COMPETITION TRIBUNAL  
REPUBLIC OF SOUTH AFRICA 
Case No: 08/LM/Feb02
In the large merger between:  
Distillers Corporation (SA) Limited       Primary Acquiring Firm
And
Stellenbosch Farmers Winery Group Ltd Primary Target Firm
______________________________________________________________
Reasons for Tribunal Decision  ­ Non­Confidential
______________________________________________________________
Finding
1. We have found that a substantial lessening of competition is likely in  
the   proprietary spirits market , one of several markets implicated in  
this   transaction.     For   the   reasons   outlined   below   we   find   that   a  
consideration of the claimed efficiency gains is not pertinent.  We have  
found   that   there   are   no   consequences   for   the   public   interest   that  
influence our finding.
2. A further hearing will be convened in order to determine an appropriate  
remedy in respect of that market in which we have found the likelihood  
of a substantial lessening of competition.
The Parties 
Primary acquiring firm 1 
3. Distillers Corporation (SA) Limited (“Distillers”) was a listed investment  
holding company, involved, through its subsidiaries, in the production  
and wholesale distribution of branded spirits, wine and ready to drink /  
flavoured alcoholic beverages (‘RTDs’ or ‘FABs’). Distillers produced,  
1  For reasons that are elaborated below, this merger has already been implemented for some  
considerable time.  Hence we try and refer to the parties in the past tense for the simple reason that they  
no longer exist.  However, we analyse the merger at the time at which it should have been notified and,  
so, a degree of confusion in the tense used is inevitable. Where consideration is given to developments  
subsequent to the notification we have tried to indicate this explicitly.
Non­Confidential version 1

marketed, sold and distributed various brandy brands (including Oude  
Meester, Richelieu, Viceroy, Klipdrift), whisky (Harrier), vodka (Count  
Pushkin), cane (Seven Seas), premium wines (including Fleur du Cap,  
Le Bonheur, Neethlingshof and Grunberger), sparkling wines (J.C. le  
Roux) and liqueurs  (Amarula Cream).   The FABs manufactured and  
distributed by Distillers included Bacardi Breezer, Bernini and Castello.  
Distillers   also   acted   as   the   South   African   agent   and   distributor   of  
international brands such as Gordon’s gin, Martini, Bacardi rum, and  
Glenfiddich whisky.
Primary target firm 
4. SFW was a producer and wholesaler of wine, spirits and alcoholic fruit  
beverages within South Africa. As a leading wine producer, it boasted  
names such as Nederburg, Zonnebloem, Graca, Chateau Libertas and  
Plaisir de Merle. Its spirit brands included Mellow­Wood brandy, Old  
Buck   gin,   Mainstay   cane   spirit   and   Romanoff   vodka.   It   had   the  
distribution rights in SA for Martell brandy 2  and Bols brandy 3. It was the  
market leader in the FABs market with brand names such as Hunter’s  
Dry, Hunters Gold, Crown, Savannah, Esprit, Montello and Manhattan. 
Shareholding structure
5. SFW   and   Distillers   were   both   controlled   by   the   same   trio   of  
shareholders.     The   two   companies   had   an   identical   shareholding  
structure:
∙ Rembrandt­KWV   Investments   (“RemKWV”)   held   60%   of   the  
shares of both parties. RemKWV is a joint holding company of  
Rembrandt   and   KWV,   in   which   each   holds   a   50%   interest.  
KWV’s interest in RemKWV is held through a listed subsidiary,  
KWV   Investments   Limited   in   which   KWV   owns   approximately  
54%;
∙ SAB   held   30%   of   both   companies   through   its   wholly   owned  
subsidiary Other Beverages Industries (Pty) Ltd (“OBI”);
∙ The general public held the remaining 10% of both companies.

∙ The general public held the remaining 10% of both companies.
2  The Martell brandy contract was subject to litigation between SFW and Seagram (now Pernod  
Ricard). In terms of the agreement SFW has the right to distribute the brand for a rolling five year  
period, provided agreed sales targets were met. The agreement was terminated by Seagram, albeit,  
contended SFW, unlawfully so. The matter was decided in Distell’s favour on appeal (SFW vs Martell  
& CIE, Supreme Court of Appeal 427/01, 6 September 2002), so that the Martell distribution rights  
remain with Distell. (T4, p. 128)
3  Subsequent to the merger, E. Snell & Co. has acquired the Bols license.
Non­Confidential version 2

Other significant participants in the production and distribution of  
alcoholic beverages
6. The merged entity’s most significant competitors in the production and  
distribution of spirits are GUDV and E. Snell & Co. GUDV is the South  
African subsidiary  of multinational  spirit  producer Diageo, which was  
established   out   of   the   merger   between   Guinness   and   Grand  
Metropolitan. GUDV has the largest market share in whisky (including  
the J&B, Johnny Walker and Bell’s brands) and vodka (Smirnoff).   It  
has smaller stakes in brandy,  gin,  and FABs.  GUDV thus competes  
primarily with Distell in the middle and upper segments of the spirits  
markets.  As elaborated below these are commonly referred to as the  
‘proprietary’ or ‘prop’ and ‘premium’ spirits.  
7. E.   Snell   &   Co   is   a   smaller   South   African   company,   which   produces  
mainly, although not entirely, ‘value­for­money’ spirits – the low­price  
end of the market  ­ including brandy (Wellingtons and Bols),  whisky  
(Two Keys and Firstwatch), vodka (Absolut), cane (Cape to Rio), Gin  
(Strettons   Deluxe   Gin)   and   an   alcoholic   fruit   beverage   or   ‘FAB’  
(Snapper). 
8. Douglas   Green   Bellingham   (DGB),   a  long   established   South   African  
company, is mainly a wine merchant, but its portfolio does encompass  
some well­know spirits brands in whisky (Balantine Finest) and brandy  
(Connoisseur).   Brown   and   Forman,   a   major   international   liquor  
company,   also   distributes   some   of   its   important   proprietary   and  
premium   spirits,   notably   Jack   Daniels   whiskey.   The   South   African  
licensee   of   Brown   and   Forman   brands   is   the   Really   Great   Brand  
Company, which also performs distribution and related sales functions  
for E. Snell & Co. Other competitors include Seagrams (whose brands  
were subsequently acquired by Pernod Ricard), African Wines & Spirits  
and a large number of wine producers.  The UK­based Bulmer, which

and a large number of wine producers.  The UK­based Bulmer, which  
has   substantial   international   interests   in   cider,   entered   the   South  
African   market   in   1999,   when   it   acquired   certain   cider   brands   from  
Gilbeys.  Bulmer exited the local market in 2002.
9. South African Breweries has a near­monopoly in beer where it enjoys a  
market   share   of   approximately   95%.     It   has   also   recently   begun  
producing FABs.
Non­Confidential version 3

The Merger
The transaction
10. On September 20, 2000 Distillers and SFW entered into an agreement  
in terms of which Distillers would acquire, subject to the approval of the  
shareholders, the assets  and  liabilities of SFW,  including the shares  
held by SFW in the issued share capital of Western Province Cellars  
Limited, SFW Holdings Limited, Bofor Properties (Pty) Ltd, and Devon  
Road Property (Pty) Ltd, and all the trade names and trademarks of  
SFW. The assets sold by SFW to Distillers included SFW’s shares in  
its   operating  companies  and   all   its   trademarks,   but   excluded  certain  
specified assets. An addendum to the sale agreement was executed  
on 9 October 2000 (A171). Pursuant to the transactions, the merged  
entity was renamed Distell Group Limited (“Distell”).
11. Two   common   shareholders,   Rembrandt­KWV   Investments   (currently  
known   as   Remgro­KWV   Investments   Limited)   and   South   African  
Breweries, held 90% of the voting equity in both acquiring and target  
firm. The remaining 10% of each firm was held by the general public. 
12. Post­merger the (simplified) share holding structure is as follows:
Remgro KWV Investments South African Breweries  Public 
shareholders
60%     30% 10%
Distell Group Limited
South African Distillers and Wines
SFW Holdings Limited  Others
History of the transaction
13. On   8   June   2000,   the   legal   representatives   of   the   merging   parties  
approached   the   Commission   and   asked   it   to   clarify   whether   the  
proposed transaction to merge the businesses of SFW and Distillers  
constituted   a  notifiable   transaction.   The   parties’   essentially   held   that  
because   of   the   common   identity   of   the   parties’   shareholders   the  
transaction constituted an internal restructuring and not a merger as  
defined in the Act. On 7 August 2000, the Commission concluded in a  
Non­Confidential version 4

letter addressed to the parties’ legal representatives, that the proposed  
transaction would not constitute a merger as defined in section 12 of  
the Act and accordingly was not notifiable in terms of section 13. Based  
upon   this   opinion,   the   parties   proceeded   to   issue   cautionary  
announcements advising of the proposed merger. 
14. In terms of an agreement dated 20 September 2000 (amended on 9  
October   2000)   the   merging   parties   effected   a   transaction   whereby  
Distillers acquired all the principal assets and liabilities of SFW. The  
purchase consideration in respect of the SFW assets, in the amount of  
R515 157 950,31, was settled through the issue by Distillers to SFW of  
55 580 000 Distillers ordinary shares in the share capital of Distillers.  
These Consideration Shares were distributed by way of a dividend in  
specie and reduction in share capital to the SFW shareholders.
15. Seagrams,   a   large   multinational   producer   of   various   alcoholic  
beverages,   subsequently   launched   an   application   in   the   Cape   High  
Court on 10 November 2000 in which it asked the court to find that the  
transaction between SFW and Distillers constituted a merger in terms  
of   the   Act.   The   applicant   sought   an   interdict   restraining   the  
respondents   from   implementing   the   merger,   alternatively   an   order  
referring the matter to the Competition Tribunal. In his judgment, Jali J  
ruled that section 65(3) made it clear that the High Court did not have  
jurisdiction   to   hear   the   matter,   insofar   as   it   related   to   competition  
matters within the exclusive jurisdiction of the competition authorities. 4
16. Bulmer   SA   (Proprietary)   Limited   (“Bulmers”)   (the   local   subsidiary   of  
another   large   multinational   producer   and   distributor   of   alcoholic  
beverages)   and   Seagram   Africa   (Proprietary)   Limited   (“Seagrams”),  
both competitors of Distell, subsequently brought an application to the

both competitors of Distell, subsequently brought an application to the  
Competition Tribunal in terms of section 62(1) of the Competition Act  
89 of 1998. The basis of the application was that the respondents failed  
to notify a transaction that the applicants contended was a merger as  
defined in terms of section 12(1) of the Act.
17. The Tribunal found that the transaction constituted a merger as defined  
in terms of section 12 of the Competition Act and ordered the parties to  
notify the merger to the Competition Commission. This judgement was  
upheld by the Competition Appeal Court on 27 November 2001. 5 
18. The   merger   was   subsequently   notified   on   12   December   2001.   The  
Competition Commission recommended in June 2002 that the merger  
be approved subject to certain conditions.   Essentially, the conditions  
recommended by the Commission  relate to  the sale of  a number of  
4  Seagram Africa (Pty) Ltd and Stellenbosch Farmers’ Winery Group Ltd, Stellenbosch Farmers’  
Winery Ltd, Distillers Corporation (SA) Ltd, 7759/00, CPD.
5   Competition Appeal Court 08/CAC/May01. 
Non­Confidential version 5

brandy and sparkling wine brands.
Rationale for the transaction
19. The parties claim that the merger will generate increased efficiencies  
that will enhance international competitiveness and shareholder value.  
In   particular,   they   argue   that,   absent   the   merger,   neither   company  
could afford the intensive marketing strategies nor effectively manage  
the supply and distribution of alcoholic products overseas. The merged  
entity, on the other hand, would, through combining marketing budgets  
and   by   cost   savings,   achieved   through   economies   of   scale   and  
reduced duplication, have a significantly greater chance of penetrating  
international   markets.   The   cost   savings   would   be   realized   by   a  
rationalization   of   support   services,   manufacturing   and   distribution  
facilities and by reductions in working capital and fixed assets. 
The Hearing
20. After the merger was referred to the Competition Tribunal in June 2002  
a prehearing conference was held July 7 th  2002, and a hearing was  
duly convened, spanning five days in total:
­ 15 August 2002
­ 16 August 2002
­ 22 August 2002
­ 9 October 2002
­ 15 November 2002
21. A   total   of   nine   witnesses   were   heard.     These   were   called   by   the  
Competition   Commission,   the   merging   parties   and   the   Competition  
Tribunal.
22. The following witnesses were called by the Competition Commission:
­ Mr. Alistair Norman Lewis, AC Nielsen South Africa
23. The following witnesses were called by the merging parties
­ Mr. John Forsyth, partner McKinsey and Company
­ Mr. Johannes Jacobus Scannel, MD of Distell
­ Mr. Valerio Doriano Toros, Distell
­ Mr. Jacobus Hendrik Visser, Distell
24. By the Competition Tribunal:
­ Mr. Michael Clifford Veysie, MD Bulmers
­ Mr. Tim Hutchinson, CE Douglas Green Bellingham
­ Mr. David Hooper, MD E. Snell & Co
­ Mr. Colin Robinson, MD Robinson Liquors
Non­Confidential version 6

Competition Evaluation
Background
25. The   alcoholic   beverages   sector   represents   to   competition   folklore   in  
South Africa, what, we imagine, the oil industry represented to those  
concerned with competitive markets in the USA at the turn of the last  
century.   Not only do we have what is, to all intents and purposes, a  
single domestic beer producer, but we have a longstanding history of  
state   intervention   in   the   production   of   wine   and   spirits,   intervention  
manifestly designed to support narrow private interests rather than the  
public interest, that is possibly unparalleled in its breadth and intensity.  
The 1982 Competition Board Report on the liquor industry notes:
 
“As   early   as   1918   a   written   agreement,   a   so­called   ‘gentlemen’s  
agreement’   was   entered   into   by   the   KWV   and   the   wine   merchants  
under which the KWV would refrain from competing with the merchants  
it   supplied.   Specifically,   the   KWV,   as   a   quid   pro   quo   for   the   co­
operation of private entrepreneurs, undertook not to compete with the  
existing interested parties in the trade in or distillation or manufacture  
of wines and spirits in Africa south of the equator.” 
26. And, further:  “In 1924 Parliament incorporated this principle in Act 5 of  
1924, so that the KWV is not allowed to sell any wine or spirits to any  
person   not   being   a   bona   fide   distiller,   wholesaler,   association   of  
distillers or wholesaler or co­operative society”. 6 
27. However, the apotheosis of anti­competitive conduct in this sector is  
surely   the   agreement   which   secured   South   African   Breweries’   beer  
monopoly   and   the   Rembrandt   Group’s   pre­eminent   position   in   the  
spirits,   particularly   the   brandy,   market.     We   refer,   of   course,   to   the  
notorious market sharing arrangement between the beer producer and  
its   counterpart   in   the   wine   and   spirits   sectors   that   saw   the   former

its   counterpart   in   the   wine   and   spirits   sectors   that   saw   the   former  
agreeing to limit its involvement in wine and spirits in exchange for an  
undertaking from the Rembrandt group to stay out of the beer market.  
To add insult to injury, KWV was allowed to take up a significant share  
of the new spirits and wine company. 7
6  Competition Board, Report No. 10, ‘Investigation into Restrictive Practices in the Supply and  
Distribution of Alcoholic Beverages in the Republic of South Africa’, 1982.
7  Nor, it seems, has South African Breweries, desisted from entering into these types of agreements.  
Recent media reports suggest that in East Africa SAB has entered into a geographical market sharing  
arrangement that saw it agreeing with subsidiaries of Diageo, another large multinational brewer, to  
exit the Kenyan market in exchange for securing a monopoly in Tanzania. Business day, ‘SAB closes  
Kenyan Brewery after Four­Year Beer War”, 15 May 2002.
Non­Confidential version 7

28. The market sharing arrangement was effected by the restructuring in  
1979 of the South African liquor industry. This involved an arrangement  
between   SAB,   SFW,   OMG   (effectively   Distillers’   predecessor)   and  
KWV,   culminating   in   the   formation   of   a   new   entity,   Cape   Wine   and  
Distillers Limited (“CWD”). 8
29. CWD   was   listed   on   the   Johannesburg   Stock   Exchange,   its   shares  
being allocated (and held) in the following manner:
­ The Rembrandt group:  30%
­ SAB: 30%
­ KWV: 30%
­ General public: 10%
30. This restructuring was designed to facilitate a split in the liquor industry  
in terms of which:
­ SAB   purchased   the   Rembrandt   Group’s   beer   interests  
(Intercontinental Breweries, or ICB, the large brewery with which  
SAB had been in a price war that year);
­ SAB agreed to pool its wine and spirits interests (including SFW  
and   Henry   Tayler   &   Ries)   via   the   CWD   and   to   limit   its  
involvement in wine and spirits to its 30% investment holding in  
CWD;
­ The Rembrandt Group sold its wine and spirits operations (the  
Distillers   Corporation   Ltd   and   the   Oude   Meester   Group   and  
thereby the retailers Western Province Cellars and Liquortown);
­ The   Rembrandt   Group   undertook   to   have   no   interests   in   the  
beer market;
­ Rembrandt   and   SAB   undertook   to   divest   their   retail   liquor  
interests; and
­ SFW, Distillers and OMG become wholly­owned subsidiaries of  
CWD.
31. This   restructuring   effectively   meant   that   SAB   sacrificed   its   wine   and  
spirits interest to CWD in return for a beer monopoly and a stake in  
CWD, whilst spirit and wine production was concentrated in the CWD,  
which   acquired   the   two   largest   producer­wholesaler   bodies,   namely  
SFW and Oude Meester Group (OMG).
32. Shortly thereafter, the Rembrandt Group and KWV pooled their shares  
in CWD (60%) in a jointly owned holding company, Rembrandt­KWV  
Investments Limited.

