Introduction to special issue on complexity in economics and finance

Abstract

This special issue of the Journal of Economic Dynamics and Control is a collection of 10 papers presented at a workshop on Complexity in Economics and Finance, held at the Lorentz center in Leiden, the Netherlands, on October 22-27, 2007. The benchmark approach to economic and financial economic theory is to assume the existence of a fully rational, representative agent that exists inside an economy with complete markets. In dynamic settings, this approach typically assumes that the representative agent maximizes her expected lifetime utility subject to a (perceived) budget constraint. One implication is that optimizing agents will be forward-looking, and hence, there will be a crucial role for expectations in individual decisions. It follows that depending on the nature of expectations, macroeconomic and financial outcomes will be impacted by the beliefs of agents. Although expectations play a primary role in macroeconomic and financial outcomes, there is no consensus on how agents actually form expectations. The Rational Expectations (RE) paradigm, still the benchmark model in economics, assumes that agents’ subjective expectations are consistent with the actual stochastic process generated by those beliefs. A large literature questions the RE hypothesis on theoretical and empirical grounds. Theoretically, it can be argued that RE are not realistic because of the strong cognitive and computational assumptions required for agents’ beliefs to be rational. To actually compute rational beliefs, agents would need to know the precise structure and laws of motion for the economy, even though this structure depends on agents’ beliefs (c.f. Evans and Honkapohja (2001)). Moreover, rational expectations models often may possess multiple equilibria, on which individual agents must coordinate. Empirical studies have shown that survey data and macroeconomic data are not consistent with rational expectations. Mankiw, Reis, and Wolfers (2003) and Branch (2004) show that survey data on inflation expectations exhibit a wide degree of time-varying heterogeneity that can not be generated by a representative rational expectation. Milani (2007) finds that a simple New Keynesian model where rational expectations are replaced by an adaptive learning rule is preferred by the data to the best fitting model solved under the assumption of rational expectations. The standard economic approach also fails to explain many important features of economic systems, e.g., the emergence of speculative bubbles and crashes in financial markets, fat tails, long memory and clustered volatility in the returns distribution of