage ([CONFIDENTIAL]) of the total listings in the cars and bakkies category. However, Mr Nikkelen acknowledged that there might be certain anomalies in the data, for example because a true private lister has created more than one identity for the same vehicle, and on that basis has incorrectly been counted as a dealer in exhibit W. However, Mr Nikkelen said there were procedures within OLX in order to capture such behaviour, and that it is normally not associated with true private listers, but rather with dealers.

 

[283.] The Commission sought to verify these results using a sample of the users that Mr Nikkelen claimed are not true private sellers. It found that a significant number of those users identified as not true private sellers were indeed private sellers, which implies that the methodology used by Mr Nikkelen was not reliable. It appeared that it incorrectly included true private sellers in this list of alleged by capturing those that had re-posted their initial listing with updated information, or who were concurrently attempting to sell other items that are not vehicles.

 

[284.] However, the Commission submitted that even if this evidence is accepted, it still shows that roughly [CONFIDENTIAL] per month on estimations are from sellers with one listing which Mr Nikkelen sought to classify as private sellers. This puts private listings at twice those of Gumtree which according to strategic plan above is in the region of [CONFIDENTIAL]. This is also before a consideration that the same grey dealer issue is likely to affect the Gumtree number too. As such, the Commission contended that this does not change the conclusion that OLX has the most private listings by some margin.

 

[285.] The Commission furthermore submitted that the sheer size of online private listings led the deal team to identify that [CONFIDENTIAL].

 

[286.] This brings us to one final argument put forward by the merging parties, being that any rival of WeBuyCars would be able to replicate this potential tie up with any of rivals, such as Gumtree. According to the merging parties, all platforms are capable of offering private listings, and there are a number of platforms that attract significantly more website views than OLX. In conclusion, the merging parties were of the view that any assistance that OLX derived from integration with WeBuyCars was likely to be modest at best, and would simply enable OLX to compete more effectively with its rivals in the market, in particular Facebook Marketplace and Gumtree. They argued that there was no evidence that any such benefit would somehow afford OLX an unassailable advantage in its market.

 

[287.] However, as counsel for Weelee indicated, aside from the fact that Gumtree is likely to be a far less rich source of leads since it is not a specialist platform, there is no obvious incentive for Gumtree to enter into such an arrangement, absent the payment of money to it. There would therefore be a substantial cost to the car- buying service associated with any such arrangement. Further, no other car- buying service is big enough to be able to afford such an arrangement at present. Finally, an arrangement between Gumtree and an external car-buying service would require a high level of integration with        systems in order to target only tired sellers and it is not clear that Gumtree would entertain such an arrangement.

 

Our analysis

 

[288.] It is trite that when assessing portfolio effects, it is not necessary to establish dominance in each of the relevant markets. What is required is an assessment of whether the proposed merger increases the market power that it would not otherwise have in each of the markets individually pre-merger; and whether this is replicable by competitors.

 

[289.] What is clear from the evidence is that the size of private listings on OLX, relative to alternatives that might be utilised by rivals of WeBuyCars, is significant. This much cannot be disputed from the evidence of private listings. In our view there appears to be a difference between generalist online platforms such as Gumtree and specialist online platforms, geared specifically to used car sales. As such Mr inclusion of generalist platforms in his calculation of        market position is incorrect and dilutes likely market position. In our view the market share estimation of [50%-60%] is a more reliable indicator of probable market position. In addition, strategic plan replicated in the extract in paragraph [267.] above shows that OLX does have a leading market position.

 

[290.] Given the relative size of OLX, WeBuyCars will be in a position to leverage off market position in private listing offerings in the following ways. The first is that by integrating the platforms, WeBuyCars will be able to generate leads from OLX by being able to target tired sellers more directly in a way that its rivals will not be able to replicate. Customers (i.e. private sellers of used vehicles) will not need to necessarily go through the system of uploading their vehicle on the WeBuyCars portal. The systems will be integrated, meaning that a customer by listing his/ her vehicle the OLX website, could receive a quote from WeBuyCars automatically.

 

[291.] The leveraging benefit to WeBuyCars is also supported by deal team calculation which showed that converting [CONFIDENTIAL] of the OLX leads to WeBuyCars will generate an additional [CONFIDENTIAL] per month for WeBuyCars. When one compares this to a firm such as GetWorth which is only able to purchase approximately [CONFIDENTIAL] in a month, these figures stand in stark contrast and emphasise the scale at which WeBuyCars will be able to grow with a WeBuyCars OLX merger. The same can be said for Weelee which facilitates the sale of roughly [CONFIDENTIAL].

 

[292.] In our view this transaction would therefore supercharge WeBuyCars in further entrenching its dominant position as a car-buying service in South Africa. There is also no clear evidence that any rival of WeBuyCars would be able to replicate such a tie-up in the market in competition with WeBuyCars. Moreover, there is currently no effective competitor to constrain the market power which WeBuyCars enjoys.

 

[293.] It is for these reasons that we are of the view that there is sufficient evidence to show the competition harm that may arise through the integration of the OLX and WeBuyCars platforms post-merger.

 

AutoTrader and WeBuyCars tie-up

 

[294.] Another area of concern related to the tie-up between AutoTrader and WeBuyCars. In this section, we first set out the market position of AutoTrader and its importance to dealers, before proceeding to consider the effect on competition that will flow from a combination of WeBuyCars and AutoTrader. The first issue being what is referred to as potential post-merger data and the second relates to [….].

 

[295.] AutoTrader is a household name in South Africa which was acquired by Naspers in 2017. It is a specialised online advertising portal which, on the one hand, allows users to browse classifieds exclusively for new and used vehicles and, on the other, serves as an advertising platform through which car dealerships,  or private motor vehicle sellers can advertise vehicles for sale online, for the potential purchase by end users of the online platform.

 

[296.] Sellers of vehicles that want to advertise their vehicles on the platform are required to pay a subscription fee, based on the number of vehicles to be listed and listing features. AutoTrader also sells banner advertising on its website, which is primarily purchased by and by car insurance and finance providers. Through Autofuzion, AutoTrader sells reports, statistics, valuations, transactional data and other analysis to assist and guide dealers to improve the quality of their listings, to upload their listings and to have a pricing advisory tool. AutoTrader also has a presence in print media selling two print publications: Commercial Trader, advertising the sale of commercial vehicles; and Agricultural Trader, advertising the sale of agricultural vehicles.

 

[297.] According to the TNS Dealer Survey[103] results prepared on behalf of AutoTrader in September 2016, AutoTrader is the largest and most well-established source of used vehicle leads in South Africa . [104] The survey suggests that AutoTrader dominates the market, in spite of the availability of a number of competitors, and in spite of the fact that AutoTrader has the highest monthly fees in the market.

