Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


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

RE: st: RE: Hausman-Taylor AR(1) estimator


From   KORAY ERCİHAN <[email protected]>
To   "[email protected]" <[email protected]>
Subject   RE: st: RE: Hausman-Taylor AR(1) estimator
Date   Fri, 7 Sep 2012 00:19:34 +0300

Dear Mark,

I first command: xi: xthtaylor ln_exports_od ln_gdp ln_sgdp dgdppc contig comlang_off ln_distcap euin i.year, endog( euin contig comlang_off)

Since xtoverid gives the following error
      xtoverid error: internal reestimation of eqn differs from original
      r(198)

I tried "xtoverid, noi" the output stata shows the Sargan Statistic chi-sq(3)=4.74 with p-value=0.1919

Then I started try different specifications

1- xi: xthtaylor ln_exports_od ln_gdp ln_sgdp dgdppc contig comlang_off ln_distcap euin i.year, endog( ln_gdp dgdppc comlang_off)
   
   xtoverid
      xtoverid error: internal reestimation of eqn differs from original
      r(198);
    
   xtoverid, noi
    Sargan statistics =     0.121
        Chi-sq(4) P-val =    0.9983 

2-xi: xthtaylor ln_exports_od ln_gdp ln_sgdp dgdppc contig comlang_off ln_distcap euin i.year, endog( ln_gdp dgdppc comlang_off) vce(bootstrap)

    xtoverid, noi gives the following result

   xtoverid, noi
      Warning - endogenous variable(s) collinear with instruments
      Vars now exogenous: __00000J __00000M __00000P __00000S __00000V __00000Y
                          __000011 __000014 __000017 __00001A __00001D __00001G
                          __00001J __00001M __00001P __00001S __00001Z __000020
      Warning - collinearities detected
      Vars dropped:       __00000O __00000R __00000U __00000X __000010 __000013
                          __000019 __00001C __00001F __00001I __00001L __00001O
                          __00001R __00000H __00000K __00000Q __00000T __00000W
                          __00000Z __000012 __000015 __000018 __00001B __00001E
                          __00001H __00001K __00001N __00001Q contig
      Warning: estimated covariance matrix of moment conditions not of full rank.
               standard errors and model tests should be interpreted with caution.
      Possible causes:
               number of clusters insufficient to calculate robust covariance matrix
               singleton dummy variable (dummy with one 1 and N-1 0s or vice versa)
      partial option may address problem.
      
      First-stage regressions
      -----------------------
      
      First-stage regression of __00001V:
      
      OLS estimation
      --------------
      
      Estimates efficient for homoskedasticity only
      Statistics robust to heteroskedasticity and clustering on pairs1
      
      Number of clusters (pairs1) = 168                     Number of obs =     2517
                                                            F( 26,   167) =  2.2e+16
                                                            Prob > F      =   0.0000
      Total (centered) SS     =  113.0830424                Centered R2   =   0.9995
      Total (uncentered) SS   =  186.3030374                Uncentered R2 =   0.9997
      Residual SS             =  .0529722933                Root MSE      =  .004611
      
