Please use this identifier to cite or link to this item: https://hdl.handle.net/10419/189342 
Year of Publication: 
2006
Series/Report no.: 
Queen's Economics Department Working Paper No. 1062
Publisher: 
Queen's University, Department of Economics, Kingston (Ontario)
Abstract: 
This paper analyzes the semiparametric estimation of multivariate long-range dependent processes. The class of spectral densities considered is motivated by and includes those of multivariate fractionally integrated processes. The paper establishes the consistency of the multivariate Gaussian semiparametric estimator (GSE), which has not been shown in other work, and the asymptotic normality of the GSE estimator. The proposed GSE estimator is shown to have a smaller limiting variance than the two-step GSE estimator studied by Lobato (1999). Gaussianity is not assumed in the asymptotic theory. Some simulations confirm the relevance of the asymptotic results in samples of the size used in practical work.
Subjects: 
fractional integration
long memory
semiparametric estimation
JEL: 
C22
Document Type: 
Working Paper

Files in This Item:
File
Size





Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.