 

[EXTRACT CONFIDENTIAL]

 

[298.] From the above, it would appear that Cars.co.za stands out as the key competitor as it is the most like AutoTrader in terms of having a large database as well as a trusted and reputable third-party dealership network. [CONFIDENTIAL]. The growth of Cars.co.za is also evident in the investigation, as revenues appear to have grown significantly from 2017 to 2018.

 

[EXTRACT CONFIDENTIAL]

 

[299.] Based on the above estimates,     major competitor is Cars.co.za. These two companies represent the dominant online platforms for consumers, with AutoTrader having [30%-40%] market share, Cars.co.za having a [30%-40%] market share and Carfind.co.za a distant third, with [0%-10%].

 

[300.] According to GetWorth, automotive online classifieds are crucial for selling cars to the public and AutoTrader and Cars.co.za account for the bulk of leads generated. AutoTrader alone accounts [CONFIDENTIAL] of the customer leads for sales of used vehicles. most recent lead figures put AutoTrader at [CONFIDENTIAL] of leads in any single month.

 

 

[301.] The Commission also submitted that the relative market positions are reflected in the prices charged by the different platforms, with both AutoTrader and Cars.co.za charging around R200 per 1 000 listings, compared to CarFind at R69 and AutoFind at R40.

 

[302.] AutoTrader itself claims that [CONFIDENTIAL] [105]

 

[303.] It is common cause between Mr Hodge and Mr Murgatroyd that AutoTrader has a strong market position in South Africa for used car listings by dealers. The point of contention was whether this position allowed AutoTrader to act independently of its competitors.

 

[304.] Mr Murgatroyd submitted that AutoTrader would not be able to act independently of competitors for two reasons. The first is that dealers have an alternative, in Cars.co.za, to whom they can turn if AutoTrader were to treat them unfavourably as a result of the merger.

 

[305.] The second is that the degree of market power that a vehicle listing platform has is dependent on the extent to which leads are unique to that platform and there is very little evidence to suggest that the majority of leads that AutoTrader generates are exclusive to AutoTrader. Mr Murgatroyd argued that the fact that [CONFIDENTIAL] (as per Mr   evidence) of customers multi-home on AutoTrader and Cars.co.za is indicative of the fact that the leads AutoTrader generates are unlikely to be unique to AutoTrader.

 

[306.] The Commission submitted that the evidence on dealer behaviour strongly suggests that dealers consider both AutoTrader and Cars.co.za to be must-have platforms. Dealers are willing to pay both platforms to advertise vehicles for sale instead of relying on their own websites and multi-homing brings in sufficiently more leads for the dealers to justify the cost.

 

[307.] Mr Morgan stated that he could not simply leave AutoTrader as he sources more than [CONFIDENTIAL] of his leads from AutoTrader and these leads would be very costly to replace. Interestingly, Mr Morgan went on to suggest that it was his belief that the leads generated from AutoTrader and Cars.co.za [CONFIDENTIAL]. He stated that:

 

[CONFIDENTIAL] [106]

 

Conclusion

 

[308.] AutoTrader is a leading player with estimate market shares at 37%. AutoTrader, in its own document referred to in paragraph [297.] above, considers itself to be dominant.

 

[309.] The evidence of Mr Morgan was that dealers view AutoTrader as must platform. This finding is borne out by the evidence on the cost of listings. If one compares the cost of listing on AutoTrader and Cars.co.za for example these two platforms are able to achieve much higher rates which also reflects their importance to dealers and the quality of their listings.

 

[310.] From a dealer perspective it is clear that the volume of leads and quality of leads are the most important considerations for dealers when choosing between platforms, hence the choice to pay more and be able to access platform.

 

[311.] This factual evidence would suggest that, contrary to Mr contentions, AutoTrader appears to be a must have platform for dealers and that any suggestion that dealers could easily replace the leads from AutoTrader with a different platform is purely speculative. To the contrary, Mr Morgan indicated that [CONFIDENTIAL], despite AutoTrader charging significantly more than the third and fourth largest competitors.

 

[312.] Having set out the above on the importance of AutoTrader in the market in general, we next discuss the effect on competition that is likely to arise from the combination of WeBuyCars and AutoTrader. More specifically we discuss the issue of self- preferencing and data.

 

Self-preferencing

 

[313.] One of the theories of harm advanced by the Commission in this case is that a collaboration between WeBuyCars and AutoTrader may result in self-preferencing. This concern arises from the fact that where a listing is placed on the site has a significant impact on the leads it generates.

 

[314.] Mr Hodge evidence indicates that AutoTrader pre-merger, already engages in preferential behaviour through the use of its different listing packages. He points out that AutoTrader has not lost dealers despite this practice.[107]

 

[315.] Mr Murgatroyd disagreed with this contention, suggesting that it is difficult to read anything into the issue of self-preferencing as it stands, given that a premium listing (which allows for a more preferential position on the website) is available to any dealer wishing to pay for it. He further pointed out, however, that the commercial feasibility of engaging in such conduct on two-sided platforms like AutoTrader is subject to a critical proviso, namely that it does not degrade the consumer experience and does not cause dealer disaffection.

 

[316.] In this regard, Mr Mienie indicated that self-preferencing would be futile because, if the consumer experience on the AutoTrader platform is degraded, consumers will leave the site, and then so will dealers. Likewise, if dealers leave the site, so will consumers.

 

[317.] He explained:

 

 a company that makes money out of preferring dealers. the way we generate additional revenue is to prefer dealers, one dealer over another, explain the premium and the standard TO and the stand-outs. And predicated on the fact that you need scale in order to do that, but  one caveat, if we prefer a particular customer and we start to hurt consumer experience, we start to hurt what the consumer sees or wants to see, we very quickly lose the trust and confidence of that consumer, which takes a long time to build up. We lose the trust and confidence of the consumer, because they start to see things that they want to see, they start to see things that you are forcing in front of them in their[108] [….].

 

[318.] Regarding dealer disaffection, Mr Minnie pointed out that dealers require AutoTrader (and other advertising platforms) to be fair and non-discriminatory in its dealings with them. He said any preferential treatment would not be acceptable to dealers because it would be discriminatory treatment of a firm that other dealers regard as a direct competitor of theirs. As he explained, essentially eaten the lunch of the car dealers in the trade-in. what [109] [….].

 

[319.] Accordingly, the merging parties contend that in order for AutoTrader to provide the services required by its dual customer base, and to compete effectively with rival advertising platforms, it must ensure that it advertises as many vehicles, from the widest variety of sellers, as possible on its platform. In this way, more purchasers are attracted to its site, which in turn attracts more sellers to its site, and so on, in a positive network loop.

