Can opinion be stable in an open network with hierarchy? An agentbased model of the Commercial Court of Paris

We develop an agent-based model of the co-evolution of social networks and opinion formation, using empirical data on the advice network of judges at the Commercial Court in Paris. We aim to identify the motives that drive agents' behavior and their implications both on the shape of the emerging network and on the possible harmonization of members' opinions over time. We make the assumptions that judges seek advice from senior colleagues who are higher up in the hierarchy, who enjoy high reputation, or who are similar to them, and test how these assumptions impact on the global dynamics of opinion diffusion and network formation, and fit with the data. Our results single out the combination of criteria that most likely guide individuals' selection of advisors and provide insight into their effects on opinion formation.