in CWD (60%) in a jointly owned holding company, Rembrandt­KWV  
Investments Limited.
33. In   1982   the   Competition   Board   (Competition   Board   Report   no.10)  
recommended   that   the   vertical   integration   in   the   liquor   industry   be  
8  Sources: ‘Conspiracy of Giants; The South Africa Liquor Industry’, by M. Fridhjon & A. Murray  
(1986), Competition Board Report No. 10 (1982),  op cit  and Bulmer affidavit in the matter between  
Bulmer, Distillers, SFW and the Competition Commission, 94/FN/Nov00.
Non­Confidential version 8

prohibited and that the merger which had taken place in 1979, giving  
rise to the formation of CWD, be reversed:
”The competition that previously existed between SAB and ICB  
and between the two largest producer­wholesalers of wine and  
spirits   SFW   and   Oude   Meester,   has   inevitably   been   either  
terminated or restricted by the restructuring”
And further:  ”The Board is convinced that the circumstances described … do  
not justify in the public interest the KWV’s acquisition of an interest in CWD”.
34. This recommendation was rejected by the government of the day.
35. However, in 1988, the then Minister of Economic Affairs supported a  
separation of the two main components of the CWD, namely SFW and  
OMG, reportedly motivated by a desire to enhance competition. The  
separation   was   effected   by   a   separate   listing   on   the   Johannesburg  
Stock   Exchange   of   SFW   and   a   new   entity,   named   the   Distillers  
Corporation SA Limited, comprising the interests of OMG. This event  
constituted a partial reversal of the 1979 restructuring that had created  
a concentration of  wine and spirit interests  within a single corporate  
structure.
36. It   is   undoubtedly   the   breathtaking   audacity   of   these   manifestly   anti­
competitive agreements and their endorsement by the political powers  
of   the   time,   that   accounts   for   the   persistence   of   anti­competitive  
structures in the alcoholic beverages sector and for the intensity of the  
disquiet articulated by consumers, distributors, the current government  
and,  in particular,  other,  inevitably  smaller,  producers at the state of  
affairs in this industry.  However, while the structure of the industry that  
has emerged as a result of these agreements undoubtedly demands  
an   unusual   degree   of   vigilance   from   the   competition   authorities,   we  
cannot use the provisions of the Competition Act to turn the clock back,

cannot use the provisions of the Competition Act to turn the clock back,  
to redeem, ex post facto, the sins of the past.    We are, regrettably,  
obliged to take the structure of the industry as we find it and, in merger  
proceedings at least, to limit our interventions to those transactions that  
result in a substantial lessening of competition.
The Relevant  Markets
The Geographic Market
37. It is common cause between Distell and the Competition Commission  
that the relevant geographic market is national.   The Tribunal agrees  
that the relevant geographic market is indeed the South African market  
and this issue will not be considered further. 
Non­Confidential version 9

The Product Market
38. In common with other contested merger proceedings, the main area of  
contention   between   the   parties   and   the   Competition   Commission  
centres on the identification of the relevant product market. In short,  
whereas the merging parties contend for a broad product market that  
encapsulates   all   alcoholic   beverages,   the   Commission   prefers   a  
narrower   definition   that   places   each   traditional   category   of   alcoholic  
beverage – brandy, whisky, wine, etc – in separate relevant markets. 
39. The Commission defines a variety of relevant product markets based  
on traditional liquor categories, including spirit type, different types of  
wine (table wine, sparkling wine and fortified wine) and a market for  
flavoured   alcoholic   beverages   or   FABs. 9  The   Commission   finds  
product overlaps in the following markets:
Table 1: Market shares and HHI changes per liquor category
Liquor category Post­merger 
Market share
(% sales value)
Post­merger HHI HHI   increase   due  
to merger
Whisky 11.8 2487 65.6
Brandy 71.7 5366 1941.1
Vodka 16.6 4565 130.3
Cane 39.4 4309 424.7
Gin 73.2 5147 1542.9
Table wine 59.5 N/a N/a
Sparkling wine 74.4 Approx 5580 Approx 2710
Fortified wine 80.8 N/a N/a
FABs 61.8 4793 861.4
Source: AC Nielsen data and Competition Commission recommendations
40. While,   as   will   be   elaborated   below,   the   Commission   found   that   the  
impact   on   competition   of   the   horizontal   overlap   in   most   of   these  
markets   is   ameliorated   by   other   factors   –   for   example   the   unusual  
9  FABs are the ready­to­drink mixes and alcoholic fruit beverages that technically encompass a range  
of liquor categories, including ales, beers, ciders, alcoholic fruit beverages, wines and spirit coolers.  
FABs characteristically boast novel tastes/flavours, fashion aspects, thirst quenching qualities and  
convenience (as they are ready to drink). (B62) The fast growing and ‘fickle’ market for FABs is

characterized by short product life cycles and rapid market share shifts. (Beverage Business Yearbook  
2001; B63)
Non­Confidential version 10

dynamism of the FABs market or the merged entity’s relatively small  
market   share   in  whisky   –   the   Commission’s   narrow,   category­based  
market definition was the basis for its finding of a substantial lessening  
of   competition   in   the   brandy,   sparkling   wine   and   gin   markets   and,  
accordingly,   for   its   recommendation   that   the   merged   entity   be  
compelled to divest itself of a number of brands in these markets. 10   
41. In   the   brandy   market   the   Commission   recommended   that   Distell   be  
compelled to dispose of 16%   of its market share whilst terminating the  
manufacture and distribution of all KWV brands. In the sparkling wine  
market the Commission recommended that Distell dispose of brands  
with a cumulative market share of between 20­30% in volume.
42. The merging parties on the other hand define the relevant market to  
include all alcoholic beverages, ranging from beer to spirits, including  
wine   and   FABs.   The   parties   find   that   Distell’s   post­merger   market  
share   in   the   national   alcoholic   beverage   market,   based   on   litres  
absolute alcohol, is 19.7%. 11 
43. Accordingly,   the   parties   have   argued   that   there   is   no   substantial  
lessening of competition in the relevant market and suggest that, even  
if there was, the efficiencies generated by the merger would offset any  
detrimental effects of the merger.
44. Much hinges then on the identification of the relevant product market.  
Unfortunately, however, the Commission has produced scant evidence  
in support of its view of the relevant product market.  For the most part,  
it elected to support its case through a critique of evidence produced by  
the parties and through examining witnesses called by the Tribunal.  It  
is not surprising then that counsel for the merging parties should have  
raised, at the commencement of the hearings, the question of onus,

raised, at the commencement of the hearings, the question of onus,  
arguing that it is not for the parties to prove that competition will not be  
substantially lessened by the merger, but rather for the Commission to  
establish the likelihood of a substantial lessening of competition.
45. However, the question of onus is not as clear­cut as the parties would  
have us believe.   The Tribunal is the decision maker in respect of all  
large mergers. It is, in other words, required to establish not whether  
some participant has discharged an onus, but, rather, whether or not  
10  The Commission has only recommended remedies in two of these markets, viz. sparkling wine and  
brandy.  In the gin market, although it found that competition had been substantially lessened it  
declined to recommend a remedy on the grounds that litigation concerning Distell’s distribution  
agreement with Gilbey’s (GUDV) was in process and its outcome would have a determinant effect on  
the merged entity’s market share. The matter is currently still subject to litigation, although agreement  
has been reached for the brand to remain with Distell until the matter has been decided by the courts.  
The Commission acknowledges however that ‘whoever ends up with control of the Gordon’s brand,  
will control the gin market’.
11  When sorghum beer sales are taken into account, the market share drops to 17.2%. 
Non­Confidential version 11

there has been a substantial lessening of competition.  This it will do on  
the   basis   of   the   evidence   and   argument   submitted   to   it,   including  
evidence   garnered   through   the   Tribunal’s   exercise   of   its   inquisitorial  
powers.   It   will,   indeed,   become   apparent   that   our   reading   of   the  
evidence and argument that we have heard has led us to a view of the  
relevant   market   distinct   from   that   of   both   the   Commission   and   the  
parties.  
46. That   having   been   said   then,   the   Commission   has   argued   for   a  
particular   conclusion,   namely,   that   the   transaction   is   likely   to  
substantially   lessen   competition,   and   the   remedies   proposed   by   it  
consequent   upon   this   finding   embody   potentially   important  
consequences. In doing this the Commission has relied almost entirely  
on a critique of the parties’ own arguments and the evidence of the  
parties’ witnesses. However, a critique, no matter how trenchant, of the  
parties’ argument and of the evidence led by them may establish that  
the   parties   are   wrong;   but   it   cannot,   on   its   own,   establish   that   the  
counter­argument   is   correct.     In   short,   for   an   adverse   finding   the  
Tribunal   must   be   satisfied   that   the   evidence   and   argument   that   has  
been presented, whether from documents discovered or oral evidence  
led by the Commission, the parties and the Tribunal, affirms that the  
transaction is likely to lead to a substantial lessening of competition.  
Evidence   presented   by   the   Commission   has   made   but   a   small  
contribution to meeting this standard.
47. To   return   then   to   the   definition   of   the   relevant   market,   the  
Commission’s finding of a substantial lessening of competition in the  
brandy   and   sparkling   wine   categories   rests   heavily   on   its   insistence

brandy   and   sparkling   wine   categories   rests   heavily   on   its   insistence  
that there is a range of narrow relevant markets defined by traditional  
liquor categories including spirit types (e.g. brandy, whisky, vodka, etc),  
different categories of wine (e.g. table wine, fortified wine and sparkling  
wine) and a separate market for FABs. 12    In short, the Commission  
avers that faced by an increase in the price of brandy, consumers will  
not switch to another spirit category or, even less will they switch to  
one of the other categories of alcoholic beverages such as wine, FABs  
or beer. It insists, in other words, that inter­category competition will not  
constrain an exercise in market power on the part of a producer whose  
market share in one or more of the separate categories increases as a  
result of this transaction.  
48. The Commission’s contentions with respect to the relevant market rest  
on   two   pillars.     The   first   is   international   jurisprudence   which,   the  
Commission pointed out, mostly supports the narrow, category based  
relevant   markets   contended   for   by   the   Commission.     Secondly,   the  
12  Note that the Commission appears to have conceded that ‘white spirits’ – gin, vodka and cane spirits  
– constitute a single relevant market.  Certainly, the Commission’s witness, Mr. Alistair Lewis,  
conceded this.  
Non­Confidential version 12

Commission insists that evidence presented to the Tribunal establishes  
the   weakness   of   inter­category   competition   in   the   South   African  
market.   As we have already intimated, the overwhelming bulk of the  
evidence was presented by the parties themselves, although important  
evidence   was   also   presented   by   witnesses   called   by   the   Tribunal.  
Hence,   the   Commission   relies   overwhelmingly   on   a   critique,   on   its  
particular   interpretation,   of   this   evidence.     The   Commission   called   a  
single witness, Mr. Alistair Lewis, an employee of AC Nielsen, the well­
known market research firm.
 
International Jurisprudence
49. The   Commission   avers   that   competition   authorities   and   courts  
elsewhere   have,   when   confronted   with   the   task   of   establishing   a  
relevant market in the alcoholic beverages market, found for narrow,  
category­based markets.     The Commission particularly relies on US  
and   European   decisions   in   two   prominent   mergers,   namely,   the  
Guinness   Plc   and   Grand   Metropolitan   Plc   (1997)   and   Pernod  
Ricard/Diageo/Seagram   Company   Ltd   (2001).   These   mergers   were  
evaluated   by   the   European   Commission 13  and,   in   the   case   of  
Guinness/Grand   Metropolitan,   also   by   the   US   Federal   Trade  
Commission.   The   Guinness/Grand   Metropolitan   merger   was   also  
investigated by the Australian ACCC. 
50. In   the   Guinness   /   Grand   Metropolitan   transaction   the   European  
Commission based its market definition on spirit type.  This conclusion  
rested,  inter alia , on the finding that spirit drinkers display brand loyalty  
within   the   category   of   choice   and   also   on   the   observation   of   well­
entrenched   ‘occasion­based’   consumption   patterns   which   renders  
substitution   on   the   basis   of   small   price   variations   unlikely.   The

substitution   on   the   basis   of   small   price   variations   unlikely.   The  
European Commission highlights the importance to competition in the  
spirits market of branding and its application to individual spirit types as  
a further justification for product markets not wider than that for each  
main spirit type, i.e. whisky (further segmented to differentiate Scotch  
whisky), brandy (further segmented to differentiate Cognac/Armagnac),  
rum, gin, vodka, tequila and flavoured spirits. 
51. Note that the European Commission conceded the possibility of market  
segmentation based on price and quality differentiation observing that  
‘a consumer who habitually drank a premium brand would not regard a  
cheaper   one   as   providing   an   adequate   substitute   in   terms   of   taste,  
image and so forth’. 14  However, this observation did not, in the view of  
the European Commission, alter the finding that placed individual spirit  
categories   at   the   centre   of   the   relevant   market   definition   but   rather  
13  EC IV/M.938 and EC COMP/M.2268.
14  EC IV/M.938, para 18.
Non­Confidential version 13

constituted  the   basis   for   a  further  segmentation,   this  time  within   the  
separate spirit categories.
52. In   its   assessment   of   the   Guinness/Grand   Metropolitan   merger,   the  
Federal Trade Commission focused on whisky and gin, and particularly  
on the premium segments within those categories.   Within the whisky  
category   it   distinguished   between   origins   (i.e.   Scotch   Bourbon   and  
Irish) as well as different quality and price categories. The definition of  
premium gin included a reference to its origin (i.e. England) and a retail  
price level, comparing   prices of specific brands. 15  
53. The   ACCC   found   limited   demand   substitution   between   various   spirit  
categories,   with   price   increments   in   a   particular   category   tending   to  
result in brand shifting rather than a reduction in sales in that category,  
supporting a relevant market definition based on the spirit categories.
54. In the Pernod Ricard / Diageo / Seagram Company Ltd transaction the  
European   Commission   cites   the   Guinness/   Grand   Metropolitan  
judgement   in   its   relevant   market   definition   and   re­emphasises   the  
importance of branding. Although the European Commission notes the  
possibility   of   defining   the   market   according   to   different   price/quality  
combinations, i.e. premium, secondary brands, private labels, low price  
etc, the market definition in this judgement is at one with the finding in  
Guinness / Grand Metropolitan. The European Commission concludes  
that  as ‘there is a  continuous price spectrum  ranging  from the most  
expensive to the cheapest’ and as ‘rebates can change the position of  
a   given   brand   in   the   spectrum’   the   price/quality   distinction   was   only  
found   applicable   to   the   exclusion   of   Cognac/Armagnac   from   the  
product market that otherwise included all brandies. 16
55. The Competition Commission insists that we would need particularly

55. The Competition Commission insists that we would need particularly  
powerful   contrary   evidence   to   row   against   the   tide   of   international  
opinion   that   has   found   narrow   category­based   markets.     We   are  
mindful   of   this.   Indeed   Section   1(3)   of   the   Act   explicitly   empowers  
those interpreting or applying the Act to have recourse to international  
jurisprudence, a wise provision given the immaturity of our jurisdiction  
relative to those who have worked with competition law and economics  
for many years.  However, whilst foreign jurisprudence may be, indeed  
it certainly has been, of great assistance in refining our understanding  
of   legal   questions   and   economic   theory   and   in   guiding   our   factual  
enquiries,   it   cannot   detract   from   the   strong   factual   basis   that   must  
ultimately underpin all efforts to determine a relevant market.   Counsel  
for the merging parties cites an extract from our decision in the large  
15  FTC press release (1997, December 15) ‘Dewar’s Scotch, Bombay Gin and Bombay Sapphire Gin  
to find new Corporate Homes under FTC Agreement’, www.ftc.gov.
16  EC COMP/M.2268, para 17.
Non­Confidential version 14

merger   between   Bromor   Foods   (Pty)   Ltd.   and   National   Brand  
Limited:17
“Defining a relevant market for consumer products is notoriously  
difficult. Delineating a relevant market for beverage products is  
especially   difficult   because   one   is   faced   with   not   only   the  
subjective   proclivities   of   consumers   but   also   the   marketing  
stratagems   of   firms   as   they   attempt   to   differentiate   their  
products in response to competitive threats. 
Beverage antitrust  cases have long been  the  subject of  bitter  
contestation over relevant market definition. On the one hand  
merging parties contend they are merely minor players fighting  
for their "share of the throat", in a market where the fizzy drink  
competes with anything that can be imbibed from fruit juices to  
tea.   On   the   other   hand   competition   regulators   argue   that   the  
fizzy drink is the relevant product market.
Ultimately each case must be determined on its own facts and  
foreign judgments can do no more than give us guidelines to  
method   for   they   cannot   serve   as   a   way   for   us   to   come   to   a  
conclusion   on   facts.   The   behavior   of   a   teenage   consumer   of  
carbonated   beverages   in   Texas   is   no   more   use   to   us   as  
evidence   than   the   behavior   of   the   French   consumer   of  
carbonated mineral water.”
56. It should be noted that the European Court of First Instance has itself  
explicitly   insisted   on   the   overriding   significance   of   a   current   factual  
enquiry   when   determining   relevant   markets   even   to   the   extent   of  
diminishing the weight of a prior finding of the European commission  
when making a subsequent relevant market determination or, what is  
the same thing, a finding of dominance:
“Second, a finding of a dominant position by the Commission,  
even   if   likely   in   practice   to   influence   the   policy   and   future

even   if   likely   in   practice   to   influence   the   policy   and   future  
commercial   strategy,   does   not   have   binding   legal   effects   as  
referred to in the IBM judgment.  Such a finding is the outcome  
of an analysis of the structure of the market and of competition  
prevailing   at   the   time   the   Commission   adopts   each   decision.  
The   conduct   which   the   undertaking   held   to   be   in   a   dominant  
position   subsequently   comes   to   adopt   in   order   to   prevent   a  
possible infringement of Article 86 of the Treaty is thus shaped  
by the parameters which reflect the conditions of competition on  
the market at a given time. 18”
17  19/LM/Feb00
18  The Coca­Cola Company v Commission of the European Communities  – Court of First Instance,  
Non­Confidential version 15

