America Becoming: Racial Trends and Their Consequences: Volume II (2001)

Cecilia A.Conrad

In 1950, Black women earned, on average, sixty cents for every dollar earned by White women.¹ Between 1960 and 1980, this wage gap disappeared. No documented racial trend between 1950 and 1980 is quite as impressive. Unfortunately, the improvement in relative earnings did not continue past 1980; and this post-1980 deterioration in relative earnings was not limited to Black women. The Hispanic/White median weekly earnings ratio for year-round, full-time workers was 0.86:1 in 1980 but only 0.75:1 in 1996. The earnings ratio for Asian² women, who have higher average earnings than White women, fell from 1.50:1 in 1988³ to 1.38:1 in 1996.

Analyzed here are two trends—the improvement in Black-White relative earnings from 1950 to 1975 and the increase in racial inequality after 1980. In the process, trends in employment and in the occupational distribution of women workers will be examined. The ultimate goal is to assess

the prospects for racial equality in labor-market status—Are we moving toward equality of opportunity and access?

The experiences of the many racial and ethnic groups represented in the U.S. labor market are individually unique; however, time, space, and data limitations preclude describing each group at the same level of detail. The guiding principle was first to identify interesting trends, then analyze the position of the groups involved in those trends. In some instances, there were too few data or there is too little available research on the experience of a particular group to offer substantive insights.

There is ample anecdotal evidence of historical employment discrimination. Before 1960, it was difficult for a Black woman to get a job as a clerical worker whatever her credentials. Goldin (1990) observes that the only employers of Black clerical workers were Black-owned insurance companies and Montgomery Ward, then exclusively a mail-order business. In a 1940 Women’s Bureau survey, more than 50 percent of employers reported that they had a company policy against hiring Black women as clerical workers (Goldin, 1990). Weaver’s classic study of the plight of Black labor documents the exclusion of Blacks from skilled blue-collar jobs, particularly in the South (Weaver, 1946). During World War II, Black women were urged to help the war effort by taking jobs as domestic servants so that White women would be free to work in manufacturing jobs (Jones, 1985b). Discriminatory employment practices and segregation by occupation and industry were extensive, and contributed to differences in pay structure and indirectly to the Black-White female pay gap

During the 1960s and the 1970s, the median earnings of Black women increased relative to earnings of White women, White men, and Black men (Figure 6–1). Black/White female wages came closest to parity in 1975; from 1980 to 1996, the earnings of Black women relative to those of White women fell. Young Black women, in particular, have seen a reduction in relative earnings (Blau and Beller, 1992; McCrate and Leete, 1994). Blau and Beller (1992) found a reduction in the wages of young Black women, following an adjustment for selectivity bias—in other words, observed wages understate the extent of the wage gap because young Black women with the largest skill deficits had dropped out of the labor market.

FIGURE 6–1 Ratio of Black women’s earnings compared to those of White women, Black men, and White men. SOURCE: March Current Population Surveys, 1967–1996.

Table 6–1 is based on census data of median-income ratios for Hispanic, Asian, and Black women. The Hispanic/White ratio increased between 1960 and 1970, but there was less of an increase than for Black women. Asian (in this instance, Japanese, Chinese, and Filipino) women had higher average earnings than White women in 1960 and retained that advantage in 1980. Filipino women had an increase in relative income greater than that of Black women.

Figure 6–2 charts the earnings of Hispanic, Asian, and Black women, as a ratio of earnings of White women, for year-round, full-time workers since 1974. For Hispanic women, the earnings ratio is relatively stable through the 1980s but decreases in the 1990s. For Asian women, data were not available prior to 1988; since then, however, relative earnings decreased through 1990, rebounded in the early 1990s, and then decreased again.

Figure 6–3 depicts gender differences in earnings. For all races, the ratios of women’s to men’s earnings increased during the 1980s and have been fairly stable since 1990. The gender gap is smaller for racial and ethnic minority groups than for Whites.

The increase in relative earnings of Black women between 1960 and 1975 was accompanied by a shift in their occupational distribution. Before 1960, a White woman without a college degree could find employment as a secretary, a sales clerk, or as a blue-collar operative. A Black woman, particularly in the South, had one option—domestic service. In 1960, more than 33 percent of all Black women worked as domestic servants (“private household workers”); only 3.2 percent of White women held these jobs (statistics cited in this paragraph are from the 1960 and 1980 Censuses). In 1960, of all women with 12 years of schooling, less than 20 percent of Black women, compared to an estimated 54 percent of White women, held jobs as clerical workers; and White women were 1.6 times more likely to be employed as blue-collar operatives than were Black women. In the South, less than 10 percent of all Black women worked as operatives; whereas, 18.5 percent of White women worked in this occupation. By 1980, however, less than 8 percent of Black women worked as domestic servants, the percentage employed as clerical workers had more than tripled, and, in the South, 14 percent of Black women worked as operatives compared to 8 percent of White women.

