Governing through Quality: Conventions and Supply Relations in the Value Chain for South African Wine

Abstract

Global value chain (GVC) analysis examines the dynamics of economic globalisation and international trade. The concept of GVC governance illustrates how ‘lead firms’ achieve certain functional divisions of labour along a value chain – resulting in specific allocations of resources and distributions of gains. In this article I argue that agri-food lead firms do not govern chains solely on the basis of buyer power, market share, and/or economies of scale or scope but also through normative work. In order to do so, I apply convention theory to the analysis of governance in the value chain for South African wine. I analyse how wine quality conventions applied in the UK are translated in South Africa into specific functional divisions of labour and supply relations, themselves underpinned by local configurations of quality conventions. The case study of wine suggests that lead firms are able to drive a value chain only when industrial and market conventions are dominant, such as in basic quality wine. These conventions are more portable and thus easier to transmit at a distance. Where other, less portable, conventions are more important in discovering quality, as in mid-range and top quality wines, the value chain is much more fragmented and less driven.

Full text

1
Governing through Quality: Conventions and Supply Relations
in the Value Chain for South African Wine
Stefano Ponte
Abstract
Global value chain (GVC) analysis examines the dynamics of economic globalisation and
international trade. The concept of GVC governance illustrates how ‘lead firms’ achieve
certain functional divisions of labour along a value chain – resulting in specific allocations of
resources and distributions of gains. In this article I argue that agro-food lead firms do not
govern chains solely on the basis of buyer power, market share, and/or economies of scale or
scope but also through normative work. In order to do so, I apply convention theory to the
analysis of governance in the value chain for South African wine. I analyse how wine quality
conventions applied in the UK are translated in South Africa into specific functional
divisions of labour and supply relations, themselves underpinned by local configurations of
quality conventions. The case study of wine suggests that lead firms are able to drive a value
chain only when industrial and market conventions are dominant, such as in basic quality
wine. These conventions are more portable and thus easier to transmit at a distance. Where
other, less portable, conventions are more important in discovering quality, as in mid-range
and top quality wines, the value chain is much more fragmented and less driven.
Introduction
Global value chain (GVC) analysis has emerged since the early 1990s as a novel methodological
tool for understanding the dynamics of economic globalisation and international trade. It is
based on the analysis of discrete value chains where input supply, production, trade and
consumption or disposal are explicitly and (at least to some extent) coherently linked. In
addition to the descriptive aspects of territoriality and input–output structure, much GVC
discussion has revolved around two analytical issues: how GVC are internally governed (in the
context of a larger institutional framework) and how enterprise-level upgrading or downgrading
takes place. In the agri-food sector these discussions have often been carried out with an interest
in how power and rewards are embodied and distributed, what entry barriers characterise GVCs
and how unequal distributions of rewards can be challenged in favour of labour, small producers
and/or developing countries.
This is the accepted version of the following article: Ponte, S. (2009)
“Governing through Quality: Conventions and Supply Relations in the Value
Chain for South African Wine,” Sociologia Ruralis, Vol. 39, No. 3, pp. 236-
257. DOI: 10.1111/j.1467-9523.2009.00484.x
It has been published in final form at:
http://onlinelibrary.wiley.com/doi/10.1111/j.1467-9523.2009.00484.x/abstract

2
In the literature GVC governance has been defined and operationalised in different ways.
The original approach to governance developed by Gereffi (1994) sees it as the process of
organising activities with the purpose of achieving a certain functional division of labour along
a value chain – resulting in specific allocations of resources and distributions of gains. Here, a
group of ‘lead firms’ drive a value chain through specific mechanisms that are related to the
nature of entry barriers and core competences. This interpretation of governance as drivenness
involves lead firms defining the terms of chain membership, the incorporation/exclusion of
other actors accordingly and the re-allocation of value-adding activities (Gereffi 1994; Ponte
and Gibbon 2005; Raikes et al. 2000; Wilkinson 2006). A second and more recent approach has
focused on unpacking how co-ordination takes place at individual nodes along a GVC (Gereffi
et al. 2005; Altenburg 2006), an approached termed elsewhere governance as coordination
(Gibbon, Bair and Ponte 2008).
In these two interpretations of governance the literature suggests that lead firms do not
necessarily control value chains through hands-on mechanisms (such as vertical integration or
highly socialised or supervised relations). Increasingly, they do so in hands-off ways by
providing specifications to their immediate suppliers (or buyers) who in turn transmit and
translate them further along the value chain (Ponte and Gibbon 2005; Sturgeon 2002). A key
factor in governing in hands-off ways is for lead firms to succeed in spreading specific quality
conventions and operational procedures of their liking. In other words, lead firms obtain or keep
their position in a GVC not only as a result of their dominant market share, superior innovation,
and/or substantial economies of scale and scope, but also through ‘normative work’.
On the basis of such observations, a third approach has emerged, that of governance as
normalisation (Gibbon, Bair and Ponte 2008). This approach underscores the discursive
dimension of the framing of buyer–supplier relations, based on a constructivist approach to the
knowledge content of transaction and the capacities of suppliers. The term ‘normalisation’ is
used not to mean ‘making things normal’ in the vernacular sense of unexceptional, but refers to
a project of re-aligning a given practice so that it mirrors or materialises a standard or norm.
Early work taking this approach (Ponte 2002; Daviron and Ponte 2005; Ponte and Gibbon 2005)
borrowed from convention theory in situating and analysing the dynamics of buyer–seller
relations in wider normative contexts. This work used convention theory to elaborate an account
both of the immediate normative environment in which value chain actors operate (that is, in
relation to their functional statuses as buyers or suppliers) and the broader normative
frameworks influencing the designations attached to the products and services they exchange
within GVCs. Such contexts provide vocabularies for describing, and prescriptions concerning,
what actions buyers should take when governing a value chain and what specific qualities
suppliers should aim for and how they should secure them.

3
This article furthers the analytical project of GVC ‘governance as normalisation’ by
examining the value chain for South African wine. To the author’s knowledge, few attempts
have been made in the literature to analyse the governance of the GVC for wine. Often, this has
been done in terms of how a specific producing country is inserted in the GVC (for Chile and
New Zealand, see Gwynne 2006, 2008; for New Zealand, see Lewis 2008; Hayward and Lewis
2008). Others have used different political economy approaches to analyse wine market
restructuring (see, among others, Pritchard 1999; Shaw 2001). Finally, other contributions have
examined specific aspects of value chain dynamics (such as learning and innovation) through a
convention theory approach (see Guthey 2008 on the northern California wine industry) but
have not linked them explicitly to governance.
In the rest of this article, first I provide a concise analysis of quality conventions and supply
relations in the value chain for wine ending in the UK (the main destination of South African
wine). Second, I scrutinise how South Africa-based wine actors transmit upstream (towards
producers) demands that are placed on them by lead firms, and how this transmission process
leads to specific functional divisions of labour and supply relations in South Africa, themselves
underpinned by local configurations of quality conventions. In both instances much of the
specific empirical evidence underpinning the analysis is available elsewhere (Ponte 2007; Ponte
and Ewert 2007).
The wine GVC is particularly interesting in this context because, in the agri-food sector, it
has the most complex and sophisticated quality infrastructure. Furthermore, it is going through a
major process of restructuring in which the battle lines are drawn along the application,
challenge and re-interpretation of different quality conventions. Recent trends in the geography
of wine production, trade and consumption, and the changes in the quality composition of
supply and demand, have been well documented elsewhere (see, among many others, Anderson
2004b; Spahni 2000; Unwin 1996). These include, in the last few decades, a dramatic fall in
production volumes and per capita consumption in traditional wine-making and -consuming
countries, such as Portugal, Spain, France and Italy. This has been partly compensated, on the
one hand, by growing production and exports in New World producing countries (Argentina,
Chile, South Africa, New Zealand, Australia and the USA) and, on the other hand, by increasing
consumption in the UK and the USA. What for centuries was considered a cottage industry is
becoming characterised by the presence of large multinational companies (Anderson 2004a). At
the same time, the level of concentration in production/processing in the industry is far behind
other sectors in the agri-food sector such as coffee, cocoa or beer. Wine retail, traditionally the
domain of small specialist shops, is now in the hands of supermarket chains, especially in
northern Europe and the UK, but increasingly in southern Europe as well. Although there are
fears of the homogenisation of styles and offerings in the wine market, this is still an industry

