Empirical study on the long-memory components in asset returns

Abstract

In this work, the presence of long-memory components in asset returns is investigated. For a stock market, it surely possesses various periodic components. Furthermore, for various markets, the relationship between power and period of cycle components follows same power law with an exponent close to -0.92, spanning in a wide scale range from several days to nearly four years. This empirical result indicates that such power relationship may be universal and will enlighten modeling of market volatility. © 2012 Springer-Verlag GmbH.