Question on RIF Decomposition Using Fernando Rios's RIF Package in Stata

[Sandeep Sharma]

Hi everyone,

I'm analyzing the formal/informal pay gap using the RIF decomposition from Fernando Rios's RIF package in Stata. Here’s the command I’m using:

oaxaca_rif ln_wage experience experience_sq education_Below_primary education_Primary education_Tenth_grade education_Secondary education_Bachelor education_Masters_above female hh_size caste_group_6_Janajati caste_group_6_Adhibasi caste_group_6_Madhesi caste_group_6_Dalit caste_group_6_Others married voc_train migrated_fr_job tot_chores_hrs urban overtime_40 if year == 2018, by(formal_employment) rif(q(10)) wgt(1) rwprobit(experience experience_sq education_Below_primary education_Primary education_Tenth_grade education_Secondary education_Bachelor education_Masters_above) swap

I'm facing a few issues and would appreciate your help:

1. Global Options: Is it possible to use global macros for variables like education and caste_group_6? When I tried, the command seemed to ignore the global values.
2. Multiple Quantiles: Can I calculate RIF for multiple quantiles at once, e.g., rif(q(10(10)90)) instead of just rif(q(10))?
3. Selection Bias Adjustment: I have a variable HH_formal (number of formal employees in the household except self) that I want to use as an instrument variable. Can I adjust for selection bias using this variable?

Thanks in advance for your help!

[FernandoRios]

Hi Sanda
some answers:
1) Not sure what you mean with global macros. I don't see any on your code. If you meant to say...use a macro like $variables instead of the long list, yes its possible. Please provide more details to understand your problem otherwise.
2) No, the command only works with one RIF at a time.
3) The theory behind RIF does not really support correcting sample selection (heckman style), And the command does not support instruments either. If you want to do something like that, you may want to create the RIFs first, then use Oaxaca option "model"
Best wishes
F

[Sandeep Sharma]

Respected Sir,

Thank you for the prompt response.
1. In the earlier one i tried with global option using the code

global education "education_Below_primary education_Primary education_Tenth_grade education_Secondary education_Bachelor education_Masters_above"
global caste "caste_group_6_Janajati caste_group_6_Adhibasi caste_group_6_Madhesi caste_group_6_Dalit caste_group_6_Others"
global exp "experience experience_sq"

oaxaca_rif $experience $education hh_size female $caste married urban voc_train migrated_fr_job tot_chores_hrs overtime_40 if year == 2018, ///
by(formal_employment) rif(q(10)) wgt(1) rwprobit($experience $education) swap

but the results in the figure didnt show anyof the global options variable ?

2. Could you be kind enough to help me understand creating RIFs first and usin Oaxaca option "model" here

[Sandeep Sharma]

Respected FernandoRios,

I was able to solve the aforementioned problem using macro variables.

Using the code
bootstrap, reps(100): oaxaca_rif ln_wage $exp $education $caste hh_size female ///
married child_12 voc_train migrated_fr_job tot_chores_hrs urban overtime_40 ///
$class_5, by(formal_employment) wgt(1) rif(q(90)) rwlogit(HH_formal dep_ratio $exp $education ) swap

I was able to get the following results, but my unexplained components seems to be because of rwg_error. How do i explain these results?

Model : Blinder-Oaxaca RIF-decomposition
Type : Reweighted
RIF : q(90)
Scale : 1
Group 1: formal_employment = 1 x1*b1 N of obs 1 = 3934
Group c: X1~>rw~>X2 or x2*b1 N of obs C = 3934
Group 2: formal_employment = 0 x2*b2 N of obs 2 = 9611

-----------------------------------------------------------------------------------------
| Observed Bootstrap Normal-based
ln_wage | Coef. Std. Err. z P>|z| [95% Conf. Interval]
------------------------+----------------------------------------------------------------
Overall |
group_1 | 5.067668 .00948 534.57 0.000 5.049088 5.086248
group_c | 4.804999 .0323876 148.36 0.000 4.741521 4.868478
group_2 | 4.746465 .008618 550.76 0.000 4.729574 4.763356
tdifference | .3212031 .0123806 25.94 0.000 .2969376 .3454687
t_explained | .2626685 .0323214 8.13 0.000 .1993197 .3260173
t_unexplained | .0585346 .0331028 1.77 0.077 -.0063456 .1234148
------------------------+----------------------------------------------------------------
explained |
total | .2626685 .0323214 8.13 0.000 .1993197 .3260173
p_explained | .1904816 .0116029 16.42 0.000 .1677404 .2132228
specif_err | .0721869 .0282784 2.55 0.011 .0167623 .1276115
------------------------+----------------------------------------------------------------
p_explained |
experience | -.0047758 .0050854 -0.94 0.348 -.0147429 .0051913
experience_sq | .011546 .0050143 2.30 0.021 .0017181 .0213738
education_Below_primary | .0035275 .003347 1.05 0.292 -.0030326 .0100876
education_Primary | -.0121284 .0067355 -1.80 0.072 -.0253297 .0010729
education_Tenth_grade | .0089924 .0036236 2.48 0.013 .0018901 .0160946
education_Secondary | .034596 .0073364 4.72 0.000 .0202168 .0489752
education_Bachelor | .0649243 .0073943 8.78 0.000 .0504316 .0794169
education_Masters_above | .0682549 .0080218 8.51 0.000 .0525325 .0839774
caste_group_6_Janajati | -.0002648 .0009929 -0.27 0.790 -.0022109 .0016814
caste_group_6_Adhibasi | -.0018221 .0020026 -0.91 0.363 -.0057472 .002103
caste_group_6_Madhesi | -.000053 .000419 -0.13 0.899 -.0008742 .0007682
caste_group_6_Dalit | .0012466 .0021478 0.58 0.562 -.0029629 .0054562
caste_group_6_Others | -.0005618 .0015989 -0.35 0.725 -.0036956 .002572
hh_size | .0013207 .0014373 0.92 0.358 -.0014963 .0041377
female | -.0008591 .0007308 -1.18 0.240 -.0022914 .0005732
married | -.0008637 .0035311 -0.24 0.807 -.0077847 .0060572
child_12 | -.0031829 .0025443 -1.25 0.211 -.0081696 .0018038
voc_train | .0049811 .0021725 2.29 0.022 .000723 .0092392
migrated_fr_job | -.0008788 .0009504 -0.92 0.355 -.0027415 .0009839
tot_chores_hrs | -.0005168 .000892 -0.58 0.562 -.0022651 .0012315
urban | .0007571 .0014649 0.52 0.605 -.0021141 .0036282
overtime_40 | .0162423 .0036001 4.51 0.000 .0091862 .0232984
------------------------+----------------------------------------------------------------
specif_err |
experience | -.2525198 .1921172 -1.31 0.189 -.6290626 .1240229
experience_sq | .1086923 .12228 0.89 0.374 -.130972 .3483566
education_Below_primary | .0206193 .0144623 1.43 0.154 -.0077263 .0489648
education_Primary | -.0220972 .0316848 -0.70 0.486 -.0841983 .0400038
education_Tenth_grade | -.0287745 .0199767 -1.44 0.150 -.0679282 .0103792
education_Secondary | -.044881 .0160603 -2.79 0.005 -.0763587 -.0134034
education_Bachelor | -.0431565 .0116825 -3.69 0.000 -.0660538 -.0202592
education_Masters_above | -.0205794 .005423 -3.79 0.000 -.0312082 -.0099506
caste_group_6_Janajati | -.0080032 .02525 -0.32 0.751 -.0574923 .0414858
caste_group_6_Adhibasi | .0017737 .0106761 0.17 0.868 -.019151 .0226984
caste_group_6_Madhesi | .0018976 .0102879 0.18 0.854 -.0182663 .0220614
caste_group_6_Dalit | .0104654 .0075088 1.39 0.163 -.0042516 .0251824
caste_group_6_Others | .0060804 .0038681 1.57 0.116 -.001501 .0136618
hh_size | -.0702909 .0840586 -0.84 0.403 -.2350428 .0944609
female | .0065985 .0176907 0.37 0.709 -.0280745 .0412716
married | .0145613 .0478601 0.30 0.761 -.0792427 .1083653
child_12 | .0649284 .0329254 1.97 0.049 .0003959 .129461
voc_train | -.0075139 .0141116 -0.53 0.594 -.0351722 .0201444
