MacNeil Agencies (Pty) Ltd v Competition Commission (121/CACJul12) [2013] ZACAC 3; [2013] 2 CPLR 416 (CAC) (18 November 2013)

THE COMPETITION APPEAL COURT OF

SOUTH AFRICA

CAC Case No: 121/CACJul12

In the matter between:

MACNEIL AGENCIES (PTY) LTD

APPELLANT And

THE COMPETITION

COMMISSION

RESPONDENT

Coram: DAVIS JP, VICTOR AJA & ROGERS AJA

Heard: 20 SEPTEMBER 2013

Delivered: 18 NOVEMBER 2013

______________

JUDGMENT

ROGERS AJA:

Introduction

The appellant (‘MacNeil’) was the 7th respondent in a complaint which the present respondent (‘the Commission’) initiated on 18 March 2008 and referred to the Competition Tribunal (‘the Tribunal’) on 2 February 2009. I shall refer to the other respondents cited in the referral by the following abbreviated names: 1st – DPI; 2nd – Petzetakis; 3rd – Marley; 4th – Swan; 5th – Amitech; 6th – Flo-Tek; 8th – Andrag; 9th – Gazelle.

The markets to which the referral relates were the manufacture and supply of polyvinylchloride (‘PVC’) and high-density polyethylene (‘HDPE’) pipe products in South Africa. The Commission alleged that the respondent firms had, in violation of s 4(1)(b) of the Competition Act 89 of 1998 (‘the Act’), colluded in these markets to fix prices and to allocate tender contracts issued by customers. The Commission alleged that the collusion started prior to 1998 (when the Act came into force) and continued until 2007.

Shortly before making the referral the Commission granted DPI conditional leniency. The other firms delivered answering affidavits in which they variously denied the allegations of collusion outright or disputed the extent of their participation or the appropriate penalty. However, by the time the trial began in the Tribunal in September 2010 Marley, Swan and Flo-Tek had reached settlements with the Commission in which they admitted their participation in prohibited conduct and agreed to pay penalties of 6% of their 2007 turnover (or affected turnover, in the case of Marley).

The trial ran for a number of days in September 2010 and January/February 2011. Argument was heard in April 2011. MacNeil, Andrag and Gazelle contended that they had not committed any prohibited practices. Petzetakis (which had just failed to beat DPI to the post in submitting a leniency application) and Amitech conceded participation in prohibited practices but denied that it

was as extensive as the Commission alleged or that they should be penalised to the extent for which the Commission contended. The

Tribunal delivered its decision on 4 July 2012. MacNeil was found to have participated in price-fixing through its attendance at three meetings which took place over the period February to October 2007. An administrative penalty of R2 million was imposed. Although the findings in respect of the other firms are not relevant to this appeal, I mention for the sake of completeness that Andrag was found to have engaged in prohibited conduct (unrelated in any way to MacNeil) but the tribunal concluded that it would be inappropriate to impose any administrative penalty; the case against Gazelle was found not to have been made out; and Amitech and Petzetakis, which in any event conceded participation in collusive conduct, were ordered to pay administrative penalties of R11,1 million and R9,92 million respectively.

MacNeil appeals to this court against the Tribunal’s findings on the merits and against the extent of the administrative penalty imposed on it. Mr Rosenberg SC, leading Mr Kelly, appeared for MacNeil at the appeal, as he did in the Tribunal; and Mr

Maenetje SC appeared for the Commission, as he did in the Tribunal.

The evidence

The evidence before the Tribunal indicated that there had been a cartel in existence for some years in the relevant markets nationally, the core members being DPI, Petzetakis and Marley. In the latter part of 2006 MacNeil entered the PVC piping market to which the cartel related. MacNeil was based in Cape Town and its customers were principally in the Western Cape. The respondent firms, apart from selling their piping products to merchants and directly to contractors, occasionally sold product to each other. MacNeil, which mainly sold its pipes to merchants, also undertook toll manufacturing for Flo-Tek. The firms with which MacNeil was alleged by the Commission to have colluded in price-fixing were DPI, Petzetakis and Flo-Tek.

Insofar as the case against MacNeil is concerned, the relevant personalities were DPI’s Andre Auret (national sales and marketing director) and Rene le Riche (who was, at the relevant time in 2007, the sales director of DPI’s Incledon-DPI

division); Petzetakis’ Trevor Lombard; Flo-Tek’s Shaun Hart; and MacNeil’s Neil Malherbe (managing director), Sean Diab (general manager) and Phillip Brink (sales manager). I shall refer to these persons by their surnames.

The Commission’s allegations in the referral affidavit, insofar as MacNeil is concerned, were sparse. In para 36.6 the Commission’s deponent alleged that during February or March 2007 a meeting was held at DPI’s Cape Town offices attended by Le Riche of DPI, Lombard of Petzetakis, Hart of Flo-Tek and Brink of MacNeil. These firms were alleged to have ‘reaffirmed the pricing principles previously agreed between them as detailed above in relation to PVC pipes, specifically PVC sewer piping’. In truth, MacNeil was not mentioned in any earlier paragraphs. The Commission went on to say, in para 37, that it was clear from the meetings mentioned earlier in the affidavit that at least until March 2007 the respondent firms colluded in fixing prices though the Commission said that it had reason to believe that the firms’ collusive conduct continued beyond March 2007. The referral affidavit then went on to deal under a separate heading with the topic of collusive tendering and market allocation. No particularised allegations were made under this head against MacNeil nor did the Commission pursue such a case against MacNeil at the hearing before the Tribunal.

In Petzetakis’ answering affidavit its deponent, Michelle Harding (its then managing director), said that she had no personal knowledge of the February/March 2007 meeting but Petzetakis admitted the Commission’s allegations. Harding referred to a confirmatory affidavit by Lombard.1

Flo-Tek and MacNeil each delivered answering papers in which they dealt with para 36.6 of the referral affidavit and with various meetings which had taken place between DPI, Petzetakis, Flo-Tek and MacNeil during 2007. These answering affidavits provided details of meetings additional to those mentioned in the referral affidavit but were exculpatory in nature – both firms contended in essence that they did not at these meetings agree to fix prices. Flo-Tek’s main answering affidavit was made by its chief financial officer, Mr C Bandaru, but a confirmatory affidavit by Hart was filed. Brink deposed to MacNeil’s

main answering affidavit and there was a confirmatory affidavit from Diab.2

Because DPI had been granted leniency, it did not deliver answering papers in response to the referral affidavit. In advance of the hearing, however, the Commission delivered witness statements by Auret and Le Riche. Le Riche’s statement referred to four meetings allegedly involving MacNeil in 2007. Auret and Le Riche testified at the Tribunal hearing

It is unclear whether Petzetakis delivered witness statements. Its principal deponent, Harding, left Petzetakis in July 2009 (before the trial commenced) and was called by the Commission as its first witness. We were told that we did not need to read her evidence, presumably because she did not claim to have attended any meetings at which MacNeil was represented. Lombard, the only Petzetakis representative who could have spoken from direct knowledge about meetings with MacNeil, did not testify nor is there any witness statement from him in the appeal record.

Because Flo-Tek had settled with the Commission by the time of the trial, no witness statements from its representatives were delivered nor was Hart, the Flo-Tek employee alleged to have been involved in meetings with MacNeil, called to testify. However,

both Mr Rosenberg and Mr Maenetje in cross-examination made reference to Flo-Tek’s answering affidavit (which, in the respects relating to Hart, had been confirmed by the latter).

MacNeil delivered witness statements by Malherbe, Diab and Brink, and they all testified at the hearing.

By way of background to the meetings involving MacNeil in 2007, it is relevant to note that Brink stated in MacNeil’s answering affidavit that it was common knowledge in the industry, at the time MacNeil entered the PVC pipe market, that DPI, the dominant

player, strongly desired to establish fixed prices and to allocate tenders. Brink had previously worked for Petzetakis and testified

that while he worked there he heard through the grapevine of meetings which were taking place among the manufacturers. He was

given price lists and informed about the discount levels.

That such collusion was happening prior to 2007 was clear from the evidence at the trial and from the Tribunal’s findings in relation to the other firms. MacNeil was a new entrant seeking to gain market share by aggressive pricing. It was thus inherently likely that the larger firms with interests in the Western Cape would seek to constrain this new disruptive force.

