FMOLS/DOLS

[unknown]

Hi,

I am carrying a cointegration estimation, through EViews ,7 between a set of I(1) variables using FMOLS and DOLS procedures (since there is some correlation troubles with the VAR method). I am wondering if I do enter my variables (in the equation specification and the deterministic regressors) in level or in first differences? In other words my questions is does Eviews tranfsorm my series into first differences when i perform FMOLS and DOLS estimation methods.

Thanks in adavnce for your help.

[EViews Glenn]

Neither method transforms into differences. Both methods estimate the cointegrating equation in original I1 space but with different methods of handling the asymptotic feedback.

[unknown]

Hi,

Thank you for your prompt reply. I know the differences between the two methods but i was not sure if I have to enter my variables into level or not in the equation specification box, since for exemple in the VAR estimation data are not transformed into differences by EViews whereas they are when using the VECM command.

[EViews Glenn]

Since both methods estimate the cointegrating relationship in levels, you should enter your variables in levels.