Statalist


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

RE: st: RE: Question on xthtaylor


From   "Schaffer, Mark E" <M.E.Schaffer@hw.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: st: RE: Question on xthtaylor
Date   Thu, 26 Jul 2007 08:32:13 +0100

Suryadipta, Rodrigo,

The way -xthtaylor- works is by transforming/creating variables, and
then using -regress- to estimate the transformed equation.  This
equation uses IV estimation.

The way -xtoverid- works is by doing the same transformation and then
using -ivreg2- to estimate it.  This generates a Sargan-Hansen overid
stat that is reported by -xtoverid-.  There's also an internal check
that the transformed regression results from -ivreg2- are the same as
those reported by the original -xthtaylor- estimation.

Calling -xtoverid- with the undocumented -noisily- option will cause it
to report this internal -ivreg2- estimation.  Suryadipta, if you do this
and -xtoverid- gets this far, you might be able to tell what's wrong or
get the overid stat from the output.

I suspect that what is causing the problem is the collinearity of yr1:
it gets dropped by -xthtaylor-, and this may be causing -xtoverid- to
get confused.  Omitting it from the estimation by hand may solve the
problem.

In any case, Suryadipta, if you can send me your dataset offlist so I
can reproduce the problem, I can fix the xtoverid bug.

--Mark

> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu 
> [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of 
> Roy,Suryadipta
> Sent: 26 July 2007 00:42
> To: statalist@hsphsun2.harvard.edu
> Subject: Re: st: RE: Question on xthtaylor
> 
> Dear Nick, Rodrigo, and Mark,
> 
> Thank you very much indeed for your suggestions- I shall 
> closely follow what Nick and Rodrigo had suggested as regards 
> the R-sq and the robust standard errors. Meanwhile here is 
> the output from -xthtaylor-, followed by
> -xtoverid- :
> 
> xthtaylor rhat2 corrupt x1 loggdp yr* logpopulat island 
> landlocked, endog(corrupt x1 loggdp)
> note: yr1 dropped due to collinearity
> Hausman-Taylor estimation                       Number of 
> obs      =      1712
> Group variable (i): id                          Number of 
> groups   =       121
>                                                  Obs per 
> group: min =         1
>                                                             
>     avg =      14.1
>                                                             
>     max =        16
> Random effects u_i ~ i.i.d.                     Wald 
> chi2(21)      =     45.12
>                                                  Prob > 
> chi2        =    0.0017
> --------------------------------------------------------------
> ----------------
>         rhat2 |      Coef.   Std. Err.      z    P>|z| 
>     [95% Conf. Interval]
> -------------+------------------------------------------------
> ----------
> -------------+------
> TVexogenous  |
>           yr2 |   .0035842   .0310919     0.12   0.908 
>    -.0573547    .0645231
>           yr3 |   .0281418   .0301447     0.93   0.351 
>    -.0309406    .0872243
>           yr4 |   .0379859    .029738     1.28   0.201 
>    -.0202995    .0962714
>           yr5 |   .0256643   .0296197     0.87   0.386 
>    -.0323892    .0837178
>           yr6 |   .0244164   .0298725     0.82   0.414 
>    -.0341325    .0829654
>           yr7 |   .0182896   .0302195     0.61   0.545 
>    -.0409396    .0775188
>           yr8 |    .017861    .030464     0.59   0.558 
>    -.0418472    .0775693
>           yr9 |   .0058436   .0305084     0.19   0.848 
>    -.0539517    .0656389
>          yr10 |   .0217139   .0307479     0.71   0.480 
>    -.0385508    .0819786
>          yr11 |   .0236914   .0310368     0.76   0.445 
>    -.0371396    .0845224
>          yr12 |  -.0067792   .0313415    -0.22   0.829 
>    -.0682073    .0546489
>          yr13 |  -.0104486   .0317474    -0.33   0.742 
>    -.0726723    .0517751
>          yr14 |  -.0093062   .0322576    -0.29   0.773 
>    -.0725299    .0539175
>          yr15 |  -.0097496   .0326688    -0.30   0.765 
>    -.0737793    .0542801
>          yr16 |  -.0071128   .0332515    -0.21   0.831 
>    -.0722844    .0580589
>    logpopulat |  -.0630136   .0467928    -1.35   0.178 
>    -.1547258    .0286985
> TVendogenous |
>       corrupt |   -.231095   .0438374    -5.27   0.000 
>    -.3170147   -.1451753
>            x1 |   .0316134   .0058632     5.39   0.000 
>     .0201217    .0431051
>        loggdp |  -.0366722   .0434294    -0.84   0.398 
>    -.1217922    .0484478
> TIexogenous  |
>        island |  -.0773091    .262035    -0.30   0.768 
>    -.5908883    .4362701
>    landlocked |   .0981921   .2297618     0.43   0.669 
>    -.3521327    .5485169
>               |
>         _cons |   1.175497   .9099886     1.29   0.196 
>    -.6080481    2.959042
> -------------+------------------------------------------------
> ----------
> -------------+------
>       sigma_u |  .88297205
>       sigma_e |  .20390525
>           rho |  .94937106   (fraction of variance due to 
> u_i)
> --------------------------------------------------------------
> ----------------
> note:  TV refers to time varying; TI refers to time invariant.
> . 
> . xtoverid
> . invalid name
> r(198);
> I am using Stata 9 and regularly update all the codes from 
> ssc. -xtoverid- ran alright when I used the example from 
> stata (using xtoverid.hlp).
> 
> Thanks again for your invaluable inputs!
> 
> Suryadipta.
> 
> 
> 
> 
> On Wed, 25 Jul 2007 22:40:40 +0100
>   "Schaffer, Mark E" <M.E.Schaffer@hw.ac.uk> wrote:
> > Suryadipta,
> > 
> >> -----Original Message-----
> >> From: owner-statalist@hsphsun2.harvard.edu
> >> [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of  
> >>Roy,Suryadipta
> >> Sent: 25 July 2007 18:44
> >> To: statalist@hsphsun2.harvard.edu
> >> Subject: st: Question on xthtaylor
> >> 
> >> Hi,
> >> 
> >> There have been a few threads on the first one, but I am  still 
> >>asking the question to find out if there is an easier solution:
> >> 
> >> 1. Is there a command to obtain robust/ cluster(robust)  standard 
> >>errors while using xthtaylor (as in xtreg, fe or xtreg, re)?
> >> 
> >> 2. Is there a command to obtain R-square while using xthtaylor?
> >> 
> >> 3. Moreover, after I run xthtaylor, I am trying to do the  
> >>overidentifiaction test with -xtoverid- but I am getting an  error 
> >>message : "invalid name". I only have time-varying  variables as 
> >>endogenous variables.
> > 
> > It's impossible to tell what's going wrong with just the 
> information  
> >you've provided.  Can you post the estimation results and xtoverid  
> >output, plus the versions of xthtaylor and xtoverid?
> > Then maybe we can
> > tell if it's a problem with your estimation, xthtaylor, or xtoverid.
> > 
> > --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
> > 
> > 
> >> 
> >> Any comment on these issues would be really helpful.
> >> Thanks as usual!
> >> Suryadipta.
> >> *
> >> *   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/
> 

*
*   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