Systemic Sustainability of the French Organic Rice and PGI Einkorn Value Chains: A Preliminary Assessment Based on Network Analysis
Abstract
:1. Introduction
2. Conceptual Background
2.1. Systemic Sustainability of Food Value Chains
2.2. Using Network and Mapping Analysis to Assess the Systemic Sustainability of Food Value Chains
2.2.1. The Contribution of Network Analysis to Value Chain Assessment
2.2.2. Value Chain and Network Governance
2.2.3. Survival Capacity of the Network
2.3. Operationalisation of Systemic Sustainability
3. Methodology
3.1. Source of Data
3.2. Information on the History and Organization of Chains
3.3. Information on Sustainability According to the Traditional Pillars
3.4. Information on Chain Governance and Survival Capacity
- Betweenness centrality:where gjk is the number of geodesics connecting jk, and gjk(ni) the number of geodesics that actor i is on.
- Clustering coefficient:where ki is the number of “adjacent actors” of actor i (actors who are directly connected to this actor i), and ni the number of edges between the ki adjacent actors of actor i.
- Distance:
- Density:
4. Results
4.1. A Highly-Centralised and Low-Density Network: The Case of the PGI Einkorn Chain
4.1.1. History and Mapping of Trade
4.1.2. Traditional Pillars of Sustainability
4.1.3. Survivability
Network Characteristics
Value Chain Governance and Survival Capacity
Summary
4.2. A Highly Centralised Organisation within a Diversified Network: The Case of the Camargue Organic Rice Value Chain
4.2.1. History and Mapping of Trade
4.2.2. Traditional Pillars of Sustainability
4.2.3. Survivability
Network Characteristics
Value Chain Governance and Survival Capacity
Summary
5. Discussion and Conclusions
Author Contributions
Funding
Conflicts of Interest
References
- Dufumier, M.; Lallau, B. Agriculture et Développement Durable; Presses Universitaires du Septentrion: Villeneuve d’Ascq, France, 2010; pp. 38–41. [Google Scholar]
- Rangnekar, D. Remaking Place: The Social Construction of a Geographical Indication for Feni. Environ. Plan. A 2011, 43, 2043–2059. [Google Scholar] [CrossRef] [Green Version]
- Raynolds, L.T. Fair Trade: Social Regulation in Global Food Markets. J. Rural Stud. 2012, 28, 276–287. [Google Scholar] [CrossRef]
- Bean, M.; Sharp, J.S. Profiling Alternative Food System Supporters: The Personal and Social Basis of Local and Organic Food Support. Renew. Agric. Food Syst. 2011, 26, 243–254. [Google Scholar] [CrossRef]
- Ministère de l’Agriculture. Ambition Bio 2017, un Nouvel Horizon pour la Bio! Available online: http://agriculture.gouv.fr/chiffres-agriculture-bio (accessed on 10 June 2017).
- Fares, M.; Magrini, M.-B.; Triboulet, P. Transition Agroécologique, Innovation et Effets de Verrouillage: Le Rôle de La Structure Organisationnelle Des Filières. Cah. Agric. 2012, 21, 34–45. (In French) [Google Scholar]
- Foster, A.D.; Rosenzweig, M.R. Microeconomics of Technology Adoption. Annu. Rev. Econ. 2010, 2, 395–424. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- Ministère de l’Agriculture. Les Enjeux de la Production Biologique en France. Available online: http://agriculture.gouv.fr/IMG/pdf/ANALYSE_CEP_50_Les_enjeux_de_la_production_d_agriculture_biologique_en_France_cle0b821c.pdf (accessed on 10 January 2015).
