Excellent Research Confirms Flaw in Earnings Premium

One of the leading efforts comes via the Postsecondary Value Commission, which in 2021 developed three core metrics quantifying the economic return on investment (ROI) for every college in the United States: Minimum Economic Return, Earnings Premium, and Economic Mobility. In this report, we explore how sensitive these three metrics are to different assumptions. Specifically, what happens when we recalculate these metrics using local as opposed to statewide earnings data?

Most colleges operate in regional markets defined not by state or national boundaries but by much smaller geographic areas.

Geography also plays a major role in shaping earnings, cost of living, and other economic outcomes in the U.S. For example, rural communities often have lower earnings than urban or suburban areas in their own states. North Carolina’s Research Triangle, home to three top tier research institutions and a robust science and technology sector, is a prime example. This region fares better in most economic measures compared to the rural Appalachian regions of the state. Likewise, the Seattle metro area, headquarters to several of the world’s largest and most profitable technology companies, has higher median incomes, higher cost of living, and lower unemployment rates compared to the state’s rural regions.

Commuting zones are somewhere in the middle—they are smaller than states, yet larger than a single city—and they are especially relevant in the context of higher education and across the social sciences where researchers commonly use commuting zones to study how employment, earnings, health, upward mobility, poverty, and education vary by where one lives. Commuting zones are statistically-derived measures of local labor markets based on the U.S. Census Bureau’s journey-to-work data.

In total, 754 unduplicated institutions in 47 states are positively affected by using local earnings when calculating the Postsecondary Value Commission’s ROI metrics. These 754 institutions would have failed to pass a given ROI metric had we only used state-level earnings. But when we use local earnings, which account for differences in cost-of-living and capture local contexts, these 754 institutions now pass. This represents about 16% of all colleges and universities in the U.S., suggesting one seemingly small tweak—using local rather than state-level earnings—can impact a nontrivial share of the nation’s colleges and universities.

Using local earnings in these calculations not only captures important statistical variation, it also has a greater effect on institutions that are most deeply tied to their local regions and economies. For instance, we find institutions benefiting most from local measures tend to be public, have broad-access missions, and serve disproportionately high shares of Pell grant recipients. Approximately one in five public institutions would pass the Postsecondary Value Commission’s ROI metrics had local earnings been used in the calculation. But using statelevel earnings makes these institutions fail. In addition to the impacts on the public sector of higher education, using local earnings also tends to benefit institutions located in rural places and places with lower-incomes, higher child poverty rates, and in many cases higher shares of people of color. More research is needed to fully understand these patterns, but we are finding evidence that geographic variation in earnings can disproportionately affect certain institutions and communities, meaning it should be accounted for when calculating economic ROI metrics.

Using local-level (rather than state-level) earnings data when reporting EP and D/E rates in the FVT framework would provide a more accurate and contextualized statistic for current or future students. For instance, if the goal is to provide students with actionable information, then a student attending a rural college who plans to live in a rural place after college might find it irrelevant to compare earnings against the statewide median. Instead, they may want to see how a program’s earnings compare to programs at other rural-located institutions or other nearby places. The same argument can be made for urban and suburban locations, or those crossing state lines. Adjusting measures for geographic contexts provides more useful information at relatively low cost and could be incorporated into the FVT framework during future negotiations and updates.

While the Postsecondary Value Commission’s metrics are not exactly the same as OBBBA’s, the Earnings Premium metric is most similar to this new law. As a result, we anticipate many public community colleges, institutions serving lower-income students, and located in lower-income communities will be disproportionately affected by comparing earnings to state-level medians rather than to local earnings. The U.S. Department of Education is undergoing negotiated rulemaking in the winter of 2025 and early 2026, where they will develop an appeals process for programs failing the new test. Perhaps programs failing the state-level test can use local-level earnings upon appeal; those passing the test with local earnings could be reinstated or otherwise not penalized.