Hello everyone
I am trying to estimate a dynamic model based on a growth regression. For that reason I applied several techniques such as FE or POLS being Arellano Bond and Arellano Bover my preferred options due to the dynamic nature of the estimation.
My panel data is an unbalanced panel covering the EU country members for a total of 11 periods. Each period accounts for 5 years to minimize endogneity.
The model looks like:
y_{i,t}-y_{i,t-s}=\beta_1y_{i,t-s} + \beta_2 \textbf{X}_{i,t-s}+ \beta_3 d_{i,t-s} + u_{i,t}
Y stands for the log real GDP , X is a set of control variables used in growth regressions and d is the initial public debt, the variable I am interested in.
These are the results for POLs and FE ( variables were corrected by hand so lags do not appear in the commands though are present):
So far results seem reasonable but once i turn to the DGMM estimators:
I limited the number of instruments and tried applying time dummies which solver the problem or the arellano bond test but adds too many instruments and insignificant results even if i limit the later. So far best specification achieved has been produced by :
I have tried to read other FAQ and I also read Roodmans paper on xtabond2 but at his point I can not figure out what's wrong.
Thank you for your time
I am trying to estimate a dynamic model based on a growth regression. For that reason I applied several techniques such as FE or POLS being Arellano Bond and Arellano Bover my preferred options due to the dynamic nature of the estimation.
My panel data is an unbalanced panel covering the EU country members for a total of 11 periods. Each period accounts for 5 years to minimize endogneity.
The model looks like:
y_{i,t}-y_{i,t-s}=\beta_1y_{i,t-s} + \beta_2 \textbf{X}_{i,t-s}+ \beta_3 d_{i,t-s} + u_{i,t}
Y stands for the log real GDP , X is a set of control variables used in growth regressions and d is the initial public debt, the variable I am interested in.
These are the results for POLs and FE ( variables were corrected by hand so lags do not appear in the commands though are present):
Code:
reg gdpcg lgdpc `exog' , r
Linear regression Number of obs = 198
F(7, 190) = 9.22
Prob > F = 0.0000
R-squared = 0.2836
Root MSE = .11154
------------------------------------------------------------------------------
| Robust
gdpcg | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
lgdpc | -.1441605 .0251156 -5.74 0.000 -.1937017 -.0946192
gg | .0000703 .0002976 0.24 0.814 -.0005168 .0006573
popgrowth | -5.159994 2.215187 -2.33 0.021 -9.529514 -.7904749
liquid_gdp | .0007237 .000325 2.23 0.027 .0000825 .0013648
inflation | .0000228 .0001006 0.23 0.821 -.0001756 .0002211
tradeop | .2444825 .1020323 2.40 0.018 .0432209 .4457441
govs_peen | -.2621709 .215216 -1.22 0.225 -.6866905 .1623486
_cons | 1.601594 .2559984 6.26 0.000 1.09663 2.106558
------------------------------------------------------------------------------
xtreg gdpcg lgdpc `exog' , r fe
Fixed-effects (within) regression Number of obs = 198
Group variable: panelid Number of groups = 28
R-sq: Obs per group:
within = 0.3109 min = 2
between = 0.2194 avg = 7.1
overall = 0.2373 max = 11
F(7,27) = 21.74
corr(u_i, Xb) = -0.4754 Prob > F = 0.0000
(Std. Err. adjusted for 28 clusters in panelid)
------------------------------------------------------------------------------
| Robust
gdpcg | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
lgdpc | -.1971971 .0296721 -6.65 0.000 -.2580791 -.136315
gg | .0012151 .000485 2.51 0.019 .00022 .0022101
popgrowth | -6.926695 2.321577 -2.98 0.006 -11.69018 -2.163213
liquid_gdp | .0008602 .0002566 3.35 0.002 .0003336 .0013867
inflation | -.000201 .0001369 -1.47 0.154 -.000482 .00008
tradeop | .2623044 .1197898 2.19 0.037 .016516 .5080927
govs_peen | -.7567418 .3109876 -2.43 0.022 -1.394836 -.118648
_cons | 2.164377 .2984268 7.25 0.000 1.552056 2.776699
-------------+----------------------------------------------------------------
sigma_u | .0597245
sigma_e | .11106113
rho | .22431816 (fraction of variance due to u_i)
------------------------------------------------------------------------------
Code:
xtabond2 gdpcg lgdpc `exog' , gmmstyle(lgdpc , collapse ) ivstyle(`exog') orthogonal twostep robust sma
> ll nolevel nodiffsargan artest(6)
Favoring space over speed. To switch, type or click on mata: mata set matafavor speed, perm.
Warning: Two-step estimated covariance matrix of moments is singular.
Using a generalized inverse to calculate optimal weighting matrix for two-step estimation.