Collusive activity by its nature tends not to be documented. There are no minutes of the meetings MacNeil attended nor emails or other correspondence relating to the discussions. By the time of the hearing the witnesses were talking about events which had occurred three or more years previously. In the case of Auret and Le Riche of DPI, their evidence concerned a broader canvas

than just the 2007 meetings with MacNeil. Complete consistency and precision of recall could not fairly be expected of them. The evidence needed to be assessed holistically to see what was established on a balance of probabilities.

There was contact between Auret of DPI and Malherbe of MacNeil in 2006 and 2007, including a meeting in June or July 2007. These interactions were not shown to have involved collusion in contravention of s 4(1)(b). Auret said in this context that prices would have been discussed among the firms ‘at a lower level’.

Apart from the meeting between Auret and Malherbe in mid-2007, Le Riche in his witness statement and oral testimony spoke of four meetings in 2007 involving MacNeil:

[a] The first meeting, allegedly attended by Diab on behalf of MacNeil, occurred in about January/February 2007. This was a meeting where DPI (Le Riche), Petzetakis (Lombard), Flo-Tek (Hart) and MacNeil (Diab) agreed not to supply product to a merchant,

Peakstar, which had secured three large-volume tenders in the Western Cape at very low prices. (In his witness statement Le Riche did not mention Diab’s presence. In his oral evidence he said that this was a detail he had forgotten but subsequently remembered. The Commission’s referral affidavit had not made reference to a meeting concerning Peakstar.)

[b] The second meeting, which he placed in February/March 2007 and was held at DPI’s Cape Town premises, was attended by the same four firms. DPI, Petzetakis and Flo-Tek were represented as before but MacNeil, so Le Riche claimed, was represented by Brink. This was a meeting where, according to him, pricing was discussed and agreed. (In broad terms, this would correspond to the meeting alleged by the Commission in para 36.6 of the referral affidavit.)

[c] Third and fourth meetings, along similar lines to the second, were held later in the year where pricing was again discussed and re-affirmed. In his witness statement Le Riche placed the third meeting in July 2007 and the fourth (last) meeting in September 2007. In his oral evidence he said the last meeting was held in November 2007. However, it would not surprise me if his recollection as to timing was somewhat hazy. There was other evidence (from MacNeil and Flo-Tek) that the third and fourth meetings occurred in about September and October 2007. One can fairly conclude that Le Riche’s third and fourth meetings were the same two meetings as those mentioned by MacNeil and Flo-Tek as having occurred in September and October 2007.

As to the first meeting (where Peakstar was discussed), there is no version in the answering affidavits of Petzetakis, Flo-Tek and MacNeil for the simple reason that no such meeting was mentioned in the referral affidavit. Lombard of Petzetakis and Hart

of Flo-Tek were not called to testify. In oral evidence Diab denied having attended any such meeting. MacNeil did in fact proceed to supply product to Peakstar.

The Tribunal accepted that it had not been proved that MacNeil was a party to the Peakstar meeting. On this basis, Le Riche was in error in implicating MacNeil in the first meeting of 2007. This finding has not been attacked on appeal. I shall thus confine

my attention to what Le Riche regarded as the second, third and fourth meetings, where – according to him – pricing was discussed. For convenience I shall refer to them as the second, third and fourth meetings even though they were the only three involving MacNeil.

The Tribunal appears to have thought that the rejection of Le Riche’s evidence regarding MacNeil’s attendance at the first (Peakstar) meeting meant that his evidence fell away altogether in regard to alleged price-fixing involving MacNeil at a

meeting in early 2007. The Tribunal instead relied on what MacNeil and Flo-Tek had said in their answering affidavits and on the evidence of Diab and Brink. Even on this evidence, assessed in the light of the inherent probabilities, the Tribunal concluded that pricing had been discussed at a meeting attended by MacNeil early in 2007 and that Diab had been told to tell Brink that MacNeil should ‘fall into line’ on pricing.

However, I do not regard Le Riche’s evidence as being irrelevant to an assessment of this – on his version, the second of four meetings involving MacNeil in 2007. If, as the evidence in its totality indicated, prices were discussed at a meeting in early 2007, this would be Le Riche’s second meeting but he would have erred in identifying MacNeil’s representative as having been Brink rather than Diab. Le Riche obviously recalled Diab as having attended a meeting early in 2007 but seems mistakenly to have placed him at the Peakstar meeting rather than the pricing meeting.

Le Riche’s evidence in his witness statement concerning what he styled the second meeting (actually the first meeting involving MacNeil) was that its purpose was ‘to attempt to ensure stabilisation of the market, which had been volatile as pricing had remained unchecked and the price of input-polymer was at an unsustainable level’.3 He said that the parties ‘re-affirmed the previously agreed pricing principles (ie the agreed-upon price list and maximum discount) in relation to PVC sewer piping’. The notion of reaffirmation was almost certainly correct in relation to Petzetakis and may also have been right for Flo-Tek. There is no evidence, however, that MacNeil had previously agreed to any pricing principles. Nevertheless, in the context of the long-standing cartel of which Le Riche had knowledge, his use of the word ‘re-affirmed’ in the context of a meeting attended by Petzetakis among others, is understandable.

In his oral evidence4 Le Riche said that the second meeting was arranged by Hart. He confirmed that the parties had specifically discussed prices and discounts. It was very clearly spelt out that the firms would not go below certain prices for product they sold directly to contractors and would not drop below certain further discounted prices when selling product to merchants.

It was put to Le Riche by Mr Rosenberg for MacNeil5 that Brink denied having been present at the meeting in early 2007 (Le Riche’s second meeting). It was also put that Diab had been present, that certain pricing proposals had been made but that no agreement was reached. In cross-examination Mr Rosenberg aligned MacNeil’s case with Flo-Tek’s answering affidavit regarding this meeting, which in summary was:6 that Le Riche had tabled proposals for the manner of pricing into the Cape Town market; that the proposals were designed to preserve DPI’s share of the Cape Town market and its profit margins; that Le Riche told Flo-Tek as a new entrant into the Cape Town market that it was expected to adhere to these proposals – that ‘this was how things worked in the Cape Town market’; and that Hart had attended the meeting primarily as a ‘fishing expedition’, was ‘non-committal’ about Flo-Tek’s involvement in the proposals ‘and did not agree to them’. Quite how Flo-Tek’s supposed absence

of agreement was communicated, if at all, was not the subject of oral evidence from Flo-Tek because Hart was not called as a witness. The supposed absence of agreement as alleged in Flo-Tek’s affidavit might mean no more than that Hart subjectively did not intend to go along with the pricing proposals. It is not without significance that Mr Rosenberg for MacNeil did not put to Le Riche that Hart had clearly distanced himself and Flo-Tek from DPI’s pricing proposals.

Le Riche remained firm, under cross-examination from Mr Rosenberg, that at the second meeting (as well as at the third and fourth meetings) the parties discussed the specifics of pricing.7 He said that his witness statement had referred only to sewer pipe pricing because most of the time was spent on setting the sewer piping prices: they were the largest piping component of civil contracts but agreement was also struck on the pricing of other PVC piping – that part was ‘easy’.8

In re-examination Le Riche said9 that he was concerned that his cross-examination by MacNeil’s counsel created an impression of confusion about the meetings. He said he had re-read his witness statement overnight which very clearly set out the four meetings which took place in 2007. The only contentious issue, he said, was that he had forgotten that Diab attended the one meeting (Le Riche was clearly referring here to the Peakstone meeting, ie the first 2007 meeting). For the rest, he said that his witness statement clearly described all four meetings exactly as they occurred. He added:

‘As far as the three other meetings with Philip Brink were concerned, we discussed prices, we discussed the price on pressure pipe, sewer pipe and all the sewer fittings. After the meeting everybody left there in agreement, I not once picked up that anybody is not in agreement. So yes, those three specific meetings on price definitely occurred according to my witness statement.’

The ‘other three meetings’ to which Le Riche was referring in the above passage were what I have referred to as the second, third and fourth meetings. We know that Le Riche was mistaken in identifying MacNeil’s representative at the second of the four meetings (the February/March 2007 meeting) as being Brink rather than Diab. This part of Le Riche’s evidence

nevertheless has a ring of sincerity and strength of feeling about it.

We know that Petzetakis admitted in its answering papers the Commission’s allegations in para 36.6 of the referral affidavit, which is essentially Le Riche’s version of the so-called second meeting. Lombard filed a confirmatory affidavit. This was an admission by Petzetakis against its own interest. Although the admission that MacNeil was represented by Brink was incorrect, the admission is not without evidential value.

