Cointegration (to difference or not to difference) NEED HELP

[kulilit]

Hi guys, I am currently doing my thesis and I am confused whether to use the first differences of my variables in Johansen Cointegration or not. I already tested the stationarity of my variables and I found out that all of them are I(1).

Should I run their first differences or just their level?


Thank you so much in advance for your help.

[darryl]

You should run Johansen's Cointegration at levels.

[Dim]

At levels.

But do you know how to apply the Pantula's Principle in order to determine the best model (with/without ternd; with/without intercept)? I mean, do you know the code for this Pantula's principle?

Thanks,

D.