A U.S. Civil Rights Commission study (Zalokar, 1990) reports indices of occupational dissimilarity for Black and White women in 1940, 1960, and 1980. The index of occupational dissimilarity measures the percentage of Black, or White, women who would have to change occupations in order for the Black-White distribution to be equalized.⁴ In 1940, 64 percent of Black women would have had to change occupations to equalize the distribution; in 1960, 51.9 percent; by 1980, 20.4 percent. Among women with 12 years of schooling, 43.7 percent would have had to change jobs in 1960 for equalized occupational distributions; 21.3 percent in 1980; 14 percent in 1990; and 16 percent in 1997.⁵

Table 6–2 lists the indices of occupational dissimilarity for Asian-American (Chinese, Japanese, Filipino), Hispanic (Mexican and Puerto Rican), and Black women. Because the index is sensitive to the definition of occupational categories, the indices for Black women are not identical to those reported by Zalokar (1990), but they are close. The indices for American women of Japanese, Filipino, and Puerto Rican background show the same pattern as for Black women. The occupational distributions have become markedly more similar since 1960. However, for women of Mexican and Chinese background, the distributions have not converged since 1960. In 1960, approximately 24 percent of Mexican-

American women would have had to change jobs to make their occupational distribution identical to that of White women; in 1990, the percentage was unchanged. For Chinese-American women, only 10.5 percent would have had to change jobs in 1960 to make the distributions similar; in 1990, it was 15 percent.

The percentage of Hispanic women employed as clerical workers increased between 1960 and 1980 as it did for Blacks, but the changes were less dramatic. In 1960, 16 percent of Mexican-American women were employed as clerical workers; in 1980, 26.2 percent; and in 1990, 25 percent (U.S. Bureau of the Census). For Puerto Rican women, the percentage employed as clerical workers increased from 13.0 percent in 1960, to 31.9 percent in 1980, to 31 percent in 1990.

Figure 6–4 shows the change in the percentage of female workers employed in administrative and clerical work, by race, between 1960 and 1996. For women of Japanese and Chinese background, and White women, the percentages in this occupation decreased. Meanwhile, percentages increased dramatically for Puerto Rican and Black women. Based on an analysis of data from National Longitudinal Surveys, Power and Rosenberg (1993) suggest that White female clerical workers have more upward mobility than Black female clerical workers. They analyzed the

FIGURE 6–4 Change, by race/ethnicity, in the percentage of women employed as clerical workers, 1960–1990. SOURCE: U.S. Censuses of the Population, 1960 and 1990.

occupational positions of young women in 1972 and 1980 and found that among the young Black women who were clerical workers in 1972, 71.6 percent were still employed as clerical workers in 1980. By comparison, only 58.5 percent of the 1972 White clerical workers remained in that occupation. Both Black and White women left clerical work primarily for jobs as professional and technical workers or as managers and administrators (Whites); however, the Black women who moved into professional and technical jobs tended to be more concentrated in low-paying, female-dominated jobs than were White women.

Changes in the occupational distribution of Asian women paralleled the changes for Whites. Figure 6–5 shows the proportion of each racial/ ethnic group employed as executives and managers. In 1960, 3.8 percent of Japanese-American women were employed as executives and managers; by 1980, this percentage had increased to 8.3 percent. For Chinese-American women, the percentage increased from 5.4 percent to 10.4 percent; and for Filipino-American women, from 1.6 percent to 6.4 percent. For White women, the percentage employed as executives and managers was 4.3 percent in 1960 and 7.8 percent in 1980.

FIGURE 6–5 Change in percentage of women employed as executives and managers, 1960–1990. SOURCE: U.S. Censuses of the Population, 1960 and 1990.

Between 1960 and 1980, the concentration of Black women working in third-sector employment—i.e., government and nonprofit organizations like educational institutions and hospitals—increased relative to that of other groups. Burbridge (1994) reports an increase of 73 percent between 1960 and 1980, while the percentage employed in the for-profit sector decreased 36 percent. In contrast, for all workers, the percentage employed in the for-profit sector decreased only 11 percent. An analysis of 1997 Current Population Survey (CPS) data, shown in Table 6–3, reveals the concentrations of Black and Hispanic women among professions and executives in third-sector employment.

Ironically, as Black women shifted out of domestic service, White women moved from housework to paid employment. Labor-force participation rates (LFPR) for White women increased rapidly between 1960 and 1980, outpacing increases for other groups. As a result, the LFPR gaps between Black and White women and between Asian and White women narrowed. The gap between White and Hispanic women increased.

Table 6–4 shows LFPR trends between 1960 and 1980 by racial group. For White women, LFPR increased from 40.3 percent in 1960, to 49.6 percent in 1980, to 57.6 percent in 1990. For Black women, LFPR increased from 49.9 percent in 1960, to 52.9 percent in 1980, to 59.3 percent in 1990. Relative decreases in LFPR for Black women were greatest for those who had never been married (who had lower LFPR than their White counterparts) and among younger women (Jones, 1985a). LFPR for women 35 and older increased at a similar rate for Blacks and Whites. Young, never-married Black women remained less likely to be employed or actively looking for work than did their White counterparts, and the gap is increasing. Filipino women’s LFPR, in 1960, was the lowest among Asian Americans (Filipino, Japanese, and Chinese). LFPR increases for Filipino women, however, rival those of White women; 36.4 percent of Filipino women were either employed or actively looking for work, compared with 40.3 percent of White women. LFPR for Filipino women increased from 36.24 in 1960 to 72.3 in 1990.

