Re: [Gretl-users] ADF, KPSS and Engle-Granger tests
> 1) In ADF test should be chosen variant with constant and linear
trend
> and
> this test should be made for the levels of variable (not differences)?
No. If you think, I suppose based on the graphic, that your series follows
a
quadratic trend, it is very probably your series really be I(2). In that
case
the ols estimators of the ADF equation in levels are inconsistent, and the
t-statistic does not follow the DF distribution. You should start with the
series in differences (+ linear trend), so testing first the second unit
root.
Thanks. But let me ask about something (integration and cointegration are
not 'my topics', so sorry if questions are trivial). What is the role of
deterministic parts in ADF regression? Why we sometimes add constant and
linear trend to ADF regression? It's still I(1) process? Couldn't be better
to test differences of variable with a constant term instead (so second
differences will be dependent variable in ADF regression)?
Greetings,
Mariusz
Poland