MacNeil’s own answering affidavit regarding this meeting (made by Brink and confirmed by Diab) was that MacNeil had been invited to attend several meetings which Brink understood to be for purposes of fixing prices in accordance with DPI’s

requirements. As to the meeting early in 2007, Diab had been invited so that Le Riche could tell him that MacNeil’s pricing was too low and that Diab should bring Brink (MacNeil’s sales manager) ‘into line’. At the meeting Le Riche threatened MacNeil that if it continued to price below the levels DPI found acceptable DPI would dump product into the market in order to cripple MacNeil. MacNeil’s answering affidavit denied that Diab ‘affirmed’ any pre-existing principles – MacNeil had not been party to any earlier agreement. There were no allegations in the answering affidavit as to how Diab had reacted to what DPI said.

In Diab’s witness statement he confirmed that he had attended the meeting in February/March 2007 at Hart’s invitation. He continued:

‘(5) I contributed little to this meeting. My only reason for being there was to pick up useful information, if made available. There was no agreement reached regarding pricing principles, nor was there any reaffirmation of any previously agreed pricing principles. MacNeil was a newcomer to the market, with an extremely small market share. There was no sensible reason for it to engage in price fixing or for it to adhere to the requirements of the large manufacturers such as [DPI]. [MacNeil] did not do so, as evidenced by the records of its sales over the period in question.

(6) As to Mr Le Riche calling on me to bring Mr Brink “into line”, I cannot be sure whether this was communicated to me by Mr Le Riche at this meeting, or whether I came to hear of this requirement on another occasion.’

This is a somewhat coy narration. Diab does not specifically say that DPI made pricing proposals but the defensive nature of the quoted passages seems to imply that pricing proposals were made, Diab’s version being that MacNeil did not agree to them at the meeting and did not in fact implement them. The statement also appears to dilute what Brink had said about this meeting in MacNeil’s answering affidavit and which Diab had confirmed in a supporting affidavit. (These discrepancies, as one would expect, were exposed in cross-examination.)

In his oral evidence Diab said that he attended the meeting because it was an opportunity for MacNeil as a new player to meet the other manufacturers and he also hoped that he might get to see DPI’s factory. He described his role as passive. The main

discussion was taking place among the other three firms. He described the conversation as ‘general day-to-day stuff’,

‘general chatter’ – there was no agreement to set prices or award contracts or allocate customers. When confronted in cross-examination with the purpose of this and other meetings as stated in Brink’s affidavit, he at first denied that Brink’s affidavit was correct and then said that he could not recall and conceded he had no basis to differ from Brink’s description of the purpose for which the meetings were called. He was taken in cross-examination to Flo-Tek’s answering affidavit. He claimed not to recall the tabling of pricing proposals by Le Riche but said he had no basis to dispute that prices were discussed. Regarding Le Riche’s instruction to Diab to bring Brink into line, Diab at first denied this but was then shown his witness statement after which he accepted that it was possibly said.

Brink could naturally not speak from personal knowledge regarding this particular meeting. He did testify, however,10 that Diab told him that DPI wanted MacNeil to fall into line on pricing and threatened to dump product if MacNeil did not comply (Brink claimed not to take the threat seriously). It must follow, despite Diab’s initial denial and subsequent purported absence of recollection in oral evidence, that these statements were made by Le Riche to Diab at the meeting.

If it is accepted that DPI made pricing proposals and told Flo-Tek and MacNeil that they would need to adhere to the ‘way

things worked’ in Cape Town, there is no evidence that Hart for Flo-Tek or Diab for MacNeil gave any external manifestation of their alleged absence of agreement. Petzetakis’ version in its answering affidavit, confirmed by its attendee Lombard, was an admission that previously agreed pricing principles had been reaffirmed. Since Petzetakis was a long-standing core member of the cartel, and since it would have been in Petzetakis’ interest to minimise the extent of its admitted involvement, the admission in my view must carry some weight. And Le Riche clearly gained the impression that there was agreement.

Questions of precise timing aside, it is common ground that there were two further meetings between the same firms in the second half of 2007 and that MacNeil was represented at these meetings by Brink. These meetings (Le Riche’s third and fourth meetings) were held at Flo-Tek’s premises. These were not mentioned in the referral affidavit and there was thus no specific traversal of the version to which Le Riche subsequently attested. However, both Flo-Tek and MacNeil in their answering affidavits volunteered information about these further meetings.

In its answering affidavit Flo-Tek placed the meetings in July 2007 and September/October 2007 respectively. Flo-Tek alleged11 that the meetings (Le Riche’s third and fourth meetings) were called because Flo-Tek and MacNeil ‘had declined to implement the proposals made by Le Riche in the February 2007 meeting’, that Le Riche tried to convince Hart and Brink to implement DPI’s proposals but that Hart ‘did not agree’ to follow the proposals made on behalf of DPI and

Petzetakis. This version would imply that at least DPI and Petzetakis (the latter represented by Lombard) were going along with

the proposals.

Brink said in MacNeil’s answering affidavit 12 that the last two meetings took place in September and October 2007 and that he attended only because of an invitation by Flo-Tek which was a substantial customer. Elsewhere in his affidavit, though, he said13 that during 2007 MacNeil was invited to attend meetings, ‘which I understood were held for the purposes of fixing prices in accordance with [DPI’s] requirements’ and that Brink refused to attend those meetings ‘save those which occurred in September and October 2007’ which he attended only because of an invitation from Flo-Tek, a substantial customer. It would thus appear from MacNeil’s answering affidavit that its version is that both Hart and Brink attended the last two meetings knowing the purpose thereof. Be that as it may, Brink said, regarding the first of the two meetings he attended (Le Riche’s third meeting), that Le Riche complained that MacNeil’s prices were ‘too cheap’, a remark to which Brink decided not to respond – he stated in the affidavit that because of the threats that had been made by Le Riche to Diab at the February/March 2007 meeting, he (Brink) was disinclined to be confrontational. Le Riche produced a DPI pricelist

and ‘suggested’ that the other firms use the list with an agreed discount structure. Le Riche also suggested that the firms keep a tender register and that Le Riche would decide the allocation of tenders. Brink stated his recollection to be that no one else at the meeting was ‘impressed’ by Le Riche’s suggestions ‘and that nobody agreed to the terms thereof’. Brink says that he certainly did not agree and that MacNeil did not in fact implement the proposals.

Brink then mentioned14 a similar meeting in October 2007 (Le Riche’s fourth meeting) ‘at which Le Riche again attempted to impose [DPI’s] requirements on MacNeil’. Again, said Brink, he did not agree.

In its answering affidavit Petzetakis did not deal with the third and fourth meetings. Petzetakis’ Harding stated that she had taken over as managing director in May 2006. She said she became increasingly uncomfortable with Petzetakis’ involvement in illicit collusion and that on 28 May 2007 she announced to a special meeting of her sales managers that Petzetakis would no longer participate in any meetings, phone calls or discussions with the other firms. She stated in her affidavit that she later learnt from the Commission’s investigations that two employees, one of whom was Lombard, had, in violation of this instruction, participated in discussions with competitors and that they were disciplined.15 Accordingly, Lombard’s participation in the alleged collusive meetings in the second half of 2007 is not inconsistent with Petzetakis’ affidavit but no particulars of the conduct appear from that affidavit nor did Lombard testify.

Le Riche’s version in his witness statement, subsequently confirmed in his oral evidence, was that these two meetings were initiated by Hart of Flo-Tek and reaffirmed the pricing principles previously agreed. At the last meeting there was also discussion ‘regarding the difficulties experienced in ensuring that the agreement regarding the pricing of pipes and fittings for the civils market was adhere to’.16

It is not in dispute that the third and fourth meetings were held at Flo-Tek’s premises. Although Flo-Tek in its answering affidavit said that Le Riche ‘chaired’ the meetings, Flo-Tek did not say that DPI initiated them. Le Riche testified that both meetings were called at Hart’s instance. Brink was invited to attend the meetings by Hart but did not say (and probably could not say) who initiated the meetings. Given that the meetings were held at Flo-Tek’s premises, it strikes me as somewhat implausible that the exculpatory version offered by Flo-Tek in its answering affidavit represents the full truth. We know that Flo-Tek subsequently reached a settlement with the Commission in which it admitted that during 2007 it had contravened s 4(1)(b) by attending a series of meetings at which agreements, arrangements or understandings to fix prices and discounts in respect of PVC pipes were reached. The settlement was made an order on 27 October 2010.

