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: xtivreg2 weak-instrument-robust inference


From   "Schaffer, Mark E" <M.E.Schaffer@hw.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: st: RE: xtivreg2 weak-instrument-robust inference
Date   Wed, 14 Jul 2010 00:29:21 +0100

Camden,

Two suggestions:

First, update your ivreg2, xtivreg2 and ranktest to the latest versions.  I can see from the output that you are using an older version of ivreg2 at least.

Second, I can see from the output that you have a duplicate variables problem - you have listed 12 variables twice, e.g., notimpactedNOBlnclaimsuspt is listed as an exogenous regressor and also as an excluded instrument.  (xt)ivreg2 is dropping the duplicates, but it might not be dropping the occurences you want.  You should specify the equation without the duplicates to be sure you are getting what you want.

Hopefully one or other other of these suggestions will solve your problem.

--Mark

> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu 
> [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of 
> Camden Roberts
> Sent: 13 July 2010 16:36
> To: statalist@hsphsun2.harvard.edu
> Subject: Re: st: RE: xtivreg2 weak-instrument-robust inference
> 
> Dr. Schaffer,
> 
> Thank you for the response, I really appreciate it.  For what it is
> worth, I have included my output below in case there are any further
> insights anyone could provide.
> 
> . xtivreg2 lntotpremclass_avg notimpactedNOBlnchget
> notimpactedNOBlnclaimsuspt notimpactedNOBllntotprem
> notimpactedNOBllntotlim  notimpactedNOBuncertainty1 notimpactedNOB03
> notimpactedNOBperiod4 notimpactedNOBperiod5 impactedNOBlnchget
> impactedNOBlnclaimsuspt impactedNOBllntotprem impactedNOBllntotlim
> impactedNOBuncertainty1 impactedNOB03 impactedNOBperiod4
> impactedNOBperiod5 (lntotlimclass_avg=notimpactedNOBlnclaimsuspt
> notimpactedNOBlntotprem notimpactedNOBllntotprem
> impactedNOBlnclaimsuspt impactedNOBlntotprem impactedNOBllntotprem
> lnavg_totalassets lnavg_discontop lnavg_estdoubtrecv lnavg_investcap
> lnavg_debttoequity lnavg_pensionexp lnavg_netprofitmargin  lnavg_quick
> notimpactedNOBuncertainty1 notimpactedNOB03 notimpactedNOBperiod4
> notimpactedNOBperiod5  impactedNOBuncertainty1 impactedNOB03
> impactedNOBperiod4 impactedNOBperiod5), first  gmm2 fe
> Warning - singleton groups detected.  1 observation(s) not used.
> Warning - duplicate variables detected
> Duplicates:    notimpactedNOBlnclaimsuspt notimpactedNOBllntotprem
> impactedNOBlnclaimsuspt impactedNOBllntotprem
>                notimpactedNOBuncertainty1 notimpactedNOB03
> notimpactedNOBperiod4 notimpactedNOBperiod5
> impactedNOBuncertainty1 impactedNOB03 impactedNOBperiod4
> impactedNOBperiod5
> 
> FIXED EFFECTS ESTIMATION
> ------------------------
> Number of groups =        14                    Obs per 
> group: min =         7
>                                                               
>  avg =      14.3
>                                                               
>  max =        16
> 
> First-stage regressions
> -----------------------
> 
> First-stage regression of lntotlimclass_avg:
> 
> FIXED EFFECTS ESTIMATION
> ------------------------
> Number of groups =        14                    Obs per 
> group: min =         7
>                                                               
>  avg =      14.3
>                                                               
>  max =        16
> 
> OLS estimation
> --------------
> 
> Estimates efficient for homoskedasticity only
> Statistics consistent for homoskedasticity only
> 
>                                                       Number 
> of obs =      200
>                                                       F( 26,  
>  160) =    33.51
>                                                       Prob > 
> F      =   0.0000
> Total (centered) SS     =  20.97721972                
> Centered R2   =   0.8448
> Total (uncentered) SS   =  20.97721972                
> Uncentered R2 =   0.8448
> Residual SS             =  3.254944993                Root 
> MSE      =    .1426
> 
> --------------------------------------------------------------
> ----------------
> lntotlimcl~g |      Coef.   Std. Err.      t    P>|t|     
> [95% Conf. Interval]
> -------------+------------------------------------------------
> ----------------
