How to Read Job-Growth Data: Which Sources Actually Predict the Rising Professions?

The 2026 labour market is calm on the surface and moving underneath. The ILO projects global unemployment at 4.9 percent, about 186 million people. Eurostat’s euro-area vacancy rate slipped to 2.1 percent in the second quarter, and Indeed’s US postings index ended 2025 six percent above its February 2020 baseline but 5.2 percent down on the year. Hiring has cooled, yet AI postings keep climbing.

This guide, part of the kurums Rising Professions 2026 hub, is a methodology primer for HR leaders and CFOs. It compares the sources everyone quotes, explains percentage versus absolute growth, postings versus employment and vacancy duration, and ends with a watchlist method.

Which job-growth data sources actually predict the rising professions?

No single source does. Official projections forecast employment a decade ahead; posting indices and LinkedIn profile data detect new titles within months; shortage lists show which roles employers cannot fill. The table compares what each measures, its cadence and where it misleads.

Two patterns appear. The sources that look furthest ahead (BLS, Cedefop, WEF) update slowest and cannot see a brand-new title; those that update weekly (Indeed, Lightcast) see every new title but cannot say how many people hold the job. Shortage lists sit between. Coverage is the second: LinkedIn covers eleven countries in 2026 and Indeed ten; neither publishes for Türkiye, China, Japan, South Korea, Poland or the Gulf, so there, as the worldwide overview of rising professions shows, planners rely on national job boards and ministries.

How do the main source types differ in what they count?

Six families exist. Projections answer “how many jobs”, vacancy surveys “how many unfilled”, posting indices “what is advertised”, profile data “who moved”, employer surveys “what firms expect”, and shortage lists “what cannot be filled”.

Official projections: BLS and Cedefop

The BLS Employment Projections of 27 August 2026 expect US employment to grow 5.9 million (3.5 percent) in 2025–35, well below the previous decade’s 10.9 percent. The BLS calls this an equilibrium scenario, not a forecast, and assumes technology follows historical patterns, a caveat for AI roles. Cedefop’s 2025 Skills Forecast projects EU employment growth of only about 0.4 percent a year to 2035, so most openings in health, care and construction come from replacement, not expansion.

Vacancy surveys: ONS, Eurostat and the IAB

Vacancy surveys ask employers directly. The ONS counted 707,000 vacancies in May to July 2026, down 19,000 on the year; Eurostat’s euro-area rate was 2.1 percent in Q2 2026, from 4.1 percent in the Netherlands to 0.5 percent in Romania. Germany’s IAB survey found 1.15 million open positions in Q1 2026 while only about 600,000 were registered with the Federal Employment Agency; analysis built on registered vacancies alone sees half the picture.

Posting indices: Indeed and Lightcast

Indeed Hiring Lab’s index is seasonally adjusted and benchmarked to 1 February 2020; at the end of 2025 US postings were 6 percent above baseline but 5.2 percent lower year on year, with AI-mentioning postings at 4.2 percent of the total. Lightcast, formerly Burning Glass, aggregates from more than 220,000 sources, deduplicates over a 60-day window (removing up to 80 percent of collected ads) and warns that postings are not vacancies.

Profile data: LinkedIn Jobs on the Rise

LinkedIn measures job starts, not adverts. The 2026 edition ranked titles by growth in jobs members began between 1 January 2023 and 31 July 2025, excluding internships and interim roles; it publishes ranks, not rates. AI engineer is first in nine of eleven countries; prompt engineer leads in India and Head of Artificial Intelligence in Germany.

Employer surveys: the WEF Future of Jobs Report

The World Economic Forum surveyed more than 1,000 large employers representing over 14 million workers. Its figures to 2030 (170 million jobs created, 92 million displaced) are extrapolated expectations, not counts. Usefully, it separates the two growth questions: big data, fintech and AI specialists lead by rate, while farmworkers lead by volume with 34 million additional jobs on a base of about 200 million.

