IS CHAOS GENERIC IN ECONOMIC DATA?

This paper examines recent developments in nonlinear science in economics. Several claims of findings of chaos in economic data are reviewed. We discuss how each claim has been revised in light of further analysis, and point out several traps for empirical researchers in economic data. These traps suggest certain methodological refinements useful for researchers analyzing very small data sets, including diagnostic tests to detect ill conditioned data, filtering data to exclude nonchaotic alternatives, and nonparametric procedures to check the precision of parameter estimates. Most specialists in the field would say there is no conclusive evidence of chaos in economic or financial data.