Statalist


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

Re: st: Why option "robust" and "cluster" doesn't work?


From   "feldman" <feldman@primal.ucdavis.edu>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Why option "robust" and "cluster" doesn't work?
Date   Wed, 24 Oct 2007 10:00:01 -0700

My mistake, you are absolutely right. I was looking at the
BE_options in the help file instead of the FE_options. 

  

----- Original Message Follows -----
From: "Thomas Jacobs" <thomasjacobs@gmail.com>
To: statalist@hsphsun2.harvard.edu
Subject: Re: st: Why option "robust" and "cluster" doesn't
work?
Date: Wed, 24 Oct 2007 10:05:34 -0500

> Alejandro,
> 
> Version 10 certainly does permit using cluster with the
> vce option after either xtreg, fe or xtreg, re.
> 
> For example, on another post I made a day or so ago (more
> complicated than necessary but is all I have handy at the
> moment):
> 
>  xi: xtreg slope15y i.moodyrating|Mslope15y swap15slope,
> fe vce(cluster mktconum)
> i.moodyrating     _Imoodyrati_0-19    (naturally coded;
> _Imoodyrati_0 omitted) i.mood~gMs~15y   _ImooXMslo_#      
>  (coded as above) note: robust covariance is not full rank
> ; F test numerator degrees of freedom is 20   instead of
> >  21
> 
> Fixed-effects (within) regression               Number of
> obs      =    166351 Group variable: mktconum           
 
>           Number of groups   =       190
> 
> R-sq:  within  = 0.2843                         Obs per
> group: min =         9 between = 0.4482               
   
>                     avg =     875.5 overall = 0.3139  
   
>                                  max =      1493
> 
> F(20,189)          =    340.39
> corr(u_i, Xb)  = 0.1384                         Prob > F
 
