Automation and output
Do robots boost productivity? US manufacturing productivity says not lately
US manufacturing productivity grew 3.4% a year for two decades. Then, over the era robots poured onto factory floors, it turned negative.
The precise figures: labor productivity growth ran at 3.4% a year from 1987 to 2007, then at negative 0.5% a year from 2010 to 2022, a slowdown of 3.9 percentage points. Total factor productivity growth fell from 1.4% a year to 0.1%.
This page traces those figures to the analysis they come from and is careful about what they do and do not prove about robots specifically.
Data covers US manufacturing labor productivity and total factor productivity, 1987 to 2022. Last reviewed by a human editor before publication.
The figures and where they come from
Each figure is rated for how safely you can cite it today. Ratings judge current usability, not whether a number was ever correct.
| Figure | What it is | Source | Citation Confidence | Notes |
|---|---|---|---|---|
| +3.4% per year | Manufacturing labor productivity, 1987 to 2007 | [A] | High | Strong productivity growth in the two decades before the financial crisis. |
| -0.5% per year | Manufacturing labor productivity, 2010 to 2022 | [A] | High | Measured labor productivity growth turned negative in the era automation accelerated. The source states it as -0.5 percent per year. |
| 3.9 points per year | Size of the slowdown | [A] | High | The drop from +3.4% to -0.5% is a slowdown of 3.9 percentage points per year, a large reversal. |
| +1.4% per year | Total factor productivity, 1987 to 2007 | [A] | Medium | TFP captures output not explained by more labor or capital, closer to a pure efficiency measure. |
| +0.1% per year | Total factor productivity, 2010 to 2022 | [A] | Medium | TFP growth fell to roughly zero over the automation era. |
Why the numbers disagree
The gap is between the promise and the aggregate. At the level of a single line, a robot can genuinely raise output per hour. At the level of US manufacturing as a whole, measured labor productivity growth went negative after 2010, and total factor productivity growth fell to near zero. Whatever robots are doing, it is not showing up as a productivity surge in the sector's aggregate numbers.
Economists do not fully agree on why, which is why the source calls it mysterious. Candidate explanations include mismeasurement, the drag of the financial crisis, weak capital investment, and shifts in what US factories make. Robots are one factor among many, and the data cannot isolate their contribution.
This is the honest limit of the claim. 'Robots boost productivity' is well supported for specific tasks and cells. 'Robots have boosted US manufacturing productivity in the aggregate' is not supported by the productivity statistics for the 2010 to 2022 period, and anyone citing a broad productivity gain from automation should reckon with that.
How to cite these figures
If you cite a productivity gain from robots, keep it at the level it is measured: a specific task, cell, or plant, not US manufacturing as a whole.
For the aggregate picture, cite the measured figures: manufacturing labor productivity growth of 3.4% a year from 1987 to 2007 versus negative 0.5% from 2010 to 2022, and name the source.
Avoid claiming automation drove an aggregate productivity boom in the 2010s. The sector's productivity statistics show the opposite, and the causes are debated.
Where people go wrong
Generalizing a single-cell productivity gain to the whole sector. The aggregate data does not show it for 2010 to 2022.
Blaming or crediting robots for the slowdown. The causes are debated and the data cannot isolate any single factor.
Treating labor productivity and total factor productivity as the same thing. They are different measures; both slowed, but they capture different things.
How we checked
The figures trace to a Federal Reserve Bank of New York analysis, which draws on Bureau of Labor Statistics productivity data. We cite the New York Fed page because the BLS productivity pages block automated verification, while the New York Fed page carries the same figures in a form our citation check can fetch and confirm.
We fetched the source and confirmed the labor-productivity and total-factor-productivity figures appear in the text, including the negative 2010 to 2022 figure and the 3.9-percentage-point slowdown.
We do not claim to explain the slowdown, because the source does not. The value of the page is that it puts a measured, sourced brake on the broad 'robots boost productivity' claim, at the aggregate level where that claim is weakest.
Full source list
Primary sources, with live links. Every figure above traces to one of these.
- [A]Federal Reserve Bank of New YorkJuly 2024
Federal Reserve Bank of New York, Liberty Street Economics, "The Mysterious Slowdown in U.S. Manufacturing Productivity"
https://libertystreeteconomics.newyorkfed.org/2024/07/the-mysterious-slowdown-in-u-s-manufacturing-productivity/
Common questions
- Do robots increase productivity?
- For a specific task or cell, a robot can raise output per hour. But US manufacturing labor productivity growth went from 3.4% a year before 2007 to negative 0.5% from 2010 to 2022, so an aggregate productivity surge from automation does not show up in the sector's measured data.
- Why did manufacturing productivity slow down?
- Economists do not fully agree, which is why the source calls it mysterious. Candidates include mismeasurement, the financial crisis, weak investment, and changes in what US factories make. The data cannot isolate the role of robots.
- Is this proof robots do not work?
- No. It is evidence that the broad claim of an aggregate productivity boom from automation is not supported for the 2010 to 2022 period. Task-level gains are real and separate.
- What is the difference between labor productivity and total factor productivity?
- Labor productivity is output per hour worked. Total factor productivity is output not explained by more labor or capital, closer to pure efficiency. Both slowed sharply over the automation era.
More data, traced to source
- Robots multiplied. US factory output barely moved. What the numbers show
The world's operating robot fleet passed 4.6 million while US industrial production sits just above its 2017 level and runs below capacity. Automation is not tracking output the way people assume.
- 'One robot destroys 5.6 jobs': what the study really measured
The viral statistic that one robot destroys 5.6 jobs comes from a real, careful study. But the 5.6 figure is a local-labor-market estimate, not a claim that each robot nationwide eliminates 5.6 jobs. Here is what the researchers actually measured.
- How many US firms actually use robots? Far fewer than you would think
The story is that everyone is automating. The US Census Bureau's own firm survey found only about 2% of US firms use robotics. A newer Census release puts it far higher. The gap is the whole story.
- A cobot's measured productivity gain: 10%, not a multiple
Collaborative-robot marketing implies large throughput gains. A time-studied assembly cell measured the actual gain at 10%. Here is the study, and why a measured number beats a marketed one.
- US manufacturing jobs: 17.1 million then, 12.6 million now
US manufacturing employed 17.1 million people at the start of the century and 12.6 million now. Here are the figures behind 'manufacturing is dying' and 'manufacturing is back,' traced to Congress's own research service.