 

[320.] Given this essential feature of AutoT business model, the merging parties submitted that it would therefore be entirely self-destructive for AutoTrader to engage in any strategy that would have the result of reducing the number of vehicles advertised on its platform, because that would disincentivise both its purchasing customers, and in turn its selling customers, from using AutoTrader as an advertising platform.

 

[321.] Further, Mr Mienie stated that WeBuyCars only has approximately [CONFIDENTIAL] of the stock on AutoTrader and therefore:

 

[self-preferencing] would damage the consumer experience. Because if you can imagine, if you have enough stock volume, if only got a couple of thousand cars available, the chances that you can crowd out the front page is hard to do, just because you have sufficient stock of that make, model, variance, price parameter, that the consumer has chosen. There are 15 listings on the first page. Now, the ability to crowd out that first page would be very difficult in consumer is going to sit back, bad consumer experience, and say, well bouncing, going somewhere else"[110]

 

[322.] [CONFIDENTIAL].

 

[323.] Further, Weelee submitted that preferencing vehicles in this way will have no noticeable effect on consumer experience, but the consistent application of the strategy across all, or most, search results will clearly bring a substantial benefit over other dealers for the simple reason that no other dealer will benefit from appearing, consistently, on the first page of all search results. Moreover, the effectiveness of the strategy will grow exponentially as WeBuyCars continues to acquire ever greater volumes of the available used car stock in the country.

 

[324.] There is no dispute that AutoTrader is the largest source of used vehicle leads in South Africa and is the leader in terms of website visits and influence on customer choice, and is a platform for dealers, in spite of its higher pricing.

 

[325.] We do not have to conclusively decide whether the merged entity would engage in self-preferencing as this is not a self-standing theory of harm. Rather, it is the cumulative effect of the various synergies that we consider to likely result in a substantial prevention or lessening of competition.

 

The combination of AutoTrader and WeBuyCars databases

 

[326.] In data driven mergers, a typical theory of harm which arises is the exclusionary effect from the combination of the databases of the merging firms. This is because these datasets need not be in the same market, as access to multiple data sources can improve the overall quality of the database and contribute to greater economies of scope. According to the Commission, this unique combination provides the merged entity with an advantage over competitors to improve on its offering in a way that cannot be matched.

 

[327.] According to the merging parties, AutoTrader already makes available all dealer and consumer data that it believes to be relevant and valuable to its dealer customers through its AutoFuzion product. In this regard, Mr Mienie explained that AutoFuzion provides a tool for uploading and analysing stock on AutoTrader. It also offers pricing and stocking tools, which enable dealers to view list pricing and stocking trends of similar vehicles from competing dealers to assist them with their pricing and stocking decisions. Mr Mienie also explained that other vertical classifieds platforms such as Cars.co.za have access to the same dealer and customer data as AutoTrader.

 

[328.] Further, from a WeBuyCars perspective, WeBuyCars does not purchase data from any third-party sources, including (publicly available) Autofuzion product. Mr van der Walt is also not aware of any data AutoTrader might have that would be of benefit to it. This is because WeBuyCars regards its own actual sales data as more comprehensive, accurate and relevant than data from AutoTrader (or any other advertising platform).

 

[329.] Mr Nikkelen likewise testified that richness of the data set that [WeBuyCars] sits on is really so the value of being handed potentially smaller and poorer data sets from platforms such as AT and OLX would be […].

 

[330.] On the basis of both Mr van der Walt and Mr submissions, Mr Murgatroyd submitted that the incremental benefit of the data from AutoTrader to WeBuyCars was not likely to be substantial.

 

[331.] Finally, the merging parties also suggested that competitors would be able to achieve the same outcome (gain the same benefit) through a partnership with cars.co.za, a rival of AutoTrader.

 

[332.] As already noted above, AutoTrader already sells some of its data, in packaged form, in a product called Autofuzion. In his testimony, Mr Mienie was adamant that AutoTrader was already monetising all the useful data in its possession and that, if there was more useful data available to it, it would monetise that too.

 

[333.] However, during cross examination Mr Mienie conceded that this was not entirely true for two reasons. The first is that Autofuzion provides historical information to dealers concerning the demand for, and list pricing of, vehicles and it gives no insight into the current position in the market. Accessing that data, directly, would clearly benefit a dealer because they would be able to price according to the current market position rather than the historical position. Secondly, AutoTrader does not monetise its personal details or their behaviour on website, such as the pages they accessed and the like.

 

[334.] According to Weelee, it is important to put this into perspective. More specifically, there are some 30 000 active AutoTrader users, who have conducted some 25 million searches and opened up nearly 40 million website pages. The mine of information to which AutoTrader has access is therefore very substantial and would provide unique insight into the preferences and buying patterns of customers. AutoTrader has this information in its possession. In Particular, it changed its privacy policy to make clear to its customers that they would be tracking their behaviour and keeping that information when AutoTrader became part of Naspers and thus gained access to Google Analytics. The merged entity would accordingly be able to benefit from access to all of data.

 

[335.] According to Weelee, the significance of this is therefore two-fold. First, the majority of people looking to buy a car are probably looking to sell one too. Therefore, this information would permit targeted and perfectly-timed advertising of services to customers that are likely to be in the market to sell their vehicle. Second, when a customer purchases his/her next vehicle, the sale of his/her current vehicle is likely to become more urgent, relatively speaking. Marketing on the sales side (for example, sending a targeted advertisement to a user advertising the sort of vehicle they are known to be interested in buying) will likely increase their need to sell their current vehicle. This marketing, combined with the targeted marketing of services, will likely generate a substantial number of leads for WeBuyCars, far more quickly than is currently the case.

 

[336.] Mr Hodge succinctly summarised the key factors that should be considered when assessing the likely effect that could emanate from the combination of WeBuyCars database with the AutoTrader database. He submitted that there were three questions to ask, namely; (1) what data does AutoTrader have?; (2) how much of this data does AutoTrader provide through the Autofuzion platform?; and (3) how useful is this data to WeBuyCars?

 

[337.] To assist in the debate during the hot tub, the economic experts broadly classified the data of interest to the transaction across two types of data, namely; (i)

and (ii) In what follows we will consider the arguments in relation to the usefulness of this data.

 

[338.] With respect to the Commission submitted that WeBuyCars would benefit from the range and quality of data. First, data contains much more detail than the basic transactional data that WeBuyCars stores on its transactions. In particular, AutoTrader gathers data for almost the full universe of dealers and actual current listing prices which reflect the current competitive pricing in the market.