- De Ponti, T.; Rijk, B.; Van Ittersum, M.K. The Crop Yield Gap between Organic and Conventional Agriculture. Agric. Syst. 2012, 108, 1–9. [Google Scholar] [CrossRef]
- Seufert, V.; Ramankutty, N.; Foley, J.A. Comparing the Yields of Organic and Conventional Agriculture. Nature 2012, 485, 229–232. [Google Scholar] [CrossRef] [PubMed]
- Stringer, M.F.; Hall, M.N. A Generic Model of the Integrated Food Supply Chain to Aid the Investigation of Food Safety Breakdowns. Food Control 2007, 18, 755–765. [Google Scholar] [CrossRef]
- Neven, D. Developing Sustainable Food Value Chains: Guiding Principles; Food & Agriculture Organization of the United Nations (FAO): Rome, Italy, 2014. [Google Scholar]
- López-ridaura, S.; Keulen, H.V.; van Ittersum, M.K.; Leffelaar, P.A. Multiscale Methodological Framework to Derive Criteria and Indicators for Sustainability Evaluation of Peasant Natural Resource Management Systems. Environ. Dev. Sustain. 2005, 7, 51–69. [Google Scholar] [CrossRef] [Green Version]
- Lee, A. Health-Promoting Schools. Appl. Health Econ. Health Policy 2009, 7, 11–17. [Google Scholar] [CrossRef] [PubMed]
- Quatro, S.A.; Waldman, D.A.; Galvin, B.M. Developing Holistic Leaders: Four Domains for Leadership Development and Practice. Hum. Resour. Manag. Rev. 2007, 17, 427–441. [Google Scholar] [CrossRef]
- Fresco, L.O. Challenges for Food System Adaptation Today and Tomorrow. Environ. Sci. Policy 2009, 12, 378–385. [Google Scholar] [CrossRef]
- IPCC. Climate Change 2007: Appendix to Synthesis Report; Baede, A.P.M., van Der Lingen, P., Verbruggen, A., Eds.; IPCC: Geneva, Switzerland, 2007. [Google Scholar]
- Talamini, E.; Ferreira, G.M.V. Merging Netchain and Social Network: Introducing the Social Netchain Concept as an Analytical Framework in the Agribusiness Sector. Afr. J. Bus. Manag. 2010, 4, 2981–2993. [Google Scholar]
- Thadakamalla, H.P.; Raghavan, U.N.; Kumara, S.; Albert, R. Survivability of Multiagent-Based Supply Networks: A Topological Perspective. IEEE Intell. Syst. 2004, 19, 24–31. [Google Scholar] [CrossRef]
- Vurro, C.; Russo, A.; Perrini, F. Shaping Sustainable Value Chains: Network Determinants of Supply Chain Governance Models. J. Bus. Ethics 2010, 90, 607–621. [Google Scholar] [CrossRef]
- Christopher, M.; Peck, H. Building the Resilient Supply Chain. Int. J. Logist. Manag. 2004, 15, 1–14. [Google Scholar] [CrossRef] [Green Version]
- Giacomini, C.; Mancini, C. Organisation as a Key Factor in LocalisedAgri-Food Systems (LAFS). Bio-Based Appl. Econ. 2015, 4, 17–32. [Google Scholar]
- Crespo, J.; Réquier-Desjardins, D.; Vicente, J. Why Can Collective Action Fail in Local Agri-Food Systems? A Social Network Analysis of Cheese Producers in Aculco, Mexico. Food Policy 2014, 46, 165–177. [Google Scholar] [CrossRef]
- Jarosz, L. Understanding Agri-Food Networks as Social Relations. Agric. Hum. Values 2000, 17, 279–283. [Google Scholar] [CrossRef]
- Lazzarini, S.; Chaddad, F.; Cook, M. Integrating Supply Chain and Network Analyses: The Study of Netchains. J. Chain Netw. Sci. 2001, 1, 7–22. [Google Scholar] [CrossRef]
- Ulanowicz, R.E.; Goerner, S.J.; Lietaer, B.; Gomez, R. Quantifying Sustainability: Resilience, Efficiency and the Return of Information Theory. Ecol. Complex 2009, 6, 27–36. [Google Scholar] [CrossRef]
- Granovetter, M. The Impact of Social Structure on Economic Outcomes Social Networks and Economic Outcomes: Core Principles. J. Econ. Perspect. 2005, 19, 33–50. [Google Scholar] [CrossRef]
- Chiffoleau, Y.; Dreyfus, F.; Stofer, R.; Touzard, J.-M. Networks, Innovation and Performance. In Vertical Markets and Cooperative Hierarchies; Springer: Dordrecht, The Netherlands, 2007; pp. 35–60. [Google Scholar]
