The $140 Billion Public-College Myth

The public's confidence in higher education is falling rapidly. The share of the public that has confidence in higher education declined from 57% in 2015 to 38% in 2026. As colleges and their advocates try to win back the public's confidence, one of their strongest arguments is that college is worth it for both students and for the country broadly.

While it is true that a college education generally pays off for the typical student, there are a few problems with this view. The biggest, and most obvious, problem to note is that there are many predictable cases in which college is not worth it from a financial perspective. A famous example is Harvard's certificate program in theater, which failed an Obama-era debt-to-earnings measure because it left graduates with excessive debt ($78,000) relative to earnings ($36,000).

Perhaps more important, however, is that the value of a college education is often overstated. A recent analysis from the Institute for Higher Education Policy (IHEP), which sought to determine the economic contribution of four-year public colleges, provides a representative example of this mistake.

The IHEP formula estimates the net benefit of a college education as the higher wages of college graduates after accounting for the net cost of college paid by students. The headline value is determined by the annual earnings of college graduates ($62,300) minus the earnings of high school graduates ($35,700) and the net cost of college ($13,000). It then multiplies the resulting figure by 10.3 million college graduates.

The resulting figure is $140 billion. Therefore, the analysis concludes, the implication is that public four-year colleges provide a net benefit of $140 billion to the economy.

While IHEP's general approach to calculating the economic benefit of college education is sound, its analysis systematically overstates benefits, understates costs, and makes several other unorthodox choices that render its estimate unreliable. There are, by my count, at least three big problems with the analysis.

First, the analysis estimates college costs as the student's cost of attendance minus grant aid but ignores two huge categories of costs. One is the costs paid by taxpayers instead of students. Across all types of colleges, the federal government provides around $4,000 in grants, work-study, tax benefits, and loan subsidies per student. Meanwhile, state and local governments provide around $12,000 per student at public colleges. The IHEP analysis should have taken both figures and adjusted them to separate out the spending for bachelor's degree students at four-year public colleges. Instead, it simply ignored them.

The second type of costs missing from the IHEP analysis are the student's opportunity costs. If a student didn't go to college, they could have been working and earning wages. These forgone wages should be counted as an (implicit) cost of college. Ignoring both these college costs means that the IHEP analysis substantially underestimates the costs of college.

Second, the benefits of college attendance are overstated. The key benefit in the analysis is the higher wages that college graduates earn relative to high school graduates. The IHEP analysis finds this value by taking the median wage of college graduates and subtracting the median wage of high school graduates. While this difference might be fine as a rough proxy for the potential financial benefits of going to college, it isn't clear that colleges cause these benefits.

Students are not randomly assigned to college or no college, and indeed college graduates tend to be more intelligent, more diligent, and more conformist than nongraduates. All of those tendencies are valued by employers, so the typical college graduate would earn more than the typical non-graduate even if they never attended college. The part of the earnings premium due to these underlying differences in the college-educated population is often called the signaling effect. There is considerable debate over how much of the difference in wages is due to signaling: some argue that signaling's share is small (30%) and some contend it is on the larger side (80%). The IHEP method takes the rather extreme position of assuming that signaling accounts for no part of the earnings premium.

Finally, college dropouts are ignored. The IHEP analysis splits the population into two categories: college graduates, who earn more but pay some costs to attend college, and non-college graduates, who don't get the higher wages from attending college but also don't pay the costs of attending college. But there is a third category, college dropouts, who don't get the (full) benefit from attending college but do pay (some of) the costs. The costs paid by these students (and taxpayers) should be included in the cost of college tab, as should any benefits (those with some college tend to earn more than those that never attended, though again, part of this is likely due to signaling). Ignoring college dropouts understates the costs of college.

A few smaller issues compound these problems: aggregate figures should multiply average, not median values, and the methodology document indicates that college costs are added rather than subtracted. These are minor issues, but they add up. Between overstated benefits and understated costs, the $140 billion figure isn't a real estimate of college's benefit to the economy. But it is, unfortunately, a number that will be repeated often in the debate over higher education.