Self-organization and the persistence of noise in financial markets

A dynamic model of financial markets with learning is demonstrated to produce a self-organized system that displays critical behavior. The price contains private information that traders learn to extract and employ to forecast future value. Since the price reflects the beliefs of the traders, the learning process is self-referencing. As the market learns to correctly extract information from the price, the market deemphasizes private information. Despite the convergence of the model towards the parameters producing efficiency, pricing deviations remain constant due to the increased sensitivity of the price to small errors in information extraction produced by the model's own convergence.