Page 1 of 1

Significance in PDLs

Posted: Sun Apr 17, 2016 8:48 am
by firefoxxp
Hi, everyone,
When I use PDLs to estimate a distributed lag model, how could I know each explanatory variable is significant or not?
As in Eviews help, PDL01 to PDL06 are all not significant to the t-statistic, The original coefficients seems also insignificant, can I ingore all these informations and only see the sum of lags? How can I judge whether the explanatory variables are significant or not?
Many thanks!

Re: Significance in PDLs

Posted: Sun Apr 17, 2016 9:42 am
by startz
Look at the representations view. You will see that an equation like

Code: Select all

LS Y C PDL(X,3,2)
is estimated as
Y = C(1) + C(2)*PDL01 + C(3)*PDL02 + C(4)*PDL03

Use View/Coefficient diagnostics/Wald to test the hypothesis

Code: Select all

c(2)=c(3)=c(4)=0

Re: Significance in PDLs

Posted: Sun Apr 17, 2016 7:55 pm
by firefoxxp
Many thanks!
I still don't know if all the PDLs are significant according to Wald test, then the related sum of lags must be significant? As the t-statistic indicate that sum of lags is not significant.
How to know the significance of the "sum of lags" at the bottom of the table?
Many thanks!