Testing for Granger causality in the presence of measurement errors
In this paper a potential problem with tests for Granger-causality is investigated. If one of the two variables under study, but not the other, is measured with error the consequence is that tests of forecastablity of the variable without measurement error by the variable with measurement error will be rejected less often than it should. Since this is not the case for the test of forecastability of the variable with measurement error by the one without there is a danger of concluding that one variable leads the other while it is in fact a feed-back relationship. The problem is illustrated by an example.
|Date of creation:||24 Sep 2004|
|Date of revision:|
|Contact details of provider:|| Postal: NHH, Department of Business and Management Science, Helleveien 30, N-5045 Bergen, Norway|
Phone: +47 55 95 92 93
Fax: +47 55 95 96 50
Web page: http://www.nhh.no/en/research-faculty/department-of-business-and-management-science.aspx
More information through EDIRC
When requesting a correction, please mention this item's handle: RePEc:hhs:nhhfms:2004_011. 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: (Stein Fossen)
If references are entirely missing, you can add them using this form.