Who Is Losing Factory Jobs to China?

Nobody Since 2013

Richard Baldwin 

 

 

Introduction.

The title sounds like a no-brainer, right? Everybody knows China has, for decades now, been stealing jobs from America and Europe and Japan, right? China shock 2.0 will be even worse than 1.0 (Autor et al, 2013) and all that, no? Well, it turns out that once again everybody hasn’t been looking at the data. When will this ‘everyone’ wise up?

Today’s Factful Friday, a short one since it’s mid-August and you all should have better things to do, shows the numbers using the OECD’s Trade in Employment dataset.

Here’s the spoiler for those in a hurry to get to the beach, bike or barbecue. The job theft story doesn’t blatantly violate the data between 2001 and 2013. China gained about 40 million factory jobs and the advanced economies lost about 15 million.

But those trends both reversed. Since 2013, China has lost 18 million manufacturing jobs – about 12% of its factory workforce at the peak. The advanced economies, as a group, have lost essentially none (+0.2 million). The jobs China shed reappeared in other emerging economies: Viet Nam, Indonesia, Nigeria, Mexico, Pakistan, India and their peers. They added 22 million new factory jobs. ILO (2025) has corroborated the trends if not the exact numbers.

China’s factory employment peaked in 2013 while advanced economies’ flatlined.

I first looked at this question in my Factful Friday of 22 December 2023, using the vintage of the OECD’s Trade in Employment data that ended in 2018. The 2025 edition extends the series to 2022 and revises the history, so today’s column is a two-and-a-half-years-on update – including one place where the new data overturns what I reported then. All figures below are from that single, consistent 2025 vintage.

 

 

Chart 1. Manufacturing employment in millions of persons, 1995–2022. Advanced economies follow the IMF World Economic Outlook definition (36 economies in the sample). Source: OECD Trade in Employment (2025 edition); author’s calculations.

The chart shows manufacturing employment in millions of persons for three groups: China, the advanced economies (IMF definition), and all other emerging economies in the dataset.

The orange line is the story. China’s manufacturing employment climbed from 108 million in 1995 to a peak of 152 million in 2013, then fell to 129 million by 2018 before a partial recovery to 134 million in 2022. Net change since 2013: −18 million jobs, a fall of about 12%.

The teal line is the counterpoint. The advanced economies lost manufacturing jobs steadily from 75 million in 1995 to about 60 million in 2010 – and then stopped losing them. From 2013 to 2022 the group’s total moved from 60.1 million to 60.3 million. A decade of net stability, spanning 36 economies, the trade war, and the pandemic.

The grey line never bends. Other emerging economies added manufacturing jobs in every sub-period: 186 million by 2022, up 22 million since 2013.

The reversal is a 2013 phenomenon, not a 1995 one.

The chart splits the 27 years at 2013 and shows the change in factory jobs for each group in each window.

 

 

Chart 2. Change in manufacturing employment, millions of jobs, over 1995–2013 and 2013–2022. Source: OECD Trade in Employment (2025 edition); author’s calculations.

Over 1995–2013 the bars match the received view: China +44 million, advanced economies −15 million, other emerging economies +38 million. Over 2013–2022 the pattern inverts for the first two groups: China −18 million, advanced economies zero. Only the third group’s bar keeps its sign.

In share terms, China held 40% of the world’s manufacturing jobs in 2013 and 35% in 2022. The advanced economies’ share was 16% in both years. The other emerging economies went from 44% to 49% – they now hold nearly half of the world’s factory jobs (UNIDO 2024).

Where the jobs went.

The chart ranks the largest national changes over 2013–2022. China’s −18 million dwarfs every other move in the dataset. The next-largest losses are Japan (−1.6 million), Brazil (−0.7 million) and Ukraine (−0.5 million). Note what is not on the losers’ side: the United States gained about 0.8 million manufacturing jobs over the period, and Germany was flat. Japan was the second biggest loser in absolute numbers at about −1.6 million.

 

 

Chart 3. Largest changes in manufacturing employment by country, 2013–2022, millions. Source: OECD Trade in Employment (2025 edition); author’s calculations.

