Kossi Messanh Agbekponou will defend his PhD in front of the following jury:

• Ariell Reshef (CNRS, Université Paris 1 Panthéon-Sorbonne), referee
• Liza Jabbour (University of Birmingham), referee
• José De Sousa (Université Paris-Assas et Sciences-Po), examinator
• Catherine Laroche Dupraz (Institut Agro Rennes-Angers), examinator
• Karine Latouche (INRAE), supervisor
• Angela Cheptea (INRAE), co-supervisor

Abstract

Over the last thirty years, there has been a significant increase in global value chain (GVC) trade. The challenge is to measure the importance of GVCs in the agri-food sector from a microeconomic perspective and identify the determinants and motivations of firms to participate in this process, in order to better account for GVCs in the design of countries' trade and industrial policies. Specifically, the thesis analyses the GVC participation premium of retailers' suppliers; the role of product quality on firms' boundary choices in GVCs; and how these choices affect their bargaining power. Theoretical elements, based on a model with heterogeneous firms, make it possible to analyze these mechanisms in detail. The empirical analyses use detailed data on French companies in the agri-food industry to     highlight the importance of intra-industry analysis of GVCs. The results show the GVC participation premium for retailers' suppliers. Secondly, it is documented that quality upgrading enables firms to participate more intensively in GVCs by performing a greater number of production stages. This leads to increased input costs and higher profits and value added. Thirdly, it is demonstrated that firms' specialization in further upstream stages of GVCs, as well as expansion along GVCs of firms producing closer to final demand, increase the division of surplus in GVCs, and that these effects are mainly due to higher quality production. Additionally, the effect of specialization is stronger when firms actually and solely undertake processing activities in GVCs, and because it reduces the hold-up problem for firms, otherwise the two effects cancel each other out.