The passage of Proposition 209, the California Civil Rights Initiative (CCRI), has signaled to many the beginning of the end for affirmative action [see Peter Schrag's "When Preferences Disappear"]. Evidence from California shows, however, that while the gap between white and minority educational achievement has narrowed, the gap between white and minority wages has continued to increase. This evidence strongly suggests that, contrary to the claims of many CCRI supporters, California's labor markets have not outgrown the need for interventions to correct bias, intentional or otherwise.
Even conservative economists acknowledge that affirmative action was responsible in the late 1960s and 1970s for raising blacks' wages and bringing them more in line with the qualifications of black workers. But whether minority (especially black) wages are still less than what their qualifications warrant is a more difficult question than it appears. Educational attainment (measured by years of schooling completed) and achievement (measured by test scores) remain higher among whites than among minority workers. White workers are also paid more. The key question is whether differences in qualifications fully justify the wage gaps. If educational admissions and labor markets now work "perfectly" and reward Californians by merit, as some opponents of affirmative action claim, there might be widespread agreement that affirmative action should have no place. And if unequal qualifications entirely explain wage inequality, it might make more sense to emphasize improving minority job applicants' educational qualifications. But if, on the other hand, improved education among minority youths does not narrow the wage gap, affirmative action or some other market interventions may be reasonable.
The trend data are dramatic and troubling. In the last two decades, the educational qualifications gap between white and minority workers has narrowed sharply, as black and Latino attainment and achievement have improved relative to that of whites. But just as dramatically the wage gap has grown. The overall wage gap has not only widened from what social scientists refer to as a "fallacy of composition"—that is, minorities have not disproportionately increased their attainment at levels (like high school diplomas) where relative wages have fallen. The wage gap has widened even for workers with the same education levels: California's minority high school graduates now earn less, compared to white high school graduates, than they used to. The same is true of minority workers who have attended some college and minority workers who are college graduates.
The mere existence of a widening wage gap, even in the face of a shrinking educational gap, does not by itself prove discrimination. Measurable education is not the only relevant qualification workers bring to employment. Some affirmative action opponents insist, for example, that even if attainment and achievement qualifications are relatively more equal, minority job applicants have inferior discipline and work habits and that this difference justifies rational distinctions in wage outcomes. Others claim that a spatial mismatch between minority applicants and the location of available jobs creates a relative oversupply of applicants for jobs minorities can reach, leading to average wages that are lower for minority workers than for white workers of similar qualifications. Still others claim that blacks and Latinos seek jobs in occupations or sectors (say, government rather than entrepreneurship) with lower average earnings.
There are good reasons to believe, however, that these explanations are incomplete. For example, while blacks may be overrepresented in California's government executive positions where wages are lower than in comparable private jobs, blacks are also overrepresented in government blue-collar jobs that pay more than comparable private employment. Similarly, the spatial mismatch hypothesis has more power to explain widening wage gaps for high school-educated workers than for college-educated workers, who are more mobile and not trapped in urban ghettos. None of this shows that affirmative action is the best solution. Other policies such as "empowerment zones" or further minimum-wage increases may be better. But the data support a strong case for some intervention.
THE EDUCATIONAL ATTAINMENT OF YOUNG CALIFORNIANS
Blacks have been closing the educational attainment gap at almost every level (the only exception being that while black females have increased their college graduation rate, white females have increased their college graduation rate even more). Between 1980 and 1995, while the high school graduation rate for California's male white workers, aged 25 to 34, increased from 94 percent to 96 percent, the graduation rate of black male workers increased from 90 percent to 95 percent. Between 1980 and 1995, the share of young white workers who had attended some college (including community colleges or technical institutes) rose only slightly, from 68 percent to 70 percent; the share of black workers with some college education increased from 57 percent to 63 percent. And while the college graduation rate of young white males in the workforce barely changed, going from 34 percent in 1980 to 35 percent in 1995, the college graduation rate for comparable blacks increased from 18 percent to 20 percent, though that still left a large absolute difference.
The educational attainment gap between black and white females also narrowed at the "high school completed" and "some college" levels. Indeed, by 1995, the high school graduation rate of employed black females (99 percent) was higher than for comparable whites. Likewise, more young female black than white workers had attended college (74 percent versus 73 percent), reversing their 1980 disadvantage (56 percent versus 62 percent).
Interpreting similar data for Latinos is difficult because many data sources do not distinguish native-born from immigrant Latinos. We do, however, have census data for native-born Latino workers in 1980 and 1990 showing that for males, the attainment gap has stayed about the same. For young Latina workers, high school graduation rates jumped from 82 percent in 1980 to 85 percent in 1990, while white female rates were unchanged. The share of young Latina workers who had attended some college jumped dramatically, from 39 percent in 1980 to 55 percent in 1990; the share of white female workers with some college also jumped, but less (from 62 percent to 75 percent), so the gap narrowed here as well. The share of young Latina workers with college degrees increased from 11 percent in 1980 to 13 percent in 1990. Although this increase occurred at a faster rate than for white females (whose college graduation rate went from 29 percent to 33 percent), the college graduate gap between white female and Latina workers remained relatively stable.
Because educational attainment is an important consideration in employers' wage decisions for young workers, we would expect that young white males, on average, would earn more than young black males in 1995—but not as much more as they did in 1980. We would also expect, based on this educational attainment data, that the wages of young black female workers would be approaching the wages of young white females; that the gap between young white female and Latina workers would have significantly narrowed; and that the wage gap between Latino and white males would remain basically unchanged. In fact, however, wage disparities between whites and minorities (both Latino and black) have grown at all levels of educational attainment.