 

[339.] Second, Naspers conducted a due diligence on WeBuyCars which found that WeBuyCars had poor technical capabilities and was incapable of providing an automated price response to the customer. The pricing function is generally complex because of the many variables that have to be considered. There are a variety of methods of determining pricing, but all depend on information from the market. A larger database set assists in reducing the uncertainty associated with pricing in this business. Mr van der Walt also accepted that the TransUnion data, which many dealers rely on, will not give the buyer an accurate picture of the actual value of the car. WeBuyCars only recently started collecting data relating to the sale of used vehicles across most of the spectrum of makes, types and models of vehicles. However, as Mr van der Walt indicated under cross-examination, the W data is currently flawed insofar as it does not currently capture inspection data, which would explain why the same vehicle may fetch quite different prices.

 

[340.] Mr Murgatroyd contended that the data that WeBuyCars would have access to, as a result of the merger, may only be incrementally more useful than Autofuzion. However, Mr Morgan did not agree. Mr Morgan explained that vast volume of data uploaded by dealers is highly detailed and contains information on the make, model, year and mileage of cars, but also the location, colour, specifications, additional features, the     name, list prices and the amount of time that the listing has spent on AutoTrader. Importantly, AutoTrader does not provide Vehicle Identification Number (VIN) data through the Autofuzion platform. This is consistent with Mr Morgan evidence. Mr Morgan testified that he had found, in running GetWorth, that there are two types of data, the and the The fumes [CONFIDENTIAL].[111]

 

[341.] In support of Mr contention that AutoTrader has valuable data, it was revealed through discovery and cross-examination that AutoTrader generates valuable data from dealers. Unlike private sellers, it was submitted that dealers registered on the platform through sales representatives. During this registration process, sales representatives collect other information from dealers such as the VIN number, variant, registration year, manufacturer colour, optional extras, photos of the vehicle, condition of the vehicle, and trade-in price. It was further revealed in the proceedings that AutoTrader has built a VIN database unmatched by any other firm in South Africa.

 

[342.] In addition, vehicle check site provides information by VIN number on the car manufacturer, transmission type, the year the car was introduced, the registration number of the vehicle, the fuel type, the warranty start and end date, the colour of the vehicle, the engine number, the body shape, the model type, the engine size, whether the car has been stolen, the vehicle accident history, the mileage history and the financing history.

 

[343.] While AutoFuzion was designed to help dealers manage stock, with faster stock turnover being a key measure of success in this industry, it does not match the level of detail that WeBuyCars would have access to post-merger, as WeBuyCars would have the view of the whole market. Furthermore, the AutoFuzion data is not in a form or shape to assist a car-buying service to price a high volume of vehicles with very different characteristics quickly and efficiently. Car buyers need the data to be presented in a form that is useful to them and not in a format that AutoTrader may consider to be useful to a dealer.

 

[344.] Customer data was loosely defined as falling into two categories, namely; (i) customer information used for targeted marketing purposes; and (ii) customer behavioural data capturing their journey on a particular site.

 

[345.] In terms of customer information used for targeted marketing purposes, the Commission submitted that it was important to highlight that it is currently possible for AutoTrader to gather the contact details of any user enquiring on a car through the platform, whether they have registered or not. According to Mr Mienie, data from registered users is quite large, standing at [CONFIDENTIAL] and growing by [CONFIDENTIAL] users per month. Using the consumer behaviour data and contact lists of AutoTrader, the Commission submitted that WeBuyCars will be able to engage in targeted advertising. This is very valuable data to car-buying businesses. AutoTrader has consumer behaviour data by IP address in the region of [CONFIDENTIAL] per month.

 

[346.] Turning to customer behavioural data, this would be information AutoTrader collects during and after a voluntary registration process that a user will opt into if he or she wishes to access additional functionality on the AutoTrader platform (such as favourite vehicles, the ability to privately list vehicles for sale, the ability to subscribe to the AutoTrader newsletter or to subscribe for vehicle alerts). This platform user behavioural information data, collected by Google Analytics, includes the number of visitors, number of sessions, the average session duration, the number of pages viewed in a session, demographic information, brands and models searched for, price ranges searched within, locations searched in, as well as filters used when searching on the platform.

 

[347.] According to Mr Hodge, data from AutoTrader could enable WeBuyCars to understand customer behaviour on a deeper level and engage in targeted marketing when customers are making decisions regarding their purchases. Further, WeBuyCars would be able to improve the quality of its products and services, increase efficiency, exploit new business opportunities and be a more target-oriented business. The Commission therefore concluded that the cost of generating leads and acquiring customers would be much higher for competitors post the merger when compared to a WeBuyCars with access to data. Sharing data between unrelated companies is, at best, difficult from a legal and risk perspective. Moreover, what is problematic is that in order to gain the benefit of the association, a competitor would likely have to pay a rival (such as cars.co.za.) for this information and that party would have to agree to a full integration of its systems with those of the car-buying service.

 

[348.] WeBuyCars would also be able to price more competitively and would be in a much better position to gauge demand for specific vehicles than any of its competitors.

 

Conclusion

 

[349.] It is apparent that the increment of data to which WeBuyCars would gain WeBuyCars access through AutoTrader, vastly exceeds the data that WeBuyCars currently stores. Indeed, Mr van der Walt acknowledged that there are limitations to database. The key missing element that is relevant to the pricing function of a car-buying business is market data, which is precisely the type of data that AutoTrader collects.

 

[350.] AutoTrader has amassed a significant amount of data in respect of the vehicles sold through its platform. The size of this data set is particularly staggering when one considers the numbers of sellers from whom this information is collected.

 

[351.] The European Report on Digital Markets, in the context of examining competition between large established platforms and new entrants, argues that:

 

 platforms possess a strong competitive advantage over new entrants because of, inter alia, network externalities and privileged access to data. Often this advantage will translate also into advantages in related[112] [….].

 

[352.] This sentiment can also be distilled in the European     discussion around the nature of the data which larger platforms would have access to:

 

to data can represent another form of competitive advantage and is differentiated among Exclusive access to a large amount of individual-level data, even when accessed anonymously e.g. to be used by a machine-learning algorithm, might provide a competitive advantage. A superior algorithm might then attract users who would contribute more and/or fresher data on an ongoing basis, thereby reinforcing the advantage.

 

Access to non-anonymous individual data including e.g. both historical and ongoing observed data may often be necessary to offer a service to an individual or would greatly increase its quality. In the context of platforms, data might furthermore have a positive feedback loop where control over some of an data increases the ability to collect more of [113]

 

[353.] In his testimony, Mr Friedman made a further point regarding how through the merged entity, WeBuyCars would be able to extract maximum advantage out of the data: through its access to FCG which would enable it to mine the data. He referred to infrastructure as like a plug and play for technology He articulated his fundamental concern with the proposed transaction as follows:

 

if you take 30 years of data from AutoTrader and data on 60 000 cars a month from OLX and combine it with      data and then throw that into a machine purpose built to crunch that data and help you accelerate your growth in an memerging market [114] (i.e. the use of infrastructure).