- Thuo, M.; Bell, A.A.; Bravo-Ureta, B.E.; Lachaud, M.A.; Okello, D.K.; Okoko, E.N.; Kidula, N.L.; Deom, C.M.; Puppala, N. Effects of Social Network Factors on Information Acquisition and Adoption of Improved Groundnut Varieties: The Case of Uganda and Kenya. Agric. Hum. Values 2014, 31, 339–353. [Google Scholar] [CrossRef]
- Lamine, C.; Meynard, J.-M.; Perrot, N.; Bellon, S. Analyse Des Formes de Transition Vers Des Agricultures plus Écologiques: Les Cas de l’Agriculture Biologique et de La Protection Intégrée. Innov. Agron. 2009, 4, 483–493. (In French) [Google Scholar]
- Kuhn, T.S. The Structure of Scientific Revolutions; University of Chicago Press: Chicago, IL, USA, 1970. [Google Scholar]
- Nemes, G.; High, C. Old Institutions, New Challenges: The Agricultural Knowledge System in Hungary. Stud. Agric. Econ. 2013, 115, 76–84. [Google Scholar] [CrossRef]
- Quiédeville, S.; Barjolle, D.; Mouret, J.-C.; Stolze, M. Ex-Post Evaluation of the Impacts of the Science-Based Research and Innovation Program: A New Method Applied in the Case of Farmers’ Transition to Organic Production in the Camargue. J. Innov. Econ. Manag. 2017, 22, 145. [Google Scholar] [CrossRef] [Green Version]
- Quiédeville, S.; Barjolle, D.; Stolze, M. Using Social Network Analysis to Evaluate the Impacts of the Research: On the Transition to Organic Farming in the Camargue. Cah. Agric. 2018, 27, 15012. [Google Scholar] [CrossRef]
- Earl, M.J.; Scott, L.A. Opinion: What Is a Chief Knowledge Officer? Sloan Manag. Rev. 1999, 40, 29. [Google Scholar]
- Scott, J. Social Network Analysis: A Handbook; SAGE Publications: London, UK, 2000. [Google Scholar]
- Wasserman, S.; Faust, K. Social Network Analysis: Methods and Applications; Cambridge University Press: Cambridge, UK, 1994. [Google Scholar]
- Vroegindewey, R.; Hodbod, J. Resilience of Agricultural Value Chains in Developing Country Contexts: A Framework and Assessment Approach. Sustainability 2018, 10, 916. [Google Scholar] [CrossRef]
- Powell, W.W.; Koput, K.W.; Smith-Doerr, L. Interorganizational Collaboration and the Locus of Innovation: Networks of Learning in Biotechnology. Adm. Sci. Q. 1996, 41, 116–145. [Google Scholar] [CrossRef]
- Rogers, E.M. Diffusion of Innovations; Free Press: New York, NY, USA, 1995. [Google Scholar]
- Wilson, N. The Supply Chains of Perishable Products in Northern Europe. Br. Food J. 1996, 98, 9–15. [Google Scholar] [CrossRef]
- Claro, D.P.; Hagelaar, G.; Omta, O. The Determinants of Relational Governance and Performance: How to Manage Business Relationships? Ind. Mark. Manag. 2003, 32, 703–716. [Google Scholar] [CrossRef]
- Storper, M.; Harrison, B. Flexibility, Hierarchy and Regional Development: The Changing Structure of Industrial Production Systems and Their Forms of Governance in the 1990s. Res. Policy 1991, 20, 407–422. [Google Scholar] [CrossRef]
- Porter, M.E. The Competitive Advantage of Nations; Free Press: New York, NY, USA, 1990. [Google Scholar]
- Dosi, G. Sources, Procedures, and Microeconomic Effects of Innovation. J. Econ. Lit. 1988, 26, 1120–1171. [Google Scholar]
- Harrison, B.; Kelley, M.R.; Gant, J. Innovative Firm Behavior and Local Milieu: Exploring the Intersection of Agglomeration, Firm Effects, and Technological Change. Econ. Geogr. 1996, 72, 233–258. [Google Scholar] [CrossRef]
- Scott, A.J. The Role of Large Producers in Industrial Districts: A Case Study of High Technology Systems Houses in Southern California. Reg. Stud. 1992, 26, 265–275. [Google Scholar] [CrossRef]
- Harrison, B. Industrial Districts: Old Wine in New Bottles? Reg. Stud. 1992, 26, 469–483. [Google Scholar] [CrossRef]
- Carayannis, E.G. Knowledge-Driven Creative Destruction, or Leveraging Knowledge for Competitive Advantage. Ind. High. Educ. 2008, 22, 343–355. [Google Scholar] [CrossRef]
- Aspara, J.; Hietanen, J.; Tikkanen, H. Business Model Innovation vs Replication: Financial Performance Implications of Strategic Emphases. J. Strateg. Mark. 2010, 18, 39–56. [Google Scholar] [CrossRef]
- Rigby, D.; Woodhouse, P.; Young, T.; Burton, M. Constructing a Farm Level Indicator of Sustainable Agricultural Practice. Ecol. Econ. 2001, 39, 463–478. [Google Scholar] [CrossRef]
- Zahm, F.; Viaux, P.; Vilain, L.; Girardin, P.; Mouchet, C. Assessing Farm Sustainability with the IDEA Method—From the Concept of Agriculture Sustainability to Case Studies on Farms. Sustain. Dev. 2008, 16, 271–281. [Google Scholar] [CrossRef]
- Schader, C.; Baumgart, L.; Landert, J.; Muller, A.; Ssebunya, B.; Blockeel, J.; Weisshaidinger, R.; Petrasek, R.; Mészáros, D.; Padel, S.; et al. Using the Sustainability Monitoring and Assessment Routine (SMART) for the Systematic Analysis of Trade-Offs and Synergies between Sustainability Dimensions and Themes at Farm Level. Sustainability 2016, 8, 274. [Google Scholar] [CrossRef]
- Pretty, J.N. Regenerating Agriculture: Policies and Practice for Sustainability and Self-Reliance; Joseph Henry Press: Washington, DC, USA, 1995. [Google Scholar]
- Gomez, A.A.; Swete Kelly, D.E.; Syers, J.K.; Coughtan, K.J. Measuring Sustainability of Agricultural Systems at the Farm Level. In Methods for Assessing Soil Quality; Soil Science Society of America: Madison, WI, USA, 1996; pp. 401–410. [Google Scholar]
- FranceAgriMer. Etude Coûts de Production Sur Le Riz Camarguais-Récolte 2011; FranceAgriMer: Limoges, France, 2011. (In French) [Google Scholar]
- Seyfang, G.; Smith, A. Grassroots Innovations for Sustainable Development: Towards a New Research and Policy Agenda. Environ. Politics 2007, 16, 584–603. [Google Scholar] [CrossRef]
- Parent, D.; Bélanger, V.; Vanasse, A.; Allard, G.; Pellerin, D. Method for the Evaluation of Farm Sustainability in Quebec, Canada: The Social Aspect. In Methods and Procedures for Building Sustainable Farming Systems; Springer: Dordrecht, The Netherlands, 2013; pp. 239–250. [Google Scholar]
- Gafsi, M.; Favreau, J. Appropriate Method to Assess the Sustainability of Organic Farming Systems. In Proceedings of the 9th European IFSA Symposium, Vienna, Austria, 4–7 July 2010. [Google Scholar]
- Halberg, N.; Verschuur, G.; Goodlass, G. Farm Level Environmental Indicators; Are They Useful? An Overview of Green Accounting Systems for European Farms. Agric. Ecosyst. Environ. 2005, 105, 195–212. [Google Scholar] [CrossRef]
- Picazo-Tadeo, A.J.; Gómez-Limón, J.A.; Reig-Martínez, E. Assessing Farming Eco-Efficiency: A Data Envelopment Analysis Approach. J. Environ. Manag. 2011, 92, 1154–1164. [Google Scholar] [CrossRef] [PubMed]
- Borgatti, S.P.; Everett, M.G.; Freeman, L.C. UCINET 6 for Windows: Software for Social Network Analysis; Analytic Technologies: Harvard, MA, USA, 2002. [Google Scholar]
- Heun, M.; Schäfer-Pregl, R.; Klawan, D.; Castagna, R.; Accerbi, M.; Borghi, B.; Salamini, F. Site of Einkorn Wheat Domestication Identified by DNA Fingerprinting. Science 1997, 278, 1312–1314. [Google Scholar] [CrossRef]
- Bogucki, P. The Spread of Early Farming in Europe. Am. Sci. 1996, 84, 242–253. [Google Scholar]
- Chabrol, D.; Bassene, J.B.; Fournier, S. Petit Épeautre de Haute-Provence: Quand l’IGP Est Le Pivot D’un Système Associant Différents Signes de Qualité. In Proceedings of the Symposium International sur L’agriculture Biologique Méditerranéenne et les Signes Distinctifs de Qualité liés à l’Origine, Agardi, Maroc, 2–4 December 2013. (In French). [Google Scholar]
- Clary, V. Projet Petit Épeautre 2014–2017; Syndicat des Producteurs de Petit Epeautre de Haute Provence: Mévouillon, France, 2013. (In French) [Google Scholar]
- Biopartenaire. Présentation. Available online: http://www.biopartenaire.com/ item/presentation.html (accessed on 4 April 2014).