57. In   short,   while   we   will   not   simply   ignore   the   US,   European   and  
Australian   findings,   the   weight   assigned   them   is   reduced   if   the  
evidence   indicates   that   the   general   features   of   the   South   African  
market   differ   significantly   from   those   that   characterize   these   other  
national   markets.     And,   if  evidence  is  adduced  that  directly  conflicts  
with the inferences drawn from the general features of the market – in  
other   words,   if,   for   example,   we   are   shown   persuasive   evidence   of  
inter­category substitution in the South African market then this must,  
at   least   in   respect   of   those   categories   that   have   been   shown   to   be  
substitutable,   surely   trump   a   decision   based   on   the   general  
characteristics of the market.  
58. That   then   is   the   sequence   of   the   argument:   we   first   consider   the  
international   jurisprudence   which   finds   overwhelmingly   for   relevant  
markets defined by liquor category; we then ask whether the evidence  
demonstrates   that   the   characteristics   of   the   South   African   market  
approximate those of the markets that generated these findings; and  
we   finally   ask   whether   there   is   evidence   of   consumer   behaviour   in  
South Africa that is at odds with the inferences drawn from the general  
characteristics of the market.  In this case, the evidence in point would  
relate to the issue of inter­category substitution.  
59. The   Competition   Commission   argument   effectively   holds   that  
consumer behaviour in the USA, Europe  and  Australia  is sufficiently  
similar to that of South African consumers of alcoholic beverages to  
justify   the   simple   importation   of   conclusions   regarding   the   relevant  
market from these jurisdictions.   It would have us accept that if brand  
loyalty   and   occasion­based   drinking   –   the   two   features   upon   which

loyalty   and   occasion­based   drinking   –   the   two   features   upon   which  
their relevant market findings are based – are prevalent in the USA,  
European and Australian markets, then they will be present in the local  
market as well.
60. However   little   evidence   has   been   presented   in   support   of   this   far­
reaching   and   essentially   factual   assertion.     And   yet   there   are   clear  
prima   facie   grounds   for   questioning   its   validity.     South   Africa’s  
particular  income  distribution  and the  absolute levels  of  poverty with  
which a large proportion of the liquor­consuming population contend is,  
per   definition,   an   extremely   powerful   determinant   of   consumption  
patterns and behaviour, particularly where ‘discretionary’ consumption  
is   concerned     –   little   evidence   is   needed   to   assert   the   massive  
disparities  between   South  Africa   and   the  highly   developed   countries  
from   whom   the   Commission   would   have   us   draw   essentially   factual  
conclusions regarding the relevant market.  
Joined Cases T­125/97 and T­127/97
Non­Confidential version 16

61. Moreover,   the   truly  unique   features  of   South  Africa’s   liquor   history  ­  
recall   that   until   relatively   recently   the   vast   majority   of   South   Africa’s  
population was, by law, not permitted to enter outlets at which spirits  
were   sold   –   throw   the   Commission’s   questionable   proposition   into  
sharp   relief.     Indeed   the   parties   argue   that   the   fact   that   the   bulk   of  
liquor purchased in South Africa is still sold and consumed in the semi­
legal shebeen environment is evidence of the unusual character of the  
South   African   market.     The   rapid   and   massive   shift   in   consumption  
away from sorghum beer to other alcoholic beverages is also a unique  
feature of the South African market.   The single witness called by the  
Commission,   Mr.   Alistair   Lewis   of   AC   Nielsen,   identified   another  
distinguishing   feature   of   alcoholic   beverage   consumption   in   South  
Africa, namely, the: 
“huge wine market, which is not so predominant in other parts of  
the world.  In other words, I’m not talking necessarily of quality  
wine.     I’m   talking  about   the  bottom   end  of   the   market,   which  
traditionally started here back in the sixties or even before that.”  
62. Lewis nevertheless insisted that, but for this distinction, which strikes  
us as rather significant, the consumption patterns in the South African  
market match those found elsewhere.
63. The   merging   parties   have,   for   their   part,   presented   evidence   of   the  
distinctive   features   of   South   African   liquor   consumption.   Hence   they  
submit – and this was not contested by the Commission ­ that South  
African   spirits   consumers,   in   contrast   with   their   developed   country  
counterparts, rarely drink spirits neat, but rather use it to add alcoholic  
content to a mixer, so that the key attribute of the spirits is, in the minds

content to a mixer, so that the key attribute of the spirits is, in the minds  
of South African consumers at least, its alcohol content rather than its  
particular taste. This would portend a greater possibility of substitution,  
at   least   between   spirit   categories.   The   parties   also   insist   that   liquor  
consumption in South Africa, as opposed to other societies in which  
similar   research   has   been   conducted,   is   less   ‘occasion   based’,   less  
structured by the time of day at, or occasion on, which it is consumed,  
an   assertion   borne   out   by   the   research   conducted   as   part   of   the  
‘brandy study’. 19    
64. The Commission purports to find support for its claim that “the broader  
South African alcoholic beverage market is not substantially different  
from the markets in the United States of America, Australia and all the  
member   states   of   the   European   Community   …”   in   research   results  
which   indicate   that   a   significant   proportion   of   the   respondents  
consumes more than one kind of drink at a particular occasion.   This, 
19  A study undertaken on behalf of Distell by the US consultancy firm, McKinsey.  The study is  
described more fully below. 
Non­Confidential version 17

the Commission argues, is evidence of occasion­based consumption  
among South African liquor consumers.  
65. However, this response confirms at best that more than one brand is  
consumed   by   a   significant   proportion   of   consumers   on   any   one  
occasion. This does not show that the occasion determines the switch  
–   indeed,   in   the   absence   of   further   evidence,   it   may   be   reasonably  
inferred   from   this   that   South   African   liquor   consumers,   in   apparent  
contrast with their counterparts in the other jurisdictions cited, display  
little brand or category loyalty. In short, this evidence may well support  
the notion of a wide relevant market, in which consumers drink more  
than one type of drink at any one occasion.
66. Moreover, Mr. John Forsyth, the parties’ expert witness, testified that  
market surveys that attempted to relate specific occasions to particular  
liquor products using a sample of South African consumers, called into  
question the notion of occasion­based drinking in South Africa:
(Paragraph omitted, contains confidential information)
67. And, Forsyth concludes: 
(Paragraph omitted, contains confidential information)
68. In   summary,   we   find   unpersuasive   the   Commission’s   reliance   on  
foreign jurisprudence in determining the relevant market.  Certainly, the  
evidence provides little justification for the uncritical application of the  
European   and   US   findings   to   South   Africa.     On   the   contrary,   the  
available   South   African   evidence   suggests   that   there   are   important  
unique features of South African liquor consumption that will have an  
important bearing on the definition of the relevant market. 
Inter­category Competition: the South African evidence
69. We have been presented with a confusing welter of evidence, some of  
it   empirical   and   much   of   it   anecdotal   drawn   from   a   combination   of  
market survey and direct experience of the market.  We are also faced

market survey and direct experience of the market.  We are also faced  
with appeals – largely emanating from the Commission ­ to ‘common  
sense’, to the personal knowledge or, at least, personal opinion, that  
many   have   of   this   mass   consumption   market.     The   proponents   of  
‘common   sense’   effectively   ask   us   to   accept   that   consumption   of   a  
particular category of alcoholic beverages is a matter of deep personal  
taste that will not be compromised by a mere increase in price.  While  
an increase in the price of tea may plausibly give rise to a switch to  
coffee, a loyal brandy – or whisky or wine or beer – consumer will, in a  
manner of speaking, simply swallow the price increase, his expressed  
Non­Confidential version 18

commitment   to   his   alcoholic   beverage   of   choice   would   inhibit   him  
switching   to   another   category   or   even   decreasing   significantly   his  
overall consumption of his chosen alcoholic beverage.
70. While   we   cannot   simply   ignore   these   conventional   wisdoms   –   the  
particular   categories   do,   after   all,   have   different   tastes   or,   in   the  
language   of   marketing   specialists,   ‘intrinsics’   –   we   cannot   base   a  
definition   of   the   relevant   market   on   these   insights   alone.     It   would  
introduce   an   intolerable   degree   of   subjectivity   and   uncertainty   into  
competition   analyses,   most   particularly   where   consumer   goods  
markets are at issue.
71. The evidence on substitutability presented by Mr. Lewis and on which  
the Commission sought to rely for its version of  the relevant market  
related,   firstly,   to   the   form   in   which   Nielsen   collected   data,   and,  
secondly, to evidence of long term stability in the market shares of the  
various   traditional   categories   within   the   broader   alcoholic   beverages  
market.
72. Lewis   testified   that   Nielsen’s   clients   in   the   liquor   industry   do   not  
generally   and   traditionally   request   information   on   liquor   sales   as   a  
share of total  spirit sales or of all  alcoholic beverages. The services  
rendered   to   AC   Nielsen   clients   in   South   Africa   include   a   bimonthly  
presentation that incorporates an overview of economic trends in the  
SA market and a brief overview of the total liquor market.  For the most  
part,   however,   Nielsen   is   required   to   report   on   the   individual   spirit  
categories.   This, the Commission averred, constituted evidence that,  
in   their   daily   practice,   the   actual   participants   in   the   liquor   industry  
themselves worked with category­specific relevant markets.  ‘Why’ it is

themselves worked with category­specific relevant markets.  ‘Why’ it is  
implicitly asked ‘would they ask for information delineated by traditional  
liquor category if they genuinely believed that their products in these  
categories competed with products in all other categories, that is, with  
products in the alcoholic beverages market?’
73. Even if we assumed that the Nielsen experience confirmed that firms  
were   predominantly   interested   in   sales   data   in   traditional   liquor  
categories,   it   is   by   no   means   clear   that   we   should   be   drawing   the  
inference sought by the Commission.   It is wholly possible to imagine  
motorcar manufacturers asking for data to be collected that would help  
identify   the   most   popular   colour   vehicle   that   they   produced   but   this  
would surely not justify a conclusion that placed red and blue vehicles  
in separate relevant markets!
74. Nor is it clear that Nielsen’s clients do, in fact, always require that data  
are   collected   in   traditional   liquor   categories   or,   when   tracked   in  
traditional categories, that they use the data in this form.  Firstly, Lewis  
Non­Confidential version 19

acknowledges that Nielsen is no longer asked to collect separate data  
on   gin,   vodka   and   cane   spirits   but   rather   to   aggregate   these   in   a  
collective   ‘white   spirits’   category.     Particularly   interesting   is   Lewis’  
acknowledgement   that   this   has   shifted   over   time,   an   admission   that  
supports   the   parties’   notion   of   a   dynamic   market   characterized   by  
significant shifts as new consumers and new products enter the market  
and tastes change.  Secondly, Lewis admits that a major Nielsen client,  
South   African   Breweries,   requires   data   of   the   overall   alcoholic  
beverages   market   and   that   it   uses   this   in   order   to   track   the  
performance of its products, beer, and, recently, FABs, vis­ à­vis spirits  
and   wine.     Again,   this   is   identified   as   a   relatively   recent   mode   of  
collecting and presenting data. 
75. The   parties   predictably   deny   that   the   form   in   which   Nielsen   data   is  
collected has any bearing on the identification of the relevant market.  
They effectively suggest that this simply reflects a convenient mode of  
organizing masses of market data.
76. Moreover, the parties insist that their broad view of the relevant market  
is shared by many of their competitors. This is however only partly true.  
Hence SAB and GUDV, the large multinational producers of a broad  
range of alcoholic beverages, both consider their various products to  
be competing in the broad alcoholic beverages market. Snell, a locally  
based participant in the alcoholic beverages market, avers that it is  ‘an  
acknowledged fact’ that ‘beer, wine, RTDs, ciders and spirits fiercely  
compete for liquor consumers’. On the other hand, Seagram, also a  
large multinational producer of alcoholic beverages, and DGB, a local  
producer and distributor, appear to support a narrow market definition.  
On the basis of its experience in a range of national markets and on a

On the basis of its experience in a range of national markets and on a  
telephone   survey   of   local   FAB   consumers,   Bulmer,   yet   another  
multinational   producer   and   recent   entrant   into   the   South   African  
market, contends that FABs are a separate market, although it is not  
clear that this conclusion is supported by the results of the survey that  
they commissioned. 20  The merging parties also find support for their  
20  The Lexecon report concludes that ‘a small but significant price increase is sustainable in the SA  
FAB market, as only a small number of consumers would substitute away from FABs to other types of  
alcohol as a result of a 10% increase’. The research methodology consisted of a telephonic survey of  
200 FAB consumers in the Gauteng area who own a cellular phone, which apparently amounts to 30%  
of these FAB consumers. The survey results demonstrate a low substitutability between FABs and  
other alcoholic drinks. The drawbacks of this approach are numerous. Firstly, the 30% of FAB drinkers  
who own a cellular phone are likely to be less sensitive to price changes than those who do not.  
Secondly, as Lexecon acknowledges, what is really measured is reported preference (obtained by  
questions around hypothetical situations), not revealed preference (i.e. the real­life response) which is  
likely to lead to over­stated results. The same survey also shows that if the favorite brand of FAB is not  
available, although 51% of respondents would purchase a different type of FAB, as much as 15% of  
the respondents would switch to a different type of alcoholic drink altogether (mostly beer and spirits,  
C103). Moreover, when faced with a price increase of 10%, 74.7% of respondents would continue to  
buy this brand, but only 3% would have bought a different type of FAB, and 10.4% of respondents  
would not have bought any alcohol (C105). The latter results suggest that brand­loyalty, rather than  
Non­Confidential version 20

broad market definition in the views of industry analysts.
77. The   parties   claim   that   the   experience   of   the   ‘brandy   study’   also  
confirms   their   view   of   the   relevant   market.     This   study   was  
commissioned by Distell and undertaken by McKinsey, the large US  
based management consultancy.  
78. Although   initially   conceived   as   a   study   of   brandy   consumption   and  
hence   dubbed   the   ‘brandy   study’,   Mr.   John   Forsyth,   a   McKinsey  
executive who testified on Distell’s behalf at the hearings, averred that  
it immediately became apparent to the McKinsey research team that  
the patterns of, and prospects for, brandy consumption could not be  
understood without locating it in a wider study of the consumption of  
alcoholic   beverages   generally.     Forsyth   testified   that   his   company  
characteristically   approached   requests   to   analyse   the   positioning   of  
particular   products   in   the   market   by   asking   consumers   a   range   of  
pertinent questions concerning the product under scrutiny.   If, in the  
course   of   their   survey   of   consumers,   the   respondents   persistently  
mentioned other products, this signalled the necessity for widening the  
scope of the study to include these other products.  
79. Hence, the ‘brandy study’ effectively became a study of brandy’s place  
within the broader alcoholic beverages market. In essence the study  
found: 
(Paragraph omitted, contains confidential information)
80. In   fact   it   found   that   a   significant   proportion   of   the   population  
(accounting for a significant part of consumption) is prepared to change  
drinking   patterns   in   response   to   a   change   in   price   or   other   product  
attributes.
81. While the finding cited in the preceding paragraph is, on the face of it,  
pertinent, we are provided with no further basis for this conclusion, nor  
are we told what liquor categories are referred to.  Moreover, changes

are we told what liquor categories are referred to.  Moreover, changes  
inspired by ‘fashion’ are likely to be of a longer­term nature – more akin  
to   a   shift   of,   rather   than   along,   a   demand   curve   –   than   are   those  
inspired by price changes.
82. This goes to some of the reservations that we have about the brandy  
study   for   the   purposes   of   conducting   a   competition   analysis.   It   is   a  
study   manifestly   designed   to   inform   Distell’s   marketing   strategy   and  
although it does, in the process, unearth insights and information that  
are   of   some   indicative   interest   in   a   competition   analysis,   it   is   not   a  
category­loyalty determines the behaviour in this category, thereby weakening the case for a separate  
FAB market. 
Non­Confidential version 21

competition analysis. That it asks ‘how can I, predominantly a producer  
of brandy and other spirits, persuade beer drinkers to come over to my  
product’   or   ‘what   steps   do   I   need   to   take   to   prevent   the   breweries  
wooing away my customers’ is not, on its own, the identification of a  
relevant market that includes both brandy (and other spirits) and beer  
(or, for that matter, wine).  
83. It   merely   acknowledges   that,   new   consumers   aside,   the   most   likely  
source of additional custom for any alcoholic beverage is to be   found  
in the ranks of existing consumers of alcoholic beverages, rather than  
in the ranks of, say, church congregants.  It simply says that those who  
are  regularly   to  found  in   a  bar   or   shebeen  –   rather  than  at   Sunday  
School   ­   constitute   the   most   fertile   ground   on   which   to   market   an  
alcoholic beverage.  Many of the patrons of a shebeen or bottle store –  
the vast majority, given current South African consumption patterns   of 
alcoholic   beverages   –   are   beer   drinkers.     Hence,   that   beer   drinkers  
should be a major target of a spirits’ producer is not surprising.   It is,  
indeed,   as   little   surprising   as   the   converse   –   we   would   expect   beer  
producers to be as much concerned with its ability to make inroads into  
the ranks of brandy consumers and this likely explains SAB’s interest in  
gathering data of alcoholic beverages consumption from AC Nielsen.  
84. At most it may reflect that in the long term battle to change tastes, to  
develop both ‘intrinsics’ and ‘extrinsics’ that would boost the long term  
prospects of a particular type of spirit – that would move the demand  
curve   for   that   particular   type   of   spirit   outwards   –   a   spirits   producer  
would target,  inter alia , beer consumers, that great rump of consumers  
who already consume alcoholic beverages.   It does not suggest that

who already consume alcoholic beverages.   It does not suggest that  
the current competitive strategy, particularly the pricing strategy, of a  
spirits producer is materially influenced by the competitive strategy of  
the beer brewers.  
85. We have already dealt with the question of ‘occasion­based’ drinking,  
the overriding basis for the European decisions that confined relevant  
markets   to   traditional   categories.     The   absence   of   evidence   of  
occasion­based   drinking   in   South   Africa   clearly   opens   the   way   to  
hypothesise   that   consumers   would   switch   between   categories   in  
response   to   price   movements   and   this   hypothesis   appears   to   be  
confirmed by the brandy study and a number of authoritative studies  
and observers.
86. The   brandy   study   finds   that   ‘..consumers   consume   products   from   a  
number of categories and freely switch between categories’
(Rest of paragraph omitted, contains confidential information)
Non­Confidential version 22