One of the lines of cross-examination pursued with Le Riche was that even on his own version he could not have understood there to be agreement on prices at the last two meetings. In that regard Le Riche said the following when testifying in chief:17

‘Our last meeting was round about November 2007 and that was the one between Flo-Tek, Petzetakis, MacNeil and ourselves. It wasn’t

really working, people weren’t adhering to it, in fact the last two meetings we had, for me it was an exercise just to go and listen to market information. We discussed prices, but I had been in the industry a long time and you know, sometimes guys shared information, which was beneficial, which worked to our advantage. So the last two meetings for me, weren’t, I realised they weren’t working and after November they just ceased. It was a waste of time.’

Importantly, though, he added that after the last of the meetings he received several telephone calls from Hart accusing DPI of not sticking to the agreed prices. Le Riche then indicated that nobody was really adhering to the prices anymore so DPI would be

‘doing our own thing from now onwards’.

When cross-examined by Petzetakis’ counsel, Le Riche said that there was still agreement reached at these meetings but ‘the relevance of the meetings started fading in the last two meetings we had’.18

When Mr Rosenberg pursued a similar theme,19 Le Riche said there ‘were still discussions all the time as far as set prices were discussed’ and that it ‘really only started unravelling at the seams in 2007, after the very last meeting where I mentioned the four manufacturers involved’ (my emphasis). He was shown certain figures which indicated that MacNeil had not followed DPI pricing. He said this did not surprise him but that the advantage of the meetings was that ‘you still set the bar pretty high and it left a lot of gap for you to operate in’ so there was definitely an advantage in having the meetings. Although by the time of last two meetings he thought it was a waste of time but

‘by still sitting in the meetings, it still gave us an indication of thinking of what the other guys felt where the prices should be, but when you sat in that meeting, some would table a price and all the other guys would sit around and say we are not happy with it; you either lift it or move it up and by that you can understand where they want to base it.’

It was put to him that he did not for one minute labour under the misapprehension that anyone was going to follow the agreed prices, to which he replied: ‘The last two meetings, no.’

However, when Mr Rosenberg later cross-examined Le Riche on the details of the last two meetings, Le Riche said the following:20

‘MR LE RICHE: Philip Brink was quiet at the meeting because he was a new entrant. It’s the first time I had met him. Whether Lombard had met him before, I am not sure. He was, as I say, quieter than especially Shaun Hart, but when we discussed prices, he would agree and he would say yes, I’m happy with that or let’s change it slightly.

ADV ROSENBERG: He says the prices were proposed or, in fact, put on the table and he merely listened. He was a youngster. He hadn’t met you. He listened and was non-committal.

MR LE RICHE: No, he was party to it all.’

….

MR ROSENBERG: … Those two meetings [ie the last two meetings] you say were meetings which were effectively not to be taken seriously for the purpose of any firm commitment to pricing.

MR LE RICHE: That was my opinion and my feeling, but the meeting went on as normal. We all sat down. We agreed on prices. We discussed exactly what the prices should be. I had been in the game quite a long time. So, I saw an advantage there to listen and hear what was going on. In fact, Trevor Lombard from Petzetakis, when we walked out of the one meeting, also said to me, you know, we gathered some information here.

ADV ROSENBERG: Well, it appears that at these two meetings at the very least everybody attended merely for the purpose of gathering some information and nobody believed that there was any commitment to setting or adhering to prices.

MR LE RICHE: There was a lot of input put in setting those prices…A lot of time was spent on it, a lot of input from each manufacturer was given and we all left the meeting in agreement that’s the price list.

ADV ROSENBERG: I put it to you that on your evidence that any meeting with Mr Brink in July 2007 and thereafter in October 2007 or any meeting with Mr Brink during that period at which ostensibly there was any kind of commitment whatsoever to pricing, was one which you would not have taken seriously.

MR LE RICHE: I think you may be twisting my words a bit, because it was a serious meeting. I am going back to my experience in the game. I knew how to play the game and I knew that let me use this as an information gathering thing, but as far as I was concerned, everybody had agreed to the prices and a lot of people were applying them, especially at tender stage. There was a clear indication we followed it and I could see it on tenders that were out there in the open marketplace. All the manufacturers were their tendering at those rates.

…

MR ROSENBERG: I say on your evidence there can be no question that MacNeil gave an undertaking as understood by you that it would be charging agreed prices.

MR LE RICHE: In those meetings with Philip Brink there was agreement and consensus reached that that is the price list and that’s what we will all stick to. He never once said he doesn’t agree. He never once said he is not sticking to it. He never once did not participate. So, when the meeting was adjourned, my understanding was there was firm agreement that the guys would implement it. It was just the last two where in my opinion it was a waste of time.’

I have already quoted a passage from Le Riche’s evidence in re-examination where he said that the second, third and fourth meetings ended on the basis that there was agreement and that Le Riche never picked up that anybody dissented.

Viewing Le Riche’s evidence in its totality, his version is clear that by words or conduct the four firms agreed with the proposed prices and discount structures though by the time of the last two meetings he (Le Riche) were starting to suspect that

the meetings were futile. The very fact that Hart, after the last meeting, phoned Le Riche to complain that DPI was not adhering to the agreed prices shows that an understanding to follow a pricing structure had, on Le Riche’s version, been reached. On his evidence, each of the participants, despite whatever private reservations they may have harboured, agreed to adhere to a price list capped discounts.

Brink’s oral testimony in chief was that he attended both of these meetings at the invitation of Hart. At the first meeting he was very quiet and basically said nothing. He arrived late for the meeting. There was already a pricelist on the table – he could not remember whose, and there was a discussion around tender pricing. He denied having committed MacNeil to charging prices with reference to any other firm’s pricelist.21 This version is at variance with his affidavit, where he said that during the meeting Le Riche produced a DPI pricelist and suggested that the list be used by the other firms as a base pricelist with an agreed discount structure.

He claimed in chief not to recall what was discussed at the second meeting but denied that there was an agreement to set prices.22 This is again at variance with his affidavit where he said that the last meeting was similar to the September 2007 meeting and that Le Riche again tried to impose DPI’s requirements on MacNeil.

Under cross-examination23 he accepted that a price list was discussed at each of the two meetings though he could not recall whether it was the same price list. He recalled that MacNeil was criticised at these meetings for its low pricing. He admitted that the toll-manufacturing arrangement between MacNeil and Flo-Tek was not discussed at the meeting. It was put to him that para 34.1 of his answering affidavit on behalf of MacNeil stated that during 2007 MacNeil had been invited to attend meetings which he understood were held for the purpose of fixing prices in accordance with DPI’s requirements. When he answered that this statement in his affidavit was incorrect, it was put to him that he was not being truthful. He conceded that although the pricing discussed at the meetings was

contrary to MacNeil’s strategy he did not explain this to the other attendees nor did he show any dissent from the discussions

that were taking place.24

The legal framework

The Commission did not allege that MacNeil’s conduct amounted to a ‘concerted practice’ for purposes of s 4(1). Its case was that an ‘agreement’ was concluded at these meetings between MacNeil and the other three firms and that the agreement constituted the direct or indirect fixing of selling prices contrary to s 4(1)(b)(i). The expression ‘agreement’ is defined in s 1 as including ‘a contract, arrangement or understanding, whether or not legally enforceable’.

In Netstar (Pty) Ltd & Others v Competition Commission of South Africa & Others 2011 (3) SA 164 (CAC) this court said the following in paras 25 and 26 after quoting the definitions of ‘agreement’ and ‘concerted

practice’:

‘(25) … A concerted practice arises from the conduct of the parties, and does not amount to an agreement. A possible example might be the type of cartel arrangement where a market leader signals a price increase by way of public announcement and, in accordance with long-standing practice in the industry, the other participants follow its lead. However, care must be taken not to confuse independent conduct with interdependent conduct. It suffices for present purposes to say that the emphasis is on the conduct of the parties. By contrast, an agreement arises from the actions of and discussions among the parties directed at arriving at an arrangement that will bind them either contractually or by virtue of moral suasion or commercial interest. It may be a contract, which is legally binding, or an arrangement or understanding that is not, but which the parties regard as binding upon them. Its essence is that the parties have reached some kind of consensus. No doubt, in many cases the same evidence may be relied upon as pointing towards either an agreement or a concerted practice. However, sight should not be lost of the fact that they are different. The definition of an agreement extends the concept beyond the contractual arrangement. However, what it requires is still a form of arrangement that the parties regard as binding upon both themselves and the other parties to the agreement. Absent such an arrangement, there is no agreement, even in the more extended sense embodied in the definition. By contrast, a concerted practice examines the conduct of the parties to determine whether it is coordinated conduct or if they are acting in concert. The absence of any arrangement between them, or any belief that they are obliged to act in that fashion, is immaterial.

