Two Axes of Fit: Credentials and Capability
To understand whether education and work fit, examine two relationships separately: a worker’s formal qualification against the education associated with a job, and the worker’s usable capability against the skills that the work actually requires. This chapter develops that distinction because it gives readers a more reliable way to classify a mismatch than treating a degree as a complete inventory of what someone can do.
The distinction is easy to state but easy to lose in practice. A graduate employed in a job that does not normally require a degree may be called overeducated. That label can be accurate as a description of the credential-to-job relationship, yet it does not establish that the graduate has unused skills, lacks motivation, or is unable to perform the job. Conversely, a worker with the expected qualification can still be underskilled for a role after a new digital system, process, or responsibility changes what the job demands.
The key is to keep the two axes connected without making them interchangeable. Formal education can develop skills, signal preparation, and open access to work. It is still only one source of capability, and a job’s credential requirement may reflect hiring convention as much as the content of the work.
The education axis has two different requirements
Overeducation and undereducation are defined by comparing educational attainment with the level of education judged necessary for a job. If a worker has more formal education than that requirement, they are classified as overeducated; if they have less, they are classified as undereducated. The comparison concerns a qualification level, not a direct test of every task the worker can perform.
That definition becomes more informative when the requirement is split into two questions: what education is needed to get the job, and what education is needed to do it? An employer may use a degree as a screening signal even when the work itself could be performed competently with a lower level of formal education. A gap between the two can indicate credential inflation or a surplus entry requirement rather than a deficit in the job’s actual content.
These two requirements can diverge through the hiring process. A qualification is visible before a candidate has demonstrated how they work, so it can act as a convenient proxy for persistence, general preparation, or expected trainability. Once the person is in the role, however, the relevant question shifts to the knowledge and tasks the job actually uses. Treating the hiring proxy as a description of job content turns a decision rule into a supposed fact about capability. It also makes it difficult to tell whether a graduate is in a genuinely low-skill job or simply in a job whose employer set a high screening threshold.
The reverse error is possible when an occupational label is treated as a complete description of every workplace carrying that label. Two positions with the same broad title may differ in autonomy, tools, customers, regulatory demands, or the amount of judgement required. An educational requirement inferred from the occupation may therefore describe a typical pattern rather than the specific work performed by one person. The qualification axis is about a relationship to a defined job, and the definition of that job matters.
Suppose an employer advertises an administrative position requiring a degree because many applicants have one, although the tasks involve procedures that do not depend on degree-level knowledge. A graduate in that position may be overeducated under a comparison with the job’s educational requirement. But the graduate may also be using their analytical, writing, coordination, or customer-service abilities fully. The credential mismatch would be real; a skills mismatch would remain unestablished.
The reverse case is equally important. A technician may lack the formal qualification conventionally associated with a role but have acquired the relevant competence through years of work and training. They could be undereducated according to a classification while being capable of doing the job. The label may still matter for regulated access, pay progression, or mobility, but it does not prove an immediate skills deficit.
How the educational requirement is inferred matters as well. Researchers may ask workers what education was needed, infer a typical level from others in the occupation, or rely on professional evaluations of jobs. These approaches can disagree because they answer slightly different questions. An occupational average may reflect historical entry patterns; a worker’s answer may reflect personal interpretation; an expert classification may become outdated as work changes. The education axis is therefore not a natural property attached to a job. It is a comparison produced from a definition and an evidence source.
This is why a degree-to-job statistic should not be treated as a verdict on the value of higher education. Education may provide knowledge, ways of reasoning, networks, eligibility for regulated work, and a foundation for later learning. A worker can receive those benefits even when their first job does not match the level of the credential. The policy question is whether the mismatch represents wasted capability, a temporary transition, a voluntary trade-off, an inflated hiring filter, or a job market that has not created enough work using the available capability. The credential alone cannot decide among those explanations.
The skills axis follows capability into work
Overskilling and underskilling use a different comparison. Overskilling describes a worker who believes they possess more skills than the current job requires; underskilling describes a worker who believes their current skills do not meet the job’s demands. The reference point is the work being performed, not the level of the worker’s diploma.
This comparison includes human capital acquired outside formal education. A person may learn a production process from a colleague, develop judgement through repeated cases, gain project-management ability through experience, or refresh a technical skill in workplace training. None of those capabilities is fully represented by the highest qualification in an education register. A degree may contribute to capability, but the capability-to-work comparison asks what the person can bring to this job at this time.