      ------------------------------------------------------------------------------
                   |               Robust
          __00001V |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
      -------------+----------------------------------------------------------------
          __00000E |   -1564.76   1123.943    -1.39   0.166    -3783.727    654.2076
          __00000J |  -.5162412   .0351062   -14.71   0.000    -.5855503   -.4469321
          __00000M |  -.8659744   .6635963    -1.30   0.194    -2.176093    .4441444
          __00000P |   3.50e-08   4.98e-08     0.70   0.483    -6.33e-08    1.33e-07
          __00000S |   8.24e-08   1.85e-07     0.44   0.657    -2.84e-07    4.48e-07
          __00000V |   8.23e-08   1.88e-07     0.44   0.662    -2.88e-07    4.53e-07
          __00000Y |   7.54e-08   1.62e-07     0.46   0.643    -2.45e-07    3.96e-07
          __000011 |   8.80e-08   2.13e-07     0.41   0.681    -3.33e-07    5.09e-07
          __000014 |   8.60e-08   2.05e-07     0.42   0.676    -3.19e-07    4.91e-07
          __000017 |  -58.81884   45.06964    -1.31   0.194    -147.7985    30.16085
          __00001A |   6.93e-08   1.88e-07     0.37   0.714    -3.03e-07    4.41e-07
          __00001D |   1.01e-07   2.67e-07     0.38   0.705    -4.25e-07    6.27e-07
          __00001G |   1.58e-07   4.28e-07     0.37   0.712    -6.87e-07    1.00e-06
          __00001J |   2.42e-07   5.92e-07     0.41   0.683    -9.26e-07    1.41e-06
          __00001M |   2.56e-07   6.34e-07     0.40   0.687    -9.96e-07    1.51e-06
          __00001P |   2.79e-07   6.87e-07     0.41   0.685    -1.08e-06    1.64e-06
          __00001S |   3.29e-07   8.34e-07     0.39   0.694    -1.32e-06    1.97e-06
          __00001Z |   .1080825   .4199995     0.26   0.797    -.7211103    .9372752
          __000020 |  -222.1497   157.0535    -1.41   0.159     -532.216    87.91648
          __00000I |    .516241   .0351062    14.71   0.000     .4469319    .5855501
          __00000L |   .8659743   .6635963     1.30   0.194    -.4441445    2.176093
          __000016 |   58.81886   45.06966     1.31   0.194    -30.16085    147.7986
          __00001U |   .9999997   9.03e-07  1.1e+06   0.000     .9999979    1.000001
          __00001X |  -4.87e-08   9.65e-07    -0.05   0.960    -1.95e-06    1.86e-06
          __00000N |    151.295   106.8069     1.42   0.158    -59.57083    362.1609
        ln_distcap |   1.402371   .9924473     1.41   0.160    -.5569892     3.36173
      ------------------------------------------------------------------------------
      Warning: estimated covariance matrix of moment conditions not of full rank.
               standard errors and model tests should be interpreted with caution.
      Possible causes:
               number of clusters insufficient to calculate robust covariance matrix
               singleton dummy variable (dummy with one 1 and N-1 0s or vice versa)
      partial option may address problem.
      ------------------------------------------------------------------------------
      Included instruments: __00000E __00000J __00000M __00000P __00000S __00000V
                            __00000Y __000011 __000014 __000017 __00001A __00001D
                            __00001G __00001J __00001M __00001P __00001S __00001Z
                            __000020 __00000I __00000L __000016 __00001U __00001X
                            __00000N ln_distcap
      ------------------------------------------------------------------------------
      Partial R-squared of excluded instruments:   0.9958
      Test of excluded instruments:
        F(  7,   167) =  2.0e+11
        Prob > F      =   0.0000
      
      First-stage regression of __00001Y:
      
      OLS estimation
      --------------
      
      Estimates efficient for homoskedasticity only
      Statistics robust to heteroskedasticity and clustering on pairs1
      
      Number of clusters (pairs1) = 168                     Number of obs =     2517
                                                            F( 26,   167) =  1.6e+12
                                                            Prob > F      =   0.0000
      Total (centered) SS     =  159.7663429                Centered R2   =   0.9998
      Total (uncentered) SS   =  160.0142873                Uncentered R2 =   0.9998
      Residual SS             =  .0301100832                Root MSE      =  .003477
      
