Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
st: -oaxaca- problem
From
"Rubil Ivica" <[email protected]>
To
<[email protected]>
Subject
st: -oaxaca- problem
Date
Thu, 28 Nov 2013 14:30:47 +0100
Dear Statalisters,
I have a problem with Ben Jann's -oaxaca-.
I run:
oaxaca lwage $exp_var_2011 [aw=w], by(pri) weight(1) nodetail
And the output below suggests the mean of lwage for the group1 (pri==0) is 3.468472, while for the group2 (pri==1) the mean is 3.296296
Model for group 1
(sum of wgt is 9.1834e+06)
Source | SS df MS Number of obs = 1836
-------------+------------------------------ F( 32, 1803) = 80.76
Model | 129.198455 32 4.03745173 Prob > F = 0.0000
Residual | 90.1351807 1803 .049991781 R-squared = 0.5890
-------------+------------------------------ Adj R-squared = 0.5818
Total | 219.333636 1835 .119527867 Root MSE = .22359
------------------------------------------------------------------------------
lwage | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
age | -.0200541 .0071478 -2.81 0.005 -.034073 -.0060351
age2 | .0195698 .0077357 2.53 0.011 .004398 .0347417
female | -.1260799 .0122342 -10.31 0.000 -.1500746 -.1020852
married | .0374127 .0130586 2.86 0.004 .011801 .0630244
educ_low | -.056286 .0237125 -2.37 0.018 -.1027928 -.0097791
educ_high | .1128072 .0197363 5.72 0.000 .0740987 .1515156
educ_mrdr | .3645071 .0339375 10.74 0.000 .2979462 .431068
exper | .0150992 .0040778 3.70 0.000 .0071016 .0230968
exper2 | -.02621 .0084305 -3.11 0.002 -.0427447 -.0096754
tenure | .0043512 .002352 1.85 0.064 -.0002618 .0089642
tenure2 | -.0000641 .0000568 -1.13 0.259 -.0001755 .0000473
supervise | .1118821 .0155912 7.18 0.000 .0813034 .1424608
urban | .0380235 .0120195 3.16 0.002 .0144498 .0615971
occ1 | .3100307 .038864 7.98 0.000 .2338074 .386254
occ2 | .1805049 .0196895 9.17 0.000 .1418882 .2191215
occ4 | -.0814861 .0203193 -4.01 0.000 -.1213381 -.0416342
occ5 | -.1697702 .0226988 -7.48 0.000 -.214289 -.1252515
occ6 | -.1982014 .0561438 -3.53 0.000 -.308315 -.0880877
occ7 | -.1408072 .0258736 -5.44 0.000 -.1915525 -.0900618
occ8 | -.1192174 .0261573 -4.56 0.000 -.1705192 -.0679156
occ9 | -.323215 .0286932 -11.26 0.000 -.3794905 -.2669395
cont_temp | -.1330346 .0280786 -4.74 0.000 -.1881046 -.0779646
mediumfirm | .0141947 .0133518 1.06 0.288 -.0119919 .0403813
largefirm | .0651125 .0137416 4.74 0.000 .0381613 .0920638
act3_2011 | -.0176042 .0229491 -0.77 0.443 -.0626139 .0274055
act6_2011 | -.0070403 .0373465 -0.19 0.850 -.0802873 .0662067
act7_2011 | .0268201 .0499898 0.54 0.592 -.0712239 .124864
act8_2011 | .0378372 .0193573 1.95 0.051 -.0001278 .0758023
act11_2011 | .0838373 .0412563 2.03 0.042 .0029222 .1647524
northwest | -.0053293 .0188184 -0.28 0.777 -.0422374 .0315789
centraleast | -.0452847 .017326 -2.61 0.009 -.0792658 -.0113036
adriatic | -.033017 .0168158 -1.96 0.050 -.0659974 -.0000366
_cons | 3.666189 .1340373 27.35 0.000 3.403304 3.929073
------------------------------------------------------------------------------
Model for group 2
(sum of wgt is 1.0569e+07)
Source | SS df MS Number of obs = 2611
-------------+------------------------------ F( 32, 2578) = 104.00
Model | 401.188724 32 12.5371476 Prob > F = 0.0000
Residual | 310.77096 2578 .120547308 R-squared = 0.5635
-------------+------------------------------ Adj R-squared = 0.5581
Total | 711.959684 2610 .272781488 Root MSE = .3472
------------------------------------------------------------------------------
