*
*  Example from section 20.3, page 767
*  Indirect least squares
*
cal 1970
open data table20-1.prn
data(format=prn,org=columns) 1970:1 1991:1
*
* Compute the two reduced form regressions. Save the coefficient vectors as bq
* and bp.
*
linreg p
# constant x
compute bp=%beta
prj phat
linreg q
# constant x
compute bq=%beta
prj qhat
*
*  Compute the ILS estimates
*
compute b1hat=bq(2)/bp(2)
compute b0hat=bq(1)-b1hat*bp(1)
disp b0hat b1hat
*
*  Do the OLS regression
*
linreg q
# constant p