87. The   Brandy   Study   also   identifies   a   blurring   of   the   division   between  
traditional liquor categories.
(Rest of paragraph omitted, contains confidential information)
88. It is also noteworthy that the consumer category whose participants are  
most wedded to the occasion­based drinking patterns identified in the  
European cases represent a very small proportion of the adult South  
African   population   and   that   large   proportion   of   the   South   African  
population   (representing   a   significant   part   of   liquor   consumption)   is  
prepared to change drinking patterns in response to a change in price  
or fashion.
89. However,   when   the   brandy   study   isolates   consumer   responses   that  
may   assist   us   in   arriving   at   precise   conclusions   regarding  
substitutability,   it  appears  that   the   actual  blurring  of  traditional   liquor  
categories is between spirit categories: 
(Paragraph omitted, contains confidential information)
90. The   assessment   by   another   group   of   market   researchers  
commissioned   by   the   parties,   Ingwe   Communications,   of   the  
abovementioned research reinforces the focus on inter­spirit category  
substitution: 
(Paragraph omitted, contains confidential information)
91. Moreover, Distell appears to have acted on these findings:
(Paragraph omitted, contains confidential information)
92. What little independent research has been presented to us appears to  
support   the   conclusions   of   the   parties’   research   insofar   as   it  
demonstrates   substitution   between   spirit   categories.   Hence  
independent   research   on   the   behaviour   of   South   African   consumers  
reported   in   the   2001   Alcoholic   Beverage   Review   concludes   that  
‘consumer   behaviour   in   the   face   of   economic   hardship   continues   to  
move from a premium brand in one category to a cheaper category

move from a premium brand in one category to a cheaper category  
rather   than   a   cheaper   brand’ .     We   were   informed   of   research   on  
product innovation that concluded that: ‘the growth of RTD’s/FAB’s has  
resulted in a blurring of the traditional product categories’. 
93. The Commission however presents evidence that purports to question  
substitution between spirit categories.  It argues that certain of the data  
presented by Nielsen, in particular those data indicating the apparent  
long­term  stability  of  the  distribution  of  market   shares  between  spirit  
Non­Confidential version 23

categories, support a category­based definition of the relevant market.  
The Commission effectively contends that if competition was occurring  
between categories, then one would expect to see movement over time  
in the shares  of the alcoholic  beverages  market  commanded  by  the  
various   categories   –   the   stability   in   these   shares   indicate   that  
competition occurs within the respective categories.
94. The Commission presents a 10­year trend in the market shares (based  
on litres sold) in narrow spirit categories:
Table 2a Trends in spirit category market shares
1993
(%)
2002
(%)
Brandy 44.1 40.2
Whisky 19.7 22.9
Cane  9.7 5.0
Gin 5.9 5.4
Vodka 14.2 14.5
Rum  2.5 4.6
Liqueurs 3.9 7.4
Source: AC Nielsen
95. However,   the   Commission’s   reliance   upon   long­term   trends   in   the  
liquor   industry   as   evidence   of   low   inter­category   substitution   has  
several drawbacks.   First, substitution in response to a price increase  
does not have to be long­term in nature in order to qualify as evidence  
of substitutability for the purposes of a competition analysis – indeed it  
is   generally   accepted   that   long­term   trends   are   rather   indicative   of  
changes in tastes, new product innovation, etc., that is, changes that  
cause  a  shift  of,   rather  than  a  movement   along,   the  demand  curve.  
Secondly, using the long­term liquor sales data runs the considerable  
risk of masking underlying and, from the perspective of a competition  
evaluation, critical shorter­term trends.  
96. In other words, it is wholly conceivable that an increase in the price of a  
particular   liquor   category,   A,   may   have   caused   a   significant   sales  
decline   in   the   short   term   as   consumers   switched   to   category   B.  
However, assume that, for whatever reason, the producer of Category  
B could not hold its prices at the relatively low level and, a year later  
was obliged to increase its prices to the level of Category A thus losing

was obliged to increase its prices to the level of Category A thus losing  
sales to that category.   If the producers of the respective categories  
then, for say  the next five years, maintained this  parity between  the  
prices of the two categories, the long term trend would indicate relative  
stability in the overall shares of the two product categories and, in the  
Commission’s view, an absence of inter­category competition, while an  
analysis   of   short   term   trends   and   events   may   indicate   a   degree   of  
Non­Confidential version 24

substitutability between the categories that placed them in the same  
relevant market. In our view year­on­year changes in market shares of  
the   various   products   in   a   declining   market   are   important   to   indicate  
substitution is occurring in that market, even if the market shares return  
to historical levels every decade or so. 
97. Indeed,   although   more   rigorous   econometric   analysis   would   be  
necessary in order to establish a structural break in the data set, it is  
nevertheless   reasonably   clear   that   there   are   at   least   two   conflicting  
trends discernable in the period under review. Hence, if one takes 1993  
and 2002 as a basis for comparison it appears that brandy lost only  
8.8% and gin 8.5% while vodka increased its share by 2.1%.  However,  
a   significantly   different   picture   emerges   if   we   distinguish   between  
1993­97 on the one hand and 1998­2002:
Table 2b Trends in spirit categories market shares
1993­1997
(%)
1998­2002
(%)
Brandy + 6.6 ­ 11.3
Gin ­ 18.6 + 10.2
Vodka ­ 15.5 + 10.7
Source: Calculations based on AC Nielsen data
98. Note that the decline (increase) in the brandy share is accompanied by  
increases   (declines)   in   vodka   and   gin.   For   the   remaining   spirit  
categories,   the   trend   is   indeed   one   of   either   long­term   decline   or  
growth:
Table 2c Trends in spirit categories market shares
1993 – 2002
(%)
Cane ­ 48.5
Rum + 84.0
Liqueurs + 89.7
Source: Calculations based on AC Nielsen data
99. Clearly for the market share data to provide any meaningful insight into  
the  extent  of  substitutability  between  categories,  they would  have  to  
read alongside price fluctuations over the same period – hence, as we  
will elaborate below, it is extremely pertinent that the break in the trend  
coincides with the Oude Meester price hike in 1997. The Commission  
has argued – and this too is examined below – that the break is caused

has argued – and this too is examined below – that the break is caused  
by   the   behaviour   of   a   statistical   outlier,   namely   Kwazulu­Natal.  
However,   eliminating   this   important   piece   from   the   overall   data   set  
requires detailed justification.   Certainly, a simple assertion based on  
the long­term trends revealed by the Nielsen data reveals little about  
Non­Confidential version 25

the competitive interplay between the traditional liquor categories.
100. Some of the witnesses have confirmed the simultaneous existence of  
stable   long­term   market   shares   and   short­term   fluctuations.     Mr.  
Hooper, the witness from Snell, identified this phenomenon:
“I think if brandy were to go up in price significantly and there  
have   been   other   instances   where   Distell   might   point   to   that  
having happened. In Natal, for example, fairly recently brandy  
went up fairly substantially in relation to Smirnoff. And Smirnoff  
picked up a lot of market share in Natal. You do get these shifts  
and you do get them in pockets, but if one looks over a medium  
period of time, I think you’d find that those shifts are there, but  
they are relatively minor. And if you take a brandy category, for  
example,   which   is   as   large   as   brandy   is   in   the   South   African  
market that the brandy price goes up out of kilter with other spirit  
categories   or   any   other   alcohol   category,   it   may   suffer,   but   it  
would suffer to a relative degree.   I don’t think you’re going to  
suddenly find brandy going from its fifty percent (50%) market  
share of the spirits market, for example, rocketing down to forty  
percent (40%) or below forty percent (40%). I don’t think you’re  
going   to   get   that   sort   of   shift   occurring   in   a   short   to   medium  
term. In the long term it is a possibility, but I don’t see it as a  
short or medium term reaction.” 21 
101. Other   witnesses   have   pointed   to   the   significance   of   shifts   in   market  
shares,   however   small,   in   a   declining   market.     Forsyth,   the   parties’  
expert witness, testified as follows:
“If the market on the other hand has been growing, then you  
could say well some of that additional consumption in alcohol  
beverages could come from  other  drinks that  they  may  have,  
but because it’s shrinking or stable then there’s probably a high

but because it’s shrinking or stable then there’s probably a high  
degree of substitution going on.”
102. Clearly, then the Nielsen data on market share do not materially assist  
in determining the extent of inter­category substitutability and do not  
make a significant contribution to the  task of identifying the relevant  
market. 
103. The   parties,   for   their   part,   insist   that   the   evidence   of   substitutability  
between   the   various   traditional   categories   supports   their   case   for  
defining   an   alcoholic   beverages   market.   They   rely   on   a   range   of  
sources   in   support   of   their   claim   that   competition   occurs   across   the  
traditional liquor categories and, accordingly, that a broad definition of  
21  Hearing transcript, 9 October 2002, Mr. Hooper, p. 89.
Non­Confidential version 26

the   relevant   market   that   includes   all   alcoholic   beverages   should   be  
accepted.
104. The   parties’   marketing   research   provides   some   interesting   data  
indicative of the vulnerability  of  spirits  to  substitution, illustrating that  
within   the   spirits   category   certain   spirits   enjoy   considerably   less  
category   loyalty   than   others   and   that   in   those   same   categories   a  
significant   proportion   of   consumers   is   highly   ‘at   risk’   to   other  
categories. 
105. We   reiterate  that,   in  our   view,  the   responses   gleaned  in  the  brandy  
study   provide   a   strong   indication   of   substitutability   within   spirit  
categories   but   it   does   not   help   in   identifying   substitutability   between  
spirits, beer and wine.  
106. However,   the  parties  have   also   attempted  more  traditional   statistical  
analyses of actual and likely responses to increases in the price of key  
spirit  brands.     Firstly,   they   have   attempted   to   calculate   own­price  
elasticities   and   cross­price   elasticities   based   on   actual   consumer  
responses to actual price movements. Secondly, they have undertaken  
a   ‘brand   price   trade   off’   study,   based   on   data   gathered   through   the  
brandy study.  This latter study, in contrast with the more conventional  
elasticity studies, is not a study of actual behaviour but is rather in the  
nature   of   a   laboratory   experiment   based   on   survey   responses   to  
hypothetical increases in the price of different spirit categories (as a  
whole) as well as particular spirit brands.
Elasticities
107. The   parties   have   attempted   to   calculate   own­price   elasticities   and  
cross­price   elasticities.   Own­price   elasticities   are   generally   used   to  
indicate market power: if the own­price elasticity is low then prices can  
be increased as the resultant decline in volume would be small, leading

be increased as the resultant decline in volume would be small, leading  
to   an   increase   in   revenue;   low   own­price   elasticities   are   therefore  
associated  with  market  power.   The  parties   have  provided  own  price  
elasticities  for  Distell’s  products, most of which  are greater than  (­)1  
and therefore price­elastic.
108. However,   the   Commission   correctly   points   out   that   the   own­price  
elasticities calculated here are unreliable, largely because the prices of  
other   brands   in   all   liquor   types   were   increased   by   varying   amounts  
during   the   period   under   review   thus   violating   the   ceteris   paribus  
condition   and   making   it   impossible   to   isolate   the   impact   of   the   own  
price   increase.   The   simple   ‘textbook’   calculation   of   elasticities  
employed here is in fact only possible in theory, as the requirement to  
hold ‘all other variables constant’ can simply never be fulfilled in reality.  
Non­Confidential version 27

Changes in income, prices of other products, product promotions and a  
potentially endless range of other variables can affect the demand for a  
certain product, particularly of ‘luxury’ items, even in the short term. 
109. This is not to say that elasticity calculations are not possible, but rather,  
that they should take the functional  form of an econometric demand  
model,   that   uses   multiple   regression,   to   estimate   the   demand   for   a  
certain   product.   Such   a   model   would   assign   values   to   different  
influences   and,   combined   with   the   appropriate   econometric   tests,  
correct for bias problems. 
110. A calculation of own price elasticities based on mere sales data that  
would   hold   up   to   scientific   scrutiny   is   therefore   highly   unlikely.   The  
approach taken here, namely comparing two annual sales aggregates  
and   two   sets   of   retail   prices,   is   simply   too   crude   to   result   in   any  
meaningful findings, thus explaining the confusing sizes and signs of  
the elasticities found. 22 
111. The   Commission   also   correctly   points   out   that   the   cross­price  
elasticities submitted by the parties are counterintuitive as only 2 out of  
8 calculations even have the correct sign for substitution ­ the result  
must be positive, indicating that a higher price of one product leads to  
higher   demand   of   its   substitutes.     A   negative   is   indicative   of  
complementary   products.   The   cross­price   elasticities   suffer   from   the  
same weaknesses as the own­price elasticities, namely that changes in  
relative prices, income and other relevant variables are not taken into  
account. 
112. In summary, the elasticities submitted by the parties are of little or no  
assistance   in   determining   the   existence   or   extent   of   inter­category  
substitution   and   should   be   disregarded.     The   inadequacies   of   the  
elasticity calculations were in fact admitted by the parties.

elasticity calculations were in fact admitted by the parties.
The Brand Price Trade Off study
113. The Brand Price Trade­Off study (based on the data collected for the  
Brandy   Study)   analyses   substitution   between   specific   brands,  
22  On a different note, the positive signs of some of the own­price elasticities are interpreted by the  
Commission to indicate a Giffen Good and therefore summarily dismissed. A Giffen Good is an  
inferior good that constitutes such a significant part of a household’s budget, that the income effect of a  
price increase is greater than the substitution effect.  Indeed this effect is typically associated with the  
Irish potato case. However, it is possible in theory, and this is confirmed by Mr. Scannell’s testimony  
that a price increase leads to a change in the ‘extrinsics’ of the product (i.e. a certain amount of snob  
value is added) so that demand actually increases. Clearly this effect does not indicate a Giffen good.  
Instead, this would suggest a shift of the demand curve as opposed to a movement along the curve (by  
means of vertical product differentiation). Given that the marketing plans and consumer segmentation  
research support the notion of snob value (e.g. RTDs and FABs are status symbols), the idea of price  
increases leading to increases in sales volumes should not be discounted.
Non­Confidential version 28

spanning   several   liquor   categories .   Consumers   are   presented   with  
different   price   combinations   for   several   liquor   brands   and   asked   to  
indicate   their   purchase   preference.   The   liquor   brands   tested   include  
several brandy brands and several brands from other liquor categories  
including vodka, whisky and beer.
According to Forsyth, the parties’ expert witness, this methodology has  
been used by marketing departments for several decades to measure  
how consumers react to price changes. Forsyth explains why this is the  
preferred methodology in market survey:
‘The way it works is the following. It’s based on a presumption  
that if you just go to someone and say what are you willing to  
pay; are you willing to pay more and what have you, you tend  
not to get very sensible answers. What you need to do is to put  
something more into it, more, as much as you can in a research  
setting,   realistic   situations   where   they   are   choosing   between  
brands at different prices, kind of like they would be in a retail  
outlet,   trying   to   choose   a   product   to   buy.   And   then   observe  
those   different   choices   and   they   will   infer   from   that   how  
important price is and how important brand is in their different  
choice.’
114. Forsyth   explains   further   that   the   BPTO   study   ‘…would   be   very  
important for me as I put together my marketing strategy, because in  
my   marketing   strategy   what   I   want   to   do   is   I   want   to   take   certain  
brands  and   target   them   at   key   segments   and   then   look   at   those  
segments   to   see   where   I’m   at   risk   and   to   understand   that   as   I   put  
together my marketing plan and my brand plans, that I have to worry  
about   the   potential   switching   to   somebody’s   other   market,   to   these  
other products.’ 
115. The   Brandy   Study   found   that   the   majority   (..%)   (confidential)  of

115. The   Brandy   Study   found   that   the   majority   (..%)   (confidential)  of  
(confidential)   drinkers   consume   more   than   (confidential)  of   alcoholic  
beverage   within   a   ..­day   (confidential)  period,   compared   to   ..%  
(confidential)   of   (confidential)   drinkers   and   that   a   high   proportion   of  
(confidential)   consumers   are   at   risk   to   other   categories.   The   BPTO  
analysed   the   substitution   between   (confidential)   and   other   alcoholic  
beverages. 
116. The main findings were that  (confidential) consumption is price elastic  
as   a   10%   price   increase   (in   all   (confidential)  brands   under   scrutiny)  
leads to a   (confidential)  market  share loss of ..%   (confidential). This  
(confidential)  market share loss corresponds to the following gains in  
the market shares of the other products:  
Non­Confidential version 29

Table 3 Market share increases per product due to 10% price
increase in a particular liquor category (confidential)
Confidential  Confidential  Confidential  Confidential  Confidential 
Confidential  
%
Confidential  
%
Confidential  
%
Confidential  
%
Confident
ial  %
Source: Calculations based on BPTO
117. These   figures   suggest   inter­category   substitution,   at   least   between  
brandy brands, on the one hand, and beer, whisky and vodka brands,  
on the other.
118. In addition,  when  the price of   (confidential)  is  increased by 10%,  as  
much as ..%   (confidential)   of its market share is lost, which does not  
translate in equivalent market share gains for the other   (confidential) 
brands   from   which   the   respondents   could   choose,   thus   clearly  
indicating inter­category substitution. 23
Table 4 Market share increases due to 10% price increase in
a particular spirit brand (confidential)
Conf. Conf. Conf. Conf. Conf. Conf. Conf. Conf.
Conf.  % Conf.  % Conf.  % Conf.  % Conf.  % Conf.  % Conf.  % Conf.  %
Source: Calculations based on BPTO
119. From the perspective of a competition analysis, the main weakness of  
the BPTO study is  the fact  that  the brands included are purposively  
selected,   i.e.   they   are   a   selection   of   brands   presumed   to   compete  
against   each   other,   and   if   an   important   competitor   is   excluded,   the  
results will suffer in accuracy. 24 In this case the selection of the brands  
and price points were done for the brandy study itself and based on  
consumer focus group information that indicated that competition was  
taking   place   between   these   brands.   The   data   sample   included   a  
statistically significant number of respondents and was performed for  
the purposes of Distell’s marketing overhaul and not for the purpose of  
any competition investigation.
120. Note also that consumer responses gathered in this type of laboratory

