estimating sur and modified wald statistic

[guptashruti]

hi,

my query is with respect to a time series data, i have three variables and using the information criterions the optimal lag length came out to be 4. the issue is that i have to employ the toda yamamoto test(1995)( based on a var analysis) which estimates the model with the optimal number of lags+the max order of integration of the series invovled. so i estimated a model with 6 (4+2) lags as one of the series was integrated of order 2. However, after running var, i dont get the option to use coefficient tests( as in it dosnt show up only in the options menu )...

i want to test the significance of the first 4 coefficients only ( modified wald statisitic)?

any idea as to how to do that?

Also i would like to estimate sur for the same , which am unable to figure out? as i read that sur gives the estimatefor mwald( modified wald statistics)

thanks :)

[EViews Gareth]

Turn your VAR into a System.

Proc->Make System.


Then estimate the system (note you can choose SUR as an estimation method). Then click on View->Coefficient diagnostics->Wald

[guptashruti]

hey
thanks much:)
this works ...

thanks a lot !

[guptashruti]

ONE MORE QUICK QUESTION....
what is the default method of estimation when one just estimates VAR?.. ( is it ols?)

[guptashruti]

I AM GETTING THE SAME COEFFICIENT ESTIMATES VIA BOTH ....

THANKS :)

[mmahbobi]

Hi,

in using TODA and Yamakoto Wald test, I know how to add restrictions in Wald, but what kind of restriction needs to be added? how do I come up with the set of restrictions?

highly appreciate quick responses
MM