4
that produces a phenomenal array of different products, which are sold under a combination of
brand names, grape variety and/or geographical indications of origin.
Finally, South Africa is instructive of the normative aspects of value chain governance for
two reasons: (1) its wine sector, due to sanctions against apartheid, was insulated from the
international market for a long period; with the opening of the South African export market in
1994 the long-simmering battles that had taken place internationally on ‘what wine quality is’
unfolded in the country in a short time; (2) South Africa has positioned itself as neither Old
World nor New World style wine producer, effectively trying to find its own normative space
and profile.
GVCs and conventions
Conventions are generally defined as a broad group of mutual expectations that include – but
are not limited to – institutions. For convention theory,1 rules are not decided prior to action, but
emerge in the process of actions aimed at solving problems of co-ordination. At the same time,
action may be tested and thus needs to be justified by drawing on a variety of criteria of justice
that are broadly accepted at a particular time. In other words, convention theory links situated
action to widely accepted normative models (Borghi and Vitale 2006). Conventions are not
fixed in time and space: they include mechanisms of clarification that are themselves open to
challenge. They are both guides for action and collective systems to legitimise those actions that
can be submitted to testing and discussion, leading to compromises and possibly defeat
(Boltanski and Thévenot 1991; Eymard-Duvernay 2006a and b; Favereau and Lazega 2002;
Ponte and Gibbon 2005; Wilkinson 1997).
Boltanski and Thévenot (1991) developed six historically based ‘worlds of legitimate
common welfare’ that draw on particular paradigms of moral philosophy. They elaborate an
account of how these worlds are embedded in the behaviour of firms on the basis of organising
principles. To these ‘worlds’ correspond different norms of qualification, of people (e.g.,
employees) and objects. Depending on what justifications are employed, one can arrive at
different conventions for organising the activities of firms (see also Boltanski and Chiapello
1999). According to Boltanski and Thévenot:
The inspirational world rests on the principle of common humanity and non-exclusion
(based on Augustine) and agreement about evaluation and action refers to grace and divine
inspiration (in firm parlance, creativity).
The domestic world is founded on the principle of dignity (Bossuet) and agreement is
founded on the basis of tradition (firms draw on the concept of loyalty).
The opinion-based world is structured around the principle of difference (Hobbes) and
objects and subjects are evaluated through the opinion of others (firms use the concept of
reputation).

5
The civic world finds its justification in the notion of common welfare (Rousseau) and
agreement is founded on the fact that individuals are sensitive to changes in common welfare
(firms are organised around the concept of representation).
The market world is based on the notion that difference is justified by sacrifice, effort or
investment (Smith), and agreement is found on the basis of market principles such as price
(firms organise themselves around the concept of competitiveness).
Finally, the industrial world is based on the existence of ‘orders of greatness’ (St Simon) and
agreement is based on objective (technical and measurable) data (firms invoke the concept of
productivity).
Convention theory does not place a hierarchical value to these worlds, nor does it portray
them as historical inevitabilities. Furthermore, at any particular time and locality multiple
justifications of action may be operating at the same time. Finally, although there is an internal
coherence in each world there are also qualifications that bridge different worlds.
On the basis of the Boltanski and Thévenot framework, Eymard-Duvernay (1989; see also
Sylvander 1995; Thévenot 1995) developed a typology of quality conventions and related forms
of coordination. Eymard-Duvernay’s main point of departure is that price is the main
management form of a particular market only if there is no uncertainty about quality. If this is
the case, differences in price are equated with quality. This characterises a market quality
convention.2 When price alone cannot evaluate quality, economic actors adopt other
conventions to solve uncertainty about quality. In a domestic convention this is solved through
trust (long-term relationships between actors or use of private brands that publicise the quality
reputation of products). In this case the definition of quality is resolved ‘internally’, and the
identity of a product is guaranteed or institutionalised in the repetition of history by its region or
country of origin or by a brand-name. In an industrial convention uncertainty about quality is
solved through the actions of an external party that determines common norms or standards and
enforces them via instrument-based testing, inspection and certification. More recently, an
additional category has been added, civic convention, where there is a collective commitment to
welfare, and the quality of a product is related to its impact upon society or the environment.
In this article, to these quality conventions, two others are added to complete the original
Boltanski and Thévenot framework: an opinion convention where uncertainty about quality is
resolved through the personal judgment (rather than objective measurement) of an actor that is
external to the exchange and has a ‘good reputation’; and an inspirational convention where the
personality of one of the actors in the exchange, his/her genius, intuition, creativity, vision or
downright weirdness substitutes for other means of assessing quality. Table 1 lists the
modalities for assessing quality that are linked to these conventions in relation to wine and
specifies the actual instruments of verification. These will be explored in detail in following
sections.

6
TABLE 1 ABOUT HERE
In recent Anglophone literature convention theory has been used chiefly as a source for
typologies of various dimensions of product quality.3 This literature has been mainly concerned
with classifying new dimensions of product quality, reflecting struggles over the content of
civic conventions; and judging these struggles as for or against a contemporary ‘capitalist’
project. This capitalist project is said to be striving not only to ‘commodify’ new product
qualities, but also to commodify information about them – what Freidberg (2003, 2004) terms
‘double fetishism’. These dynamics are said to lead to a hollowing out of organic, fair trade, and
other sustainability labels (Raynolds 2000, 2002, 2004; Barham 2002; Renard 2003).
Alternatively, convention theory has been used to highlight the putative emergence of
‘alternative food networks’ based on locality and domestic conventions (Murdoch and Miele
1999; Murdoch et al. 2000).
Until recently convention theory had not specified the boundaries and overlaps of
conventions across different levels of generality and spheres of action. This led to some degree
of confusion on the applicability of conventions in terms of reach (and degrees of power) and of
geographical or organisational extension. Distinctions between conventions of quality (Eymard-
Duvernay 1989) and wider normative models of organisational ‘best practice’ (Boltanski and
Chiapello 1999) were drawn only implicitly. This has now been partly clarified by some of the
key convention theorists in a programmatic paper presented at the 2003 Paris conference on
‘Conventions et institutions’ (Eymard-Duvernay et al. 2006), where the authors extend the
concept of a horizontal plurality of forms of co-ordination to a vertical plurality that allows the
explanation of ‘more localised’ spheres of action (but do not mention ‘more generalised’ ones).
This approach, according to its promoters, incorporates variation in levels of generality and
justification within conventions. In other words, the plurality of possible worlds of justification
of action drawn upon within certain spheres coexists with more localised and familiar
understandings of what is just.
In this article, convention theory is employed to unpack the dynamics of governance in
GVCs, but instead of explicitly discriminating between broader and more proximate levels of
normative influence in the same sphere of action, focus is placed on: (1) how quality
conventions that apply to a particular segment4 of a value chain (the UK end-market of the wine
value chain) shape its governance; and (2) how these are ‘translated’ upstream into specific
supply relations and functional divisions of labour in the South African segment of the wine
chain.
Space limitations militate against examining the whole cycle of formation, challenge, test
and reconfiguration of quality conventions in the wine value chain and the overlapping of
values, practices and institutions brought into play by an array of actors throughout this cycle.
The more modest focus is instead on instruments of verification, which make visible and

