Dear all,
I estimated the following dynamic panel model on economic growth obtaining the following results:
xtabond gdppc_gr if eu27==1 & year>=2000, pre(educ socialexp, lag(0,1)) pre(ln_gdp_pc, lag(0,1)) lags(1)
Arellano-Bond dynamic panel-data estimation Number of obs = 151
Group variable (i): countryn Number of groups = 26
Wald chi2(4) = 100.52
Time variable (t): year Obs per group: min = 5
avg = 5.807692
max = 6
One-step results
------------------------------------------------------------------------------
D.gdppc_gr | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
gdppc_gr |
LD. | -.034262 .0731906 -0.47 0.640 -.1777129 .1091889
educ |
D1. | .1479119 .1160591 1.27 0.203 -.0795596 .3753835
socialexp |
D1. | -1.032387 .1780187 -5.80 0.000 -1.381297 -.6834765
ln_gdp_pc |
D1. | 6.809475 2.934207 2.32 0.020 1.058535 12.56042
_cons | -.2916244 .1720441 -1.70 0.090 -.6288246 .0455758
------------------------------------------------------------------------------
Sargan test of over-identifying restrictions:
chi2(77) = 147.76 Prob > chi2 = 0.0000
Arellano-Bond test that average autocovariance in residuals of order 1 is 0:
H0: no autocorrelation z = -3.41 Pr > z = 0.0007
Arellano-Bond test that average autocovariance in residuals of order 2 is 0:
H0: no autocorrelation z = -0.68 Pr > z = 0.4989
As you see the lagged dependent variable is not significant, what does it mean? the lagged value used as an instrument is not good?
Moreover the Sargan test gives no overidentification problems, but the AB test at the 2nd order is not significant. How can I interpret them?
thanks
Cristina
*
* 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/
