workplace. As a result, our analysis is somewhat speculative and perhaps contributes as much by posing research questions as by answering them. What we can establish is that the long-term increasing diversity of the U.S. workforce has been mirrored in occupations in recent years. In terms of both demographic characteristics and pay, workers in the occupations that we currently use to classify the workforce are increasingly diverse, making it more likely that such occupations include men and women, whites and blacks, more-tenured and less-tenured workers, and high-wage and low-wage workers. This increased diversity within occupations constitutes indirect evidence that the correspondence of current occupational classifications with the jobs that workers do is breaking down, since it is plausible that the increasing diversity of workforce characteristics and wages is reflected in increasing diversity of work.

Nonetheless, the evidence is limited in two important ways. First, it is indirect. It is possible, for example, that a more diverse workforce now does the same jobs that more homogeneous workers performed in the past, in which case this increasing diversity need not pose any challenge to occupational classification. However, at least from the perspective of economics, we would regard this latter possibility as far more plausible if there were not growing variance of wages within occupations; if wages ultimately reflect productivity, growing variance of wages is an indicator of growing variance of productivity, which in turn seems likely to be linked to increased diversity of work.

Second, we can document increasing variability in the type of work done within occupations as we now describe them. We do not attempt to assess the implications (presumably, the costs) of any failure of occupational classification systems to adapt to this increasing variability. On one hand, we could speculate that private businesses in a competitive environment find other ways to organize work to best utilize their workforce. On the other hand, occupational classification is also important in institutions that do not compete in the market—such as public organizations involved in training and career planning and the military.