Statalist The Stata Listserver


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

st: RE: treatreg (specification of indepvars_t)


From   "Schaffer, Mark E" <M.E.Schaffer@hw.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: treatreg (specification of indepvars_t)
Date   Mon, 5 Mar 2007 22:54:20 -0000

Richard, 

> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu 
> [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Richard Lo
> Sent: 05 March 2007 00:59
> To: statalist@hsphsun2.harvard.edu
> Subject: st: treatreg (specification of indepvars_t)
> 
> Dear Stata-listers, 
> 
> Although I have read some materials on the treatment effect 
> model, I am still not sure whether I should specify a 
> structural or a reduced-form equation for the 
> treatment(selection) equation. According to the textbooks of 
> Wooldridge (2002) and Greene (2003), it seems that a 
> reduced-form treatment equation should be used. Though in 
> Maddala (1983, p120-122), it is not clear which form is 
> specified for the probit equation.
> {It seems that the example given in the on-line help [
> the command in (1)]specifies the excluded variables
> only, which would not produce a reduced-form results
> for the treatment equation [as the command in (2)
> does]. 
> 
> (1)treatreg ww wa cit, treat(wc=wmed wfed) twostep
> (2)treatreg ww wa cit, treat(wc=wa cit wmed wfed) twostep  
> 
> As a result, I am puzzled at the correct way to
> specify the variables (indepvars_t) for the treatment
> equation. The results will be different depending on
> whether only excluded variables or all the exogenous
> variables are specified in the treatreg command.

As I understand it, the difference between (1) and (2) is that (2) uses
no exclusion restrictions in the selection ("treatment") equation:
everything exogenous appears on the RHS.  I think this is what you mean
by "reduced form".

I think this is just another version of the usual trade off between
efficiency and robustness that you get in system estimation.  If the
exclusions from the treatment equation are correct, both are consistent
but (1) is more efficient than (2); but if they aren't correct, (1) is
inconsistent but (2) remains consistent.

Cheers,
Mark

> [Unlike ivreg command which would produce the same
> results when we specified either way. ]
> 
> Would be grateful that someone can give me some
> guidance on this issue. 
> 
> Thanks you for your kind help in advance.
> 
> Richard Lo
> 
> 
> 		
> ___________________________________________________________ 
> All New Yahoo! Mail - Tired of unwanted email come-ons? Let 
> our SpamGuard protect you. http://uk.docs.yahoo.com/nowyoucan.html
> *
> *   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 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index