[Adelio Antunes]

Dear Dr Rios,

Thanks to you and to Prof Baum for the rif_package.

I am currently working on a household health expenditure dataset to assess the determinants of inequality on a (“negative”) binary outcome: financial hardship. After discovering your package, I decided to use it for my purposes. Unfortunately, I am having difficulties in interpreting the results of “rifhdreg”.

The examples in the package documentation and your recent working paper with Dr de New provided some clarity on the interpretation of quantile distribution and Gini measures. However, I cannot get my head around the interpretation of the results of the regression on the Erreygers Concentration index (CI); which value is negative for my sample.

The table below includes the output of the regression and the means of the independent variables. How should one interpret the results of the dummy and continuous variables? Would this be correct:

• A transfer of 1 percentage point of all the households from rural to urban areas (i.e. from 0.24 to 0.25) results in a decrease of the CI from -23.687 to -24.0324(=-23.687-0.01*34.54); resulting in an increase in inequality of 1.46% (0.01*-23.687/-34.54).
• An increase of 1 percentage point in exemptions results in an increase of the CI from -23.687 to -23.343(=-23.687+0.344); resulting in a decrease in inequality of -1.45% (0.344/-23.687).

___________________________________________________________________________________
Results                                     |    mean    |    coef      |    se
___________________________________________________________________________________
Independent variables    
Urban (vs Rural) area = 1, Urban            |    0.24    |    -34.54    |   (24.39)
Exemptions [per 100 persons]                |   10.19    |      0.344*  |    (0.202)
Hospitalisations [per 100 persons]          |   12.28    |      0.162   |    (0.288)
Per capita: health expenditure [in PPP]     |   36.56    |     -0.271***|    (0.0670)
Per capita: injuries [per 100 persons]      |    0.63    |      0.808   |    (1.625)
Constant                                    |    0.24    |    -11.36    |   (10.17)
___________________________________________________________________________________
Mean RIF                                                 |    -23.687   |     
Observations                                             |    100       | 
R-squared                                                |    0.089     |     
___________________________________________________________________________________
*** p<0.01, ** p<0.05, * p<0.1               

I embedded below a data extraction and simplified syntax if needed.

