Long-Term Memory in Stock Market Returns: International Evidence

A lot of recent work has addressed the issue of the presence of long memory components in stock prices because of the controversial implications of such a finding for market efficiency and for martingale models of asset prices used in financial economics and technical trading rules used for forecasting. This paper examines the presence of long memory in the stock returns of seven countries, namely Japan, Korea, New Zealand, Malaysia, Singapore, the USA and Australia. The classical and modified rescaled range tests, the semiparametric test proposed by Geweke and Porter-Hudak, the frequency domain score test proposed by Robinson and its time-domain counterpart derived by Silvapulle, are applied to these returns in order to detect the long memory property. Evidence suggests that the Korean, Malaysian, Singapore and New Zealand stock returns are long-term dependent, indicating that these two markets are not efficient. The results of this study should be useful to regulators, practitioners and derivative market participants, whose success precariously depends on the ability to forecast stock price movements. Copyright @ 2001 by John Wiley & Sons, Ltd. All rights reserved.