Page 1 sur 18

Wine industry Supply Chain (WSC) modeling: an Argentina-France comparison

Saglietto Laurence

University of Nice Sophia Antipolis – GREDEG CNRS France

Corresponding author: laurence.saglietto@gredeg.cnrs.fr

Fulconis François University of Avignon - CRET-LOG / Aix-Marseille University

France

Bédé Didier University of Toulouse III, Research Center LGCO

France

De Almeira Goes Joice University of Nice Sophia Antipolis – GREDEG CNRS

France

Forradellas Raymundo Universidad Nacional de Cuyo – CEAL

Argentina

Abstract The purpose of this research is to understand the Wine industry Supply Chain (WSC) structure from a social networks approach. This analysis will identify the appropriate measure of performance of the WSC. To achieve that objective, this paper will at first tackle the supply chain modeling on the one hand through a generic approach, and on the other hand we will propose a WSC representation from a social networks analysis (SNA). In order to assess the WSC structure, we consider three levels: triads, groups and the whole network. We will then be able to answer the following question: what is the WSC structure? Secondly, a comparative study of the WSCs will be realized between the New World and the Old World ones using data from Mendoza in Argentina (MA) and South-Eastern France (SEF) wine region. The paper presents a conceptual model of the WSC. This model will be assessed through qualitative studies among different members located in Mendoza and in South-Eastern France. This paper proposes a framework contributing to understand worldwide WSCs structure, mapping tools and some indicators to assess the reasons of their strength or weakness. The tools discussed may be reused to map any WSC in the world. Thus, this paper offers insights into the pre-requisite conditions for a thorough analysis of the performance criteria and potential improvements of WSC through a method leading to creation of value for organizations. Keywords: Supply Chain, Wine Industry, Structure, Social networks, Argentina, France,.

Page 2 sur 18

1. Introduction

To understand the complexity of international wine trade, several approaches are possible: for instance, a geographical approach (regarding the evolution of production and consumption areas directly related to export flows), or an economic approach (since the 60s, the production is increasing and its structure has changed with the emergence of new markets). The stakes are high, given the profits generated by this industry. Mariani et al. (2012) explore the changes which occurred in the geography of international trade. The aggregated wine import flows were divided into five groups of countries according to their role in the international market: - major importers1: Germany, UK and USA, the three countries which have long been the main export

destinations for wine, - small traditional importers: 12 countries, - small non-traditional importers: 56 countries, - major Mediterranean exporting countries: France, Greece, Italy, Spain, Portugal, - other exporters: New-Zealand, Australia, Argentina, Chile, South Africa. France and Argentina belong to different categories due to their geographical position. They are not exposed to the same economical fluctuations. Mediterranean exporting countries have lost market share in favor of other exporting countries that have increased their market share. This is partly due to the preferences of many consumers for imported wines as well as on the transfer of production from small units to multinational groups. Thus, the international distribution network has changed to fit to the new international structure of demand. A recent phenomenon has emerged: intermediation. Intermediation has developed in countries other than the consumer and producer ones (e.g. Singapore for example). Henceforth, the traditional competitive relationships among intermediaries and vertical relationships in the supply chain have been changed. Today, the number of intermediaries has increased, be it market-makers or match-makers. Mariani et al. (2012) show clearly the phenomenon of re-export through intermediaries, as following: "by combining statistics on wine production and exports it is possible to identify some exporting countries without domestic production or with a production lower than their exports. By filtering wine exporters in the way a set of 33 re-exporting countries was identified". Beside these geographical or economic approaches, more territorial (Demossier, 2011) and managerial approaches (Fernandez-Olmos et al., 2009) focus on the performance of the Wine industry Supply Chain (WSC). It became an important issue for many parts of both the old and new world. Indeed, the international wine trade is based on a complex industry insofar as the wine is produced according to specific quality standards (in different regions) and, for being exported requires specific conditions. WSC is complex for several reasons: - the nature of the product (wine is available in red, rosé, white, sweet and effervescent), - the large number of heterogeneous actors (they belong to different professional categories: grower, producer,

transporter, distributor, retailer, cellar...), - the relationships between the actors are complex (e.g. the existence of a multi-level system for the distribution

cycle that differs depending on the country), - the requirements of the end consumers are contextualized and cultural, continued pressure from local and

international competitors intensifies competition on the planet, - finally, the legal constraints of the production and distribution makes the task even more difficult.

Thus, mapping the WSC is a complex exercise. According to Garcia (2009), the dynamics of WSC, "is very complex not only due to quantity of actors who participates to fulfill the customer’s requirement, but also because the SC integration, which is different according to the culture, agricultural and industrial practices of each country". As any extended logistic chain, it requires "a synchronization of all activities" (Min and Zhu, 2002), the information sharing through cooperation in the ultimate aim of creating value for all actors. Appreciating the timing and all physical and information flows of the WSC requires a modeling of flow, relationships, links and actors. "In order to understand the WSC it is necessary to count with a model that represents its operations, its relationships and its dynamics" (Garcia, 2009). We propose to model the heterogeneous actor's network of WSC from a social network analysis (SNA). This analysis will then be mobilized to conduct a study of the overall structure of the WSC to show how the shape of the network is explanatory of strategic orientations. The structure is both determining framework and determined by the action. To achieve this goal, we first discuss the modeling of the supply chain through a generic approach and then propose, from a social networks approach, a representation of the WSC (2). In order to appreciate the structure of the WSC, three different levels are considered: the triads, the groups and the complete network. We will be able to provide answers to the following question: What is the structure of the WSC? We will then present the research design used (3). It consists in comparing the WSC of Mendoza in Argentina (illustrating the "new world") to the South-East of France (illustrating the "old world"). In this regard, the model we propose will be

1 Large importers: three countries which have long been major destinations for wine exports.

Page 3 sur 18

appreciated through qualitative studies. Finally, the main results will be presented and discussed (4). This paper presents an analysis which contributes to a better understanding the WSC in the world, mapping tools and information to assess the reasons for their strength or weakness. This paper offers insights into the pre-requisite conditions for a thorough analysis of performance criteria and possible improvements in the WSC.

2. From the modeling of supply chain to the modeling of Wine industry Supply Chain (WSC)

What is the structure of the WSC? To answer this question we will discuss the modeling of the supply chain through a generic approach (2.1). Then we will propose, from a social networks approach, a modeling of the WSC (2.2). 2.1. Modeling the supply chain

