Chaotic learning equilibria
成果类型：
Article
署名作者：
Schönhofer, M
署名单位：
University of Bielefeld
刊物名称：
JOURNAL OF ECONOMIC THEORY
ISSN/ISSBN：
0022-0531
DOI：
10.1006/jeth.1999.2509
发表日期：
1999
页码：
1-20
关键词：
摘要：
Chaotic temporary equilibrium trajectories under least squares learning are shown to exist in a standard version of the overlapping generations model. Theses trajectories show forecast errors that never vanish. Thus we do not find convergence to an equilibrium, in which agents have perfect foresight. Complex trajectories can even be found in the case of savings functions, which are monotonically decreasing in the expected gross rate of inflation. In this case no cyclical or chaotic trajectories with perfect foresight exist. Thus learning alone is a source of chaotic dynamic. Journal of Economic Literature Classification Numbers: D83, D84. (C) 1999 Academic Press.