Among the many concerning consequences of Donald Trump’s election is the very real threat that the American public education system, which has educated generations of school children, will be privatized.

Donald Trump campaigned on a program to employ $20 billion in federal funds to promote school choice, including private school vouchers. Mike Pence has a track record on the issue, having pushed for privatization of public education in Indiana. And Trump’s nominee for Education Secretary, Betsy DeVos, has long championed private school vouchers throughout the country.

One of many causes for concern about private school vouchers—alongside questions of academic effectiveness, access, and funding—is the effect that such policies have on socioeconomic and racial segregation.

School integration, by socioeconomic status and race, provides a powerful lever for promoting social mobility and social cohesion. Low-income students who attend socioeconomically-mixed schools are as much as two years ahead of low-income students who attend high poverty schools on the National Assessment of Educational Progress in fourth grade mathematics. And a growing body of research suggests that all students benefit from attending racially diverse schools and that “diversity makes us smarter.”

Trump’s proposal to devote federal funds to private school vouchers could turn back the clock on racial and socioeconomic integration in our nation’s schools. Here are four reasons for concern:

1. Many of the private schools in the United States that would be eligible to receive public money through a voucher program serve a disproportionate percentage of white and wealthy students.

As of 2012, 72 percent of students in private schools were white, compared to 52 percent of public school students. In forty-three out of fifty states, white students in private schools are more likely than those in public schools to attend schools with at least 90 percent white enrollment. In South Carolina, where the gap between public and private enrollment in over 90 percent white schools is greatest, 63 percent of white students in private schools attend such schools, compared to just 5 percent of white students in public schools. But it’s not just in the south where we see large disparities. In New Jersey, 45 percent of white private school students are in schools with at least 90 percent white enrollment, compared to just 15 percent of white public school students.

These patterns are in part rooted in a history of deliberate segregation. During the desegregation of public schools in the South from the 1950s–1970s, many white Christian families started private schools catering to white families and often refusing to admit nonwhite students. Some of these “segregation academies” exist today, and although overt racial discrimination is not allowed, many remain segregated in practice. Because private schools can place conditions on admissions or enrollment, several independent academies are either unable to or refuse to accommodate students with IEP plans or disabilities, place limitations on student expression that disproportionately punish LGBTQ students and students of color, have parental commitment requirements that often disadvantage low-income families, or bar pregnant or parenting teens from enrolling. State administration of voucher programs might fail to guarantee access and inclusion of diverse and disadvantaged students in private schools that receive public funds.

Data on private school enrollment also support a pattern of “white flight” and “wealthy flight” from the public sphere. A 2002 study found “the strongest predictor of white private enrollment is the proportion of black students in the area.” In 2012, private school parents were almost twice as likely to come from relatively wealthy families as public school parents. According to the U.S. Census Bureau, 60 percent of private school parents made more than $75,000 a year (the Bureau’s highest income bracket), compared to 32 percent of public school parents.

2. In theory, means-tested voucher programs might provide an opportunity for low-income and minority students to integrate private schools, but in practice, private school voucher programs are unlikely to benefit the most disadvantaged students.

Even when school choice programs target disadvantaged students, low-income families may be limited from participating when scholarship funds do not meet full tuition costs or transportation is not provided. Moreover, it has long been established that choice programs, because they require access to knowledge about opportunities, tend to benefit the most advantaged subset of students eligible.

There is also good reason to believe that federally funded means-tested vouchers programs would face political pressure to morph into a universal voucher program. Leading voucher advocate Terry Moe, for example, argues that to get a foot in the door, advocates “need to pursue programs targeted at disadvantaged children (at least in the short run).” Once voucher programs gain a foothold and appear less risky to the public, stage two is to universalize them. Says Moe: “Most Americans believe that vouchers (if adopted) should eventually be made available to virtually all children.”

Finally, not all private schools are created equal, with many consistently underperforming academically when compared to public schools. Researchers point out that, when controlling for demographic differences, student achievement on mathematics testing is equal to or lower than achievement on that same testing in public schools. Instruction and teacher quality is not necessarily better in private schools, although their students are more likely to have out-of-classroom experiences that translate into better academic outcomes.

3. The great majority of private schools in the United States are religiously affiliated, raising concerns that private school vouchers would result in sorting students by religious background.

Nearly 80 percent of private school students attend religious schools. Voucher programs, therefore, often create an unavoidable dilemma. As Richard Just has noted, “The inane logic of vouchers would leave us with a stark choice: either become a country that pays religious institutions to proselytize to children of other faiths, or become a country that educates children of different religions separately.”

4. When other countries have tried school vouchers, segregation has increased.

Research on large-scale private school vouchers and similar school choice programs in Sweden, Chile, and The Netherlands suggest that in each case those programs have increased socioeconomic and racial segregation in schools. A 2013 study on Chile found that school segregation by socioeconomic status (SES) increased in Chile as market-based reforms grew, and private schools, including those receiving vouchers, were more segregated than public schools for both low-SES students and high-SES students. A 2015 study on Sweden found that the use of private school vouchers was associated with increased segregation of immigrant versus native Swedish students. And in The Netherlands, a marketplace approach to education has contributed to high and growing levels of segregation of students by educational disadvantage.

As the nation grows increasingly fragmented by race, religion, class, and gender, the undermining of public education—one of the few remaining sources of social cohesion—should raise alarm bells for liberals and conservatives alike.