120. Note also that consumer responses gathered in this type of laboratory  
experiment   are   likely  to  be   biased  towards  overstating  the   extent   of  
substitution.  That is, consumer responses tested in this way are likely  
to be stronger than their responses to a real life price hike. In real life  
consumers will not necessarily remember the prices they paid prior to a  
price increase, nor be able or willing to compare the prices of the near­
23  (Footnote omitted, contains confidential information). 
24  If an important competitor is excluded this would lead to an understatement of the elasticity. Note  
that the respondent is given the option of choosing ‘none’ of the selected brands selected for the  
purpose of the study as the likely winner in the event of a price increase.
Non­Confidential version 30

substitutes this closely. 
121. Other aspects of the methodology employed in this study are also open  
to criticism. The model does not allow for the possibility of consumers  
reducing alcohol consumption due to a price increase (either reducing  
their consumption of a particular brand or by switching out of alcoholic  
beverages completely) as consumers are expected either to continue  
buying their brand or to buy an alternative alcoholic beverage, and this  
stylisation   of   consumer   behaviour   will   further   amplify   the   expected  
response.
122. However,   these   shortcomings   notwithstanding,   the   BPTO   data   is  
impressive   and   the   findings,   albeit   open   to   qualification,   informative  
from a competition perspective.  We are, at least, satisfied that they are  
made on the basis of consumer research.   After all is said and done,  
consumers indicating their preferences based on their reserve prices  
constitutes   the   basis   of   how   demand   curves   are   constructed.  
Moreover,   whilst   maintaining   the   ceteris   paribus   condition,   the  
methodology   mimics   consumer   behaviour   in   a   SSNIP   test,   thereby  
predicting   substitutability.   The   approach   itself   is   interesting   in   that   it  
provides a manageable alternative for researching substitutability, one  
less involved than econometric modelling, and which, when prudently  
conducted, may be less prone to bias and other data related defects. 
123. On   balance   then   the   BPTO   appears   to   be   a   useful,   albeit   far   from  
flawless, measure of substitution.  However, while this particular study  
does   demonstrate   that,   in   the   face   of   a   generalised   increase   in   the  
price of a particular spirit   type,  consumers will substitute other liquor  
categories,   these   are,   by   and   large,   spirit   categories.     And,   while   it  
indicates that in the face of an increase in the price of a particular spirit

indicates that in the face of an increase in the price of a particular spirit  
brand, consumers will substitute other brands, again these are mostly  
spirit brands.   This may be a consequence of the range of possible  
substitutes   selected   –   but   while   it   constitutes   further   evidence   of  
substitution between spirit categories, it does not persuade us that beer  
and spirits are substitutable categories of alcoholic beverage. 
The Kwazulu­Natal Experience
124. In most substitution analyses the research possibilities would end here,  
with a hotchpotch of evidence of varying reliability, which, on a balance  
of probability, suggests that South African consumers consider various  
sprit categories as substitutes, but without a real case study, without,  
that is, the opportunity to analyse a sufficiently significant event that  
would   allow   for   a   more   confident   conclusion   on   substitution.  
Fortunately in this case there is a real life event that can be analysed  
and on which we place greatest reliance and which represents a major  
Non­Confidential version 31

challenge to the Commission’s already weak findings.  We refer to the  
dramatic   shifts   that   occurred   in   the   liquor   market   of   Kwazulu­Natal  
province   in   1997/8.   As   the   tables   presented   by   AC   Nielsen   show   a  
decline in the market share of brandy of some 8.6% was accompanied  
by a rise in the market share of vodka in the order of 34.0%.
125. The shifts between the two categories have been largely attributed to  
movements in two brands, namely Oude Meester brandy and Smirnoff  
vodka. When, in 1997, the price of Oude Meester increased to above  
R30,   large   decreases   in   market   share   ensued,   and   these   volumes  
were   lost   mainly   to   Smirnoff   vodka.   Both   the   Commission   and   the  
parties   attributed   this   evidence   of   substitution   to   Oude   Meester  
breaking a psychological price barrier by pricing itself at the thirty Rand  
mark.
126. On the face of it this provides powerful support to those who argue that  
consumers  will   substitute  between  categories  in  response  to  a  price  
increase.   But   why   does   this   response   appear   to   be   confined   to  
Kwazulu­Natal and does that fact allow us to follow the Commission  
and simply treat Kwazulu­Natal as a statistical outlier, an anomaly? In  
support of its claim to have this evidence treated as an anomaly, the  
Commission   in   fact   presented   data   that   excluded   Kwazulu­Natal,  
thereby   demonstrating   that   absent   the   data   from   this   province   the  
evidence of substitution is not significant.
127. The   witness   for   the   Commission,   Mr.   Lewis,   alludes   to   the   specific  
attributes   of   the   Kwazulu­Natal   liquor   market,   indicating   that   Oude  
Meester was consumed mainly by black consumers: 
Manoim: “But you have no explanation as to why Natal should  
be different in that respect.” 
Lewis: “Well I hope there are not too many people from Natal,

be different in that respect.” 
Lewis: “Well I hope there are not too many people from Natal,  
but historically Natal has always done its own thing. And what’s  
interesting, and again I’m not positive, because as I say it’s not  
my field of expertise, but I’m sure one witness will be able to say  
something   later   ­   that   the   Oude   Meester   brand   was  
predominantly consumed by the black market, not by the Asian,  
although a fair amount was obviously, but it was predominantly  
a black brand. So we have seen brands within that sector of the  
market change overnight. Now you can classically look at Lion.  
You can look at what’s happening to Castle at the moment. You  
can go through a list and when markets turn they turn quickly in  
that field.”
128. Lewis has in fact hit the nail on the head by indicating that the Oude  
Non­Confidential version 32

Meester drinkers in Kwazulu­Natal belong predominantly to a particular  
consumer demographic. It appears that province has a particularly high  
concentration of one particular consumer segment (as identified by the  
Brandy   Study)   who   switched   en   masse   as   the   result   of   the   Oude  
Meester   price   moving   above   R30.   This   segment   of   consumers   is  
characterised as:
(Paragraph omitted, contains confidential information)
129. Kwazulu­Natal   has   a   significantly   greater   concentration   of   this  
consumer segment than any of the other provinces.   It follows that if  
the rest of the consumers in this category are spread equally over the  
other 8 provinces, then one would expect the aggregate figures minus  
Kwazulu­Natal   to   indicate,   on   the   face   of   it,   a   considerably   lower  
degree of substitution.   In fact the responses of consumers in any of  
the other provinces may mirror that of their Kwazulu Natal counterparts  
but the extent and direction of the substitution is only clearly revealed  
when   Kwazulu   Natal,   the   locale   of   the   largest   concentration   of   this  
segment, is included in the aggregate data.   The inclusion of Kwazulu  
Natal does not distort the data – on the contrary the exclusion of the  
provincial data would vastly understate the data by excluding the bulk  
of the affected consumer segment.
130. Note that by excluding the Kwazulu­Natal numbers, the brandy­vodka  
substitution becomes less obvious, but the pattern of rising shares in  
gin and vodka when brandy is on the decline remains clear, particularly  
in the period 1998­2002, the period following the Oude Meester price  
increases:
Table 5 Trends in market shares of spirit categories 1993 –
2002 (brandy, gin, vodka)
1993 – 1997
(%)
1998 – 2002
(%)
Brandy + 1.1 ­ 9.6
Gin ­ 20.6 + 17.6
Vodka + 13.4 + 8.0
Source: Calculations based on AC Nielsen data
131. For the remaining categories, the trend is again one of either long­term  
decline or growth:

decline or growth:
Table 6 Trends in market shares of spirit categories 1993 –
2002 (cane, rum, liqueurs)
1993 – 2002
(%)
Non­Confidential version 33

Cane ­ 41.4
Rum + 78.6
Liqueurs + 86.4
Source: Calculations based on AC Nielsen data
132. The   Commission   argues   that   the   KZN   data   shows   that   the   price  
increases post­1997 occurred when the pricing was on the elastic part  
of   a   generally   price   inelastic   demand   curve   and   that   it   merely  
demonstrates that market power was exercised ‘a bit too far’. However,  
Oude   Meester   is   one   of   many   brandies   in   the   proprietary   (medium­
priced)   market   segment,   so   that   the   switch   to   an   alternative   liquor  
category   (rather   than   another   brandy   brand   within   the   broad   price  
segment) is evidence of inter­category substitution.
133. Obviously there are  some important caveats.   The  type of statistical  
‘eyeballing’   engaged   in   here   does   not   prove   causality.   Causality  
between the rise in the price in Oude Meester and the rise in Smirnoff  
sales   can   only   be   established   scientifically   when   a   reliable   demand  
model   of   the   sort   described   earlier   is   constructed.   However,   Oude  
Meester’s ‘small, but significant, non­transitory’ price increase coupled  
with the concentration in Kwazulu­Natal of the consumer segment at  
which   Oude   Meester   was   directed,   persuades   us   to   treat   this  
experience   as   the   single   most   plausible   piece   of   evidence   for   inter­
category substitution on the record. 
134. We   should,   in   concluding   this   discussion   of   the   Commission’s   and  
parties’ views of the relevant market, comment in greater depth on the  
Commission’s   insistence   that   where   inter­category   substitution  
occurred   –  as   in   the   Oude   Meester/Smirnoff   example   –  it   reflects   a  
case   of   monopolistic   pricing   taken   a   step   too   far,   a   case   of   the  
consequences   of   pricing   on   the   price­elastic   portion   of   a   generally

consequences   of   pricing   on   the   price­elastic   portion   of   a   generally  
inelastic demand curve, a phenomenon referred to as the ‘Cellophane  
Fallacy’ and first explored in the much cited case of   United States v.  
E.I. du Pont de Nemours and Co .25  In that case, it was argued that  
cellophane   was   not   a   separate   relevant   market   since   it   competed  
directly   and   closely   with   flexible   packaging   materials   such   as  
aluminium foil, wax paper and polyethylene. It was not recognised that  
a high own­price elasticity may mean that a firm is already exercising  
monopoly power and that, as the sole supplier of cellophane, Du Pont  
was   likely   to   have   set   prices   at   monopolistic   levels   above   which  
alternative products became substitutes. 
135. In   the   matter   before   us   the   Commission’s   contention   of   ‘cellophane  
fallacy pricing’ is based on its insistence that there is little difference in  
the   cost   of   production   between   a   ‘value   for   money’   brandy   and   a  
25  U.S. v. E.I. du Pont de Nemours & Co (1956 351 US 377).
Non­Confidential version 34

premium   and   proprietary   brandy.     On   the   basis   of   this   argument   it  
concludes that all brands in the higher ‘proprietary’ and ‘premium’ are  
priced at supra­competitive levels. The Commission concludes that the  
prices   should   be   adjusted   to   ‘competitive   levels’   before   determining  
whether they are substitutable with other competitively priced products.  
This argument serves to underpin the Commission’s narrow relevant  
market definition. 
136. Admittedly,   the   ‘cellophane   fallacy   effect’   should   be   borne   in   mind  
when   using   demand   elasticities   to   determine   market   power   or   the  
extent   of   the   relevant   market,   as   any   profit­maximising   firm   with   a  
degree of market power will set prices at a level where demand for its  
product   is   elastic   (otherwise   it   would   raise   prices   further).   Using  
elasticities   that   are   based   on   elastic   demand   pricing   are   therefore  
misleading, as substitution at monopoly prices is much more likely than  
at competitive levels. As Bishop & Walker remark: 
“The   mere   fact   that   at   the   monopoly   price,   a   monopolised  
product faces demand substitutes does not mean that the firm  
producing   the   product   has   no   market   power.”   It   is   therefore  
generally   argued   that   in   non­merger   inquiries   observed   own­
price elasticities may understate the degree of market power.” 26 
137. For  merger enquiries  however  the elasticity of  demand refers to the  
elasticity at the prevailing price level as this assessment is concerned  
with future market power. The Commission is therefore not justified in  
suggesting   that   the   prices   should   be   adjusted   to   competitive   levels  
before price elasticities are used. 
138. The   Commission’s   argument   that   the   observed   decreases   in,   for  
instance, Oude Meester brandy sales are due to price increases that

instance, Oude Meester brandy sales are due to price increases that  
“pushed it towards the elastic part of Oude Meester’s normally inelastic  
demand curve”, is based on the production cost of Klipdrift brandy and  
a reference to the brandy report which indicated that:
(Paragraph omitted, contains confidential information)
139. According to the Commission, the Tribunal further “heard evidence that  
there is little difference in the cost of production between a ‘value for  
price’ brandy and a premium or proprietary brandy”. The Competition  
Commission   submits   that   brands   in   the   proprietary   and   premium  
categories are priced well above competitive levels, and that therefore  
their   prices   should   be   adjusted   to   competitive   levels,   before  
determining   whether   they   are   substitutable   with   other   competitively  
26 ‘ Economics of E.C. Competition Law: Concepts, Application and Measurement’ Sweet and  
Maxwell, 1999
Non­Confidential version 35

priced products. In the absence of this adjustment, interpretations of  
pricing behaviour will fall prey to the cellophane fallacy.
140. There are two problems with this line of reasoning. Firstly, there is no  
evidence presented by the Commission that establishes that it is price  
inelasticity (reflected in a movement along the demand curve) rather  
than vertical  differentiation (reflected in a shift of the demand curve)  
that   allows   for   higher   prices   ­   the   evidence   from   the   parties   merely  
indicates   that   substitution   takes   place.   The   Commission   attempts   to  
argue   that   any   pricing   above   marginal   cost   provides   evidence   that  
demand is inelastic.  In fact pricing at marginal cost only occurs when  
all   the  conditions   of  perfect  competition  are  met   –  it  is  a  theoretical  
artefact.  Pricing above marginal cost is evidence of an imperfection in  
the   market   –   notably,   in   this   instance,   product   heterogeneity   or  
differentiation ­ that enables the exercise of market power, but it fails to  
provide any evidence on the elasticity of demand in market analysis,  
and, hence, of cellophane fallacy­type pricing behaviour.
141. Secondly, the ‘admission’ by the parties that there is little difference in  
the production cost of the different types of brandy does not include  
marketing expenditure or image­driven strategic pricing.  Note also that  
the parties have presented evidence that demonstrates that the costs  
of   producing   a   premium   or   ‘prop’   (that   is   high   and   medium   priced)  
brandy   is  higher   than  that   of  producing  a   value  (that  is,   low   priced)  
brandy.     We   have,   in   short,   been   presented   with   no   evidence   on  
marginal cost – we have simply been presented with an assertion to  
the effect that the differential in price between two brands of brandy  
establishes   that   the   higher   priced   brandy   is   priced   at   a   supra­
competitive price.

competitive price. 
142. Moreover, it should be borne in mind that the cellophane fallacy was  
identified in a situation where the supplier was the sole manufacturer of  
a particular product so that an exercise of market power would lead  
consumers   to   consider   alternative   products   with   different  
characteristics. The situation is very different in the South African liquor  
market.   Oude   Meester   is   one   of   many   brandies   in   the   proprietary  
(medium­priced)   market   segment   to   which   consumers   could   have  
switched in response to a price increase of Oude Meester. Therefore,  
the switch to an alternative liquor category altogether, more plausibly  
indicates genuine substitution between spirit categories rather than a  
market power­induced switch to products that are not really substitutes.  
This point was noted by the parties:
 
Mr.   Rogers :   “Indeed   the   existence   in   a   market   of   significant  
branding and the availability of brands across a wide range of  
prices   is   incompatible   with   the   application   of   the   so­called  
Non­Confidential version 36

cellophane fallacy.”
143. The   Commission   itself   acknowledges   that   overly   narrow   market  
definitions may penalise vertical product differentiation:
Adv. Pretorius : “In the United States monopolisation complaints  
are not recognised in respect of single product categories. The  
reason for this is that it seems to create markets which are too  
narrow and therefore punish the success of a company makes  
him   dominant   or   monopoly,   who   has   successfully   branded   its  
products   …   who   has   differentiated   his   products   from   other  
products through successful branding. The FTC has a tendency  
to   define   narrow   product   margins   often   resulting   in   the  
punishment  of  companies  that  have successfully  differentiated  
their products from very similar products.” 
144. However, the Commission nevertheless opposes a segmentation of the  
market that reflects the wide price differentiation within each category  
or group of categories because, it insists, there are no differences in  
the intrinsics of the products involved:
Adv.   Pretorius :   “We   submit   that   these   types   of   distinctions  
protect   consumers   who   could   unnecessarily   because   what  
prevents   them   from   informing   themselves   better   about   the  
content of the brandy or the content of what is in the superstore.  
And   places   an   undue….   places   an   undue   premium   on   social  
status such as in these instances that people often drink Bells  
because they would like their friends to see that they drink Bells.  
Although   the   content   remains   the   same.   On   the   basis   of   the  
hypothetical monopolist to distinguish between these categories  
in   our   view   is   protecting   a   lazy   consumer   and   a   consumer  
preoccupied with his social status in the community. If we accept  
… if we accept that there is no quality difference or the quality  
difference is limited and in any event not to the extent that it is  
double in price.”

