RE: st: ivreg2 doesn't report Kleinbergen-Paap statistic

["Schaffer, Mark E" <M.E.Schaffer@hw.ac.uk>]

Eva, Mukkades,

A couple of ideas:

Can you confirm which versions of -ivreg2- and -ranktest- you have installed?  We updated both recently.  I see that Eva says she has updated -ivreg2- but I'm not sure she also updated -ranktest-.  The latter is what calculates the K-P stat.

Eva, what happens if you use -xtivreg2- to do your fixed-effects estimation?  That is,

xtivreg2 y x1 x2 x3 x4 x5 x6 (x7 = z1 z2), cluster(x8) gmm2s endogtest(x7) first

where the data are -xtset- accordingly.  It might be that the -partial()- option is the culprit.

Can either of you show us the actual output?

--Mark (-ivreg2- coauthor)

> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu 
> [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of 
> Mukaddes Yanik
> Sent: 24 April 2012 13:46
> To: statalist@hsphsun2.harvard.edu
> Subject: Re: st: ivreg2 doesn't report Kleinbergen-Paap statistic
> 
> I am experiencing a similar problem. I don't think it is 
> particularly related to clustering because I do get this 
> statistic in some other specifications that are also 
> estimated with clustering. But when I try some other 
> variables, Kleibergen-Paap rk LM statistic is missing. In my 
> case, I have two endogenous variables and when 
> (Kleibergen-Paap rk LM statistic) is missing, F test of 
> excluded instruments is not calculated either (it is reported 
> as 0.00) even though Angrist-Pischke multivariate F test of 
> excluded instruments is reported and significantly different 
> from zero. 
> 
> 
> --- On Tue, 4/24/12, SPRING Eva <eva.spring@unifr.ch> wrote:
> 
> > From: SPRING Eva <eva.spring@unifr.ch>
> > Subject: st: ivreg2 doesn't report Kleinbergen-Paap statistic
> > To: "statalist@hsphsun2.harvard.edu" 
> <statalist@hsphsun2.harvard.edu>
> > Date: Tuesday, April 24, 2012, 7:08 AM Hello, I am trying 
> to estimate 
> > a model where I have one endogenous variable and two excluded 
> > instruments. The equation I type into Stata looks the following way:
> > ivreg2 y x1 x2 x3 x4 x5 x6 FixEffect* (x7 = z1 z2),
> > cluster(x8) gmm2s partial(FixEffect*)  endogtest(x7) first
> > 
> > I was hoping to test the identification of my specification 
> > (underidentification test, weak identification test) with the 
> > Kleinbergen-Paap rk statistics. But unfortunately, Stata doesn't 
> > report this statistic. I have 450 observations, 1 endogenous 
> > regressor, 2 excluded instruments, 6 included
> > (exogenous) regressors, 118 variables are partialed out (the
> > FixEffect* are country dummies) and I have 131 clusters. I 
> work with 
> > an unbalanced panal dataset. After performing some tests myself, I 
> > found out that the problem lies in the clustering. If I 
> don't cluster, 
> > Stata reports the Kleinbergen-Paap statistics. However, this would 
> > yield wrong results and when I cluster, then I cannot test the 
> > validity of my instruments/results as Stata refuses to report the 
> > statistics I need to do this. I also updated all packages 
> of ivreg2, 
> > but it does not change the problem. I work with Stata/SE 12.1.
> > 
> > What can I do? Is there another way to proceed? What can I 
> do in order 
> > to get the Kleinbergen-Paap statistic? Or is there an 
> alternative to 
> > this statistic? Do I have too many clusters or what is the problem?
> > I would be very grateful for your help as I've been reading several 
> > Stata documents (among them Baum, Schaffer, Stillmann 2003, 
> 2007) and 
> > I've been searching the net for the past week without finding any 
> > answer or suggestion.
> > Greets,
> > Eva
> > 
> > 
> > *
> > *   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/
> > 
> 
> *
> *   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/
> 


-- 
Heriot-Watt University is the Sunday Times
Scottish University of the Year 2011-2012

We invite research leaders and ambitious early career
researchers to join us in leading and driving research
in key inter-disciplinary themes. Please see 

http://www.hw.ac.uk/researchleaders

for further information and how to apply.

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


*
*   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/