- Cahier des Charges de la Dénomination «Riz de Camargue» Homologué par Arrêté du 18 Décembre 2013. Available online: https://www.legifrance.gouv.fr/jo_pdf.do?numJO=0&dateJO=20140107&numTexte=28&pageDebut=00191&pageFin=00191 (accessed on 5 July 2018). (In French).
- Mouret, J.-C. L’évolution Des Pratiques Agricoles En Relation Avec La Mise En Œuvre D’une Riziculture Durable. In 40 ans de Recherche au Service de la Gestion en Camargue: Actes du Forum Scientifique du 24 Septembre 2010; Parc Naturel Régional de Camargue: Bouches-du-Rhône, France, 2011. (In French) [Google Scholar]
- Chataigner, J.; Mouret, J.C. Recherche et Production Rizicole En France. Cah. Options Méditerranéennes 1997, 24, 117–126. (In French) [Google Scholar]
- Delmotte, S.; Lacombe, C.; Couderc, V.; Mailly, F.; Mouret, J.C.; Lopez-Ridaura, S.; Barbier, J.M. Obstacles, Levers and Impacts of Organic Farming Development in Camargue. Innov. Agron. 2013, 32, 213–226. [Google Scholar]
- Janssen, M.A.; Bodin, Ö.; Anderies, J.M.; Elmqvist, T.; Ernstson, H.; McAllister, R.R.J.; Olsson, P.; Ryan, P. Toward a Network Perspective of the Study of Resilience in Social-Ecological Systems. Ecol. Soc. 2006, 11, 15. [Google Scholar] [CrossRef]
- Bell, G.G. Clusters, Networks, and Firm Innovativeness. Strateg. Manag. J. 2005, 26, 287–295. [Google Scholar] [CrossRef]
- Newman, L.; Dale, A. Network Structure, Diversity, and Proactive Resilience Building: A Response to Tompkins and Adger. Ecol. Soc. 2005, 10, resp2. Available online: https://www.jstor.org/stable/pdf/26267768.pdf?refreqid=excelsior%3Aee545a164b0e3025b8119974710a9604 (accessed on 5 July 2018). [CrossRef]
- Krugman, P.R. Geography and Trade; Leuven University Press: Leuven, Belgium, 1991. [Google Scholar]
- Bouncken, R.B.; Gast, J.; Kraus, S.; Bogers, M. Coopetition: A Systematic Review, Synthesis, and Future Research Directions. Rev. Manag. Sci. 2015, 9, 577–601. [Google Scholar] [CrossRef]
- Xueming, L.; Slotegraff, R.J.; Pan, X. Cross-Functional “Coopetition”: The Simultaneous Role of Cooperation and Competition within Firms. J. Mark. 2006, 70, 67–80. [Google Scholar]
- Narver, J.C.; Slater, S.F. The Effect of a Market Orientation on Business Profitability. J. Mark. 1990, 54, 20–35. [Google Scholar] [CrossRef]
- Bodin, Ö.; Crona, B.I. The Role of Social Networks in Natural Resource Governance: What Relational Patterns Make a Difference? Glob. Environ. Chang. 2009, 19, 366–374. [Google Scholar] [CrossRef]
Influence of the Central Actor (s) | |||
---|---|---|---|
Low | High | ||
Density of the actor network | Low | “Transactional” | “Dictatorial” |
High | “Acquiescent” | “Participative” |
Actors Interviewed from the PGI Einkorn Value Chain | Actors Interviewed from the Camargue Rice Value Chain | |
---|---|---|
Upstream actors | - 7 elected representatives of the PGI einkorn union - 19 producers of PGI einkorn | - 4 fully-organic producers - 3 partially-organic producers (share of fields farmed organically) - 9 conventional producers |
Downstream actors | - 2 approved PGI shellers/millers - 1 PGI einkorndistributer (Euro Nat)- 6 Einkorn sellers | - 4 storage companies (SARL Thomas, Madar, Biocamargue and SudCéréales) - 2 packers (BioSud, Soufflet group) |
Organisations | - Biopartenaire - Coop COCEBI - Bio82 | - Camargue Natural Reserve - Chamber of Commerce and Industry |
Pillars of Sustainability | Indicators of Sustainability | Meaning of the Indicators | Data Collected |
---|---|---|---|
Socio-economic | Net producer margin [54,55]. | This indicator highlights notions of ethics (fair compensation) and performance. | Organic farmers were first asked to describe their crop rotation from the previous year. We then asked the yields and sales prices for the different crops. The net producer margin was then calculated by deducing production costs from the turnover (yield * sales prices). In the case of the Camargue, production costs for organic rice were derived from a study conducted by FranceAgriMer [56]. For the other crops, the following hypotheses were made: maize (€1800/ha), sorghum (€1700/ha), conventional wheat (€1700/ha), organic wheat (€1500/ha), alfalfa (€1500/ha), and temporary grassland (€1200/ha). As for the Einkorn case, the production costs were calculated based on the interviews with farmers: costs of all inputs applied and cost of machinery use (for all operations). |