 

[354.] In support of this concern, Mr Friedman referred to an article, which he thought had been written in May 2019, by Pedro Castello, the head of analytics and machine learning at FCG, on how to build, develop and deploy a machine learning model to predict car prices, using neural networks. Mr Friedman also referred to how FCG had essentially based their whole business model on this pricing function [….].

 

[355.] As Mr Hodge stated: reality is that if we look at the FCG experience in other is clever data work helps you. You can […]. [115]

 

[356.] Later in his testimony, he went on to refer to pricing development team of stating that    algorithms and ability to analyse data, find halo customers, all of that, seems a lot more [….].

 

[357.] On the issue of barriers to entry, the concern remained that the WeBuyCars/AutoTrader tie-up would create a formidable competitor. As Mr Friedman submitted:

 

that data and even with the a relationship with AutoTrader and its dealers, unlimited capital, all these things together, when you take all these pieces of the puzzle and put them together it is going to be harder for people to raise capital to compete with Naspers. No one is going to want to compete with WeBuyCars as a part of this model. It is going to make it very difficult, if not impossible to compete with [116] [….].

 

[358.] We are of the view that, post-merger, the integration of the WeBuyCars and AutoTrader databases would allow the merged entity to be in a vastly advantageous position in its ability to price accurately. This advantage is likely to be further enhanced following integration with FCG. Despite Mr van der assertions to the contrary, we must conclude that this amount of data would be invaluable to the merged entity. The unassailable advantages which access to this data would confer upon WeBuyCars (already dominant in the car-buying services market) are, in our view, incontrovertible.

 

Other advantages associated with the transaction

 

[359.] There are other advantages that arise as a result of this transaction which could serve to entrench the dominance of WeBuyCars. These include the following: FCG Technology, Naspers Media and capital resources.

 

FCG technology

 

[360.] This theory of harm concerned how the merger would benefit WeBuyCars through access to technology from FCG and/or Naspers.

 

[361.] According to Mr Nikkelen, OLX believed that WeBuyCars was a company, whereas FCG was a company, and that WeBuyCars might therefore benefit from the addition of FCG technology in its operations. As he stated in his evidence [CONFIDENTIAL] rather than electronically.

 

[362.] In terms of the documentary evidence it was view, arising from the due diligence that it performed on WeBuyCars, [CONFIDENTIAL]. It was stated in the Naspers ICR WeBuyCars, under the heading [CONFIDENTIAL].

 

[363.] However, as Mr Nikkelen went on to say at the hearing, Mr van der Walt did not believe in the benefits of   technology and that, given that OLX wanted Mr van der Walt to continue to run WeBuyCars post-merger, interactions with him a sort of pull nature in which we hoped to inspire him to see the value of that rather than a push nature or prescriptive

 

[364.] Mr van der Walt confirmed in his evidence that, during the due diligence process, representatives of WeBuyCars had high-level discussions with representatives of OLX and FCG to assess       technology systems and to ascertain whether they would add value to WeBuyCars operations. Mr van der Walt, who clearly has a preference for operating manually and in a traditional way, submitted:

 

[CONFIDENTIAL] [117]

 

[365.] Mr van der Walt explained that WeBuyCars had also considered electronic auctioning platform, but that it had since developed its own one which was ready for launch at the time of his evidence.

 

[366.] In summary, Mr van der position on the FCG technology was the following:

 

 

my mind and my world, I think technologies out there going to rapidly, radically all over night change us into this formidable competitor going to kill the market. I just see that happening. Yes, we would love to have technologies that helps us take better photos, do online auctions quicker, although we have developed those. We have an evaluation app that we believe is on par with available out there. So, I see how the merger is going to be a significant threat to other players in the [….] .[118]

 

[367.] While Mr Nikkelen acknowledged that machine learning and artificial intelligence are an important part of future strategy, he submitted that these benefits are largely aspirational and potential at this stage, with very little practical impact currently. He also pointed out that the development of machine learning and artificial intelligence techniques forms of the vision of any technology and is not unique to Naspers as WeBuyCars itself was developing artificial intelligence capabilities of its own.

 

[368.] The Commission disagreed with the merging parties. According to the Commission, while Mr van der Walt tried to downplay the potential benefit from the Naspers technology (including FCG) in his testimony, he cited the benefits of tying up with a technology company when interviewed for Radio 702 (which took place prior to his testimony and thus carries significant weight, in our view) in which he said: I think many people know this, but Naspers owns various other car- buying companies overseas. So, also sitting on certain technologies that like to adopt in our business and therefore it made sense to a deal with them .[119]

 

[369.] In addition, he conceded, under cross examination, that he was not intimately familiar with FCG technology (and accordingly not well placed to say that it would not benefit him) and that WeBuyCars was, in fact, [CONFIDENTIAL]. Mr van der Walt also [CONFIDENTIAL].

 

[370.] It was, in the view, entirely logical that FCG technology would benefit WeBuyCars. FCG is a global car-buying company that is able to put far more resources into platform development. FCG has indicated that it has a 65-person team dedicated to product and technology and has invested extensive amounts of capital in a global platform that can be adapted to local needs. Aside from the benefits which a WeBuyCars/FCG tie up could bring, WeBuyCars could never hope to independently match       investments, given   diverse global business offerings.

 

Leveraging off Media

 

[371.] As mentioned above, marketing spend is critical to drive leads which, in turn, drive purchases. The more money spent on marketing; the more leads generated. It is unsurprising then that marketing is one of the largest cost items for a car-buying service given that share of voice relative to competitors is a key feature of how car- buying services compete.

 

[372.] Insofar as WeBuyCars is concerned, it was estimated that its marketing spend per purchase was approximately [CONFIDENTIAL] and this amount was potentially higher for competitors. Further, pre-merger total marketing spend accounted for [CONFIDENTIAL]. The WeBuyCars [CONFIDENTIAL]. WeBuyCars uses a mixture of TV, radio, billboard as ATL advertising. It uses Google and Facebook for digital marketing purposes.

 

[373.] The concern raised by the Commission was that post-merger, WeBuyCars would have access to cheaper advertising from Naspers group companies (including Media24 which is  largest media group with both publishing assets and leading digital online websites). This is because [CONFIDENTIAL].

 

[374.] According to the merging parties, this was a spurious theory because advertising and marketing expenditure on both AutoTrader and Media24 is trivial. marketing spend primarily goes to radio advertising and billboard advertising, because WeBuyCars believes that its advertising will be most effective if it reaches consumers while they are in their cars.