      ------------------------------------------------------------------------------
                   |               Robust
          __00001Y |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
      -------------+----------------------------------------------------------------
          __00000E |    294.893   57.42321     5.14   0.000      181.524     408.262
          __00000J |   .0115646   .0334573     0.35   0.730    -.0544891    .0776183
          __00000M |  -6.176372    .482563   -12.80   0.000    -7.129082   -5.223662
          __00000P |  -3.91e-09   1.21e-08    -0.32   0.748    -2.79e-08    2.00e-08
          __00000S |  -3.39e-09   4.52e-08    -0.07   0.940    -9.26e-08    8.58e-08
          __00000V |  -3.43e-09   4.57e-08    -0.07   0.940    -9.36e-08    8.68e-08
          __00000Y |  -3.38e-09   3.94e-08    -0.09   0.932    -8.11e-08    7.43e-08
          __000011 |  -3.51e-09   5.17e-08    -0.07   0.946    -1.06e-07    9.86e-08
          __000014 |  -3.47e-09   4.97e-08    -0.07   0.944    -1.02e-07    9.46e-08
          __000017 |   18.64497   2.276844     8.19   0.000     14.14987    23.14008
          __00001A |  -4.17e-09   4.56e-08    -0.09   0.927    -9.41e-08    8.58e-08
          __00001D |  -3.76e-09   6.46e-08    -0.06   0.954    -1.31e-07    1.24e-07
          __00001G |  -2.99e-09   1.04e-07    -0.03   0.977    -2.09e-07    2.03e-07
          __00001J |  -2.05e-09   1.44e-07    -0.01   0.989    -2.87e-07    2.83e-07
          __00001M |  -2.25e-09   1.55e-07    -0.01   0.988    -3.08e-07    3.03e-07
          __00001P |  -1.90e-09   1.68e-07    -0.01   0.991    -3.33e-07    3.29e-07
          __00001S |  -1.50e-09   2.03e-07    -0.01   0.994    -4.03e-07    4.00e-07
          __00001Z |  -.7893339   .1997606    -3.95   0.000    -1.183716   -.3949522
          __000020 |   40.77849   7.974215     5.11   0.000     25.03523    56.52176
          __00000I |  -.0115646   .0334572    -0.35   0.730    -.0776182    .0544891
          __00000L |   6.176372    .482563    12.80   0.000     5.223662    7.129082
          __000016 |  -18.64498   2.276856    -8.19   0.000    -23.14011   -14.14985
          __00001U |  -3.81e-09   2.21e-07    -0.02   0.986    -4.39e-07    4.32e-07
          __00001X |          1   2.34e-07  4.3e+06   0.000     .9999995           1
          __00000N |  -27.50739      5.438    -5.06   0.000    -38.24347    -16.7713
        ln_distcap |  -.2601827   .0505418    -5.15   0.000    -.3599658   -.1603996
      ------------------------------------------------------------------------------
      Warning: estimated covariance matrix of moment conditions not of full rank.
               standard errors and model tests should be interpreted with caution.
      Possible causes:
               number of clusters insufficient to calculate robust covariance matrix
               singleton dummy variable (dummy with one 1 and N-1 0s or vice versa)
      partial option may address problem.
      ------------------------------------------------------------------------------
      Included instruments: __00000E __00000J __00000M __00000P __00000S __00000V
                            __00000Y __000011 __000014 __000017 __00001A __00001D
                            __00001G __00001J __00001M __00001P __00001S __00001Z
                            __000020 __00000I __00000L __000016 __00001U __00001X
                            __00000N ln_distcap
      ------------------------------------------------------------------------------
      Partial R-squared of excluded instruments:   0.9971
      Test of excluded instruments:
        F(  7,   167) =  3.3e+12
        Prob > F      =   0.0000
      
      First-stage regression of __000021:
      
      OLS estimation
      --------------
      
      Estimates efficient for homoskedasticity only
      Statistics robust to heteroskedasticity and clustering on pairs1
      
      Number of clusters (pairs1) = 168                     Number of obs =     2517
                                                            F( 26,   167) =        .
                                                            Prob > F      =        .
      Total (centered) SS     =   .001764719                Centered R2   =   0.0149
      Total (uncentered) SS   =  .0017968437                Uncentered R2 =   0.0325
      Residual SS             =  .0017384285                Root MSE      =  .000835
      