Yield level and variability [55]. | Output per hectare and variability over years. | Organic farmers were asked directly about the yields of the different crops. The yield variability was evaluated in a more qualitative way: organic farmers were asked to state whether they are regularly faced with significant production falls; and if yes to what degree and how often (on average). | |
Dependence on the CAP, understood here as a way to account for the level of farm autonomy [52,54]. | Level of support from the CAP in relation to the level of income. A decline in farmed areas following a decrease in the CAP might significantly weaken the value chain. | This factor was only considered in the Camargue case. We used the French website Telepac where subsidies given to organic farmers are publicly available. The same sample as the organic farmers interviewed was used. The dependence on the CAP was calculated as the difference between the net margin with and without CAP payments, both for the rotation and for rice only. | |
Commitment to sustainable and alternative production schemes [57]. | Producers not heavily involved may represent a weakness for a value chain; a high level of commitment relies on non-economic rationales. | We asked organic farmers to rank their motivations to produce organically, from the most to the least important. They could choose from the following items: respect for the environment, social pressure, resilience of the farm, better revenue, and other reasons (specified where applicable). In the Camargue case, organic farmers were also asked whether they could easily abandon their organic production mode in the event of economic difficulties, e.g., decrease in the selling price. | |
Quality of life [52]. | Ethics and human development. | Organic farmers were asked to give a general quality of life score from 0 to 10 (10 being the highest possible score) in the Camargue case. In the einkorn case, farmers were asked to say if their quality of life is average, good or very good. | |
Transmissibility of farms [58,59]. | Viability of the sector and transmission of know-how to future generations; the share of farms with a successor in mind was calculated. | Organic farmers were asked whether they already know who will take over the farm. The percentage of positive and negative answers was then calculated. | |
Environment | TI (toxicity index) [60]. TFI (treatment frequency index) [61]. | Assessed through the use of chemical inputs, which is zero in the case of einkorn. For the rice value chain, both organic and conventional farmers were asked to explain their practices in terms of chemical treatments: the name of the products and the dosage applied. The certified dosage as well as the toxicity (LD50) of the molecules used was found on the French E-Phy website. |
Definition of the Attributes | Indicators Used to Measure Attributes | Tools and Methods Used | |
---|---|---|---|
Attributes of value chain governance | |||
Centrality | Propensity of a central stakeholder to connect unconnected members. | By the indicator of betweenness centrality, which refers to the extent to which an agent can play the role of a “broker” or “gatekeeper” with a potential for control over others. | Ucinet Software. Graph analysis [62]. |
Density | Relational thickness. | This attribute is also an indicator, which can be calculated based on a matrix of actors’ relationships. | Ucinet Software. Graph analysis. |
Attributes of the survival capacity of the value chain | |||
Robustness | Ability to cope with attacks. | The clustering coefficient and betweenness centrality are used to identify the central organisation(s) of the network. The size of the larger component is estimated by dividing the aggregate degrees of the actors involved in it by the total sum of degrees from the overall network. | Ucinet Software. Graph analysis. |
Responsiveness and distance | Rapidity of responses to shocks. | Calculated by the indicator of distance (a short distance between actors is conducive to a high degree of responsiveness). | Ucinet Software. Graph analysis. |
