The debate over affirmative action in higher education, largely dormant since 2003, when the Supreme Court affirmed racial preferences at the University of Michigan School of Law, may soon come roaring back. A new challenge to racial preferences at the University of Texas at Austin is currently before the Fifth Circuit Court of Appeals, and could land on the Supreme Court’s docket late next year. Supporters of affirmative action are anxious, given that the swing vote on the Roberts Court, Justice Anthony Kennedy, joined a majority in striking down racial school integration plans in 2007 and dissented in the Michigan case. And President Obama, who has said that his own daughters do not deserve affirmative action preferences, cannot be counted as an unambiguous supporter.
As the debate heats up again, both supporters and opponents of affirmative action are likely to find ammunition in Thomas J. Espenshade’s and Alexandria Walton Radford’s book. A scholarly analysis of new data from eight highly selective public and private universities, it examines how affirmative action actually works today, how well beneficiaries perform academically, and, importantly for the upcoming court case, how well alternatives to affirmative action—such as class-based preferences—might work in creating racially diverse campuses. Although the book is generally well done, it seriously underestimates the potential for class-based affirmative action to produce racial diversity. Well-designed plans that recognize broad differences between black and white poverty will likely become the new face of affirmative action in the coming years.
The authors provide a fascinating peek inside the admissions process at several unnamed universities, which are roughly representative of the top fifty in the U.S. News & World Report’s rankings. In the class entering in 1997, the authors find, racial preferences were far more significant than the proverbial “tiebreaker” at the selective private universities studied. Espenshade and Radford report that blacks received an admissions boost worth 310 SAT points compared with whites. Hispanics received a 130 point preference, and Asians received a 140 point penalty compared with whites. Taken together, on average, an Asian American student must score a whopping 450 points higher on the combined math and verbal sections of the SAT to have the same chance of being admitted as an African American applicant.
By contrast, Espenshade and Radford find that socioeconomically disadvantaged students receive a much smaller preference. Students identifying themselves as “lower class” received a 130 point boost over middle class students, who in turn had a 30 point advantage over upper class students. And while selective private universities favor low-income minorities over wealthy minorities, controlling for academic ability, low income whites are three times less likely than upper middle class whites to be admitted. This may be true, they say, because universities want to “save their scarce financial aid dollars for students who will help them look good on their numbers of minority students.”
Not surprisingly, given these policies, certain schools that have made strides on racial diversity have not achieved much socioeconomic diversity. In 2004, for example, as the University of Michigan was celebrating its victory in the Supreme Court, this national symbol of racial diversity had more students from families making in excess of $200,000 per year than families earning less than the national median of $53,000 a year. Moreover, Espenshade and Radford show that socio-economic stratification at the institutions studied is getting worse, not better, and it is occurring particularly among racial minorities, as wealthier black and Hispanic students replace lower income minorities from earlier eras.
Given the lower credentials of under-represented minorities entering selective universities, it is not surprising that the authors find that black and Hispanic students perform less well academically than Asians and whites. More disturbing is their confirmation of earlier studies that showed minority students on campus “underperforming” what their entering credentials would predict. In the selective universities studied, the median class rank for black students was at the 20th percentile, compared with the 32nd percentile for Hispanics, 52nd percentile for Asian Americans, and 57th percentile for whites. The very good news, however, is that Espenshade and Radford find that minority students are more likely to graduate from more selective schools, dispelling the idea that “overmatched” minority students are hurt by affirmative action at the undergraduate level.
Finally, Espenshade and Radford discuss what would happen if universities abandoned race-based affirmative action and instituted instead a system providing a leg up to socioeconomically disadvantaged students of all races or one providing automatic admissions to the top 10 percent of high school graduating classes. The analysis is important: it goes to the heart of the ongoing challenge to affirmative action at the University of Texas. Moreover, the issue of whether race-neutral alternatives might produce sufficient racial diversity was one that Justice Kennedy pressed in his dissent in the Michigan case.