      ------------------------------------------------------------------------------
                   |               Robust
          __000021 |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
      -------------+----------------------------------------------------------------
          __00000E |  -2.690367          .        .       .            .           .
          __00000J |  -.0096089          .        .       .            .           .
          __00000M |   .1063051          .        .       .            .           .
          __00000P |   2.23e-12          .        .       .            .           .
          __00000S |   2.61e-11          .        .       .            .           .
          __00000V |   2.59e-11          .        .       .            .           .
          __00000Y |   2.28e-11          .        .       .            .           .
          __000011 |   2.81e-11          .        .       .            .           .
          __000014 |   2.74e-11          .        .       .            .           .
          __000017 |   .2186006          .        .       .            .           .
          __00001A |   1.70e-11          .        .       .            .           .
          __00001D |   3.31e-11          .        .       .            .           .
          __00001G |   6.25e-11          .        .       .            .           .
          __00001J |   1.02e-10          .        .       .            .           .
          __00001M |   1.08e-10          .        .       .            .           .
          __00001P |   1.20e-10          .        .       .            .           .
          __00001S |   1.43e-10          .        .       .            .           .
          __00001Z |   .0015887          .        .       .            .           .
          __000020 |   -.392503          .        .       .            .           .
          __00000I |   .0096089          .        .       .            .           .
          __00000L |  -.1063051          .        .       .            .           .
          __000016 |  -.2186007          .        .       .            .           .
          __00001U |  -1.53e-10          .        .       .            .           .
          __00001X |  -3.55e-11          .        .       .            .           .
          __00000N |   .2439816          .        .       .            .           .
        ln_distcap |   .0025431          .        .       .            .           .
      ------------------------------------------------------------------------------
      Warning: estimated covariance matrix of moment conditions not of full rank.
               standard errors and model tests should be interpreted with caution.
      Possible causes:
               number of clusters insufficient to calculate robust covariance matrix
               singleton dummy variable (dummy with one 1 and N-1 0s or vice versa)
      partial option may address problem.
      ------------------------------------------------------------------------------
      Included instruments: __00000E __00000J __00000M __00000P __00000S __00000V
                            __00000Y __000011 __000014 __000017 __00001A __00001D
                            __00001G __00001J __00001M __00001P __00001S __00001Z
                            __000020 __00000I __00000L __000016 __00001U __00001X
                            __00000N ln_distcap
      ------------------------------------------------------------------------------
      Partial R-squared of excluded instruments:   0.0125
      Test of excluded instruments:
        F(  0,   167) =        .
        Prob > F      =        .
      
      
      
      Summary results for first-stage regressions
      -------------------------------------------
      
      Variable    | Shea Partial R2 |   Partial R2    |  F(  7,   167)    P-value
      __00001V    |     0.9933      |     0.9958      |     2.0e+11       0.0000
      __00001Y    |     0.9776      |     0.9971      |     3.3e+12       0.0000
      __000021    |     0.0122      |     0.0125      |           .            .
      
      NB: first-stage F-stat cluster-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(5)=2.77     P-val=0.7353
      Kleibergen-Paap rk Wald statistic           Chi-sq(5)=.        P-val=     .
      
      Weak identification test
      Ho: equation is weakly identified
      Kleibergen-Paap Wald rk F statistic                    .
      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(7,167)= 16.41     P-val=0.0000
      Anderson-Rubin Wald test     Chi-sq(7)=116.70    P-val=0.0000
      Stock-Wright LM S statistic  Chi-sq(7)=149.96    P-val=0.0000
      
      NB: Underidentification, weak identification and weak-identification-robust
          test statistics cluster-robust
      
      Number of clusters             N_clust  =        168
      Number of observations               N  =       2517
      Number of regressors                 K  =         22
      Number of instruments                L  =         26
      Number of excluded instruments       L1 =          7
      
      IV (2SLS) estimation
      --------------------
      
      Estimates efficient for homoskedasticity only
      Statistics robust to heteroskedasticity and clustering on pairs1
      
