Long memory in forex markets: an empirical analysis

Purpose This study aims to analyze the long memory properties of daily/monthly log return series of thirteen currency pairs, offering insights into the predictability or randomness of return and volatility in forex markets. Design/methodology/approach The study uses returns, absolute returns and squared return data for these currencies over different time periods, applying the Hurst Exponent analysis, Local Whittle Estimator and GPH estimator for long memory estimation. Findings The results indicate no evidence of long memory in most currency pairs, suggesting long-term market efficiency. However, there is some inconclusiveness in the case of JPYUSD and AUDUSD, where two studies show no long memory, but the Hurst Exponent indicates its presence. Research limitations/implications The study is limited to a specific set of currency pairs and time periods, and future research could expand this analysis to more currencies and longer timeframes. Practical implications The findings have implications for investors and policymakers, indicating that most currency markets exhibit long-term efficiency, with information shocks not persisting in prices over time. Social implications The study’s insights can contribute to a better understanding of forex market dynamics, potentially informing policies and strategies in international trade and finance. Originality/value This study compares long memory across a diverse set of currency pairs, adding to the existing literature by providing a comprehensive analysis of long memory in forex markets.