/* Modeling ARMAX time series: ARIMA model with an ARMA disturbance process */

/* No need to tsset this data but under most conditions you will need to do so */

use http://www.stata-press.com/data/r9/friedman2, clear

/* Let's model personal consumption expenditures (consump) and money supply (m2) */

arima consump m2, ar(1) ma(1), if tin(, 1981q4)

/* We might be concerned that our specification has led to disturbances that are */
/* heteroskedastic or non-Gaussian, therefore we will refit the model with a robust option */

arima consump m2, ar(1) ma(1) robust, if tin(, 1981q4)

/* Note substantial increase in the estimated standard errors and that the MA term */
/* specified in model 1 is now only marginally significant. */