* Example generated by -dataex-. For more info, type help dataex
clear
input float(outcome urban) double(freecare hospitalised) float(oop_ppp2011_pc injury_pc) double windex str5 psu double hhw float newstratum
1 1 0 0  106.6383         0   -.6653426762509831 "02017" 325.0031  21
1 0 0 0  33.93037         0   -.9940431241995353 "06041" 384.0734  62
0 1 0 0 36.192394         0   1.3772819525025548 "05010" 281.9622  51
0 0 0 0  3.393037         0  -1.5624445999821204 "17018" 428.6849 172
0 1 1 0         0         0   -.3868521018730237 "24006"  91.9543 241
0 0 0 0 12.441135         0    .5485493166751617 "18016" 206.1392 182
1 1 0 0 129.72711         0    -.477384960306328 "08034" 349.1909  81
0 0 0 0 14.081104         0   -.4793959245152082 "03042" 326.9764  32
0 0 0 0 1.3572148         0   .33021259980077117 "20029" 337.9224 202
0 0 1 0         0         0   -1.399081326933255 "14039"   465.28 142
0 1 1 0         0         0  -.32727381027949654 "06042" 318.2305  61
1 1 0 0  88.21896         0   1.4256173738302529 "12005" 260.4666 121
1 0 0 0 27.144297         0  -.34453240547787434 "03012" 299.3531  32
0 0 0 0  .6786074         0   -.5196965315819949 "03037" 243.4908  32
0 0 0 0 1.1310123         0  -.32550264283464536 "10022" 278.4566 102
0 1 0 0  5.089556         0   1.8056082002656537 "08060" 343.8567  81
0 1 0 0 .25447777         0    .8508176461046831 "18008" 174.4553 181
0 0 0 0  3.393037         0   -.6874367257691888 "14004" 465.3886 142
1 0 0 0  54.28859         0   -.7032819004174158 "20036" 330.7335 202
0 0 0 0 13.572148         0    .4458127053740737 "21043" 383.7216 212
1 0 0 0   83.4687         0   -.9198050380755813 "17046" 325.3251 172
1 0 0 0  54.28859         0   -.4634501674724758 "18024" 190.0133 182
0 1 0 0 .50895554         0     1.52427832451893 "21007" 438.8106 211
0 1 0 0  3.393037         0    1.415592204856456 "13021" 129.8951 131
0 1 0 0  6.592186         0    .4756382414458382 "01028" 376.4169  11
0 0 0 0 16.965185         0  -.11751882264082486 "17009" 304.0917 172
0 1 0 0  77.55513         0   2.0590904430614922 "12061" 649.9344 121
0 0 0 0  9.670156         0   -.2383199576340929 "21043" 383.7216 212
0 0 1 0         0         0  -.47450342632685916 "15002" 297.8626 152
1 1 0 2 101.79111         0   -.7405601489787412 "17027"  493.183 171
0 1 0 0  9.694391         0    .5118759227271751 "05014" 441.5103  51
0 1 0 0  50.89556         0    3.052973628870931 "02014"  310.817  21
1 0 0 0 16.965185         0  -1.1729055417514695 "22005" 294.7763 222
0 1 0 0 1.5268667         0     .860137215195725 "17032" 177.3949 171
0 1 0 0  54.28859         0   1.3457775437745838 "19010" 201.6258 191
0 1 0 0 2.2620246         0   1.0858306852204567 "19011" 170.6817 191
0 1 0 0  5.089556         0   -.7616071307673543 "24008"  82.3284 241
0 0 0 0 4.0716443         0  -.48414949508664645 "17005" 416.9502 172
0 0 0 0 16.965185         0 -.027396483334440776 "17023" 414.3553 172
1 0 0 0 101.79111         0   -.4271800169886727 "06038" 388.4588  62
0 0 0 0  6.786074         0   -.4931124118055506 "01014" 396.9657  12
0 0 0 0  4.847196         0  -.16262726190923735 "16013" 234.6134 162
0 0 0 0  6.022641         0  -.20062561335083703 "17012" 433.5184 172
0 0 0 0 2.2620246         0    .7387068063805651 "15008" 219.3244 152
1 0 0 0  33.93037         0   -.7081652499218105 "07006" 353.0623  72
0 0 0 0  42.41296         0    .9477729275093842 "09015" 166.6411  92
0 0 0 0  4.071645         0   -1.326667630984104 "02002" 430.9372  22
0 0 0 0  9.048099         0    -.509134866640311 "14025" 474.2442 142
0 0 0 0 13.572147         0    -.288705236932152 "06024" 385.1814  62
0 0 0 0  11.02737         0    .5440175900545187 "03043" 494.7862  32
0 0 2 0   8.14329         0  -.10708125360690628 "07006" 353.0623  72
0 1 0 0  3.223385         0    .5556016644151214 "01025" 364.8701  11
1 0 0 0  50.89556         0  -.49740898115228355 "16020" 208.5324 162
0 0 0 0 1.8850206         0  -1.1249240316384668 "16021" 145.4773 162
1 1 0 1 22.620247         0   -.9725244311511236 "20033" 249.0567 201
1 0 0 0  67.86074         0   -.6796091517527205 "17047" 408.7627 172
0 0 0 0  6.786074         0   -.8707770259839175 "06034" 377.2153  62
0 0 0 0  3.393037         0  -.00499992656713616 "02009" 415.8561  22
0 1 0 0 16.588182         0   2.0983921231359437 "25046" 210.7115 251
0 0 0 0 1.5834173         0    .8482866654772797 "03028" 383.8385  32
0 0 0 0 4.2412963         0  -1.3092995834052046 "06040" 374.1129  62
0 0 0 0  3.393037         0   -.6057290653249331 "07035" 360.6338  72
1 0 0 0   36.6448         0    .6471823873256822 "07013" 253.4089  72
1 0 0 0  79.17086         0  -1.2387824476145446 "16007"  191.709 162
0 0 0 0 1.6965185         0    .6539780550681383 "07008" 347.1472  72
1 0 0 3 168.52084         0   .06081976649467262 "02007" 275.1536  22
0 0 0 0 4.4109483         0  -.48726285886212756 "07007" 333.1601  72
0 0 0 0 16.965185         0  -.15277320997312518 "25043" 230.3937 252
1 0 0 0  81.43289         0    .2355084184970267 "06010" 383.1199  62
0 0 0 0  .9048099         0   .41244154627889124 "21032" 401.5766 212
0 0 1 1         0         0   -.9744729721056214 "14057" 465.4036 142
0 1 0 0 1.3572148         0    1.331114095603382 "15032" 211.4474 151
0 0 0 0         0         0  -1.2370125216897607 "10002" 279.2827 102
0 0 0 0  16.28658         0   .20590989997474568 "04015" 343.5223  42
1 0 0 1  54.28859         0   .40296103286780943 "15025" 310.4955 152
0 0 0 0  9.048099         0   -.6681651115018918 "03050" 443.1528  32
1 0 0 0 21.715435         0    -.585522282008272 "06035" 435.9285  62
1 0 0 1  904.8099         0   -.7303145877819389 "13016" 245.9765 132
0 0 0 0 16.965185         0   2.0257376351641216 "03002" 406.7226  32
0 0 0 0 18.096197         0    .3212616903201595 "14011" 517.8802 142
0 1 0 0 12.441135  .3333333   .35537333442783675 "05032" 573.3594  51
0 1 1 0 .16965185         0   2.6794682832144145 "05011" 288.0924  51
1 0 0 0  67.86074         0    -.781898741362883 "08020" 432.6883  82
1 0 0 0  33.93037         0   .10507800015825447 "01008" 672.4203  12
1 0 0 0  84.82593         0  -.07319974674366442 "15009" 297.1825 152
0 0 2 0         0         0   -.7185192770766162 "16008" 218.8445 162
0 1 0 0 10.179111         0  -1.1508447752977298 "15017"  288.911 151
1 1 0 0  67.86074         0    .7439100234892032 "12085" 663.2095 121
1 0 0 1 226.20247         0  -.42621138336921416 "14021" 489.3474 142
0 0 1 0         0         0  -.05922603648259058 "14006" 465.4653 142
0 0 0 0 15.834172         0   -.4451799770448089 "17013" 428.8367 172
0 0 0 0  6.786074         0   -.2391803573136902 "02047" 475.5555  22
0 0 0 0 11.310123         0    .5357359651864448 "03004" 412.5359  32
0 0 0 0 1.1310123         0   .12142710919226243 "14049" 445.7798 142
1 0 0 1  201.5464         0  -1.0138132342656636 "07016" 362.2616  72
1 0 0 0 27.144297         0     .155596722067095 "06045" 360.0294  62
0 0 0 0         0         0   -.9694810242369618 "17006" 453.8851 172
1 0 0 1  38.17167 .08333334   1.4545317730887652 "17046" 325.3251 172
0 0 0 0 19.001009         0   .41772459678200513 "03039" 406.3593  32
1 0 0 0  47.50252         0   -.7190750614795979 "20014" 328.6322 202
end
 
