Can the UNC System’s Return-On-Investment Findings Be Trusted?

Design errors may be limiting an otherwise useful study’s utility.

A report on the financial returns on a UNC education, commissioned two years ago, was finally released in Nov. 2023. The study was done by the private consulting firm Deloitte, working with the RPK Group and the Burning Glass Institute, and included results from 765 undergraduate and 599 graduate programs on the 16 campuses of the UNC System, offered from 2015 to 2020. The data permit three types of comparisons: earning differentials for graduates across institutions, earning differentials across disciplines, and earning differentials for those studying and then staying to work in the state of North Carolina.

While the results are controversial and somewhat difficult to understand fully, critical information can be gleaned concerning several important educational issues. These conclusions are summarized below, with an indication of the limitations and ambiguities that seem to be attached to the study.

Return on Educational Investment

Of primary interest to users of this report are the comparisons of lifetime earnings among students in multiple educational categories: (1) those who receive an undergraduate degree compared to those without an undergraduate degree and (2) those who receive a graduate degree compared to those who receive only an undergraduate degree.

The summary results of the study, including all programs and campuses, are the following:

The UNC study contains major unanswered questions about the discounting of future earnings gains.Lifetime earnings for those with a UNC-System bachelor’s degree are $1.2 million compared to $628,000 for those without a degree. This earnings differential of $572,000 must be reduced by the cost of the degree itself to provide a net return on the college-education investment.

Using a “calculated student investment” of approximately $78,000, the study found a net return of $494,000. Ninety-four percent of the system’s undergraduate programs had a positive differential for graduates.

For the graduate-degree holder, lifetime earnings are $2.1 million, or slightly less than $1 million more than the holder of only a bachelor’s degree can hope to receive ($2.1 million minus $1.2 million). Given an additional cost estimate of $70,000 for a graduate degree, the study found an earnings differential near $900,000 for graduate-degree holders. Ninety-one percent of the system’s graduate programs had a positive differential for those receiving the graduate degree.

The above earnings, according to Deloitte, were calculated using 2021 dollars in order to standardize the price level and remove any gain for later graduates (from a higher price level) or loss for earlier ones (from a lower price level).

While these two summary conclusions seem impressive and attractive at a glance, the UNC study contains major unanswered questions about the process of discounting future earnings gains and the percentage rate of return. These concepts are illustrated in a separate study on returns on a college education recently performed by three economists at the Federal Reserve Bank of St. Louis.

According to the St. Louis study, higher education continues to offer attractive returns to its customers. Indeed, “the rates of return on a college education in 2020 were: 24.1% for white men, 22.7% for white women, 14.2% for Black men, 13.5% for Black women, 35.9% for Asian men and 31.1% for Asian women.”

The St. Louis authors, however, seem to have adjusted their earnings increments in two ways that the Deloitte-UNC study failed to consider. First, they introduced the concept of present value, a necessary adjustment if one wishes to calculate a true investment return. Because future earnings have a lower current or present value than do earnings received now, one can’t simply treat tomorrow’s money as the equivalent of today’s. This is highly relevant to return-on-investment studies, since much of the “return” in question happens after the age of 40, perhaps 20 years after graduation. For instance, at an interest rate of 5 percent and an average rate of inflation, a $30,000 gain payable in 20 years has a present value of only $11,308 (or $30,000 divided by 1.05 to the 20th power). Thus, comparisons between the earnings of students with degrees and those without degrees will be overstated in the absence of discounting.

Without relative comparisons, the desirability of UNC graduates’ dollar returns is highly ambiguous.It is unclear, in the Deloitte-UNC study, which calculations used discounting appropriately and which did not. (A Deloitte representative with whom I spoke confirmed the ambiguity.) If future earnings were not discounted, then there is an overstatement of the earnings gain.

Second, unlike the St. Louis study, which provided rates of return on college completion (e.g., 24.1 percent for white men), the Deloitte-UNC study provides simple dollar values for earnings gains for graduates. Without relative comparisons, the desirability of these dollar returns is highly ambiguous. The St. Louis study concluded, persuasively, that the rates of return on college education were considerably higher than the rates of return on investment in financial assets. No such claims seem possible in the Deloitte-UNC study. For instance, is the rate of return on a UNC education less than, equal to, or greater than the return on an investment in the S&P 500?

