Abstract (EN)

This article investigates the time evolution of the integration of derivative markets through the graph-theory. We focus on three categories of underlying assets: energy and agricultural products, as well as financial assets. Integration is seen as a necessary condition for systemic risk to appear. In order to fully understand it, this phenomenon is comprehended through a three-dimensional approach: observation time, spatial relationships and term structures. Such an approach indeed gives the possibility to investigate prices shocks appearing and propagating in the physical as well as in the paper markets. In order to understand the underlying principles and dynamic behavior of our system, we select specific tools of statistical physics. We first use minimum spanning trees as a way to filter the information contained in the graph. We then study the topology of the filtered networks in order, first to see how they are organized, second to quantify the degree of randomness in this organization. Lastly, the time dependent properties of the trees are examined. On an economic point of view, the emerging taxonomy is meaningful, which is a key justification of the use of our methodology. Moreover, we observe an increasing integration of the markets through time, as well as a dominance of spatial over maturity integration.