Unemployment rates are higher for Black and Hispanic women than for White women. Between 1975 and 1990, the gap between Black and White women increased, as Figure 6–6 illustrates. Hispanic women, both Mexican-American and Puerto Rican, have exceptionally high rates of unemployment compared to White women.

FIGURE 6–6 Unemployment rates for women, age 20 and older. SOURCES: Black (1954–1971), Bureau of Labor Statistics, Series LFU21001722, http://www.bls.gov; Black (1972–1999), Bureau of Labor Statistics, Series LFU21001732, http://www.bls.gov; White, Bureau of Labor Statistics, Series LFU21001712, http://www.bls.gov; Hispanic, Bureau of Labor Statistics, Series LFU21000050, http://www.bls.gov.

Neoclassical economic theory emphasizes the importance of human capital in explaining differences in earnings among individuals. The principal tool used to investigate the role of human capital is a regression model. In the standard regression model, the dependent variable is the natural logarithm of earnings, and the set of explanatory variables includes measures of schooling, work experience, and indicators of marital and family status. Controls for occupation and industry may also be included. This estimated model can then be used to explain differences in mean wages by comparing the mean characteristics of two groups. For

example, if an additional year of schooling increases earnings by $1,000 per year, and White women have, on average, one additional year of schooling than Black women, then difference in schooling explains $1,000 of the observed earnings difference between Black and White women. Once all relevant productivity characteristics have been included in the model, the remaining pay difference is unexplained and attributable either to unmeasured differences in productivity (that could reflect premarket discrimination) or to labor-market discrimination.⁷ Using this methodology, a decrease in the earnings gap between Black and White women might be explained by either (1) changes in the measured productivity characteristics of Black and White women, (2) changes in the return to productivity characteristics, (3) changes in unmeasured productivity, or (4) a decrease in labor-market discrimination.

This methodology has some important limitations. First, the omission of relevant productivity characteristics, or the measurement of some characteristics with error, can bias results. For example, many studies use a measure of potential experience (age minus years of schooling minus six) when data about actual experience are unavailable. For Black and White women of the same age, potential experience might be identical, but actual experience could be quite different. This is a particular problem for the analysis of long-term trends, given racial differences in LFPR. The primary source of data, the Public Use Micro Samples (PUMS) of the decennial U.S. Census, does not have information about actual work experience. In addition, before 1990, the only measure of education available in PUMS was years of formal schooling completed. Hence, with PUMS data, racial differences in the returns to schooling might reflect differences in the rate of degree attainment. If, for the same number of years of schooling, White women are more likely to have earned a degree, then an additional year of schooling will have a bigger effect on their earnings. CPS data have similar problems.

A second limitation is the potential for sample-selection bias. Individual decisions about whether to participate in the labor market will affect the observed structure of pay. For example, if only high-ability White women participate in the wage sector, but high-, medium-, and low-ability Black women work, the observed returns to schooling are likely to be higher for White than for Black women. With a simple-regression model, a researcher might attribute this difference to labor-

market discrimination when, in fact, it reflects differences in productivity. Unfortunately, there are no easy solutions to this problem. Manski (1989) offers a technical discussion of the selection problem and discusses the limitations of the standard solution, a two-stage estimation procedure proposed by Heckman (1976). The results of this procedure are very sensitive to model specification. Alternative approaches, such as the use of robust, nonparametric estimators, have other limitations. According to Greene (1997:983), “the issue remains unsettled.”

For the most part, studies reviewed in the following section do not correct for sample-selection bias. Several authors report that they attempted the standard, Heckman two-step estimation procedure: Cunningham and Zalokar (1992) report in a footnote that they found no evidence of sample-selection bias when they attempted the procedure; McCrate and Leete (1994) described their results using the Heckman procedure as “problematic.” Blau and Beller (1992) did correct for selection bias in trends in earnings ratios, but not in estimated-wage equations.