Capability is consequently relational. The same worker can be well matched in one role, overskilled in a role with narrow or repetitive tasks, and underskilled in a role that introduces unfamiliar tools. The worker has not changed merely because the job comparison changed. What changed is the relationship between the person’s capabilities and the requirements of the work. This is also why a skills diagnosis should name the task or capability at issue rather than speaking of a person as simply “high-skilled” or “low-skilled.”
The comparison is also about more than a single quantity called “skill level.” The ILO review notes that the European Skills and Jobs Survey asked about eleven specific areas as well as overall skills. The overall measure had correlations of only 0.19 to 0.26 with mismatches in particular technical or transversal skills. That does not make the overall question useless. It indicates that different components of work can produce different experiences of fit.
For example, a worker might be highly capable in literacy, analysis, and problem-solving but lack proficiency with a newly introduced software system. They could be overskilled in the broad sense that the job uses only a fraction of their reasoning ability, yet underskilled for one immediate technical task. Another worker may have the required technical knowledge but too few opportunities to use communication or planning skills. A binary matched-versus-mismatched label would conceal the distinction between capability, skill use, and job requirement.
This multidimensionality changes what a sensible response looks like. A technical deficit may call for practice, mentoring, or training tied to a task. Underuse of analytical ability may call for broader responsibilities, more autonomy, or a different job design. Neither response follows simply from knowing the worker’s degree level. A qualification can help identify a learning foundation, but it cannot tell an employer which capability is unused or which task is difficult without more direct evidence.
The source of the judgement must also be kept visible. Most large datasets do not directly observe a specialist evaluating every worker’s competence against every task. They rely on survey answers, and the wording shapes the result. A question about unused skills may capture underutilisation. A question about being able to handle more demanding duties may capture perceived capacity. A question about needing further training may express a desire for future progression rather than an inability to perform current duties. These are related experiences, but they are not identical measurements.
The relationship between the two axes becomes clearer in the ESJS evidence. Among workers classified as overeducated, 44% considered themselves overskilled when they were hired, 18% considered themselves underskilled, and 38% considered their skills matched to the job. Among undereducated workers, 27% considered themselves underskilled at hiring. The figures show partial overlap and substantial divergence. They do not provide a conversion rule such as “one degree level equals a fixed amount of skill.”
The same pattern appears when the full workforce is considered. Ten percent of workers in the survey were both overeducated and overskilled, while 26% were matched in education but overskilled. The second group is especially important: their qualifications fit the educational expectation, but the job still did not use their capability fully. If an analyst looked only at degrees, that underutilisation would disappear.
At the other extreme, the small share of workers who were both undereducated and underskilled in this high-income EU sample should not be interpreted as evidence that educational deficits never matter. It shows only that the two conditions did not coincide often in that dataset and definition. Work experience, employer training, task allocation, and the way requirements were measured can all affect the relationship.
The diagram is deliberately asymmetric. A qualification can contribute to capability, but it does not contain every capability a worker has. Experience and training can contribute as well. The job supplies the context in which a match is assessed, because the same person may be well matched for one role and under- or overskilled for another. A qualification-to-job comparison and a capability-to-job comparison can therefore produce different results without either being logically mistaken.
What to ask before naming the problem
A degree-to-job example becomes evidence of a skills problem only after several questions have been answered. First, what does the job actually require: a credential for legal access, a credential as a hiring signal, or knowledge and skills used in daily tasks? Second, what evidence describes the worker’s capability: a qualification, a demonstrated task, a worker report, an employer assessment, or a record of training? Third, when is the comparison made—at recruitment, after onboarding, or after the work has changed? Fourth, who is making the judgement, and what interests or limits shape that observation?
These questions are not bureaucratic additions to an otherwise obvious conclusion. They determine the object being measured. If the concern is credential inflation, the relevant evidence may compare hiring requirements with job content and graduate outcomes. If the concern is a worker’s technical deficit, the relevant evidence may examine specific tasks, training, and performance. If the concern is unused capability, the analyst must look at task allocation, autonomy, progression, and opportunities to apply skills. A single degree statistic cannot answer all three.
The most defensible conclusion is therefore modest but useful: education and skills mismatch are related dimensions of labour-market fit, not synonyms. Formal education can shape capability and influence access to jobs, while capability is also formed and tested through experience, training, and the organisation of work. The next question is how researchers turn these dimensions into statistics—because the method used to infer a job requirement or a worker’s skill can change the apparent size and location of the problem.