Shortage lists: Australia, the UK, Canada and Germany

Jobs and Skills Australia’s Occupation Shortage List rated 293 of 1,022 occupations (29 percent) in shortage in 2025, down from 333. The UK Migration Advisory Committee recommended in July 2026 that 28 occupations, down from 52, keep Temporary Shortage List access; that list and the Immigration Salary List expire on 31 December 2026. Canada’s 2026 Express Entry categories add researchers, senior managers and military recruits and drop agriculture. Germany’s agency counted 157 shortage occupations for 2025, led by care, construction and trades, with software development dropping off.

Why does percentage growth tell a different story from absolute growth?

Percentage growth rewards small occupations; absolute growth rewards large ones. A role with 50,000 holders can double and still add fewer jobs than a million-strong role growing five percent. The percentage shows momentum, the absolute number shows hiring volume.

Nurse practitioners grow fastest at 41.0 percent, yet the BLS names services for the elderly and persons with disabilities as the largest job-creating industry (625,400 positions), and private healthcare and social assistance adds more than 2.2 million jobs, 37 percent of the total. An HR team reading only the fastest-growing list plans for a few thousand clinicians and misses the care workers it will actually recruit.

Decline works the same way. The BLS expects office and administrative support to shrink 4.0 percent, losing 752,100 jobs; Cedefop projects coal at minus 10.7 percent a year, a steep rate on a small base. Neither is collapse; both leave time to redeploy, as the guide to the skills behind 2026’s rising jobs explores.

What is the difference between job postings, vacancies and employment?

Employment counts people in jobs; vacancies count positions employers are trying to fill; postings count advertisements. One vacancy can generate several postings, many vacancies are never advertised online, and a posting surge may reflect recruiter behaviour rather than new jobs.

Occupational skew is the first problem. The Greater London Authority found that in 2019 managers and professionals were 50.1 percent of UK online postings but only 22.0 percent of vacancies in the Employer Skills Survey. A nursing home filling care roles by word of mouth and a bank advertising one analyst post on five boards look very different in posting data and almost identical in a vacancy survey.

Duplication is the second: Lightcast removes up to 80 percent of collected ads as duplicates, so raw counts overstate demand several times over. Divergence is the third: Indeed found US tech postings 34 percent below their 2020 level at the end of 2025 while tech postings mentioning AI were 45 percent above it. Both are true for different slices, which is why LinkedIn’s job-start data is a useful cross-check: it counts people who moved, not adverts.

How should you read vacancy duration and fill rates as shortage signals?

Vacancy duration and fill rate measure friction, not volume. A role can be small and slow-growing yet still a shortage if employers take months to fill it. Germany’s Federal Employment Agency and Jobs and Skills Australia build their shortage ratings on these indicators.

The German Engpassanalyse defines Vakanzzeit as the median days between an employer’s desired start date and the vacancy’s closure, for filled positions only. It sits alongside job seekers per vacancy, the occupational unemployment rate, the exit rate from unemployment, median pay growth and the change in the foreign-employee share, which has doubled from about 7 to 14 percent in shortage occupations since 2014.

Australia uses the fill rate, with a reading below 67 percent indicating likely shortage. Its 2025 report shows easing: the average fill rate rose from 68.7 to 70.9 percent and applicants per vacancy climbed from 13.5 in 2020–21 to 27.5 in 2024–25, of whom only 4.3 were suitable. Yet 139 occupations had been in shortage every year since 2021, 71 of them trades. More applicants did not fix roles that need licences and years of training.

Two ratios complete the toolkit. The ONS reported 2.5 unemployed per vacancy in April to June 2026, up from 2.3; the IAB reported 264 unemployed per 100 openings in Q1 2026, 13 more than a year before. Both markets are loosening, so the roles that stay hard to fill are the true rising professions of 2026.

How is LinkedIn Jobs on the Rise computed, and where can it mislead?