difference is limited and in any event not to the extent that it is  
double in price.”
The Relevant Market: a third perspective
145. As   already   indicated,   we   do   not   believe   that   the   Commission   has  
adequately supported its argument for a number of narrow, category­
based   relevant   markets.     The   extensive   evidence   and   argument  
presented by the parties casts severe doubt on the Commission’s case  
and   has   contributed   to   our   rejection   of   the   latter’s   version   of   the  
relevant market.  This is certainly true where spirits are concerned – on  
the evidence before us we are able to conclude that there is sufficient  
substitutability between the various spirits categories to include them in  
Non­Confidential version 37

the same relevant market, although, as elaborated below we believe  
that this substitutability occurs within broad price bands.   As we shall  
indicate,   the   Commission’s   arguments   for   placing   the   various   spirit  
categories in separate relevant markets is plausible only with respect to  
the highest price band. 
146. However,   although   we   are   persuaded   that   there   is   considerable  
substitutability between the various spirit categories, we nevertheless  
reject the parties’  argument for a single alcoholic beverages market,  
particularly insofar as it presupposes the inclusion of beer in the same  
market as spirits.   The parties have attempted to rely on the brandy  
study and their calculation of the ‘Brand Price Trade Off’ to establish  
the substitutability of beer and spirits but, as we have already pointed  
out, this evidence is not wholly persuasive.  
147. While the parties will no doubt insist that it is not for them to establish  
the substitutability of beer and spirits, nor is it acceptable to make a  
wholly   novel   and   counter­intuitive   claim   to   the   Tribunal   without  
establishing a factual basis for the claim.   Spirits and beer are, to be  
sure, both beverages and both contain alcohol but their ‘intrinsics’ are  
otherwise distinct, their production processes are equally distinct, and  
we are aware of no single instance in which the claim that they belong  
to the same relevant market has been made, let alone upheld.   Fish  
paste and beluga caviar are both commonly spread on crackers and  
both have some relationship to fish, but this does not make a claim to  
place them in one market at all plausible.
148. We   note   too   that   breweries   and   distilleries   are   still,   by   and   large,  
housed within wholly separate corporate entities.   Indeed when South  
African Breweries and the various distilleries entered into their market  
sharing  arrangement they did so on  the basis of a division between

sharing  arrangement they did so on  the basis of a division between  
beer, on the one part, and, on the other, spirits and wine.  In the recent  
past the brewers and the distillers have entered into competition in the  
very diverse FABs market.  Indeed we have indicated that we believe  
that   the   evidence   presented   in   this   transaction   suggests   that   the  
competitive   interface   between   spirits,   FABs,   wine   and   beer   is  
considerably   more   complex   than   that   contended   for   by   the  
Commission.    However,  we  do not  think that  it supports  the parties’  
argument for a single relevant market comprising all of these various  
categories.     We   find,   instead,   that   there   are,   implicated   in   this  
transaction: 
 Three   relevant   spirits   market     ­   a   ‘premium’   spirits   market,   a  
‘proprietary’ or ‘prop’ spirits market and a ‘value’ spirits market.   The  
basis for this three­way division is the vast price differential between  
the   various   spirits   brands.     This   is   further   elaborated   immediately  
Non­Confidential version 38

below. 
 We find secondly that there is a wine market.  Indeed, consistent with  
the   basis   on   which   we   have   divided   the   spirits   markets,   there   are  
undoubtedly several wine markets whose boundaries reflect the vast  
price differential between premium wine, at one end of the spectrum,  
and   low   value,   relatively   inexpensive   wine,   at   the   other   end   of   the  
spectrum.   However, we have been provided with insufficient evidence  
to identify with any confidence the location of the boundaries between  
brand or label price bands. 
 We are not convinced that there is sufficient evidence to arrive at a  
definitive   conclusion   regarding   FABs.     This   is   clearly   a   very   broad  
ranging category with certain brands and sub­categories, for example  
the various ciders, possibly in the same market as beer, while others  
are more likely to compete with spirits, while yet others may well be  
best situated in one or other of the wine markets.  While it is precisely  
the varied nature of the products collectively referred to as FABs, that  
allows them to act as a bridge between spirits, wine and beer, it is also  
this quality that resists attempts to pigeonhole FABs in a single relevant  
market.     
149. Spirits   are,   in   our   view,   the   potential   area   for   concern   in   this  
transaction.     As   already   elaborated,   we   reject   the   Commission’s  
argument   in   favour   of   placing   the   traditional   spirit   categories   in  
separate relevant markets.  We believe that there is sufficient evidence  
of inter­category substitution to support a finding that places different  
categories of spirits in the same market.  However, we do not believe  
that  there  is  a  single  spirits market.    Instead,  we find three relevant  
spirits markets. 
150. However, in contrast with the Commission’s view, our spirit markets are  
not segmented by the category of spirits, by, as it were, the contents of

not segmented by the category of spirits, by, as it were, the contents of  
the bottle, but rather by the vast price differential within each of the  
various  categories.    Hence,  dare  we  say  it,  just   as  ‘common  sense’  
would reject an argument for placing a multi­million Rand Rolls Royce  
in the same market as the humble Opel Corsa, so too does it rebel at  
the notion that Remy Martin competes with Wellington (or Cape to Rio)  
or Chivas Regal with First Watch (or Russian Bear).  However, as we  
shall elaborate below, we do think that the parties’ have, through their  
evidence and argument on substitutability, made out a sufficient case  
for a relevant market that includes, for example, the likes of a brandy  
like Wellington, a whisky like First Watch, a cane spirit like Cape to Rio  
and a vodka like Russian Bear because they belong in the same broad  
price brand, just as we accept a second relevant market that includes,  
for example, Smirnoff vodka, Gordon’s gin, Klipdrift brandy and Bells  
whisky.  
Non­Confidential version 39

151. In other  words, we are  persuaded  that  there is  substitution between  
traditional   spirit   categories   but   we   are   equally   persuaded   that   this  
substitution   will   take   place   between   spirits   in   the   same   broad   price  
band.  We do not claim to be able to identify precisely the boundaries  
of   each   of   those   bands   but,   just   as   in   previous   mergers   in   other  
consumer   goods   industries,   we   were   satisfied   to   accept   the   LSM  
categories in common usage in those sectors, so here are we satisfied  
to   accept   the   broad   categories   delineated   as   value,   proprietary   and  
premium   that   are  commonly  used   in  the   trade.     Hence  faced  by  an  
increase   in   the   price   of   Wellington   a   consumer   may   well   seek   the  
comfort of another value brandy brand, but so too may the consumer in  
question turn to First Watch whisky.  However, the affected consumer  
is unlikely to substitute a proprietary or premium brand of any category.  
By the same token an increase in the price of Smirnoff vodka may well  
force a hitherto loyal customer into the hands of Gordon’s gin or Bells  
whisky   but   it   is   unlikely   to   have   him   reaching   down   into   the   value  
brands or ascending into the premium sector.
152. Let us be clear that these are spirits markets. That is, we do not believe  
that   inter­category   substitution   is   unlimited.     That   is,   while   we   are  
persuaded that there is sufficient evidence of substitution between the  
spirit categories in the same broad price band to justify several price­
segmented spirits markets,  we do not believe that  there  is  sufficient  
evidence to include FABs, wine or beer in any of these markets.  
153. An   immediate   qualification   is   required   here:   we   are   persuaded   that  
there   is   sufficient   evidence   that   changing   tastes,   new   products   and

there   is   sufficient   evidence   that   changing   tastes,   new   products   and  
rapidly evolving demographics are injecting new dynamic features into  
the   broad   alcoholic   beverages   market.     For   example,   we   are  
persuaded that FABs not only constitute an important new category of  
alcoholic beverage, one that has clearly made inroads at the expense  
of   other   liquor   categories,   but   that   they   also   constitute   an   important  
bridge between consumers of spirits and beer, spirits and wine, and  
wine   and   beer.   Or,   by   way   of   a   second   example,   we   believe   it  
reasonable to hypothesise that inexpensive wine is part of the same  
relevant   market   as   the   lowest   price   spirit   band   where,   several  
witnesses have pointed, there is particularly strong evidence of inter­
category   substitutability.     But   we   do   not   have   sufficient   evidence   to  
broaden these relevant markets in this way – hence we will restrict our  
relevant spirits market to the traditional spirit categories although we  
will factor the interplay between FABs and spirits into our assessment  
of competition in the middle segment spirits market – the ‘proprietary  
market’ – and we will consider the influence of inexpensive wine in our  
assessment   of   the   bottom   segment   –   the   ‘value   segment’   –   of   the  
spirits markets.  
Non­Confidential version 40

154. Our   argument   for   a   relevant   market   that   is   based   on   inter­category  
substitution and income or living standard based segmentation draws  
on more than common sense. 
155. We have already dealt with the evidence on substitution between the  
traditional   spirit   categories   at   some   length   and   have   concluded   that  
there is sufficient evidence to factor this phenomenon into the definition  
of the relevant market.
156. As for the market segmentation based on price differentials, consider  
how   deeply   embedded   in   the   thinking   of   industry   participants   is   the  
three­way distinction between ‘value’ brands, ‘proprietary’ brands and  
‘premium’   brands.     Most   of   the   witnesses   who   participated   at   the  
hearing   articulated   their   marketing   strategies   and   that   of   their  
competitors in terms of this approach. 
157. While   these   consumer   segments   are,   by   no   means,   absolutely  
coterminous   with   income   level   (nor,   we   should   note,   is   the   LSM  
concept   conventionally   used   in   market   research),   there   is   clearly   a  
close relationship between branding strategy and price simply because  
income and the related ability to participate in the market is a critical,  
the most critical, element in the construction of the consumer segment  
at   which   a   particular   brand   is   targeted.     Certainly,   a   fundamental  
marketing   strategy   is   to   drive   consumers   up   the   brand   ladder,   to  
persuade them to desert the lower reaches of the market in favour of  
higher value brands.  But there are limits imposed on this strategy – as  
the saying goes ‘the poor will always be with us’, the rich will always be  
willing   to   pay   a   premium   for   consuming   the   ‘extrinsic’   qualities   that  
identify them as a group distinguishable from the rest of society, and  
entrepreneurs will constantly strive to segment their markets.

entrepreneurs will constantly strive to segment their markets.  
158. The occasions on which the distinction between the value, proprietary  
and premium categories informed the various analyses presented to us  
by   all   the   witnesses   are   too   numerous   to   list.     Seagrams   and   DGB  
presented us with a detailed map locating the various spirit brands in  
the   three   segments   of   value,   proprietary   and   premium.     Snell’s  
competitive advantage (and, we are told, its Achilles heel) manifests  
the   existence   and   significance   of   this   market   segmentation   most  
starkly.     Snell   is   the   dominant   producer   and   distributor   of   ‘value’  
brands,   those   brands  serving  the   bottom   end   of   the  market.     It   has  
developed highly successful brandy, whisky and white spirits brands.  
But it is struggling to transform one or more of its brands from the low  
margin value segment to the higher margin proprietary segment.    In  
recent years it has been anchored to the value segment by Distell’s  
efforts to penetrate that segment.  
Non­Confidential version 41

159. Clearly,  while  the  low   value  segment  of  the market  does  not,  on its  
own, particularly attract Distell, by entering Snell’s turf, Distell increases  
the   already   considerable   risks   that   attach   to   Snell’s   attempts   to   re­
brand   an   established   successful   low   value   brand   –   Snell   is  
simultaneously   taking   on   the   likes   of   Distell   and   GUDV   in   the  
proprietary sector, while, through migrating a successful value brand, it  
eases the ability of these formidable competitors to penetrate the value  
segment.     Indeed,   this   predicament,   which   we   re­examine   below,  
appears to be the basis of Snell’s objection to the transaction.
160. What emerges then is a plausible third view of the relevant market, one  
distinct from the narrow category­based market definition contended for  
by   the   Commission,   but   also   clearly   distinguishable   from   the   overly  
broad definition advanced by the parties.   It is a view of the relevant  
market   that   takes   account   of   the   particular   features   of   liquor  
consumption   in   South   Africa   which,   for   present   purposes,   is   most  
starkly manifest in a blurring of the distinction between various spirit  
types. 
161. Moreover,   our   view   of   the   relevant   market   recognizes   the   central  
significance of branding in the marketing of alcoholic beverages. Inter  
alia, the evidence of substitution between Oude Meester brandy and  
Smirnoff   vodka,   indicates   that   increasingly   spirit   manufacturers  
compete   on   image   and   branding,   on   the   ‘extrinsics’   of   the   product,  
rather than on spirit category, the ‘intrinsics’. This means that, when  
faced   with   a   significant   price   increase,   a   consumer   that   prefers   a  
branded brandy is likely to switch to another branded spirit, possibly a  
white spirit or even a FAB, of equal ‘extrinsic’ appeal rather than to a

white spirit or even a FAB, of equal ‘extrinsic’ appeal rather than to a  
less expensive, but less ‘extrinsically’ satisfying, product in the same  
traditional spirit category. 
162. Contrary to the conclusions reached in the other jurisdictions that have  
examined   liquor   mergers,   we   view   the   rise   in   the   importance   of  
branding   –   something   acknowledged   by   all   participants   in   these  
hearings   ­   as   contributing   to   a   decrease   in   the   significance   of   the  
differentiation   between   traditional   categories.     Furthermore,   the  
successful penetration of FABs, has introduced a dynamic element into  
the market that has contributed to this blurring of categories.    
163. We find then that there are three relevant spirits markets implicated in  
this   transaction   each   bounded   by   the   broad   price   categories   of   the  
products that populate these markets.  We will use the language of the  
trade   to   delineate   these   markets   and   so   dub   them   the   value,  
proprietary and premium markets.  
164. We   are   aware   that   where   the   blurring   of   product   categories   is  
Non­Confidential version 42

concerned   our   conclusions   are   likely   to   apply   more   strongly   to   the  
value   and   proprietary   markets   than   to   the   premium   market.     Most  
market   participants   acknowledge   occasion­based   drinking   and  
category loyalty in respect of the premium market. At the opposite end  
of   the   spectrum,   the   most   extreme   example   of   category   blurring   is  
apparent in the value  market segment.  Clearly this  segment attracts  
the most price­conscious consumers who would more readily substitute  
between   different   spirit   types   and,   further   research   may   well   reveal,  
may   even   substitute   between   spirits   and,   because   of   this   country’s  
unusually large market for inexpensive wine, wine.
165. This caveat is supported by the evidence of Mr. Tim Hutchinson, the  
witness from DGB, who, while insisting on the salience of traditional  
categories in the premium segment, conceded that switching between  
categories is more pronounced in the value segment:
Mr. Hutchinson :   “Well I think if you did a lot of research on a  
good loyal brandy consumer, he’s a brandy drinker. You know,  
you are a gin drinker. There can be arguments in the market like  
Natal   that  the  consumer  there  in the  emerging market  swaps  
over between gin and vodka. There could be an argument that  
says that as whisky gets more expensive at the bottom­end of  
the market, you might have people who, on just pure economic  
grounds, are gravitating towards brandy, just from a retail selling  
price point. 27”
Is there likely to be a substantial lessening of competition?
166. The Commission has determined that wine and FABs produced by the  
merging   parties   belong   to   separate   relevant   markets.     Distell   has   a  
significant share of each of these markets.
167. Despite Distell’s significant share of the FABs market, the Commission  
has not found a substantial lessening of competition in that market.  It

has not found a substantial lessening of competition in that market.  It  
accepts   the   parties’   argument   that   FABs   constitute   a   particularly  
dynamic market characterized by fickle consumer behaviour and rapid  
new entry and innovation.  The Commission avers that these features  
of   the   FABs   market   will   undermine   any   attempt   on   the   part   of   the  
merged   entity   to   exercise   market   power.     Bulmer,   the   multinational  
FABs producer, has offered evidence purporting to demonstrate that  
FABs’   consumers   are   willing   to   absorb   significant   price   increases  
before switching to alternative beverages.
168. We  note  also  that   the  FABs  category  is  extremely  diverse  – certain  
27  T3: p. 162.
Non­Confidential version 43

FABs are effectively mixed spirits, others are cider products, while a  
third  category  is  akin  to  fortified  wine.    It   is by  no means clear that  
these are part of a single market.  Further evidence may well show that  
some should be placed in one of the spirits markets, that cider should  
be placed in the beer market and that the remainder should be in the  
wine market. Moreover, the evidence regarding substitutability between  
FABs and other alcoholic beverages is conflicting. The Lexecon study  
performed for Bulmer suggests that FABs consumers are very brand­
loyal, but there is little to suggest that they are equally category­loyal,  
and that, when faced with a price increase in a particular FAB brand,  
consumers   will   switch   predominantly   to   other   FABs.   Moreover,   the  
parties   marketing   research,   as   well   as   several   witnesses   and   other  
industry   sources,   suggest   that   substitution   between   FABs   and   other  
liquor categories is likely.
169. For   all   these   reasons   we   accept   the   Commission’s   finding   that   the  
transaction will not lead to a substantial lessening of competition  
in the FABs market.   
170. The Commission has divided the wine market into separate categories,  
notably table wine, fortified wine and sparkling wine.  We note that, as  
with the various spirits categories, there is a significant divergence in  
the   prices   of   different   wine   brands   or   labels   and   this   will   certainly  
dictate   a   further   segmentation   of   the   wine   market   into   a   number   of  
markets defined by a price band.  However, as already noted, we have  
been   provided   with   insufficient   overall   data   on   the   wine   market   to  
identify the boundaries of these segments.
171. Despite   high   post­merger   shares   in   the   table   wine   market   and   the  
possible  anti­competitive  effects   of   the  merger,   the  Commission   has

possible  anti­competitive  effects   of   the  merger,   the  Commission   has  
found that the merger ‘could be justified by the ability of the parties to  
become internationally more competitive’. While we are sceptical of the  
Commission’s   conclusions   on   efficiencies   –   there   appears   to   be   no  
reason   why   inter­firm   cooperation   on   international   wine   marketing  
requires a merger ­ we have not been provided with sufficient evidence  
to make a finding on the competitive consequences of the merger in  
this market. We note however persuasive evidence on the low entry  
barriers   in   this   market   and   the   existence   of   competing   cooperatives  
and   numerous   wine   estates   and   will   accordingly   not   investigate   this  
matter further.  
172. In the fortified wine market, the Commission argues that since there is  
‘no   growth   taking   place   in   this   market’,   it   is   highly   unlikely   that   the  
parties would be able to obtain market power through this transaction.  
Again we have not been provided with sufficient evidence to sustain a  
clear finding in this market – we are, indeed, yet to be persuaded that  
Non­Confidential version 44

fortified wine constitutes a separate relevant market.  We note also the  
relatively   insignificant   size   of   this   market   compared   to   the   total   of  
alcoholic   beverages   and   will   accordingly   not   investigate   this   matter  
further. 
173. The Commission has identified a separate relevant market for sparkling  
wine, arguing that its consumption is reserved for particular occasions  
for   which  other   wine  does   not   substitute.    The   parties,  on  the  other  
hand, have argued that the link between the consumption of sparkling  
wines, on the one hand, and, on the other hand, particular occasions is  
breaking down.  We note too that the merger combines SFW’s 31,5%  
pre­merger share with Distillers 43,7% share, of which the lion’s share  
(40,6% of the total) is accounted for by a single brand, JC le Roux, with  
the   rest   of   the   market   populated   by   a   myriad   of   very   small   brands.  
Short of an outright prohibition, this makes it very difficult to construct  
an   effective   remedy.       However,   again   we   note   the   inadequate  
evidentiary   basis   for   definitively   determining   the   existence   of   a  
separate relevant market in sparkling wine. We note also that sparkling  
wine represents a minimal part of the transaction under examination  
and that entry barriers are relatively low as effectively any current wine  
producer is a potential sparkling wine producer and entry barriers into  
wine production per se are low. Accordingly, we will not examine this  
issue further.
  
