Inverse Mills raito estimation
Posted: Wed Sep 20, 2017 7:31 pm
Hello
I have used the Heckit in Eviews 9 but the output is very doubtful, so I decided to estimate the first-stage probit model first and then put Inverse Mills ratio in the second equation estimated with OLS - kind of old style before Heckit was not in Eviews.
I found and used a code to compute Inverse Mills raito in old posts (my probit first-stage equation is called "femaledirfirststage":
eqfemaledirfirststage.makeresid ordinary
eqfemaledirfirststage..makeresid(g) general
eqfemaledirfirststage..fit yprob
series imills2 = general*(1-yprob)/ordinary
But the program returns the message "ordinary is not defined", and the iMills2 series have N/A in all the fields. How do I fix the above code?
Also, is it true that inverse mills ratio is simply residuals series automatically stored in Eviews from estimation of the first-stage probit model? Thanks a lot for any advice!!!
I have used the Heckit in Eviews 9 but the output is very doubtful, so I decided to estimate the first-stage probit model first and then put Inverse Mills ratio in the second equation estimated with OLS - kind of old style before Heckit was not in Eviews.
I found and used a code to compute Inverse Mills raito in old posts (my probit first-stage equation is called "femaledirfirststage":
eqfemaledirfirststage.makeresid ordinary
eqfemaledirfirststage..makeresid(g) general
eqfemaledirfirststage..fit yprob
series imills2 = general*(1-yprob)/ordinary
But the program returns the message "ordinary is not defined", and the iMills2 series have N/A in all the fields. How do I fix the above code?
Also, is it true that inverse mills ratio is simply residuals series automatically stored in Eviews from estimation of the first-stage probit model? Thanks a lot for any advice!!!