We deal with the important question of estimating the long-run variance of a stationary sequence. We derive the asymptotic properties of a generalized Newey-West type of estimator in the case of a linear I(d) process. The results show that the bias and asymptotic distribution of the generalized Newey-West estimator depend on the memory parameter d. If the series has long memory then the estimator might even have a non-Gaussian limit distribution. The optimal bandwidth parameter q minimising MSE is derived. Theoretical results explain the large bias observed in simulation studies with arbitrarily chosen q. An alternative estimator is suggested. It has an asymptotic Gaussian distribution and bias which do not depend on d. The estimator is easy to apply and can be used to construct confidence intervals. Simulations confirm the theoretical findings.
Download Info
To our knowledge, this item is not available for
download. To find whether it is available, there are three
options:
1. Check below under "Related research" whether another version of this item is available online.
2. Check on the provider's web page
whether it is in fact available.
3. Perform a search for a similarly titled item that would be
available.
Publisher Info
Paper provided by Department of Economics, University of York in its series Discussion Papers with number
05/19.
Length: Date of creation: Date of revision: Handle: RePEc:yor:yorken:05/19
Contact details of provider: Postal: Department of Economics and Related Studies, University of York, York, YO10 5DD, United Kingdom Phone: (0)1904 433776 Fax: (0)1904 433759 Email: Web page: http://www.york.ac.uk/depts/econ/ More information through EDIRC
For technical questions regarding this item, or to correct its listing, contact: (Paul Hodgson).