Testing for long memory volatility of Chinese stock markets with FIGARCH model

Abstract

In this study, the long memory property in the volatility of Chinese stock markets is examined. For this purpose, we applied two semi-parametric tests (GPH and LW) and the FIGARCH model, to four Chinese market indices: Shanghai A, Shanghai B, Shenzhen A and Shenzhen B. The estimates from the GPH test and the LW estimator support the Taylor effect in absolute and squared returns, implying that the long memory property exists in the volatility of four Chinese stock markets. Also, the FIGARCH model is better equipped to capture the long memory volatility process than the GARCH and IGARCH models. In particular, the FIGARCH (1, d, 0) model is found to provide a good representation of four Chinese stock returns. Finally, the Student-t distribution outperforms the normal one in capturing leptokurtosis in residuals.