Hello everyone,

I am estimating a model where my dependent variable is economic growth and all my independent variables are initial values of explanatory variables. As my data is split into 5 year periods the initial values mean the following: for period1971-1975 the initial value of 1971, for period 1976-1980 the initial value of 1976 and so on. As my T is very large I have to restrict the number of lags used as instruments in order to reduce the instrument proliferation problem. I did this by using the command lag (. 1) in xtabond 2 as shown below. However, can somebody tell me if restricting the lags in this way is correct? My idea was to use only the most recent lags as instruments as they are correlated with endogenous regressors. I believe restricting the lags by using lag (2 2) would mean that I am going far back in time and the instruments used are not that correlated with endogenous regressors.

Any help would be very highly appreciated!



xtabond2 gr lgrowth prcreditBI prcreditBI2 log_school linfl log_govsize log_trade td* if period<9, gmm(L( lgrowth prcreditBI2 prcreditBI log_school log_govsize log_trade linfl ), lag(. 1)) iv(td*) two robust
Favoring space over speed. To switch, type or click on mata: mata set matafavor speed, perm.
td7 dropped due to collinearity
td9 dropped due to collinearity
td10 dropped due to collinearity
td11 dropped due to collinearity
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: Country_ Number of obs = 550
Time variable : period Number of groups = 111
Number of instruments = 92 Obs per group: min = 1
Wald chi2(14) = 250.80 avg = 4.95
Prob > chi2 = 0.000 max = 8
------------------------------------------------------------------------------
| Corrected
gr | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
lgrowth | -.8822967 .5095106 -1.73 0.083 -1.880919 .1163258
prcreditBI | 5.378966 2.381949 2.26 0.024 .7104322 10.0475
prcreditBI2 | -2.647564 1.282077 -2.07 0.039 -5.160388 -.1347397
log_school | 1.784351 .6929808 2.57 0.010 .4261338 3.142569
linfl | -.5492066 .2257515 -2.43 0.015 -.9916714 -.1067419
log_govsize | -2.582938 .8458974 -3.05 0.002 -4.240867 -.9250099
log_trade | 1.083419 .5700486 1.90 0.057 -.0338554 2.200694
td1 | 2.051786 .7271167 2.82 0.005 .6266636 3.476909
td2 | 2.951726 .686082 4.30 0.000 1.60703 4.296422
td3 | 2.113944 .6746269 3.13 0.002 .7916992 3.436188
td4 | 1.920597 .56364 3.41 0.001 .8158825 3.025311
td5 | -.1515009 .4335273 -0.35 0.727 -1.001199 .6981969
td6 | .5356309 .284069 1.89 0.059 -.0211341 1.092396
td8 | .1037828 .3646904 0.28 0.776 -.6109971 .8185628
_cons | 8.777158 4.18574 2.10 0.036 .5732579 16.98106
------------------------------------------------------------------------------
Instruments for first differences equation
Standard
D.(td1 td2 td3 td4 td5 td6 td7 td8 td9 td10 td11)
GMM-type (missing=0, separate instruments for each period unless collapsed)
L.(L.lgrowth L.prcreditBI2 L.prcreditBI L.log_school L.log_govsize
L.log_trade L.linfl)
Instruments for levels equation
Standard
td1 td2 td3 td4 td5 td6 td7 td8 td9 td10 td11
_cons
GMM-type (missing=0, separate instruments for each period unless collapsed)
D.(L.lgrowth L.prcreditBI2 L.prcreditBI L.log_school L.log_govsize
L.log_trade L.linfl)
------------------------------------------------------------------------------
Arellano-Bond test for AR(1) in first differences: z = -3.35 Pr > z = 0.001
Arellano-Bond test for AR(2) in first differences: z = -1.79 Pr > z = 0.074
------------------------------------------------------------------------------
Sargan test of overid. restrictions: chi2(77) = 103.88 Prob > chi2 = 0.022
(Not robust, but not weakened by many instruments.)
Hansen test of overid. restrictions: chi2(77) = 83.25 Prob > chi2 = 0.293
(Robust, but weakened by many instruments.)