Unfortunately, Espenshade and Radford present an uncharacteristically superficial analysis of this complex issue. Citing only two sources of data, they flatly conclude that “class-based affirmative action is not a satisfactory substitute for race-based affirmative action.” First, they note that racial diversity has declined at public universities in California, Florida, and Washington, all of which have banned affirmative action and sought to institute alternatives. Second, their simulations using data from highly selective universities find that eliminating the preference for blacks and Hispanics would result in sharp declines in representation, and that class-based affirmative action would only partially offset these decreases. If students were admitted solely by SAT scores at the private universities studied, blacks and Hispanics would have a combined representation of 3.4 percent. Providing preferences to lower and working class applicants would raise the combined representation to 10.2 percent. But this result would represent a substantial decline from the current combined representation of 16.2 percent. (At the public universities studied, the combined black and Hispanic representation would decline from 9.3 percent to 6.2 percent.) The big beneficiaries would be lower-class and working-class whites and Asians.
But the reality is far more complicated than the authors suggest. Elite public schools in places such as California did see a decline in the proportion of black and Latino students after affirmative action was barred, but this does not tell us what would happen if a national ban were put in place. Public universities in California have been at a stark competitive disadvantage in attracting black and Hispanic students because the nation’s private universities (in California and other states), as well as public universities in most states, could continue to provide racial preferences, and so many talented minority students stopped applying to places like Berkeley and UCLA. (In 1997, after California’s ban on affirmative action was instituted, black applications fell at UCLA by 13.3 percent and at UC Berkeley by 7 percent.)
Moreover, some universities have been successful in maintaining racial diversity despite a ban on using race in admissions. At UT Austin, which was forbidden by a lower court from considering race between 1997 and 2003, a class-based affirmative action plan, combined with a plan to automatically admit students in the top 10 percent of their high school class, resulted in a net increase in racial diversity. In 1996, prior to the ban on race, African Americans and Latinos made up 18.8 percent of the freshman class. In 2004, under the socioeconomic and top 10 percent plan, the combined representation had risen to 21.4 percent. (A white applicant named Abigail Fisher sued in 2008 after UT Austin decided to add race back into its admissions system following the Supreme Court’s green light to do so in the Michigan case.)
Likewise, Espenshade and Radford’s simulation of class-based affirmative action at selective private universities is deeply flawed. For one thing, the preference simulated is given to students who self-identify as “lower class” or “working class,” blunt categories that paper over important differences between the daily challenges of poor white students and poor black students. In large measure because of slavery and segregation, for example, black median household wealth is less than 10 percent of white wealth, even though the income gap is much smaller; black household income stands at 62 percent of white income. And owing to racial discrimination in the housing market, low income blacks are much more likely to live in concentrated poverty than poor whites.
Incorporating wealth and neighborhood poverty into a class-based affirmative action program is the right thing to do on the merits—because growing up with little net worth and in a disadvantaged neighborhood certainly presents extra obstacles for a student—and it will also have the effect of boosting the racial dividend of class-based affirmative action. Both of these factors are well known to social scientists. It is odd that Espenshade and Radford do nothing to incorporate these considerations into their simulation.
Likewise, Espenshade and Radford apply the socioeconomic preference to a narrow band of the applicant pool—the bottom 10.6 percent—and would boost their representation to just 16.9 percent at private universities and 6.4 percent at public institutions. By contrast, another study by Anthony Carnevale and Stephen Rose concluded that it would be possible to increase the representation of the bottom economic half from 10 percent currently to 38 percent at the nation’s most selective 146 institutions without decreasing graduation rates at all.
The shift from race-based preferences to class-based preferences will likely occur in the near future, as a conservative Supreme Court and a liberal president converge on the idea that affirmative action should extend to economically disadvantaged students of all races rather than increasingly affluent minority students. Espenshade and Radford are right that blunt and anemic class-based preferences will not preserve racial diversity, but smartly thought out programs that acknowledge broad differences between blacks and whites of similar income should be far more successful. The situation is not hopeless, even if it has for too long been misunderstood.