Mapping the supply chain of a sector or a business has now become an essential tool for analyzing the performance and reducing costs, while creating financial value. This need for modeling, responds to the vertical disintegration (in the 90s), which has led firms towards a new organization that link in a network of supply chains. Based on this observation, research in supply chain showed interest in the design of supply chain and the development of supply chain management instruments. There are several definitions of supply chain and supply chain management, highlighting the diversity of models. For example, on one hand, Stevenson (1999) defines the SC "as a sequence of supplier warehouses, operations, and retail outlets". He differentiates between two kinds of supply chains one relating to goods and the second relating to services and on another hand the Supply Chain Council defines the SC as follows: "The supply chain encompasses every effort involved in producing and delivering a final product, from the supplier’s supplier to the customer’s customer" (Supply Chain Council, 2002). Similarly, the SCM has several definitions. According to Handfield & Nichols (1999), "the SCM includes managing information systems involved in sourcing and procurement, production, scheduling, order processing, inventory management, warehousing, distribution and customer service", and according to Chen & Paulraj (2004, pp. 119-120) "the term of SCM has been used to explain the planning and control of materials and information flows as well as the logistics activities not only internally within a company but also externally between companies". To return to SC, its boundaries lie at the heart of the debate, since there are several types of supply chain. In this regard, Chen & Paulraj (2004) analyzed and synthesized over 400 articles from diverse disciplines. They deduced from the results of their study, several key points: "the entire domain of the concept (SC) is very extensive and cannot be covered in just one study" (Chen & Paulraj, 2004, p. 135). "While studies in organizational structure in general have not been lacking, research addressing the network structure conducive to supply chain performance has been very limited" (Chen & Paulraj, 2004, p. 125). In this context, supply chain is represented as a linear approach (suppliers – purchasing – production – distribution – customer). Aware of the limitations presented by these initial definitions, Caddy & Helou (1999, 2007) propose to apply the principles of the general systems theory (Yourdon, 1989; Von Bertalanffy, 1993) to the supply chain and its management in order to appreciate the interaction of the supply chain components. They seek to model the way supply chain operates as a system. "Supply chains are composed of components: people, organizations, technological structure, information flows, flows of physical good so, and flows of intangible services" (Caddy & Helou, 2007, p. 322). The system includes interconnected parts which affect each other. This representation is very abstract and ignores explicitly consumers and suppliers. However, by studying how subsystems are interconnected, or which are the differences or the lack of articulation between subsystems, the feedback mechanisms or the interactions with the environment, the performance of SC system is approached. Henceforth, an approach in terms of open systems seems more appropriate to address the global supply chain. Liu (2011) starts from the same postulate, arguing that the supply chain is a system made of components. However, in order to represent it, he does not refer to the general systems theory but to the modern logistics supply chain theory. The purpose of this representation is to identify areas that may benefit from a reduction of logistics costs and increase efficiency and profits. This theoretical approach makes the link between management and logistics. The representation of supply chains takes ultimately the great ideas of the previous approaches. The model presents the complex reality of the supply chain. To approach this complexity Cooper et al. (1997) or Lambert et al. (1998) seek to capture both the complexity and reality by proposing a modeling of supply chain network with three structures: the type of supply chain partnership (primary and secondary partners), the structural dimensions of a supply chain network (horizontal and vertical structure), and the characteristics of process links among supply chain partners (managed process links, monitored process links, not managed process, and non-member process links). Garcia (2009) and Garcia et al. (2012) propose to use a modeling language or network presentation to build "a generic model of the Supply Chain which represents all the possible instances. In this way, some authors have focused on the interests of a representation of the supply chain by the Social Network Analysis (SNA) (Lazzarini et al., 2001 ; Ellram et al., 2006 ; Carter et al., 2007 ; Mueller et al., 2007 ; Borgatti & Li, 2009 ; Kim et al., 2011 ; Galaskiewicz,

Page 4 sur 18

2011).The majority of these authors explain why theories from social networks are particularly relevant. They present the key elements that make up a network, indicators and analyzes mobilizable: nodes level (degree, closeness centrality and betweenness centrality) and network level (density, centrality, complexity)… However, few of them go beyond in mapping a supply chain because they have recognized the difficulty of collecting network-level data in supply network (for a complete state of the art see Kim et al., 2011). Lazzarini et al. (2001) demonstrate the interest to mobilize a SNA by the fact that supply chain analysis suggests vertical interdependencies that require "a systemic understanding of resource allocation and information flow between firms". The interpretation of supply chain by an SNA is interesting because "supply chains are not really linear chains but most often expansive networks" (which explains the neologism “supply networks” or “net chains”) (Lazzarini et al., 2001; Mueller et al., 2007; Kim et al., 2011).For Carter et al. (2007), SNA "can be applied both within and between organizations in a supply chain". SNA analyses the structure of a network and map the relation among a group of actors. The relations can represent linking, communication, service or products. SNA describes and analyses the interrelationships of units or nodes within a network. In fact, SNA contribute to maps the complexity of supply chain. Supply chain consists of several nodes, which can be call "actors". Each actor is a character, a link of the chain, a part played by a performer. Our work follows this path, while trying to go further. Indeed, our goal here is to assess and map all levels of network modeling of the WSC, based on SNA (Granovetter, 1985; Wasserman, 1994; de Nooy et al., 2005; Knoke & Yang, 2008). According to Kim et al., (2011, p. 195) "the term of “network” into supply chain management research represents a pressing need to view supply chains as a network for firms to gain improved performance, operational efficiencies, and ultimately sustainable competitiveness". Theories of social networks allow taking into account all the social interactions that may influence the choice of actors of the supply chain and supply chain as a whole, because each actor is not isolated but is part of a set of relationships, which influence their choices. Thus, we agree with Borgatti & Li (2009, p. 8), who argue that "the ego network concept is probably the closest to a supply chain theorist’s intuitive understanding of a supply network, but it is not a perfect match". We will refer to a "meso-social" level analysis, i.e. expanded to actors’ relations. Three levels of analysis are to be considered for modeling the supply chain: the triad (Snijders, 2001; Opsahl, 2013), the group (Newman & Park, 2003; Borgatti & Li, 2009; Balasundaram et al. 2011) and the complete network (Newman & Park, 2003; Opsahl, 2013). Regarding the triad, the analysis focuses on the relationship formed by three actors, interconnected within the structure in which they fit and with respect to all other network triads. Analysis cares about all the strategies underlined by the triads. Each actor can have a different broker role. Let us note that the dyad was not chosen because we consider here that a network starts with three players. Regarding the group, it's a set of actors who know each other and maintain privileged relations of same nature. The group is composed of relatively homogeneous actors who react similarly to the dissemination of standards, for example. They maintain relatively dense interlinks. Concerning the complete network, it's a network restricted by relevant socially instituted borders with an a priori internal network cohesion. Existing borders are those of the supply chain and relationships are dense. 2.2. Modeling the wine industry supply chain (WSC) with social networks analysis (SNA)

In general, research in logistics offers several possible analysis units such as individuals, functions, companies or organizations, sectors or industries, interfaces or relationships, chains where the emphasis is made on flows between companies or projects, processes or information systems (Fabbe-Costes et al., 2013).The three level of

analysis for modeling WSC include the following items: triads, groups and complete network. Nodes are composed by several actors and ties are only "contractual relationship" of several types: purchase relationship (wine grape grower or producer and raw materials supplier, grape grower and producer sells wine grapes), contractual outsourcing relationship (freight forwarder), relationship selling, relationship norms (implementing the same standard quality). In our study, it is important to have only one unit of analysis to assess the three levels of WSC. The unit we have chosen is the "contractual relationship" allowing us to go beyond the mere supply network. Our model is a simplified representation of contractual relations between actors in a generic form that each country will have to take ownership. We seek to reflect the diversity of actors and not their number. Then each actor is actually a community. The goal is a simplified representation, a global multi-dimensional and coherent view (for a detailed presentation of the actors, see Table 1 below).

Page 5 sur 18

ACTORS DEFINITIONS (adapted to GS1, 2005; Garcia 2009; Garcia et al., 2012; and the authors)

Raw material supplier

This actor provides Wine Producers and Fillers/Packers with all the supplies needed for wine-making or filling and packing. The main activities are: receive new orders from wineries and/or Fillers/Packers, prepare orders, send supplies to the wineries and Fillers/Packers, store supplies, etc.

