"xttest0" after "xtivreg, re ec2sls"

[Josep Rialp]

Dear all,

My name is Josep Rialp.

I'm using "xtivreg, re ec2sls" and I would like to check if random effects estimations are more convenient than a pooled approach (when I implement "xtivreg, fe" the F test to compare fixed effects versus the pooled option is provided directly).

I'm trying to run "xttest0" command but the comment that I receive is the folllowing:

"last estimates not found"

I also obtain this message when I run "estimates store random" after running "xtivreg, re ec2sls".

I wonder if someone has some advice or suggestion for me.

Thank all of you in advance for your attention.

Josep

[Carlo Lazzaro]

Josep:
welcome to this forum.
If your code does not include -fvvarlist- notation, you can check what is the best specification for your panel regression model via the community-contributed command -xtoverid- (just type -search xtoverid- to spot and install it):

. use https://www.stata-press.com/data/r16/nlswork
(National Longitudinal Survey.  Young Women 14-26 years of age in 1968)

. xi: xtivreg ln_w age not_smsa (tenure = union birth south), re ec2sls

EC2SLS random-effects IV regression             Number of obs     =     19,007
Group variable: idcode                          Number of groups  =      4,134

R-sq:                                           Obs per group:
     within  = 0.0893                                         min =          1
     between = 0.2276                                         avg =        4.6
     overall = 0.1651                                         max =         12

                                                Wald chi2(3)      =    2358.86
corr(u_i, X)       = 0 (assumed)                Prob > chi2       =     0.0000

------------------------------------------------------------------------------
     ln_wage |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
      tenure |   .0617952   .0024766    24.95   0.000     .0569412    .0666492
         age |  -.0023111   .0008107    -2.85   0.004    -.0039001   -.0007221
    not_smsa |  -.2252714   .0081015   -27.81   0.000      -.24115   -.2093928
       _cons |   1.650852   .0192063    85.95   0.000     1.613208    1.688495
-------------+----------------------------------------------------------------
     sigma_u |    .328017
     sigma_e |  .64359089
         rho |  .20619825   (fraction of variance due to u_i)
------------------------------------------------------------------------------
Instrumented:   tenure
Instruments:    age not_smsa union birth_yr south
------------------------------------------------------------------------------

. xtoverid

Test of overidentifying restrictions:
Cross-section time-series model: xtivreg ec2sls  
Sargan-Hansen statistic 679.160  Chi-sq(7)    P-value = 0.0000

.

[Josep Rialp]

Thank you very much Carlo!!!

[Vivek Sarati]

Josep:
welcome to this forum.
If your code does not include -fvvarlist- notation, you can check what is the best specification for your panel regression model via the community-contributed command -xtoverid- (just type -search xtoverid- to spot and install it):

. use https://www.stata-press.com/data/r16/nlswork
(National Longitudinal Survey. Young Women 14-26 years of age in 1968)

. xi: xtivreg ln_w age not_smsa (tenure = union birth south), re ec2sls

EC2SLS random-effects IV regression Number of obs = 19,007
Group variable: idcode Number of groups = 4,134

R-sq: Obs per group:
within = 0.0893 min = 1
between = 0.2276 avg = 4.6
overall = 0.1651 max = 12

Wald chi2(3) = 2358.86
corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000

------------------------------------------------------------------------------
ln_wage | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
tenure | .0617952 .0024766 24.95 0.000 .0569412 .0666492
age | -.0023111 .0008107 -2.85 0.004 -.0039001 -.0007221
not_smsa | -.2252714 .0081015 -27.81 0.000 -.24115 -.2093928
_cons | 1.650852 .0192063 85.95 0.000 1.613208 1.688495
-------------+----------------------------------------------------------------
sigma_u | .328017
sigma_e | .64359089
rho | .20619825 (fraction of variance due to u_i)
------------------------------------------------------------------------------
Instrumented: tenure
Instruments: age not_smsa union birth_yr south
------------------------------------------------------------------------------

. xtoverid

Test of overidentifying restrictions:
Cross-section time-series model: xtivreg ec2sls
Sargan-Hansen statistic 679.160 Chi-sq(7) P-value = 0.0000

.

Hi Mr. Carlo, what does significant P-value for Sargan Hansen statistic means for overidentification?

[Carlo Lazzaro]

Vivek:
a rejecton of the null

...casts doubt on the validity of the instruments.

(quoted from the help file of the community-contributed module -xtoverid-).

As an aside, please call me Carlo, just like all on (and many more off) this list do. Thanks.

[Vivek Sarati]

Vivek:
a rejecton of the null (quoted from the help file of the community-contributed module -xtoverid-).

As an aside, please call me Carlo, just like all on (and many more off) this list do. Thanks.

Hi Carlo. Thanks.

Is there any endogeneity or other postestimation tests for Xtivreg Random effect in state?

[Vivek Sarati]

stata*