lwage | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
age | .0107747 .0074768 1.44 0.150 -.0038865 .0254359
age2 | -.0144506 .0091347 -1.58 0.114 -.0323627 .0034616
female | -.1835176 .0160742 -11.42 0.000 -.2150373 -.1519979
married | .0333863 .0174478 1.91 0.056 -.0008269 .0675994
educ_low | -.0968793 .0258114 -3.75 0.000 -.1474925 -.0462661
educ_high | .2377385 .0258013 9.21 0.000 .1871451 .2883319
educ_mrdr | .22368 .0745542 3.00 0.003 .0774877 .3698722
exper | -.0066592 .0041738 -1.60 0.111 -.0148436 .0015252
exper2 | .0313363 .0100652 3.11 0.002 .0115996 .051073
tenure | .0062553 .002891 2.16 0.031 .0005863 .0119242
tenure2 | -.0001353 .0000808 -1.68 0.094 -.0002937 .0000231
supervise | .1942685 .0211429 9.19 0.000 .1528097 .2357273
urban | .0106154 .015441 0.69 0.492 -.0196627 .0408935
occ1 | .3826859 .047162 8.11 0.000 .2902067 .4751651
occ2 | .0739356 .031726 2.33 0.020 .0117246 .1361465
occ4 | -.1942338 .0286501 -6.78 0.000 -.2504134 -.1380542
occ5 | -.3055466 .0269873 -11.32 0.000 -.3584657 -.2526276
occ6 | -.3949566 .1190838 -3.32 0.001 -.6284662 -.1614469
occ7 | -.200156 .0276926 -7.23 0.000 -.2544581 -.145854
occ8 | -.3148345 .0281448 -11.19 0.000 -.3700233 -.2596457
occ9 | -.3004927 .0349444 -8.60 0.000 -.3690147 -.2319707
cont_temp | .0053779 .0237008 0.23 0.821 -.0410967 .0518525
mediumfirm | -.0285624 .0184601 -1.55 0.122 -.0647605 .0076357
largefirm | .0670266 .0181104 3.70 0.000 .0315143 .102539
act3_2011 | -.0501554 .0207554 -2.42 0.016 -.0908543 -.0094565
act6_2011 | .0325907 .0276224 1.18 0.238 -.0215735 .086755
act7_2011 | -.049103 .0215923 -2.27 0.023 -.091443 -.006763
act8_2011 | .3988161 .0325846 12.24 0.000 .3349214 .4627107
act11_2011 | .1960192 .0360301 5.44 0.000 .1253683 .26667
northwest | -.1211181 .0244692 -4.95 0.000 -.1690994 -.0731368
centraleast | -.216864 .0238503 -9.09 0.000 -.2636316 -.1700963
adriatic | -.0748019 .0218918 -3.42 0.001 -.1177293 -.0318746
_cons | 3.261302 .1328525 24.55 0.000 3.000794 3.52181
------------------------------------------------------------------------------
Blinder-Oaxaca decomposition Number of obs = 4447
Model = linear
Group 1: pri = 0 N of obs 1 = 1836
Group 2: pri = 1 N of obs 2 = 2611
------------------------------------------------------------------------------
lwage | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
overall |
group_1 | 3.468472 .008098 428.31 0.000 3.4526 3.484344
group_2 | 3.296296 .0102489 321.62 0.000 3.276209 3.316384
difference | .1721757 .0130621 13.18 0.000 .1465746 .1977769
explained | .1374541 .0146018 9.41 0.000 .1088351 .166073
unexplained | .0347217 .01552 2.24 0.025 .004303 .0651404
------------------------------------------------------------------------------
Here's the puzzle. The means for groups 1 and 2 from the above output table must be equal to the means obtained by running:
bysort pri: sum lwage
Yet, they unfortunately do not correspond, as the following output suggests:
-> pri = 0
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
lwage | 1836 3.352651 .3396782 1.809842 4.643055
-> pri = 1
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
lwage | 2611 3.088002 .4094408 1.992164 5.142046
As you can see, the same observations are used for estimating the wage regressions within -oaxaca- and for the computation of the means using
bysort pri: sum lwage
Is there anybody who knows where the problem comes from?
Thanks.
ivica
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/