7
translate the complex and hidden negotiations, interactions and representations that lay behind
the crystallisation of a convention (or the overlap of conventions) at a particular time and
position in the value chain. Verification (or testing) is the moment when justifications are
rendered explicit, where socio-technical devices are pulled in and different kinds of knowledge
and expertise recruited. The aim is thus to show how certain quality conventions and attached
instruments of verification translate into supply relations and divisions of labour that are
employed to drive value chains in particular ways.
Quality conventions and governance in the GVC for wine ending in the UK
This section provides a concise picture of the overlap and tensions between the different quality
conventions that operate in the wine GVC ending in the UK, followed by a short discussion of
how these conventions influence GVC governance. This picture arises from an analysis of
interviews carried out by the author with South Africa wine operators (and subsidiaries of UK
importers and marketers based in South Africa) on perceptions of how buyers in the UK
measure the quality of wine. This was supplemented by an interpretive reading of the discourses
on quality in wine trade and consumer-oriented magazines (see methodological appendix for
details). What is provided here is a simplified, snapshot picture, which does not do justice to the
array of enrolments of expertise and knowledge, or to the complex combination of advertising,
marketing and technical (winemaking and viticultural) tools that are behind measuring quality
and justifying one quality convention or another.
The overlap of quality conventions in the value chain for wine ending in the UK is
represented in Figure 1 in the form of a quality pyramid of wine. The pyramid is divided into
three perceived quality brackets for top quality, mid-range and basic wines. Inside the pyramid,
for each bracket, the main instruments of verification of quality are noted. To each of these
correspond one or a combination of quality conventions, indicated on the left side of the
pyramid. Each convention is mentioned in order of descending importance in each quality
bracket. It is worth remembering that different conventions almost always overlap, their
configurations change in time, and that they are always contested, with the possible exception of
those lubricating the basic wine market (at least in the UK).
FIGURE 1 ABOUT HERE
Top wines
Both the trade literature and interview data suggest that, for top quality wines, one of the main
instruments of verification of quality is endorsement by a respected (or renowned) wine writer,
judge or publication. Following the original Boltanski and Thévenot (1991) framework, this
translates into a quality convention based on opinion. So far, the quality convention literature
has mobilised only some of Boltanski and Thévenot’s worlds and not others. Wine provides the
opportunity to complete their palette of conventions. In an opinion convention, the judgment of
quality rests on the aesthetic approach of the endorser towards wine, the perceived

8
independence of the endorsers from industry interests and their own preferences. Main factors
for top wines are ratings by US wine writer Robert Parker, and by influential publications such
as Wine Spectator and Decanter. When location or terroir are at play (in domestic conventions,
see below) in top quality wine, such parameters are also mediated by ‘wine experts’ – given that
even knowledgeable consumers need to be guided through the great diversity of wines available
in the market. A brand alone will not be able to measure quality either in this quality bracket.
Which experts, and in what markets, come to be recognised as being of ‘worthy opinion’ is
the topic of entire books (see McCoy 2005). Also, opinion makers will not have an impact on
quality unless producers make wines that appeal to their palate. Therefore, the rise and
consolidation of one particular aesthetic of wine instead of another will also depend on those
actors, for example, flying winemakers such as Michel Rolland, who can deliver portable
solutions to winemakers and viticulturists around the world and do the hidden (some say ‘dirty’)
work of shaping an opinion convention into real world practices (on flying winemakers, see
Lagendjik 2004).
Terroir, the specific combination of soils and microclimate in a particular vineyard or
property, is another important factor in the quality evaluation of top wines, but less so in
English-speaking markets than in continental Europe. Terroir, as the French insist, relates to a
domestic convention, where intimate knowledge of the land, and long-term and repetitive fine-
tuning of practices and varieties embed into the wine the natural elements of land and climate.
Such a convention was for a long time dominant in assessing top quality wine in Old World
producer countries, but it has been challenged by increased consumption in English-speaking
countries such as the UK and the USA, and by the rise of ‘opinion kings and queens’ (such as
Robert Parker in the USA and Jancis Robinson in the UK). Quality as terroir is built upon the
idea that good wine can come only from specific locations and only after centuries of
experimentation and close knowledge of the land. Viticulture and winemaking skills are
directed to ‘letting the land’ speak – the best winemaking in this tradition is one that is as non-
interventionist as possible. The still unresolved conflict between a domestic quality convention
and one rooted in opinion has perhaps been the defining feature of the normative power struggle
in the top quality wine industry over the last few decades, with signs that the opinion convention
is becoming more dominant.
Finally, some wines are reputed to be of top quality because of the unique ‘personality’ of
either the wine, its presentation, the winemaker, or the property behind the wine – this includes
weird, ‘visionary’ offerings, ‘mad winemakers’, alternative labelling or a specific story behind
any of these. Here is another convention that had been disregarded in the quality convention
literature, even though it has been given a prominent treatment in the original Boltanski and
Thévenot discussion of worlds of justification. Such a convention is related to inspiration –
creation, innovation, vision, uniqueness and even elements of alchemy. In terms of quality

9
conventions, this means that a wine’s quality is assessed against the properties of those making
it (or marketing it), where their expertise is not necessarily based on technical competence or on
tasting prowess, but on their capability to enact an ‘artistic moment’. An inspiration convention
is part performance (in interviews with trade magazines, for example; or at the property with a
prospective buyer) and part embedded in design (a unique label, a funky name for the wine or a
special bottle shape). It may indeed be based on technical skills in winemaking (viticulture is
not fancy enough to make an impression in the same way), provided that they are used in unique
or alchemical ways. For example, the success of wines produced by South African maverick
Charles Back can only be explained by the humour he inspires and his boisterous persona. This
applies equally to commercial wines (such as tongue-in-cheek wines ‘Bored Doe’ or ‘Goats do
Roam’) and to top quality offerings based on the life histories of some of his grape suppliers
written down on the bottle label.
Mid-range wines
For mid-range wines the combination of quality conventions and of influential actors is even
more complex. Endorsement by wine critics (and an opinion convention) is still a factor, but it is
a less determinant one than in top quality wines. South African wine operators of all sizes argue
that good scores from wine writers and publications help to sell wine, as do stickers and medals
awarded in international competitions. But they also claim that what is more determinant in this
range of wines is a combination of (and often conflict between) geographical origin, brands and
grape varieties. A normative war has taken place in the wine world in the last few decades in
this quality bracket too, but here it is based on the fault lines of two different groups of
instruments of verification of quality: geographical origin/terroir (supported by Old World wine
producers) versus brand/grape variety (supported by New World wine producers), with clear
indications that the latter is gaining ground on the former. One exception to this simplified
divide is Champagne, where both the geographical indication and brand names are important for
marketing.
Indications of geographical origin are instruments of a domestic convention, but are less
precise and ‘immediate’ than ‘terroir’ and can relate to fairly large and internally diverse
regions. They seek to transmit a sense of (sometimes romanticised) connection with a place,
where trust is embedded in the specific geography, in a processing system that is typical of that
area, and in the people who carry it out. At the same time there are vast differences in the wine-
producing world on what an indication of geographical origin actually means and how it is
operationalised. In some countries, including South Africa, generic indications of geographical
origin actually transmit information more akin to a brand than a territory. Therefore, the
expertise and knowledge enrolled in the process of selling quality to buyers are closer to
marketing and branding. Instead, in smaller indications of geographical origin, geographic
information system data, soil sampling, macro-geographical and climatic information and

