Why the Underemployment Number Depends on the Question
Underemployment among US workers with bachelor’s degrees is the subject here: are graduates using their education in the occupations where they work? The immediate focus is why serious reports can produce sharply different answers to that question. The useful outcome is not choosing the most reassuring or alarming percentage. It is learning to ask what a number classifies, which workers it includes, and what evidence it uses before drawing a conclusion about graduates or their jobs.
The same graduate can be counted in different ways
Underemployment sounds like a simple condition. A worker has more education or skill than their job requires, so the education is being underused. In the Georgetown Center on Education and the Workforce report Rethinking Underemployment: Are College Graduates Using Their Degrees?, the underlying idea is the underutilization of human capital: a mismatch between a worker’s educational attainment and the education requirements of the job. The report focuses on people with bachelor’s degrees, although the broader problem can occur at other education levels.
That definition is meaningful, but it does not yet tell us how to observe a mismatch. A person’s degree is visible in a survey. A job can be assigned to an occupation. What is much harder to observe at national scale is the actual work performed: the responsibilities a particular employer gives a particular employee, the level of judgment involved, the skills used each day, and whether the role is a temporary landing place or a durable career match.
This is also not the same question as unemployment. Unemployment asks whether someone who wants work currently has a job. Underemployment in this discussion asks whether the education a working person has is being used or rewarded in the job they hold. A worker can be employed and still feel that their education is not being used. Conversely, a worker can hold a job whose occupational label suggests a lower entry credential while performing specialized work and receiving a substantial return for a bachelor’s degree.
Other labor-market measures answer still different questions. For example, the BLS U-6 series includes people marginally attached to the labor force and people working part time for economic reasons. Those measures matter, but they are outside this education-and-occupation analysis. Keeping the questions separate is important: otherwise a discussion of graduates’ job fit starts absorbing every form of labor-market difficulty.
The stakes are not merely verbal. If graduates’ education is genuinely underused, individuals may lose earnings and the economy may misallocate human capital. If a broad occupation-level rule calls productive, well-matched work “underemployment,” the resulting alarm can obscure where the real mismatch is concentrated. Measurement therefore sits between the public claim and the policy response. Before asking how large the problem is, we need to know what the estimate is treating as evidence.
Why the headline percentages disagree
The report brings the problem into view by showing how much the answer moves when the measuring rule or the population changes. Using a BLS-based approach, it classifies a bachelor’s-degree worker as underemployed when they work in an occupation that BLS assigns an entry requirement below a bachelor’s degree. Under that rule, 47 percent of workers ages 22–23 with bachelor’s degrees are counted as underemployed when all workers are included, and 43 percent are counted when the analysis is limited to full-time, full-year workers. Among workers ages 25–54, the corresponding figures are 39 percent and 37 percent.
Those four figures are not four measurements of exactly the same thing. The younger group is meant to approximate people within a year of graduation; the older group represents prime-age workers. “All workers” includes people employed part time or not for the full year, while the full-time, full-year group removes some effects of irregular or seasonal employment. The report emphasizes the latter comparison to avoid treating part-time or seasonal work as if it were automatically evidence that education and job requirements do not align.
Change the classifier and the estimate changes again. Across the approaches it examines, the report describes recent or prime-age bachelor’s-degree underemployment estimates ranging from about 24 percent to 52 percent. The spread does not mean that researchers are simply making up numbers. It means that the numbers embody different judgments about what counts as a degree-level job. One method asks what education is typically needed to enter an occupation. Another asks what education the people currently working in that occupation actually have. A third adds the earnings associated with a degree within the occupation.
The denominator matters as much as the numerator. A percentage for recent graduates cannot be compared casually with a percentage for all prime-age graduates. A figure that includes part-time workers is not answering the same question as one limited to full-time, full-year workers. A result based on a particular data period also reflects the labor market of that period, including hiring conditions and the pace at which graduates move into better-matched roles.
The number is therefore not a free-floating property of “college graduates.” It is the output of a chain: a population is selected, occupations are coded, an educational comparison is chosen, and a threshold determines who falls on either side. If any link changes, the estimate can move while the underlying workers have not.
Three questions hiding inside one label
The first question is about the typical route into an occupation. The Bureau of Labor Statistics assigns occupations a “typical education needed for entry,” ranging from no formal educational credential to a doctoral or professional degree. The assignment is a useful starting point for comparing occupations. But “typical” is doing important work. The category describes the education most workers need to enter the occupation; it does not say that every role inside it has the same requirements, that every current worker has only that level of education, or that the entry-level category captures what becomes appropriate later in a career.
The second question is about the workers who are actually there. An occupation may contain people with high school diplomas, associate’s degrees, bachelor’s degrees, and graduate degrees because it includes different responsibilities, employers, regions, or career stages. The report finds that 27 percent of occupations classified as “high school level” employ more workers with a bachelor’s degree than workers with a high school diploma. Just over one in five workers in occupations assigned a high school diploma or less for entry hold a bachelor’s degree or higher. Those findings do not prove that every graduate in such an occupation is well matched. They do show why the occupation cannot be treated as an educationally uniform bucket.
This observed distribution leads to the realized-matches lens. It asks whether bachelor’s-degree holders make up a majority or a plurality of workers in an occupation. The answer can recategorize an occupation that BLS labels below the bachelor’s level. This is closer to a record of realized hiring, but it carries its own interpretation: a high share of graduates may reflect employer demand, changing skill needs, limited graduate supply, credential inflation, or a weak labor market that pushes degree holders into jobs that do not truly require their education.
The third question is whether the degree appears to have economic value within the occupation. Workers with bachelor’s degrees in BLS high-school-level occupations have median earnings of $62,000, compared with $45,000 for workers with a high school diploma, a 38 percent premium in the report’s full-time, full-year prime-age sample. That difference suggests that some graduates may be bringing valued capabilities to jobs that the entry label alone would call below degree level. It is evidence worth considering, not a direct measurement of tasks or a guarantee that the work is satisfying.
These lenses keep the primary subject in view while changing the evidence used to understand it. The subject is not “BLS categories,” “worker distributions,” or “wages” in isolation. It is the relationship between graduates and the occupations they enter. The categories and statistics are instruments for examining that relationship, each with a different blind spot.
An occupation-level estimate also cannot settle every individual case. Broad occupational groupings do not capture every job title or every difference in seniority and specialization. Requirements can vary by sector and region. The education a current worker has is not necessarily the education their employer required, and a wage premium may reflect experience, geography, role composition, bargaining conditions, or other factors as well as education-linked productivity. The available evidence can show patterns across groups; it cannot directly observe whether a particular graduate uses college-level skills every day, enjoys the role, or will remain in it.
That boundary should make the statistics more useful, not less. A population estimate can identify where the labor market deserves attention. It can reveal that a transition is difficult for many recent graduates, that some occupations contain much more educational diversity than their entry labels suggest, or that degree holders receive different returns within the same broad occupation. It cannot turn those patterns into a fixed judgment about one person.
The central question for the rest of this series is consequently precise: when a bachelor’s-degree holder works in an occupation assigned below the bachelor’s level, which part of the relationship are we measuring? The entry path, the education of workers already hired, or the value reflected in their earnings? Once that question is made explicit, disagreement among percentages becomes something to investigate rather than a reason to choose a headline by instinct.