st: AW: Use of xtabond2

["Dithmer, Jan" <jdithme@food-econ.uni-kiel.de>]

Dear Mavilde,

you should first work through the paper from the author of the routine xtabond2, which describes the program in great detail:

David Roodman, 2009. "How to do xtabond2: An introduction to difference and system GMM in Stata," Stata Journal, StataCorp LP, vol. 9(1), pages 86-136, March.

The generalized inverse to calculate the optimal weighting matrix is generated automatically.
The warning "Number of instruments may be large relative to number of observations" points to the fact
that your number of instruments is higher than your number of groups as a rule of thumb to indicate that there
may be problems of instrument proliferation.

To understand the potential problems coming with a large instrument set, read:

David Roodman, 2009. "A Note on the Theme of Too Many Instruments," Oxford Bulletin of Economics and Statistics, Department of Economics, University of Oxford, vol. 71(1), pages 135-158, 02.


Best, Jan


-----Ursprüngliche Nachricht-----
Von: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] Im Auftrag von Mavilde Modesto
Gesendet: Sunday, June 09, 2013 12:36 AM
An: 'Sergiy Radyakin'
Cc: owner-statalist@hsphsun2.harvard.edu
Betreff: st: Use of xtabond2

Dear Statalist,

I am a new user of Stata and I do not Know enough to solve those Two "Warnings" I have got  running xtabond2. 

The first one I do not understand because I use 49 instruments but I have got 510 observations.

About the second warning anyone can help me indicating which command should I use to produce and use the generalized inverse to calculate optimal weighting matrix for two-step estimation and how to proceed?

I would be so grateful!

Here are the results I have got:

xi: xtabond2 logGVApc l.logGVApc logPubInvpc logProductiv
i.yearid,gmm(logGVApc,lag(2 2)) iv(i.yearid) robust twostep small/*2 lag*//*considering Public Investment as exogenous*/ i.yearid _Iyearid_1-18 (naturally coded; _Iyearid_1 omitted) Favoring speed over space. To switch, type or click on mata: mata set matafavor space, perm.
_Iyearid_18 dropped due to collinearity

Warning: Number of instruments may be large relative to number of observations.
Warning: Two-step estimated covariance matrix of moments is singular.
Using a generalized inverse to calculate optimal weighting matrix for two-step estimation.
Difference-in-Sargan/Hansen statistics may be negative.

Dynamic panel-data estimation, two-step system GMM Group variable: id Number of obs = 510 Time variable : Periodo Number of groups = 30 Number of instruments = 49 Obs per group: min = 17 avg = 17.00 F(19, 29) = 967.14 Prob > F = 0.000 max = 17

logGVApc Coef. Corrected Std. Err. t P>t [95% Conf. Interval]

logGVApc
L1. .682456 .1063249 6.42 0.000 .4649971 .8999149

logPubInvpc -.0131582 .033408 -0.39 0.697 -.0814852 .0551688 logProductiv .3087351 .1130752 2.73 0.011 .0774704 .5399997
_Iyearid_2 -.0068182 .0563709 -0.12 0.905 -.1221097 .1084733
_Iyearid_3 .0011161 .0454337 0.02 0.981 -.0918064 .0940385
_Iyearid_4 -.0114692 .0372671 -0.31 0.760 -.0876889 .0647506
_Iyearid_5 .0045814 .0361524 0.13 0.900 -.0693586 .0785214
_Iyearid_6 -.0350082 .03002 -1.17 0.253 -.096406 .0263896
_Iyearid_7 -.0075198 .0267703 -0.28 0.781 -.0622712 .0472316
_Iyearid_8 -.0195811 .028722 -0.68 0.501 -.0783242 .039162
_Iyearid_9 .0031761 .020015 0.16 0.875 -.0377591 .0441114 _Iyearid_10 .0003316 .0215934 0.02 0.988 -.0438319 .044495
_Iyearid_11 .0088895 .0172075 0.52 0.609 -.0263038 .0440827
_Iyearid_12 .0194218 .0134674 1.44 0.160 -.0081222 .0469658
_Iyearid_13 .0252198 .0125812 2.00 0.054 -.0005116 .0509512
_Iyearid_14 .0273918 .0257642 1.06 0.296 -.0253019 .0800856
_Iyearid_15 .011291 .0230309 0.49 0.628 -.0358124 .0583944
_Iyearid_16 .0043055 .0191247 0.23 0.823 -.0348089 .0434199
_Iyearid_17 .0141141 .0165234 0.85 0.400 -.0196801 .0479083 _cons -.2379976 .3225759 -0.74 0.467 -.8977395 .4217443 Instruments for first differences equation Standard
D.(_Iyearid_2 _Iyearid_3 _Iyearid_4 _Iyearid_5 _Iyearid_6 _Iyearid_7
_Iyearid_8 _Iyearid_9 _Iyearid_10 _Iyearid_11 _Iyearid_12 _Iyearid_13
_Iyearid_14 _Iyearid_15 _Iyearid_16 _Iyearid_17 _Iyearid_18) GMM-type (missing=0, separate instruments for each period unless collapsed) L2.logGVApc Instruments for levels equation Standard
_Iyearid_2 _Iyearid_3 _Iyearid_4 _Iyearid_5 _Iyearid_6 _Iyearid_7
_Iyearid_8 _Iyearid_9 _Iyearid_10 _Iyearid_11 _Iyearid_12 _Iyearid_13
_Iyearid_14 _Iyearid_15 _Iyearid_16 _Iyearid_17 _Iyearid_18 _cons GMM-type (missing=0, separate instruments for each period unless collapsed) DL.logGVApc

Arellano-Bond test for AR(1) in first differences: z = -2.60 Pr > z = 0.009 Arellano-Bond test for AR(2) in first differences: z = 1.33 Pr > z = 0.185

Sargan test of overid. restrictions: chi2(29) = 68.55 Prob > chi2 = 0.000 (Not robust, but not weakened by many instruments.)

Hansen test of overid. restrictions: chi2(29) = 17.64 Prob > chi2 = 0.951 (Robust, but weakened by many instruments.) Difference-in-Hansen tests of exogeneity of instrument subsets:
GMM instruments for levels

Hansen test excluding group: chi2(13) = 17.64 Prob > chi2 = 0.172 Difference (null H = exogenous): chi2(16) = 0.00 Prob > chi2 = 1.000
iv(_Iyearid_2 _Iyearid_3 _Iyearid_4 _Iyearid_5 _Iyearid_6 _Iyearid_7
_Iyearid_8 _Iyearid_9 _Iyearid_10 _Iyearid_11 _Iyearid_12 _Iyearid_13
_Iyearid_14 _Iyearid_15 _Iyearid_16 _Iyearid_17 _Iyearid_18) Hansen test excluding group: chi2(13) = 10.44 Prob > chi2 = 0.657 Difference (null H = exogenous): chi2(16) = 7.20 Prob > chi2 = 0.969

Mavilde Modesto
Universidade Católica Portuguesa



*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/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/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/