Other Considerations of Earnings Bias

For various other reasons, there can be both overstatements and understatements of the actual net earnings gain from a degree. A major overstatement of the dollar benefit can result from an understatement of the cost of providing education. In the Deloitte analysis, a college-costs figure of approximately $78,000 per year was used, a figure that represents the approximate direct annual outlay per undergraduate. This $78,000 represents a mean or average figure for a course of study.

Yet the student does not pay the full cost of his education. A significant portion of the individual cost is offset by private, federal, and state grants that lead to specific scholarships for students and general cost reductions (like library grants) that benefit all students. The true net earnings gain per student should be reduced by the inclusion of these cost reductions.

Now, for other reasons, the Deloitte earnings gain is understated. First, college graduates historically have had lower rates of unemployment than have non-graduates. Thus, the risk-adjusted earnings that the college graduate receives merit an upward boost from the reduced likelihood of being unemployed. Second, college graduates tend to have greater mobility, making a graduate of, say, UNC-Chapel Hill more likely to move to another state if wages in her area of study are higher elsewhere than in North Carolina. The earnings of these higher-salary, migrating employees are not covered in the Deloitte study, which captures only those individuals who are working in North Carolina, thus reducing the average gain in earnings enjoyed by UNC-System graduates. Third, some difficult-to-measure gains, like scheduling flexibility and higher fringe benefits, are not captured in the Deloitte-UNC study.

The earnings of higher-salary, migrating employees are not covered in the Deloitte study.Additionally, there are social benefits of higher education that accrue to all members of society. These include improved public health and reduced crime, both of which are positively correlated with higher ed. Usually, societal studies of the gains from education include some estimate of these benefits.

Thus, the earnings differentials reported in the Deloitte-UNC study are compromised, pushed up by the absence of discounting and pushed down by other factors detailed above. The result of these distortions is that one has reduced confidence in the study’s measure of the added value of a UNC-System education and even less confidence in the estimated value-added of specific programs, whether undergraduate or graduate.

Variance in Returns Across Disciplines

There is growing demand in North Carolina for workers with higher education and technical skills. At the same time, there has been some slowdown in enrollment in higher ed. The positive earnings results presented in the Deloitte report should encourage the youth of the state to enroll in college.

These results, however, vary significantly across programs and disciplines, even as they tend to mirror the earnings distribution among professions in the United States. For instance, as NC Newsline’s Joe Killian reports, “The highest return [for undergraduate degrees] was in science, technology, engineering, and mathematics (STEM) fields. […] Degrees in Biotechnology were found to be the most lucrative, with a median lifetime return of more than $3.2 million.”

With such variance across programs, a legitimate fear could arise that investment in academic programs will be guided by these results. This can be shortsighted for two reasons. First, as indicated above, the actual returns on a college degree are subject to error and inadequate methodology. To make decisions with such data can lead to poor judgments. Second, there are often interrelationships among programs. As an example, a program in Classical Greek at the undergraduate level may be needed to support a strong graduate Classics program. Assume that the earnings increment from the undergraduate Greek program is low, while the earnings increment from the graduate Classics program is high. Eliminating the undergraduate Greek program places the strong graduate Classics program in jeopardy. Thus, short-sighted decisionmaking could result from simply looking at the incremental earnings of a particular program.

Great care must be used if North Carolinians wish to fine-tune their educational resource allocation using these data. Very poor results could occur from simplistic reviews of incremental earnings.

Very poor results could occur from simplistic reviews of incremental earnings.Economic Mobility

Finally, the Deloitte report shows that the UNC System contributes significantly to economic mobility. For nearly 90 percent of students from low-income families (defined as an income of $17,900 or less), there was upward movement of at least one “income band” upon the completion of an undergraduate degree. The report estimates that “forty-two percent of all low-income students rose four income bands[,] leading to a yearly income of $91,300 or greater after 20 years.” A reasonable conclusion is that the UNC System, by encouraging access to its programs, is helping economically disadvantaged students have a significantly better lifestyle. Even with the data ambiguities discussed above, the positive equity effects of the state’s educational system are important and prevalent.


After two years of work, the Deloitte working group has provided a “comprehensive” study of the economic returns on education in the UNC System. The immediate question for policymakers is whether the limitations of the study can be addressed and corrected to make it more useful.

Clark Ross has taught economics at Davidson College since 1979. He served as dean of the faculty for nearly 15 years.