Grape Grower

The Grape Grower is responsible for the production and harvest of the grapes. The main activities of grape growers are: planting the grapes, cultivating and pruning the vines, eliminating the inadequate vineyards, fertilizing the vineyards, controlling plagues, harvesting grapes, etc.

Wine producer

Wine producers are responsible for receiving grapes, the elaboration, manufacture and/or blending of wine products. In general, the main activities to elaborate wine are: receiving and weighing the grapes, crushing, stemming and pressing juice, addition of sulfite and decanting, addition of yeast, fermentation, refrigeration, clarification and stabilization, temperature control, preparation for bottling, maturation in bottle, etc.

Cooperative winery

A group of grape producers join their resources to own a winery at a lower cost per unit by maximizing the production volume. Grape producers bring their grape at the winery cooperative which commercializes their wines to numerous customers: trader, restaurateur, retailer and mass market distribution industry.

Bulk wine distributor

The bulk distributor is responsible for reception, storage, dispatch, processing, sampling and analysis of bulk wine. The bulk distributor receives bulk wine from the Wine Producer. The wine is usually pumped into transport containers such as road tankers or barrels. When the wine arrives at the “tank farm”, the bulk distributor checks the receiving documents and takes samples for tasting and analysis. He approves or rejects the wine (if rejected, the wine returns to the nominated source).

Transit cellar

The transit cellar is responsible for the reception, storage, dispatch, processing, sampling and analysis of bulk wine. The transit cellar can be part of the filler/packer company (geographically separated or not) or can be outsourced. The transit cellar receives bulk wine from bulk distributors in different kinds of containers. During the transit cellar stage, the wine is prepared for onward sale and filling. It is loaded for transit to the customer and is accompanied by all the appropriate documents.

Filler/Packer

Fillers/Packers are responsible forth reception, analysis, filling, packing and dispatch of finished goods. The Filler/Packer receives containers of bulk wine from the wine producer, and then the wine is filled into different kinds of packages. Consumer units, such as bottles, bag-in-box, tetra packs, etc. are produced from The wine batches supplied.

Freight Forwarder

This actor organizes the shipment planning, which is the process of choosing shipment frequencies and deciding for each shipment which orders should be assigned. It also includes the safe and efficient movement of goods on behalf of an exporter, importer or another company or person, sometimes including dealing with packing and storage. Typical activities include: researching and planning the most appropriate route for a shipment (taking into account the nature of the goods, cost, transit time and security), arranging appropriate packing (taking into account climate, terrain, weight, nature of goods and cost) and delivering or warehousing of goods at their final destination.

Freight operator

They supply service for transporting goods from the Winery to the Importer or to other actors (distributor, wholesaler, retailer, etc.), by air, through airline services, by sea through shipping lines or by road and rail through different operators. The courier could be an express/parcel carrier trucking company, an ocean liner, a railroad or an air carrier/ integrator

Importer

This actor buys goods from the Wine Producer and is responsible for the reception, storage, inventory management and dispatch of finished goods, which receives from the Freight Forwarder through the Freight Operator. The Importer sales and delivers finished goods to the Wholesaler or Distributor of the destination country depending on the distribution channel used in the country.

Finished goods

distributor

This actor is responsible for the reception, storage, inventory management and dispatch of finished goods, as well as re-packing and re-labeling as per specific customer requirements required

Wholesaler

The Wholesaler receives pallets and cartons from the Finished Goods Distributor and picks and dispatches goods to the retails stores. They put new orders to the Finished Goods Distributor, to the Importer and may also buy directly from the winery.

Retailer

The Retailer receives finished goods from the Finished Goods Distributor or the Wholesaler depending on the distribution channel. The retailer sells consumer units (bottles, cartons) to the end consumer. The different sales’ channels are: hyper/supermarket, liquor stores, drugs stores, specialist store, hotels, restaurants, catering, clubs, etc.

End customer

This is the end customer of the SC. End customers may buy finished goods directly from some wineries, or they can make an indirect order of new products when they go to the store or supermarket and chose some kind of wine.

Table 1. Actors of SC : Enterprise with different functions

Page 6 sur 18

2.2.1. The triad

In a triad each central actor has a different role. Table 2 below provides an illustration of different possible roles.

Triad Example of role

for the central actor in the triad

Freight Operator / Importer / End customer Intermediate

To turn them into its own advantage, dealings between networks members

Freight Forwarder / Freight Operator / Importer Broker

To facilitate or control the flow

Filler/Packer / Finished Good Distributor / Wholesaler Coordinator

To reconcile several options of network members and align goals

Wholesaler / Retailer / End customer Agent

To distribue ressource

Transit / Filler/Packer / Finished Good Distributor Hub

To facilitate or control the flows of supply across the whole network

Table 2. Roles of central actors Being inserted into a network facilitates the solution of problems and reduces costs (transaction, coordination, opportunity...). For example, adhering to standard GS1 (2005) allows each actor in the supply chain to be connected to the upstream and downstream actors by a following up of standard. The GS1 system of numbering and bar coding applies all along the WSC: from the grape grower to the retailer, to facilitate traceability. As an illustration, the grape grower is responsible for the production. Each plot or block of vines is identified with Global Location Number (GLN), which is allocated by the grape grower. After the transformation from juice to wine, the wine producer adds to GLN identification, a product identification (GTIN), a shipping container identification (SSCC) and the quantity of wine dispatched (AI). When the bulk distributor received the wine, he adds a new identification, and so forth. Normative social relationships help bringing favorable factors to the firms performance (in terms of trust, traceability and information transfer), which are not found in the adjustment by the market. 2.2.2. The groups

It brings together actors who maintain relatively dense interlinks, because they are linked in a production chain, processing or services. In the WSC, there are three groups which have functioning, rules and strategies both specific but linked within a complete network: the production group, the distribution on "national market" and the distribution on "international market". The production group includes the following actors (Pretorius & Høj, 2005; standards, 2005; Garcia, 2009): the grape grower, the wine producer, the bulk wine distributor, the transit cellar and the filler/packer. Upstream group also maintains relations with the raw materials supplier and downstream with cellar door sales. However, this group is not the same, nor as complete in all countries. There are several possibilities based either on dyads (wine producer - filler/grape grower-packer or cellar door sales) or triads (grape grower-wine producer - filler/packer).

Fig. 1. Production group2

2 Legend: Arrows on all figures of this paper symbolize the chosen unit of analysis, i.e. the "contractual relationship".

Page 7 sur 18

According to Garcia (2009), "from the point of view of the relationship between order arrival and production release, production systems can be classified into make to stock / by to stock or make to order / buy to order […]. And for the packing several solutions: bottle to order, label to order, pack to order, deliver to order" (Garcia, 2009, p. 42). The distribution on "national market" includes the following actors (Min & Zhou, 2002; GS1 standards, 2005; Garcia, 2012): the finished good distributor, the wholesaler and the retailer. Upstream the group also maintains relations with the filler/packer and downstream with the end customer.

Fig. 2. Distribution on national market This modeling of the distribution group on "the national market", is having variations between countries, that range from the dyad (distributor – end customer or retailer - end customer or wholesaler - end customer or cellar door sales - end customer) to the triad (distributor - retailer - customer or end distributor - wholesaler - end customer or wholesaler - retailer - end customer). The distribution on "international market" includes the following actors (Min & Zhou, 2002; Garcia et al., 2012): the freight forwarder, the freight operator and the importer. Upstream the group also maintains relations with the filler/packer and downstream with the end customer.

