Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: st: RE: ivreg2 & endogenity


From   "Schaffer, Mark E" <M.E.Schaffer@hw.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: st: RE: ivreg2 & endogenity
Date   Thu, 2 Aug 2012 10:40:47 +0100

Ozgur,

The Hansen J stat is zero because the equation is exactly identified, as it states in the output.  You can get an explanation of this in any econometrics textbook.

The interpretation of the underidentification test is briefly discussed in the ivreg2 help file.  If you want more details, have a look at the references in the help file, starting with Baum et al. 2007.  In practice, though, people rarely comment on the underidentification stat, particularly when there is evidence that identification is not weak.

--Mark

> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-
> statalist@hsphsun2.harvard.edu] On Behalf Of Ozgur Ozdemir
> Sent: 02 August 2012 09:08
> To: Stata
> Subject: FW: st: RE: ivreg2 & endogenity
> 
> 
> Hi Mark,
> 
> thanks for your quick response. unfortunately I am new to ivreg2 and still
> learning. I have some more questions. I put the ivreg2 output below. I have 4
> variables which might be potentially endogenous and I am using the lags of
> them as instruments. in summary, it seems, the endog test result is giving p
> value of 0.2470 and as far as I know that is based on the regression anmong
> the error term and the exegenous variables so I should not be worried about
> endogentiy. Wald F for weak identification is also more than 10 and again to
> my knowledge rule of thumb if it is more than 10 that means the instruments
> are not week. but not sure how I should comment on the underidentification
> test and also Hansen J statistics ? thanks in advance.
> 
> 
> . xi : ivreg2 y (x1 x2  x3 x4= L.x1 L.x2 L.x3 L.x4) x5 x6 x7 x8 x9 x10 x11 x12 x13
> x14 x15  i.year  industry1 industry2 industry3 industry4 i
> > ndustry5 industry6 industry7 industry8 industry9 industry10 industry11
> > industry12 industry13 industry14 industry15, ffirst robust endog(x1
> > x2 x3 x4)
> i.year            _Iyear_2000-2010    (naturally coded; _Iyear_2000 omitted)
> Warning - collinearities detected Vars dropped:       _Iyear_2010 industry15
> 
> 
> Summary results for first-stage regressions
> -------------------------------------------
> 
> Variable    | Shea Partial R2 |   Partial R2    |  F(  4,  5455)    P-value
> x1          |     0.3792      |     0.5499      |      608.90       0.0000
> x2          |     0.3894      |     0.5678      |      672.62       0.0000
> x3          |     0.2277      |     0.2304      |      243.86       0.0000
> x4          |     0.4141      |     0.4148      |      114.91       0.0000
> 
> NB: first-stage F-stat heteroskedasticity-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)=643.07   P-val=0.0000
> Kleibergen-Paap rk Wald statistic           Chi-sq(1)=934.00   P-val=0.0000
> 
> Weak identification test
> Ho: equation is weakly identified
> Kleibergen-Paap Wald rk F statistic               231.84 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(4,5455)=10.41     P-val=0.0000 Anderson-Rubin Wald test     Chi-
> sq(4)=41.94     P-val=0.0000 Stock-Wright LM S statistic  Chi-sq(4)=41.14     P-
> val=0.0000
> 
> NB: Underidentification, weak identification and weak-identification-robust
>     test statistics heteroskedasticity-robust
> 
> Number of observations               N  =       5494 Number of
> regressors                 K  =         39 Number of instruments                L  =         39
> Number of excluded instruments       L1 =          4
> 
> IV (2SLS) estimation
> --------------------
> 
> Estimates efficient for homoskedasticity only Statistics robust to
> heteroskedasticity
> 
