Résumé
Background and objectives. In recent years, public authorities and society at large have expressed a growing concern about sustainability issues, which have officially been captured by the signature of the 2030 agenda and the formulation of the 17 Sustainable Development Goals that guide international development cooperation operations. As a consequence of these trends, new needs are emerging in terms of assessment of value chains including agricultural value chains. On one hand, the actors of the value chains increasingly look for assessment methods that can measure the performances and impacts of their operations not only in terms of competitiveness but also in terms of contribution to sustainable development, in order to comply with private standards, national or international regulations and also to anticipate reputational risks. On the other hand, public and private actors and international donors need to evaluate their policies and programs in support of agricultural chains' development not only in terms of their potential as economic multipliers and generators of foreign currency, but also in terms of their capacity to contribute to the sustainable development of the territory where they are based. As an institution focused on applied research for agricultural development, CIRAD has long lasting records of partnership with a large range of players of agricultural value chains, both in a local and national context or at the global level (Griffon 1990; Fabre et al., 1997, Temple et al. 2007). Over the decades its field of investigation went beyond agronomic issues per se, encompassing socio-economic and environmental challenges from the production level down to the supply of agricultural processed products. As part of an institution in charge of supporting public and private decision making, CIRAD staff is challenged by its partners on how to evaluate the impact of, and contribution to sustainable development of value-chains. CIRAD has accumulated a significant capital of expertise in the field of agro-technical analysis, economic analysis, Life Cycle Assessment (LCA) and more recently developed new capacities in the evaluation of social impact of agri-chains (Feschet, 2014). This asset provides the ground to participate fruitfully to the assessment of value chains' contribution to sustainable development. The purpose of this paper is to discuss the objectives pursued and propositions made by a task force of CIRAD scholars working on a harmonized conceptual framework for evaluating the contribution of agri-chains to sustainable development. Issues. Historically, Value Chains have been assessed on the basis of their financial (from the entrepreneur perspective) or economic (from the policy maker perspective) performances using a range of benefit/cost indicators (Fabre 1994, Monke, 1989). Since the seventies the development and methodological consolidation of Life Cycle Analysis (LCA) offers a sound ground to assess the environmental impacts of the VC. The social and commonly called third dimension of the sustainability framework has been so far much more difficult to capture through methodologies that are unanimously recognized and applied by professionals (Seuring 2012). The Corporate Social Responsibility (CSR) relies on a battery of indicators defined within a specific context and is relevant for global value chains driven by large and formal companies and their suppliers with reputational and organizational interest in implementing a CSR strategy. However, beyond the specificity of value chains where this approach is applied, it also lacks a proper conceptual framework linking the human interventions with actual impacts on social well-being. Besides, there are very few attempts to build an integrated measure of the contribution of a value chain to sustainable development. Most of the analytical tools developed for an integrated assessment of the sustainability in the past decade are conceived at a macro-scale such as the Genuine Progress Indicator or the Sustainable Net Benefit Index (country or global) (Lawn, 2003, Anieleski, 2001). They cannot be straightforwardly adapted to an integrated assessment of value-chains that include only a limited share of resources allocation and thus cannot capture the total impact, direct within the systems and induced effects (outside the system at national, global scale) in environmental, economic and social terms. At the value-chain scale, in line with the establishment and increasing dissemination of CSRs by value-chain drivers and their suppliers, a number of indicators are defined, measured and reported but they do not provide or proceed from an articulated/integrated measure of the impact of the companies on sustainable development. They rather constitute a list of indicators that take into account the various dimensions in parallel. Furthermore, these tools are by construction limited to socio-economic and institutional aspects where a formal corporation can be accountable for their impacts and performances. The various attempts to build up a framework that can be applied and provide a sound base for assessing VC performances, or contribution to sustainable development have so far been very limited (Hassini, 2012).Challenges and working hypothesis. Based on a first range of brainstorming sessions among CIRAD scholar involved in assessing the impacts of VC along different dimensions of sustainable development, the paper aims to present working hypotheses and challenges that the group has so far identified putting them into perspective with current literature. These key challenges are briefly presented in the following. Which definition of sustainability. While there are clear references for assessing individually each dimension of sustainability, a combined or harmonized conceptual framework requires defining clearly what sustainable development means and how the different dimensions interact. A metric of value chain impact on sustainable development can only be operational in decision making terms (choice, tradeoffs, and ranking) if it relates to an explicit conception of sustainability. Shifting from a global scale to the scale of a value chain. Assuming that impact on sustainable development cannot only be thoroughly assessed at the global level (considering the socio-biosphere as a closed system), a value chain as an open subsystem cannot be sustainable as such, so we rather aim at assessing its contribution to sustainable development. The delineation, the mapping of the value-chain perimeter plays a key role in locating the point where the value chain sub-systems interact with “outside the VC system”. The functional analysis of the VC, how the systems operate and interact with the outside of the system are key steps in the different approaches and can be an entry point for developing an articulated methodology addressing the different dimensions of the sustainable development. In LCA, one key challenge remains the modelling of local impacts at the scale of the whole value chain which spreads over the world. Important efforts have been made to spatialize the impact assessment especially for eutrophication, acidification, water deprivation impacts etc. This is particularly important for agri-food chains which have important contributions to such local and regional impacts for which the local parameters play a key role. Impacts / performances. Another key aspect of the evaluation approaches relate to the level at which the indicators are defined and how close they are from the actual object to protect. In the LCA approach, one should evaluate the potential impacts of a human intervention and the indicators should have a clear and scientific link with the final object to protect or “area of protection”. Conversely, in other approaches, indicators might be defined close to the human intervention with no explicit link to the final effect on human well-being or ecosystems. Static versus dynamic approach and applicability. Distinction should be made between assessing the performance/impacts of a VC for a given configuration of the system observed at a specific period and having the capacity to include a temporal/dynamic perspective. Modeling of flows, input/output feedback measures, is certainly the most comprehensive way to assess the impact of a system such as VC on sustainable development (Ukidwe. 2011), but this option is highly demanding in terms of data and skills and might not be easily applied in contexts where data and analytical time is limited. An assessment limited to a number of indicators is likely more feasible but will require an explicit and sound analytical framework explaining how different categories of indicators are articulated. The starting point of the working group with the systems definition and identification of convergence in terms of methodology will be developed through a collective ontology process and is further presented in a companion paper submitted to the conference (Martin et al., 2016). (Texte intégral)