10
expertise are enrolled to characterise quality and operationalise the demarcation of
administrative boundaries.
Brands in mid-range wines are often accompanied by the indication of one or more grape
varieties. The approach here is to help the consumer recognise and appreciate the differences
between varieties and to hook onto a specific brand to sort through a myriad of offerings and
variations. Marketing and advertising knowledge is essential here, rather than knowledge of the
terroir, the macro-geographical traits of the area or the opinion of a renowned wine writer. From
a convention theory point of view, brands operate as a domestic quality convention as well, in
the sense that repetition of experience builds trust and the brand name becomes a substitute for
quality. However, as argued elsewhere (Gibbon and Ponte 2005), the mechanisms of
transmission of information are very different for brands than for the repetition of interpersonal
relations and intimate knowledge of places.
An increasingly important element at play in the evaluation of mid-range wines is their civic
content – as attested by the success of organic wines, and now also fair trade wines in UK, USA
and Scandinavian markets (South Africa was the first and still is the largest exporter of such
wine), The quality of wine in this case is related the impact of its production on society or the
environment. At the same time, many labels and certifications related to environmental
management (ISO 14000, organic, biodynamic, biodiversity) and social impacts (Fair Trade and
the Wine Industry Ethical Trade Association) have tended to codify and formalise these
concerns in ways that resemble industrial convention procedures (see also du Toit 2002).
Auditing and certification provide instruments to objectify civic concerns that tend to obfuscate
the civic content and frame it in the realm of fairly standardised technical devices such as
inspection, form filling and documentary flow control. For this reason, in Figure 1 this
convention is labelled civic/industrial rather than just civic.
Basic wines
Basic wines are evaluated for quality in quite a different way from mid-range wines. The
complex configuration quality conventions and processes of contestation described above leave
the ground for a much simpler constellation. Retailers decide what quality is and everyone else
follows suit. For retailers the first, and most important step, is that a basic material quality is
assured. Three elements in delivering basic quality in wine are needed: (1) intrinsics and
packaging; (2) codified solutions to food safety and (3) logistics. These are features of an
industrial convention (see details in Ponte 2007) that entails the enrolment of technical expertise
and skills at the levels of logistics, clean winemaking (as opposed to unique or fancy) and
‘Smart viticulture’ (inspired by the influential Australian consultant Richard Smart).
South African operators report that UK retailers communicate very specific demands on
intrinsics and packaging to their suppliers when buying basic quality wine: they tell them what
to bottle, what kind of label and cork to use, the weight and shape of the bottle and the recycling

11
possibilities. Specifications on intrinsics at this level of quality can be measured or described
easily for purchasing purposes. These factors are then translated into specific clean winemaking
techniques and further upstream into systematic viticultural practices.
In recent years, however, the package of basic quality that needs to be provided by suppliers
has become more demanding and goes well beyond the intrinsics of the wine. One of the main
UK retailers, for example, is implementing retail-ready packaging in wine, which entails
unloading from the pallet to shelf in one move. Retailers are also moving towards screw-cap and
synthetic closures to minimise returns for corked wine. In terms of food safety, in addition to
meeting EU food safety rules, suppliers are increasingly under pressure to conform to food
safety and quality management procedures through British Retailer Consortium, International
Food Standard and/or ISO 22000 certifications.
In relation to logistics, UK retailers are working towards lead times of 8 weeks on
promotion. As a result, UK-based agents and marketers are trying to exert more direct control
over logistics. Previously they sold wine free on board to retailers; now, some have started
selling ‘in-bond delivery’ in the UK. This way, retailers can place a call with a lead time of 3
days for delivery. Retailers are seeing themselves increasingly as shelf-space providers.
Suppliers can log into the retailer’s supply management system and monitor movements in retail
space and stocks, and order replenishment themselves. These processes entail an increasingly
enrolment of packaging and logistics expertise and knowledge in wine production and
marketing.
Once the basic quality step is cleared, then price and promotions (instruments of a market
convention) are the shortcuts for signalling quality. Social and environmental certifications do
not play a major role in this price bracket, nor do personality, geographical origin or terroir.
Wine is offered under a brand, often with the indication of a combination of varieties, but price
is more important than brand recognition. External endorsements may play a role but promotion
is much more important. So, domestic and opinion conventions do not play important roles.
Governance
In the following, a brief sketch of governance in the GVC for wine ending in the UK is
provided, based on analysis carried out in more detail elsewhere (Ponte 2007). What emerges is
a picture of a GVC that is, in its basic quality configuration, retailer-driven and highly driven in
a hands-off manner on the basis of strict demands on basic quality, price and (at least for the
UK) promotional support. The middle range strand5 is less driven, enjoys better margins for
suppliers, does not have a clear group of lead firms and allows for more input by suppliers on
the determination of quality. The top quality strand is placed in between the other two quality
strands in terms of their levels of drivenness, with external actors (wine critics and publications)
vying for prominence in driving the chain against elite producers.

12
The nature of quality conventions and their overlap in the three price brackets are key
determinants of different configurations of GVC governance. The conventions that operate in
the top price bracket make it difficult to drive a value chain completely from one functional
position: proximity and uniqueness are not elements that are easily translated and spread both
functionally and territorially, although some elements can be codified and opinion can be spread
more easily. Thus, this strand of the GVC for wine exhibits a middle range degree of
drivenness. On the basis of the better portability of the opinion quality convention, it is actually
increasingly driven by wine critics and specifically through Robert Parker’s scores. Because the
100-point Parker scale has been adopted by other influential publications a creeping element of
an industrial convention is helping to give the impression of the ‘objectification’ of tasting,
which also makes it easier to communicate information about quality, thus increasing the level
of drivenness. Interestingly, Parker does not buy or sell wine and he does not have direct
shareholdings in wineries or the trade. His empire is based on selling his opinion (McCoy
2005). The value chain is thus (partially) driven by an external actor. Although quality in this
segment is also to some extent producer-determined (quality specifications do not follow a
register provided by retailers, but are the result of tradition and/or innovation in wineries), it is
an open secret that producers (often through the enrolment of specialist consultants)
increasingly try to shadow the new aesthetics of wine that Parker and others have promoted.
In mid-range wines, even though domestic conventions are the principal instruments of
quality evaluation, a reliance on branding would in theory make it easier to communicate
quality information along the value chain (and in end-markets through advertising). However,
there is no clear driver in this strand of the GVC: branded wine producers have not yet achieved
a dominant position, and wine critics, wine marketers and retailers still have influence on its
governance. There is no clear locus from where explicit functional divisions of labour are
imposed either. Retailers and branded wine marketers (and producer/marketers) are mutually
dependent – the first need successful brands to sell; the latter need a retail outlet, a contact with
the consumer; both need to pay attention to wine critics’ judgments.
In basic wines the industrial/market nature of dominant quality conventions makes it much
easier for lead firms (retailers in this case) to specify quality information requirements and to
transmit these to their first-tier suppliers and beyond. The GVC in this strand is highly driven by
retailers, which also set all elements of a quality profile.
Wine quality, supply relations and value chain governance in South Africa
The wine industry is an important contributor to the economy of the western Cape region of
South Africa. Perhaps even more important than its direct economic impact on employment and
foreign exchange generation is the unique position of wine (and to some extent wine tourism) in
generating images of South Africa abroad. It is therefore surprising that this industry has been
the subject of a relatively limited academic literature. Much of this literature has been generated