The studies cited in this section, although not numerous,⁸ use different sources of data and different sets of explanatory variables.⁹ Nonetheless, they reach remarkably similar conclusions about the contributing factors: essentially that the improvement in relative earnings between 1950 and 1980 is explained primarily by (1) the expansion of employment opportunities outside of private-household service; and (2) a decrease in labor-market discrimination either because of increasing competitiveness of the U.S. economy or enforcement of equal employment opportunity (EEO) laws, or both. Improvements in educational attainment also contributed, but were not the major factor (Zalokar, 1990; Cunningham and Zalokar, 1992; Carlson and Swartz, 1988).¹⁰

Behind the expanded opportunities and decrease in labor-market discrimination are three possible explanations. The first is the convergence of unmeasured productivity characteristics, such as the quality of schooling [wage equations estimated by Zalokar (1990) did not include controls for the quality of schooling]. If Black women, on average, attended lower quality schools, then a year of schooling might have a smaller effect on wages for Black women than for White women. If school quality improved for Blacks between 1960 and 1980, differences in the structure of pay could be expected to disappear. A second explanation is a decrease in racial discrimination in wage setting—i.e., when, for a specific job, an employer offers a Black woman lower wages than are offered to a White woman with the same credentials. Legislation at both the state and federal levels made discrimination in wage setting illegal in the 1960s. The third explanation is the convergence of occupational distributions. Findings from Zalokar (1990), Cunningham and Zalokar (1992), and Blau and Beller (1992) suggest that the decrease in occupational segregation between 1960 and 1980 was a major contributing factor to the increase in relative wages of Black women.

Albelda (1986), basing findings on a regression analysis of the determinants of changes in occupational distribution from 1962 to 1981, attributes the convergence of Black-White occupational distributions to educational attainment and to structural shifts in the economy. Zalokar (1990:53) concludes:

Enforcement of EEO regulations and affirmative action were also factors in the decrease in racial discrimination after 1960. Title VII of the Civil Rights Act of 1964 prohibited discrimination on the basis of sex and race in any aspect of employment. In 1965, Executive Order 11246 required that all federal contractors have affirmative action plans specifying goals and timetables for increasing the representation of women and minorities in their work force. Although there is some debate about the magnitude of the impact of these laws, they appear to have increased employment opportunities for Black women (Leonard, 1990; Heckman and Payner, 1989; Betsey, 1994). Fosu (1992) attributes to Title VII a significant increase in Black women’s occupational mobility from 1965 to 1981.

Although a consensus may have been reached about the convergence in earnings of Black and White women, there is no consensus about the divergence during the 1980s and 1990s. McCrate and Leete (1994), Anderson and Shapiro (1996), and Blau and Beller (1992) find evidence of a divergence in pay structure, particularly in the rates of return to education, but disagree about the causes. Darity et al. (1996) and Blau and Beller (1992) report decreasing, or no, wage discrimination.

Using PUMS data for persons 25 to 54 years old, Darity et al. (1996:420) conclude “there is no systematic evidence of discriminatory differentials affecting the incomes” of Black women after 1980. Blau and Beller (1992) report a decrease in the unexplained wage differential for older women (20 or more years of potential work experience) between 1971 and 1988, but note also an increase for younger women (especially those with 0 to 9 years of experience). Using CPS data, they compare wages earned by an older Black woman with those earned by an older White woman with the same credentials, and estimate that in 1971 the ratio was 80.6:1; in 1981, 90.8:1; and in 1988, 91:1. In contrast, for a Black woman with 0 to 9 years of potential experience, compared to a White woman with equivalent credentials, the ratio was 96.2:1 in 1971; 94.4:1 in 1981; and 93.2:1 in 1988. For the 10-to-19-years-of-experience cohort in Blau and Beller’s study (1992), the unexplained wage differential decreases between 1971 and 1981, but increases between 1981 and 1988, perhaps the result of unmeasured characteristics, perhaps indicative of a decrease in school quality for Blacks.

McCrate and Leete (1994:180) express skepticism of this explanation because young Blacks had “recently at least held their own relative to Whites in measures of reading proficiency, scientific knowledge and math achievement.”¹¹ Using data from the 1977 and 1986 NLS, they report a similar divergence in pay structures for women ages 23 to 28; in particular, the rate of return to education for Black women decreased relative to that of White women. Anderson and Shapiro (1996), in an analysis of NLS data for women ages 34 to 44, report that wages were closer to parity and that more of the gap in wages could be explained by differences in characteristics in 1980 than in 1988.

The fact that unexplained differentials go in different directions for

different experience groups raises questions about the role of labor-market discrimination in addition to the possible contribution of unmeasured characteristics, such as the quality of schooling. Leonard (1990) argued that the enforcement of EEO regulations effectively ended in the late 1970s; hence, employers with tastes for discrimination might have been able to exercise them with relative impunity. However, Blau and Beller (1992) prefer to attribute the changes to unmeasured characteristics rather than to discrimination.

Bound and Dresser’s (1999) results show that college-educated Blacks experienced a greater drop in relative earnings than did women with less than a college degree. They attribute some of the relative decrease in returns to a college degree to the greater occupational mobility of college-educated White women. However, this explanation begs the question, Why are White women more upwardly mobile?

England et al. (1999) attribute racial differences in occupations to differences in skills. Following Neal and Johnson (1996), England et al. use Armed Forces Qualification Test (AFQT) scores as a measure of cognitive skill. Comparing AFQT scores with data on young women from the National Longitudinal Survey of Youth (NLSY), they conclude that differences in cognitive skill contribute to wage differences between Black and White, and Hispanic and White women. Furthermore, they argue that differences in cognitive skill are correlated with occupations. Black women and Hispanics occupy jobs that require fewer cognitive skills than jobs occupied by White women.