      Number of clusters (pairs1) = 168                     Number of obs =     2517
                                                            F( 22,   167) =   206.75
                                                            Prob > F      =   0.0000
      Total (centered) SS     =  1617.186606                Centered R2   =   0.6816
      Total (uncentered) SS   =  1651.722694                Uncentered R2 =   0.6883
      Residual SS             =  514.8476128                Root MSE      =    .4543
      
      ------------------------------------------------------------------------------
                   |               Robust
          __00000F |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
      -------------+----------------------------------------------------------------
          __00001V |   1.570001   .2646353     5.93   0.000     1.047539    2.092462
          __00001Y |  -1.466572   .4010023    -3.66   0.000    -2.258259   -.6748846
          __000021 |  -128.1563   112.4465    -1.14   0.256    -350.1562    93.84362
          __00000E |  -13.86337   14.29162    -0.97   0.333     -42.0789    14.35216
          __00000J |  -.6627835   .4548794    -1.46   0.147    -1.560839    .2352717
          __00000M |   .2074742   .0617934     3.36   0.001     .0854772    .3294712
          __00000P |   .0609254   .0493406     1.23   0.219    -.0364863    .1583371
          __00000S |   .0825058   .0825237     1.00   0.319    -.0804183    .2454299
          __00000V |   .0995982   .0854717     1.17   0.246     -.069146    .2683424
          __00000Y |   .2799819   .0823949     3.40   0.001     .1173121    .4426518
          __000011 |   .3832506   .0983113     3.90   0.000     .1891575    .5773438
          __000014 |   .4761469   .0998267     4.77   0.000     .2790621    .6732318
          __000017 |   .6613286   .1012605     6.53   0.000      .461413    .8612443
          __00001A |   .7126187   .1066265     6.68   0.000      .502109    .9231284
          __00001D |   .7121479   .1173461     6.07   0.000     .4804749    .9438209
          __00001G |   .6235858   .1556122     4.01   0.000     .3163652    .9308064
          __00001J |   .3948429   .2041333     1.93   0.055    -.0081716    .7978574
          __00001M |   .4205724   .2167743     1.94   0.054    -.0073988    .8485436
          __00001P |   .4553149   .2401146     1.90   0.060    -.0187363    .9293661
          __00001S |   .2717456   .2840648     0.96   0.340    -.2890753    .8325665
          __00001Z |  -1.177078   3.488846    -0.34   0.736    -8.065006     5.71085
          __000020 |  -1.021229   1.758372    -0.58   0.562    -4.492731    2.450274
      ------------------------------------------------------------------------------
      Underidentification test (Kleibergen-Paap rk LM statistic):              2.771
                                                         Chi-sq(5) P-val =    0.7353
      ------------------------------------------------------------------------------
      Weak identification test (Kleibergen-Paap rk Wald F statistic):              .
      Stock-Yogo weak ID test critical values:  5% maximal IV relative bias    13.95
                                               10% maximal IV relative bias     8.50
                                               20% maximal IV relative bias     5.56
                                               30% maximal IV relative bias     4.44
      Source: Stock-Yogo (2005).  Reproduced by permission.
      NB: Critical values are for Cragg-Donald F statistic and i.i.d. errors.
      ------------------------------------------------------------------------------
      Warning: estimated covariance matrix of moment conditions not of full rank.
               overidentification statistic not reported, and
               standard errors and model tests should be interpreted with caution.
      Possible causes:
               number of clusters insufficient to calculate robust covariance matrix
               singleton dummy variable (dummy with one 1 and N-1 0s or vice versa)
      partial option may address problem.
      ------------------------------------------------------------------------------
      Instrumented:         __00001V __00001Y __000021
      Included instruments: __00000E __00000J __00000M __00000P __00000S __00000V
                            __00000Y __000011 __000014 __000017 __00001A __00001D
                            __00001G __00001J __00001M __00001P __00001S __00001Z
                            __000020
      Excluded instruments: __00000I __00000L __000016 __00001U __00001X __00000N
                            ln_distcap
      Dropped collinear:    __00000O __00000R __00000U __00000X __000010 __000013
                            __000019 __00001C __00001F __00001I __00001L __00001O
                            __00001R __00000H __00000K __00000Q __00000T __00000W
                            __00000Z __000012 __000015 __000018 __00001B __00001E
                            __00001H __00001K __00001N __00001Q contig
      Reclassified as exog: __00000J __00000M __00000P __00000S __00000V __00000Y
                            __000011 __000014 __000017 __00001A __00001D __00001G
                            __00001J __00001M __00001P __00001S __00001Z __000020
      ------------------------------------------------------------------------------
      xtoverid error: internal reestimation of eqn differs from original
      r(198);


