st: Oaxaca: error with 2 large sets of categorical variables

["Yigit Aydede" <yigit.Aydede@SMU.CA>]

Hello

I appreciate if somebody would help me on this.  

I use oaxaca on my data set.  It's a simple wage decomposition.  I generated two large sets of categorical variables: 48 categories for the occupation and 30 for the industry.  Both sets have the base category.  To generate these sets I used the tab command (i.e. tab occupation, gen (occ)).  My oaxaca command is as follows

oaxaca lnwage age agesq occ2-occ48 ind2-ind30, by(female) weight(1) detail noisily categorical(occ? ind?)

The error message:  dropped coefficients or zero variances encountered
specify -relax- to ingnore
r(499);

And
 _error():  3300  argument out of range
      oaxaca_normalize():     -  function returned error
                 <istmt>:     -  function returned error

Every time the result has an arbitrary omitted variable.  I checked the dummies, nothing is wrong with them.  To test them I put one set of dummies at a time. That is when I use the following

oaxaca lnwage age agesq occ2-occ48, by(female) detail noisily weight(1) categorical(occ?)
OR
oaxaca lnwage age agesq  ind2-ind30, by(female) detail noisily weight(1) categorical(ind?)

both works very well without any error

To test further, I used oaxaca.dta that comes with the command.

I generated isco dummies as stated in the help and education by using the tab command.  I got 9 dummies for isco and 10 dummies for education

When I use them together, I have similar error messages.  And no problem if I use one of the series (disco or edu)

I appreciate fi somebody would help me on this.

All the best,
--Yigit 


Yigit Aydede
Department of Economics
Saint Mary's University
Halifax, NS, B3H 3C3
Canada

T: (902) 420-5673
F; (902) 420-5129



*
*   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/