Testing for time-varying long-range dependence in volatility for emerging markets

This paper tests whether volatility for equity returns for emerging markets possesses long-range dependence. Furthermore, the assertion of whether long-range dependence is time-varying is checked through a rolling sample approach. The empirical results suggest that there exists long-range dependence in emerging equity returns' volatility and also that it is time-varying. This assertion also holds true for Japan and the US, which are considered more developed markets. Moreover, these results are robust to "shuffling" the data to eliminate short-term autocorrelation. Therefore, they suggest that the class of GARCH processes, which are currently employed to analyze volatility of financial time series, is misspecified. (C) 2004 Elsevier B.V.. All rights reserved.