I am using the xtabond2 to draw advantage of the twostep System GMM approach in order to estimate my dynamic panel data model. I am considering two groups of countries: The European Union (28 countries) and the Central and Eastern Europe (10 countries) both groups from 1990 - 2014 with an unbalanced dataset. Hence, N1 = 700, N2 = 250 and T = 25.
The point is to estimate the effect of investments in renewable energy for the two groups of countries on (ln) of gross domestic product pr. capita. Since I assume that the model is dynamic I also use the first three lags of ln GDP pr. capita as independent variables. My linear regression model takes the form:
LnGDPc,t = β1RECc,t + β2LnGDPc,t-1 + β3LnGDPc,t-2 + β3LnGDPc,t-3 + αc + µt + ɛc,t,
RECc,t is gross inland renewable energy consumption and my coefficient of interest.
To estimate this model I run the following code:
Code:
xtabond2 L(0/3).LnGDP REC tdum4-tdum25, /// gmm(L.LnGDP, lag(1 4) collapse equation(diff)) /// gmm(REC, lag (1 2) collapse equation(diff)) /// ivstyle(tdum4-tdum25, equation(level)) twostep robust
I hereby instrument by two lags of REC as first-difference and four lags of the first lagged dependent variable as first-difference. This works well when I estimate the model for the European Union countries (N=28). However, when I run the same regression for the Central and Eastern European countries (N=10) I obtain the following output:
Code:
Dynamic panel-data estimation, two-step system GMM ------------------------------------------------------------------------------ Group variable: country1 Number of obs = 220 Time variable : year Number of groups = 10 Number of instruments = 28 Obs per group: min = 22 Wald chi2(25) = 2.01e+07 avg = 22.00 Prob > chi2 = 0.000 max = 22 ------------------------------------------------------------------------------ | Corrected LnGDP | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- LnGDP | L1. | 1.06558 .0589094 18.09 0.000 .9501197 1.18104 L2. | 0 (omitted) L3. | -.0614096 .059251 -1.04 0.300 -.1775394 .0547201 | REC | -2.52e-06 1.69e-06 -1.49 0.135 -5.82e-06 7.87e-07 tdum4 | -.044087 .010539 -4.18 0.000 -.0647431 -.0234308 tdum6 | .0105556 .0108727 0.97 0.332 -.0107544 .0318657 tdum7 | -.008127 .0136221 -0.60 0.551 -.0348258 .0185717 tdum8 | 0 (omitted) tdum9 | 0 (omitted) tdum10 | -.028836 .0148624 -1.94 0.052 -.0579658 .0002938 tdum11 | 0 (omitted) tdum12 | 0 (omitted) tdum13 | 0 (omitted) tdum14 | 0 (omitted) tdum15 | 0 (omitted) tdum16 | 0 (omitted) tdum17 | 0 (omitted) tdum18 | .0298336 .0065639 4.55 0.000 .0169685 .0426987 tdum19 | -.0188079 .0147789 -1.27 0.203 -.0477741 .0101583 tdum20 | 0 (omitted) tdum21 | 0 (omitted) tdum22 | -.057359 .0440904 -1.30 0.193 -.1437746 .0290566 tdum23 | 0 (omitted) tdum24 | 0 (omitted) tdum25 | 0 (omitted) _cons | 0 (omitted) ------------------------------------------------------------------------------ Instruments for first differences equation GMM-type (missing=0, separate instruments for each period unless collapsed) L(1/2).REC collapsed L(1/4).L.LnGDP collapsed Instruments for levels equation Standard tdum4 tdum5 tdum6 tdum7 tdum8 tdum9 tdum10 tdum11 tdum12 tdum13 tdum14 tdum15 tdum16 tdum17 tdum18 tdum19 tdum20 tdum21 tdum22 tdum23 tdum24 tdum25 _cons
My question: How do I come around this problem in Stata? I have read threads like https://www.statalist.org/forums/for...m-time-dummies[\URL] but I cannot find the answer here.
Any help is appreciated. Thank you in advance.
Best regards,
Emil F. Mahler.
0 Response to Problem With Xtabond2 In Dynamic Panel Data Concerning Omitted Lagged Coefficients Reported in Output
Post a Comment