*
* Example 8/2/4 from pp 411-413
*
open data adv_sale.dat
data(format=free,org=columns) 1 36 sales adv
*
* Estimate the model with an AR(1) error term
*
boxjenk(ar=1,maxl,regressors) sales
# constant adv{0 to 4}
*
* The results of the transfer function estimation are slightly different, likely
* because of different handling of expansion of the denominator polynomial.
*
boxjenk(ma=1,maxl,constant,inputs=1) sales
# adv 1 1 0