Long-Term Memory in Realized Volatility: Evidence from Chinese Stock Market

摘要

In this paper, we examine the long-term memory in realized volatility with different time scales based on high frequency data of Shanghai Stock Exchange Composite Index (SSECI). We choose R/S analysis method to calculate Hurst exponents of long-term memory, and use ARFIMA model to estimate and forecast. Our results show that long-term memory in realized volatility becomes strong as time scales increases. We also found that the realized volatility is best measured and forecasted by one-minute interval.