File(s) under permanent embargo

Long memory in financial markets: a heterogeneous agent model perspective

journal contribution
posted on 01.07.2018, 00:00 authored by M Zheng, Ruipeng LiuRuipeng Liu, Y Li
During last decades, studies on asset pricing models witnessed a paradigm shift from rational expectation and representative agent to an alternative, behavioral view, where agents are heterogeneous and boundedly rational. In this paper, we model the financial market as an interaction of two types of boundedly rational investors — fundamentalists and chartists. We examine the dynamics of the market price and market behavior, which depend on investors' behavior and the interaction of the two types of investors. Numerical simulations of the corresponding stochastic model demonstrate that the model is able to replicate the stylized facts of financial time series, in particular the long-term dependence (long memory) of asset return volatilities. We further investigate the source of the long memory according to asset pricing mechanism of our model, and provide evidences of long memory by applying the modified R/S analysis. Our results demonstrate that the key parameter that has impact on the long memory is the speed of the price adjustment of the market maker at the equilibrium of demand and supply.

History

Journal

International review of financial analysis

Volume

58

Pagination

38 - 51

Publisher

Elsevier

Location

Amsterdam, The Netherlands

ISSN

1057-5219

Language

eng

Publication classification

C Journal article; C1 Refereed article in a scholarly journal

Copyright notice

2018, Elsevier Inc.