Fig. 3. Distribution on international market In case of wines distributed on the international market, the wine is distributed through the importer. Several variants are possible, for example, the freight forwarder may be conventional freight forwarder or LSP (LLP or 4PL). In all cases all are match-makers. The importer buys wine from different wineries and sells it using the system with 3 levels of service, which is composed of finished goods distributor, wholesaler and retailer. The producer uses the service to coordinate freight shipping. Forwarder, in turn, uses the services of a freight operator for wine shipping. The match-makers do not trade. They relate both buyers and sellers. They are recruited to facilitate the transaction when the cost of information search and transaction is too high for partners. These intermediaries (middlemen) through their network allow reducing costs. Freight forwarder may be conventional or market-makers forwarders. In the latter case market-makers buy and / or sell for their own account (from selling price and offer price they set call). 2.2.3. Complete network

It is a network which expresses itself in terms of: nodes representing the different actors of the WSC, contractual relationship constituting our unit analysis and trust coordination links and norms. Strong ties are representative of local integration, while weak ties provide the overall network integration. The morphology of this network expresses structure contextual relations, inseparable of the strategic behavior of each actor and the role of each individual within this structure. The role is thus a complex construct that is similar to the concept of status.

Page 8 sur 18

Fig. 4. Complete network Thus, the filler/packer can be seen as a "bridge" between the three groups. It is a point of passage between two groups who are or have themselves access to different resources. Burt (2005) has shown that this position of "intermediary" can bring to the concerned party two types of profits or profit opportunities. The benefits of control, on the one hand, which runs between two groups of contacts and benefits in information and communication, on the other hand.

3. Research design

The research we are conducting is supported and funded by the EcosSud France/Argentina program. It is a program of scientific and academic cooperation with the Spanish-speaking America. It depends on the Department of Foreign Affairs and the Ministry of National Education, higher education and research (http://www.univ-paris13.fr/cofecub-ecos/ecos-sud). Our study focuses on the South-East region of France and the region of Mendoza, Argentina. They are both renowned worldwide for the quality of their wine. The production and exports of wines from these two regions are also very important. Regarding the retained search protocol, it is composed of a variety of actors, many semi-structured interview guides and literature studies (internal and external). The same research protocol was followed in Argentina and France. As a first step, the Protocol has been tested and applied in Argentina. The data have been published in Garcia (2009) and Garcia et al. (2012). In a second step, the Protocol has been applied identically in France (in 2014). The time difference between the first step and the second has been sidelined by a return to the field with six new case studies. The results can then be compared. We also examine two importers. Selection of the companies: from national and international database we identified companies that correspond to the activities that we are considering, namely: grapes growing (0121Z), wine-making (1102B) and production of wines, including sparkling wines (1102A). And, among these companies, we selected only those who exercise a part of their activity abroad. Regarding Mendoza Argentina (MA) case, as per INV database (2013), we identified 1280 Argentine companies corresponding to the three codes of the International Classification of Economic Activities (NACE 0121Z, 1102A, 1102B) among which 1047 are active internationally, and 998 are in the Mendoza region (4 zones). The great South-East of France (SEF) encompasses three wine regions (Languedoc-Roussillon, Provence-Alpes-Côte d'Azur and Rhône-Alpes). From the DIANE database, we identified 1486 French firms corresponding to the three codes of the French Activities nomenclature (0121Z, 1102A, 1102B) whose 355 are active internationally. Among them, 73 are located in the greater SEF. The table 3 below summarizes the process of selection of companies and selected case studies.

Selection of the companies

Argentina French

Database used INV DIANE

Wine regions selected

- High Zone of the Mendoza River - North Zone - East Zone - Valle de Uco

- Languedoc-Roussillon - Provence - Vallée du Rhône

Page 9 sur 18

Criteria for identifying companies

Classification of Economic Activities (NAC) - 0121Z: grapes growing - 1102A: manufacture of sparkling wines - 1102B: wine-making * International business

* Classification of Economic Activities (NAF rev. 2) - 0121Z: grapes growing - 1102A: manufacture of sparkling wines - 1102B: wine-making * International business

Number of companies involved

998 73

Studies websites of these companies

Case studies 6 companies

(see information1) 6 companies: 2 per wine region

(see information2)

Case studies 2 importers (see information 3)

Table 3. Process of selection of companies

Information 1. Case studies in Argentina

Argentina – Region of Mendoza

General information Winery 1 Winery 2 Winery 3 Winery 4 Winery 5 Winery 6

Real Annual Production (hectoliter)

1.000.000 12.000.000 3.750.000 840.000 6.000.000 1.000.000

Main Production Strategy

Label to order

Label to order

Bottle to order

Make to stock and label to order (second label)

Make to stock (bottle+ label+packing)

Bottle to order

Percentage of production for exportation

60 45 60 90 60 98

Nb of customers - international

60 60 40 25 70 25

Most important destination country

Mexico, Brazil, USA,

China

USA, Holland, Canada

Brazil USA UK, Canada,

Mexico USA, Brazil

Resources utilization

Filling line capacity (bottle/hour)

2000 12.000 4500 2500 30.000 2500

Wine aging in bottle capacity used (bottles)

300.000 60.000 352.500 600.000 - 40 000

finished products warehouse capacity (bottles)

72.000 1.200.000 660.000 600.000 - 92.400

Source : adapted to Garcia et al., 2012.

Page 10 sur 18

Information2. Case studies in France

General information Winery 1 Winery 2 Winery 3 Winery 4 Winery 5 Winery 6

Real Annual Production (hectoliter)

80 000 52 000 6 000 3 000 35 000 31 244

Nb bottle/year 8 350 000 6 930 000 500 000 61 000 4 700 000 3 765 000

Main Production Strategy

Bottle to order

Bottle to order

Make to stock and

label to order

(second label)

Bottle to order

Bottle to order

Make to stock and

label to order

(second label)

Percentage of production for exportation

16 15 34 75 16 16

Nb of country 20 17 70 80 7 3

Most important destination country

Switzerland,

Belgium, Scandinav

ia, USA

England, other

European countries,

USA, Canada,

Brazil

All Europe, USA,

Canada, Mexico,

Hong Kong, United Arab Emirates...

Europe, Latin

America, South Africa

Denmark, Ireland,

Germany, Belgium,

Netherlands, Switzerland, New Zealand

Brazil, Hong Kong,

Germany

Resources utilization

Filling line capacity (bottle/hour)

2000B/h 4500B/h 3000B/h 3000B/h 4000B/h 7200B/h

Wine aging in bottle capacity used (bottles)

15 000 12 000 1200 800 3500 20 000

finished products warehouse capacity (bottles)

1 300 000 600 000 20 000 5 000 253 000 300 000

Information 3. Importers

Importer 1 Importer 2

Turnover in 2013 30 millions/R$ 40 millions/R$

Importer Nationality Brazil Brazil

Main provider countries Germany, Argentina, Australia, Chili, Spain, France, Italy, Portugal, Uruguay, South Africa.

