Page 1 of 1

Breusch-Godfrey Serial Correlation LM

Posted: Mon Sep 22, 2008 1:34 pm
by tg128
how can I use the Breusch-Godfrey Serial Correlation LM test in EView 6, what do I suppose to put in the blank area please?

also for the ARCH LM test, what is WGTRESID? is that standarized residual or just squared error terms??

How can I make standarized residual series please?

Posted: Mon Sep 22, 2008 2:24 pm
by EViews Gareth
Which "blank" area are you talking about? If you're using the Breusch-Goodfrey Serial Correlation LM Test from the menus, there is no "blank" area...

WGTRESID are the residuals, weighted by whatever weight series you used in the original equation.

Posted: Tue Sep 23, 2008 8:39 am
by tg128
Which "blank" area are you talking about? If you're using the Breusch-Goodfrey Serial Correlation LM Test from the menus, there is no "blank" area...

WGTRESID are the residuals, weighted by whatever weight series you used in the original equation.
Sorry, I meant the Breusch-Pagan-Godfrey test...

Posted: Tue Sep 23, 2008 8:47 am
by EViews Gareth
Which "blank" area are you talking about? If you're using the Breusch-Goodfrey Serial Correlation LM Test from the menus, there is no "blank" area...

WGTRESID are the residuals, weighted by whatever weight series you used in the original equation.
Sorry, I meant the Breusch-Pagan-Godfrey test...
Whichever regressors you wish to use in the auxiliary regression for the BPG Test. The standard, text-book, approach is to use the regressors from the original equation (and EViews usually adds those variables to that white box by default).

thanks

Posted: Tue Sep 23, 2008 8:54 am
by tg128
thanks