AN EXCHANGE: ON THE APPROPRIATE SPECIFICATION OF HUMAN CAPITAL MODELS .

Item request has been placed! ×
Item request cannot be made. ×
loading   Processing Request
  • Additional Information
    • Abstract:
      This article presents the views of the author on a paper by Alan Blinder which attempted to quantify the separate components of the wage differential between males and females that are attributable to differential endowments and to differences in employers' evaluations to these characteristics. Using data compiled by the Michigan Survey Research Center, he estimates two sets of equations to account for the sample variation in the log of male and female hourly wages - a reduced form equation containing 13 family background variables plus age, regional, labor market, mobility, and health parameters and a structural equation with schooling, age, age squared, occupation, union membership, and job tenure variables. From the results obtained with the latter specification. Blinder concludes that men earn much larger wage increments for advancing to higher educational levels. We have shown that the use of age rather than work experience in structural wage equations is likely to result in a biased estimate of the difference in the sex-specific rates of return to schooling favorable to males. While the sensitivity of the estimated magnitudes of the individual components of the male-female wage differentials to the specification of the earnings function employed has been demonstrated, our results do not contradict Blinder's conclusion that a considerable portion of the wage differential between men and women cannot be explained simply by differing endowments of schooling or postschool training.