Money & Finance › Economics & schools of thought › GDP: measuring the size of the economy
Productivity: output per hour of work
· Concept
Mainstream viewMost economists and practitioners agree.
How much output the economy gets from each hour of work, the main long-run source of higher wages and living standards.
See it on the Money & Finance map →How it works
Labor productivity is real output divided by hours worked. It rises with better machines and software, more skilled workers, and smarter ways of organizing work. US nonfarm business productivity grew about 2.8% a year from 1948 to 1973, slowed to about 1.5% until 1995, ran near 3% in the internet years to 2004, then slowed again. Small differences compound: at 2% a year, output per hour doubles in about 35 years; at 1%, it takes 70.
Why it matters
As Paul Krugman put it, 'Productivity isn't everything, but in the long run it is almost everything.' It caps how fast wages, profits and GDP can grow without inflation.
Think of it like
Like a management company running 1,000 units with the staff that once ran 600: same hours, more output.
Facts, rules and data
BLS publishes quarterly productivity data back to 1947. Robert Solow's 1987 quip, 'You can see the computer age everywhere but in the productivity statistics,' named the lag between invention and measured gains.
Debates and risks
Whether AI lifts productivity growth sharply, as the internet did in the late 1990s, or slowly and unevenly, is a central economic question of the 2020s.
Connections (3)
▸The AI stock boom and Big Tech concentration bets on future gains in Productivity: output per hour of workExperts disagree
The connection
AI stock prices embed an expectation that AI will make workers far more productive, lifting profits for chipmakers, cloud firms and users. If productivity growth accelerates, as in the late-1990s internet era, prices may prove justified; if gains come slowly, the boom could deflate, as the dot-com bust did in 2000 despite the internet's later payoff.
When and why
From ChatGPT's November 2022 launch onward; Big Tech committed hundreds of billions of dollars a year to data centers in 2024–26.
How solid is the link?
Company forecasts and analyst models tie valuations to AI-driven efficiency; early studies show sizable gains in tasks like coding and customer support, but little yet in national data.
Debates and risks
Estimates of AI's boost to annual US productivity growth range from under 0.1 point (Daron Acemoglu) to about 1.5 points (Goldman Sachs); nobody knows which is closer.
▸Productivity: output per hour of work drives long-run growth of GDP: measuring the size of the economyMainstream view
The connection
GDP grows from two sources: more hours worked and more output per hour. With workforce growth slowing, productivity drives most long-run gains in GDP per person. US output per hour has risen roughly fivefold since 1948, which is why living standards climbed even as average working hours fell.
When and why
Robert Solow's 1956–57 studies showed most growth in US output per hour came from technology and efficiency, not more capital per worker; he won the 1987 Nobel Prize.
How solid is the link?
BLS and BEA data: output growth equals hours growth plus productivity growth, an accounting identity, and Solow's growth accounting is standard practice.
Debates and risks
Why measured productivity growth slowed after 2005 despite smartphones and the cloud, whether from mismeasurement or a real slowdown, is still debated.
▸Creative destruction (Joseph Schumpeter) raises Productivity: output per hour of workMainstream view
The connection
Much productivity growth comes not from existing firms improving but from resources shifting: efficient newcomers grow, laggards shrink or close, and workers and capital move up. Studies of US retail found most productivity gains in the 1990s came from new, more efficient chain stores replacing older single shops.
When and why
US economy-wide, measured since the 1980s with Census business microdata; John Haltiwanger and colleagues documented the pattern.
How solid is the link?
Studies by Lucia Foster, Haltiwanger and C.J. Krizan (2001, 2006) attribute most retail productivity growth to entry and exit; manufacturing studies find similar effects.
Debates and risks
US startup rates fell for decades before a post-2020 surge in new business filings; whether that surge revives productivity growth is unclear.