Oaxaca decomposition

[Biswa bhusan]

Dear user, I want to do oaxaca decomposition following https://www.stata-journal.com/articl...article=st0151
However, i have to correct selection in my specification. Therefore, i used following code by going through https://www.stata.com/statalist/arch.../msg00036.html
use oaxaca.dta
reg lnwage educ if female==0
estimates store male
reg lnwage educ if female==1
estimates store female
oaxaca male female

But unfortunately i am getting following error. variable male not found. Can any one help me?

[Marcos Almeida]

You have written more than 2 dozens posts, hence I assume you may be aware of the FAQ recommendations. Just as a remind, please try to provide data/output/command under code delimiters.


That said, the command seems according to the recommendation.

I wonder whether you have the updated version.

[Biswa bhusan]

Marcos, I am not aware of copyrights to download data and share directly. The data on which i wrote, can be obtain by typing "use oaxaca.dta". Priory to this you need to use "ssc install oaxaca". I am using Stata 13.0 MP. I am providing do file. The command does not seem to correct. Somewhere, mistake is there. Please revert to this problem.

The Output are:

reg lnwage educ if female==0

Source | SS df MS Number of obs = 751
-------------+------------------------------ F( 1, 749) = 131.04
Model | 25.5751083 1 25.5751083 Prob > F = 0.0000
Residual | 146.182033 749 .195169604 R-squared = 0.1489
-------------+------------------------------ Adj R-squared = 0.1478
Total | 171.757142 750 .229009522 Root MSE = .44178

------------------------------------------------------------------------------
lnwage | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
educ | .0753016 .0065781 11.45 0.000 .0623879 .0882153
_cons | 2.55059 .0793699 32.14 0.000 2.394776 2.706404
------------------------------------------------------------------------------

.
. estimates store male

.
. reg lnwage educ if female==1

Source | SS df MS Number of obs = 683
-------------+------------------------------ F( 1, 681) = 80.21
Model | 23.3493995 1 23.3493995 Prob > F = 0.0000
Residual | 198.250744 681 .291117099 R-squared = 0.1054
-------------+------------------------------ Adj R-squared = 0.1041
Total | 221.600144 682 .324926897 Root MSE = .53955

------------------------------------------------------------------------------
lnwage | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
educ | .078652 .0087823 8.96 0.000 .0614085 .0958956
_cons | 2.383336 .1007802 23.65 0.000 2.185459 2.581214
------------------------------------------------------------------------------

.
. estimates store female

.
. oaxaca male female
variable male not found
r(111);

[Marcos Almeida]

I wonder whether this is what you wish:

. oaxaca lnwage educ, by(female)

Blinder-Oaxaca decomposition                    Number of obs     =      1,434

           1: female = 0
           2: female = 1

-------------------------------------------------------------------------------
       lnwage |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------+----------------------------------------------------------------
Differential  |
 Prediction_1 |   3.440222   .0174724   196.89   0.000     3.405977    3.474467
 Prediction_2 |   3.266761   .0218257   149.68   0.000     3.223984    3.309539
   Difference |   .1734607   .0279579     6.20   0.000     .1186642    .2282573
--------------+----------------------------------------------------------------
Decomposition |
   Endowments |   .0457899   .0112163     4.08   0.000     .0238064    .0677734
 Coefficients |   .1296214   .0264739     4.90   0.000     .0777334    .1815094
  Interaction |  -.0019506   .0064022    -0.30   0.761    -.0144987    .0105976
-------------------------------------------------------------------------------