The authors present various matching estimators of the impact on earnings for
individuals who attended public and private job training programs in Colombia.
The authors estimate propensity scores by controlling for the wide variety of
personal and socioeconomic background variables of those individuals. The effect
of training, measured by the mean impact of the treatment on the treated, shows
that: (i) for youths, no institution has a significant impact in the short or long run
except private institutions for males; the scope of the data, however, limits the
reliability of the result; (ii) for adult males, neither SENA nor the other public
institutions have a significant impact in the short or long run; (iii) for SENAtrained
adult females there are positive but not significant impacts in the short run
and greater and close to significant effects in the long run. All other public
institutions have a higher impact that is significant in the long-run; (iv) for adults
trained at private institutions there are large and significant effects in both the
short and long run, but for adult males in the short run the effects are smaller and
only barely significant. In addition, neither short nor long courses provided by
SENA seem to have a significant impact on earnings. In general, females benefit
more from both short and long courses than males. Finally, a cost-benefit analysis
shows that under the assumption of direct unitary costs equal to SENA, private
institutions are more profitable than public institutions, which are in turn more
profitable than SENA.