1.- The endogenous variable varies between 0 and 1: f_n_p1p2p3_corrupcion
2.- There is a data panel: 17 cross sections and 10 time series
3.- There are 26 exogenous variables
4.- The Wald Insigma test concludes that we cannot accept the homoscedastic model
5.- In the fracreg probit I have included all the exogenous variables in the heteroscedasticity equation
6.- What model would you choose?
7.- Fracglm estimation:
. fracglm f_n_p1p2p3_corrupcion $xlist_nivel_base, link(p)
Fractional Probit Regression Number of obs = 2,007
Wald chi2(24) = .
Prob > chi2 = .
Log pseudolikelihood = -886.57796 Pseudo R2 = 0.1947
---------------------------------------------------------------------------------------------
| Robust
f_n_p1p2p3_corrupcion | Coef. Std. Err. z P>|z| [95% Conf. Interval]
----------------------------+----------------------------------------------------------------
paro | .0447631 .0030547 14.65 0.000 .0387759 .0507503
educ | -4.242257 .6703192 -6.33 0.000 -5.556058 -2.928456
internet | .0386991 .0013827 27.99 0.000 .0359892 .0414091
pibpc | .0257263 .0067201 3.83 0.000 .0125552 .0388975
ciclo | -.0030377 .0007722 -3.93 0.000 -.0045512 -.0015243
provincias | .0081609 .0072343 1.13 0.259 -.0060181 .0223398
habitantes | 1.44e-07 3.69e-08 3.89 0.000 7.13e-08 2.16e-07
denspob | 6.57e-07 .0001983 0.00 0.997 -.0003881 .0003894
pobdep | 12.64284 1.672184 7.56 0.000 9.365421 15.92026
pobinm | -2.607802 .5587866 -4.67 0.000 -3.703004 -1.512601
ubisigpol | .0460005 .0345048 1.33 0.182 -.0216276 .1136286
d_ideo_cuadra_votante_gob_3 | -.046812 .0105974 -4.42 0.000 -.0675825 -.0260415
mayabs_numero | .0910656 .0349888 2.60 0.009 .0224888 .1596424
indfp | -1.353811 .1811062 -7.48 0.000 -1.708773 -.9988495
aligob_numero | .0490593 .0235465 2.08 0.037 .002909 .0952097
partelec | -.0040318 .0019283 -2.09 0.037 -.0078112 -.0002524
legcons | .0351732 .0078346 4.49 0.000 .0198176 .0505288
gp | -6.82e-08 1.23e-08 -5.55 0.000 -9.24e-08 -4.41e-08
saldopre | -8.06e-08 1.09e-08 -7.39 0.000 -1.02e-07 -5.92e-08
deuda | 1.60e-08 2.24e-09 7.13 0.000 1.16e-08 2.04e-08
gi | -1.50e-07 3.07e-08 -4.88 0.000 -2.10e-07 -8.96e-08
gc | -4.40e-08 1.73e-08 -2.54 0.011 -7.79e-08 -1.01e-08
presfisc | -.0044704 .0297088 -0.15 0.880 -.0626986 .0537577
autfin | -.8489541 .1637078 -5.19 0.000 -1.169816 -.5280927
remanente | 5.41e-11 1.46e-11 3.71 0.000 2.55e-11 8.26e-11
_cons | -4.449184 .50938 -8.73 0.000 -5.44755 -3.450817
---------------------------------------------------------------------------------------------
8.- Fracreg probit estimate:
fracreg probit f_n_p1p2p3_corrupcion $xlist_nivel_base, het($xlist_nivel_base)
Iteration 2584: log pseudolikelihood = -874.49287
Fractional heteroskedastic probit regression Number of obs = 2,007
Wald chi2(19) = 4.51
Prob > chi2 = 0.9997
Log pseudolikelihood = -874.49287 Pseudo R2 = 0.0297
---------------------------------------------------------------------------------------------
| Robust
f_n_p1p2p3_corrupcion | Coef. Std. Err. z P>|z| [95% Conf. Interval]
----------------------------+----------------------------------------------------------------
f_n_p1p2p3_corrupcion |
paro | .0004344 .0003959 1.10 0.273 -.0003415 .0012104
educ | -.0034432 .01302 -0.26 0.791 -.0289619 .0220755
internet | .0002027 .0001974 1.03 0.304 -.0001841 .0005895
pibpc | .0002006 .000211 0.95 0.342 -.000213 .0006142
ciclo | -.0000231 .0000245 -0.94 0.346 -.0000712 .0000249
provincias | .0000824 .0001384 0.60 0.552 -.0001889 .0003537
habitantes | 1.01e-09 1.69e-09 0.60 0.548 -2.29e-09 4.32e-09
denspob | 4.79e-06 6.20e-06 0.77 0.440 -7.37e-06 .0000169
