Statalist The Stata Listserver


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: st: Re:


From   "Schaffer, Mark E" <M.E.Schaffer@hw.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: st: Re:
Date   Thu, 23 Mar 2006 17:59:19 -0000

I think this might be a case of a "forbidden regression", something that
comes up on Statalist from time to time:

http://www.stata.com/statalist/archive/2005-05/msg00158.html
http://www.stata.com/statalist/archive/2003-11/msg00795.html 

--Mark

> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu 
> [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of 
> Arne Risa Hole
> Sent: 23 March 2006 17:31
> To: statalist@hsphsun2.harvard.edu
> Subject: Re: st: Re:
> 
> I suppose an alternative approach would be to do something like this:
> 
> use http://fmwww.bc.edu/ec-p/data/hayashi/griliches76.dta, 
> clear qui regress iq med kww expr tenure rns smsa predict 
> double iq_hat gen double intact = iq_hat*expr qui regress lw 
> iq_hat intact s expr tenure rns smsa replace iq_hat = iq 
> replace intact = iq*expr predict double res, residual gen 
> double res2 = res^2 qui sum res2 scalar iv_mse = 
> r(mean)*r(N)/e(df_r) matrix b = e(b) matrix V = 
> e(V)*(iv_mse/e(rmse)^2) ereturn post b V ereturn display
> 
> i.e. get the forecast y2hat (iq_hat in the example) from the 
> first-stage regression and interact this with the exogenous 
> RHS var x1 (expr in the example). Then you replace the 
> interaction with y2*x1 before calculating the residuals/MSE.
> 
> Austin's suggestion is probably the better one though, but 
> this seems to me to be ok.
> 
> Cheers
> Arne
> 
> On 23/03/06, Austin Nichols <austinnichols@gmail.com> wrote:
> > If your endog RHS var y2 is interacted with an exog RHS var 
> x1, then 
> > you have a "new" endog RHS var y2x1, and you may need additional 
> > excluded instruments.  Use -ivreg- as suggested.
> >
> > On 3/23/06, Paolo Caruso <P.Caruso@warwick.ac.uk> wrote:
> > > Thanks for your replies.
> > >
> > > The reason why I want to exclude the variables is because 
> they are interactions (based on the variable that I am 
> instrumenting) therefore, it can not be valid to include them 
> in the first stage. If there is any other way of avoiding 
> this problem then please let me know.
> > >
> > > Thank you for the commands but the actual problem I am 
> having is if it is possible to do the same thing apart from 
> with more than one instrumented variable and what additional 
> commands I would have to do in order to ensure the correct 
> t-stats are reported.
> > >
> > > Regards,
> > >
> > > Paolo
> > >
> > > >>> austinnichols@gmail.com 03/23/06 14:08 PM >>>
> > > It is worse than "not a good idea" to pick and choose among exog 
> > > vars to include in the first stage--it is not the IV 
> estimator, and 
> > > you would have to derive consistency results for your new 
> homemade 
> > > estimator based on different assumptions. You are better off 
> > > sticking with -ivreg- or -ivreg2- to be sure. If you want 
> to include
> > > d7usq=d7unit^2 as a new endogenous RHS variable, you now have two 
> > > endog variables, and you may need more excluded 
> instruments--but the 
> > > products (with exog vars) and powers of your existing excluded
> > > instrument(s) are candidates.
> > >
> > > On 3/23/06, Arne Risa Hole <arnehole@gmail.com> wrote:
> > > > You can try the following code:
> > > <snip> This replicates:
> > > >
> > > > ivreg lw expr tenure rns smsa (iq s=med kww)
> > > >
> > > > If you want to remove some of the instruments simply drop them 
> > > > from the first stage regressions. Note that not using all the 
> > > > exogenous variables as instruments is probably not a 
> good idea - 
> > > > see http://www.stata.com/support/faqs/stat/ivreg.html.
> > > >
> > > > On 23/03/06, Paolo Caruso <P.Caruso@warwick.ac.uk> wrote:
> > > > > This is a question with regards to IV's.
> > > > >
> > > > > I am using STATA in order to do a 2SLS. STATA does 
> not allow you to specify which variables to include or not 
> include in the first stage. This means that I have to 
> complete the regression in the separate two stages.
> > > > >
> > > > > I know how to adjust the MSE when I have one variable 
> that I want to instrument, however, I do not know how to do 
> it when I have more than one.
> > > > >
> > > > > The command lines in stata to adjust the MSE for one 
> variable are as follows:
> > > > <snip> ereturn display
> > > > >
> > > > > where d7unit is the variable that I want to 
> instrument and loginc is the dependant variable.
> > > > >
> > > > > I would also like to instrument d7unit^2 and this is 
> where I am unsure what the relevant command lines would be in 
> order to ensure that the correct MSE is used.
> > > > >
> > > > > Any help would be much appreciated.
> > >
> > > *
> > > *   For searches and help try:
> > > *   http://www.stata.com/support/faqs/res/findit.html
> > > *   http://www.stata.com/support/statalist/faq
> > > *   http://www.ats.ucla.edu/stat/stata/
> > >
> > >
> > > *
> > > *   For searches and help try:
> > > *   http://www.stata.com/support/faqs/res/findit.html
> > > *   http://www.stata.com/support/statalist/faq
> > > *   http://www.ats.ucla.edu/stat/stata/
> > >
> >
> > *
> > *   For searches and help try:
> > *   http://www.stata.com/support/faqs/res/findit.html
> > *   http://www.stata.com/support/statalist/faq
> > *   http://www.ats.ucla.edu/stat/stata/
> >
> 
> *
> *   For searches and help try:
> *   http://www.stata.com/support/faqs/res/findit.html
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/
> 
> 

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2020 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index