Abstract: Experimental markets exhibit regularities that are fundamentally unexpected and challenging from a neoclassical viewpoint: that a market of few haggling traders, endowed with only local and private information, could converge to an optimal equilibrium is indeed a surprising result, because the dominant theory assumes infinitely many price-taking traders endowed with perfect knowledge of market supply and demand. In this talk, we shall present a theory of the market mechanism rooted in classical economics that is a more natural framework for understanding the experimental findings, and we derive from this theory some insights on bubbles in durable or asset markets.