Procedure for properly estimating an AR(p)

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

Procedure for properly estimating an AR(p)

p_giannuzzo
This post has NOT been accepted by the mailing list yet.
Dear stata users,
I'm trying to estimate an autoregressive process AR(p).
Following the literature:
-I checked if the series is stationary or not running the augmented Dickey-Fuller test (as I expected, the series results stationary);
-I computed the optimal number of lag using DFgls test.
In your opinion, is this procedure correct?
Do I have to use other tests for unit-roots (Dickey-Fuller test, KPSS test, .... or is the augmented Dickey-Fuller test fine to this end?
Do I have to use other methods for computing the optimal number of lag, as, for instance, the command varsoc, or is the dfgls command ok?
Thanks for answering and sharing my problem!