174. Our   evaluation   of   the   transaction’s   impact   on   competition   is   thus  
restricted   to   the   three   spirits   markets.     There   are   the   ‘value   spirits  
market’,   the   ‘proprietary   (or   ‘prop’)   spirits   market’   and   the   ‘premium  
spirits market’.
175. We have calculated market shares to the best of our ability although  
we have been hampered by significant data imperfections.

we have been hampered by significant data imperfections.
Table 7 Merging parties market shares in different price
segments
Shares per category SFW Distillers Cumulative
Value 21.4% 8.5% 29.9%
Proprietary 7.7% 37.6% 45.3%
Premium 16.4% 26.1% 42.5%
Source: Calculations based on AC Nielsen data
176. In summary, the above table shows that the merged entity commands  
a   share   of   approximately   29,9%   of   the   value   market,   45,3%   of   the  
proprietary  market   and  42,5%  of  the  premium  market.  However,  the  
impact   of   the   merger   differs   between   the   three   relevant   markets.  
Hence the table above illustrates that the pre­merger SFW enjoyed a  
significantly greater presence in the value segment (21,4%) than did its  
erstwhile   rival,   Distillers   (8,5%).   This   is   also   marked   in   the   ‘prop’  
Non­Confidential version 45

segment where Distillers (37,6%) – largely because of its commanding  
presence   in   brandy   –   overshadowed   SFW   (7,7%).     In   the   premium  
segment, on the other hand, the pre­merger entities were, with SFW at  
16,4% and Distillers at 26,1%, of similar size and so the accretion of  
market   share   is   greater   than   in   respect   of   the   other   two   segments,  
although   here   too   Distell   owes   the   lion’s   share   of   its   position   to  
Distillers’ contribution.  
177. One possible conclusion that might be drawn from these data is that  
because the pre­merger value and proprietary markets were dominated  
by   SFW   and   Distillers   respectively,   competition   has   not   been  
substantially  lessened   by  the  merger.     We   are,   however,   inclined  to  
treat this argument cautiously – although the accretion, and hence the  
‘lessening of competition’, may appear modest when a firm with very  
large market share merges with a firm enjoying a smaller share, this  
may,   by   the   same   token,   reflect   the   removal   of   one   of   the   few  
remaining sources of competition. 
178. This is of particular concern in a market structured along the lines of  
the spirits markets that we are currently scrutinising.  There is not, after  
all,   vibrant   competition   beyond   the   walls   of   Distell.     In   the   value  
segment   there   is   Snell   with   a   very   large   market   share.     In   the  
proprietary segment, GUDV looms very large indeed, Out of the total  
spirits   market   (i.e.   value,   prop,   and   premium   combined)   GUDV   is  
estimated   to   have   a   market   share   of   23   ­   29%. 28  Most   of   GUDV’s  
products   are   in   the   prop   and   premium   markets,   including   powerful  
whisky brands (including J&B, Johnny Walker and Bell’s) and vodka  
brands (Smirnoff) as well as smaller stakes in brandy, gin, and FABs.  
This is graphically illustrated in Annexure ‘A’.

This is graphically illustrated in Annexure ‘A’.
179. In   the   context   of   this   highly   concentrated   market   structure   even   a  
relatively small pre­merger market share may represent one of the last  
hopes for robust competition. 
180. The following two tables show the composition of the spirit portfolio of  
each   of   the   parties   prior   to   the   merger.     The   first   of   these   tables  
illustrates the particularly powerful position of the pre­merger Distillers  
in proprietary and premium brandy brands.  This contrasts with SFW’s  
position as revealed in the succeeding table.   SFW clearly did not, in  
any of the other spirit categories, enjoy an equivalent position to that  
held   by   its   merger   partner   in   brandy.     It   is,   instead,   more   broadly  
spread across the various liquor categories with noteworthy positions in  
value white spirits and whisky as well as premium brandy.
28  Based on a submission by E. Snell & Co, which contains the 29% estimate. GUDV’s submission  
contains the similar, albeit lower, number of 23%.
Non­Confidential version 46

Table 8 Distillers’ market share per liquor category
Distillers   shares  
per category
Brandy White spirits Whisky Rum* Total spirits
Value 16.2% 5.5% 3.1% 0.0% 8.5%
Proprietary 77.4% 27.9% 2.4% 0.0% 37.6%
Premium 46.1% 0.6% 3.5% 24.8% 26.1%
  Source: Calculations based on AC Nielsen data
* Based on assumptions
Table 9 SFW’s market share per liquor category
SFW   shares   per  
category
Brandy White spirits Whisky Rum* Total Spirits
Value 20.7% 28.4% 28.3% 0.0% 21.4%
Proprietary 11.0% 14.6% 0.0% 0.0% 7.7%
Premium 24.2% 0.0% 18.8% 0.0% 16.4%
Source: Calculations based on AC Nielsen data
* Based on assumptions
181. The   following   table   shows   the   market   shares   in   various   spirit   type   /  
price level categories of the merged entity.
Table 10 Distell’s market share per liquor category
Distell   shares   per  
category
Brandy White 
spirits
Whisky Rum* Total spirits
Value 36.9% 33.9% 31.4% 0.0% 29.9%
Proprietary 88.4% 42.6% 2.4% 0.0% 45.3%
Premium 70.2% 0.6% 22.3% 24.8% 42.5%
Source: Calculations based on AC Nielsen data
* Based on assumptions
182. In summary then, in the proprietary and premium spirit markets Distell  
owes   its   position   largely   to   Distillers’   dominance   of   the   brandy  
category.   In the value segment Distell’s strength derives largely from  
SFW’s contribution.  The upshot is that while Distell’s market shares in  
each of the relevant markets under scrutiny is, on the face of it, cause  
for   concern,   this   is  somewhat   ameliorated   by   an   examination   of   the  
relatively small additions to market share achieved in consequence of  
the merger. 
183. This inference – an inference that, we reiterate, is to be drawn with  
considerable caution ­ is reinforced by the fact that both of the parties  
to this  transaction  had  identical  shareholders.   While  the transaction  
clearly resulted in a change of control and hence required notification  
and evaluation, it remains nevertheless that their identical ownership

and evaluation, it remains nevertheless that their identical ownership  
structures   undoubtedly   compromised   the   intensity   of   pre­merger  
competition between the parties.
Non­Confidential version 47

184. We   are   required,   then,   to   examine   additional   features   of   the   post­
merger   markets,   before   concluding   definitively   whether   or   not  
competition has in fact been substantially lessened.  We will show that  
this further examination reveals that whereas in the value and premium  
markets   there   is   not   a   substantial   lessening   of   competition,   in   the  
proprietary market, on the other hand, competition is compromised by  
the transaction.  We will examine each of these in turn.
The value market
185. In the value segment Distell enjoys a market share of slightly under  
30%.  The bulk of this share – of the order of 21,4% ­ was contributed  
to   the   merged   entity   by   SFW.     The   accretion   is   small   but   not  
inconsequential – the merger is clearly removing a competitor of some  
significance from the value segment.
186. However,   there   are   other   factors   that   ameliorate   the   impact   of   the  
merger on competition.
187. Firstly, Distell is not the pre­eminent presence in this segment.   This  
position is clearly occupied by Snell which owns, inter alia, the most  
successful   brand   in   value   whisky   (First   Watch),   value   brandy  
(Wellington VO), value vodka (Russian Bear), value cane (Cape to Rio)  
and   value   Gin   (Strettons   Deluxe   Gin).     Snell   has,   with   considerable  
success, focused on this low value segment.   Snell has achieved its  
powerful position in this market remarkably quickly and, although price  
rather than branding has been the key determinant of market share in  
this segment,  it  appears  that it  has managed to  establish  significant  
brandy   and   whisky   brands   which   has   enabled   it   to   charge   a   small  
premium over its competitors’ brands in the value market.
188. However,   margins   in   the   value   market   are   extremely   tight.     Indeed,  
Snell’s success partly resided in the reluctance of established distillers

Snell’s success partly resided in the reluctance of established distillers  
like   Distell   and   GUDV   to   participate   in   this   segment.   Having  
established a powerful beachhead in the low value market, Snell is now  
attempting   to   migrate   some   of   its   successful   value   brands   into   the  
proprietary market where margins are significantly greater.  
189. Distell now appears determined to penetrate the value market. It is not  
clear precisely why Distell has decided to enter this market, although  
clearly   Snell’s   success   and,   particularly,   the   prospect   that   this   may  
ease   its   ability   to   penetrate   the   proprietary   sector,   must   have  
influenced Distell.
190. Distell’s   entry   into   the   value   market   clearly   represents   a   significant  
Non­Confidential version 48

threat to Snell.  Not only does it represent a direct challenge to Snell on  
its own terrain, but it increases the risk to Snell of migrating one of its  
successful value brands into the proprietary market.   Snell avers that  
Distell requires the merger in order to successfully prevent entry to the  
proprietary   market.     Snell   argues   that   the   value   market   with   its   low  
margins represents a major challenge to a large company supporting a  
relatively  huge  overhead  structure.  Snell   argues that  the  merger  will  
enhance   Distell’s   market   power   in   the   high   value   proprietary   and  
premium market segments and that the additional returns gleaned from  
the transaction will subsidise its penetration of the value segment. Snell  
effectively   suggests   that   the   merger   will   enable   Distell   to   engage   in  
predatory behaviour in the value segment cushioned by the returns on  
its dominant position in the proprietary and premium segments.  Snell  
argues   that   aided   by   its   predatory   capacity,   Distell   will   ultimately  
triumph in the value market, while simultaneously preventing Snell from  
moving   its   established   value   brands   into   the   proprietary   segment.  
Upon   successfully   driving   its   competitors,   notably   Snell,   out   of   the  
value segment, Distell will desert that low margin segment in favour of  
a committed presence in the prop and premium markets having further  
limited the prospect of new entry in those segments.
191. However, while we accept that aspects of this scenario are plausible,  
we do not believe that this portends a lessening of competition in the  
value   segment.     On   the   contrary,   it   will   raise   the   competitive  
temperature   in   that   segment.     Snell   has   already   demonstrated   an  
ability   to   raise   somewhat   the   prices   of   its   key   brands   in   the   value  
segment and this will be constrained by Distell’s entry into that market.

segment and this will be constrained by Distell’s entry into that market.  
We do not accept the argument that this will ultimately enable Distell to  
exercise market power in the value segment.   Snell remains a robust  
competitor   in   that   segment   and   even   if   Distell   enjoys   considerable  
success in the value market, low entry barriers, clearly established by  
Snell’s proven ability to establish itself there, will constrain any exercise  
of market power by Distell, just as it already constrains Snell. 
192. We should underline our view that entry barriers in this segment are  
low   principally   because,   in   contrast   with   the   other   two   market  
segments,  penetration of the  value  market does not presuppose the  
considerable expense and lead times required in brand building.
193. We should also note that even though we have ultimately elected to  
exclude wine from our relevant market definition in the value segment,  
we are persuaded that the unusually significant production of low value  
wine in South Africa, will nevertheless constrain pricing practices in the  
value   segment,   a   segment   where   price,   above   all,   determines  
consumer behaviour.  
Non­Confidential version 49

194. We accordingly find that the merger will not result in a substantial  
lessening of competition in the value spirits market.
The proprietary market
195. However,   as   already   indicated,   we   are   less   sanguine   regarding   the  
outlook for competition in the proprietary segment.
196. Distell’s post­merger market share in this sector is considerable.   We  
accept that Distillers enjoyed a considerable pre­merger market share  
of this segment.  However, the merger with SFW has added important  
brands to the Distillers’ armoury, notably the brandy brands, Martell 5  
Star and Mellow Wood 5, and Mainstay cane and vodka.
197. Crucially, an analysis of key competitive drivers in this segment, does  
not, in contrast with the value segment, ameliorate our concerns.  We  
refer in particular to our conclusions regarding barriers to entry and the  
dynamic features of the market.
198. Barriers   to   entry   in   the   proprietary   spirits   market   are   significant.     In  
particular,   successful   penetration   of   this   sector   clearly   requires  
considerable investment in brand building.   Snell’s concerns and the  
difficulties it has experienced in entering this market are clear evidence  
of this.  Distell’s concern, to keep all of its brands out the hands of its  
competitors   –   as   evidenced   by   its   resolute   defence   of   its   license   to  
distribute the Martell brand – is a further indication of the importance of  
brand recognition for entry into this segment.   Clearly, the owners of  
the agency brands fear that Distell would use its distribution licence to  
undermine the brand before it reverts to its owner. 
199. Note   also   that   in   this   segment   although   successful   branding   is,   in  
contrast   with   the   value   segment,   important   and   the   higher   margins  
available in this segment are testament to this, the margins appear to  
be significantly lower than in the premium segment.  The upshot of this

be significantly lower than in the premium segment.  The upshot of this  
is that for successful penetration of the proprietary segment not only is  
investment in branding important but high volumes are crucial as well.  
It   appears   that   a   premium   brand   earns   such   high   margins   that   it   is  
wholly possible to sustain market entry on the basis of very low sales  
volumes.  Where the proprietary brands are concerned it appears that  
not   only   must   the   brands   be   solid,   but   they   must   also   achieve  
considerable sales volumes.   This is graphically depicted in Annexure  
A. This twin requirement makes for very high barriers to entry indeed.   
200. We have also considered the dynamic features of this market.  In this  
respect, FABs are  the most significant  element in  our consideration.  
Both the pricing level and the ‘extrinsics’ associated with FABs, ensure  
Non­Confidential version 50

that   the   ‘demographics’   of   its   consumers   overlap   with   those   of   the  
consumers of proprietary spirits brands.  While we have declined, after  
substantial consideration, to include FABs in the relevant market, we  
nevertheless   do   believe   that   they   are   capable   of   acting   as   a  
considerable   influence   on   the   behaviour   of   participants   in   the  
proprietary spirits market.   Many of the strongest FABs are, after all,  
simply ‘ready­to­drink’ spirits, mixed spirits.  Others – for example, the  
cider FABs ­ are likely to act as something of a bridge between spirits  
and   beer,   while,   still   others   are   clearly   designed   to   appeal   to   wine  
consumers. FABs are thus significant not merely because certain of the  
products in this market may already be in direct competition with spirits  
but   because   others   may   ultimately   come   to   build   bridges   between  
spirits and wine, and, even, spirits and beer. Part of the extraordinary  
success of the FABs category is undoubtedly attributable to its ability to  
draw   new   consumers   –   women,   it   is   suggested   ­   into   the   alcoholic  
beverages market.  But it is equally clear that FABs draws a significant  
measure   of   its   support   from   the   ranks   of   spirits,   beer   and   wine  
consumers. 
201. Just   as   we   have   determined   that   inexpensive   wine   may   influence  
competitive behaviour in the value segment, so, are we satisfied, that  
FABs similarly influence the behaviour of participants in the proprietary  
segment.   However, far from ameliorating our concerns regarding the  
impact on competition, this factor serves to exacerbate our concerns.  
Distell enjoys a very powerful position in the FABs market ­ a market  
share of 70.8% in 2000 and 61.8% in 2001.  Had we elected to include  
FABs in the relevant market, Distell’s market share would have been  
considerably larger,  rising,  we estimate,  to 51,8%  of the prop  spirits

considerably larger,  rising,  we estimate,  to 51,8%  of the prop  spirits  
and FABs market.
202. We   accordingly   find   that   the   merger   will   substantially   lessen  
competition in the proprietary spirits market.
The premium market
203. Distell’s   post­merger   market   share   in   the   premium   spirits   market   is  
substantial   and   the   accretion   is   notably   larger   than   in   any   other  
segment.  Moreover, the relatively affluent consumers in this category  
are more likely to conform to the brand­, category­ and occasion­loyal  
customer   stereotype   identified   by   the   Commission   than   is   any   other  
group of consumers.  However, by the same token they are particularly  
sophisticated and well­informed consumers, characteristics that enable  
them to resist more easily attempts to exercise market power.
204. We   note   also   that   certain   of   the   key   Distell   brands   in   the   premium  
Non­Confidential version 51

market are agency brands. We refer to Martell VO brandy and Scottish  
Leader whisky. Martell VO is the most successful brand in the premium  
brandy stable, commanding a 24.2% share on its own, compared to  
Oude   Meester   with   23.2%.   Scottish   Leader   is   SFW’s   only   premium  
Scotch (with a premium whisky market share of 18.8%), which is added  
to   Distillers   modest   sales   of   Glenfiddich   (3.5%   of   premium   whisky  
sales), itself an agency brand. Distell’s licenses to distribute Martell VO  
and   Scottish   Leader   are   both   subject   to   sales   targets   and   periodic  
renegotiation.   This   not   only   renders   Distell’s   competitive   position  
vulnerable   to   the   multinational   owners   of   those   brands   should   they  
choose  to enter the South African market  directly  –  and the lengthy  
court   battles   over   both   the   key   brandy   and   gin   brands   that   Distell  
distributes is indicative of precisely this intention – but it also means  
that,   should   we   find   a   substantial   lessening   of   competition   in   this  
market, there is no readily available remedy.  Certainly, it would not be  
possible for us to order Distell to dispose of a brand that it does not  
own. 
205. Brands are all­powerful in the premium market and they are owned by  
powerful multinationals. Brands are, as already noted, also important in  
the   proprietary   market   but   here   South   African   brands   –   notably,  
although  not  exclusively,  brand   brands  –  are   a  significant   presence.  
However,   in   the   premium   market   there   is,   to   our   knowledge,   not   a  
single South African brand of note. Even a company with the resources  
commanded by Distell would have difficulty developing a new premium  
brand   of   whisky   or   brandy   or   any   of   the   white   spirits.     Accordingly  
Distell is only likely to advance in this market through the medium of