The appropriateness of the AFQT score as a measure of cognitive ability is the subject of ongoing debate.¹² If the test is racially biased, as some critics suggest, then the AFQT score is a proxy for race and its inclusion will bias against a finding of discrimination. If the AFQT score does measure a productivity-relevant attribute, its exclusion will bias in favor of finding discrimination. Even if differences in AFQT scores do explain differences in wages, no study has investigated the ability of AFQT scores to explain the erosion of relative wages over time.

Earnings differentials between Hispanic and Asian and White women appear to be explained primarily by differences in characteristics. Darity

et al. (1996) and Carlson and Swartz (1988) report no evidence of racial wage discrimination against Asian women; and Reimers (1985a), Carlson and Swartz (1988), Darity et al. (1996), and Baker (1999)¹³ report little evidence of racial wage discrimination against Hispanic women. Hence, it is reasonable to expect that changes in productivity characteristics explain the decrease in relative wages.

Asian women have higher average earnings than White women because they have higher educational attainment and more work experience, and they are more concentrated in professional occupations (Carlson and Swartz, 1988). Some Asian women may have even higher returns to productivity than White women.¹⁴ Hence, the most likely explanation for the decrease in relative wages of Asian women is a decrease in the size of their human capital advantage. White women have increased their rates of college completion relative to Asian women; and as more White women engage in paid labor, Asian women may lose their edge in work experience.

Reimers (1985a), using data from the 1976 Survey of Income and Education (SIE), attributes the difference in pay between Hispanic and White women to differences in education, family size, and to immigrant status. More recent studies corroborate her findings (Carlson and Swartz, 1988; Darity et al., 1996; Baker, 1999). But there are mixed results regarding the effect of English language ability on earnings. Carlson and Swartz (1988) obtained a statistically significant coefficient on English ability in an earnings equation for Mexican-American women, but not for Cuban-American or Puerto Rican women. In Reimers’ estimated equations, the coefficient on English proficiency is never statistically different from 0. Baker (1999), in a study that focused on Mexican-American women in the South-west, found a positive and statistically significant effect of English ability on earnings in 1980 but not in 1990.

Immigration complicates any analysis of Hispanic/White wage ratio trends. Because immigrant status depresses wages (Reimers, 1985a; Baker, 1999), changes in the percentage of immigrants in the Hispanic population would affect the wage ratio. In addition, the presence of Hispanic immigrants in a labor market might affect the wages of Hispanic women

and White women differently. Baker (1999) analyzed the wages of Mexican-American, Black, and other women workers in Southwestern labor markets and found evidence that an increase of Mexican immigrants depressed the wages of U.S.-born Mexican women, had no effect on the wages of U.S.-born Black women, but increased the wages of U.S.-born White women.¹⁵

Given the diversity of experiences among Hispanics, it is conceivable that the decline in relative earnings reflects a change in the composition of the Hispanic female labor force. However, this explanation seems unlikely for several reasons. (1) If immigrants are excluded, the trend in relative wages persists; the Hispanic/White wage ratio decreased for U.S.-born Mexican women and for mainland-born Puerto Rican women after 1980. (2) The percentage of Puerto Rican women workers who are island-born decreased since 1980, but there was no improvement in the relative wage ratio. (3) Baker’s analysis of the impact of the size of the immigrant population on the wages of U.S.-born Mexican women estimates a smaller effect in 1990 than in 1980.

A more plausible explanation for the erosion of relative wages for Hispanic women is a change in productivity characteristics. Mexican-American and Puerto Rican women have increased their educational attainment; however, the gap in college-completion rates continues to increase. Comparable statistics for Mexican-American women having earned a bachelor’s degree or higher, compared with White women, are as follows: in 1970, 2.3 percent compared with 13.8 percent; in 1980, 4.9 percent compared with 21.6 percent; in 1990, 6.2 percent compared with 27.9 percent. The gap increased 3 percent between 1980 and 1990 and 8 percent between 1970 and 1990.¹⁶ Given the growing premium paid for college diplomas, it is likely this gap in educational attainment contributed to the decrease in relative earnings of Mexican-American women.

In theory, a woman’s labor-force participation decision is influenced by (1) the wage she can expect to receive; (2) the productivity of her

nonmarket time, as reflected by number of young children, marital status, and living arrangements; and (3) her nonlabor income, including the earnings of her spouse and government transfers. Empirical findings are generally consistent with this theoretical model. The presence of young children and a high-wage earning spouse tend to reduce LFPR among married women; however, factors associated with higher wages, such as education and good health, tend to increase LFPR. Government transfer programs, such as Aid to Families with Dependent Children, tended to reduce formal participation among potentially eligible women.

These factors, however, do not provide a full accounting of racial differences in LFPR among married women (Reimers, 1985b; Goldin, 1977). Using data from the 1976 SIE, Reimers estimates that differences in characteristics explain 95 percent of the gap between U.S.-born White and foreign-born Hispanic married women, and 83 percent of the gap between U.S.-born White and U.S.-born Hispanic married women. Differences in education, family size, language ability, and age of children contribute to lower LFPR in these groups. Differences in characteristics explain 77 percent of the gap between U.S.-born White married women and foreign-born Asians.

Differences in characteristics do not appear to explain historical gaps in LFPR between U.S.-born White, Asian, and Black married women; this gap can be explained by differences in parameters. In other words, given equal education, family size, language ability, and ages of children, the propensity to work is higher for Black and Asian than for White married women. Reimers suggests that cultural factors contribute to this otherwise unexplained differential.

A factor contributing to the slow increase in LFPR of Black women during the 1960s and 1970s was the decrease in LFPR among young single Black women (Jones, 1985a). Jones attributes the gap among young women to differences in the demand for their labor services; she asserts that the relative decrease in Black female LFPR “appears to be a response to inadequate employment opportunities with reasonable wages” (1985a:27). Jones’ hypothesis is consistent with the increased demand for skill since the late 1970s. Indeed, in some occupations that opened up to Black women in the 1960s and early 1970s, total employment contracted in the 1990s. Bank tellers provide a case in point; technological change has reduced employment in this occupation just as the number of minority women with these jobs expanded.

It is possible that the factors contributing to the decrease in employment rates among young Black males also contribute to the decrease in labor-force participation for young Black women—lack of skills, geographic concentration in inner cities, low-income communities, and discrimination (Holzer, 1999). Although studies of employers suggest more

positive attitudes toward Black women than Black men, young Black women still must overcome skill deficits and the problem of spatial mismatch (Kirschenman and Neckerman, 1991; Holzer, 1996). McLafferty and Preston (1992), in a study of Northern New Jersey, found that Black and Hispanic women have longer commuting times and less localized labor markets than White women. Thus geographic concentration of work outside of central cities may pose an obstacle to employment for young women; childcare needs and concerns about personal safety may discourage them from traveling long distances late at night or early in the morning.

The decrease in the relative LFPR of young Black women may also be linked to their higher rates of teen motherhood, leading younger Black women to have lower relative rates of labor-force participation when they are young and higher relative rates in their late 20s and early 30s. Hispanic women probably confront similar obstacles to employment.

The occupational distributions of Black and White women converged between 1960 and 1980, but this trend is unlikely to continue. Initially, because of their more extensive historical participation in the labor force and the barriers to employment they confronted, Black women may have benefited more from the enforcement of EEO laws than White women. Indeed, Epstein (1973) argued that Black women enjoyed an advantage relative to both Black men and White women because they represented a “double-minority” —employers concerned about meeting affirmative action goals and timetables could count a Black woman as two minorities. In addition to this “advantage,” the growth in demand for clerical workers and service workers in the 1960s and 1970s created opportunities for Black, White, and Hispanic women on a somewhat equal basis.

Labor in the 1980s, however, favored college-educated workers, and professional and technical workers (Bound and Johnson, 1995; Juhn and Murphy, 1995). White and Asian women, with higher rates of college completion, were able to take advantage of emerging new opportunities; most Black and Hispanic women were not. Yet, even college-educated Black women were increasingly concentrated in low-wage occupations. Occupational comparisons for Black and White college graduates in 1989 are as follows: 29.5 percent of Blacks were clerical workers, compared with 14.5 percent of Whites; 47 percent of Blacks were managerial and professional workers, compared with 63.2 percent of Whites. Bound and Dresser’s (1999) study suggests that 18 percent of the increase in the wage

gap between Black and White college graduates can be attributed to increasing concentration of Black college graduates in low-wage occupations. Table 6–5 provides percentages of occupational distributions of White, Black, Hispanic, and Asian and Pacific Islander women in March 1997.

Future research needs to explore the factors that restrict occupational mobility for Black women and Hispanics. This analysis should investigate decision making by individuals as well as institutional constraints; and the labor-market experiences of Asian groups should not be neglected. Analyses of the labor-market status of Asian women are limited in number; and researchers tend to lump Asians into a single group despite the diversity of characteristics among different national groups.

The trends in occupations suggest that wage inequality is likely to continue to increase, as it has since 1990. The only factors that might offset this trend are a shift in the relative demand for skilled labor or an increase in the college-completion rates of Hispanic and Black women. If the demand for college-educated workers decreases relative to that for other workers, the relative wages of college-educated labor may decrease.

In addition, as noted earlier, Black and Hispanic women have higher rates of unemployment and lower rates of labor-force participation than

White and Asian women. These differences in employment rates will contribute to a relative deficit in experience. Black and Hispanic women also continue to have lower rates of college enrollment, which will further limit the demand for their labor. Finally, welfare reform has increased the supply of low-skilled labor, and minority women appear to be having a more difficult time moving off the welfare rolls (DeParle, 1998).

These observations suggest several questions for future research. (1) What are the obstacles to college completion for Hispanic and Black women? McElroy (1996) and Cardoza (1991) suggest that one obstacle is the higher incidence of teen motherhood among Hispanics and Black women; but, it is not clear whether early childbearing deters college attendance or vice versa. Incentives for early childbearing are poorly understood and require further investigation. (2) How do the employment obstacles confronting young women differ from those confronting young men? For example, neighborhood segregation is likely to impact women differently from men, but few studies have examined the spatial mismatch hypothesis for young women. (3) How do the employment prospects of young men affect decision making by young women? Although several authors have suggested a link between male employment status and family formation, none has explored the potential link between employment prospects of young men and schooling and employment choices of young women. (4) To what extent does discrimination by employers limit opportunities? Racial differences in employment rates and in occupational distribution might reflect differences in tastes or in unobserved productivity characteristics. On the other hand, the persistence of occupational and employment disparities may reflect institutional factors, including discrimination in the employment process. One method of assessing the role of discrimination is the employment audit in which matched pairs of testers (one minority, one White) respond to advertised job openings to develop comparisons of treatment at each stage of the application process. Although there have been a number of employment audits involving Black and Hispanic men, there have been no employment audits involving minority women.

Are we moving toward equality of opportunity and access? The answer is uncertain. Once women are in the same jobs, they appear to earn the same pay. However, there persist racial differences not only in the distribution of women by occupations but also in employment status; and the widening gap in college-completion rates and the deterioration in the relative returns to schooling among younger women threaten future progress.

Special thanks to Kenneth Chay, William Spriggs, William Darity, Margaret Hwang, and Rebeccca Blank for their advice as this manuscript was in progress. It has also benefited from the suggestions of an anonymous referee.

Albelda, R. 1986 Occupational segregation by race and gender, 1958–1981. Industrial and Labor Relations Review 39:404–411.

Altonji, J., and R.Blank 1999 Race and gender in the labor market. In Handbook of Labor Economics, Volume 3C, O.Ashenfelter and D.Card, eds. New York: North Holland.

Anderson, D., and D.Shapiro 1996 Racial differences in access to high-paying jobs and the wage gap between Black and White women. Industrial and Labor Relations Review 49(2):273–286.

Baker, S. 1999 Mexican-origin women in southwestern labor markets. In Latinas and African American Women at Work: Race, Gender and Economic Inequality, I.Brown, ed. New York: Russell Sage Foundation.

Betsey, C. 1994 Litigation of employment discrimination under Title VII: The case of African American women. American Economic Review 84(2):98–107.

Blau, F., and A.Beller 1992 Black-White earnings over the 1970s and 1980s: Gender differences in trends. The Review of Economics and Statistics 74(2):276–286.

Bound, J., and L.Dresser 1999 Losing ground: The erosion of the relative earnings of African American women during the 1980s. In Latinas and African American Women at Work: Race, Gender and Economic Inequality, I.Brown, ed. New York: Russell Sage Foundation.

Bound, J., and G.Johnson 1995 What are the causes of rising wage inequality in the United States. Federal Reserve Bank of New York Economic Policy Review 1(1):9–17.

Brown, L, and I.Kennelly 1999 Stereotypes and realities: Images of Black women in the labor market. In Latinas and African American Women at Work: Race, Gender and Economic Inequality, I. Brown, ed. New York: Russell Sage Foundation.

Burbridge, L. 1994 The reliance of African-American women on government and third-sector employment. American Economic Review 84(2):103–107.

1997 Black women in the history of African American economic thought: A critical essay. In A Different Vision: African American Economic Thought, Volume One, T. Boston, ed. New York: Routledge.

Cardoza, D. 1991 College attendance and persistence among Hispanic women: An examination of some contributing factors. Sex Roles A Journal of Research 24:133–147.

Carlson, L., and C.Swartz 1988 The earnings of women and ethnic minorities. Industrial and Labor Relations Review 41:530–552.

Corcoran, M., C.Heflin, and B.Reyes 1999 The economic progress of Mexican and Puerto Rican women. In Latinas and African American Women at Work: Race, Gender and Economic Inequality, I.Brown, ed. New York: Russell Sage Foundation.

Cunningham, J., and N.Zalokar 1992 The economic progress of Black women, 1940–1980: Occupational distribution and relative wages. Industrial and Labor Relations Review 45:540–555.

Darity, W., D.Guilkey, and W.Winfrey 1996 Explaining differences in economic performance among racial and ethnic groups in the USA: The data examined. American Journal of Economics and Sociology 55(4):411–426.

DeParle, J. 1998 Shrinking welfare rolls leave record high share of minorities. The New York Times July 27:A1+.

England, P., K.Christopher, and L.Reid 1999 Gender, race, ethnicity, and wages. In Latinas and African American Women at Work: Race, Gender and Economic Inequality, I.Brown, ed. New York: Russell Sage Foundation.

Epstein, C. 1973 Positive effects of the multiple negative: Explaining the success of Black professional women. American Journal of Sociology 78(4):912–935.