(26) … The case for a concerted practice is based on evidence that assesses the nature of the conduct of the firms said to be party to the practice. By contrast, the case for an agreement examines whether an agreement as defined was concluded, and that focuses on the existence of consensus between the parties. Even where reliance is placed on the same evidence in support of these distinct cases, it requires separate evaluation…’

Mr Maenetje submitted that the Tribunal correctly found in this case that the requisite consensus existed but argued that to the extent that Netstar restricted the wide definition of ‘agreement’ in the Act it would be incorrect and could be departed from.

I do not consider that the passages I have quoted from Netstar call for reconsideration. This court, in distinguishing between a ‘concerted practice’ and an ‘agreement’ held that the essence of the latter was consensus – some form of arrangement that the parties regarded as binding on each other. In my view, consensus is indeed the essence of the statutory definition of ‘agreement’ – consensus is inherent in the words ‘contract, arrangement or understanding’ (whether or not legally enforceable). Consensus sufficient to constitute a ‘contract, arrangement or understanding’ must be proved on a balance of probability (see s 68) before a finding can be made in terms of s 4(1) that the firms have committed a prohibited practice in the form of an ‘agreement’. This naturally does not mean that the consensus need amount to a contract at private law. Particularly in regard to the per se prohibitions in s 4(1)(b), the parties would, by the very illicit nature of their arrangement, not contemplate legal enforcement. They need not even have agreed upon a punishment mechanism. Furthermore, the content of the consensus need not, I venture to suggest, rise to the level of precision sufficient to satisfy the requirement of certainty applicable to private law contracts, ie the precision needed to defeat an argument that the alleged agreement is void for vagueness.

In the present case, the question is whether consensus was reached to fix prices (whether directly or indirectly) at the three 2007 meetings attended by MacNeil. Although the precise content of the price lists and discounts structures which were discussed at each of the three meetings were not established, it was never suggested that the proposals, made with reference to price lists and involving the capping of discounts with reference to those price lists, were insufficient to amount to price-fixing if there

was consensus on them. The degree of precision required for consensus is thus not an issue in this appeal. The question is

whether consensus on the proposals existed by the end of each meeting.

Because Diab and Brink portrayed themselves as passive participants at the meetings in question, the Tribunal in the present case referred to its earlier decision in Competition Commission v Aveng (Africa) Ltd t/a Steeledale & Others [2012] ZACT 32 where, in para 16, the Tribunal quoted with approval the following passage from Aalborg Portland A/S v Commission of the European Communities,25 decided by the European Court of Justice in January 2004 (case citations omitted):

‘(81) According to settled case-law, it is sufficient for the Commission to show that the undertaking concerned participated in meetings at which anti-competitive agreements were concluded, without manifestly opposing them, to prove to the requisite standard that the undertaking participated in the cartel. Where participation in such meetings has been established, it is for that undertaking to put forward evidence to establish that its participation in those meetings was without any anti-competitive intention by demonstrating

that it had indicated to its competitors that it was participating in those meetings in a spirit that was different from theirs…

(82) The reason underlying that principle of law is that, having participated in the meeting without publicly distancing itself from what was discussed, the undertaking has given the other participants to believe that it subscribed to what was decided there and would comply with it.’

There are statements to similar effect in later European competition case law. We were referred in argument, for example, to the relatively recent decision of the General Court (7th Chamber) in DKKK Kaisha & Others v European Commission delivered in February 2012 where, in para 52, the statement in paras 81-82 of Aalborg was essentially repeated. The court added the following in para 53 (case citations again omitted):

‘It must be pointed out in this regard that the notion of publicly distancing oneself as a means of excluding liability must be interpreted narrowly. In order to disassociate itself effectively from anti-competitive discussions, it is for the undertaking concerned to indicate to its competitors that it does not in any way wish to be regarded as a member of the cartel and to participate in anti-competitive meetings. In any event, silence by an operator in a meeting during which an unlawful anti-competitive discussion takes place cannot be regarded as an expression of firm and unambiguous disapproval. A party which tacitly approves of an unlawful

initiative, without publicly distancing itself from its content or reporting it to the administrative authorities, effectively

encourages the continuation of the infringement and compromises its discovery…’ .

The tribunal dealt with this topic later in its judgment in relation to the case against Andrag (see paras 87-93). In that context reference was made not only to Aveng but also to United States law. The Tribunal quoted the proposition by Areeda and Hovenkamp26 that

‘... there will be an agreement even though the challenged arrangement falls short of forming a contract because, for example, the parties declare an intention not to be legally bound, each party reserves the right to abandon the venture at will and without notice, offer and acceptance are not fully in accord, or the understanding is too vague to allow a court to enforce it (even though it were not illegal)’.

In dealing with Andrag, the Tribunal said, with reference to Netstar, that on the Tribunal’s interpretation of that judgment there had to be ‘an element of consensus’ and there had to be some ‘form’ of arrangement in order for there to be an agreement. Netstar did not, said the Tribunal, decide ‘what form that consensus needs to take or what facts would suffice to infer consensus’. The Tribunal observed that the latter issue had been considered by the courts in the United States and Europe which had determined that an agreement may be inferred to exist in certain circumstances for anti-trust purposes, even though they might be insufficient for a conclusion that an agreement existed at common law. The Tribunal then said the following in paras 97 and 98:

’(97) Contract law is concerned with the private consequences of agreements; competition law is concerned with the public consequences. Formal offer and acceptance, whilst rigid concepts in the common law of contract, become more supple in the competition law form of the concept. Within the ambit of competition law, in a particular context, where there is a duty to speak, silence or an ambivalent answer may suffice as well as the witnessed signature, to signify assent.

(98) Thus when competitors meet, even on an unintended occasion for some, once the conversation moves to proposals of an unlawful agreement, those attending must repudiate the proposal by conduct in unambiguous terms, however awkward it may be to do so, lest the other firms present reasonably infer that the accused firm had assented.’

In my respectful view, the Tribunal’s observation that offer and acceptance are ‘rigid concepts’ in our common law of contract is not correct. Express offer and acceptance are not required – the formation of contractual consensus may

be proved in a variety of ways and often it will be unnecessary and difficult to isolate the distinct elements of offer and acceptance.

I also do not understand the need to ‘re-interpret’ the lucid language in the passages I have quoted from Netstar. In my opinion it is unnecessary for purposes of the present appeal to determine whether this court’s approach to passive attendance and public distancing should be exactly as reflected in the above passages from European case law or exactly in line with United States law. One is ultimately concerned with the factual question whether a sufficient consensus was achieved to constitute an ‘agreement’ as defined in our Act and as explained in Netstar. One should not substitute, for this relatively simply stated enquiry, rules from cases which may reflect no more than a manner of assessing evidence or which incorporate strictures not flowing naturally from the language of our Act.

That having been said, the basic rationale of the European and American cases, namely that passive participation without public distancing is sufficient because it creates in the minds of the other participants the belief that the passive participant has subscribed to the arrangement and intends to comply with it, is not inconsistent with South African law. It has long been accepted in our private law of contract that a person cannot escape from an apparent agreement merely because his subjective intention differed from the apparent agreement. This is known as the doctrine of quasi-mutual assent. In Sonap Petroleum (SA) (Pty) Ltd v Pappadogianis 1992 (3) SA 324 (A) at 239F-240B the court said that in various earlier decisions our courts had adapted, for purposes of the facts of their respective cases, the well-known dictum of Blackburn J in Smith v Hughes (1871) LR 6 QB 597 at 607:

‘If, whatever a man’s real intention may be, he so conducts himself that a reasonable man would believe that he was assenting to the terms proposed by the other party, and that other party upon the belief enters into the contract with him, the man thus conducting himself would be equally bound as if he had intended to agree to the other party’s terms’.