3- xi: xthtaylor ln_exports_od ln_gdp ln_sgdp dgdppc contig comlang_off ln_distcap euin i.year, endog( ln_gdp ln_sgdp dgdppc comlang_off)

xtoverid, noi detected collinearities

Similar output above and its overidentication test result;
    Sargan statistics:  0.113
    Chi-sq(3) P-val =    0.9903

xtoverid, noi cluster(pairs1) also says "warning: estimated covariance matrix of moment conditions not of full rank." This also happened when command xthtaylor with vce(bootstrap). Both of them don't report Sargan test statistics.

In that sense, what should I do? can I believe my Sargan test after xtoverid although collinearity exists? what about collinearity since I can't use either bootstrap or xtoverid, cluster(pairs1). Can I use Sargan statistics after xtoverid, noi for my Hausman-Taylor overidentification test. 

I will appreciate if you help me.



Thanks






> Subject: st: RE: Hausman-Taylor AR(1) estimator
> Date: Thu, 6 Sep 2012 20:45:27 +0100
> From: [email protected]
> To: [email protected]
> 
> Koray,
> 
> You'll need to show us the output with the error message about the lack
> of full rank etc. before we can offer advice about what it means and how
> to deal with it.
> 
> You can use xtoverid to get cluster-robust SEs; this is a way of
> addressing the within-panel autocorrelation. This post briefly
> describes how:
> 
> http://www.stata.com/statalist/archive/2011-03/msg00414.html
> 
> --Mark
> 
> > -----Original Message-----
> > From: [email protected] [mailto:owner-
> > [email protected]] On Behalf Of KORAY ERCIHAN
> > Sent: 06 September 2012 17:30
> > To: [email protected]
> > Subject: st: Hausman-Taylor AR(1) estimator
> > 
> > Dear Statalist,
> > 
> > The issues related to Hausman and Taylor estimator have been indicated
> in
> > Statalist but I couldn't find a solution for my problem.
> > I am using panel data including 168 bilateral trade relations under
> 1993-2007
> > years. I want to estimate Hausman-Taylor estimator (HT). Since I find
> > autocorrelation in my data, I want to estimate HT AR(1) too.
> > 
> > I estimated two HT models for my variables. One of them (HT1) is
> having 3
> > variables in endog part and the other (HT2) has 4 variables in endog
> part. I
> > have 7 explanatory variables at hand and I'm using time dummies as
> well. The
> > overidentification test results; 0.11 with p-value: 0.9903 and 0.12
> with p-
> > value: 0.9983 for HT1 and HT2 respectively. However, I obtained these
> test
> > results after commanding "xtoverid, noi" but when I use bootstrap
> option for
> > xthtaylor, it gives an error about the lack of full rank for
> covariance matrix
> > then I can't get the overidentification test result.
> > 
> > Do you think I can rely on my overidentification results since their
> p-values
> > are very high? and how can I solve the autocorrelation problem?
> > 
> > I will very grateful if you share your knowledge.
> > 
> > Thanks
> > 
> > Koray
> > 
> > *
> > * 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/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index