Page 1 of 1

Autocorrelation Test in a Seemingly Unrelated Regression SUR

Posted: Mon Mar 23, 2015 11:12 am
by thomas84
Hi everyone,

I have estimated a system of 15 specifications using seemingly unrelated regression. Looking at the estimation output some of the specifications show DW values of 1.4 which indicates autocorrelation. How do I test for this in a SUR? The only thing I could find under Residual Diagnostics is "System Residual Portmanteau Tests for Autocorrelations". However, I do not know what this test does and how to interpret the results. Normally I use a Breusch-Pagan test, however, this is not an option in the SUR estimated system. What can I do to test for autocorrelation in the individual specifications?

Best
Thomas
"

Re: Autocorrelation Test in a Seemingly Unrelated Regression

Posted: Tue Mar 31, 2015 4:51 am
by thomas84
Hey again,

does anyone have an idea please? I cannot find anything related to this in Eviews manual or text books. Thank you.

Thomas