See also, for example, Pillay & Another v Shaik & Others 2009 (4) SA 74 (SCA) paras 55-60; and see Christie The Law of Contract in South Africa 6th Ed at 10-12.

In the present context, another relevant consideration is that under certain circumstances our law imposes on a person a duty to speak, and a failure to do so where the duty exists may amount to an objective manifestation of consent, regardless of the subjective intention of the silent party (Christie op cit at 70-71 and the case there cited). The question whether a duty to speak exists in particular circumstances is ultimately, it seems to me, a question of policy and fairness. In relation to delictual misrepresentation by silence in a contractual setting the duty to speak is part of the wrongfulness enquiry which is avowedly policy-laden (see ABSA Bank Ltd v Fouche 2003 (1) SA 176 (SCA) paras 4-5). I do not see why the position should be different when the question, for purposes of quasi-mutual assent, is whether a party by his silence has represented that he acquiesces. In McWiliams v First Consolidated Holdings (Pty) Ltd 1982 (2) SA 1 (A) Miller JA said that acquiescence might be inferred where, in ‘ordinary commercial practice and human expectation’ a firm repudiation would be ‘the norm’ (10E-F). In the context of competition law, competing firms may have occasion to meet and discuss matters which are entirely innocent. However, if a firm’s representative attends a meeting of competitors knowing that collusive activity will be discussed, or if he finds after arrival at the

meeting that collusive activity is being proposed, I have little difficulty in saying that in general the representative would be under a duty to distance himself from the proposals under discussion, either by leaving or by stating that he wants no part of them; in other words a ‘firm repudiation’. In the private law of contract, the duty to speak (or perhaps more accurately, the peril of remaining silent), where it exists, has to do with fairness to the other party to the arrangement; where one is dealing with attempts by firms to reach an ‘agreement’

prohibited by s 4(1), on the other hand, the duty to speak is not concerned so much with fairness to the other firms but with the harm that could be caused to members of the public if the other firms should proceed to conduct themselves on the reasonable

assumption that the passive attendee assented to their proposals. Cartels are after all the most egregious form of anti-competitive

conduct.

Ultimately, though, it is a question of fact whether the passive attendee’s conduct is such as reasonably to create in the minds of the other participants his assent to the proposals under discussion (and it is the creation of that reasonable impression which poses the risk of competition harm). I do not exclude the possibility that there may, in a particular case, be surrounding circumstances which show that the other participants could not reasonably have understood the passive attendee to have assented to their proposals. That conclusion necessitates a careful assessment of the facts to which I now turn.

Evaluation

The evidence must, as I have said, be viewed holistically. Contrary to Mr Rosenberg’s submission, I do not accept that Flo-Tek’s answering affidavit should have been left out of account because Hart did not give oral evidence. The fact that he was not called (conceivably either the Commission or MacNeil could have called him) is a factor to be borne in mind in assessing the weight of the affidavit evidence. The same goes for Petzetakis’ answering papers. The Tribunal does not operate as a court of law in accordance with strict rules of evidence (see, for example, s 52(2) and s 53(3)). The answering affidavits of

the respondent firms constitute evidential material. In the present case, confirmatory affidavits were filed by the individuals who had personal knowledge of the relevant facts (Lombard and Hart).

In my view, the evidence I have summarised establishes on a balance of probability that pricing structures were proposed and discussed at three meetings in 2007, the participants being DPI, Petzetakis, Flo-Tek and MacNeil (these would be Le Riche’s second, third and fourth meetings). Diab represented MacNeil at the first of the meetings, Brink at the other two. Le Riche’s oral

evidence was that pricing was discussed and agreed at each meeting. Petzetakis admitted this in relation to the first meeting, which was the only one alleged in the referral affidavit; and admitted that after May 2007 Lombard attended further collusive meetings. Given that DPI and Petzetakis were long-standing members of the cartel, there is no reason to doubt that at least Le Riche and Lombard would have expressed concurrence with the setting of prices.

MacNeil’s answering affidavit indicated that the meetings to which MacNeil was invited in 2007 were initiated for the purpose of fixing prices in accordance with DPI’s requirements. Brink acknowledged that it was common knowledge in the industry that DPI wished to have fixed prices in the market, and Brink was aware, from his previous employment with Petzetakis, that the manufacturers were accustomed to meeting with each other. Diab and Brink offered no plausible explanation for their attendance at the meetings other than to discuss pricing. Although they were invited to attend on each occasion by Hart and although Flo-Tek was a substantial customer, neither Diab nor Brink maintained that they thought they were going to a bilateral meeting to discuss the tolling arrangement between MacNeil and Flo-Tek. Indeed, the February/March 2007 meeting took place at DPI’s premises, so Diab could not have thought this.

Diab tried to portray the meeting he attended in February/March 2007 as amounting to no more than ‘general chatter’. The Tribunal found him to be a poor witness, and a reading of the transcript does not bring me under any different impression. Although the Tribunal was inclined to discount Le Riche’s evidence concerning this particular meeting, I have already explained why, in my view, his testimony was indeed germane, even though Le Riche misidentified MacNeil’s representative at the meeting. The fact that pricing proposals were tabled and discussed was in line with Le Riche’s evidence and with the answering affidavits of Petzetakis and Flo-Tek. Flo-Tek subsequently conceded liability in circumstances where the only allegations against it were the same as those levelled against MacNeil. MacNeil’s own answering papers were to the effect that MacNeil’s low pricing was discussed and that Diab was told to see to it that MacNeil was brought into line. Despite Diab’s evasion on this point in oral evidence, Brink confirmed that Diab told him this. The ‘line’ that MacNeil was to toe was clearly the price lists and discounts structures which, according to the other evidence, were discussed at the meeting. Diab did not

claim to have articulated any dissent or to have done anything else to show that he would not tell Brink about the ‘line’

which was to be followed.

In my opinion, the evidence established on a balance of probability that Diab, by his failure to protest at the meeting of February/March 2007, created the reasonable impression in the minds of the other attendees that MacNeil would go along with the proposals. DPI and Petzetakis probably gave their express support. Flo-Tek’s later admission of liability and payment of a fine, and the fact that after the subsequent two meetings Hart was badgering DPI about non-compliance, suggests that Flo-Tek may also have expressed concurrence. There is at any rate no evidence from either Le Riche or Diab that Hart articulated any opposition. A

statement that a particular attendee ‘did not agree’ to a proposal is a vague and generalised conclusion – one is more concerned with the objective manifestations of the participant’s reaction to a proposal than to a statement which could amount to no more than the attendee’s subjective mental attitude. Similarly, to say that a participant was ‘not impressed’ by a proposal is merely an assertion as to his state of mind. More is required in order to establish that there was a firm repudiation.

As to the other two meetings, Brink (against whom the Tribunal made no express credibility finding but whose evidence in transcript comes across as evasive on essential details) acknowledged that Diab had reported to him that MacNeil was expected to fall into line. Brink’s answering affidavit for MacNeil stated the purpose of the two meetings he attended. The fact that he did so at Flo-Tek’s invitation is neither here nor there. It is inherently implausible that he would have attended the meetings

without knowing their purpose. He did not suggest that he was informed in advance of any legitimate business to be discussed. Why would he attend, one may ask rhetorically, if he did not intend to participate in pricing discussions? I can understand that privately he may have intended to disregard the proposals but he certainly did not attend with a view to stating clearly that position because, on his own evidence, he did not do so. One is driven to conclude on the probabilities that he did not want to be seen to be breaking ranks, since MacNeil, as a relatively new player with a small market share, might have been vulnerable to retaliation. Brink himself said that he did not wish to be confrontational with Le Riche because of the dumping threat which had previously been made. It appears, further, that Hart, who represented Flo-Tek, which was a major customer of MacNeil, was probably a supporter of the proposals, contrary to the picture painted in Flo-Tek’s answering papers. The last two meetings took place at Flo-Tek’s premises, and it was Flo-Tek that got MacNeil involved. As I have said, there is no evidence that

any other legitimate business was scheduled for discussion. Le Riche’s unchallenged testimony is that Hart phoned him after

the last meeting to complain that DPI was not adhering to the pricing arrangements. This indicates that Hart understood an

arrangement to have been reached (as we know, Flo-Tek did in the event settle with the Commission and pay a substantial fine). Brink did not testify that Hart made any protest at the meeting against the proposals; and Brink did not claim himself to have said anything to distance MacNeil. The defence was in essence a contention that there was no agreement because Brink did not expressly say that he agreed and because privately he had no intention of complying. But, on a preponderance of probability, that is not the impression which his conduct reasonably created in the minds of the other attendees in circumstances where, as I have found, there was a duty on Brink’s part to speak if he was against the illicit arrangement.

