Author(s): Mikhail Anufriev, Economics Discipline Group, UTS Business School, University of Technology, Sydney. Cars Hommes, University of Amsterdam and Tomasz Makarewicz, University of Amsterdam
Date of publication: July 2015
Working paper number: 29
Abstract: We study a model in which individual agents use simple linear first order price forecasting rules, adapting them to the complex evolving market environment with a smart Genetic Algorithm optimization procedure. The novelties are: (1) a parsimonious experimental foundation of individual forecasting behaviour; (2) an explanation of individual and aggregate behavior in four different experimental settings, (3) improved one-period and 50-period ahead forecasting of lab experiments, and (4) a characterization of the mean, median and empirical distribution of forecasting heuristics. The median of the distribution of GA forecasting heuristics can be used in designing or validating simple Heuristic Switching Model.