> not~Blnchget |  -.0414577    .021775    -1.90   0.059    
> -.0844612    .0015458
> notimpacte.. |   .0924537   .0536503     1.72   0.087    
> -.0135004    .1984079
> notimpacte.. |    -.23722   .0588693    -4.03   0.000     
> -.353481    -.120959
> n~Bllntotlim |   .6098831   .0625331     9.75   0.000     
> .4863864    .7333799
> notimpacte.. |  -.0047151   .0557134    -0.08   0.933    
> -.1147434    .1053133
> notimpac~B03 |  -.3623428   .0795988    -4.55   0.000    
> -.5195426    -.205143
> not~Bperiod4 |  -.2170987   .0764569    -2.84   0.005    
> -.3680935   -.0661038
> not~Bperiod5 |  -.2226961   .0985589    -2.26   0.025    
> -.4173403   -.0280519
> impactedN~et |  -.0674339    .041296    -1.63   0.104    
> -.1489894    .0141215
> impactedN~pt |   .2319278   .1127642     2.06   0.041     
> .0092296    .4546259
> impactedNO.. |  -.2321764   .0859004    -2.70   0.008    
> -.4018213   -.0625315
> i~Bllntotlim |    .664471    .104069     6.38   0.000     
> .4589449     .869997
> impactedN~y1 |  -.0693409   .0858992    -0.81   0.421    
> -.2389833    .1003014
> impactedN~03 |  -.3001218   .1302113    -2.30   0.022    
> -.5572763   -.0429673
> impactedNO~4 |  -.0044267   .1007324    -0.04   0.965    
> -.2033633    .1945098
> impactedNO~5 |   -.134695   .1098232    -1.23   0.222    
> -.3515849    .0821949
> n~Blntotprem |   .4110068   .0483461     8.50   0.000     
> .3155279    .5064857
> i~Blntotprem |   .3657071   .0679038     5.39   0.000     
> .2316038    .4998105
> lnavg_tota~s |  -.0293143   .1880292    -0.16   0.876    
> -.4006535    .3420249
> lnavg_disc~p |  -.0392525   .0422026    -0.93   0.354    
> -.1225985    .0440936
> lnavg_estd~v |   .0115162   .0420492     0.27   0.785    
> -.0715268    .0945593
> lnavg_inve~p |   .0876884   .1661616     0.53   0.598    
> -.2404643    .4158411
> lnavg_debt~y |  -.0431072   .0301739    -1.43   0.155    
> -.1026977    .0164833
> lnavg_pens~p |   .0984751   .0529745     1.86   0.065    
> -.0061444    .2030946
> lnavg_netp~n |  -.0193907    .020282    -0.96   0.340    
> -.0594457    .0206643
>  lnavg_quick |   .0204199   .0260387     0.78   0.434    
> -.0310039    .0718437
> --------------------------------------------------------------
> ----------------
> Included instruments: notimpactedNOBlnchget notimpactedNOBlnclaimsuspt
>                       notimpactedNOBllntotprem notimpactedNOBllntotlim
>                       notimpactedNOBuncertainty1 notimpactedNOB03
>                       notimpactedNOBperiod4 notimpactedNOBperiod5
>                       impactedNOBlnchget impactedNOBlnclaimsuspt
>                       impactedNOBllntotprem impactedNOBllntotlim
>                       impactedNOBuncertainty1 impactedNOB03 
> impactedNOBperiod4
>                       impactedNOBperiod5 notimpactedNOBlntotprem
>                       impactedNOBlntotprem lnavg_totalassets 
> lnavg_discontop
>                       lnavg_estdoubtrecv lnavg_investcap 
> lnavg_debttoequity
>                       lnavg_pensionexp lnavg_netprofitmargin 
> lnavg_quick
> --------------------------------------------------------------
> ----------------
> Partial R-squared of excluded instruments:   0.4127
> Test of excluded instruments:
>   F( 10,   160) =    11.24
>   Prob > F      =   0.0000
> 
> 
> 
> Summary results for first-stage regressions
> -------------------------------------------
> 
> Variable    | Shea Partial R2 |   Partial R2    |  F( 10,   
> 160)    P-value
> lntotlimclas|     0.4127      |     0.4127      |       11.24 
>       0.0000
> 
> Underidentification tests
> Ho: matrix of reduced form coefficients has rank=K1-1 
> (underidentified)
> Ha: matrix has rank=K1 (identified)
> Anderson canon. corr. N*CCEV LM statistic   Chi-sq(10)=76.75  
>  P-val=0.0000
> Cragg-Donald N*CDEV Wald statistic          Chi-sq(10)=130.68 
>  P-val=0.0000
> 
> Weak identification test
> Ho: equation is weakly identified
> Cragg-Donald Wald F-statistic                      11.24
> 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(0,160)= .         P-val=     .
> Anderson-Rubin Wald test     Chi-sq(0)=.         P-val=     .
> Stock-Wright LM S statistic  Chi-sq(0)=186.00    P-val=     .
> 
> Number of observations               N  =        200
> Number of regressors                 K  =         17