South Africa, Germany, Argentina, Australia, Austria, Chili, Croatia, Slovenia, Spain, USA, France, Greece, Hungary, Italy, New Zealand, Portugual, Uruguay

Main buyer countries Brazil Brazil

Sales volume Brazil 1,8 millions/bottles 2,5 millions/Botlles

Main purchase mode Producer Caves

Producer Caves

Incoterm Ex Works Ex Works

Importation way Sea Sea

Port in France Le Havre Le Havre

Port in Argentina Buenos Aires Buenos Aires

Port in Brazil Santos – Sao Paulo Santos – Sao Paulo

Domestic transportation in Brazil

Road Transport - Service providers

Road Transport - Own carriers

Page 11 sur 18

4. Results and discussion The results of our investigations allow to present the WSC in Argentina (Mendoza) and France (South-East) compared with a summary of which is given below (see Figures 5-7 and information 4) theoretical model.

Fig. 5. Theoretical model: the complete network of WSC

Fig. 6. MA model: Mendoza Argentina network of WSC

Fig. 7. SEF model: South East of France network of WSC

Page 12 sur 18

Information4. Clustering and centrality

Theorical model WSC Argentina WSC French WSC

0,343 0,188 0,480

The weighted overall clustering coefficient. It is the weighted mean of the clustering coefficient of all the actors each

one weighted by its degree.

The clustering coefficient of an actor is the density of its open neighborhood. It is the number of pairs of neighbors connected by edges number of pairs of neighbors

Degree Centrality= It is the maximum of percentage of centrality that you can have. Closeness Degree centrality = high score is more central. Closeness is calculated by taking the average of all paths or all trails and is more realistic. Betweenness = it has typical a high level of variance. Eigenvector = power, information, influence are similar centrality measures and count the number of walks. Short link is weighted more heavily. They basically measure influence.

The theoretical model (figure 5) emphasizes the existence of three main actors: wine producer, filler/packer and finish goods distributors (cf. information 4). In contrast, the AM model emphasizes only one actor (wine producer) and SEF model two actors: cooperative winery and grape grower. In France as in Argentina grape grower and wine producer are estate wineries: private domains. MA and SEF models are more integrated in their production network that the theoretical model, but relatively equivalent at the international distribution network. As regards the national distribution network model SEF is the most different from the theoretical model. The production network (figures 6 & 7), in MA and SEF models is composed of a central hub which facilitates or controls the flows of supply across the whole national and international network. This interface area takes advantage from its spatial position and its communication infrastructure. In MA model (figure 6), the wine producer is the central hub and it integrates several roles. The winery produces grapes from its own vineyards and it has the same facility so it covers grape grower, bulk distributor, transit cellar, and filler/packer role. Wine producer have their own vineyards and produces only through these vineyards, in order to maintain the quality of wines (Garcia, 2009). In some cases, grapes are bought to other grape growers. In these cases, grape growers are in relations with wineries to define the quantity and variety of grapes needed for the production. The harvest is made once a year and it has to be done in the right time when the grapes are in their optimal maturity (Ferrer et al., 2008).In Mendoza, private domains own parcels of vineyards in different areas of the region, locating at different altitudes and offering different varieties. So, once the harvest completed, grapes are transported to wineries using an internal or external means of transport. Using the same basic process, transforming juice to wine may require numerous steps. Many procedures and operations are performed. Depending on the type of wine that is going to be produced the process and the number of steps are different. In the SEF model (figure 7), two actors play the role of a central hub: the grape grower and the cooperative winery (in the South-East of France cooperative wineries and private domains are divided territory in equal proportion):

Page 13 sur 18

The grape Growers are often domains very famous. They have their own vineyards and produces only through these vineyards. Their private domains are concentrated in the same location. They integrate several production functions: wine producer, bulk wine distributor, transit cellar, filler/packer and finish goods distributor.

The grape grower is the most important node within the WSC because the final quality of the wine is directly related to the quality of the grapes (Ferrer et al., 2008). Grape growers and cooperative winery can also export grapes to foreign countries and they can sell grapes directly to end customer.

The cooperative winery includes a large group of grape growers (20 to 300 members). It integrates several

production functions: wine producer, bulk wine distributor, transit cellar and filler/packer. The organization and equipment of the cooperative also influences wine quality (several distribution factors are important): the barrels where wine is aged, the winemaking method, the wine maker, the winery equipment, and the bottling plan (Adamo, 2004). The cooperative winery receives grapes (could be already crashed) from different and various grape growers. Therefore, it can receive different qualities of grapes. It cannot refuse to buy grapes delivered by a cooperator-adhering. But, to maintain the best quality, the French cooperative winery sets up several rules: it varied prices depending on the quality, it imposes strict specifications loads and chooses a closed structure i.e. it won’t accept all grape growers of the region. Most cooperative wineries make very large volumes of wine compared to specific areas. In SEF model, cooperatives wineries have a very important place in the WSC.A cooperative winery is a specific company founded and managed by growers to ensure in common: the use of production tools, but also packaging, storage, marketing or processing of the products of their farms and their supply of fertilizers and other inputs. It works on the principle that all growers are members of the cooperative and therefore, it belongs to them. In the SEF model, cooperative wineries are a tradition. There are as much independent cooperative wineries as wine producers in Mendoza. Mendoza has 29 cooperatives (which include 5000 grape growers) for 16,000 grapes growers. Many old companies are located in Mendoza and the South East. Three quarters of the companies of the Southeast were created during the 90s. In AM and SEF models, all production functions are performed by: the cooperative winery, the grape grower or the wine producer. Therefore there is no middleman and control is global. This structure eases continuous improvement initiatives. It focuses on the quality, safety, health, environment and sustainable development. The functioning of the organization is validated and recognized certifications such as :

ISO 9001 (quality certification) (SEF et MA models),

ISO 14001 (environmental certification) (SEF et MA models),

IFS or International Food Standard (referential of food safety) (SEF model),

BRC or British Retail Consortium (referential of food safety) (SEF model),

VDD or Vignerons en Développement Durable (SEF model),

HCCP or Hazard Analysis Critical Control Point(AM model),

ASPBA or Argentina Sustainability Protocol by Bodegas de Argentina (AM model).

In the interests of sustainability of their heritage, grape growers of SEF model and wine producers of MA model take care of their environment. They develop organic part of their vineyards and are focused on sustainable development including environmental, social and economic aspects:

Environmental aspects (SEF and MA models): soils conservation, reduction of water and fuel consumption, systematic waste sorting and recycling, reduction in pollutant emissions, carbon footprint, actions fostering biodiversity and upkeep of the countryside, practicing rainwater collection…,

Socio-economic aspects (SEF model): setup rooftop photovoltaic panels to improve energy cost and limits the carbon footprint, lighter bottle, collecting and recycling cork or plastic stopper, computerized vineyard owner specification and intranet use,

Socio-economic aspects (MA model): the protocol of sustainability has a strong humanistic approach, encouraging the provision of social benefits for employees and child-care and nutrition for the families of seasonal workers. For example, one of Argentina winery has a permanent workforce of 530 people. To facilitate the retention of families working in the cellar, and given the scarcity of places of learning, the company built the Cultural and Recreational Center where social activities and staff training and develop their families.