Distell is only likely to advance in this market through the medium of  
additional agency agreements.  This, as we have already indicated, is  
only   feasible   if   the   large   multinational   brand   owners   do   not   wish   to  
exploit the South African market directly.  
206. Moreover,   even   if   the   multinational   brand   owners   prefer   an   agency  
arrangement with a South African distributor, Distell is, by no means,  
the only effective distribution agent available in this country and there is  
no   reason   to   expect   that   a   large   multinational   would   favour   Distell,  
potentially a competitor in the international market, over other efficient  
South African distributors such as Snell or DGB.  The relationship that  
has developed between Snell and Brown and Forman, the US based  
owner of, inter alia, Jack Daniels, the powerful premium whiskey brand,  
is evidence of this. The existence of these effective local distribution  
channels   coupled   with   the   importance   of   well­established   brands  
means that entry into the premium market is relatively easy.
207. We accordingly find that the transaction is not likely to cause a  
substantial   lessening   of   competition   in   the   premium   spirits  
market.
Non­Confidential version 52

208. In  summary  we have determined  that the transaction  will   result  in a  
substantial   lessening   of   competition   in   the   proprietary   spirits   market  
only.
Countervailing Efficiency Gains
209. We   are,   in   the   event   of  an  adverse  finding  on  competition  grounds,  
obliged to examine whether there are any efficiency gains arising from  
the merger that would outweigh  the negative impact on  competition.  
The   parties   have   presented   us   with   a   massive   quantity   of   evidence  
regarding   efficiencies.     The   Commission   has   disputed   much   of   this  
evidence   –   in   particular   it   has   argued   that   many   of   the   efficiencies  
claimed could have been achieved without merging the two entities.
210. However, we do not believe that a finding on efficiencies is necessary  
in   this   instance.     In   our   view,   the   efficiency   defence   is   intended   to  
resurrect an anti­competitive merger that, absent the claimed efficiency  
gains, may fall to be prohibited.  This is clearly not contemplated in this  
instance.  We have found a substantial lessening of competition in one  
of   the   many   markets   implicated   in   this   transaction.     Any   potential  
remedy imposed, whether behavioural or structural, will be confined to  
that market compromised by this transaction.   It is well nigh impossible  
to   isolate   the   efficiency   gains   that   refer   to   this   market   specifically,  
although, suffice to say that we are firmly of the view that none of the  
efficiencies claimed will be sacrificed by this finding or by any of the  
remedies   that   flow   from   it.     Accordingly,   this   will   not   be   examined  
further.
Public Interest
Introduction 
211. In this case both the merging parties and the unions have relied on the  
public   interest   grounds   set   out   in   the   Act   in   section   12   A   (3).   That  
section states:

section states: 
When   determining   whether   a   merger   can   or   cannot   be   justified   on  
public   interest   grounds,   the   Competition   Commission   or   the  
Competition Tribunal must consider the effect that the merger will have  
on – 
a) a particular industrial sector or region;
b) employment;
c) the   ability   of   small   businesses,   or   firms   controlled   or   owned   by  
Non­Confidential version 53

historically disadvantaged persons, to become competitive; and
d) the ability of national industries to compete in international markets.
    
212. The parties rely on subparagraphs (a) and (d) of subsection (3) ­ they  
state   in   broad   terms   that   the   merger   creates   an   internationally  
competitive   firm   and,   ironically,   relying   on   what   some   of   their  
competitors have stated during the hearing, and a submission from the  
DTI, that the merged firm will be good for the industry as a whole. 29 
This   they   assert   justifies   the   merger   on   substantial   public   interest  
grounds.
213. The unions rely on subparagraph (b) of subsection (3) and assert that  
job losses occasioned by the merger are so great that they will have an  
adverse effect on employment and hence this justifies its prohibition on  
public interest grounds.
214. Thus   the   public   interest   asserted   pulls   us   in   opposing   directions.  
However, the legislation expressly allows for this. Just as the legislation  
may   allow   the   public   interest   to   resurrect   a   merger   that   will   harm  
competition   it   also   contemplates   a   situation   where   a   public   interest  
ground may justify the prohibition of a merger even if a merger does  
not have an anti­competitive effect. 30 
215. All   of   this   presupposes   that   we   have   determined   what   the   public  
interest is and that it is not, to borrow a phrase from Robert Bork, a  
“policy at war with itself”.
216. In   this   case,   multiple   public   interest   grounds   are   implicated   which  
suggest contradictory outcomes. 
217. How do we approach this contradiction? In the first place each public  
interest ground asserted, should be viewed in isolation to see if it is  
29  Paras 359­360 and 363, parties’ written heads of argument. “Mirroring the view of Mr. Hutchinson,  
Edward Snell & Co Ltd (‘Snell’) stated in its submission to the Commission  that the merger was not

itself a bad thing and that: ‘it could be deemed highly desirable for our country and for the wine and  
spirits industry as a whole, to have a local, globally competitive, profitable and visionary business  
leading the industry.’ Mr. Hooper, the managing director of Snell repeated this view at the Tribunal  
hearing. He stated that: ‘for the South African liquor industry to have a globally competitive business  
heading up the industry, given the aggregations that have taken place internationally, in itself is not a  
bad thing.’  Most significantly, the Department of Trade and Industry (‘DTI’) (which may, in terms of  
section 18(1) of the Act make representations on any public interest ground referred to in Section 12A  
(3)) has also endorsed the role of Distell as charting the way for South African alcoholic beverages in  
the international market. In its submission dated 21 May 2002, the DTI supported the unconditional  
approval of the merger and stated the following in its recommendations: ‘From an industry and  
enterprise development point of view, it is noted that the merger will substantially improve the  
international market presence of the South African wine and spirits industry’.
30  This is the effect of the word “otherwise” in Section 12A(1)(b) when read with 12A (3)(b).
Non­Confidential version 54

substantial. Assuming that the answer to that is in the affirmative and  
that   more   than   one   ground   exists   which   contradicts   the   other/s   we  
need to see, first, if they can be reconciled and, if not, we must then  
balance them and come to a net conclusion. 
218. The language of subsection 12A(1)(b) states:
“..otherwise, determine whether the merger can or cannot be justified  
on substantial public interest grounds by assessing the factors set out  
in subsection (3).”
219. This   language   is   consistent   with   both   a   balancing   approach   and   a  
reconciliation   approach   and   we   therefore   need   not   choose   between  
either.
220. In most cases, however, we suspect that balancing will be unnecessary  
as is illustrated by the facts of this case.
221. This is because a contradiction can be of two types. One is when the  
opposite   interests   collide   head   on,   like   two   medieval   knights   in   a  
jousting contest where only one can emerge as the victor and continue  
his course. This is where we are faced with a stark choice of whether to  
prohibit or to approve.
222. The other situation is where the opposite interests avoid one another  
like two vehicles bypassing each other in opposite directions on a dual  
lane   road.   They   pass   one   another   without   affecting   their   respective  
opposite courses.
223. In   this   case,   unless   we   find   that   the   employment   grounds   justify  
prohibition, the unions’ public interest and those of the parties do not  
collide ­ hence we have no jousting contest and the opposing interests  
can travel past one another on the road in their separate lanes.
224. We will examine the employment claims first.
225. Since   this   transaction   was   first   made   known   to   them,   the   unions,  
FAWU and NUFBWSAWU,   have been concerned about the numbers  
of   jobs   that   would   be   lost.   The   unions   were,   for   this   reason,   active

of   jobs   that   would   be   lost.   The   unions   were,   for   this   reason,   active  
participants   in   the   litigation   that   took   place   over   notification   that   we  
referred   to   earlier   in   the   background   section.   The   unions   also  
intervened in these present proceedings and, with the consent of the  
merging parties, were present throughout the hearing, put questions to  
certain witness and made legal submissions.
226. The parties’ evidence is that the actual employment effects are minimal  
Non­Confidential version 55

and offer the following statistics.
227. Prior to the merger, as at 30 June 2000, the total number of employees  
at   both   firms   was   5828.Once   final   production   decisions   had   been  
implemented, which would determine the extent of job losses, 1 882  
jobs were lost. Of these only 1414 represented actual terminations, the  
balance representing either jobs that were frozen or not filled once they  
became   vacant   or   whose   occupants   were   re­assigned   to   other  
positions in the Distell group. Thus according to Distell the effective job  
loss figure is 1 414 and that included in this figure are 630 people who  
accepted   voluntary   retirement   and   621   who   accepted   voluntary  
retrenchment. This left a remainder of 164 people who were forcibly  
retrenched.
228. Distell   goes   on   further   to   state   that   only   16   of   the   164   employees  
forcibly   retrenched,   challenged   this   action   and   declared   a   dispute.  
Some cases have been settled, but to date those that have proceeded  
for determination have all been resolved in favour of Distell.
229. Distell  also details  a lengthy process of negotiation and consultation  
with   unions   over   the   retrenchments.   It   states   that   the   retrenchment  
package, which included payment of four weeks pay for every year of  
service, exceeded the minimum requirement of the law (one week) and  
the previous policy of both companies (two weeks). The unions insisted  
on six weeks.
230. Given   this   factual   background,   Distell   argues   that   once   one   has  
properly analysed the effects of the terminations one should only take  
into   account   the   164   forced   job   losses,   and   that   this   number   in   the  
context of the overall merger, does not constitute a substantial public  
interest ground for prohibiting the merger.
231. The unions did not dispute Distell’s figures or its claim to have followed  
the procedures outlined above, but they are less sanguine about the

the procedures outlined above, but they are less sanguine about the  
prospects of those who took voluntary packages and whom Distell has  
accordingly removed from the ranks of those adversely affected by the  
merger.   Many   employees,   argue   the   unions,   took   the   voluntary  
package because they believed they had no other choice or needed  
the   money   in   the   short   term.     Accordingly,   they   should   still   be  
considered as having been adversely affected by the transaction. With  
the   inclusion   of   this   class   of   employee   the   numbers   of   employees  
adversely   affected   by   the   merger   is   significantly   increased   and   the  
impact on the public interest concomitantly larger. 31
31  However,  Distell take this analysis, whose point of departure is to look at the real motive of  
the employee exercising the choice, and turns it on its head. Distell argues that many  
employees offered retrenchment refused initially as a hold­out strategy as they knew they  
would continue to be paid whilst their dispute was resolved and that they would still collect the  
Non­Confidential version 56

232. We have in previous decisions indicated that we do not exercise our  
public   interest   determinations   in   a   void. 32  Parliament   has   in   many  
instances   enacted   legislation   that   deals   quite   specifically   with   the  
issues referred to in section 12 A (3) and employment is no exception.  
Indeed we observed in the  Shell/Tepco case that in many respects our  
jurisdiction in these areas is secondary, as these other statutes and the  
institutions that they create, are better placed and resourced to deal  
directly and effectively with these issues than are we, given that our  
discretion   is   described   in   section   12   (A)   (3)   at   a   high   level   of  
abstraction and generality. 33 
233. In this case the unions are asking us to consider prohibiting the merger  
on   public   interest   grounds   and,   alternatively,   to   consider   a   more  
generous package to those employees who are to be retrenched. They  
further argue that if we impose a divestment remedy that it be made a  
condition   of   such   divestiture   that   the   affected   employees   are  
transferred with the assets.
234. Let us take the alternative request first. If we are to second guess the  
fairness of the retrenchments and the appropriateness of the package  
offered we would be doing precisely what we have previously said we  
should not – acting as a second­ hand regulator, post facto, when the  
appropriate   body,   on   the   evidence   before   us,   has   already   made  
determinations and was in a better position to do so.
235. The first request, that is, that the merger be prohibited because it has  
an   adverse   effect   on   employment   amounting   to   a   substantial   public  
interest ground is more difficult to evaluate.
236. Beyond  requiring  that  public interest  grounds be  “ substantial”  before  
they   qualify   for   assessment,   the   legislation   offers   no   criteria   as   a

they   qualify   for   assessment,   the   legislation   offers   no   criteria   as   a  
yardstick   for   their   evaluation,   unlike   with   the   competition   evaluation,  
where criteria are enumerated in section 12A(2).
237. Our previous approach, although not expressly articulated in this way,  
has been to suggest that, where there are other appropriate legislative  
instruments   to   redress   the   public   interest,   we   must   be   cognisant   of  
them in determining what is left for us to do before we can consider  
whether   the   residual  public   interest,   that   is   that   part   of   the   public  
same package that those who had accepted voluntary retrenchment had got, even if it was  
resolved against them. On this approach the number adversely affected may even be less  
than 164.
32See Unilever Plc and other and Robertson’s Foods Pty Ltd and others (CT 55/LM/Sep01 4 April  
2002 at paragraph 43 and Shell South Africa (Pty)Ltd and Tepco Petroleum (Pty)Ltd (CT 66/LM/Oct  
01 22 February 2002) at paragraph 58.
33  Large merger between Shell & Tepco, Competition Tribunal decision: 66/LM/Oct01.
Non­Confidential version 57

interest not susceptible to or better able to be dealt with under another  
law, is substantial.
238. Thus in the case where the public interest asserted is employment, if it  
could be demonstrated that the merger specific employment effects are  
so adverse and that no other law or regulator can remedy them, then  
we would be obliged to intervene to either prohibit or set conditions on  
an approval.
239. In   this   case   we   are   advised   that,   at   worst,   the   merger   specific  
employment losses are 1 414 jobs, if we accept the unions argument  
that all voluntary retirements and retrenchments should be counted. On  
the   other   hand   if   Distell   is   right   and   only   forced   retrenchments   are  
counted the employment effect may be less than 164.
240. How many jobs must be lost before one has grounds for a substantial  
public interest? The legislature wisely does not seek to answer that for  
us, nor can we assume that it should be a uniform figure for all mergers  
­ it would depend on the context. In the context of this merger Distell  
maintains that the total jobs lost ­ 164 on its approach ­ is less than 3 %  
of the aggregate pre­merger work force. 34
241. If, however we take the total number of jobs lost, forced and voluntary,  
then that figure becomes 24%  ­ certainly significant.
242. However, we know that even if this figure is significant the test is not  
the number of jobs lost but the substantial effect on employment. That  
means that the effect can be ameliorated by a retrenchment package  
and   not   simply   job   retention.   The   package   offered   by   Distell   seems  
reasonable measured against the requirements of the law and its own  
past   practice.   The   fact   that   only   16   employees   disputed   the   final  
package,   out   of   the   1414   terminated,   suggests   that   by   and   large  
employees felt the same way. In our view, although the loss of jobs is  
significant, the overwhelming acceptance of voluntary retirement and

significant, the overwhelming acceptance of voluntary retirement and  
retrenchment   packages   has   lessened   the   adverse   effects   and   we  
cannot   conclude   that   on   balance   the   final   effect   on   employment  
constitutes   a   substantial   public   interest   ground   that   would   justify  
prohibition of the merger. 
243. The   fact   that   the   reasonableness   of   the   package   was   a   subject   of  
protracted   negotiation   and   that   when   opposed   by   some   employees,  
subsequent adjudication by a specialist body has, on all occasions to  
date, found in favour of the company, suggests that there are also no  
grounds for us to impose, in the public interest, any further condition on  
34  Distell state that even if the merger had not proceeded there would have been job losses as SFW  
was planning to downsize its distribution and sales functions, although they given no figures for this.  
(See C 280)
Non­Confidential version 58

the company.
244. The remaining question of whether we should insist that employment  
obligations   should   follow   any   divestment   will   be   deferred   to   our  
remedies hearing.
Conclusion on Public Interest
245. Here   there   are   potentially   two   stages   of   balancing.   First   within   the  
public interest category, recall the choice of approaches between the  
joust versus the bypass. The second depends on whether we come to  
the conclusion that a substantial public interest has emerged in which  
case we must decide whether that would alter our conclusion on the  
competition grounds (Section 12 A (2)).
246. In respect of the first stage, it is not necessary for us, given our finding  
on   the   employment   issues,   to   consider   the   parties’   public   interest  
claims   as   we   are   not   faced   with   the   need   to   balance   or   reconcile  
multiple public interest claims – that is, the employment consequences  
of the transaction have not led to an adverse finding on public interest  
grounds   and,   so,   do   not   have   to   be   balanced   against   any   positive  
impact that may be found on the parties’ public interest claim.
247. Moreover, we do not need to evaluate the parties public interest claim  
because,   even   if   it   did   qualify   for   recognition,   it   would   not   alter   the  
outcome of our finding on the competitive criteria ­ we have determined  
that  the merger  should  not  be  prohibited on competition criteria and  
hence   considerations   that,   like   the   public   interest   claim,   go   to  
resurrection are not pertinent. 
248. Note   also   that   the   public   interest   benefits   to   the   industry   and   the  
country   that   the   parties   claim   are   not   negated   by   any   conditional  
remedies that might be imposed in the affected markets, and, hence,  
no evaluation of them is needed if no balancing is required between the

no evaluation of them is needed if no balancing is required between the  
competitive   concerns   and   the   public   interest   asserted.   The   same  
justification for our not evaluating the alleged efficiency gains is equally  
applicable here.
Finding
249. We   find   that   the   transaction   will   result   in   a   substantial   lessening   of  
competition in the proprietary spirits market.  For the reasons outlined  
Non­Confidential version 59

above,   this   finding   is   not   affected   by   considerations   of   efficiency   or  
public interest.
                                                  19   March   2003
D. Lewis                 Date
Concurring: M Holden and N Manoim
Non­Confidential version 60

Non­Confidential version
Annexure  A
61