over several months centred around the turn of 2023. Metrics relating to both concepts reflect
forecast structural changes in employment across companies, economies, industries and roles.
Turnover induced by employees moving between jobs for personal reasons is not included.
Fractional metrics
Respondents aggregated roles included in the jobs taxonomy to six groups:
- roles representing a large proportion of the organization’s workforce with a stable
employment outlook for the next five years; - roles representing a large proportion of the organization’s workforce which are expected to
grow in the next five years; - roles representing a large proportion of the organization’s workforce which are expected to
be increasingly redundant in the next five years; - specialised and strategic roles which are crucial to the organization, represent a small proportion
of the organization’s workforce today and are expected to have a stable employment outlook
in the next five years; - specialised and strategic roles which are crucial to the organization, represent a small proportion
of the organization’s workforce today and which are expected to be increasingly important in the
next five years; and - specialised and strategic roles which are crucial to the organization, represent a small proportion
of the organization’s workforce today and which are expected to be increasingly redundant in the
next five years.
Respondents allocated up to three roles from the jobs taxonomy to each of the six groups. One of
the three roles in the three specialised and strategic categories could be specified by a free-text field.
Free-text fields were subsequently allocated to jobs in the jobs taxonomy where possible. Metrics on
roles are only published in the report when they meet statistical criteria in a given sample.
Respondents subsequently allocated workforce fractions to each of the above groups of jobs,
both at present and as predicted for 2027. These workforce fractions were used to calculate two
metrics: estimated net growth between 2023 and 2027 and estimated structural labour-market churn
from 2023 to 2027, for the labour forces pertaining to roles in the jobs taxonomy and industries in the
industry taxonomy. In both cases, the fractional increase or decrease between 2023 and 2027
pertaining to a job or industry is compared to its workforce fraction in 2023. In the calculation of
net growth, increases and decreases are added and subtracted in the numerator, and divided
by the sum of workforce fractions in 2023 in the
denominator, across all responses pertaining to a particular role or industry. In the case of churn,
the calculation is identical, except for taking the absolute value of decreases, so that all terms in the
numerator are positive. In both cases, changes in the stable workforce fraction are neglected in the
numerator, to respect the respondent’s indication that this workforce is stable. Stable workforce
fractions from 2023 are nevertheless included in the denominator, to ensure that responses indicating a
stable workforce fraction appropriately suppress the magnitude of net growth and churn.
Reweighted metrics
ILO data were then used to translate the forecast fractional net growth for each role into estimates
of the number of jobs that will be created or displaced between 2023 and 2027. ILO estimates
(excluding their modelled estimates) of the number of employees in each occupational category of
ISCO08 level 2 were used as a basis for the number of employees working at the time of publication.
To approximate the number of employees in each occupation of the jobs taxonomy used in
the Future of Jobs Survey, the jobs taxonomy (a modified and extended version of the O*NET
SOC occupational classification) was mapped to the ISCO08 occupational taxonomy used in the
ILO data by modifying and extending the map developed by the U.S. Bureau of Labor Statistics,
which connects O*NET SOC level 4 and ISCO08 level 2. Ambiguities arising due to differences in the
granularity of job titles were resolved using Future of Jobs Survey data to estimate the relative numbers
of employees based on the number of times jobs were selected by respondents. Estimates of present
employment were then multiplied by the fractional net growth estimates obtained from the survey, to
estimate net growth worldwide in units of millions of employees.
Using this method, the Future of Jobs dataset described in Chapter 3 corresponds to 673 million
employees. By comparison, the ILO dataset used in the analysis accounts for 820 million employees.
The remaining 147 million employees correspond to roles for which the Future of Jobs Survey did
not collect sufficient data to reliably estimate net growth. Data on employees rather than general
employment was used as organisations responding to the Future of Jobs Survey maintain workers in
formal rather than informal employment.
The available ILO dataset which boasts an occupational employment breakdown with level
2 granularity in ISCO08 is smaller than modelled ILO estimates of a total of 1.739 billion employees
worldwide when country-level data gaps are extrapolated, and smaller still than the estimated
3.283 billion workers in either formal or informal employment. Extrapolating beyond this sample of
673 million employees would require unfounded assumptions regarding the structure of labour
markets, but readers may use these normalizations to perform rough estimations of the expected full
Future of Jobs Report 2023 65