Mr Rosenberg, who did not in principle challenge the Tribunal’s legal approach based on the European cases, said that liability on the strength of passive attendance and non-distancing presupposes that the passive firm attended a meeting where some agreement was actually reached (ie between others). I think this misses the point that, in circumstances such as the present, the very fact that an ‘agreement’ (within the broad definition of the Act) has been struck can be inferred from silence because there was a duty to speak and because silence in the face of such duty may create the reasonable impression of consensus (in private law, the doctrine of quasi-mutual assent). This could be so even where the discussion is bilateral and the one party is

silent. As it happens, the meetings with which we are concerned here were multilateral, and, on the probabilities, the attendees must have been aware of the purpose of the meetings. There is also no reason to doubt that at least DPI and Petzetakis, and quite possibly Flo-Tek as well, positively assented to the proposed arrangements, and that the proposals to which they thus assented were intended to encompass all four participants. (This assumes in favour of MacNeil that its representatives were indeed passive listeners. The evidence of Le Riche indicates otherwise, and Diab conceded in cross-examination that he had made some contribution. It would be remarkable, I think, if one of four attendees at a meeting should say absolutely nothing when pricing was being discussed.)

Mr Rosenberg argued that it was implausible that the other participants would have believed that MacNeil concurred in the proposals, given that the cartel in the industry had collapsed by 2007. We do not have the material before us to assess that question in regard to other firms nor in relation to the cartel of which the long-standing core members were DPI, Petzetakis and Marley, except

that we know that Petzetakis’ managing director instructed her sales managers in May 2007 to refrain from further participation and that Lombard was one of two sales managers who subsequently breached this instruction. The question is not whether MacNeil was part of some other pre-existing cartel which was still in existence in 2007 or even whether the 2007 conduct in which MacNeil was implicated should be described as a cartel. It is enough that there is evidence that there were three meetings at which, on a balance of probability, an agreement, understanding or arrangement on price-fixing was reached and where the parties either said or gave the impression that they would comply.

The same goes for Mr Rosenberg’s argument that Le Riche on his own version had no confidence in the utility of such meetings by September/October 2007. I have already given my assessment of Le Riche’s evidence on that score.

I thus do not consider that there are grounds for interfering with the Tribunal’s findings on the merits of the case.

Penalty

Turning to the administrative penalty of R2 million imposed on MacNeil, the Tribunal arrived at this figure by following the six-step procedure laid down in its Aveng decision, a procedure in turn adopted in the light of this court’s approach and suggestions in Southern Pipeline Contractors & Another v Competition Commission [2011] 2 CPLR 239, where reference was made to the European Guidelines of 2006.27 In summary, the Tribunal applied this approach as follows: step 1 – the affected turnover (PVC sales) was found to be R29,2 million (taken from MacNeil’s 2007 year); step 2 – the base amount was set at 15% of the affected turnover, ie R4,38 million (half of the of 30% reserved for the most egregious cases); step 3 – a multiplier of 7/12ths was used (on the basis that MacNeil was only involved in collusion for about seven months), yielding a figure of R2,5 million; step 4 – it was not necessary to round down the figure in step 3 because it did not exceed the s 59(2) cap; step 5 – the figure of R2,5 million from step 3 was then reduced to R2 million because mitigating factors were thought to outweigh aggravating circumstances; and step 6 – the figure thus determined (R2 million) was within MacNeil’s total turnover for the most recent financial year for which the Tribunal had figures (2009), so no further reduction was required.

Although Mr Rosenberg’s main heads of argument did not foreshadow an attack on the six-step approach in principle, he delivered a supplementary note in which he argued that Southern Pipeline Contractors contemplated a simpler process in which, after determining affected turnover in the first step, one simply selected a percentage of the affected turnover after balancing all the factors specified in s 59(3). He said that step 2 of the Tribunal’s Aveng approach only had regard to the objective gravity of the prohibited conduct and thus potentially set the base amount too high.

The imposition of administrative penalties is a matter entrusted by s 59(1) to the discretion of the Tribunal. The Tribunal is entitled, for the guidance of interested parties, to indicate the process of reasoning it will ordinarily follow in determining a penalty, provided the guidance is not inherently flawed and provided the Tribunal appreciates that it always retains a discretion which it cannot fetter, thus allowing itself to depart from its own guidelines in appropriate circumstances (cf Kemp NO v Van Wyk 2005 (6) SA 519 (SCA) para 1). Although this court in Southern Pipeline Contractors adopted a more abbreviated process of reasoning in disposing of that appeal, this did not preclude the Tribunal from subsequently adopting for itself the guidance stated in Aveng. Although the process of reasoning has been broken down by the Tribunal into additional steps, the steps in their totality permit all the factors listed in s 59(3) to be taken into account. In step 2 the Tribunal considers those of the s 59(3) factors which concentrate on the objective features of the contravention while in step 5 the Tribunal accommodates the s 59(3) factors which are individual to the respondent firm. Although the objective features of the prohibited conduct may result in a relatively high amount after the application of steps 2 and 3, step 5 places no limit on the extent to which the Tribunal can reduce the overall penalty where in its view mitigating factors outweigh aggravating factors. (On the other hand, step 6 imposes an ultimate cap on the extent to which the amount in step 5 can be increased where in the Tribunal’s view aggravating factors outweigh mitigating factors.) Mr Rosenberg was unable to explain to us why his preferred approach, in which one simply determines in the

round a percentage to apply to affected turnover, would necessarily yield different or fairer results.

Mr Rosenberg’s attack on the Tribunal’s six-step guidelines was intertwined with a contention that in the present case this had led the Tribunal, in step 2, to treat the prohibited conduct as being part of a cartel which had existed for a

number of years, even though MacNeil itself was only drawn into a price-fixing conduct during 2007. This inevitably resulted, so it was argued, in a relatively high amount in step 2. I accept that it would have been a misdirection for the Tribunal in step 2 to have had regard to broader prohibited collusive conduct to which MacNeil was not itself a party. However, I do not think the Tribunal fell into that error. It is quite clear from para 46 of the Tribunal’s reasons that in step 2 it had regard only to the limited regional collusive conduct constituted by the three meetings between the four particular firms in question held over the period February to September 2007. And in step 3 the Tribunal recognise the limited duration of the prohibited conduct by reducing the base amount by 7/12ths.

Mr Rosenberg, for the rest, argued that the penalty here was too severe, having regard to dicta in Southern Pipeline Contractors and Federal-Mogul Aftermarket Southern Africa v Competition Commission & Another [2005] CPLR 50 (CAC) that there needed to be proportionality between the sanction and the firm’s degree of blameworthiness and that the penalty should not only advance the object of deterrence but also not lose sight of fairness to the offending firm.28

In pressing this argument, Mr Rosenberg pointed out that the penalty of R2 million was R600 000 more than MacNeil’s total profit of R1,4 million for the 2007 year. He also listed various features in mitigation (that MacNeil was a new entrant in 2006/2007, was pricing below market prices, that its market share was small and its level of profitability low, that its involvement was regional not national, that it did not implement the collusive prices, and that there was no evidence of harm to consumers).

Mr Maenetje for his part said that there was no basis for interference, reminding us of the statement in Federal-Mogul that this court does not enjoy an unfettered discretion to substitute its own view for the Tribunal’s assessment and determination of an administrative penalty and that interference is permissible only when the Tribunal has acted capriciously or on a wrong principle or not brought an unbiased judgment to bear on the question or has not acted for substantial reasons.

