连老师你好:
我用ivprobit做了回归之后,用mfx进行边际分析,但是两个的估计结果一样,不知道什么原因,见下文的回归结果
. xi:ivprobit health_hh age age2 gender edu marr sons child_h age_h edu_h /*
> */ family_h house asset_h earth_h /*
> */ distance i.pro_h (home_h= vhome) if age>=60,cluster(hhid)
i.pro_h _Ipro_h_0-4 (naturally coded; _Ipro_h_0 omitted)
Fitting exogenous probit model
Iteration 0: log likelihood = -686.16163
Iteration 1: log likelihood = -617.04511
Iteration 2: log likelihood = -616.41998
Iteration 3: log likelihood = -616.41751
Iteration 4: log likelihood = -616.41751
Fitting full model
Iteration 0: log pseudolikelihood = -229.37133
Iteration 1: log pseudolikelihood = -228.18686
Iteration 2: log pseudolikelihood = -227.08996
Iteration 3: log pseudolikelihood = -227.07296
Iteration 4: log pseudolikelihood = -227.0729
Iteration 5: log pseudolikelihood = -227.0729
Probit model with endogenous regressors Number of obs = 1009
Wald chi2(19) = 140.71
Log pseudolikelihood = -227.0729 Prob > chi2 = 0.0000
(Std. Err. adjusted for 722 clusters in hhid)
------------------------------------------------------------------------------
| Robust
| Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
home_h | -2.144181 .9717439 -2.21 0.027 -4.048764 -.2395976
age | -.3553932 .1010821 -3.52 0.000 -.5535104 -.1572759
age2 | .0023075 .0006922 3.33 0.001 .0009509 .0036641
gender | .2611572 .0838192 3.12 0.002 .0968746 .4254399
edu | .0371354 .014208 2.61 0.009 .0092882 .0649826
marr | -.3059207 .1261024 -2.43 0.015 -.5530769 -.0587646
sons | .0050542 .034729 0.15 0.884 -.0630133 .0731218
child_h | -.2005358 .126321 -1.59 0.112 -.4481203 .0470488
age_h | -.013136 .009285 -1.41 0.157 -.0313342 .0050623
edu_h | .0058887 .0200565 0.29 0.769 -.0334213 .0451987
family_h | .1470308 .072189 2.04 0.042 .005543 .2885187
house | .004012 .0024494 1.64 0.101 -.0007888 .0088128
asset_h | .002786 .0145158 0.19 0.848 -.0256644 .0312365
earth_h | .0078532 .0119595 0.66 0.511 -.0155871 .0312934
distance | -.0113015 .0081062 -1.39 0.163 -.0271893 .0045863
_Ipro_h_1 | .5544208 .1654738 3.35 0.001 .2300981 .8787435
_Ipro_h_2 | .3816545 .1800984 2.12 0.034 .0286682 .7346409
_Ipro_h_3 | -.0029485 .154539 -0.02 0.985 -.3058394 .2999424
_Ipro_h_4 | .140648 .1683701 0.84 0.404 -.1893513 .4706473
_cons | 13.23127 3.782055 3.50 0.000 5.818578 20.64396
-------------+----------------------------------------------------------------
/athrho | .3807797 .1908364 2.00 0.046 .0067472 .7548122
/lnsigma | -1.802532 .0237501 -75.90 0.000 -1.849082 -1.755983
-------------+----------------------------------------------------------------
rho | .3633844 .1656368 .0067471 .6380111
sigma | .1648808 .0039159 .1573816 .1727373
------------------------------------------------------------------------------
Instrumented: home_h
Instruments: age age2 gender edu marr sons child_h age_h edu_h family_h
house asset_h earth_h distance _Ipro_h_1 _Ipro_h_2 _Ipro_h_3
_Ipro_h_4 vhome
------------------------------------------------------------------------------
Wald test of exogeneity (/athrho = 0): chi2(1) = 3.98 Prob > chi2 = 0.0460
.
.
. mfx
Marginal effects after ivprobit
y = Fitted values (predict)
= -.21120026
------------------------------------------------------------------------------
variable | dy/dx Std. Err. z P>|z| [ 95% C.I. ] X
---------+--------------------------------------------------------------------
home_h | -2.144181 .97176 -2.21 0.027 -4.0488 -.239561 .233392
age | -.3553932 .10108 -3.52 0.000 -.553512 -.157274 67.8484
age2 | .0023075 .00069 3.33 0.001 .000951 .003664 4657.37
gender*| .2611572 .08382 3.12 0.002 .096875 .42544 .539148
edu | .0371354 .01421 2.61 0.009 .009288 .064982 3.82458
marr*| -.3059207 .1261 -2.43 0.015 -.553077 -.058765 .714569
sons | .0050542 .03473 0.15 0.884 -.063014 .073123 3.1447
child_h | -.2005358 .12632 -1.59 0.112 -.44812 .047049 .979187
age_h | -.013136 .00928 -1.41 0.157 -.031334 .005062 37.3676
edu_h | .0058887 .02006 0.29 0.769 -.03342 .045197 7.89717
family_h | .1470308 .07219 2.04 0.042 .005545 .288516 5.2111
house | .004012 .00245 1.64 0.101 -.000789 .008813 16.2653
asset_h | .002786 .01452 0.19 0.848 -.025663 .031235 1.87709
earth_h | .0078532 .01196 0.66 0.511 -.015587 .031293 2.26518
distance | -.0113015 .00811 -1.39 0.163 -.027189 .004586 6.69348
_Ipro_~1*| .5544208 .16547 3.35 0.001 .230098 .878743 .219029
_Ipro_~2*| .3816545 .1801 2.12 0.034 .028668 .734641 .161546
_Ipro_~3*| -.0029485 .15454 -0.02 0.985 -.305839 .299942 .285431
_Ipro_~4*| .140648 .16837 0.84 0.404 -.189351 .470647 .166501
------------------------------------------------------------------------------
(*) dy/dx is for discrete change of dummy variable from 0 to 1