Dynamic panel-data estimation, two-step difference GMM
------------------------------------------------------------------------------
Group variable: panelid Number of obs = 170
Time variable : period Number of groups = 28
Number of instruments = 16 Obs per group: min = 1
F(7, 28) = 6.26 avg = 6.07
Prob > F = 0.000 max = 10
------------------------------------------------------------------------------
| Corrected
gdpcg | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
lgdpc | -.1052355 .0465194 -2.26 0.032 -.2005262 -.0099447
gg | .0006776 .0006514 1.04 0.307 -.0006568 .0020121
popgrowth | -6.033822 4.094238 -1.47 0.152 -14.42049 2.352844
liquid_gdp | -.0001061 .0006921 -0.15 0.879 -.0015238 .0013115
inflation | -.0000753 .000187 -0.40 0.690 -.0004584 .0003077
tradeop | .1676943 .1480974 1.13 0.267 -.1356696 .4710582
govs_peen | -.786176 .4152976 -1.89 0.069 -1.636875 .0645225
------------------------------------------------------------------------------
Instruments for orthogonal deviations equation
Standard
FOD.(gg popgrowth liquid_gdp inflation tradeop govs_peen)
GMM-type (missing=0, separate instruments for each period unless collapsed)
L(1/11).lgdpc collapsed
------------------------------------------------------------------------------
Arellano-Bond test for AR(1) in first differences: z = -3.49 Pr > z = 0.000
Arellano-Bond test for AR(2) in first differences: z = 2.75 Pr > z = 0.006
Arellano-Bond test for AR(3) in first differences: z = -2.57 Pr > z = 0.010
Arellano-Bond test for AR(4) in first differences: z = 2.01 Pr > z = 0.045
Arellano-Bond test for AR(5) in first differences: z = -1.26 Pr > z = 0.208
Arellano-Bond test for AR(6) in first differences: z = 1.34 Pr > z = 0.180
------------------------------------------------------------------------------
Sargan test of overid. restrictions: chi2(9) = 40.73 Prob > chi2 = 0.000
(Not robust, but not weakened by many instruments.)
Hansen test of overid. restrictions: chi2(9) = 15.93 Prob > chi2 = 0.068
(Robust, but weakened by many instruments.)
Code:
xtabond2 gdpcg lgdpc L.lgdpc `exog' , gmmstyle(lgdpc L.lgdpc , collapse lag(4 .) ) ivstyle(`exog') twost
> ep robust small ///
> orthogonal nolevel nodiffsargan artest(3)
Favoring space over speed. To switch, type or click on mata: mata set matafavor speed, perm.
Warning: Two-step estimated covariance matrix of moments is singular.
Using a generalized inverse to calculate optimal weighting matrix for two-step estimation.
Dynamic panel-data estimation, two-step difference GMM
------------------------------------------------------------------------------
Group variable: panelid Number of obs = 141
Time variable : period Number of groups = 27
Number of instruments = 18 Obs per group: min = 0
F(8, 27) = 69.73 avg = 5.22
Prob > F = 0.000 max = 9
------------------------------------------------------------------------------
| Corrected
gdpcg | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
lgdpc |
--. | -.7664274 .1356846 -5.65 0.000 -1.044829 -.4880256
L1. | .4334559 .0944323 4.59 0.000 .2396968 .627215
|
gg | .0016123 .0009237 1.75 0.092 -.0002831 .0035076
popgrowth | -3.16039 1.677249 -1.88 0.070 -6.60182 .2810399
liquid_gdp | .0002678 .0012761 0.21 0.835 -.0023505 .0028861
inflation | -.0050883 .0016433 -3.10 0.005 -.0084601 -.0017164
tradeop | .5624018 .165768 3.39 0.002 .2222739 .9025297
govs_peen | -1.425762 .3308342 -4.31 0.000 -2.104578 -.7469468
------------------------------------------------------------------------------
Instruments for orthogonal deviations equation
Standard
FOD.(gg popgrowth liquid_gdp inflation tradeop govs_peen)
GMM-type (missing=0, separate instruments for each period unless collapsed)
L(4/11).(lgdpc L.lgdpc) collapsed
------------------------------------------------------------------------------
Arellano-Bond test for AR(1) in first differences: z = -1.57 Pr > z = 0.117
Arellano-Bond test for AR(2) in first differences: z = -1.77 Pr > z = 0.077
Arellano-Bond test for AR(3) in first differences: z = -0.23 Pr > z = 0.818
------------------------------------------------------------------------------
Sargan test of overid. restrictions: chi2(10) = 9.40 Prob > chi2 = 0.495
(Not robust, but not weakened by many instruments.)
Hansen test of overid. restrictions: chi2(10) = 7.58 Prob > chi2 = 0.670
(Robust, but weakened by many instruments.)
Thank you for your time
Comment