Japanese manufacturing went from first in the OECD to twentieth

The Japan Productivity Center publishes an international comparison every year. The 2025 edition contains the trend line.

Japanese labour productivity per hour is $60.10, or ¥5,720 at purchasing power parity. That is 28th of 38 OECD countries. The OECD average is $79.40 and the United States is $116.50.

The movement went: 21st in 2018, 28th in 2020, 28th in 2024. Real growth was negative 0.6%, which ranks 33rd.

Per worker, Japan is at $98,344 and 29th. That is last in the G7 and 54% of the American level.

The manufacturing decline is the interesting part

More telling than the headline rank is what happened in manufacturing.

In 2000, Japanese manufacturing labour productivity ranked first in the OECD. It was 7th in 2005 and 2010, and has sat between 15th and 20th since 2015. The 2025 edition puts it at $80,411 and 20th.

An industry that fell from first to twentieth in 25 years sits at the centre of Japanese exports.

One qualification: indexed to 2019 as 100, Japan reads 102.7%, the second best figure in the G7. The recent period is not where the loss happened. It happened between 2000 and 2015.

What changed in that window

One thing that changed is that the competitive axis in manufacturing moved from the product to the experience around the product.

Once several countries could build hardware of equivalent specification, the differentiation moved into software, support, update frequency and how much effort it takes to get started.

The 0.8% figure from the previous article, the share of Japanese manufacturers crediting design for their success, describes an industry that did not register that shift while it was happening.

There is also the question of how fast software gets built. The 2026 white paper puts Japanese usage of generative AI for code generation at 7.0% against 35.4% in the United States.

In an industry where the software content of products keeps rising, that is a five times difference in the speed of producing the part that now differentiates.

Productivity is not a working hours question

Productivity discussions tend to drift toward reducing hours, but the numerator and the denominator are separate problems.

Cut hours and, if the value produced falls by the same proportion, the ranking does not move. Japan's real growth of negative 0.6%, ranking 33rd, says the numerator is not growing.

There are only a few ways to grow it.

• Produce more with the same people
• Put yourself in a position to charge more for the same thing
• Stop building things that should not be built

The third has the largest effect. The third is a decision problem, not a technology problem.

What you can check internally

The national ranking is too large to act on, so reduce it to something local.

Take everything you shipped in the last year and calculate what share is actually being used. The time spent building the rest came straight out of the numerator.

Then measure elapsed time from concept to release. Where it is long, find where the waiting happens. Usually it is not the work. It is approvals and rework.

Review both quarterly. You cannot move the national ranking. You can move these two.