In the early 20th century, regardless of your family’s income, you could be fairly confident that enrolling in college would provide a solid bump to your wages a few years down the line. And back then, the returns were the same for both rich and poor students.

That’s no longer the case. Wealthy college-goers now earn a premium nearly three times larger than the one earned by low-income students, according to economists Zachary Bleemer and Sarah Quincy, who collated more than a century of government records and surveys to measure the returns of education over time. For low-income students, those returns are half as large as they once were.

In other words, higher education has become regressive, widening class divisions by delivering far greater returns to wealthy students than to their low-income peers.

The primary reason for the shift: a decades-long policy failure that funneled poor students away from four-year research universities and into two-year community colleges and for-profit institutions.

Community colleges exploded in popularity in the 1960s and ’70s. Seeing them as an inexpensive way to increase access to education, both Republican and Democratic administrations poured resources into creating two-year institutions that would — they hoped — get poor people to college who otherwise wouldn’t enroll.

“Too many people have fallen prey to the myth that a four-year liberal arts diploma is essential to a full and rewarding life,” President Richard M. Nixon said in a 1970 address to Congress. “In fact, other forms of postsecondary education — such as a two-year community college or technical training course — are far better suited to the interests of many young people.”

The goal sounded noble. But rather than bringing more students into higher education, the rise of community colleges created a second-tier alternative — one that ended up drawing low-income students away from four-year public universities without meaningfully increasing overall enrollment.

Between 1980 and 2000, the shares of rich students enrolled in community colleges and for-profit institutions increased by 3 and 5 percentage points, respectively. But the share of poor students increased by much more: 12 percentage points at community colleges and 11 percentage points at for-profit universities.

Not all colleges are created equal. Each additional year of schooling adds to the college-going premium: transfer rates are low, which means that students at community colleges simply do less school. Just 16 percent of students obtained a bachelor’s degree six years after enrolling in community college.

Plus, quality is eroding at the colleges and universities that enroll more students from the bottom third of the income distribution, largely because those institutions have suffered more from cutbacks in government funding. Student-to-faculty ratios have ballooned, graduation rates have plummeted and dropout rates for low-income students have increased.

Of course, community college is a good fit for some. Two-year programs are affordable, and they offer flexibility for students balancing work or family. But tuition levels don’t explain why poor students have shifted away from four-year universities — community college has always been about half as expensive as four-year institutions, on average.

Between 1960 and 2010, the value of attending community college halved relative to public research universities. For-profit institutions are worse: They confer a wage premium nearly 25 percentage points lower than public research universities.

In recent times, the two-year sector has shrunk considerably, thanks in part to a wave of regulations in the 2010s that curbed the explosion of for-profit institutions. Nevertheless, the main problem persists: Poor students continue to enroll in lower-value institutions than their rich peers.

Institutional quality isn’t the whole story. Within the same university, poor students are now much more likely than rich students to choose majors with lower rates of return. In the 20th century, rich and poor students were equally likely to get humanities degrees. Today, a majority of humanities students are poor.

Here, too, educators bear the brunt of the blame: Many universities gatekeep high-value majors. For example, engineering schools often have separate applications, and they require higher test scores and grade point averages. Some “STEM” programs also expect students to arrive having completed prerequisite coursework — the kind more readily available to wealthy students in well-resourced high schools.

For decades, we’ve built a system that sets up poor students to fall behind. It won’t be easy to reverse this trend, but here are a couple of ways to start.

First, we need to stop treating two-year colleges as an equal alternative to four-year degrees. Instead, large well-resourced research and private universities should expand access to low-income students. One option: offer low-income applicants the same admissions preference that many elite colleges extend to legacy students. Four-year universities could also grow in-person enrollment and create inexpensive online offerings to lure more students from the two-year sector.

Universities can also better communicate to students the value of each major. Studies show that students overestimate earnings for humanities and business majors while underestimating them for engineering. When given accurate wage data, students become more likely to choose majors with more wage potential.

It’s still better to enroll in college than not. But higher education should work well for everyone — not just the students who come from wealthy families.