Investment in human capital and the nonwhite-white unemployment differential

Curtis Lloyd Gilroy, Binghamton University--SUNY


This is a study of the differential incidence of unemployment between white and nonwhite male workers in the U.S. labor force. The general question to which this study addresses itself is to what extent can the higher incidence of nonwhite compared to white unemployment at a point in time and over a period of time be explained by differences in certain characteristics of the respective labor forces.

To measure and compare the relative unemployment burdens of white and nonwhite workers, the unemployment differential will be investigated. The differential can be of two types: the relative differential, denoted by the ratio of the nonwhite to white unemployment rates (nonwhite/white), and the absolute differential defined by the difference between the nonwhite and white rates (nonwhite minus white). Neither is a completely satisfactory measure by itself. To show this, consider for example, the situation in which nonwhites and whites initially had unemployment rates of 8 and 4 percent respectively. If both rates decreased by the same absolute amount (say 2 percentage points), the relative unemployment differential would rise from 2.0 to 3.0 while the absolute differential would remain at 4.0. If both the white and nonwhite rates were halved (whites from 4 to 2 percent and nonwhites from 8 to 4) the relative differential would remain unchanged at 2.0 and the absolute differential would decline from 4.0 to 2.0. Finally, if the white rate were reduced by 4 percentage points and the nonwhite rate by 3 percentage points, the relative differential would increase from 2.0 to 4.0, while the absolute differential would fall from 4.0 to 3.0.

Since this study is focusing on the incidence of unemployment among nonwhites in relation to that of whites, the relative unemployment differential will be emphasized. In comparing the changing incidence of nonwhite and white unemployment, however, the absolute differential is appropriate. Both differentials will be used in the analysis to follow. Nonwhites will suffer a disproportionate unemployment burden if the relative differential is greater than one and the absolute differential is positive.

In the comparison of the nonwhite-white unemployment differential the extent to which the following three propositions are true shall be investigated:

  1. Unemployment rates for nonwhite workers consistently exceed those for white workers of like characteristics
  2. The cyclical variability of unemployment rates for nonwhite workers, at least in absolute terms, consistently exceeds that for comparable white laborers
  3. Cycle troughs and peaks of unemployment rates for nonwhite workers tend to lead and lag respectively the corresponding troughs and peaks for white laborers of like characteristics.

Stated another way, these propositions hold that nonwhite workers experience a disproportionate share of the unemployment burden. To what extent the burden of unemployment falls on nonwhite relative to comparable white workers will be the degree to which these propositions hold true.