A1229
Title: On the identification of long and short memory components with applications to volatility
Authors: Tommaso Proietti - University of Roma Tor Vergata (Italy) [presenting]
Alessandra Luati - Imperial College London (United Kingdom)
Shelton Peiris - University of Sydney (Australia)
Gnanadarsha Dissanayake - Health and Policy Research Associates (Sri Lanka)
Abstract: The purpose is to develop a methodology for deriving the autoregressive and moving average coefficients of long memory models, with a focus on ARFIMA and Gegenbauer processes. The approach is based on the cepstrum of the long memory process, which allows for a systematic decomposition of the impulse response function into long memory and short memory components. This decomposition provides a separation between persistent dynamics and short-run fluctuations. The method is compared to existing approaches, such as the Beveridge-Nelson decomposition, and its application is illustrated in analyzing volatility processes and their roughness.