>                                                       Number of obs =     5494
>                                                       F( 38,  5455) =    79.40
>                                                       Prob > F      =   0.0000 Total (centered)
> SS     =  10145.90481                Centered R2   =   0.3783 Total (uncentered)
> SS   =        20003                Uncentered R2 =   0.6846 Residual
> SS             =  6308.043746                Root MSE      =    1.072
> 
> ------------------------------------------------------------------------------
>              |               Robust
>            y |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
> -------------+----------------------------------------------------------
> -------------+------
>           x1 |   .4208927   .1533017     2.75   0.006     .1204269    .7213586
>           x2 |  -1.401601   .4784664    -2.93   0.003    -2.339377   -.4638237
>           x3 |   .5369656   .1013611     5.30   0.000     .3383016    .7356296
>           x4 |   .0150771   .1499296     0.10   0.920    -.2787796    .3089337
>           x5 |   .1971591   .0604691     3.26   0.001     .0786418    .3156764
>           x6 |   1.610886   .0505437    31.87   0.000     1.511822    1.709949
>           x7 |   .0033538   .0020311     1.65   0.099     -.000627    .0073346
>           x8 |  -.0961036   .0331389    -2.90   0.004    -.1610547   -.0311525
>           x9 |  -.0323573   .0179385    -1.80   0.071    -.0675162    .0028015
>          x10 |   .1767141   .0582928     3.03   0.002     .0624622    .2909659
>          x11 |    .066056   .0259094     2.55   0.011     .0152745    .1168374
>          x12 |  -.0238117   .0089914    -2.65   0.008    -.0414345   -.0061888
>          x13 |    .403002   .0753469     5.35   0.000     .2553248    .5506792
>          x14 |  -16.08782   6.892246    -2.33   0.020    -29.59637   -2.579264
>          x15 |   .1660394   .0168025     9.88   0.000     .1331071    .1989717
>  _Iyear_2001 |   -.789929   .0726803   -10.87   0.000    -.9323797   -.6474783
>  _Iyear_2002 |   -.722612   .0704957   -10.25   0.000    -.8607811    -.584443
>  _Iyear_2003 |  -.5715681   .0716072    -7.98   0.000    -.7119156   -.4312205
>  _Iyear_2004 |  -.4202427   .0726249    -5.79   0.000    -.5625849   -.2779005
>  _Iyear_2005 |  -.2821438   .0764437    -3.69   0.000    -.4319707   -.1323169
>  _Iyear_2006 |   -.329362    .073308    -4.49   0.000    -.4730431   -.1856809
>  _Iyear_2007 |  -.2143846   .0716904    -2.99   0.003    -.3548952    -.073874
>  _Iyear_2008 |   -.239012   .0728622    -3.28   0.001    -.3818194   -.0962047
>  _Iyear_2009 |  -.1307312   .0772359    -1.69   0.091    -.2821108    .0206484
>    industry1 |  -.2370206   .0844791    -2.81   0.005    -.4025966   -.0714446
>    industry2 |  -.4670082   .0802934    -5.82   0.000    -.6243804    -.309636
>    industry3 |   .3948047    .101585     3.89   0.000     .1957018    .5939076
>    industry4 |   .0529189   .0808414     0.65   0.513    -.1055273    .2113651
>    industry5 |   .1684467   .0417607     4.03   0.000     .0865974    .2502961
>    industry6 |  -.9214761   .1204198    -7.65   0.000    -1.157494   -.6854577
>    industry7 |   -.632609    .077908    -8.12   0.000    -.7853058   -.4799122
>    industry8 |  -.7662738   .0570088   -13.44   0.000     -.878009   -.6545386
>    industry9 |   .1785934   .0623475     2.86   0.004     .0563946    .3007922
>   industry10 |   .3139048   .0646155     4.86   0.000     .1872608    .4405488
>   industry11 |   .6298456   .0783227     8.04   0.000     .4763358    .7833553
>   industry12 |   .3870012   .0724386     5.34   0.000     .2450241    .5289784
>   industry13 |  -.7891128   .0845906    -9.33   0.000    -.9549075   -.6233182
