Juan M. Rodriguez Poo, Stefan Sperlich, and Ana I. Fernandez,
"Semiparametric Three Step Estimation Methods For Simultaneous Equation
Systems", Journal of Applied Econometrics, Vo. 20, No. 6, 2005, pp.
699-721.

The source of the data is the "Encuesta de Poblacion Activa" (EPA), the
Spanish Labor Force Surveys. These surveys have been carried out on a
quarterly basis since 1975 and are collected by the National Bureau of
Statistics (INE). They cover several thousand households and contain
information about ALL individuals therein aged over 16.  From these surveys,
the National Bureau of Statistics randomly selected a cross-section, in the
second quarter of 1990, providing additional information about some
variables that were considered relevant for labor market participation
analysis. In this paper, we consider a subsample (cross sectional, of people
interviewed at the same time) of 1010 individuals participating in the labor
market, 612 workers and 398 non-workers. Some basic statistics can be found
in the above mentioned paper.

All data are in the file psf-data.txt, an ASCII file in DOS format. There
are 14 columns, the first 12 of which contain data that were actually used
in the paper. This file is zipped in the file psf-data.zip.

The variables are

 SEX    1 if male
 AGE1   1 if age 16 to 19
 AGE2   1 if age 20 to 25
 AGE3   1 if age 26 to 35
 AGE5   1 if older than 45
 EDUC1  1 if elementary school
 EDUC2  1 if high school
 EDUC3  1 if university
 URATE  unemployment rate of district
 SINGLE 1 if single
 NOHH   1 if not household head
 WAGE   earnings per hour in Pesetas
 ----   information not provided (these variables were not used in 
        the final paper version)