> Number of instruments                L  =         26
> Number of excluded instruments       L1 =          0
> 
> 2-Step GMM estimation
> ---------------------
> 
> Estimates efficient for homoskedasticity only
> Statistics consistent for homoskedasticity only
> 
>                                                       Number 
> of obs =      200
>                                                       F( 17,  
>  169) =    15.61
>                                                       Prob > 
> F      =   0.0000
> Total (centered) SS     =   31.0461138                
> Centered R2   =   0.3902
> Total (uncentered) SS   =   31.0461138                
> Uncentered R2 =   0.3902
> Residual SS             =  18.93051061                Root 
> MSE      =     .319
> 
> --------------------------------------------------------------
> ----------------
> lntotpremc~g |      Coef.   Std. Err.      z    P>|z|     
> [95% Conf. Interval]
> -------------+------------------------------------------------
> ----------------
> lntotlimcl~g |   2.326373   .2109605    11.03   0.000     
> 1.912898    2.739848
> not~Blnchget |   .0933379   .0477317     1.96   0.051    
> -.0002145    .1868902
> notimpacte.. |  -.2315362   .1232914    -1.88   0.060     
> -.473183    .0101106
> notimpacte.. |   .6563093   .1128591     5.82   0.000     
> .4351095    .8775091
> n~Bllntotlim |  -1.552661   .1828108    -8.49   0.000    
> -1.910963   -1.194358
> notimpacte.. |   -.032614   .1162611    -0.28   0.779    
> -.2604816    .1952536
> notimpac~B03 |   .7392444   .1539487     4.80   0.000     
> .4375105    1.040978
> not~Bperiod4 |    .276797   .1151437     2.40   0.016     
> .0511194    .5024745
> not~Bperiod5 |   .2025452   .1635763     1.24   0.216    
> -.1180585    .5231489
> impactedN~et |   .1760424   .0889913     1.98   0.048     
> .0016226    .3504622
> impactedN~pt |  -.4269839   .2455661    -1.74   0.082    
> -.9082846    .0543169
> impactedNO.. |   .7653095   .1609672     4.75   0.000     
> .4498195    1.080799
> i~Bllntotlim |  -1.730799   .2568545    -6.74   0.000    
> -2.234225   -1.227374
> impactedN~y1 |   .1134413   .1793928     0.63   0.527    
> -.2381621    .4650447
> impactedN~03 |   .6359842   .2498756     2.55   0.011      
> .146237    1.125731
> impactedNO~4 |  -.1672743   .1807224    -0.93   0.355    
> -.5214837    .1869351
> impactedNO~5 |   .1720849   .2060627     0.84   0.404    
> -.2317907    .5759604
> --------------------------------------------------------------
> ----------------
> Underidentification test (Anderson canon. corr. LM 
> statistic):          76.755
>                                                    Chi-sq(10) 
> P-val =   0.0000
> --------------------------------------------------------------
> ----------------
> Weak identification test (Cragg-Donald Wald F statistic):     
>           11.241
> Stock-Yogo weak ID test critical values:  5% maximal IV 
> relative bias    20.74
>                                          10% maximal IV 
> relative bias    11.49
>                                          20% maximal IV 
> relative bias     6.61
>                                          30% maximal IV 
> relative bias     4.86
>                                          10% maximal IV size  
>            38.54
>                                          15% maximal IV size  
>            20.88
>                                          20% maximal IV size  
>            14.78
>                                          25% maximal IV size  
>            11.65
> Source: Stock-Yogo (2005).  Reproduced by permission.
> --------------------------------------------------------------
> ----------------
> Sargan statistic (overidentification test of all 
> instruments):          12.918
>                                                    Chi-sq(9) 
> P-val =    0.1664
> --------------------------------------------------------------
> ----------------
> Instrumented:         lntotlimclass_avg
> Included instruments: notimpactedNOBlnchget notimpactedNOBlnclaimsuspt
>                       notimpactedNOBllntotprem notimpactedNOBllntotlim
>                       notimpactedNOBuncertainty1 notimpactedNOB03
>                       notimpactedNOBperiod4 notimpactedNOBperiod5
>                       impactedNOBlnchget impactedNOBlnclaimsuspt
>                       impactedNOBllntotprem impactedNOBllntotlim
