TL;DR
Get everyday essentials delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
A November 9, 2026, article from Thorsten Meyer AI examines why roughly four in five U.S. workers are employed in services, drawing on labor statistics and economist William Baumol’s theory of uneven productivity growth. It argues that the shift began well before U.S. factory employment peaked in 1979, while noting that AI’s effect on service jobs remains unsettled.
About four in five U.S. workers were employed in services in 2025, according to International Labour Organization estimates cited in a November 9 article by Thorsten Meyer AI. The article argues that the shift was underway decades before manufacturing employment peaked in 1979, and uses economist William Baumol’s theory of uneven productivity growth to explain why labor-intensive services remain a growing part of the economy.
The article reports that services made up 79.6% of U.S. employment in 2025 under the ILO measure. A separate calculation from the Bureau of Labor Statistics’ Current Employment Statistics, which counts payroll jobs outside farming and excludes self-employed workers, puts service-providing industries at 62% in 1939, 67% in 1967, 72% in 1979 and 86% in 2025. These figures use different populations and should not be treated as the same series.
The timeline challenges a common shorthand that services expanded mainly because factories declined. U.S. manufacturing employment reached nearly 20 million jobs in June 1979, according to the article’s account of earlier coverage, but the services share had already grown for decades. It continued to rise after that factory employment peak.
The explanation centers on Baumol’s cost disease: productivity can rise rapidly in work that can be mechanized or reorganized, while activities requiring direct human time may be harder to speed up without changing the service. The article applies this distinction to performances, education and health care, then asks whether AI can change the productivity outlook for service work.
How 80% Ended Up In Services
About four in five U.S. workers were employed in services in 2025. The shift began decades before manufacturing employment peaked, and AI’s effects on service jobs remain unsettled.
A long shift, not a sudden switch
A separate BLS payroll-job measure shows the service share rising before and after manufacturing employment peaked.
Why some services cost more
William Baumol’s theory links rising costs to uneven productivity growth across types of work.
- Technology can help produce more with fewer labor hours.
- Manufacturing and some service tasks can be reorganized or mechanized.
- A live performance still requires musicians’ time.
- Education and health care often depend on direct human effort.
Price paths diverged
Official U.S. price changes from 2000 to 2025 illustrate very different trends; they do not prove Baumol’s theory caused them.
What the figures can—and cannot—say
Different worker groups
The 79.6% ILO estimate and BLS payroll shares use different populations and methods, so they are not one continuous series.
One explanation, not the whole story
Baumol’s theory describes productivity pressure; regulation, financing, demand and policy can also shape costs.
Job effects remain open
A customer-support study found higher productivity, but does not show how AI will change employment across services.
Why Service Jobs Keep Growing
The figures matter because they describe a long-running change in where people work, not simply a recent consequence of factory closures. If productivity improves faster in manufacturing than in many face-to-face services, the relative importance of services can rise even while factories continue producing goods. The article’s figures point to a shift that spans several generations of workers.
Baumol’s theory also offers one explanation for why costs can rise in services. In a live performance, for example, cutting the musicians’ time could change the product itself. When wages rise across the economy, employers in activities that cannot readily reduce the labor required may face higher costs without matching productivity gains. The theory describes a pressure on costs; it does not establish that every price increase is unavoidable or caused by that pressure.
The distinction has practical consequences for households and policymakers. Health care, education and repairs are services people rely on, and changes in their prices affect budgets and public spending. The article cites U.S. health spending at 18.0% of the economy and Germany’s at 12.4%, while cautioning implicitly that shared productivity pressures cannot by themselves explain the difference between national systems.
Popular right now
LG S90TR 7.1.3‑Channel OLED Evo TV Matching Home Theater Soundbar (Black)
Samsung 65-Inch QN70FA Neo QLED 4K Smart TV with 8-Foot WireLogic Ruby Ultra High Speed HDMI Cable (2025)
Roborock RockNeo Q110H Wire-Free Robot Lawn Mower
As an affiliate, we earn on qualifying purchases.
From Factory Output to Human Time
The article traces the argument to William Baumol and William Bowen’s 1966 book, Performing Arts: The Economic Dilemma, which examined financial pressures in theater, opera, music and dance. Their example was a 45-minute Schubert string quartet: four musicians still require about three hours of combined labor to perform it. The concert takes much the same human effort even as other industries find ways to produce more with fewer labor hours.
Baumol developed the broader idea in his 1967 paper, “Macroeconomics of Unbalanced Growth: The Anatomy of Urban Crisis.” He contrasted industries where labor is mainly an input that can be reduced through technology with activities where labor is closely tied to the service’s value. He described the division as a matter of degree, not an absolute split; retail and transport, for instance, have adopted technologies and practices that change how much labor particular tasks require.
To illustrate different price paths, the source gives official U.S. price changes from 2000 to 2025: a quality-adjusted television price fell 98%, while hospital services rose 274%, college tuition 188% and car repair 143%. Haircuts and car repairs, it notes, also rose substantially, showing that health insurance and student loans are not the only possible factors behind service prices. The figures are presented as comparisons, not proof that Baumol’s theory alone caused those changes.
“No one has yet succeeded in decreasing the human effort expended at a live performance of a 45-minute Schubert quartet much below a total of three man-hours.”
— William Baumol and William Bowen, quoted from their 1966 book
What the Employment Figures Cannot Show
The employment shares show how the workforce is distributed, but they do not by themselves prove why the distribution changed. The BLS payroll figures and ILO estimate cover different worker groups and use different methods; the article says the BLS series excludes farm workers and self-employed people. The precise comparison depends on which measure is used.
The source offers Baumol’s theory as an explanation, not a complete account of every service industry or price increase. Differences in regulation, financing, market structure, demand and public policy may also matter. The comparison between U.S. and German health spending does not isolate which factors account for the gap.
The article also raises, rather than resolves, the question of AI’s effect on service employment. It cites a study of 5,172 customer-support agents in which an AI assistant increased issues resolved per hour by 15% on average. That result concerns productivity in the studied support work; it does not establish how many jobs will be eliminated, created or changed across the service sector.
AI’s Next Test in Service Work
The article frames the next issue as whether AI can raise productivity in work traditionally resistant to automation, and what happens to workers if it does. Higher output per hour could reduce the labor required for some tasks, but the cited customer-support study alone cannot tell readers whether that will reduce staffing, increase service capacity or shift workers into other roles.
Further evidence would need to track employment, wages, hours and service quality across industries, alongside productivity changes. The source does not give a date for a follow-up finding or a forecast for when AI might materially change the overall service-sector workforce. For now, the employment shares establish the scale and history of the shift; the labor-market consequences of AI remain open.
Key Questions
What does “80% in services” mean?
It refers to the article’s cited ILO estimate that 79.6% of U.S. workers were employed in services in 2025. A separate BLS payroll-job calculation gives a different share because it covers a different group of workers.
Did service employment grow only after factory jobs declined?
No. The BLS series cited in the article puts services at 62% of nonfarm payroll jobs in 1939, rising to 72% in 1979, the year U.S. manufacturing employment peaked. The share kept increasing afterward.
What is Baumol’s cost disease?
It is an economic explanation for why costs may rise in activities where productivity is difficult to increase. If a service still requires much the same amount of human time while wages rise elsewhere, its costs can face upward pressure.
Does the article show that AI will eliminate service jobs?
No. It cites a study where an AI assistant raised issues resolved per hour by 15% on average among 5,172 customer-support agents. That result does not show how employment will change across the wider service economy.
Source: Thorsten Meyer AI