 

[375.] However, as the Commission pointed out, this simply reflects what WeBuyCars is currently finding economic to use, rather than being a reflection of the reach and impact Media24 could have for WeBuyCars post-merger. According to the Commission, post the merger, WeBuyCars would (at little or no cost to itself) be able to dominate the share of voice. This would be particularly true of online advertising which is currently expensive for most car-buying services but which represents a large part of their spend.

 

[376.] In the  view, the reduced marketing spend post-merger would increase share of voice and free up its resources which it could redirect to social media and online marketing campaigns. This would raise the costs for its competitors and generate barriers to entry and expansion by pushing them down the rankings and reducing their lead generation.

 

[377.] Internal documentation reveals that WeBuyCars was alive to the possible benefits accruing through their potential relationship with Naspers. Correspondence [CONFIDENTIAL].[120]

 

Access to Naspers capital resources

 

[378.] Although WeBuyCars is well resourced and may not need capital for its current expansion plans, the Commission submitted this was not indicative of the extent to which the merged entity would benefit from greater access to capital.

 

[379.] Weelee submitted that with more capital accessible to WeBuyCars post-merger, it would be harder for its competitors to access capital than is currently the case. Mr Hodge stated that this is a phenomenon which has been observed in technology markets globally and is known as the     In essence, the investment of a big technology company into a business area dries up other venture capital funding for entrants due to the low probabilities of succeeding in that market. More capital also has a benefit insofar as it may assist in fighting off future competition which WeBuyCars does not currently face. In terms of how capital is used, Mr Nikkelen confirmed that Naspers would seek to use the retained earnings to drive growth aggressively in the business as this is in line with business strategy.

 

[380.] According to the merging parties, this theory is vague, speculative and unsubstantiated by the evidence before the Tribunal. It is common cause that WeBuyCars is not currently capital-constrained and that is not expected to change in the foreseeable future. In these circumstances, so the merging parties argued, the Commission does not attempt to explain how, and to what extent, the introduction of Naspers as a majority shareholder will make access to capital more limited for new entrants, nor does it explain in what way that would increase the scope for anti-competitive practices by WeBuyCars.

 

[381.] Furthermore, Mr Murgatroyd explained in his evidence that this argument assumes that WeBuyCars is not currently subject to effective competition, and that the threat of entry or expansion by smaller rivals is all that may constrain it. He submitted that WeBuyCars already competes with a number of large and well-capitalised firms in the relevant market, and accordingly any additional access to capital would be pro-competitive.

 

[382.] What is clear is that the proposed merger will give WeBuyCars access to substantial capital resources in addition to the current resources at its disposal.

Whilst we conclude that there is not sufficient evidence in relation to the theory as advanced by the Commission, nothing turns on this.

 

Overall conclusions on the portfolio effects

 

[383.] Internal OLX documents show that of the [CONFIDENTIAL]. The merging parties sought to downplay this number to a figure closer to [CONFIDENTIAL], This is significant when compared to the next closest competitor (Gumtree) with 22 000 private listings and serves as a strong indication of the benefit which would accrue to the merged entity post transaction if each one of these [CONFIDENTIAL] listings could serve as potential leads to WeBuyCars. This, in addition to own estimates [CONFIDENTIAL]. Such a benefit could not be replicated in the market where the next largest advertising platform has half the listings and the next largest car buying service purchases less than 10% of the vehicles per month.

 

[384.] Further, Autotrader as the self-described largest and most well-established source of used vehicle leads in South Africa, with the ability to market to over 6 million visitors to its website each month, would through the proposed transaction amass 30 years of intricate data on both the cars and consumer behaviour for [CONFIDENTIAL]. It also has a vast network of dealer customers. Both Friedman and Morgan testified that the data available to Autotrader is superior to data generally available to dealers on AutoFuzion. Such data enables more precision in pricing for the purchase (and sale) of vehicles, as well as access to an established dealership network to on-sell these vehicles to. No competitor in the market today is reasonably equipped to compete with such a tie-up.

 

[385.] The technological abilities of FCG were not in dispute. With the afore-mentioned data would come the flow of expertise of FCG in the use of technology in the used car industry, enjoying a wealth of success in developing markets with far less data and far less of a substantial brand base. Include in this, the benefit accruing of a parent company able to give preferential advertising rights and immeasurably deep pockets and it is apparent that no other market player would be able to replicate the complete package of synergies which would arise in this transaction. The combination of AutoTrader, OLX and WeBuyCars, together with FCG, would raise barriers to entry into and expansion in the car-buying services market and tip the probabilistic scales towards an entrenched WeBuyCars in this market.

 

[386.] We therefore conclude that the merger is likely to result in a substantial prevention and lessening of competition as a result of the portfolio effects arising from this merger.

 

[387.] We have accordingly concluded that both the portfolio effects of the proposed merger and the removal of a potential and close competitor, FCG, from the relevant market will give rise to a substantial prevention and lessening of competition.

 

[388.] However, even if one were to conclude that the portfolio effects arising from the proposed transaction did not result in a substantial prevention or lessening of competition, the portfolio effects analysis above confirms that it is the Naspers portfolio (which includes its strong balance sheet and other businesses namely, OLX, AutoTrader and Media24) which affords substantial competitive benefits to a car-buying service, whether WeBuyCars or FCG. This analysis of the Naspers portfolio, and how it can be used by a car-buying service, thus reinforces our conclusion on the first theory of harm that the proposed merger will likely give rise to a substantial prevention and lessening of competition because it removed a potential and close competitor, FCG (in combination with Naspers), from the relevant market. The proposed merger is therefore first and foremost prohibited on the basis of this first theory of harm.

 

Probabilistic nature of merger assessment

 

[389.] In reaching our conclusion about the adverse portfolio (or conglomerate) effects of the proposed transaction, we have also considered the question of whether one should have regard only to the immediate or short-term anticipated effects of the proposed transaction on competition or also to the longer-term effects on competition. As an ex-ante regulator and bearing in mind the probabilistic nature of merger analysis, we find that we are required to make our assessment having regard to how we see the market in question evolving in the future.

 

[390.] Globally, competition authorities continue to grapple with the network effects of mergers in digital markets. These network effects present a similar theory of harm as conglomerate effects in physical markets, yet they are difficult to identify and to understand given the fast-pace evolution and interconnectivity of digital markets.

