A1398
Title: Recent extensions of short- and long memory volatility and duration models implemented with R
Authors: Oliver Kojo Ayensu - Paderborn University (Germany) [presenting]
Yuanhua Feng - University of Paderborn (Germany)
Dominik Schulz - Paderborn University (Germany)
Abstract: Recent advancements in short- and long-memory extensions of volatility (GARCH) and autoregressive conditional duration (ACD) models and their implementation in R are explored to enhance the understanding of dynamic behaviours in financial returns and non-negative time series. The study begins with a brief review of established GARCH and ACD models, then examines novel long memory GARCH variants, including fractionally integrated Log-GARCH (FILog-GARCH), modulus asymmetric FILog-GARCH (MAFILog-GARCH), and members of the "double power modulus EGARCH class", such as modulus modified EGARCH (MEGARCH) and modulus Log-GARCH (MLog-GARCH). Additionally, two new ACD models, namely the adjusted FIACD and long-memory Box-Cox ACD, are introduced. Details on their theoretical properties, estimation methods, and practical applications are investigated. A comprehensive review of relevant R packages is provided, offering practical guidance for implementing established and emerging models. This study serves as a valuable resource for researchers and practitioners seeking to apply advanced volatility modeling techniques using R.