THE EDUCATIONAL ACHIEVEMENT OF YOUNG CALIFORNIANS
Perhaps the problem is the quality of schooling. Even though young minority workers today have attended more school than in the past, they may not be able to hold jobs because their academic achievement did not improve commensurate with their level of schooling. The evidence from test scores shows, however, that an overall narrowing of the attainment gap has been matched by a corresponding narrowing of the achievement gap.
Scholastic Aptitude Test (SAT) scores can be misinterpreted if used to describe the overall quality of schooling because SAT test takers are self-selected. Declining average SAT scores may simply mean that a larger and less elite group of students chose to take the test. But SAT trends can suggest how the academic preparation of those who went to college compares from year to year. If a growing proportion of a group such as black 17-year-olds take the SAT test, the average black score might be expected to decline because the average would reflect a larger (and thus less elite) group of black students. If the average score rose despite an expansion of the pool, the increase would be all the more impressive.
Between 1976 (the year SAT scores were first reported by race and ethnicity) and 1987 (the most recent year in which the 25-year-old college graduates whose wages we report would have taken the SAT for college admission), the average verbal score was stable (456 to 453), as was the average math score (494 to 499), for California's white students. For California's black students, however, average verbal scores went from 331 to 359, while average math scores went from 354 to 388. The number of black test takers rose from 5,800 in 1976 to 7,100 in 1987, while the number of all black 17-year-olds in California remained roughly the same. Thus, the increase in average scores for black test takers almost certainly represented genuine improvement. Similarly, average total SAT scores for Mexican-American students rose from 772 to 793 during a time when significant expansion of the Latino test-taking pool also made an increase in average scores much more difficult to achieve.
Most minority students who did not go on to a four-year college did not take the SAT test. They did, however, take the National Assessment of Educational Progress (NAEP), which is given to a representative sample of all 17-year-olds. NAEP scores represent a range of reading abilities. For example, 200 reflects an ability to make inferences based on uncomplicated passages; 250 reflects an ability to reach generalizations from literature, science, and social studies passages; 300 reflects an ability to summarize and explain relatively complicated information.
Nationally, average NAEP reading scores of white 17-year-olds went from 291 in 1971 to 295 in 1988. For blacks, the gain was greater, from 239 in 1971 to 274 in 1988. For Latinos, the gain was from an average of 252 in 1975 to 271 in 1988. Thus, while average white scores remained higher in 1988, the average scores of all three groups in 1988 reflected the ability to make inferences and reach generalizations from passages dealing with literature, science, and social studies.
In math, white NAEP scores were unchanged nationally, but black and Latino scores increased substantially. In 1978, the average black mathematics score was 37 percentile points lower than the average white score. By 1990, the gap had been reduced to 21 points. On the verbal test, the gap was reduced from 41 percentile points to 24. All told, the gap between white and black NAEP scores was reduced by about 40 percent from the 1970s to 1990. The white-Latino gap was reduced from 31 percentile points to 27 points on the math test, and from about 32 percentile points to about 17 points on the verbal test.
Separate NAEP data are not available for race and ethnic score trends in California, but we do know that western regional trends were similar to the nation's. It is likely, therefore, that there was relative improvement in achievement for California's minority 17-year-olds who did not go to college.
THE RELATIVE WAGES OF YOUNG CALIFORNIANS
Once the education gap between minority and white workers narrows, the wage gap should also narrow. But the gap in wages between young white male workers and young black male workers has actually widened. In California, young black male wages fell from 84 percent of young white male wages in 1980 to 77 percent in 1995. Black females earned about as much as white females in 1980 but only 86 percent as much in 1995.
Because minority test scores have improved relative to whites among high-school educated workers and those with "some college," the wage gaps within these groups should have narrowed. But in each case minority wages fell in California as a share of white wages. For black males with high school education only, wages fell from 82 percent of wages of similarly educated white workers, to 79 percent. For those with "some college," the drop was from 88 percent to 83 percent. And for college graduates, the drop was from 94 percent to 86 percent.
For black females, relative wages also dropped. Whereas in 1979, young working black females earned 6 percent more than young white females with similar educational attainment, by 1989 their relative wages had fallen to 4 percent below the white level. Young black females with "some college" earned 4 percent more in 1979 but 6 percent less in 1989; and those with four-year college degrees earned 6 percent more than comparable whites in 1979, but 8 percent less in 1989. Young Latino workers also lost ground to white workers at each of the comparable education levels, from 1979 to 1989.
Sample data from the Current Population Survey, although not strictly comparable to census data, extends the picture closer to the present. It shows that while the wage gap did stabilize for some education groups between 1990 and 1995, it continued to widen for others. Nowhere did it narrow, despite continued narrowing of the educational attainment and achievement gaps. For example, California's young black males with only high school education saw relative wages fall from 84 percent of comparably educated young whites' wages in 1990 to 74 percent in 1995. Relative wages also fell for black females with "some college" and for Latina females at all levels of educational attainment.
While no single statistic proves the case, the broad trends are remarkably consistent: Educational attainment and achievement of minority 17-year-olds, relative to comparable whites, improved steadily—and in some cases dramatically—from the mid-1970s to the late 1980s, yet wages did not. There may be other—or better—solutions than affirmative action, but affirmative action is one way to push wages in the direction of being more consistently and rationally related to workers' qualifications. Unfortunately, the vote in California was not designed to substitute a better remedy, but no remedy at all.