The Baseline: What BLS Says an Occupation Requires to Begin
For workers with bachelor’s degrees, the first step in measuring underemployment is often a Bureau of Labor Statistics label: what education is typically needed to enter the occupation? This chapter examines how that label is constructed and how it becomes a baseline estimate of mismatch. The practical goal is precision: readers should be able to use the BLS category as evidence about an occupation’s usual entry path without mistaking it for a complete description of every graduate’s job.
“Typical” is the key word in the BLS system
The BLS system does not begin by asking how much education the average worker in an occupation has. It asks what education most workers typically need to enter that occupation. That distinction is easy to miss because the answer appears as a neat category attached to a broad occupational group. The category is an entry point, not a portrait of everyone who works there.
BLS organizes employment projections around three related kinds of information. The first is typical education needed for entry. Its eight levels run from no formal educational credential through a high school diploma or equivalent, some college without a degree, a postsecondary nondegree award, an associate’s degree, a bachelor’s degree, a master’s degree, and a doctoral or professional degree. The second is related work experience, which can be none, less than five years, or five years or more. The third is typical on-the-job training, ranging from none or short-term training to apprenticeships, internships or residencies, and training that lasts more than a year.
These categories describe a typical path to competency and entry. They do not claim that a person with less education can never enter an occupation, that everyone with more education is misallocated there, or that the work stops changing after the entry point. They also do not say that every job title inside a broad occupation has the same level of responsibility. A senior specialist and a newly hired worker may share an occupational code while doing meaningfully different work.
The word “typical” also matters when more than one route exists. BLS assigns the path it judges to be most common rather than creating a separate category for every route. The BLS example of registered nurses illustrates the trade-off. The typical entry path is recorded as a bachelor’s degree, with no related work experience or on-the-job training requirement, while the Occupational Outlook Handbook notes that an associate’s degree in nursing or a diploma from an approved nursing program can also lead into the occupation. A single category makes the national system usable, but it compresses real variation.
The system has a historical boundary as well. BLS first used the current classification for the 2010–20 projections, replacing an older system used from 1995 through 2008. BLS warns that the two systems are not comparable. A percentage calculated by combining estimates from the old and current systems would therefore look more continuous than the underlying definitions allow.
How an expert assignment becomes a baseline count
BLS assignments are informed judgments, not guesses and not simple counts of degrees. Economists combine quantitative information from several sources, including educational-attainment data from the Census Bureau’s American Community Survey, education and training information from O*NET, and postsecondary completion data from the National Center for Education Statistics. They also evaluate qualitative information from educators, employers, workers in the occupation, training experts, professional and trade associations, and unions.
That combination is necessary because no single dataset directly records the full entry process. Survey data can show what education workers have. O*NET can provide information about work and training requirements. Completion data can show the supply of people emerging from education programs. Employers and training experts can clarify licensing, customary preparation, and how a role is actually staffed. BLS economists use those inputs to choose the typical entry category for each occupation.
This design gives the system several strengths. It offers a common vocabulary across hundreds of occupations. It can distinguish a role that usually begins with a bachelor’s degree from one that commonly begins with an associate’s degree or less. It is also explicit enough that researchers can apply the same rule to a defined population instead of making an individual-by-individual judgment based on impressions.
But the same design creates the baseline’s main limitation. Expert assignment and worker data answer different questions. An expert may identify the usual education needed to enter an occupation, while a survey may show that many current workers have more education than that. Neither observation alone tells us whether the additional education is being used in a more senior role, rewarded by an employer, obtained for personal reasons, or left underused.
The classification can also lag the labor market. Employer expectations may change as technology, regulation, or organizational practice changes. Requirements may differ across industries and regions. A national category can remain useful for comparison while failing to capture a newer local hiring norm or a specialized position. BLS itself notes that entry assignments may not match the education distribution of current workers because of underemployment, individual choice, upskilling, and changing entry requirements.
The right interpretation is therefore neither “BLS knows the exact education requirement of every job” nor “BLS categories are arbitrary.” The system is a structured estimate of the typical route into an occupation. Its value lies in making that route visible and comparable. Its weakness is that a typical route is not the same thing as the full set of roles and matches contained in the occupation.