In general, sustainable agriculture from SEF wine-producers is based on an internal charter inspire by the ISO 14001 certification and the “sustainable agriculture” reference document. For the Argentinean wine-producers, they use the Argentina Sustainability Protocol by Bodegas de Argentina. Considering the national market network, MA network uses the finish goods distributor as coordinator (figure 6). He respects the specific role of each. In contrast, in the SEF network is the cooperative winery or the grape grower who assumes the role of coordinator in place of finish goods distributor. The network is more limited. The coordinator proposes several options for wholesalers and retailers and aligns the goals of each of them. He is

Page 14 sur 18

responsible for ensuring balanced relationship between actors of the network. He knows perfectly members’ needs and interests in order to ensure the operation of an effective network. The coordinator organizes activities to satisfy the end customer. He provides the support necessary to build relationships and weaves the network by actively supporting connections between network members and initiatives. In MA model (figure 6), the finish goods distributor is the most important actor between wine producer and the end customer. He is in relations with most of the wholesalers and retailers. He is in charge of the reception, storage, inventory management and dispatch of finished goods, as well as re-packing and re-labeling as required (GS1, 2005). He may also re-pack and re-label the products as per specific customer requirements. The Wholesaler receives pallets and cardboard boxes from the finished goods distributor and picks and dispatches goods to the retails stores. Regarding to the main production strategies, none of the twelve studied wineries keep the "(re)pack to order", in which the wine is already bottled and labeled, and when the order comes in the only thing to do is to pack the bottles and deliver them. In the same vein, only one winery (winery 5 in the Argentina case) seems to keep a production on stock, in other words in a "push" mode; in this case bottles are already filled, labeled, covered, packed and ready to be picked and delivered. Instead, a production strategy of bottles in a "pull" mode appears to be preferred by the five others Argentine wineries and the six French wineries studied. Indeed, the "bottle to order" strategy (including the filling, covering, labeling, packing, preparation and delivering of the order when it comes in) is adopted by two Argentine wineries (3 and 6) and by four French wineries (1, 2, 3 and 5). Concerning the "label to order" strategy (in this case the wine is already bottled and covered en mode "push ", but when the order comes in, the labeling and packing processes begin and finished goods are delivered en mode "pull"), it is being implemented by the wineries 1, 2 and 4 in Argentina and the wineries 3 and 6 in France. The two predominant strategies, "bottle to order" and "label to order", reflect the will of the wine producers, Argentine as French, to propose products in adequacy with the client’s expectations depending on their actual purchase orders and their current needs. For this purpose, they prefer a manufacturing process close of postponement or mass-customization, well-known process in the automotive or IT industry. This one allows to link up economies of scale objectives embodied by a mass-production upstream the supply chain (especially concerning the winemaking) with, downstream of the supply chain, a greater adaptation to demand by proposing differentiated product to customers and finalized depending on their purchase order. In SEF model, it is the grape owner or the cooperative that endorses the role of finish goods distributor. Due to the country size, French distribution network is shorter than Argentina one’s. Yet, French wine industry actors are active in the entire distribution channel: large retailers, online sale, direct sales in the cellar, hotels, exportation and direct sale to the end customer. They are also engage actively on online social network, in particular Facebook or Twitter. They are personally committed to distribution in addition of intermediaries. They also try to diversify their activities by offering regularly new wine combinations. Thus, end customers may buy finished goods directly from some grape growers, cooperative wineries and by Internet. Instead, the wine producer on the MA model do not use Internet website. They are dealing with specialized actors representing various Argentine wines. "La Bodega de Argentina"3 promoting wine on a large national scale whereas Argentine wine associations promoting it outside the country. The wine producer does a lot of television advertising and outdoor advertising instead of the SEF companies. Argentina targets young people with television and outdoor advertising. Wine, as soda, is sometimes sold in can. Argentina also constantly works on labels by using English consultant. The differences among the marketing approach can be explained by the input code differences between the old world and the new world. Wines of the old world are built on blends and terroir whereas input codes from the new world are built on grape variety (Dubois, 2013).In Mendoza, each grape variety grows at a given altitude. Altitude is the "Argentine terroir" since soils, temperature variability and irrigation are different. The wine producer in Mendoza uses it as a powerful marketing tool. For example, one of the Argentina wine producer, the Chardonnay, the white Sauvignon and the Merlot are planted at 1.000m and the Malbec at 1.300m… By contrast, in the Southeast France, companies are renowned for their blends. There is an adequacy between grape variety used on terroir and their assembly procedure with specific wine making modes (Dubois, 2013). Blends skillfully shuffle wines from different parcels and grape variety. This milestone gives the flavor and all the specific characteristics of each wine. For instance, across the French case studies here is a distinctive blend: for red grape variety (30% Grenache, 30% Syrah, 30% Mourvèdre, 10% Cabernet Sauvignon) and for white grape variety (35% Marsanne, 35% Roussane, 30% Grenache Blanc).

3 It’s an organization regrouping more than 200 cellars in Argentina. It promotes Argentine wines and serves the interests of producers. It is an association exclusively financed by producers. 45% of the collected funds are use for wine promotion.

Page 15 sur 18

Considering the international market network (figures 6 &7), the wine is frequently distributed to external market through importers. Importers buy wine from different wine producers, cooperative wineries or grapes growers and sell it using the 3tiers system (AM model) and 2 tiers system (SEF model) for delivering, which is composed of a freight operator or a freight forwarder to coordinate the shipment. For the external market (AM model), the wine producer "sends shipments through the freight operator using Valparaiso’s port or Buenos Aires’s port, and sometimes trans-shipments are made in a port in the middle of the route. The wine producer sells following the FOB (Free On Board) strategy which, considering the definition from the Incoterms, means that the winery sells and delivers the finished goods to the ship in the port of boarding, which will be transported to the destination country" (Garcia, 2009). Sometimes, for the distribution of the finished goods for overseas market, the wine producer works with a freight forwarder, who manages all the movements of goods using the services of shipping lines as freight operators" (Garcia, 2009). In AM model, the freight operator can be directly in relation with the wine producers when it exports grape juice and bulk wine and with the finish good distributor when he exports bottles. When the transaction is more complex to organize, a complementary actor comes into play: the freight forwarder. Sometimes the freight forwarder provides itself the physical transportation service of goods to destination countries (SEF model). The importer is responsible for the reception, storage, inventory management and dispatch of finished goods. Through the comparison between the importation processes of two Brazilian importers (Tables Information 3), we observe that MA model and SEF model use the same actors for exporting to Brazil. 1 producer: make the product available for shipments. 2 the freight operator: importers use subcontractors for the transporting towards Brazil, the sea transportation is used, as wine is considered high-value product stringent transport conditions shall be met, to ensure product integrity until final delivery to the buyer country. The busiest harbours for exporting are: Le Havre, France and Buenos Aires, Argentina. In Brazil, Santos is the most used. 3 Importer: receives the products for storage and distribution across the internal market. Importer 1 uses service providers and importer 2 uses its own fleet. 4 End Costumer: Supermarkets, wine reseller, restaurants, etc. It shall be considered that importations from France and Argentina differ in terms of lead-time, regulations, taxes and import volume. As per UVIBRA (2013) Brazil imported 3.2 million bottles from France and 12.7 million bottles from Argentina in 2013. In this case Argentina has a commercial and logistic asset over France. In AM model, the freight operator has a broker role, as the freight forwarder in SEF model. A broker is an actor who can mobilize its social capital by acting as a "broker" of information and complex flows between two actors (importer/producer) , thus providing access to new opportunities. This broker also reaps control benefits of being the facilitator of information and complex flows between contacts. The freight forwarder organizes the shipment planning (choosing shipment frequencies, the assignation of each shipment and the safety). He arranges the most appropriate means of transport, using the services of shipping lines, airlines or road and rail freight operators. It organizes the link between the different carriers and ensures the continuity of transport and all related administrative operations related thereto, customs regulations, administrative, and financial, insurance etc. In some cases, the freight forwarding company itself provides the service. In this case, the freight forwarder is also the freight operator. He supplies service for transporting goods. In France, the two main utilize ports to export wine are the port of Le Havre, primarily used by Bordeaux and Bourgogne wine producers, and the port of Marseille, mainly used by Provence-Alpes-Côte d’Azur wine producers. The Atlantic Port of Bordeaux endure a significant decline since several years explained, according to freight forwarders, by the deterioration of transportation offer of the port and by a strong competition from the container ports (Guerrero, 2010). The choice of the port to export wine is not necessarily a matter geographic proximity with the production area but is more relative to the attributes of the port in term of access to worldwide foreland, minimal access time, even if the handling cost is more important, and to pre-existing customs agreement within country when the choice of the port is located outside the French border (essentially the port of Antwerp and Rotterdam). For both AM and SEF models the quality of the wine as well as the distribution to the end consumer in the destination country is of primary importance. If problems occur during transportation, it is the wine producer, the grapes grower or the cooperative winery that should respond to the customer. AM model is growing stringent efforts are made to export, in particular towards neighboring countries (Brazil, Mexico…) and the North American market where it takes advantage of the large Hispanic community to commercialize their products. Argentines have an important promoting policy for exportation and support wine producer to sell their wines abroad. According to Giuliani et al. (2011), Argentina has made many innovations that have allowed producers to be in tune with the demands of the international market. SEF Model mainly exporting throughout Europe, North and South America and Asian countries. France is the leading country concerning the wine exportation with 24,1% market share but suffer a slight decrease from 2007 (-2,4 points) for a total of 14,5 millions hectoliters in 2013. The United State is the first importer of French wine with 17,1% market share follows by the United Kingdom, China,

