Long Memory Volatility Persistence in High Frequency Precious Metals Returns

Abstract

Using high frequency data, this paper examines the long memory property in the conditional volatility of the precious metals return series at different time frequencies using FIGARCH models. Very significant long memory characteristics have been detected in absolute returns by using Semiparametric local Whittle estimation of the long memory parameter. Estimation of the long memory parameter across many different data sampling frequencies gives consistent estimates of the long memory parameter, indicating that the series are exactly to show some degree of self-similarity. Results indicate that the long memory property remains quite consistent across different time frequencies for both unconditional and conditional volatility measures. This study is useful for investors and traders (with different trading horizons) and it can be used in predicting expected future volatility and in designing and implementing trading strategies at different time frequencies.