Dépt. Logique des Usages, Sciences Sociales et de l'Information

Téléphone : 02 29 00 11 17
Télécopie : 02 29 00 10 30
Courriel :
IMT Atlantique
Vue aérienne


Research activity


I am interested in an approach that views social and economic systems as a complex system. Macro-properties emerge from the micro-properties and the structure of interactions matters! Many independant individual actions lead to emerging properties which retroact on individual behavior.

Modeling in social sciences is a difficult task since individual act strategically based on information, beliefs, preferences, constraints some of which are non observable to the modeller. However, on one hand, with mathematical models, one gains in rigor, in simplicity by parsimony, and use of logics and mathematical tools. On the other hand, social sciences face (probably) a higher degree of complexity than natural sciences. Atoms are not strategic whereas humans are! This renders the task particularly hard and the bridging step between reality and the models difficult. But to deal with this approach, two methods which are experimental economics (great for teaching ) and agent-based modelling have become a useful support during the past two decades.

Experimental economics is not simply a tool for validating models. It can be a useful tool for designing institutions especially market institutions. It can give useful insights for a behavioral foundation of individual behavior in social interactions and has founded a new approach coined as "behavioral", economics ( Look at Camerer's Website for exemple). Rationality and individualism methodology is a major appraoch in economics. Behavioral economics has provided many new insights to this approach including the study of social preferences, models of reciprocity, emotions, norms or expectation formation in finance and may fill the gap between economics and sociology (see leading people on this approach such as Bowles, Fehr , Gintis ...).

A second important aspect of the problem is the structure of interaction that is how, when and with whom people interact. A major proponent of this approach is Alan Kirman. An implication follows directly: one cannot understand the economy by saying that it is the result of the interaction of representative agents (a representative consumer and producer for example for the study of a market). Interaction produces complexity; Agent based modelling (ABM) (Leigh Tesfatsion's Website) is particularly useful when the model is too complex for mathematical analytical tractability and for enriching simple models. In finance, it has been used to reproduce stylised facts of financial assets (such as fat tails, volatility clustering,...). A book which links behavioral finance to ABM (Levy, Levy and Solomon)

Not to say that they have changed my approach and pushed me to the way of more realistic behavioral and structural assumptions for modelling. I am lucky to collaborate with people with different academic backgrounds on different topics: social interactions (Mirta Gordon Jean Pierre Nadal, Denis Phan ) or markets (Eric Darmon , Olivier Brandouy).

I am involved in two ongoing research projects dealing with complexity and social modelling and one in agent based modelling and ontology.


Technopôle Brest-Iroise - CS 83818 - 29238 Brest Cedex 3 - France
Tél : 33 (0)2 29 00 11 11 - Fax : 33 (0)2 29 00 10 00