Théorie comportementale du portefeuille. Intérêt et limites

This paper deals with the recent developments of portfolio choice theory, which has been dominated by the classical mean/variance approach for half a century. In a first section we present some market anomalies put to light by many empirical studies. Academic researches on the limitations of the expected utility theory have given rise to new models, called ?behavioral?, such as those by Arzac and Bawa [1977] and Shefrin and Statman [2000]. They are presented in section 2. By relying on two examples, we stress on the consequences of these approaches for optimal portfolio selection. Despite the seemingly ability of these models to explain empirical phenomena, we argue and expose reasons against their practical use by portfolio?s managers. Classification JEL : G11.