Farkas, G., and K.Vicknair 1996 Appropriate tests of racial discrimination require controls for cognitive skill: Comment on Cancio, Evans, and Maume. American Sociological Review 61 (August):557– 560.

Fosu, A. 1992 Occupational mobility of Black women, 1958–1981: The impact of post-1964 antidiscrimination measures. Industrial and Labor Relations Review 45(2):281–294.

Goldin, C. 1977 Female labor force participation: The origin of Black-White difference, 1870 and 1880. Journal of Economic History 37(1):87–108.

1990 Understanding the Gender Gap: An Economic History of American Women. New York: Oxford University Press.

Greene, W. 1997 Econometric Analysis. Upper Saddle River, N.J.: Prentice Hall.

Heckman, J. 1976 The common structure of statistical models of truncation, sample selection, and limited dependent variables and a simple estimator for such models. Annals of Economic and Social Measurement 5:479–492.

Heckman, J., and B.Payner 1989 Determining the impact of federal antidiscrimination policy on the economic status of Blacks: A study of South Carolina. American Economic Review 79(1):138–177.

Holzer, H. 1996 What Employers Want: Job Prospects for Less-Educated Workers. New York: Russell Sage Foundation.

1999 The barriers to higher employment rates among African Americans. In Job Creation Prospects and Strategies, W.A.Leigh and M.C.Simms, eds. Washington, D.C.: Joint Center for Political and Economic Studies.

Jones, B. 1985a Black women and labor force participation: An analysis of sluggish growth rates. Slipping Through the Cracks: Review of Black Political Economy 14(special issue):11– 31.

Jones, J. 1985b Labor of Love, Labor of Sorrow: Black Women, Work, and the Family from Slavery to the Present. New York: Basic Books.

Juhn, C., and K.Murphy 1995 Inequality in labor market outcomes. Federal Reserve Bank of New York Economic Policy Review 1(1):26–34.

Kirschenman, J., and K.Neckerman 1991 ‘We’d love to hire them, but…’: The meaning of race for employers. In The Urban Underclass, C.Jencks and P.Peterson, eds. Washington, D.C.: The Brookings Institution.

Leonard, J. 1990 The impact of affirmative action regulation and equal employment law on Black employment. The Journal of Economic Perspectives 4(4):47–64.

Manski, C. 1989 Anatomy of the selection problem. Journal of Human Resources 24:343–360.

Maume, D., A.Cancio, and T.Evans 1996 Cognitive skills and racial wage inequality: Reply to Farkas and Vicknair. American Sociological Review 61 (August):561–564.

McCrate, E., and L.Leete 1994 Black-White wage differences among young women, 1977–86. Industrial Relations 33:168–183.

McElroy, S. 1996 Early childbearing, high school completion, and college enrollment: Evidence from 1980 high school sophomores. Economics of Education Review 15:303–324.

McLafferty, S., and V.Preston 1992 Spatial mismatch and labor market segmentation for African American and Latina women. Economic Geography 68(4):406–431.

Neal, D., and W.Johnson 1996 The role of premarket factors in Black-White wage differences. Journal of Political Economy 104:869–895.

Power, M., and S.Rosenberg 1993 Black female clerical workers: Movement toward equality with White women? Industrial Relations 32(2):223–237.

Reimers, C. 1985a A comparative analysis of the wages of Hispanics, Blacks, and non-Hispanic Whites. In Hispanics in the U.S. Economy, G.Borjas and M.Tienda, eds. New York: Academic Press.

1985b Cultural differences in the labor supply among married women. American Economic Review 75:252–255.

Rodgers, W., III, and W.Spriggs 1996 What does the AFQT really measure: Race, wages, schooling and the AFQT score. The Review of Black Political Economy 24(4):13–46.

U.S. Bureau of the Census 1963a U.S. Census of the Population: 1960. Subject Reports. NonWhite Population by Race, Final Report PC (2)-1C. Washington, D.C.: U.S. Government Printing Office.

1963b U.S. Census of the Population: 1960. Subject Reports. Puerto Ricans in the United States. Final Report PC(2)-1D. Washington, D.C.: U.S. Government Printing Office.

1963c U.S. Census of the Population: 1960. Subject Reports. Persons of Spanish Surname. Final Report PC(2)-1B. Washington, D.C.: U.S. Government Printing Office.

1963d U.S. Census of the Population: 1960. Detailed Characteristics: United States Summary. Final Report PC(1)-1D. Washington, D.C.: U.S. Government Printing Office.

1973a U.S. Census of the Population: 1970. Subject Reports. Japanese, Chinese and Filipinos in the United States. Final Report PC(2)-1G. Washington, D.C.: U.S. Government Printing Office.

1973b U.S. Census of the Population: 1970. Detailed characteristics: United States Summary. Final Report PC(2)-1G. Washington, D.C.: U.S. Government Printing Office.

Weaver, R. 1946 Negro Labor: A National Problem. New York: Harcourt, Brace and Company.

Zalokar, N. 1990 The Economic Status of Black Women: An Exploratory Investigation. Washington, D.C.: U.S. Commission on Civil Rights.