Page 16 sur 18

Germany and Singapore. Brazil is the main market target over the coming years for the South America region (Ubifrance, 2013). Argentina is ranked among the top ten wine exporters in the world, from 2010 to 2012, the country wine exportations increased by over 47% with a total of 3.2 million hectoliters in 2013, the USA is the largest importer of Argentinian wine with 23,9% market share, followed by the UK, Canada, Brazil and Russia. Brazil increased wine imports by 34,4% in 2012. Argentina is the second largest wines exporter to Brazil with a market growth of 32,86% in 2012 (Ibravin, 2013). The results presented above could be more detailed about the practices of companies interviewed. However, we intentionally limited the results to the elements for identifying social ties between different actors. Indeed, for modeling the WSC, we used only one unit of analysis: the "contractual relationship". These results simplify the wine SC representation by showing only the actors in a generic form. We seek to reflect the diversity of actors and not their number. Then, each actor seen as community enables us to underline the social relations between the actors and their specific role. The contractual and social relations represent the strength that binds the WSC members, which belong to a social network. This strength may change over time and space. These links ensure actors integration, social cohesion and to acquire a social identity. Modeling the WSC can contribute to appreciate the sensitive points for a better control (overall planning framework and cost efficiency).On this regards, the actions of the SEF and AM are significant sign: traceability, carbon footprint, use of ICT to control the company and exchange information, centralization of logistics processes at one site reducing delays between order and shipment, integration of some facets of sustainable development linked to respect of the product. We have identified relevant actors that have an essential intermediary role in the WSC network. Their clustering rate is high and a strong degree of centrality is observable (see information 5). Markowsky et al. (1988) show a positive link between centrality and power in networks. For Degenne & Forsé (1994) “actors are central if they are strongly connected to the other members of the network, they are peripheral if the connection is weakly”. As such, the wine producer in the MA model and the cooperative winery and the grape owner in the SEF model possess a high capacity to coordinate activities in the SC. In the theoretical model, four actors are essential in the WSC: grape grower, wine producer, filler/packer, and finished good distributor. In the AM model, four actors are also essential in the WSC, yet they are different actors: wine producer, finished good distributor, freight operator and importer. In the SEF model only three actors are essential: grape grower, cooperative winery and end customer. The structuring difference indicates some WSC steering distinction. The theoretical model focuses on the various stages of the production process, central to the quality traceability. In the AM model, priority is given to a more global integration (production/national market/international market). The production process is highly concentrated, facilitating traceability and coordination of each stage. There is only one actor: the wine producer. It has a significant central hub role to play. There is no triad. To some extent, it looks like an ego network. Relating to the distribution on the different markets, the national market and international market are controlled by the finish good distributor. For example, it is the role played by « Bodega de Argentina ». It is a coordinator that controls the distribution on the national and international market via the freight operator. The latter is also a broker in relation with the importer. For SEF model a high value is placed on the proximity with the end consumer. It is the most integrated WSC. The production is on the hands of the grape owner or in the hands of the cooperative winery as a central hub. The distribution process on the international market is operated by a broker, in the traditional sense, i.e. the freight forwarder that manages the complex contractual relations or far-off destinations. Otherwise they directly work with the importer. The SC internalization through direct on-line business, from the producer vineyards or from the cooperative, between the producer and the end consumer gradually becomes an integral part of the strategy of the wine-producing companies and their customer service. It constrains to manage a relational and organizational back-office (transport, stocks, predicting sales…). Several logistic models coexist according to the type of wines.

5. Conclusion In this article, we presented the various approaches for understanding the complexity of international wine trade, as well as the reasons why this market is considered as complex. We first reviewed the main works modeling generic SC, then established a model of WSC through the social networks analysis (SNA). Comparison studies between the industries of the South East of France (SEF) and Mendoza Argentina (MA) were performed. We use the SNA modeling to analyze, structure the network and map out the relationships between the group of actors in WSC. Three levels of analysis of social networks were used: triads, groups and complete network. All actors in the chain have been identified and classified based on their contractual relationships with each other. Those actors then regroup in communities, as a chain links with each a role to play. We performed a comparison between the processes of production, national and international distributions and sustainability between the theoretical model and the models of each country (SEF and MA). The results of the comparisons between the three models were

Page 17 sur 18

showed and explained using the description of each investigation throughout this study.

The contributions of this work include obtaining a new structure and vision of modeling of WSC through the approach of social networks. This model will help industries to identify inefficient and efficient processes in their chain. Through this model and proposed framework, wine industries can improve the understanding of worldwide chain where they are inserted, improving their relationships, strategies and objectives. We suggest for future works the following studies to be developed: benchmarking through a comparative study of other countries studying in more detail the practices of the companies interviewed. The realization of a correlation study between the model presented and the process optimization of SC. The use of this model as standard in other industries of the sector and food and beverage.

6. References Adamo, C. (2004). A global perspective of the wine supply chain: the case of Argentinean wineries and the US market,

Doctoral dissertation, Massachusetts Institute of Technology. Balasundaram, B., Butenko, S., & Hicks, I. V. (2011). Clique relaxations in social network analysis: The maximum

k-plex problem. Operations Research, 59(1), pp. 133-142. Borgatti, S. P., & Li, X. (2009). On Social Network Analysis in a Supply Chain Context. Journal of Supply Chain

Management, 45(2), pp. 5-22. Burt, R., S. (2005). Brokerage and Closure: An Introduction to Social Capital, Oxford University Press, New York. Caddy, I., & Helou, M. (1999, November). Supply chain and supply chain management: towards a theoretical

foundation. In Conference Proceedings, Second International Conference on Managing Enterprises, Newcastle, pp. 17-20.

