Differences in earnings, skills and labour market experience among young black and white men

Abstract

This study examines the role of racial differences in skills and labour market experience on recent earnings differences between young black and white men. Our analysis of the 2000 sample of the National Longitudinal Survey of Youth, 1979 Cohort (NLSY79), indicates that nearly all of the earnings gap between black and white men can be accounted for with a relatively parsimonious empirical model. In particular, it finds that approximately 44% of the racial earnings gap results from higher average skill and work experience levels of white men.

Notes

¹ Potential experience is defined as the difference between a worker's age and level of schooling completed, less the age when the worker began school, usually assumed to be 5 or 6 years old (Mincer, 1974).

² The notion of a full-time worker is based on the US Department of Labor's definition of year-round full-time: 35 or more hours per week, 50 or more weeks per year (1750 or more hours per year). Blau and Kahn (1997) used a similar measure, however, they use 1500 hours as the full-time equivalent cutoff level.

³ The 2000 NLSY79 cohort is comprised of workers between the ages of 35 and 43. The samples consist of workers with positive wage and salary income. Full-time military personnel and individuals who are self-employed, enrolled in school and missing information on annual hours worked were excluded. Beginning in 1994, the NLSY is conducted every two years. Thus, between 1994 and 2000 there are three missing calendar years of data on labour market hours: 1994, 1996 and 1998. Due to the small sample sizes, and to the potential interaction between race and ethnicity, Hispanic workers were excluded from the analysis.

⁴ Here one adopts the standard technique of using white men as the ‘favoured’ group, thus assuming their regressions represent the ‘discrimination free’ wage structure. Other specifications for the discrimination free wage structure (see Cotton, 1988) did not yield significant differences in our conclusions. These estimates are available from the author upon request.

⁵ The results are more relevant to measuring post-market, as opposed to pre-market, discrimination. For a discussion of the evidence of potential pre-market discrimination against blacks, see Neal and Johnson (1996).