 
* rescale variables to obtain meaningful coefficients
            
local Varlist freecare hospitalised injury_pc
 
             foreach Var of var `Varlist' {
             replace `Var'=`Var'*100
             }
      
label values outcome NOYES
label def NOYES 0 "No", modify
label def NOYES 1 "Yes", modify
label values urban URBAN
label def URBAN 0 "Rural", modify
label def URBAN 1 "Urban", modify
label var outcome "Outcome"
label var urban "Urban (vs Rural) area"
label var freecare "Exemptions [per 100 persons]"
label var hospitalised "Hospitalisations [per 100 persons]"
label var oop_ppp2011_pc "Per capita: health expenditure [in PPP]"
label var injury_pc "Per capita: injuries [per 100 persons]"
label var windex "Wealth index"
label var hhw "household weight"
label var newstratum "Stratum"
 
svyset psu [pweight=hhw], strata(newstratum)singleunit(certainty) 
 
* Distribution measure: Erreygers normalised concentration index
conindex outcome, rankvar(windex) limits(0 1) err bounded svy
 
mean urban freecare hospitalised oop*1_pc injury_pc [aw=hhw]
 
********************* RIF Determinants analysis ********************           
local DepVar outcome
 
local IndVar i.(urban) c.(freecare hospitalised oop*1_pc injury_pc)
 
rifhdreg `DepVar' `IndVar', rif(eindex(windex) lb(0) ub(1)) retain(outcome_rif) replace iseed() scale(100) svy
outreg2 using ".\temp.xls", excel label sideway replace onecol

[Rui Yang Mia]

Dear Prof. Fernando Rios,

I am writing to seek your expertise regarding some difficulties I've encountered in estimating the gender wage gap using different methodologies in Stata.

I have used three different approaches to estimate the gender wage gap: the oaxaca_rif command, Unconditional Quantile Regression (UQR) with log annual wages as the dependent variable, and a manual calculation using detailed summaries of log wages by gender across various percentiles (state command: sum logwage if female==0/1, detail).

The oaxaca_rif command provides a difference between two groups (group1 and group2). Could you please explain what this difference represents? Why might it significantly differ from the coefficients obtained via UQR for females?

Additionally, the wage gaps I manually calculated from detailed summaries do not align with the results from the oaxaca_rif results. Each method seems to yield a different magnitude of the gender wage gap.

I am greatly confused by these variations and would greatly appreciate your insights on why these differences might occur and how each method approaches the estimation of wage gaps.

Thank you very much for your time and assistance. I look forward to your expert guidance.

Many thanks,

[FernandoRios]

I cannot say much since I do not see the numbers that you see nor the sample itself
some of the potential explanations could be sample restrictions due to missing data in your explanatory variables.
other than that not sure what else could be causing the problem