On the Identification of Codependent VAR and VEC Models
In this paper we discuss identification of codependent VAR and VEC models. Codependence of order q is given if a linear combination of autocorrelated variables eliminates the serial correlation after q lags. Importantly, maximum likelihood estimation and corresponding likelihood ratio testing are only possible if the codependence restrictions can be uniquely imposed. However, our study reveals that codependent VAR and VEC models are not generally identified. Nevertheless, we show that one can guarantee identification in case of serial correlation common features, i.e. when q=0, and for a single vector generating codependence of order q=1.
|Date of creation:||15 Sep 2010|
|Date of revision:|
|Contact details of provider:|| Postal: Universitäts-Str. 31, D-93040 Regensburg|
Phone: +49 941 943-2392
Fax: +49 941 943-4752
Web page: http://www-wiwi.uni-regensburg.de/
More information through EDIRC
When requesting a correction, please mention this item's handle: RePEc:bay:rdwiwi:16477. See general information about how to correct material in RePEc.
For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: (Gernot Deinzer)
If references are entirely missing, you can add them using this form.