Caddy, I., & Helou, M. (2007). Supply chains and their management: Application of general systems theory. Journal of Retailing and Consumer Services, 14(5), pp. 319-327.

Carter, C. R., Ellram, L. M., & Tate, W. (2007).The use of social network analysis in logistics research. Journal of Business Logistics, 28(1), pp. 137-168.

Chen, I. J., & Paulraj, A. (2004). Towards a theory of supply chain management: the constructs and measurements. Journal of operations management, 22(2), pp. 119-150.

Cooper, M. C., Lambert, D. M., & Pagh, J. D. (1997). Supply chain management: more than a new name for logistics. International Journal of Logistics Management,8(1), pp. 1-14.

Degenne, A., & Forsé, M. (1994). Les réseaux sociaux, Armand Colin, Paris. Demossier, M. (2011). Beyond terroir: territorial construction, hegemonic discourses, and French wine culture.

Journal of the Royal Anthropological Institute, 17(4), pp. 685-705. deNooy, W., Mrvar, A., & Batagelj, V. (Eds.). (2005). Exploratory social network analysis with Pajek, Vol. 27. Cambridge

University Press. Dubois S. (2013). Lecture géopolitique d'un produit alimentaire mondialisé : le vin, Revue internationale et

stratégique, 1(89), pp18 à 29. Ellram, L.M. Tate, W.L. Carter C.R. (2006). Product-process-supply chain: an integrative approach to three-

dimensional concurrent engineering.International Journal of Physical Distribution & Logistics Management, 37(4), pp. 305–330.

Fabbe-Costes, N., Fulconis, F., & Koulikoff-Souviron, M. (2013). "La recherche en management logistique : la question des unités d’analyse et du périmètre", Chapitre 5, in N. Fabbe-Costes & G. Paché (Coord.), La logistique : une approche innovante des organisations, PUP-Presses Universitaires de Provence, Aix-en-Provence, pp. 49-60.

Ferrer, J. C., Mac Cawley, A., Maturana, S., Toloza, S., & Vera, J. (2008). An optimization approach for scheduling wine grape harvest operations.International Journal of Production Economics,112(2), pp. 985-999.

Fernández‐Olmos, M., Rosell‐Martínez, J., & Espitia‐Escuer, M. A. (2009). Vertical integration in the wine industry: a transaction costs analysis on the Rioja DOCa. Agribusiness, 25(2), pp. 231-250.

Galaskiewicz, J. (2011). Studying supply chains from a social network perspective. Journal of Supply Chain Management, 47(1), pp. 4-8.

Garcia, F. A. (2009), Modelado y Medición de Performance Logística en la Industria del Vino, Maestrià en logistica, Universidad National de Cyuo, Mendoza.

Garcia, F. A., Marchetta, M. G., Camargo, M., Morel, L., & Forradellas, R. Q. (2012).A framework for measuring logistics performance in the wine industry.International Journal of Production Economics, 135(1), pp. 284-298.

Giuliani, E., Morrison, A., & Rabellotti, R. (Eds.). (2011). Innovation and technological catch-up: the changing geography of wine production. Edward Elgar Publishing.

Granovetter, M. (1985). Economic action and social structure: the problem of embeddedness, The American Journal of Sociology, 91(3), November, pp. 481-510.

Guerrero, D. (2010). Les aires d'influence des ports de la France: entre réseau et gravitation. Thèse pour l’obtention du grade de Docteur en Géographie. Université Paris-Diderot - Paris VII, Paris.

Page 18 sur 18

GS1 (2005), "wine supply chain traceability.GS1 application guideline". www.GS1.org. Handfield, R. B., & Nichols, E. L. (1999).Introduction to supply chain management (Vol. 1). Upper Saddle River, NJ:

prentice Hall. Ibravin – Instituto Brasileiro do vinho. http://www.ibravin.org.br/dados-estatisticos. INV- Instituto Nacional de Vitivinicultura (2013).

http://www.inv.gov.ar/PDF/Estadisticas/Exportaciones/2014/ComparativoHLFOB.pdf Kim, Y., Choi, T. Y., Yan, T., & Dooley, K. (2011). Structural investigation of supply networks: A social network

analysis approach. Journal of Operations Management, 29(3), pp. 194-211. Knoke, D., & Yang, S. (2008). Social network analysis, Vol. 154. Sage. Lambert, D. M., Cooper, M. C., &Pagh, J. D. (1998). Supply chain management: implementation issues and research

opportunities. International Journal of Logistics Management,9(2), pp. 1-20. Lazzarini, S. G., Chaddad, F. R., & Cook, M. L. (2001). Integrating supply chain and network analyses: the study of

netchains. Journal on chain and network science, 1(1), pp. 7-22. Liu, L. (2011). Research on the Management System of enterprises using Modern Logistics Supply Chain Theory.

Procedia Engineering, 24, pp. 721-725. Markowsky, B., Willer, D., & Patton, T. (1988). Power relations in exchange networks. American sociological review,

53(2), pp. 220-236. Mariani, A., Pomarici, E., & Boatto, V. (2012). The international wine trade: Recent trends and critical issues. Wine

Economics and Policy, 1(1), pp. 24-40. Min, H., & Zhou, G. (2002). Supply chain modeling: past, present and future.Computers& Industrial

Engineering,43(1), pp. 231-249. Mueller, R. A., Buergelt, D., & Seidel-Lass, L. (2007, February). Supply chains and social network analysis. In 1st

International European forum on innovation and system dynamics in food networks, Sage, pp. 1-12. Newman, M. E., & Park, J. (2003). Why social networks are different from other types of networks. Physical Review

E, 68(3). Observatorio Vitivinicola Argentino (2014) http://www.observatoriova.com/2014/08/mapa-de-

establecimientos-vitivinicolas-en-argentina/ Opsahl, T. (2013). Triadic closure in two-mode networks: Redefining the global and local clustering coefficients.

Social Networks, 35(2), pp. 159-167. Pretorius, I. S., & Høj, P. B. (2005). Grape and wine biotechnology: challenges, opportunities and potential

benefits. Australian journal of grape and wine research, 11(2), pp. 83-108. Snijders, T. A. (2001). The statistical evaluation of social network dynamics. Sociological methodology, 31(1), pp. 361-

395. Stevenson, W.J. (1999). Production/Operations Management. 6th Ed., McGraw-Hill, Boston. Supply Chain Council (2002). http://www.supply-chain.org Ubifrance (2013). Fiche Secteur Vins et Spiritueux 2013. http://www.ubifrance.fr/vins-spiritueux-

boissons/userdocument.axd?id=001b1305737a. Uvibra (2013) "Uniao brasileira de viticultura".

http://www.uvibra.com.br/pdf/import_vinhos_espumantes_2008_mai2014.pdf Von Bertalanffy, L. (1993). General system theory. George Braziller. Traduction française: Théorie générale des systèmes,

Dunod, Paris. Wasserman, S. (1994). Social network analysis: Methods and applications (Vol. 8). Cambridge University Press,

Cambridge. Yourdon, E. Modern structured analysis. 1989. Englewood Cliffs, New Jersey, pp. 78-94. Acknowledgements This work was supported by cooperation program ECOS-sud/MINCyT that provided the initial framework for the collaboration between University of Nice Sophia-Antipolis – GREDEG CNRS, France and Universidad Nacional de Cuyo – Mendoza, Argentina (Project code N° A13H02). Special thanks to Brazilian importers for providing the necessary data for this research.