The rule behind the familiar headline
Once an occupation has a BLS entry-level education assignment, the BLS-based approach applies a simple comparison. If a worker has a bachelor’s degree but works in an occupation BLS assigns an entry requirement below a bachelor’s degree, the worker is counted as underemployed under that approach. In the Georgetown report’s grouping, “below bachelor’s” includes occupations assigned no formal educational credential, a high school diploma or equivalent, some college without a degree, a postsecondary nondegree award, or an associate’s degree.
The rule is easy to reproduce. Start with a worker’s reported education. Find the occupational category attached to their job. Compare the worker’s bachelor’s degree with the category’s typical entry level. If the category is lower, count a possible mismatch; if it is a bachelor’s level or higher, do not count the worker as underemployed under this particular rule.
The simplicity is exactly why the approach is widely used. A reader can understand the classifier without knowing how a person’s tasks are scored. A researcher can apply it to a large survey. A policymaker can compare occupations using one national standard. It provides a first signal about where graduates may be working below the education level commonly associated with their occupation.
The rule also assigns the same status to many different situations. Consider three bachelor’s-degree workers in one occupation classified below the bachelor’s level. The first may have accepted an entry job temporarily while searching for a better match. The second may hold a senior or specialized role that the broad occupation includes but the entry assignment does not describe. The third may be in a position where a degree is not needed and where the worker’s education is genuinely underused. The BLS-based count places all three on the same side of the line because it has no person-level information about their tasks or progression.
That does not make the first signal meaningless. It makes the signal conditional. The count says that the worker’s educational attainment is above the occupation’s typical entry assignment. It does not, by itself, say that the worker lacks responsibility, that the employer does not value their degree, or that the job is a poor long-term choice.
There is another important distinction between entering and remaining in an occupation. BLS’s category is designed around the beginning of the path. A worker may acquire experience, move into a more demanding role, or use advanced capabilities after being hired. Conversely, an occupation may raise its expectations over time without the category immediately capturing every change. A measure built on entry requirements is therefore particularly informative about access to an occupation, less informative about the full career lifecycle inside it.
The data linkage adds technical choices that affect the result. In the Georgetown analysis, ACS 2018–22 five-year data are connected to BLS educational requirements through a crosswalk. Some ACS occupational codes map to multiple SOC occupations with different requirements, so the analysis uses 2023 BLS employment shares to weight those possibilities and select the most relevant requirement. The sample uses workers with positive earnings who are not in the military and not enrolled in school or college; workers with graduate degrees are excluded from the bachelor’s-degree analysis. These choices make the estimate more coherent, but they also show why the resulting percentage belongs to a specified dataset and period rather than to “graduates” in the abstract.
The report emphasizes full-time, full-year workers when comparing underemployment rates because part-time and seasonal work can affect annual earnings for reasons unrelated to education/job fit. That choice does not erase irregular employment; it isolates one question from another. A count among all workers and a count among full-time, full-year workers can both be valid if their populations are named clearly.
What the baseline can and cannot establish
The BLS-based approach is best understood as a screening lens. It can identify a population whose educational attainment exceeds the typical entry assignment of their occupation. It can reveal differences between recent graduates and prime-age workers when the same rule is applied to each group. It can also provide a common starting point for comparing the results with other methods.
It cannot establish that the occupation has no graduate-level work. BLS categories are broad, and detailed occupational codes do not capture every job title. They can hide differences in seniority, sector, geography, employer preference, and specialization. Nor can the baseline distinguish a degree that is unrecognized from a degree that is being used productively but is not required for entry.
A careful reader should therefore keep three statements separate:
- The occupation’s typical entry assignment is below a bachelor’s degree. This is what the BLS category says.
- A worker in that occupation holds a bachelor’s degree. This is what the worker-level data say.
- The worker is underemployed in the fuller sense of having their college-level capabilities underused. This is a further interpretation that the first two statements do not prove on their own.
The gap between the second and third statements is where the rest of the measurement debate begins. Evidence about the education of other workers can show whether the occupation is more diverse than its entry label suggests. Evidence about earnings can show whether graduates receive a return within the occupation. Evidence about tasks, roles, and career movement would speak more directly to individual skill use. Each addition can change the interpretation, but each introduces its own assumptions.
The baseline should survive that scrutiny. Discarding it would lose a consistent reference point for the question of how people enter occupations. Accepting it as a final verdict would mistake a national occupational proxy for a description of an individual job. Its proper role is narrower and more useful: it identifies a possible education/job mismatch and tells us where to look for richer evidence.