 

[391.] In an article by Ezrachi and Modrall,[121] they state the following:

 

Notable are network effects which facilitated the rise of key platforms, gatekeepers, and the creation of new ecosystems. Also noteworthy is the tipping of some markets in favour of the dominant [….]. [122]

 

[392.] They further state that:

 

with potential and innovation competition, challenging transactions based on conglomerate theories of harm can raise difficult issues of proof. Nonetheless, where activities can be broken down into a number of markets and their market power assessed using well-recognised antitrust tools, conglomerate theories of harm offer a framework for assessing the likely impact of a large company acquiring a start- up active in a related market. In such a case, instead of foreclosure (for example), the harm could be framed as making it harder for existing or future competitors to attract customers away from the [….]. [123]

 

[393.] The Commission argued that the market structure and competition dynamics will likely be shaped in the next 5-10 years. As they stated in their heads of argument, with reference to Mr Hodge expert report:

 

WeBuyCars may continue to hold the dominant position in future in spite of entry by others, the future market structure and competitive dynamics is likely to be determined over the next 5-10 years, the time line the Naspers models to reach that more mature stage. This is the context to the discussion of the second theory of harm [the portfolio/ conglomerate theory of harm], and whether this transaction is likely to tip the probabilistic scales more towards an entrenched WeBuyCars with substantial market power and away from a market where there may be more chance of some successful challengers. Simply removing a large potential entrant in the form of OLX/FCG already tips the scales materially as discussed in the last section [the removal of a potential competitor], but we also even absent that entry, the merger tips those scales. [124]

 

[394.] Based on the evidence before us on the relative positions of each of the Naspers businesses i.e. OLX and AutoTrader, as well as WeBuyCars, we are of the view that the competition concerns are not speculative. In evolving markets such as the car-buying service market through online platforms where data will play an increasingly important role in the competitive dynamics, we are required to apply caution in our assessment so as not to further entrench dominance, thereby raising barriers to entry or expansion for new and existing competitors, as they would be unlikely to be able to attract customers away from the merged entity.

 

[395.] In our view, the creation of a combined WeBuyCars/ OLX/ AutoTrader / FCG entity would enhance existing dominant position to a point where it would essentially be unassailable vis-à-vis its rivals. Approving such a transaction in our view would tip the probabilistic scales towards an entrenched WeBuyCars which could create irreversible harm to potential successful challengers.

 

Conditions

 

[396.] The merging parties put up two sets of conditions during the merger proceedings. The latest set (exhibit AA) was introduced through the evidence of Mr Nikkelen and was not therefore canvassed with Mr van der Walt or any of the witnesses, but they are clearly inadequate to address our finding on the portfolio effects of the proposed transaction in several different respects.[125]

 

[397.] We say this for the following reasons:

 

397.1.    The first is that the Commission would not be able to effectively monitor the tendered conditions:

 

397.1.1.  Clauses 3.1 of the proposed conditions required that AutoTrader and OLX not divulge any commercially valuable data or information to WeBuyCars.

 

397.1.2.  Clause 3.2 required that if there was any use of OLX or AutoTrader to refer customers to WeBuyCars, this referral service should be made available on identical terms to any dealer who wished to subscribe to it. The lack of clarity on what would be considered valuable aside, we found it doubtful that the Commission would have the ability to conduct the necessary audit one would have to carry out to determine whether there was a cross pollination of vehicle and consumer data.

 

397.1.3.    Clause 3.3 of the conditions requires that AutoTrader should provide dealers with access to platform on commercially reasonable and non-discriminatory terms to advertise pre-owned motor vehicles; again this condition would be impracticable for the competition authorities to monitor.

 

397.1.4.    The same is true of the non-discriminatory pricing condition (clause 3.3(c)), which only applies to which are economically equivalent and which otherwise entail equivalent quantities and quality of The discretion granted to AutoTrader in this regard is so wide that the Commission would simply be unable to establish that WeBuyCars is being unduly favoured by AutoTrader.

 

397.2.   clause 3.2 of the conditions can easily be circumvented because the terms of any referral service offered by OLX and AutoTrader are left to them to determine.

 

397.3.   clause 3.5 of the conditions is, in addition to being impossible to monitor, deficient inasmuch as it contains no provisions which ensure equality of pricing between advertisements purchased by WeBuyCars and those purchased by other dealers. WeBuyCars [CONFIDENTIAL].[126]

 

397.4.   Finally, all of the proposed conditions are limited in time to four years. The problem with this is that the harm which might otherwise be occasioned post- merger, is merely postponed to a point later in time. The merger parties may well use that period to achieve the full integration necessary to benefit fully from the data sets on which they sit.

 

Conclusion

 

[398.] In conclusion we were of the view that the proposed merger would likely substantially prevent or lessen competition in the relevant market for car-buying services. The substantial prevention or lessening of competition would arise through the removal of a potential entrant, FCG.

 

[399.] Furthermore, the merger would likely lead to a substantial prevention or lessening of competition through the creation of synergies through the complementary businesses in Naspers Group (i.e.; OLX and AutoTrader) that may be leveraged to entrench existing dominance and raise barriers to entry to other players, whilst also providing some reciprocal benefits to those Naspers businesses too, which may entrench their own positions. These benefits cannot be replicated by competitors.

 

[400.] It is important to mention that this was not a typical potential entry case. In this matter, it was the acquirer that was the potential entrant. It however chose to abandon this option once it had identified the opportunity to purchase the dominant firm in the market, WeBuyCars. This however does not make a difference to our analysis of the competition effects of the proposed transaction.

 

[401.] FCG represented a formidable entrant, it had the funding, the technology and a proven track record of entering into emerging markets. Moreover, the evidence indicated that OLX/FCG was both a likely and formidable entrant into the South African market absent the transaction. This is based on documentary evidence [CONFIDENTIAL].

 

[402.] We therefore found that the likely fierce competitive consequences of entry was what most likely resulted in the transaction materialising, as WeBuyCars sought a deal at favourable terms to OLX in order to avoid competing head-on with FCG.

 

[403.] With respect to the second theory of harm, it is apparent from the post-investment strategy for WeBuyCars that there was a stated strategy for WeBuyCars to exploit the synergies within the Naspers group as a whole so as to strengthen the WeBuyCars position in the market. The synergies assessed included the access to customer leads from OLX as well as the access to        powerful data intelligence and potentially preferential treatment on the platform. OLX is a market leader for online generalist platforms, while AutoTrader is a must have platform. In addition, we also considered broader benefits which would accrue to WeBuyCars through being a part of the Naspers group such as Naspers technology and media. Overall, we found that no other competitor in the market would be able to replicate the synergies identified in the transaction.

 

[404.] The probabilistic nature of merger assessment, particularly in evolving markets such as the car-buying service market where conglomerate effects concerns arise as a result of the proposed transaction, requires a cautious approach so as not to further entrench a dominant position in the market.

