I'm having a puzzle trying to decompose the change in log-wage over percentiles between 2007 and 2017. Specifically, I'm focusing on P10, P50 and P90 and I have that the unexplained component always dominates and it is significant along all percentiles. However, it is mostly the case that none of the coefficient in the detailed unexplained is significant. As an example:
Code:
oaxaca_rif log_gross occupation sector isced emp_status pl200 region contract_type [pw=rb050] if gender==1, by(overall) rif(q(10)) relax wgt(0) rwlogit(occupation sector isced emp_status pl200 region contract_type) Model : Blinder-Oaxaca RIF-decomposition Type : Reweighted RIF : q(10) Scale : 1 Group 1: overall = 0 N of obs 1 = 7462 Group c: X2~>rw~>X1 or x1*b2 N of obs C = 7786 Group 2: overall = 1 N of obs 2 = 7786 ---------------------------------------------------------------------------------- log_gross | Coef. Std. Err. z P>|z| [95% Conf. Interval] -----------------+---------------------------------------------------------------- Overall | Group_1 | 9.281226 .0297112 312.38 0.000 9.222993 9.339459 Group_c | 9.479692 .0231539 409.42 0.000 9.434311 9.525073 Group_2 | 9.531197 .0181483 525.18 0.000 9.495627 9.566767 Tdifference | -.249971 .0348155 -7.18 0.000 -.318208 -.1817339 ToT_Explained | -.0515046 .0117615 -4.38 0.000 -.0745567 -.0284526 ToT_Unexplained | -.1984663 .0370411 -5.36 0.000 -.2710656 -.1258671 -----------------+---------------------------------------------------------------- Explained | Total | -.0515046 .0117615 -4.38 0.000 -.0745567 -.0284526 Pure_explained | -.0522429 .0086994 -6.01 0.000 -.0692934 -.0351923 Specif_err | .0007382 .0072891 0.10 0.919 -.0135481 .0150245 -----------------+---------------------------------------------------------------- Pure_explained | occupation | .004348 .0015658 2.78 0.005 .0012791 .007417 sector | -.0003758 .0006203 -0.61 0.545 -.0015916 .00084 isced | .0186843 .0033935 5.51 0.000 .0120331 .0253355 emp_status | -.0384481 .0047426 -8.11 0.000 -.0477433 -.0291528 pl200 | .0206592 .0023426 8.82 0.000 .0160678 .0252506 region | -4.69e-07 .000039 -0.01 0.990 -.0000769 .000076 contract_type | -.05711 .005093 -11.21 0.000 -.0670921 -.047128 -----------------+---------------------------------------------------------------- Specif_err | occupation | .0026133 .0292909 0.09 0.929 -.0547958 .0600224 sector | .0220972 .0177351 1.25 0.213 -.012663 .0568574 isced | .0202463 .0352649 0.57 0.566 -.0488716 .0893642 emp_status | -.104401 .0804261 -1.30 0.194 -.2620333 .0532314 pl200 | -.0564017 .020493 -2.75 0.006 -.0965672 -.0162362 region | -.0075232 .0208354 -0.36 0.718 -.0483597 .0333134 contract_type | .0152354 .0353048 0.43 0.666 -.0539608 .0844317 _cons | .1088717 .1135587 0.96 0.338 -.1136992 .3314426 -----------------+---------------------------------------------------------------- Unexplained | Total | -.1984663 .0370411 -5.36 0.000 -.2710656 -.1258671 Reweight_err | .0078208 .0162171 0.48 0.630 -.0239641 .0396057 Pure_Unexplained | -.2062871 .0335048 -6.16 0.000 -.2719553 -.1406189 -----------------+---------------------------------------------------------------- Pure_Unexplained | occupation | -.1129848 .102314 -1.10 0.269 -.3135167 .087547 sector | -.0740689 .0565395 -1.31 0.190 -.1848843 .0367465 isced | -.0262418 .121197 -0.22 0.829 -.2637835 .2113 emp_status | .2850504 .267775 1.06 0.287 -.239779 .8098799 pl200 | .0325169 .0640682 0.51 0.612 -.0930545 .1580884 region | .0836086 .0650197 1.29 0.198 -.0438277 .2110449 contract_type | -.1249659 .1243426 -1.01 0.315 -.3686729 .1187411 _cons | -.2692016 .3899048 -0.69 0.490 -1.033401 .4949978 -----------------+---------------------------------------------------------------- Reweight_err | occupation | -.0003377 .0011815 -0.29 0.775 -.0026534 .0019781 sector | -.0000911 .0005099 -0.18 0.858 -.0010904 .0009083 isced | -.002396 .0026608 -0.90 0.368 -.0076111 .0028191 emp_status | .0062846 .0094418 0.67 0.506 -.0122211 .0247903 pl200 | .0009365 .0029674 0.32 0.752 -.0048795 .0067525 region | -9.14e-06 .000083 -0.11 0.912 -.0001718 .0001535 contract_type | .0034335 .0092499 0.37 0.710 -.014696 .0215631 ----------------------------------------------------------------------------------
Furthermore, once doing on the Gini coefficient, the characteristics effects are dominant, what's the reason for the reverse in magnitude of coefficient vs characteristics effect?
thanks for the support
0 Response to Oaxaca-RIF decomposition
Post a Comment