>   industry14 |  -.9416489   .1512921    -6.22   0.000    -1.238176   -.6451219
>        _cons |   -1.77156   .2013795    -8.80   0.000    -2.166256   -1.376863
> ------------------------------------------------------------------------------
> Underidentification test (Kleibergen-Paap rk LM statistic):            643.068
>                                                    Chi-sq(1) P-val =    0.0000
> ------------------------------------------------------------------------------
> Weak identification test (Kleibergen-Paap rk Wald F statistic):        231.843
> Stock-Yogo weak ID test critical values:                       <not available>
> ------------------------------------------------------------------------------
> Hansen J statistic (overidentification test of all instruments):         0.000
>                                                  (equation exactly identified)
> -endog- option:
> Endogeneity test of endogenous regressors:                               5.419
>                                                    Chi-sq(4) P-val =    0.2470 Regressors tested:    x1 x2
> x3 x4
> ------------------------------------------------------------------------------
> Instrumented:         x1 x2 x3 x4
> Included instruments: x5 x6 x7 x8 x9 x10 x11 x12 x13 x14 x15 _Iyear_2001
>                       _Iyear_2002 _Iyear_2003 _Iyear_2004 _Iyear_2005
>                       _Iyear_2006 _Iyear_2007 _Iyear_2008 _Iyear_2009 industry1
>                       industry2 industry3 industry4 industry5 industry6
>                       industry7 industry8 industry9 industry10 industry11
>                       industry12 industry13 industry14 Excluded instruments: L.x1 L.x2
> L.x3 L.x4 Dropped collinear:    _Iyear_2010 industry15
> ------------------------------------------------------------------------------
> 
> 
> 
> 
> kind regards
> Ozgur Ozdemir
> T: +44 (0) 75 0332 9865
> E: ozdemirozgur@hotmail.com
> Skype : ozgurozdemir2005
> 
> 
> > Subject: st: RE: ivreg2 & endogenity
> > Date: Tue, 31 Jul 2012 19:07:04 +0100
> > From: M.E.Schaffer@hw.ac.uk
> > To: statalist@hsphsun2.harvard.edu
> >
> > Ozgur,
> >
> > You can do either - it's up to you. In the first case, your null is that both x4
> and x5 are exogenous. In the second case, you are doing two separate tests;
> in one, the null is that x4 is exogenous; in the other, the null is that x5 is
> exogenous.
> >
> > It's analogous to testing the coeffs on, say, x1 and x2. You can test that they
> are jointly 0, or you can do separate tests of whether one or the other is 0.
> >
> > HTH,
> > Mark
> >
> > > -----Original Message-----
> > > From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-
> > > statalist@hsphsun2.harvard.edu] On Behalf Of Ozgur Ozdemir
> > > Sent: 29 July 2012 17:40
> > > To: Stata
> > > Subject: st: ivreg2 & endogenity
> > >
> > >
> > >
> > > Hi,
> > >
> > > I would like to test the endogenity of 2 variables using their
> > > 1-year lags as instruments. The model is
> > >
> > > ivreg2 y x1 x2 (x4 x5 = L.x4 L.x5), r
> > >
> > > the variables x4 and x5 might be endogenous.
> > >
> > >
> > > and would like to test whether x4 and/or x5 is endogenous. However
> > > not sure if I need to do test together
> > > ivreg2 y x1 x2 (x4 x5 = L.x4 L.x5), r endogtest(x4 x5)
> > >
> > >
> > > or separately one by one like the following
> > > ivreg2 y x1 x2 (x4 x5 = L.x4 L.x5), r endogtest(x4)
> > > ivreg2 y x1 x2 (x4 x5 = L.x4 L.x5), r endogtest(x5)
> > >
> > >
> > >
> > >
> > > kind regards
> > > Ozgur
> > >
> > > *
> > > * 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 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/
> 
> *
> *   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 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/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index