Re: [Gretl-users] ARIMA v. OLS seasonal difference
On Wed, 15 Sep 2010, Mike Pfeiff wrote:
Is there any obvious reason why using the ARIMA feature (without any
AR
or MA terms) with a seasonal difference=1 would yield different
coefficients that using OLS and manually seasonally differencing the
independent and dependent variables?
No. Can you post an example script that shows differences in
results? We have some test cases on file where we get identical
results via the two methods, but maybe we're overlooking
something.
Ah, one point that may or may not be relevant: if you include a
constant in an ARIMA model we don't difference it away to nothing;
it's therefore equivalent to including a linear trend in an I(0)
model.
Allin Cottrell