inequality has been attributed to changing demands for skills performed on the job.

This suggests that indirect evidence can be obtained on whether skills or tasks performed in jobs are changing by looking at changes in wage or earnings inequality within occupations. Research suggests that within detailed occupations, there is relatively little variation in educational attainments. Thus, an increase in inequality within detailed occupations is unlikely to be related to changes in wage or earnings differentials associated with education. Rather, such an increase in inequality is probably due, at least in part, to changing demands for particular skills performed by workers within an occupation.

Such changes are likely to indicate one of three things. First, workers may be entering the occupation who perform new skills or tasks not previously integral to the occupation. Second, the nature of the skills or tasks required of workers in the occupation may be changing, with those who can ably perform the newly required skills and tasks earning labor market rewards, and vice versa. Third, there could simply be growing dispersion in the types of workers (differentiated by skills and tasks) in many occupations.

In all three cases, however, the evidence of rising inequality within occupations would suggest that the specific or "typical" work performed within the occupation has changed. In contrast, if the increase in inequality is largely an across-occupation phenomenon, it may be related to the same factors that have changed wage and earnings differentials by schooling, experience, etc., but it would not provide indirect evidence of changing and more variable work within occupations.

Table 2.8 reports evidence on changes in wage and earnings inequality overall, within occupations, and between occupations. These changes are also calculated over the 1983 to 1991 period, which are both relatively high unemployment years, using the Current Population Survey. The first thing the table reveals is that earnings inequality has not risen, but wage inequality has risen (i.e., an increase in the log variance of .024 indicates a 2.4 percent increase in the variance). For the purposes of this investigation, wages are of more interest than earnings, as they reflect the price of a unit of labor and do not reflect changes in hours or

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement