[Gretl-users] SARIMA(X), some theoretical and pratictal questions

[Allin Cottrell]

On Thu, 13 Aug 2009, alisson rocha wrote:

> i have two questions:
>
> 1) I´m studying a sarima(x) model about beer consumption with a
> dummy variable to a local holiday that happen in february or
> march,and the dummy aim to control this (1 in the month with
> this particular holiday and 0 without it). In gretl, can i
> estimate this model using the 'command' below(using the dummy as
> exogenous variable)? arima p d q; P D Q; dependent_variable
> dummy_variable
>
> 2) The other is a model about net revenues from a local oil
> company, where i want to use a sarimaX with brent price(the
> barrel) as exogenous variable. In this case, both variables(net
> revenues and brent price) have to be stationary? In gretl, can i
> estimate it using the command below? arima p d q; P D Q;
> dependent_variable brent_oil

Yes, you can do these things.  However, since there is some
ambiguity over the exact definition of ARIMAX models you should
probably take a look at Section 20.2 ("ARIMA models") of the Gretl
User's Guide to see what gretl is estimating for you.

Allin Cottrell