13
by sociologists and a few geographers examining how the restructuring of the wine and fruit
industries has impacted on labour practices at the farm level (Ewert and Hamman 1999; du Toit
and Ewert 2002; Barrientos and Kritzinger 2004; Kritzinger et al. 2004; Ewert and du Toit
2005). A sub-component of this literature explicitly examines the impact of ethical trade
initiatives and other ‘empowerment’ processes on both the actual conditions of labour on the
farms and the space for manoeuvre for more radical policy options (du Toit 2002; McEwan and
Bek 2006; Bek et al. 2007). Others have tackled the process of ‘Black economic empowerment’
in the wine industry and its actual and potential impact on the industry as a whole, and not only
at the farm level (Williams 2005; du Toit et al. 2007). To the author’s knowledge there has been
no published attempt to examine the political economy of the wine value chain as a whole in
South Africa and in connection to the dynamics of wine markets globally. While Shaw (2001)
laid out some analytical pointers and provided a basic architecture for such a project, this has
not been followed up (with the partial exception of Vink et al. 2004).
In the rest of this section, I examine in brief how quality conventions that characterise the
GVC for wine ending in the UK are translated into specific wine and grape supply practices and
functional divisions of labour in South Africa, and what quality conventions underpin such
practices.6 In simplified terms, wine and grape procurement in South Africa can be described as
following four-layer model: (1) own-wine production and grape procurement (both hands-on
and hands-off); (2) procurement of ready-made wine with a hands-on management model; (3)
procurement of ready-made wine with a hands-off management model; and (4) the spot wine
market.
Different kinds of actors in South Africa engage in procuring grapes and wine in different
ways. Small wineries tend to stick to own-grape growing and winemaking. However, even they
are increasingly engaged in buying and/or selling grapes from/to others to match the quality
profiles of the portfolio of wines they want to offer. Producer–wholesalers and marketers are
striving to move away from own grape growing and in some cases even winemaking. Large
producer co-operatives do not have such an option and are increasingly holding stock (and
facing higher risks) on behalf of others in the value chain.
FIGURE 2 ABOUT HERE
Figure 2 is a simplified representation of the mix of procurement systems that wine
marketers, producer–wholesalers and private cellars operate as they move down the quality
pyramid (see Ponte 2007 for more details). The divisions in the pyramid are not as clear-cut as
the graphic representation suggests – overlaps and a combination of other systems also apply.
Large operators in South Africa handling wines in all quality ranges operate not just one model,
but an overlap of different procurement systems. At the same time, across the quality range,
they are generally striving to move away from systems sitting at the top of the pyramid and
towards those sitting at the bottom, as is explained below. The modalities chosen and the

14
general movement from one to another are often the result of demands that are placed on South
African operators by their buyers upstream – definitively in the basic quality bracket, but also
partly in mid-range and top quality wines.
Supply systems and quality conventions for the production of top quality wine
To deliver top quality wine, winemakers source the grapes they need either internally in the
same company or from trusted grape farmers rather than blending ready-made wine. In this
quality bracket good scores from wine critics are becoming important to signal quality to
consumers. This has led many top winemakers in South Africa to adopt new cellar techniques to
suit particular wine styles that are preferred by influential critics (although other winemakers
still stick to more traditional ways of doing things). In the 1990s, in the first years after the end
of apartheid, such techniques and knowledge were mainly acquired either from foreign
consultant winemakers or through extended visits in other winemaking countries. More recently
they have become part of the local knowledge base provided by public or private service
providers to the wine industry.
But winemaking itself will not be able to deliver a critic-friendly wine profile unless the
grapes delivered at the cellar are of the right quality. This means that winemakers need to
translate specific wine aesthetics (say, Robert Parker’s predilection for ripe, full colour and
high-alcohol wines) into particular grape characteristics and foster the adoption of specific
viticultural management techniques in the grape farms. As a result, grape growers need to tap
into new expertise provided by consultant viticulturists (or train their own viticulturists to
particular vineyard management systems). In this framework of close co-operation the winery
and grape growers often agree on specific practices and fine-tune them during the year (such as
timing and type of desuckering and pruning; if a new vineyard is established, agreements will be
made on the density, position and aspect of new plantings). At harvest grapes will be accepted at
the winery only when they have reached the exact degree of ripeness required by the
winemaker. Such demanding interactions require a high level of engagement between wineries,
viticulturist and grape growers, with vertical integration being the extreme end of such
engagement.
In this kind of environment, the sourcing agreements between grape growers and
buyers/users of grapes are clearly underpinned by a domestic quality convention (see Figure 2).
Quality is ‘discovered’ though close engagement and proximity or with repeated personal
interaction between the grape farmer, the viticulturist and the winemaker. When terroir comes
into play as an important quality trait in consumer markets, local supply agreements are also
based on either vertical integration (grapes are sourced from the same property that makes the
wine) or very close supervision of grape growers. Therefore, we can argue that quality demands
based on opinion and domestic conventions in consumer countries are translated in South Africa

15
in supply systems underpinned mainly by domestic conventions and instruments of verification
(see Figure 2).
Supply systems and quality conventions for the production of medium quality wine
Moving down the quality bracket to mid-range wines, it can be seen that some of the practices
and interactions between winemakers, viticulturists and grape farmers become more codified
and standardised. In mid-range wines, both grapes and ready-made wine are sourced for
winemaking and blending. Grape growers that set aside blocks for the production of medium
quality wine (they are usually booked by the buyer) will still occasionally consult a viticulturist
for their management, but they are unlikely to try to fine-tune vineyard management throughout
the year with their buyers.
In mid-range wines made from these grapes, where (more generic) geographical indications,
brands and grape varieties are important in consuming countries, it is essential that year after
year consistency is achieved. This implies that the same viticultural practice is carried out,
rather than the best practice given a particular wine profile, and that the right grape variety is
available in the right quantities. This entails a more hands-off relation between grape growers
and wineries and an industrial type of quality convention, based on a more formulaic
management system – including indicators of maximum yield per hectare, for example, and
more dependency on ex post objective measurement of grape quality, rather than ex ante co-
management. But even when engagement by the buyer becomes less hands-on, domestic
elements of trust and engagement remain (for example, the winemaker will make several visits
to grape farms close to harvest time). Thus, as shown in Figure 2, this combination of grape
procurement is said to be underpinned by a compromise between industrial and domestic quality
conventions.
If the wine marketer, producer–wholesaler or private cellar uses ready-made wine for
blending and further cellaring (rather than using their own grapes or buying in grapes for
winemaking), hands-on procurement systems will usually be applied for mid-range quality
wine. This means that, at the beginning of the season, the wine buyer (a marketer or producer–
wholesaler) books a tank from the private cellar or co-operative that will make the wine. The
buyer provides detailed specifications on variety, style, wooding regime and bio-chemical
profile, and the buyer’s winemaker is involved in the vinification process and/or blending. Some
level of engagement is also carried out on the farms where the grapes will be sourced from to
make that particular wine (which entails at the very least block identification and some general
routine of viticultural management). But the level of hands-on engagement is not as close as it is
in the case of grape-buying in general. Although elements of a domestic convention remain in
wine sourcing, other quality parameters are set through industrial-type measurement devices and
procedural systems (thus, Figure 2 shows a domestic/industrial convention for wine
purchasing).