As Mr Rosenberg acknowledged, the mitigating factors listed in his submissions were taken into account by the Tribunal (though insufficiently, he submitted). This occurred in two stages of the Tribunal’s reasoning: in setting the base amount at 15% rather than 30% of affected turnover; and in reducing the base amount from R2,5 million to R2 million.29 (In terms of para 25 of the EU Guidelines it is recorded that the Commission will, in cases involving horizontal price-fixing, market-sharing and output-limitation agreements, include in the base amount a sum of between 15% and 25% of the affected turnover, regardless of the duration of the firm’s participation in the infringement – this in order to deter undertakings from entering into such agreements.) There is a further feature of the Tribunal’s assessment which was favourable to MacNeil. The Tribunal used a multiplier of 7/12ths in stage 3 of the Aveng procedure which reduced the base amount of affected turnover from R4,38 million to R2,5 million. The Tribunal may have been justified in taking a full year into account – the period March to October 2007 is eight months, and the Tribunal’s approach assumed in favour of MacNeil that the prohibited conduct ceased with the last meeting, whereas quite plausibly the collusion gave rise to ongoing effects beyond October 2007, even if MacNeil itself was disregarding the discounts fixed at the meetings. (I note in passing that para 24 of the EU Guidelines states that, in determining the multiplier, periods of less than six months will be counted as half a year and periods longer than six months but shorter than one year will be counted as a full year.)

The penalty of R2 million, expressed with reference to the statutory cap of 10% specified in s 59(2), was (as the Tribunal observed in footnote 28 of its decision) only 2,8 % of MacNeil’s most recent turnover (taken as its 2009 year in the absence of more recent information).

As to the criticism that the penalty exceeded MacNeil’s 2007 profit, the six-step approach which the Tribunal now follows and the similar procedure followed by the Commission in Europe work with turnover, not profit. This does not mean that profitability is irrelevant – the Tribunal is enjoined to take into account, as one of the factors relevant to the assessment of a penalty, the level of profit derived from the contravention, and in the present case the Tribunal indeed took this into account when it recognised, as mitigating factors, that MacNeil had not implemented cartel prices, that it priced below the other firms and that its level of profitability in 2007 was low (its after-tax profit margin in 2007 was about 3%). This does not mean, however, that a firm’s total profit, or its profit from the contravention, should be utilised in the calculation of penalties. As was stated in Southern Pipeline Contractors (para 61) the statutory cap of 10% of full turnover for the most recent financial year has been set lest a penalty should be ‘destructive

of the offending party’s business’. Of course, s 59(2) is inevitably a blunt instrument to achieve this constraint, and no doubt the Tribunal could receive and take into account evidence that a penalty arrived at along conventional lines would jeopardise the firm’s viability (cf para 35 of the EU Guidelines). Such a case was not advanced by MacNeil at the hearing nor did Mr Rosenberg make that argument to us.

I can understand why competition regulators would be reluctant to work with annual profit in determining penalties. While an assessment of turnover or affected turnover may sometimes present difficulty, it will usually be possible to extract the number with relative ease from the firm’s annual accounts or financial records. Profit, by contrast, depends on the accounting policies of the company and the precise way in which its affairs are structured. A company’s profit may be reduced by the payment of management fees or interest to a related company. Different companies may follow differing depreciation policies. And what ‘profit’ does one use? Here, MacNeil’s operating profit of R2 314 681 for 2007 exceeded the penalty though its profit after financing costs was R1 432 972 and its profit after tax R1 432 972.30 Another consideration militating against the use of annual profit figures is that, unless the company’s entire turnover is affected turnover, it would often be a matter of considerable difficulty to determine what profit the company had generated from affected turnover. Some method of allocating common items of expenditure would need to be determined. (It would not be rational to work with the overall profit on affected and unaffected turnover – the overall profit might be reduced by some unrelated business activity.) The utilisation of profit in quantifying administrative penalties would thus lead to considerable complexity.

Given the object of deterrence, there is also no reason why a penalty should not exceed a firm’s annual profit. The Aveng methodology, which is not dissimilar to that contained in the European Guidelines of 2003, does not have an underlying assumption that the product of the procedure will be an amount lower than the firm’s annual profit. Indeed, in many instances the

percentage adopted in arriving at the base amount of affected turnover (in this case, 15%) could be expected to exceed the firm’s

profit margin. One must also bear in mind that collusive firms may operate inefficiently. The harm that consumers feel is related to the amount they spend on the collusive firms’ products or services (turnover), not to the firms’ profits.

Despite these various considerations which might be thought to support the penalty ultimately imposed by the Tribunal, there were weighty mitigating factors. A fine of R2 million strikes me as disproportionate, having regard particularly to the fact that MacNeil was a new entrant with a small market share and that MacNeil priced aggressively and did not, despite its ostensible

acquiescence at meetings, implement the prices discussed at those meetings. The difference between the penalty I would have imposed

and the one imposed by the Tribunal is sufficiently striking to lead me to conclude that the Tribunal misdirected itself in its

assessment of the mitigating circumstances. It must be remembered that in determining the base amount in step 2 the Tribunal has allowed itself a wide range – up to 30% for the most egregious prohibited practices. In the OFT guidelines (mentioned in footnote 27 above) the upper ceiling in determining the base amount is 10%. While I have no difficulty with the larger range (which accords with the EU Guidelines), the use of relatively high percentages in step 2 might give rise to disproportionate penalties if adjustments for significant mitigating factors in step 5 are too miserly. In the present case I consider that the base amount of R2,5 million should have been reduced by R1,25 million in step 5 to give appropriate effect to the mitigating circumstances. This would yield a final penalty of R1,25 million, which differs sufficiently from the penalty imposed by the Tribunal to justify interference on appeal.

Despite the outcome in this particular case, participants in cartel behaviour would do well to remember that while a significant weight of net mitigating circumstances might justify greater reductions than those suggested by the Tribunal’s approach in the present matter, a significant weight of net aggravating circumstances might well justify greater increases in penalties.

Conclusion

It follows that I would allow the appeal to the limited extent of substituting for the penalty imposed by the Tribunal a penalty of R1,25 million. Since MacNeil has failed on the merits but achieved substantial success regarding the penalty, I think it would

be just and equitable for each party to pay its own costs on appeal.

______

ROGERS AJA

DAVIS JP & VICTOR AJA AGREED

APPEARANCES

For Appellant: Adv Sean Rosenberg SC and Mr L Kelly

Instructed by:

Smith Tabata Buchanan Boyes

2nd Floor Buchanan Chambers

Cnr Warwick & Pearce Road

Claremont

For Second Applicant: Adv NH Maenetje

Gildenhuys Malatji Inc

GLMI House

Harlequins Office Park

164 Totius Street

Groenkloof

1Lombard's confirmatory affidavit is not in the appeal record.

2Again, the confirmatory affidavits of Hart and Diab are not in the appeal record.

3Para 51 at 5/451.

49/813-818.

510/886.

6See para 36.6 of Flo-Tek's answering affidavit at 2/252-3.

710/887.

810/896-7.

911/1009-1010.

1012/1100-1102.

11Paras 36.6.22 - 36.6.26 at 3/254.

12Paras 29.4 - 29.8 at 4/269-270.

13para 34.1 at 4/271-272.

14Para 29.9 at 4/270.

15Paras 24-28 at 1/50-52.

16Paras 54-57 at 5/452-3.

179/826-827.

189/841.

1910/870-879.

2010/899-906.

2112/1079-82.

2212/1082-84.

2312/1084-1099.

24See particularly at 12/1084-85 and 12/1095.

25 [2005] 4 CMLR 251.

26Areeda and Hovenkamp Antitrust Law – An Analysis of Antirust Principles and their Application 2nd Ed para 1406c.

27Guidelines on the method of setting fines imposed pursuant to Article 23(2)(a) of Regulation No 1/2003 (2006/C210/02). For a discussion of these guidelines, see Whish & Bailey Competition Law 7th Ed at 276-279. In the United Kingdom the Office of Fair Trading has adopted similar guidelines for the imposition of penalties under s 38 of the Competition Act 1998 c.41 (op cit at 410-413). The current OFT guidelines make provision for a maximum base amount of 10% (not 30%) of affected turnover in its step 1 though it was remarked in Kier Group plc v OFT [2011] CAT 3 para 109 that the OFT had confined itself quite narrowly and it was observed, with reference to the EU Guidelines (where the maximum amount is 30%), that a more generous range would ‘provide more headroom at the outset’. The CAT said that this could be borne in mind when the OFT came to review its guidelines. It should also be noted that the OFT guidelines contain a step 3 (absent from the EU Guidelines) in which the base amount in step 1 as multiplied for duration in step 2 can be increased to achieve the policy objective of deterrence – this is apart from adjustments in step 4 for mitigating and aggravating factors.

28See para 9 of Southern Pipeline Contractors and para 72 of Federal-Mogul.

29The amount at the end of stage 4 would have been R2 570 353: affected turnover of R29 375 468 [record 13/1119] x 15% x 7/12ths.

3013/1166.