Good day folks,
In conducting unit root tests, we have to choose among certain specifications (intercept, trend and intercept, none). How do we determine which of these to choose?
I was told that you can tell from the Correlogram or the graph. :?
Unit Root tests specifications
Moderators: EViews Gareth, EViews Moderator
Re: Unit Root tests specifications
Hello,
You must test all. And for each, if the trend of the constant is not significative, so you must test the next model.
So, in level, you test in the order :
-trend and intercept;
-then intercept;
-then none.
If in intercept, for example, the intercept is good (C has a t-statistic superior to his critical value), but ADF is not good, so it's not good. And you do a new test in first difference with constant and you check that all is good. etc...
The critical values for intercept and trend : Table .doc.
You must test all. And for each, if the trend of the constant is not significative, so you must test the next model.
So, in level, you test in the order :
-trend and intercept;
-then intercept;
-then none.
If in intercept, for example, the intercept is good (C has a t-statistic superior to his critical value), but ADF is not good, so it's not good. And you do a new test in first difference with constant and you check that all is good. etc...
The critical values for intercept and trend : Table .doc.
Return to “Econometric Discussions”
Who is online
Users browsing this forum: No registered users and 2 guests