16
When social and environmental concerns are at play in consumer markets, these are often
translated into certification and auditing procedures at the production level (both for grapes and
wine). As a result, growers and wineries have to modify their operations, but these changes are
rarely the result of their own initiative (with the main exception of fair trade wine, of which the
certification process and content actually originated in South Africa; see Kruger and du Toit
2007). They are responses to demands placed on them in consumer markets. The conventions
underpinning the supply solution to these quality aspects in South Africa are based on objective
(or at least codified and formalised) measurements and procedures. Thus, they are more akin to
an industrial convention rather than resting on civic concerns per se. This is why civic
conventions do not appear in Figure 2 as key quality conventions that underpin supply systems
in South Africa for mid-range quality wine.
Supply systems and quality conventions for the production of basic wines
For basic wines marketers, producer–wholesalers and private cellars tend to buy and blend
ready-made wine, usually from co-operatives and ex-co-operatives, on the basis of a sample of
ready-made wine without too much input (in a hands-off manner), especially once a commercial
relation has been established for a few years. The requested wine needs to be ‘-clean and
drinkable and match a generic style. Consistency of quality year after year is still as important
as for medium-range wines, but even more important is that within the same season all bottlings
are homogeneous. Guaranteeing a minimum volume is absolutely necessary.
Price plays a key role in signalling quality in consumer markets for basic wines, which
means that all sorts of devices are used at the level of supply to cut down on costs and optimise
processes. The rules of engagement are clear and simple for suppliers. For these wines, the
marketer or producer–wholesaler provides the wine cellar with a generic recipe (itself often
negotiated between the marketer and the retailers’ wine buyer in the importing country, together
with a full breakdown of costs along the chain) and only limited input in the winemaking
process. Volumes will be drawn by the marketer or producer–wholesaler as needed within a 6–7
month period (and bottling will be done at the very last minute). In contrast, in the past
marketers purchased wines from wine cellars as soon as possible to establish control over the
wine. They made one blend for the whole year just after harvest. Currently, they need to keep
the wine at the cellar as long as possible. Smaller batches are blended and with more precise
specifications. Quality management in procurement in this case is based on elements of an
industrial convention, but trust and personal relations still play a role between seller and buyer
to ensure minimum volumes and timely delivery (thus, Figure 2 indicates an industrial/domestic
convention). Finally, some large producer–wholesalers and marketers also procure wine through
spot-buying. In this case the buyer visits the cellar, tastes wines from selected tanks and decides
whether or not to buy. Quality management in procurement is carried out through a mechanism
that is close to a market convention.

17
In conclusion, in the case of basic wines, market and industrial conventions that apply in
consumer markets are translated themselves into largely industrial and market conventions
regulating supply systems, with residual domestic conventions due to the importance of
guaranteeing volumes in a country that has only few large producers and limited scope for
expansion of viticultural land.
Governance
The South African segment of the value chain is characterised by producer–wholesalers and
marketers as the main lead firms, although their power over other actors in the country is limited
by their own need to deliver volume and homogeneous quality to their importers. At the same
time they are rationalising their supply base, shadowing what retailers are themselves doing in
the UK. They occupy a key position in South Africa because they translate quality conventions
that apply in consumer markets for different levels of quality into actual sourcing practices,
themselves underpinned by specific quality conventions and instruments of verification. For top
quality wines they rely on mainly domestic quality conventions to source the best grapes and
wine. For mid-range wines they have devised more industrial-type conventions based on
codified and routinised tools, combined with some level of repetitive interaction and hands-on
engagement (and domestic quality conventions). As they move down to basic quality wines,
they have increasingly been able to provide simple, ready-made recipes and clear parameters to
their suppliers, based on industrial and market conventions of quality.
In terms of governance marketers and producer–wholesalers are reshaping the functional
division of labour in the wine value chain in South Africa by
• generally trying to move away from grape growing towards buying-in grapes with hands-on
management
• trying to divest from winemaking as well, or to move from hands-on to hands-off
management
• partially integrating some downstream functions in logistics, inventory management and
replenishment through joint ventures with importing country agents.
In this reconfiguration of the functional division of labour, inventory and risk are pushed
upstream (in terms of volume and time) all the way to producer co-operatives and other large
cellars and to grape growers.
Conclusion: governing through quality
In this article convention theory provided an entry point to the theoretical discussion of
‘governance as normalisation’ in GVC analysis. Three different quality strands of the wine
value chain ending in the UK were examined. The three strands showed different degrees of
drivenness and mechanisms of governance, underpinned by specific quality conventions and
instruments of verification in consumer markets. These translate in specific configurations of

18
supply relations at the production level, themselves based on other, more localised, quality
conventions. The article highlighted disjunctures and fragmentations in GVC governance and
conflictual and unresolved definitions and applications of ‘quality’. Yet, a number of general
traits can still be observed and theorised upon.
The wine value chain is highly driven only in its basic quality strand, where industrial and
market conventions underpin its governance. Here, the value chain is characterised by supply
relations at the production level (in this case, South Africa) that offer little flexibility, low
margins and demanding volumes and logistics. Wine value chain strands that are governed
through other quality conventions, especially territorial-based domestic conventions (such as
geographical origin and terroir) and inspiration conventions tend to be less driven because such
conventions are less portable and thus more difficult to be operated at a distance by one group
of lead firms. In such situations, producers have more flexibility/control over production
processes and also enjoy better returns. Where opinion conventions play an important role (such
as in top quality wine), an intermediate degree of driving (by an external actor, such as a wine
critic, judge or publication) is exhibited, due to the easier portability of such a convention.
These value chain strands can offer substantial rewards to producers but also operate in an
environment of high risk (a bad score from Parker can ruin a cellar).
These observations show that while the discussion of governance in global value chains
remains important, it is equally important to examine how driving mechanisms are underpinned
by specific quality conventions and instruments of verification and how they are transmitted and
translated upstream (or downstream) in the value chain from where lead firms are functionally
situated. It is also essential to undo chains into strands and possibly compare similar quality
strands in different chains.
The analysis of quality conventions helps clarifying the extent to which drivenness can be
exercised in value chains (in wine, lead firms are able to drive the value chain only when
portable conventions are dominant, such as in the basic quality strand). Quality conventions can
also clarify how and why the same players apply different driving mechanisms when they
operate in different strands of value chains, and how and when lead firms are able to drive a
value chain (or a segment of it) with limited or no hands-on involvement beyond their
immediate suppliers. Finally, the analysis of quality conventions can help delineate how lead
firms do not govern chains solely on the basis of buyer power, market share, and/or economies
of scale or scope, but also through normative work.
Methodological appendix
This article is based on material gathered through fieldwork carried out in South Africa from
June to November 2005, plus a continuous survey of trade and consumer-oriented wine
magazines from January 2005 to October 2008. Part of the fieldwork in South Africa consisted
of 99 interviews with 74 entities that are directly or indirectly involved in the production of