>         =    0.0000
> 
> (Std. Err. adjusted for 190 clusters in mktconum)
> 
> Robust
> slope15y       Coef.   Std. Err.      t    P>t     [95%
> Conf. Interval]
> 
> Mslope15y    .8421038   .1025912     8.21   0.000     
> 639733    1.044475 _ImooXMslo_1   -.2351284   .0403382   
> -5.83   0.000    -.3146995   -.1555574 _ImooXMslo_2  
> -.6510598   .1111743    -5.86   0.000    -.8703616   
> -.431758 _ImooXMslo_3   -.5825345   .1362484    -4.28  
> 0.000    -.8512975   -.3137716 _ImooXMslo_4   -.8162742  
> 1011095    -8.07   0.000    -1.015722   -.6168261
> _ImooXMslo_5   -.8239694    .102474    -8.04   0.000   
> -1.026109   -.6218296 _ImooXMslo_6   -.8419379   .1045057 
>   -8.06   0.000    -1.048085   -.6357904 _ImooXMslo_7  
> -.8438201   .1048161    -8.05   0.000     -1.05058  
> -.6370603 _ImooXMslo_8   -.8060444       .105    -7.68  
> 0.000    -1.013167   -.5989219 _ImooXMslo_9   -.8368847  
> 1060561    -7.89   0.000     -1.04609    -.627679
> _ImooXMsl~10   -.8185907   .1042997    -7.85   0.000   
> -1.024332   -.6128495 _ImooXMsl~11   -.7526583   .1029494 
>   -7.31   0.000    -.9557356   -.5495809 _ImooXMsl~12  
> -.8299127   .1012824    -8.19   0.000    -1.029702  
> -.6301234 _ImooXMsl~13   -.7959784   .1073662    -7.41  
> 0.000    -1.007768   -.5841882 _ImooXMsl~14    -.854819   
> 102767    -8.32   0.000    -1.057537   -.6521013
> _ImooXMsl~15   -.7806554   .1084974    -7.20   0.000   
> -.9946768   -.5666341 _ImooXMsl~16   -.9274818   .1939613 
>   -4.78   0.000    -1.310089   -.5448747 _ImooXMsl~17   
> -.842949   .1033795    -8.15   0.000    -1.046875  
> -.6390232 _ImooXMsl~18    .0061745   .1018397     0.06  
> 0.952    -.1947139    .2070629 _ImooXMsl~19   -.6648868  
> 1018489    -6.53   0.000    -.8657935   -.4639801
> swap15slope   -.1419077    .030575    -4.64   0.000   
> -.2022198   -.0815956 _cons    .0048621   .0003458   
> 14.06   0.000     .0041799    .0055442
> 
> sigma_u   .00455044
> sigma_e   .01011707
> rho   .16826102   (fraction of variance due to u_i)
> 
> On 10/24/07, feldman <feldman@primal.ucdavis.edu> wrote:
> > But if you use xtreg, fe the vce() option doesn´t
allow
> > cluster so that still doesn´t answer why cluster after
> > xtreg, fe was allowed in Stata 9 but not in Stata 10.
> >
> > Alejandro
> >
> > ----- Original Message Follows -----
> > From: "E. Paul Wileyto" <epw@mail.med.upenn.edu>
> > To: statalist@hsphsun2.harvard.edu
> > Subject: Re: st: Why option "robust" and "cluster"
> > doesn't work?
> > Date: Tue, 23 Oct 2007 13:40:51 -0400
> >
> > > Try using the more general vce option.  I got
> > > surprised by this earlier  as well, and found that
> > > while cluster() still works for many of the  routines,
> > > vce will often work when cluster will not.
> > >
> > > I can see why the Stata Gods have deemed this good. 
> > > It subsumes  clustering, bootstrapping, jackknife, and
> > > all manner of methods under  one option.
> > >
> > > Paul
> > >
> > > feldman wrote:
> > > > But why was the cluster option available in Stata 9
> > > > and is not in Stata 10?
> > > >
> > > > ----- Original Message Follows -----
> > > > From: "Schaffer, Mark E" <M.E.Schaffer@hw.ac.uk>
> > > > To: <statalist@hsphsun2.harvard.edu>
> > > > Subject: RE: st: Why option "robust" and "cluster"
> > > > doesn't work?
> > > > Date: Sun, 21 Oct 2007 21:10:39 +0100
> > > >
> > > >
> > > >> Ngoc Anh Vo Thi,
> > > >>
> > > >>
> > > >>> -----Original Message-----
> > > >>> From: owner-statalist@hsphsun2.harvard.edu
> > > >>> [mailto:owner-statalist@hsphsun2.harvard.edu] On
> > > Behalf >>> Of  Maarten buis
> > > >>> Sent: 21 October 2007 10:49
> > > >>> To: statalist@hsphsun2.harvard.edu
> > > >>> Subject: Re: st: Why option "robust" and "cluster"
> > > >>> doesn't work?
> > > >>> --- Ngoc Anh Vo Thi <vtnanh@yahoo.com> wrote:
> > > >>>
> > > >>>> I'm estimating a panel data model with xtreg, fe
> > > using >>>>
> > > >>> Stata 8.2, and
> > > >>>
> > > >>>> want to obtain robust standard errors and test
> > > stat. >>>>
> > > >>> However, when I
> > > >>>
> > > >>>> add the "robust"/ "cluster" option after fe, it
> > > >>>>
> > > >>> doesn't  work. It was
> > > >>>
> > > >>>> said "Option robust not allowed". I'm wondering
> > > >>>>
> > > >>> whether  there's a bug
> > > >>>
> > > >>>> with my version of Stata.
> > > >>>>
> > > >>> There isn't a bug in your version of Stata, that
> > > option >>> is  not allowed. If you type -help xtreg-
> > > you will find >>> a list of  legal options, and
> > > -robust- isn't in that >>>
> > > >> list, nor is it in Stata 10.
> > > >>
> > > >> It's available in Stata 9, but that doesn't help
> > > someone >> who has only Stata 8.2 available.
> > > >>
> > > >> -xtivreg2- will do you what you want, though - it
> > > requires >> only Stata 8.2.  You'll also need to
> > > install -ivreg28-, >> which, as the "8" suggests, also
> > > runs in Stata 8.2.  Just >> -findit xtivreg2- and
> > > -findit ivreg28- from within Stata, >> and follow the
> > > links to install. >>
> > > >> But - note that Stock and Watson (2006), in an NBER
> > > >> technical working paper
> > > >>
> > > >> http://www.nber.org/papers/t0323
> > > >>
> > > >> show that the heteroskedasticity-robust covariance
> > > >> estimator is not consistent with fixed effects. 
> > > The >> cluster-robust estimator is OK, though.
> > > >>
> > > >> HTH.
> > > >>
> > > >> 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
> > > >>
> > > >>
> > > >>
> > > >>> Hope this helps,
> > > >>> Maarten
> > > >>>
> > > >>> -----------------------------------------
> > > >>> Maarten L. Buis
> > > >>> Department of Social Research Methodology Vrije
> > > >>> Universiteit  Amsterdam Boelelaan 1081
> > > >>> 1081 HV Amsterdam
> > > >>> The Netherlands
> > > >>>
> > > >>> visiting address:
> > > >>> Buitenveldertselaan 3 (Metropolitan), room Z434
> > > >>>
> > > >>> +31 20 5986715
> > > >>>
> > > >>> http://home.fsw.vu.nl/m.buis/
> > > >>> -----------------------------------------
> > > >>>
> > > >>>
> > > >>>
> > > >>>
> > > >>
> > >
> > >
> __________________________________________________________
> > > >>      >>> _ Want ideas for reducing your carbon
> > > footprint? Visit >>> Yahoo!  For Good >>>
> > >
> > >
> http://uk.promotions.yahoo.com/forgood/environment.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/ >
> > > >
> > >
> > > --
> > > E. Paul Wileyto, Ph.D.
> > > Assistant Professor of Biostatistics
> > > Tobacco Use Research Center
> > > School of Medicine, U. of Pennsylvania
> > > 3535 Market Street, Suite 4100
> > > Philadelphia, PA  19104-3309
> > >
> > > 215-746-7147
> > > Fax: 215-746-7140
> > > epw@mail.med.upenn.edu
> > > http://mail.med.upenn.edu/~epw/
> > >
> > > *
> > > *   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/
> >
> 
> 
> 
> -- 
> Thomas Jacobs
> 
> *
> *   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