Flexibility | Number of alternative pathways. | Number of alternative suppliers and outlets for producers and storage companies. | Map of actors and in-depth interviews with actors. As explained earlier, the actors interviewed were asked to specify their suppliers and outlets. The question asked to measure the adaptivity was as follows: “If this actor were to collapse, would it be easy to change partner? (All suppliers and outlets were evaluated). The numbers of “easy” alternatives were then calculated. In the Camargue case, actors were also asked whether changing from one partner to another would have a low, moderate or high impact on their business. |
Adaptivity | Degree of facility to change value chain partners. | Capacity to easily change suppliers and outlets for producers and storage companies. |
Network Factors of Systemic Sustainability | Result (Value or Characterisation) |
---|---|
Density | 0.14 |
Distance | 2.39 |
Betweenness centrality SARL Tofagne | 170 |
Betweenness centrality (average) | 14 |
Robustness | Low |
Responsiveness | Average |
Flexibility | Low |
Adaptivity | Low |
Governance model | “Dictatorial” |
Network Factors of Systemic Sustainability | Result (Value or Characterisation) |
---|---|
Density | 0.54 |
Distance | 1.8 |
Betweenness centrality SARL Thomas | 184 |
Betweenness centrality SudCéréales | 934 |
Betweenness centrality (average) | 53 |
Cluster coefficient (“SudCéréales–SARL Thomas–BioSud” cluster) | 8.14 |
Cluster coefficient (average) | 5.37 |
Robustness | High |
Responsiveness | Rather high |
Flexibility | High |
Adaptivity | Average |
Governance model | “Participative” |
Pillars of Sustainability | Indicators of Sustainability | Results in the PGI Einkorn Value Chain | Results in the Camargue Organic Rice Value Chain |
---|---|---|---|
Socio-economic | Net producer margin | Good | |
Yield level and variability | Low level and variability | Average level and variability | |
Dependence on the CAP | Not relevant | Low for organic and average for partially-organic producers | |
Commitment to the production mode | High | High for organic producers and low for partially-organic producers | |
Quality of life | Good | ||
Transmissibility of farms | Not prepared by the majority of producers | ||
Environment | TI (toxicity index) | Null (due to the PGI specifications) | Significant reduction for partially-organic farmers |
TFI (treatment frequency index) |
Attributes of Governance and Survival Capacity of Value Chain | Results in the PGI Einkorn Value Chain | Results in the Camargue Organic Rice Value Chain |
---|---|---|
Value chain governance | ||
Centrality | High | High |
Density | Low | High |
Governance model | “Dictatorial” | “Participative” |
Survival capacity of value chain | ||
Robustness | Low | High |
Responsiveness | Average | Rather high |
Flexibility | Low | High |
Adaptivity | Low | Average |
© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Share and Cite
Quiédeville, S.; Bassene, J.-B.; Lançon, F.; Chabrol, D.; Moustier, P. Systemic Sustainability of the French Organic Rice and PGI Einkorn Value Chains: A Preliminary Assessment Based on Network Analysis. Sustainability 2018, 10, 2344. https://doi.org/10.3390/su10072344
Quiédeville S, Bassene J-B, Lançon F, Chabrol D, Moustier P. Systemic Sustainability of the French Organic Rice and PGI Einkorn Value Chains: A Preliminary Assessment Based on Network Analysis. Sustainability. 2018; 10(7):2344. https://doi.org/10.3390/su10072344
Chicago/Turabian StyleQuiédeville, Sylvain, Jean-Baptiste Bassene, Frédéric Lançon, Didier Chabrol, and Paule Moustier. 2018. "Systemic Sustainability of the French Organic Rice and PGI Einkorn Value Chains: A Preliminary Assessment Based on Network Analysis" Sustainability 10, no. 7: 2344. https://doi.org/10.3390/su10072344
APA StyleQuiédeville, S., Bassene, J. -B., Lançon, F., Chabrol, D., & Moustier, P. (2018). Systemic Sustainability of the French Organic Rice and PGI Einkorn Value Chains: A Preliminary Assessment Based on Network Analysis. Sustainability, 10(7), 2344. https://doi.org/10.3390/su10072344