Hello Morten, Here you use a ADL(1,1) for teaching purposes. However in sometimes an ADL(1,1) may not suffice. I have a question: Suppose that you have quarterly data and you therefore choose your maxlag to be 4. By using some information criterion or generel-to-specific method, where you check for lag significance, you get that you should include the first and the third lag for your dependent variable and only include the second lag for the independent variable. Is this still feasible? would you be able to construct an ECM from that? or do you need the lags to be in chronological order and thus you keep some of the lags even though they are insignificant? I can't find any literature that address this Thank you very much.
![Seungmin "그렇게, 천천히, 우리(As we are)" | [Stray Kids : SKZ-PLAYER]](http://i.ytimg.com/vi/kAzmhLHePqU/mqdefault.jpg)
thank you this video ,
+kamel AKRICHE You're welcome. Hope you found it useful.
Hello Morten,
Here you use a ADL(1,1) for teaching purposes. However in sometimes an ADL(1,1) may not suffice. I have a question: Suppose that you have quarterly data and you therefore choose your maxlag to be 4.
By using some information criterion or generel-to-specific method, where you check for lag significance, you get that you should include the first and the third lag for your dependent variable and only include the second lag for the independent variable.
Is this still feasible?
would you be able to construct an ECM from that?
or do you need the lags to be in chronological order and thus you keep some of the lags even though they are insignificant?
I can't find any literature that address this
Thank you very much.