What is a skill mismatch? Definitions, measures and disputations

Huge strides are being made in the measurement of skill demand. The ETF’s Skills Lab seminar, The Use of Big Data to Support Anticipation for the Green and Digital transition (April, 2023), revealed the way in which data science techniques have been able to massively increase the amount of information available, in a timely manner, on the specific skills required to work in a wide variety of jobs. This type of information has the potential to reduce the level of skills mismatch.

Nevertheless, there remains a need to monitor the level of skill mismatches because of their potential to impose substantial costs on individuals (such as unemployment, reduced lifetime earnings), employers (reduced output and growth), and the economy as a whole where they prove persistent.

What is a skills mismatch? It is, at first glance, a simple question but one which proves difficult to answer.[1] First, because there is a need to make a distinction between shortages and surpluses. Secondly, there is a need to consider the population of interest. Is the focus on the workplace or the external labour market? If the former, then the interest will be in whether employees possess sufficient proficiency to fulfil the requirements of their current job, something which is usually determined by the employer though the employee may have a view on this as well. If the interest is in the external labour market then the focus will be upon recruitment and whether are too few or too many applicants with the required skills, experience, or qualifications sought for a particular job. Thirdly, there are the types of skills possessed by individuals to consider. For instance, are their skills at the right level or in subject areas relevant to a particular job?

On the basis of the above, it is relatively straightforward exercise to devise a classification of skill mismatches with respect to either the workplace of the external labour market. So far, so easy. But matters soon become much more complicated when attention turns to measurement. There needs to be clear measures of the concepts of interest. In the first instance, how is skill to be measured? This proves to be challenging. In his seminal article What is Skill? Attlewell comments: “… like so many common sense concepts, skill proves on reflection to be a complex and ambiguous idea”.[2] There is insufficient space here to comment further on the definition of skill other than to say that until relatively recently skill has been mainly measured or, more precisely, quantified with respect to occupation and qualification. Both measures are, to the say the least, imperfect.

Occupation groups together jobs which are more alike to one another then they are to another set of jobs. It stands to reason that any one occupational group, depending upon the degree to which data are aggregated, is likely to include a wide variety of jobs and therefore a wide variety of skill needs. At best occupation provides an imprecise measure of skills. Qualification possibly provides an even vaguer measure of skill in that signals that an individual has successfully completed a programme of learning which may be relevant to a particular job but not necessarily that a person has the skills and know-how to do the job in practice.

Despite these reservations, occupation and qualification are commonly used measures of skill in large part because data pertaining to both tend to be readily available and accessible (for example, from labour force surveys). Accordingly, there are indictors which purport to say something about skill mismatches which are constructed on the basis that individuals’ qualifications and occupations reveal something about the extent to which their skills are matched to their current jobs or ones for which they have applied. For instance, a relatively common measure of mismatch is the share of workers in an occupation who are over- or under-qualified where over or under are typically defined with reference to the average level of qualification held in that occupation. That individuals on average might be over- or under-skilled for that occupation is not taken into account.

Surveys of employers and employees, respectively, have sought to provide more detailed evidence on mismatches. Employer surveys, for example, are able to provide information on:

• recruitment difficulties: - i.e. the extent to which vacancies prove hard-to-fill because there is an insufficient number of applicants with the skills, experience or qualifications required (a potential indicator skill shortages); and
• internal skill gaps – i.e. the extent to which those employed in a particular workplace, or those working in a particular occupation within a workplace, are fully proficient at their existing jobs.

Similarly, surveys of employees / workers ask questions about the extent to which a worker feels under- or over-skilled in their current job or parts thereof (i.e. whether they feel that they lack certain skills or possess skills which are not sufficiently utilised in their view). Indubitably, survey data provide valuable insights into the extent of skill mismatches but they rely almost wholly upon self-reporting by employers or individuals. And employers or individuals may have their own reasons for over or understating the level of mismatch. One study suggested that, in actuality, many claims that vacancies remain unfilled because of a shortage of applicants with the skills required were more to do with the terms and conditions of employment on offer rather than being anything to do with the supply of skills.[3] A timely reminder if one were needed to be cautious when interpreting data which purports to be about skill mismatches.

With the passage of time, approaches to the measure of skill mismatches have improved. The European Skills and Job Survey, for instance, poses a series of questions to workers about the way in which their skills are matched to different aspects of their current job.[4] And employer surveys provide the means to cross-check claims of skill shortages with reference to whether those reporting shortages respond in some way. If a skill shortage is significant, then one might expect employers to respond by, say, providing (more) training to existing or new employees (cf. the reskilling and upskilling agenda), or by changing their approach to recruitment.

There are measures of mismatch which are more objective in the sense that they do not rely on employers’, workers’, or anyone else’s views. Differential occupational wage growth, for example, will indicate something about the extent to which skills are in short- or over-supply, presuming that up to date wage data are available. One might expect wage levels to respond to supply, but this is likely to be the case over the short-run only in countries where collective bargaining and other labour market institutions play little or no role in wage setting. In other words, differential occupational wage growth will provide an indication of skill mismatches only where the market exclusively sets wage levels.

The commentary is not designed to put off researchers from measuring skill mismatches. Measuring the extent, causes and implications of skill mismatches is clearly important given the costs mismatches potentially impose upon individuals, employers and the economy as a whole where they prove persistent. Rather the purpose here is to point to the need to be circumspect when measuring skill mismatches and interpreting the results. It also points to the need for a multi-faceted, multi-indicator approach to the measurement of skill mismatches.

[1] A detailed commentary can be found in: McGuinness, S., Pouliakas, K., and Redmond, P. (2018). ‘Skills Mismatch: Concepts, Measurement and Policy Approaches’. Journal of Economic Surveys, 32(4), 985–1015

[2] Attlewell, P. (1990) “What is Skill?” Work and Occupations, Vol. 17, No.4, pp.422-448

[3] Cedefop (2015) Cedefop (2015). Skills, qualifications and jobs in the EU: the making of a perfect match? Evidence from Cedefops European skills and jobs survey. Luxembourg: Publications Office

[4] Cedefop (2022). Setting Europe on course for a human digital transition: new evidence from Cedefop’s second European skills and jobs survey. Luxembourg: Publications Office. Cedefop reference series No. 123