19
grapes and wine, and their processing and marketing, and/or related inputs and services. In
terms of direct players, the main focus of fieldwork was at the marketing/exporting level and, to
some extent, at the cellar level. All nine producers–wholesalers and marketers/exporters that sell
more (or around) 1 million cases of wine per year were interviewed, plus five smaller ones
(selling between 120,000 and 800,000 cases each). The 14 entities interviewed account for
marketing almost 280 million litres of wine, or 44 per cent of the 2005 production of wine. In
the category of private cellars, given the large number of cellars in operation, the aim was to
cover a small number of medium-scale cellars (four producing between 100,000 and 250,000
cases each) and of small-scale cellars or estates (ten producing between 4,000 and 90,000 cases
each). In addition to this, 20 per cent of existing co-operatives and ex-co-operatives were
interviewed (12 in all). The co-operatives covered in the study account for a production of
around 180 million litres of wine, almost one-third of total production of wine in South Africa
in 2005. In relation to grape growers, given the limited resources available, only a small group
(10) was interviewed to provide a general idea of how contractual relations, quality assessment
and agronomic practices were changing. All three categories of growers were covered (in
addition to private cellars who have their own vineyards): independent grape growers (without
cellar facilities), growers that are part of a co-operative and growers that are shareholders in an
ex-co-operative. Finally, interviews were carried out with retail chains (three of the top four
retailers in South Africa), wine writers and judges, providers of services to the industry and
government entities that regulate the production, trade and retailing of wine. The UK segment of
the value chain for wine was examined on the basis of interviews with agents based in South
Africa that deal directly with the UK wine market and through the analysis of wine trade
publications (Harpers, EuroWine, The Drinks Business, Impact International, The Australian &
New Zealand Grapegrower & Winemaker, Drinks International, Wineland, Wine Business
Monthly, Grape, and Wine and Spirit International) and consumer-oriented wine magazines
(Wine [South Africa], Wine Spectator and Decanter). All information presented in this article,
unless otherwise stated, derives from the primary material gathered through interviews. The
identity of companies and individuals has been withdrawn to comply with the statement of
confidentiality that was offered to interviewees.
References
Altenburg, T. (2006) Governance patterns in value chains and their development impact.
European Journal of Development Research 18 (4) pp. 498–521
Anderson, K. (2004a) Introduction. Pp. 3–13 in K. Anderson ed. The world’s wine markets:
globalization at work (Cheltenham: Edward Elgar)
Anderson, K. ed. (2004b) The world’s wine markets: globalization at work (Cheltenham:
Edward Elgar)

20
Barham, E. (2002) Towards a theory of values-based labeling. Agriculture and Human Values
19 (4) pp. 349–360
Barrientos, S. and A. Kritzinger (2004) Squaring the circle: global production and the
informalization of work in South African fruit exports. Journal of International
Development 16 (1) pp. 81–92
Batifoulier, P. and G. de Larquier (2001) Introduction – De la convention et de ses usages. Pp.
9–31 in P. Batifoulier ed. Téorie des conventions (Paris: Economica)
Bek, D., C. McEwan and K. Bek (2007) Ethical trading and socioeconomic transformation:
critical reflections on the South African wine industry. Environment and Planning A 39
(2) pp. 301–319
Boltanski O. and L. Thévenot (1991) De la justification. Les économies de la grandeur
(Justification: economies of scale) (Paris: Gallimard)
Boltanski, L. and E. Chiapello (1999) Le nouvel esprit du capitalisme (The new spirit of
capitalism) (Paris: Gallimard)
Borghi, V. and T. Vitale (2006) Convenzioni, economia morale e analisi sociologica
[Conventions, moral economy and sociological analysis] Sociologia del lavoro 104 pp.
7–34
Daviron, B. and S. Ponte (2005) The coffee paradox: Global markets, commodity trade and the
elusive promise of development (London: Zed books).
du Toit, A., S. Kruger and S. Ponte (2007) De-racialising exploitation: ‘Black Economic
Empowerment’ in the South African wine sector. Journal of Agrarian Change 8 (1) pp. 6-
32.
Ewert, J. and A. du Toit (2005) A deepening divide in the countryside: restructuring and rural
livelihoods in the South African wine industry. Journal of Southern African Studies 31
(2) pp. 315–332
Ewert, J. and J. Hamman (1999) Why paternalism survives: globalisation, democratisation and
labour on South African wine farms. Sociologia Ruralis 39 (2) pp. 202–221
Eymard-Duvernay, F. (1989) Conventions de qualité et formes de coordination. Revue
Economique 40 (2) pp. 329–359
Eymard-Duvernay, F. ed. (2006a) L’économie des conventions, methodes et resultats. Tome 1:
Débats (Conventions, methods and results in economics: debates) (Paris: La
Découverte)
Eymard-Duvernay, F. ed. (2006b) L’économie des conventions, methodes et resultats. Tome 2:
Développements (Conventions, methods and results in economics: developments)
(Paris: La Découverte)

21
Eymard-Duvernay, F., O. Favereau, A. Orléan, R. Salais and L. Thévenot (2006) Valeurs,
coordination et rationalité: trois thèmes mis en relation par l’économie des conventions
[Values, co-ordination and rationality: three themes linked by the economics of
conventions] Pp. xxx–xxx in F. Eymard-Duvernay ed. L’économie des conventions,
methodes et resultats. Tome 1: Débats (Conventions, methods and results in economics:
debates) (Paris: La Découverte)
Favereau, O. and E. Lazega eds (2002) Conventions and structures in economic organization
(Cheltenham: Edward Elgar)
Freidberg, S. (2003) Culture, conventions and colonial constructs of rurality in south-north
horticultural trades. Journal of Rural Studies 19 (1) pp. 97–109
Freidberg, S. (2004) French beans and food scares: culture and commerce in an anxious age
(New York: Oxford University Press)
Gereffi, G. (1994) The organization of buyer-driven global commodity chains: how US retailers
shape overseas production networks. Pp. 95–122 in G. Gereffi and M. Korzeniewicz
eds, Commodity chains and global capitalism (Westport, CT: Greenwood Press)
Gereffi, G., J. Humphrey and T. Sturgeon (2005) The governance of global value chains.
Review of International Political Economy 12 (1) pp. 78–104
Gibbon, P., J. Bair and S. Ponte (2008) Governing global value chains: An introduction.
Economy and Society 37 (3) pp. 315-338.
Gibbon, P. and S. Ponte (2005) Trading down: Africa, value chains and the global economy
(Philadelphia: Temple University Press).
Gibbon, P. and S. Ponte (2008) Governing value chains: From governance to governmentality?
Economy and Society 37(3) pp. 365-392.
Guthey, G.T. (2008) Agro-industrial conventions: some evidence from northern California’s
wine industry. The Geographic Journal 174 (2) pp. 138–148
Gwynne, R.N. (2006) Governance in the wine commodity chain: upstream and downstream
strategies in New Zealand and Chilean wine firms. Asia Pacific Viewpoint 47 (3) pp.
381–395
Gwynne, R.N. (2008) UK retail concentration, Chilean wine producers and value chains. The
Geographic Journal 174 (2) pp. 97–108
Hayward, D. and N. Lewis (2008) Regional dynamics in the globalizing wine industry: the case
of Marlborough, New Zealand. The Geographic Journal 174 (2) pp. pp. 124–137
Kritzinger, A., S. Barrientos and H. Rossouw (2004) Global production and flexible
employment in the South African horticulture: experiences of contract workers in fruit
exports. Sociologia Ruralis 44 (1) pp. 17–39

