Bounded Rationality and Learning in Complex Markets

Abstract

This chapter reviews some work on bounded rationality, expectation formation and learning in complex markets, using the familiar demand-supply cobweb model. We emphasize two stories of bounded rationality, one story of adaptive learning and another story of evolutionary selection. According to the adaptive learning story agents are identical, and can be represented by an ``average agent'', who adapts his behavior trying to learn an optimal rule within a class of simple (e.g. linear) rules. The second story is concerned with heterogeneous, interacting agents and evolutionary selection of different forecasting rules. Agents can choose between costly sophisticated forecasting strategies, such as rational expectations, and freely available simple strategies, such as naive expectations, based upon their past performance. We also confront both stories to laboratory experiments on expectation formation. At the end of the chapter, we integrate both stories and consider an economy with evolutionary selection between a costly sophisticated adaptive learning rule and a cheap simple forecasting rule such as naive expectations. Keywords: complex systems, behavioral economics, evolutionary selection, nonlinear dynamics.