Determination of Lag for the Cointegration Test

A forum for questions and answers regarding the CATS cointegration procedure
upani
Posts: 55
Joined: Wed Jun 25, 2014 3:31 am

Determination of Lag for the Cointegration Test

Unread post by upani »

Dear All,

How to choose optimal lags in the Johansen cointegration tests?

With regards,
Upananda
TomDoan
Posts: 7814
Joined: Wed Nov 01, 2006 4:36 pm

Re: Determination of Lag for the Cointegration Test

Unread post by TomDoan »

upani wrote:Dear All,

How to choose optimal lags in the Johansen cointegration tests?

With regards,
Upananda
I would suggest using @VARLAGSELECT on the levels (not differenced) data. Since the lags in CATS are the augmenting lags, you would subtract one from what that chooses.
upani
Posts: 55
Joined: Wed Jun 25, 2014 3:31 am

Re: Determination of Lag for the Cointegration Test

Unread post by upani »

If i want to check autocorrelation for each lag and arch test, how will i do that?
TomDoan
Posts: 7814
Joined: Wed Nov 01, 2006 4:36 pm

Re: Determination of Lag for the Cointegration Test

Unread post by TomDoan »

upani wrote:If i want to check autocorrelation for each lag and arch test, how will i do that?
@MVQSTAT for multivariate serial correlation and @MVARCHTEST for multivariate ARCH.

Note that CATS includes quite a few diagnostics.
Post Reply