Working Hours: the productivity converged, the time didn't
The rich world shortened the working year for 150 years, then the US stopped and Europe didn't. Americans now work 454 hours a year more than Germans — while producing almost exactly the same amount per hour.
For a century and a half the rich world shortened the working year, relentlessly and almost everywhere. Then, around 1980, the United States stopped. Since then it has shed 52 hours a year — about one hour a week — while Germany shed 421 and France 289. An American now works 454 hours a year more than a German: about eleven forty-hour weeks. It has not always been that way round. In 1950 Germans worked more, by 405 hours; the lines crossed in 1973.
The obvious inference is that Europe chose leisure and pays for it in output. The second dataset says otherwise. German output per hour was a quarter of the American level in 1950 and is 98.8% of it today — $82.50 against $83.53, a difference indistinguishable from zero in any PPP comparison. So a German worker produces about 74% as much per year as an American and about 99% as much per hour. Essentially the whole income-per-worker gap is the length of the working year, not the productivity of the work. Both countries got the productivity growth. One took it as income, the other took it as time.
Two caveats to carry. These are hours per worker, not per full-time worker, so part of the cross-country gap is more people working part-time rather than everyone working less — and we could not measure that here. And nothing on this page identifies a cause: working-time regulation, union density and industry mix are all plausible drivers, and none of them is in the data.
Data: Feenstra et al., Penn World Table (2025) and Huberman & Minns (2005), via Our World in Data (CC BY).
Then it stopped — but only in some places
Since 1980 the United States has shed 52 hours — about one hour a week, across 43 years — and it did not even fall monotonically: hours rose from 1,840 in 1980 to 1,888 in 2000 before drifting back. Over the same span Germany shed 421, France 289.
The gap is now 454 hours, roughly eleven forty-hour weeks. It has not always run this way: in 1950 Germans worked 405 hours more than Americans, and the lines crossed in 1973.
Two reputations invert here. Japan — the country whose language gave us karoshi — has worked fewer hours per worker than the United States every year since 1998, and is 135 below it now. South Korea has fallen 37% from its 1969 peak of 3,040 hours, the largest reduction in the dataset.
@andrewdun · 1h ago
Meanwhile, output per hour converged
GDP per hour worked, PPP-adjusted, in constant 2021 international dollars. German output per hour was 25% of the American level in 1950, 67.5% in 1980, and 98.8% in 2023 — $82.50 against $83.53. It first passed 95% in 1994 and has stayed there.
The honest reading of a 1.2% gap in a PPP series is indistinguishable, not "Germany is 98.8% as productive". Ireland and Norway are deliberately absent: Irish GDP per hour is distorted by multinational profit-shifting and Norway's by petroleum rents.
@andrewdun · 1h ago
The gap is time, not talent
Both measures on one surface — hours on the left axis, output per hour on the right. The lines do opposite things. Hours diverge from parity in the early 1970s to a 454-hour gap. Output per hour converges from 4:1 to level.
Multiply them out and a 2023 American worker produces roughly $149,000 a year against a German's $110,000 — Germany at about 74%. Per hour, Germany is at 99%. Essentially the entire income-per-worker gap is the length of the working year. The familiar story — that Europe is poorer because it is less productive — runs backwards on this measure. Both places got the productivity growth; they spent it differently, one as income and one as time.
The caveat that matters most: this is hours per worker, not per full-time worker. A country can cut its average by moving people into part-time work without anyone's week getting shorter, and part-time shares differ a lot between these countries. Some of the gap is that. And nothing here identifies a cause — working-time law, union density, vacation mandates and industry mix are all plausible and none of them is in this data.
@andrewdun · 1h ago
Discussion of this dashboard
Threads here are about the dashboard as a whole — its framing, what's missing, whether the constituent indicators tell a coherent story. Per-indicator discussion lives under each panel above.
Log in to start a discussion about this dashboard.
No discussions yet. Be the first.
Annual hours worked per worker. In 1870 an American worked about 3,096 hours a year and a German about 3,284. By 2023: 1,789 and 1,335 — falls of 42% and 59%.
This is the baseline everything else is measured against, and it is the part nobody disputes. Note where most of it happened: roughly two-thirds of the American decline was already done by 1950. Pre-1950 figures are historical reconstruction (Huberman & Minns) rather than measurement — directionally solid, precisely soft.
@andrewdun · 1h ago