>                       impactedNOBuncertainty1 impactedNOB03 
> impactedNOBperiod4
>                       impactedNOBperiod5
> Excluded instruments: notimpactedNOBlntotprem impactedNOBlntotprem
>                       lnavg_totalassets lnavg_discontop 
> lnavg_estdoubtrecv
>                       lnavg_investcap lnavg_debttoequity 
> lnavg_pensionexp
>                       lnavg_netprofitmargin lnavg_quick
> Duplicates:           notimpactedNOBlnclaimsuspt 
> notimpactedNOBllntotprem
>                       impactedNOBlnclaimsuspt impactedNOBllntotprem
>                       notimpactedNOBuncertainty1 notimpactedNOB03
>                       notimpactedNOBperiod4 notimpactedNOBperiod5
>                       impactedNOBuncertainty1 impactedNOB03 
> impactedNOBperiod4
>                       impactedNOBperiod5
> --------------------------------------------------------------
> ----------------
> 
> .
> Thank you all again.
> 
> Respectfully,
> 
> Camden
> 
> 
> On Tue, Jul 13, 2010 at 11:12 AM, Schaffer, Mark E
> <M.E.Schaffer@hw.ac.uk> wrote:
> > Camden,
> >
> >> -----Original Message-----
> >> From: owner-statalist@hsphsun2.harvard.edu
> >> [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of
> >> Camden Roberts
> >> Sent: 13 July 2010 00:49
> >> To: statalist@hsphsun2.harvard.edu
> >> Subject: st: xtivreg2 weak-instrument-robust inference
> >>
> >> Hello all,
> >>
> >> I am attempting to use the xtivreg2 command and have a
> >> question about the weak-instrument-robust inference
> >> statistics.  In particular, when I estimate the model I
> >> receive the following missing statistics:
> >>
> >> 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(0,159)= .         P-val=     .
> >> Anderson-Rubin Wald test     Chi-sq(0)=.         P-val=     .
> >> Stock-Wright LM S statistic  Chi-sq(0)=186.00    P-val=     .
> >>
> >> With the exception of these missing statistics, the other
> >> tests for identification seem to suggest that the model is
> >> identified.  The output for these tests is below:
> >>
> >> --------------------------------------------------------------
> >> ----------------
> >> Partial R-squared of excluded instruments:   0.4127
> >> Test of excluded instruments:
> >>   F( 10,   160) =    11.24
> >>   Prob > F      =   0.0000
> >>
> >> Summary results for first-stage regressions
> >> -------------------------------------------
> >>
> >> Variable    | Shea Partial R2 |   Partial R2    |  F( 10,
> >> 160)    P-value
> >> lntotlimclas|     0.4127      |     0.4127      |       11.24
> >>       0.0000
> >>
> >> Underidentification tests
> >> Ho: matrix of reduced form coefficients has rank=K1-1
> >> (underidentified)
> >> Ha: matrix has rank=K1 (identified)
> >> Anderson canon. corr. N*CCEV LM statistic   Chi-sq(10)=76.75
> >>  P-val=0.0000
> >> Cragg-Donald N*CDEV Wald statistic          Chi-sq(10)=130.68
> >>  P-val=0.0000
> >>
> >> Weak identification test
> >> Ho: equation is weakly identified
> >> Cragg-Donald Wald F-statistic                      11.24
> >> See main output for Cragg-Donald weak id test critical values
> >>
> >> Underidentification test (Anderson canon. corr. LM
> >> statistic):          76.755
> >>                                                    Chi-sq(10)
> >> P-val =   0.0000
> >>
> >> If anyone has an idea of why I may be having this problem, I
> >> would greatly appreciate any assistance you could provide.
> >
> > You haven't given us the regression output, but my guess 
> would be that
> > you are using the cluster-robust covariance estimator, and 
> that have as
> > many or more excluded instruments than you have clusters (or panel
> > units).
> >
> > The artificial regression used for weak-ID-robust inference involves
> > testing all the excluded instruments.  If you have more excluded
> > instruments than clusters, you will run into problems because the
> > cluster-robust vcv won't be full rank.  In that case, the 
> maximum number
> > of constraints you can test will be the number of clusters 
> (minus 1, I
> > think), which isn't enough.
> >
> > But this is just a guess!
> >
> > --Mark
> >
> >> Thank you all,
> >>
> >>
> >> Camden Roberts
> >> *
> >> *   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 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 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