At a time when the president seeks to exacerbate racial and class cleavages on a regular basis, providing educational settings where students of different racial, ethnic, economic, and religious groups can come together to learn is paramount. But there are fair and unfair ways to achieve that diversity in higher education, and which path universities take matters. Choosing the wrong way is one factor in making the election of demagogues who feed on resentment more, not less, likely.
Four years ago, in my personal capacity, I signed on as an expert witness in a lawsuit challenging Harvard University’s use of racial preferences in its admissions process. Another expert in the case documented the considerable anti-Asian bias at Harvard, which has already garnered widespread media attention. My task, on the other hand, has been to explain how using alternatives—such as socio-economic preferences—could create a diverse student body without relying on racial preferences. (Court papers were just made public.) Socio-economic preferences can open the doors to impressive young students—including underrepresented minorities and first-generation students—who have overcome considerable obstacles in life and are now largely shut out of selective colleges. In addition, the Supreme Court has long held that if workable strategies can create the educational benefits of diversity, they should be employed, as there are costs—such as reinforcing negative stereotypes—to classifying individuals by race.
What I’ve found is that Harvard starts with a system heavily biased toward the wealthy, then adds in racial preferences to compensate. To begin with, the university preferences students in the know, heaping advantage on already privileged and disproportionately white groups. It admits those who are on a “Dean’s Interest” list and other preferred candidates through a backdoor channel known as the Z-list, in which just 2 percent in recent years were black and 1.2 percent economically disadvantaged. It gives a leg up to those who apply through the early—rather than the regular—admissions process, even though Harvard acknowledged in 2006 that such a policy “operates at cross purposes with our goal of finding and admitting the most talented students from across the economic spectrum.” (Harvard reinstated early admissions in 2011.) Harvard also preferences the children of faculty and alumni. According to Harvard’s own internal analysis, being a legacy boosts a candidate’s chances of admission by 40 percentage points, compared to a 9 percentage-point increase for low-income students.
Harvard’s preference for legacy students shuts the door to others trying to gain a foothold. For the classes of 2007 to 2016, Harvard had more legacy students than it did first-generation college students. Although 68 percent of American adults 25 and older lack a four-year college degree, an average of only 10 percent of Harvard students in the classes of 2007 to 2016 were first-generation college students. If black students, who represent 15 percent of the newest Harvard class, were as underrepresented in the Harvard population as first-generation college students have been over the years, they would account for just 2.25 percent of the student body.
Other leading colleges try to foster racial and socio-economic diversity by recruiting from the nation’s brightest community college students. Amherst College, for example, enrolls between 12 and 15 community college transfers annually. If Harvard, with a much larger student body, adopted a similarly scaled program, it would admit 44 to 55 community college transfer students annually. In fact, over a six-year period, Harvard admitted only two such students.
The results of this process are highly predictable. According to a 2017 study by Stanford University’s Raj Chetty, in recent years, Harvard has had as many students come from the top 1 percent of the income distribution as the bottom 60 percent. More come from the top 10 percent by income than the bottom 90 percent. Twenty-three times as many come from the top income quintile as the bottom quintile.
Because Harvard’s system is so heavily biased against nonwealthy students, who are disproportionately black and Hispanic, Harvard puts itself in the unnecessary position of having to employ significant racial preferences in order to achieve its diversity goals. According to Harvard’s own analysis, for example, the size of the preference provided solely to blacks is more than twice the preference provided to students of any race from families making less than $60,000 a year. Even here, though, Harvard tilts its racial preferences toward the well-off: 71 percent of underrepresented minorities admitted to Harvard are socio-economically advantaged, even though less than 32 percent of blacks and Hispanics nationally would qualify as advantaged by Harvard’s system.
There is a much better way to create diversity. Instead of adding racial preferences to counteract the effects of an admission process titled toward the rich, Harvard could create a system that meaningfully recognizes that students of many different backgrounds overcome tremendous hurdles in building their high school academic, extracurricular, and athletic records. Because race and income are deeply intertwined in American society, black and Hispanic students will still disproportionately benefit from such a program. In 2016, black median family income was 57.8 percent of white median family income, and Hispanic income was 62.9 percent of white income. The disparities in neighborhood opportunities are even greater: While 6 percent of white youth live in neighborhoods with more than 20 percent poverty rates, 66 percent of black youth live in such neighborhoods. As Sean Reardon of Stanford University has found, although the black/white achievement gap used to be twice as large as the rich/poor achievement gap, today the reverse is true: The economic achievement gap is twice as large as the racial gap.
How would this approach work at Harvard? Because my colleagues and I had access to data from more than 160,000 students who applied for admission over six cycles, we could employ Harvard’s holistic system of rating students by academic, extracurricular, athletic, and personal criteria to carefully model the results of socio-economic affirmative action programs. In one of the simulations, we turned off the effects of Harvard’s preferences for race, legacy status, early admission, faculty children, and economic disadvantage. In its place, we instituted a larger socio-economic preference that is about half the magnitude of Harvard’s existing preference for athletes. In this simulation the overall share of underrepresented minorities increased from 28 percent to 30 percent. The share of first-generation students increased from 7 percent to 25 percent. Academic preparation of the admitted students in the simulation remained superb, with the average SAT in the class declining just 1 percentage point (from the 99th percentile to the 98th) and high school GPA remaining exactly the same.
More diversity can be achieved with less divisiveness. Racial preference policies likely help explain why 60 percent of white working-class Americans say that “discrimination against whites has become as big a problem as discrimination against blacks.” Socio-economic preferences in college admissions, by contrast, would benefit working-class whites alongside working-class black, Latino, and Asian students. In an era of racial strife and antagonism, class-based affirmative action could reduce racial division while opening up gated communities in higher education to promising students from all walks of life.