Estimation of the long memory parameter in non stationary models: A Simulation Study - Groupement de Recherche en Économie Quantitative d'Aix-Marseille Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2011

Estimation of the long memory parameter in non stationary models: A Simulation Study

Résumé

In this paper we perform a Monte Carlo study based on three well-known semiparametric estimates for the long memory fractional parameter. We study the efficiency of Geweke and Porter-Hudak, Gaussian semiparametric and wavelet Ordinary Least-Square estimates in both stationary and non stationary models. We consider an adequate data tapers to compute non stationary estimates. The Monte Carlo simulation study is based on different sample size. We show that for d belonging to [1/4,1.25) the Haar estimate performs the others with respect to the mean squared error. The estimation methods are applied to energy data set for an empirical illustration.
Fichier principal
Vignette du fichier
DTGREQAM2011_24.pdf (545.5 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

halshs-00595057 , version 1 (23-05-2011)

Identifiants

  • HAL Id : halshs-00595057 , version 1

Citer

Mohamed Boutahar, Rabeh Khalfaoui2. Estimation of the long memory parameter in non stationary models: A Simulation Study. 2011. ⟨halshs-00595057⟩
140 Consultations
915 Téléchargements

Partager

Gmail Facebook X LinkedIn More