Statalist The Stata Listserver


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

st: RE: Re: xtivreg2


From   "Schaffer, Mark E" <[email protected]>
To   <[email protected]>
Subject   st: RE: Re: xtivreg2
Date   Wed, 24 Jan 2007 01:20:49 -0000

Joe,

> -----Original Message-----
> From: [email protected] 
> [mailto:[email protected]] On Behalf Of Zou Hong
> Sent: 23 January 2007 10:08
> To: [email protected]
> Subject: st: Re: xtivreg2
> 
> Dear Kit,
> 
> Many thanks for your kind suggestion.
> I have installed xtivreg2 and run the test as follows.
> 
> ttset code year
> xtivreg2 y1 x1 x2 x3 (y2 = x4 x5), fe robust endog(y2)
> 
> I am suspicious that y2 is endogenous to y1 and want to test 
> this using a fixed-effect estimation with robust standard 
> errors. Is the above setup correct?
> 
> Can I claim that the above conducted the DWH test using a 
> fixed-effect estimation with roubt standard errors?

Yes, but your terminology isn't quite right.  The C test (or GMM
distance test) is more general than the DWH test.

> Having run the models, I found:
> 
> Warning - singleton groups detected.  214 observation(s) not used.

Regarding your question about singletons - these are groups which have
only one member.  In a fixed effects estimation, they drop out
completely.  Put another way, if you dropped these observations and
re-ran your fixed effects estimation, your estimates would be exactly
the same.  It's easy to see why - the fixed effects estimator explains
the "within-group" variation, and when a group has 1 member, there is no
"within-group" variation to explain.

HTH,
Mark


> 
> FIXED EFFECTS ESTIMATION
> ------------------------
> Number of groups =       492                    Obs per group: min = 
> 2
>                                                                avg =
> 3.7
>                                                                max =
> 7
> 
> (regression output here)
> Test statistic(s) not robust
> --------------------------------------------------------------
> ----------------
> Cragg-Donald F statistic (weak identification test): 
> 1.907
> Stock-Yogo weak ID test critical values:  5% maximal IV relative bias
> 13.91
>                             10% maximal IV relative bias     9.08
>                             20% maximal IV relative bias     6.46
>                             30% maximal IV relative bias     5.39
>                             10% maximal IV size             22.30
>                             15% maximal IV size             12.83
>                             20% maximal IV size              9.54
>                             25% maximal IV size              7.80
> Test statistic(s) not robust
> Source: Stock-Yogo (2005).  Reproduced by permission.
> ---------------------------------------------------------Hanse
n J statistic 
> (overidentification test of all instruments):         3.738
>          Chi-sq(2) P-val =    0.1543
> -endog- option:
> 
> Endogeneity test of endogenous regressors: 
> 0.308
>                  Chi-sq(1) P-val =    0.5786
> Regressors tested:    y2
> 
> what is the meaning of "singleton groups" and how to deal with it?
> 
> Thanks a lot
> Joe
> 
> ----- Original Message -----
> From: "Kit Baum" <[email protected]>
> To: <[email protected]>
> Sent: Tuesday, January 23, 2007 11:17 AM
> Subject: st: re: ivreg2 update and fixed-effect estimator
> 
> 
> > Joe said
> >
> > 1. where to get the latest version of ivreg2? I always 
> found it  difficult 
> > to identify the version of an ado file from the internet. 
> The one I  found 
> > from the net is version 02.1.14. But searched from the web, 
> the latest
>  version seems to be 2.1.18. How to update the files if 
> ivreg2 is always 
> installed
> >
> > 2. Under ivreg2, how to test the exogeneity of a regressor under a 
> > fixed-effect (or random-effects) estimator?
> >
> > 3. how to request a DWH test using a robust standard error?
> >
> >
> >
> > 1. The latest version of ivreg2 (or of just about any user-authored 
> > software) is available using the ssc or adoupdate commands. 
> It is  always 
> > straightforward to check the version number: ssc type  
> ivreg2.ado will 
> > display it, as will "which ivreg2" after installation.
> >
> > 2. You need to use xtivreg2 (a 'wrapper' for ivreg2, also 
> available  from 
> > ssc) and the orthog() or endog() option. The help file 
> explains  them.
> >
> > 3. The last paragraph of Section 5 of Baum, Schaffer, 
> Stillman, SJ 3 (1) 
> > [also available in preprint form as BC WP 545] speaks to 
> this  point. A 
> > robust DWH test is available with orthog() or the newer  
> option endog(). 
> > endog() is not described in that paper, but is  described 
> in the help 
> > file. A followup paper describing these  additional features is in 
> > preparation.
> >
> > Kit
> >
> > Kit Baum, Boston College Economics
> > http://ideas.repec.org/e/pba1.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/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 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index