 

[405.] Although the merging parties tendered conditions to address any potential harm arising from the transaction, we found that the proposed conditions do not address the harm occasioned by the merger, either as regards the removal of an effective competitor, or vis-à-vis the adverse portfolio effects which would arise as a result.

 

[406.] We accordingly found that the merger should be prohibited.

 

PART C

 

Discovery Application Reasons

 

DISCOVERY APPLICATION

 

[408.] During Mr Nikkelen examination in chief on 24 October 2020, Mr Wilson enquired as to the breakdown between dealers and private listings of used vehicles on OLX. [127]This was relevant to inter alia determining the potential benefit WeBuyCars/FCG would accrue through the generation of leads from OLX. Mr Nikkelen responded that only about a quarter of sellers are true private sellers. Mr Wilson asked whether Mr Nikkelen had performed an exercise to ascertain this fact or whether it was just a figure derived from a business sense.

 

[409.] Pursuant to this question from Mr Wilson, Mr Majenge, counsel for the Commission, interjected and indicated that the evidence put forward by Mr Nikkelen did not form part of his witness statement and that the Commission reserved its rights in relation to this line of questioning.

 

[410.] Mr Nikkelen indicated to Mr Wilson that his statement was based on an exercise he had done that was not shared with the Commission. However, he was willing to share the data. Counsel for the Commission, at this point placed on record that the data referred to by Mr Nikkelen had not been discovered, argued that this resulted in prejudice and again reserved the rights to have the information struck from the record if required.

 

[411.] On 25 October 2019, the Commission requested a ruling on the above point that Mr Nikkelen had sought to introduce evidence during his examination in chief that was not part of the merging discovery.

 

[412.] We decided to admit the evidence on the basis that (i) all underlying data must be provided to the Commission and Weelee; (ii) the Commission and Weelee be afforded the opportunity to prepare on the discovered document / data; and (iii) Mr Nikkelen should be made available to be cross-examined on the data provided, if necessary. Further, we made provision for further analysis by the experts on the issue prior to the proceedings, should it be required.

 

[413.] The Commission wrote to the merging parties on the evening of 25 October 2019 requesting the underlying data set, an explanation of the rules, evidence of the decision rule being applied to generate the rules and an affidavit from any analyst involved in constructing this. In this context, the functionally refers to the rules implemented by a data analyst during his/her assessment of an underlying data set. It would ultimately describe to the Commission how the analyst went about assessing an underlying data set to come out with the results Mr Nikkelen relied upon in his testimony.

 

[414.] On 28 October 2019, the merging parties filed an affidavit deposed to by Ms Karina Quibell a data analyst at Naspers. In this affidavit, Ms. Quibell stated that the raw data used for the purposes of this assessment runs into petabytes and is not capable of being transferred electronically or even physically via storage media.

 

[415.] On 29 October 2019, the Commission addressed further correspondence to the merging parties requesting documents and data outputs from FCG and/or OLX in other markets (i.e. other countries) where FCG has sought to identify grey dealers including the initial screening test applied to the OLX data, the subsequent screening tests applied to the initial list of potential grey dealers and the final results of how many of the initial screen were confirmed as grey dealers. Grey dealers was a term used to denote individuals on OLX who were not necessarily advertising to sell cars under the banner of a traditional dealer, but who were also not true private sellers seeking to sell their own cars.

 

[416.] The Commission also required an affidavit from an appropriate FCG representative in each country other than South Africa for which results were generated as to the completeness of the document discovery and the results.

 

[417.] The merging parties opposed the broader request. They submitted that the further requests (i.e. that related to other jurisdictions and the underlying dataset) was not necessary to interrogate the data set out in exhibit W. [128]

 

[418.] In response, the Commission filed a notice of motion with the Tribunal on 7 November 2019, requiring that the merging parties discover:

 

Documents and data outputs from FCG and/or OLX in other markets where FCG has sought to identify grey dealers as referred to in slide 8 of Discovery item 4. 10 (FCG & OLX: Owning the used car ecosystem together- Dec18), including the initial screening test applied to the OLX data, the subsequent screening tests applied to the initial list of potential grey dealers and the final results of how many of the initial screen were in fact confirmed as grey dealers. The Commission also requires an affidavit from an appropriate FCG representative in each country for which results were generated as to the completeness of the document discovery and the results within 1 day of the order. (emphasis added)

 

[419.] The Commission argued that exhibit W sought to minimise the number of true private sellers on the OLX platform in South Africa. This was in stark contrast to what had been reflected in strategic documentation. In addition, the Commission had found that the test run by the merging parties to determine whether a listing on OLX SA was by a private individual or a dealer had yielded certain false positives, i.e. had counted truly private sellers. Thus, so it was argued, in order to test the veracity of the information contained in exhibit W, the Commission required the requested underlying documentation.

 

[420.] The merging parties argued that the so-called screening tests and data outputs from screening tests in other jurisdictions i.e. other countries than South Africa in relation to grey dealers was not relevant or probative of the inquiry into the usefulness and reliability of exhibit W. In addition, it would not address the issue of the presence of false positives. These factors, compounded by the prejudice which would accrue in requiring the merging parties to collate the results across several jurisdictions, militated for the dismissal of the discovery application.

 

[421.] We dismissed the request for discovery for the following reasons.

 

[422.] The Commission was unable to show how the additional documentation / data, relating to other jurisdictions i.e. other countries in which OLX is active, would assist the Tribunal in deciding the reliability of exhibit W. The Commission had argued that exhibit W and Mr accompanying testimony were not reliable inter alia because the methodology and dataset utilised yielded a sufficient number of false positives and anomalies (including an instance where a person selling DJ equipment and a car was listed as a dealer) to render the exercise unreliable. In light of this, there appeared to be no utility in discovering the underlying data as it would not take matters any further.

 

[423.] Furthermore, the Commission failed to sufficiently demonstrate that the prejudice and delays in acquiring the further discovery relating to other jurisdictions was outweighed by the sought documents or their import and impact on the case.

 

 

Mr AW Wessels and Prof. I Valodia concurring

 

Tribunal Case Managers:     Alistair Dey-Van Heerden &

Kgothatso Kgobe.

 

Tribunal Economists:             Karissa Mothoo-Padayachie;

Lumkisa Jordaan

 

For the Merging Parties         Adv. Jerome Wilson SC

and Adv. Anisa Kessery

Instructed by Nick Altini and Leana

Engelbrecht of Herbert Smith Freehills

 

For Weelee:                           Adv. W Luderitz SC assisted by

Adv. A Coutsoudis and Adv. G Marriot

Instructed by Anthony Norton and Anton Roets of

Nortons Inc.

 

For the Commission:              Bukhosibakhe Majenge,

Nelly Sakata, and

Romeo Kariga.

---