Statalist


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

st: RE: xtivreg2 with clustered standard errors - very high first stage F


From   "Schaffer, Mark E" <M.E.Schaffer@hw.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: xtivreg2 with clustered standard errors - very high first stage F
Date   Thu, 27 Mar 2008 08:46:29 -0000

Meg,

> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu 
> [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of 
> Meghana Ayyagari
> Sent: 26 March 2008 21:16
> To: statalist@hsphsun2.harvard.edu
> Subject: st: xtivreg2 with clustered standard errors - very 
> high first stage F
> 
> Hi,
> 
> I am using xtivreg2 to estimate a fixed effects model. My 
> data is a cross-section of firms across countries. My concern 
> is that while the instruments seem to pass all the tests of 
> weak instruments, the first stage F-stat (and the 
> corresponding Kleibergen-Paap Wald rk F
> statistic) is extremely high making me wonder if there is 
> something wrong here. This seems to be the case only when I 
> ask for clustered standard errors. Could someone please help 
> me interpret this? I am posting the output below.

In your example, you have a single endogenous regressor.  This is
helpful, because in this special case, the Kleibergen-Paap rk statistic
is identical to an F test of the excluded instruments in the first-stage
regression.  In practice the small-sample dof corrections vary across
-regress-, -ranktest-, -ivreg2-, etc., but that's irrelevant in your
case.

So the easiest way to check what is going on is to run the first-stage
regression by hand using -cluster-, which I think in your case means

xtivreg2 x1 z1, cluster(cnum) fe small

and do an F test of z1.  You should get a very similar F stat to the K-P
rk F stat reported by -xtivreg2-.

If not, please contact me off-list and we'll try to work out what's
going on.

Cheers,
Mark

Prof. Mark Schaffer
Director, CERT
Department of Economics
School of Management & Languages
Heriot-Watt University, Edinburgh EH14 4AS
tel +44-131-451-3494 / fax +44-131-451-3296
email: m.e.schaffer@hw.ac.uk
web: http://www.sml.hw.ac.uk/ecomes


> 
> Thanks
> 
> Meg
> 
> xtivreg2 gro size (x1=z1), cluster(cnum) fe ffirst liml FIXED 
> EFFECTS ESTIMATION
> ------------------------
> Number of groups =        79                    Obs per 
> group: min =       15
>                                                               
>  avg =      78.9
>                                                               
>  max =       378
> 
> Summary results for first-stage regressions
> -------------------------------------------
> Variable    | Shea Partial R2 |   Partial R2    |  F(  1,    
> 78)    P-value
> gcf         |     0.0585      |     0.0585      |    17074.85 
>       0.0000
> 
> NB: first-stage F-stat cluster-robust
> Underidentification tests
> Ho: matrix of reduced form coefficients has rank=K1-1 
> (underidentified)
> Ha: matrix has rank=K1 (identified)
> Kleibergen-Paap rk LM statistic             Chi-sq(1)=50.27   
>  P-val=0.0000
> Kleibergen-Paap rk Wald statistic           
> Chi-sq(1)=17296.53 P-val=0.0000
> 
> Weak identification test
> Ho: equation is weakly identified
> Kleibergen-Paap Wald rk F statistic             17074.85
> See main output for Cragg-Donald weak id test critical values
> 
> Weak-instrument-robust inference
> Tests of joint significance of endogenous regressors B1 in 
> main equation
> Ho: B1=0 and overidentifying restrictions are valid
> Anderson-Rubin Wald test     F(1,78)=  3.98      P-val=0.0497
> Anderson-Rubin Wald test     Chi-sq(1)=4.03      P-val=0.0448
> Stock-Wright LM S statistic  Chi-sq(1)=3.46      P-val=0.0628
> *
> *   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/
> 

-- 
Heriot-Watt University is a Scottish charity
registered under charity number SC000278


*
*   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