migrated_fr_job | .024735 .0187609 1.32 0.187 -.0120356 .0615057
tot_chores_hrs | .0345042 .0153636 2.25 0.025 .0043921 .0646164
urban | .0220398 .0447773 0.49 0.623 -.0657221 .1098017
overtime_40 | .1000314 .0427643 2.34 0.019 .016215 .1838479
_cons | .1530761 .2033535 0.75 0.452 -.2454895 .5516417
------------------------+----------------------------------------------------------------
unexplained |
total | .0585346 .0331028 1.77 0.077 -.0063456 .1234148
rwg_error | .0591314 .0176128 3.36 0.001 .024611 .0936517
p_unexplained | -.0005967 .0278452 -0.02 0.983 -.0551722 .0539788
------------------------+----------------------------------------------------------------
p_unexplained |
experience | .1843201 .2026105 0.91 0.363 -.2127893 .5814295
experience_sq | -.0705743 .1274907 -0.55 0.580 -.3204514 .1793028
education_Below_primary | -.0494321 .0180199 -2.74 0.006 -.0847504 -.0141138
education_Primary | -.0113247 .0350248 -0.32 0.746 -.0799721 .0573227
education_Tenth_grade | .0082796 .0213674 0.39 0.698 -.0335998 .050159
education_Secondary | .0270716 .0166596 1.62 0.104 -.0055805 .0597238
education_Bachelor | .0223758 .0124007 1.80 0.071 -.0019291 .0466807
education_Masters_above | .0118625 .0055761 2.13 0.033 .0009337 .0227914
caste_group_6_Janajati | .0067591 .0274324 0.25 0.805 -.0470075 .0605256
caste_group_6_Adhibasi | .0207149 .0141563 1.46 0.143 -.007031 .0484608
caste_group_6_Madhesi | .0155193 .011343 1.37 0.171 -.0067126 .0377513
caste_group_6_Dalit | -.0112718 .0114743 -0.98 0.326 -.0337611 .0112174
caste_group_6_Others | .0021749 .0040045 0.54 0.587 -.0056738 .0100237
hh_size | -.000825 .099804 -0.01 0.993 -.1964372 .1947873
female | .0247833 .0161223 1.54 0.124 -.0068158 .0563825
married | -.0720443 .0497996 -1.45 0.148 -.1696497 .0255611
child_12 | -.0379877 .0379743 -1.00 0.317 -.1124158 .0364405
voc_train | .0157376 .0116607 1.35 0.177 -.0071171 .0385922
migrated_fr_job | -.0127601 .0128728 -0.99 0.322 -.0379903 .0124702
tot_chores_hrs | -.0172621 .0150423 -1.15 0.251 -.0467445 .0122202
urban | .0450509 .0447755 1.01 0.314 -.0427074 .1328092
overtime_40 | -.1160712 .0553688 -2.10 0.036 -.224592 -.0075504
_cons | .0143069 .2128732 0.07 0.946 -.4029168 .4315306
------------------------+----------------------------------------------------------------
rwg_error |
experience | -.0124304 .0089372 -1.39 0.164 -.0299471 .0050863
experience_sq | .0076593 .0080572 0.95 0.342 -.0081326 .0234512
education_Below_primary | .0015351 .0013197 1.16 0.245 -.0010514 .0041217
education_Primary | .0019948 .002249 0.89 0.375 -.0024132 .0064027
education_Tenth_grade | .0018398 .0017182 1.07 0.284 -.0015278 .0052073
education_Secondary | .0012097 .0018796 0.64 0.520 -.0024743 .0048936
education_Bachelor | .0044832 .0022615 1.98 0.047 .0000507 .0089156
education_Masters_above | .0014743 .0012272 1.20 0.230 -.0009309 .0038795
caste_group_6_Janajati | -.0005822 .0025124 -0.23 0.817 -.0055064 .004342
caste_group_6_Adhibasi | -.0005554 .0032702 -0.17 0.865 -.0069649 .005854
caste_group_6_Madhesi | .0000311 .0011247 0.03 0.978 -.0021732 .0022355
caste_group_6_Dalit | .0068043 .0038806 1.75 0.080 -.0008015 .0144101
caste_group_6_Others | .0018378 .0015812 1.16 0.245 -.0012614 .0049369
hh_size | .0026571 .0081153 0.33 0.743 -.0132487 .0185628
female | -.0008137 .0015791 -0.52 0.606 -.0039087 .0022814
married | -.0001292 .0010145 -0.13 0.899 -.0021175 .0018592
child_12 | .0047858 .0035477 1.35 0.177 -.0021675 .0117391
voc_train | .0040873 .0035942 1.14 0.255 -.0029572 .0111318
migrated_fr_job | -.0114658 .0059964 -1.91 0.056 -.0232185 .0002869
tot_chores_hrs | -.0038975 .0019915 -1.96 0.050 -.0078008 5.74e-06
urban | -.0002638 .0014127 -0.19 0.852 -.0030325 .002505
overtime_40 | .0488701 .0117921 4.14 0.000 .0257579 .0719823

[FernandoRios]

You need to work on your re weighting model.