magnitude. Large economies whose workers could not be included include China, India and Indonesia,
however data from these economies is reflected in fractional estimates reported as percentages,
according to the data coverage of the Future of Jobs Survey.
The estimates of the number of employees per sector which can be found in the Industry Profiles
(p183-236) are based on the full dataset of 1.739 billion employees worldwide. This calculation is
described in the user guide to the profiles (p82).
Metrics relating to the Global
Skills Taxonomy
Three survey questions probed the present importance of skills, their expected evolution in
importance from 2023 to 2027, and the strategic focus organizations will assign to upskilling and
reskilling their workers in skills from 2023 to 2027. In each case, respondents selected skills from a
list of 26 “cross-functional” skills from the Global Skills Taxonomy (see Table A2). No attempt was
made to represent skills and knowledge which is specialised within industries or occupations. In the
case of the first and third questions, skills data are reported as both raw shares of companies and
derived estimates of the mean normalised share of each skill per respondent. While the former metrics
have the benefit of simplicity, the latter metrics give equal statistical weight to all respondents and sum
to 100% across the full Global Skills Taxonomy.
Skill importance
The relative current importance of skills was estimated using the question: “What are the
core skills workers currently need to perform well in the key roles with a stable outlook?”. The
resulting data are reported as both a raw share of companies selecting each skill and a derived
metric. The derived metric accounts for the fact that respondents were permitted to select as many or
as few skills as they wished. In this derived analysis, each skill receives an equal share of a respondent’s
statistical weight, and all respondents are assigned the same overall weight. The share per skill is
then the mean taken across the respondents in a sample. This “normalized skill share” may be treated
as a rough approximation of the cross-functional skill set of workers today.
Skill evolution
The evolving importance of skills was estimated using the question: “For the key roles with a stable
outlook, would you expect an increase or decrease in the use of the following skills?”. Respondents
specified all 26 skills as increasing, decreasing or stable in importance over the next five years.
Across the report these figures are often presented alongside the net difference between the share
of respondents classifying the skill as increasing in importance and the share classifying it as
decreasing in importance, to obtain a single metric per skill.
Reskilling focus
The reskilling focus ofstrategic importance surveyed organizations apply to skills was gauged using the
question: “Keeping in mind your current strategic direction, please select the skill clusters on which
you are focusing your organisation’s reskilling and upskilling efforts in the next five years.” For this
question, respondents ranked an unlimited number of the 26 skills according to their importance. These
ranks were aggregated using the Borda method to calculate the “Reskilling ranking” which is used to
order the Skill Profiles. The calculation is outlined in the user guide to the Skill Profiles (p255). This
aggregated ranking is also used to order the skills in Figure 4.5, which also represents the difference
in this ranking with respect to the ranking of skills by their importance, as represented above, and in
Figure 4.2.
An equivalent analysis to that for skill importance is also performed. At this stage, the rankings assigned
by respondents are neglected, and the selected skills are treated equally. The data are now reported
as both a raw share of companies selecting each skill and a derived metric. The derived metric
accounts for the fact that respondents were permitted to select as many or as few skills as they
wished. Each skill receives an equal share of a respondent’s statistical weight, and all respondents
are assigned the same overall weight. The share per skill is then the mean taken across the respondents
in a sample. This “normalized strategy share” may be treated as a rough approximation of the mean
composition of organizations’ strategic priorities and reskilling needs from 2023 to 2027.Future of Jobs Report 2023 66