22
Kruger, S. and A. du Toit (2007) Reconstructing fairness: fair trade conventions and worker
empowerment in South African horticulture. Pp. 200–219 in L.T. Raynolds, D. Murray
and J. Wilkinson eds, Fair trade: the challenges of transforming globalization (London
and New York: Routledge)
Lagendijk, A. (2004) Global ‘lifeworlds’ versus local ‘systemworlds’: how flying winemakers
produce global wines in interconnected locales. Tijdschrift voor Economische en
Sociale Geografie 95 (5) pp. 511–526
Lewis, N. (2008) Constructing economic objects of governance: the New Zealand wine
industry. Pp 103–120 in C. Stringer and R. Le Heron eds, Agri-food commodity chains
and globalising networks (Aldershot: Ashgate)
McCoy, E. (2005) The emperor of wine: the rise of Robert M. Parker, Jr. and the reign of
American taste (New York: HarperCollins)
McEwan, C. and D. Bek (2006) (Re)politicizing empowerment: lessons from the South African
wine industry. Geoforum 37 (6) pp. 1021–1034
Murdoch, J. and M. Miele (1999) ‘Back to nature’: changing ‘worlds of production’ in the food
sector. Sociologia Ruralis 39 (4) pp. 465–483
Murdoch, J., T. Marsden and J. Banks (2000) Quality, nature and embeddedness: some
theoretical considerations in the context of the food sector. Economic Geography 76 (2)
pp. 107–125
Ponte, S. (2002) The ‘latte revolution’? Regulation, markets and consumption in the global
coffee chain. World Development 30 (7) pp. 1099-1122.
Ponte, S. (2007) Governance in the value chain for South African wine. TRALAC Working
Paper 2007/9 (Stellenbosch: Trade Law Centre for Southern Africa).
Ponte, S. and P. Gibbon (2005) Quality standards, conventions and the governance of global
value chains. Economy and Society 34 (1) pp. 1-31.
Ponte, S. and J. Ewert (2007) South African wine: an industry in ferment. TRALAC Working
Paper 2007/8 (Stellenbosch: Trade Law Centre for Southern Africa).
Pritchard, B. (1999) The regulation of grower–processor relations: a case study from the
Australian wine industry. Sociologia Ruralis 39 (2) pp. 185–200
Raikes, P., M.F. Jensen and S. Ponte (2000) Global commodity chain analysis and the French
filiére approach: comparison and critique. Economy and Society 29 (3) pp. 390–417
Raynolds, L.T. (2000) Re-embedding global agriculture: the international organic and fair trade
movements. Agriculture and Human Values 17 (3) pp. 297–309
Raynolds, L.T. (2002) Consumer/producer links in fair trade coffee networks. Sociologia
Ruralis 42 (4) pp. 404–424
Raynolds, L.T. (2004) The globalization of organic agro-food networks. World Development 32
(5) pp. 725–743

23
Renard, M.C. (2003) Fair trade: quality, market and conventions. Journal of Rural Studies 19
(1) pp. 87–96
Shaw, T.M. (2001) South Africa and the political economy of wine: from sanctions to
globalizations/liberalizations. Pp. 89–105 in S.J. MacLean, F. Quadir and T.M. Shaw
eds, Crises of governance in Asia and Africa (Aldershot: Ashgate)
Spahni, P. (2000) The international wine trade (2nd edn) (Cambridge: Woodhead).
Storper, M. and R. Salais (1997) Worlds of production: the action frameworks of the economy
(Cambridge, MA and London: Harvard University Press)
Sturgeon, T. (2002) Modular manufacturing networks: a new American model of industrial
organisation. Industrial and Corporate Change 11 (3) pp. 451–496
Sylvander, B. (1995) Convention de qualité, concurrence et coopération: Cas du label rouge
dans la filière volaille. [Quality conventions, competition and co-operation: a case study
on the red label for poultry]. Pp. 73–96 in G. Allaire and R. Boyer eds, La grande
transformation de l’agriculture (Paris: INRA-Economica)
Thévenot, L. (1995) Des marchés aux normes (From markets to norms). Pp. 33–52 in G. Allaire
and R. Boyer eds, La grande transformation de l’agriculture (Paris: INRA-Economica)
Toit, A. du (2002) Globalising ethics: social technologies of private regulation and the South
African wine industry. Journal of Agrarian Change 2 (3) pp. 356–380
Toit, A. du and J. Ewert (2002) Myths of globalisation: private regulation and farm worker
livelihoods on western Cape farms. Transformation 50 pp. 77–104
Unwin, T. (1996) Wine and the vine: an historical geography of viticulture and the wine trade
(London: Routledge)
Vink, N., G. Williams and J. Kirsten (2004) South Africa. Pp. 227–251 in K. Anderson ed. The
world’s wine markets: globalization at work (Cheltenham: Edward Elgar)
Wilkinson, J. (1997) A new paradigm for economic analysis? Economy and Society 26 (3) pp.
305–339
Wilkinson, J. (2006) Fish: a global value chain driven onto the rocks. Sociologia Ruralis 46 (2)
pp. 139–153
Williams, G. (2005) Black economic empowerment in the South African wine industry. Journal
of Agrarian Change 5 (4) pp. 476–504
Stefano Ponte
Danish Institute for International Studies
Strandgade 56
Copenhagen 1401
Denmark
e-mail: spo@diis.dk

24
1 Convention theory is also known as the economics of conventions. The term ‘convention theory’ is
used in this article to avoid confusion with game theoretical approaches (also referred to as
‘economics of conventions’ in France) that focus on conventions as outcomes of strategic interaction
motivated by personal interest. What is meant by ‘convention theory’ here is thus the interpretive and
normative approach to conventions (see Batifoulier and de Larquier 2001).
2 The term ‘convention’ is used here to also cover the convention theory meaning of ‘co-ordination’.
This is to avoid confusion with the GVC meaning of ‘co-ordination’.
3 An exception to this trend is work by Storper and Salais (1997) that builds upon the original Boltanski
and Thévenot framework and focuses on forms of industrial organisation.
4 Segments are large vertical chunks of GVCs, for example, the part of the value chain from the point of
production to export, or from import to retail.
5 Strands are specific (and parallel) typologies of GVCs in which governance may differ from the
mainstream GVC under consideration (or from other strands, if there is no clear mainstream).
Governance of specific strands still refers to ‘whole length’ GVCs, unless combined with segments.
Strands may differ because of different product characteristics (e.g., specialty coffee); a different
institutional configuration (e.g., the presence of an auction that cuts the GVC); or a different end-
market/origin of production (e.g., clothes retailed in Europe or in the USA; bananas grown in Latin
America).
6 Due to space limitations only the export-oriented segment of the South African wine value chain can
be examined here. The domestic market represents around 50 per cent of the total value of wine sales
(for more details, see Ponte and Ewert 2007).