LinkedIn ranks job titles by growth in member job starts over three years. It is fast, cross-national and based on real moves, but it reflects LinkedIn’s member base, rewards new titles and publishes ranks without rates, so it signals novelty rather than scale.

Three biases follow. First, the member base: profile-keepers work in offices, so care workers, electricians and drivers, the roles that dominate German and Australian shortage lists, are under-counted. Second, title inflation: a job called “data scientist” in 2023 and “AI engineer” in 2025 shows explosive growth in the new title even if headcount barely moved. Third, the three-year lookback makes lists volatile.

Where LinkedIn publishes, the top of the list is nearly identical, from the United States and the UK to Australia and Brazil, which says as much about the platform as about each economy. Elsewhere, other signals take over: Zhaopin reported AI-agent postings up 455 percent year on year in China; JobKorea counted AI postings up 128 percent in South Korea; and Japan’s MHLW openings-to-applicant ratios of 3.76 for care work and 7.72 for structural trades surface roles no profile-based list would show.

How do you build a country watchlist from these sources?

Build it in five layers: scale from projections, speed from postings and profile data, scarcity from vacancy duration and fill rates, supply from training and immigration data, and price from salary guides. A profession scoring on three layers belongs on the list.

Registered nurses, for example, score on scale (BLS, Cedefop replacement demand), scarcity (the BA’s strongest shortage group; Australia’s persistent-shortage list) and supply (Canada’s health-care category). AI engineers score on speed and price but not yet on scarcity, since no agency publishes vacancy duration by title. The country-by-country salary comparison in this series supplies the price layer for all twenty-one markets.

What should HR and workforce planners do with this data in 2026?

Treat every list as one input, label its unit and window, and triangulate before acting. These six steps make a quarterly routine a small team can run.

1. Anchor on official projections. Record both percentage and absolute change from the BLS or Cedefop.
2. Overlay speed. Check Indeed and LinkedIn; flag titles rising in postings but absent from projections as possible relabelling.
3. Test for friction. Check vacancy duration, fill rate or unemployed-per-vacancy ratios. Fast-growing and hard to fill is a shortage; fast-growing and easy to fill is not a premium.
4. Map the policy calendar. Australia reviews annually, Canada each February, the UK by 31 December 2026.
5. Benchmark pay against two sources. Compare a recruiter guide with an official median before approving a premium.
6. Invest in supply where friction is structural. For nursing, electrical trades and semiconductor roles, fund apprenticeships and sponsored study; the buyer’s guide to certifications and courses and the MBA & Business Education directory list providers.

Frequently Asked Questions

Sources and method

1. Employment Projections 2025–35 — BLS, 2026
2. Jobs on the Rise 2026, US — LinkedIn News, 2026
3. Skills Forecast dataset — Cedefop, 2026
4. Vacancies and jobs in the UK: August 2026 — Office for National Statistics, 2026
5. Euro area job vacancy rate at 2.1% — Eurostat via Hellenic Shipping News, 2026
6. Engpassanalyse 2025 press release — Bundesagentur für Arbeit, 2026
7. Indikatoren der Engpassanalyse — Statistik der Bundesagentur für Arbeit
8. IAB-Stellenerhebung Q1 2026 — dpa via onvista, 2026
9. January 2026 Labor Market Update — Indeed Hiring Lab, 2026
10. Job Posting Analytics (JPA) Methodology — Lightcast
11. Understanding online job postings data — Greater London Authority
12. Future of Jobs Report 2025 — World Economic Forum, 2025
13. Employment and Social Trends 2026 — International Labour Organization, 2026
14. 2025 Occupation Shortage List: Key Findings Report — Jobs and Skills Australia, 2025
15. MAC Temporary Shortage List review — Morgan Lewis, 2026

Method: we read each source’s methodology notes and latest release, recorded what it measures and its cadence, and cross-checked country findings against the kurums country guides, attributing each figure to its original publisher. Figures checked November 2026.