Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: Re: Clarification request


From   Kit Baum <baum@bc.edu>
To   mabrouk abir <mabrouk_abir@yahoo.fr>
Subject   st: Re: Clarification request
Date   Wed, 3 Mar 2010 10:10:58 -0500

1) No, the restriction is not that cov(Y,Z) = 0, it is that cov(Y,Z | X) = 0. That is, once you introduce Xs, the Zs should not enter the equation directly.

2) Yes.

3) There is no need to do so, as often discussed on this list (see Mark Schaffer's posts). Consistent estimates of an IV equation with an endogenous dummy or count variable can be computed by using 2SLS, as long as the usual set of assumptions for IV are satisfied. There are also good reasons not to do 2SLS 'by hand' by computing hat-values from a Poisson or negbin regression.


On Mar 3, 2010, at 3:15 AM, mabrouk abir wrote:

> Dear Kit,
> I would like please to discuss some aspects with you.
> I have relied upon a simultaneous equations model in order to take into account of the possibility of reverse causality. Thus I adopt the 2SLS estimation approach. I would like to ask you some questions :
>  
> 1. I see that many authors don't consider the basic assumption that Z (instruments) must don't explain Y (dependent variable) otherwise (Cov (Y, Z) = 0)? I've understood that in the case of simultaneous equations, I must not take account of this assumption in the first stage. Is it true?
> 2. If I use a "Control Function Approach" (which is equivalent to two-stage-least-squares) to estimate a simultaneous equations model, are the instruments the same of those chosen in the 2SLS approach ?
>  
> 3. In a "Control Function Approach", can I use Poisson regression instead of SLS, in each stages, to estimate a simultaneous equations model. This because I adopt a count measure for an endogenous variable of the model ?
>  
> Thank you very much for reading my mail. I am looking forward to hearing from you soon.
> Yours sincerly
> Abir Mabrouk
>  
>  
>  
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 


Kit Baum   |   Boston College Economics & DIW Berlin   |   http://ideas.repec.org/e/pba1.html
                              An Introduction to Stata Programming  |   http://www.stata-press.com/books/isp.html
   An Introduction to Modern Econometrics Using Stata  |   http://www.stata-press.com/books/imeus.html


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   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   |   Site index