Overview

The OECD's occupational-proximity measure uses skills named in online job postings to identify destination-relevant overlap across occupations. It weights skills by specificity, compares shared weighted skills with the destination's requirements, and produces a directional score from zero to one. In Italy, the United Kingdom, and the United States, higher proximity is positively associated with observed occupational transitions; the measure performs better in larger regional markets in Italy and the United Kingdom. The relationship is not causal and the score describes occupational pairs rather than individuals. Local vacancies, credentials, skill depth, job quality, and worker constraints are still needed to turn a plausible map into a feasible pathway.

Learning outcomes

The score measures destination coverage, not the probability that a worker will be hired.
Specificity weighting and directionality make rare shared destination skills more informative than generic overlap.
Positive validation in Italy, the United Kingdom, and the United States is an association between occupational pairs, not causal proof or an individual forecast.

Keywords

Occupational proximityJob postingsSkill overlapDirectionalityVacanciesMobility

From Transferable Skills to a Realistic Pathway

The local problem identified in the green transition is not only where vacancies exist, but how to identify occupations that a worker could plausibly enter from their current one. This chapter examines the OECD's occupational-proximity measure, which uses skills named in online job postings to map destination-relevant overlap. The practical value is diagnostic: a high score can identify a pathway worth investigating, but only observed movement, local vacancies, credentials, and worker circumstances can turn similarity into feasibility.

What a job posting adds to a transition analysis

Sector and job title are useful labels, but they often hide transferable work. Two jobs in different sectors may involve similar reporting, maintenance, coordination, or problem-solving tasks. Two jobs with similar titles may demand different tools or levels of responsibility. Formal education narrows the picture further: it records a credential or level, not the full set of capabilities that an employer currently asks a worker to use.

The OECD measure starts from online job postings collected and processed by Lightcast. Job advertisements supply occupational labels, locations, and skill keywords that employers mention in current demand. For each occupation in each country, the study aggregates commonly cited skills into a country-specific skill set. The resulting representation is not a list of what every incumbent can do. It is a picture of the requirements visible in the vacancy market at the time of measurement.

That distinction makes the data useful for the entry-margin problem. If employers have changed the requirements for routine or green-task work, current postings can reveal capabilities that older occupational classifications do not. They can also show overlap across sectors. A worker leaving an exposed manufacturing occupation may have no obvious “same-sector” destination, yet their experience with quality control, equipment, scheduling, or safety may connect to jobs elsewhere.

The same data introduce a selection problem. Online postings are not a census of all hiring: some employers and occupations appear more often than others, and an advertisement may describe skills a firm hopes to acquire rather than the complete skills used by current employees. Posting language can also vary across countries. In non-English-speaking settings, an additional translation step is used, and Lightcast's processing choices may affect which keywords appear.

For that reason, the OECD uses a fixed set of skills for an occupation within each country rather than trying to estimate a thin regional skill set for every occupation. This improves stability when local posting counts are small, but it means that the measure captures country-level demand patterns more directly than every local variation. The data are a useful demand-side lens, not a direct assessment of a particular worker's competence.

Why common skills should count less

A simple overlap measure would count the number of skills that two occupations share. That would make a common skill such as teamwork carry the same weight as a specialised skill such as developing a particular kind of code or operating a specialised process. The OECD measure instead gives more weight to skills mentioned in fewer occupations. A skill that appears almost everywhere provides little information about why two occupations are close; a skill concentrated in a smaller set of occupations can reveal a more specific relationship.

Published by DeepMox

FAQ

What does the OECD occupational-proximity score measure?

It measures how much of a destination occupation's weighted, posting-based skill requirements are represented in the origin occupation's skill set, from zero to one.

Why is the proximity score directional?

It compares shared weighted skills with the destination's total requirements, so origin-to-destination coverage need not equal the reverse.

Why can a high proximity score still fail to produce a job transition?

It does not measure skill depth, credentials, local vacancies, employer hiring, job quality, or worker constraints; the destination may not be available or reachable.

Previous11 minGreen Transition: Local Labour-Market Options

Next11 minPipeline-and-Place Diagnosis for Structural Change

OECD Occupational Proximity for Job Transitions

From Transferable Skills to a Realistic Pathway

What a job posting adds to a transition analysis

Why common skills should count less

FAQ