Bifurcations routes and spectral analysis of agents behaviour

Abstract

This diploma thesis presents a model (Brock and Hommes 1998) which is a form of evolutionary dynamics that is called Adaptive Belief Systems in a simple present discounted value (PDV) asset pricing model. Agents can choose from a finite set of different beliefs of predictors of the future price of a risky asset. Predictor choice is (boundedly) rational in the sense that, at each date, most agents choose the predictor generating the highest past performance. Different tools will be used for studying this model, mainly spectral analysis including Periodogram and Bispectrum. It is important to find out how much these individual types influence deviation of prices from fundamental and how much energy is transferred between investors types.