pobdep | .1106881 .1109048 1.00 0.318 -.1066814 .3280575
pobinm | -.0045794 .0083798 -0.55 0.585 -.0210035 .0118447
ubisigpol | -.0000995 .0002998 -0.33 0.740 -.0006871 .0004881
d_ideo_cuadra_votante_gob_3 | -.0001631 .0001693 -0.96 0.335 -.000495 .0001687
mayabs_numero | .0009665 .001033 0.94 0.349 -.0010582 .0029911
indfp | -.0057786 .0055872 -1.03 0.301 -.0167293 .0051722
aligob_numero | .0000267 .0002772 0.10 0.923 -.0005166 .00057
partelec | .0000385 .0000312 1.23 0.217 -.0000226 .0000996
legcons | .0001257 .0001334 0.94 0.346 -.0001357 .0003871
gp | 5.00e-11 2.55e-10 0.20 0.845 -4.50e-10 5.50e-10
saldopre | -4.54e-10 4.74e-10 -0.96 0.338 -1.38e-09 4.75e-10
deuda | -5.19e-11 6.34e-11 -0.82 0.413 -1.76e-10 7.23e-11
gi | -2.66e-09 2.79e-09 -0.96 0.339 -8.12e-09 2.80e-09
gc | -7.87e-10 1.17e-09 -0.67 0.500 -3.07e-09 1.50e-09
presfisc | 7.26e-06 .0004806 0.02 0.988 -.0009347 .0009492
autfin | -.005323 .0047033 -1.13 0.258 -.0145414 .0038954
remanente | 3.23e-13 3.87e-13 0.84 0.403 -4.35e-13 1.08e-12
_cons | -.0620295 .0547684 -1.13 0.257 -.1693736 .0453146
----------------------------+----------------------------------------------------------------
lnsigma |
paro | -.0505771 .004836 -10.46 0.000 -.0600555 -.0410987
educ | -5.317301 1.390939 -3.82 0.000 -8.043491 -2.59111
internet | .013611 .0031864 4.27 0.000 .0073657 .0198563
pibpc | -.064425 .0141549 -4.55 0.000 -.0921681 -.0366818
ciclo | .0003744 .0012394 0.30 0.763 -.0020548 .0028036
provincias | .0125249 .0167076 0.75 0.453 -.0202213 .0452711
habitantes | 5.95e-08 1.02e-07 0.58 0.561 -1.41e-07 2.60e-07
denspob | -.0003132 .0003525 -0.89 0.374 -.0010041 .0003777
pobdep | 1.743201 2.66053 0.66 0.512 -3.471342 6.957744
pobinm | 2.233358 .7799517 2.86 0.004 .7046803 3.762035
ubisigpol | .1406856 .1034532 1.36 0.174 -.062079 .3434502
d_ideo_cuadra_votante_gob_3 | .0150148 .01456 1.03 0.302 -.0135223 .0435519
mayabs_numero | -.0523526 .1215787 -0.43 0.667 -.2906424 .1859373
indfp | -.3644068 .4378658 -0.83 0.405 -1.222608 .4937944
aligob_numero | -.0043678 .0753021 -0.06 0.954 -.1519572 .1432216
partelec | -.0230958 .0066734 -3.46 0.001 -.0361754 -.0100161
legcons | -.0019291 .0153492 -0.13 0.900 -.0320129 .0281547
gp | -8.98e-08 3.66e-08 -2.46 0.014 -1.61e-07 -1.81e-08
saldopre | -8.07e-09 2.55e-08 -0.32 0.752 -5.81e-08 4.20e-08
deuda | 4.76e-08 5.54e-09 8.59 0.000 3.68e-08 5.85e-08
gi | 1.60e-07 8.77e-08 1.83 0.068 -1.18e-08 3.32e-07
gc | 6.43e-08 6.20e-08 1.04 0.300 -5.73e-08 1.86e-07
presfisc | .2029782 .0573729 3.54 0.000 .0905293 .3154271
autfin | 1.58897 .4407269 3.61 0.000 .725161 2.452779
remanente | 7.28e-11 2.45e-11 2.97 0.003 2.47e-11 1.21e-10
---------------------------------------------------------------------------------------------
. test [lnsigma]
( 1) [lnsigma]paro = 0
( 2) [lnsigma]educ = 0
( 3) [lnsigma]internet = 0
( 4) [lnsigma]pibpc = 0
( 5) [lnsigma]ciclo = 0
( 6) [lnsigma]provincias = 0
( 7) [lnsigma]habitantes = 0
( 8) [lnsigma]denspob = 0
( 9) [lnsigma]pobdep = 0
(10) [lnsigma]pobinm = 0
(11) [lnsigma]ubisigpol = 0
(12) [lnsigma]d_ideo_cuadra_votante_gob_3 = 0
(13) [lnsigma]mayabs_numero = 0
(14) [lnsigma]indfp = 0
(15) [lnsigma]aligob_numero = 0
(16) [lnsigma]partelec = 0
(17) [lnsigma]legcons = 0
(18) [lnsigma]gp = 0
(19) [lnsigma]saldopre = 0
(20) [lnsigma]deuda = 0
(21) [lnsigma]gi = 0
(22) [lnsigma]gc = 0
(23) [lnsigma]presfisc = 0
(24) [lnsigma]autfin = 0
(25) [lnsigma]remanente = 0
Constraint 25 dropped
chi2( 24) = 874.29
Prob > chi2 = 0.0000
0 Response to Why are there more significantly nonzero parameters in the fracglm estimate, than in the fracreg probit?
Post a Comment