I am estimating the impact of institutional integrity on carbon emissions using quantile regression for an unbalanced dataset that runs from 1984 to 2016 for about 123 countries. I'm interested in quantile regression to view the relationship at different points in the conditional distribution of y.
Because T spans over 30 years, therefore i have differenced my variables. Thus 'd' in front of every variable.
Definition of variables"
dlrgdp = Log of real per capita GDP, dlrgdp2 = squared dlrgdp, dfrleg = institutional integrity, dfrleg2 = squared dfrleg, dlpopden = log of population density, dlenguse = log of energy use, dfr_lrgdp = interaction term between institutional integrity and GDP
Code:
qregpd dlpccarb dlrgdp dlrgdp2 dfrleg dfrleg2 dlpopden dlenguse dfr_lrgdp dfr2_lrgdp2 , quantile(.25) id(cty) fix(year) optimize(mcmc) noisy draws(1000) burn(100) arate(.5) qregpd dlpccarb dlrgdp dlrgdp2 dfrleg dfrleg2 dlpopden dlenguse dfr_lrgdp dfr2_lrgdp2 , quantile(.75) id(cty) fix(year) optimize(mcmc) noisy draws(1000) burn(100) arate(.5)
Code:
Quantile Regression for Panel Data (QRPD) Number of obs: 3231 Number of groups: 120 Min obs per group: 3 Max obs per group: 30 ------------------------------------------------------------------------------ dlpccarb | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- dlrgdp | .5437956 .0590741 9.21 0.000 .4280124 .6595787 dlrgdp2 | -.0184208 .0035656 -5.17 0.000 -.0254094 -.0114323 dfrleg | -2.656076 .1897324 -14.00 0.000 -3.027944 -2.284207 dfrleg2 | 1.623832 .0635615 25.55 0.000 1.499253 1.74841 dlpopden | -.3040617 .0167634 -18.14 0.000 -.3369172 -.2712061 dlenguse | .8296468 .0032042 258.93 0.000 .8233667 .8359269 dfr_lrgdp | .2209018 .0216103 10.22 0.000 .1785463 .2632573 dfr2_lrgdp2 | -.0129949 .0008071 -16.10 0.000 -.0145767 -.011413 ------------------------------------------------------------------------------ No excluded instruments - standard QRPD estimation. MCMC diagonstics: Mean acceptance rate: 0.473 Total draws: 1000 Burn-in draws: 100 Draws retained: 900 Value of objective function: Mean: -74.3432 Min: -84.6212 Max: -70.4597 MCMC notes: *Point estimates correspond to mean of draws. *Standard errors are derived from variance of draws.
These results are too good to be true. Now I have 3 questions related to these results.
1). How can I check the validity of these results? Is there any post estimation command that verifies that these results are correct?
2). Would it be correct to use -qplot- in panel data ? if no, then what command would be the most appropriate to generate a graph similar to that generated by -qplot-?
3). Is there a command similar to grqreg to generate graphs over quantiles in a panel data setting?
0 Response to QREGPD post estimation and graphs
Post a Comment