11

Due to the recent stock market crises, many economists have been keeping aloof from the « core» of financial theory, whose main contention is that stock market prices are at any moment the best possible estimate of the « true» value of the issuing firms. Most of those economists have brought forward, in a way or another, the hypothesis that markets suddenly get crazy. This contribution, based on both observa~tion of the recent financial history (the crash of the « New Economy» securities) and results from « experimental economics», and using the complementary concepts of « mimicry» and « convention » to account for « self-referential dynamics», is opening a third way of thinking the efficiency problematic: investors fix their strategy based on the estimates of the other participants, even if they strongly believe they are wrong.