The gainers are a roll of populous emerging economies: Viet Nam +4.4 million, Nigeria +3.5 million, Indonesia +3.5 million, Mexico +2.2 million, Pakistan +2.0 million, with Türkiye, India, Morocco, Egypt and Thailand each adding around 1 to 1.5 million.

An interesting note, maybe for a future Factful Friday, is how not all these gainers are classic industrial exporters. Some of them, I’m thinking India and Nigeria, may be doing it on the strength of their domestic markets. The ones who are industrialising by joining GVCs are benefiting from the ‘Flying Geese’ pattern (Akamatsu 1962) of moving labour-intensive stages out of China as its wages rise.

The dataset counts jobs; it does not say which factory took over whose work. The next chart gets closer to that question.

Export-linked factory jobs show the same reversal, more sharply.

The chart shows trade-linked manufacturing jobs – factory employment supported by foreign final demand, the dataset’s measure of jobs that exist because of exports.

 

 

Chart 4. Manufacturing employment embodied in foreign final demand (trade-linked factory jobs), millions of persons, 1995–2022. Source: OECD Trade in Employment (2025 edition); author’s calculations.

On this measure China’s peak came earlier, in 2008, at 54 million. By 2013 it was 48 million; by 2022, 40 million – a fall of 8 million, or 16%, since 2013. China’s share of the world’s export-supported factory jobs fell from 37% in 2013 to 30% in 2022, even as its share of world manufactured exports kept rising. Other emerging economies added 14 million trade-linked factory jobs over the same period and now hold just over half of the world total. The advanced economies were, again, flat at about 26 million.

The China numbers are worth a second look because they pair with a fact from my Factful Friday of 24 July 2026: China’s share of world manufactured exports rose strongly after 2013. Rising export share, falling export-linked factory employment. The two are reconciled by productivity: Chinese factories produce more export value per worker every year, so the same or larger export volume supports fewer jobs.

Lies, damn lies and statistics.

Four caveats.

First, the jobs are counted in persons, not full-time equivalents or hours. A shift towards longer hours per worker – or the reverse – would not show up.

Second, the TiM figures are model-based estimates built on national accounts and inter-country input-output tables, not a headcount survey. The trade-linked split in Chart 4 in particular inherits the assumptions of the input-output framework. And what counts as a “manufacturing job” is itself a moving target: a good deal of measured decline in rich countries is contracting out and reclassification rather than lost activity (Fort et al, 2018).

Third, Chinese official employment data carry a standing asterisk. The direction and rough size of the post-2013 decline are corroborated by independent work, but any single year’s level should be treated as indicative. Policy-based evidence making is not as uncommon as some may believe.

Fourth, the endpoint is 2022 and that was a pandemic-recovery year. The 2018–2022 stretch includes the trade war and Covid; the partial rebound in China’s line over 2020–2022 may be revised.

Summary and concluding remarks.

Since 2013, the answer to the title question is “China.” China lost something like one eighth of its factory jobs; the advanced economies lost none.

World factory employment did not fall; it edged up by 4 million. The jobs moved down the wage ladder, to Viet Nam, Indonesia, Nigeria, Mexico, Pakistan, inter alia. They, together, added 22 million. On the trade-linked measure the reversal is sharper still: China’s export-supported factory employment is down 16% since 2013 and its share of the world total has fallen from 37% to 30%.

None of this makes China a small manufacturer. It still employs 134 million factory workers – more than twice the advanced economies combined – and its share of world manufactured exports has kept rising. What has changed is the direction of travel. China’s factory employment is now doing what the advanced economies’ did decades earlier: shrinking as productivity rises and as labour-intensive stages migrate to lower-wage locations.

Closing remarks.

The policy debate in the advanced economies is still organised around a picture in which factory jobs drain from the rich world to China.

It is time to change the metric, not the debate.

China is, has, and probably will continue to eat the lunch of G7 manufacturers, but this shows up in shifting production, not shifting jobs. The reason is simple. China is experiencing very rapid productivity growth that has opened a huge alligator-jaws gap between its factory output and its factory jobs.

And that’s it for another Factful Friday!

References.

Akamatsu, K. (1962). A historical pattern of economic growth in developing countries. The Developing Economies, 1(s1), 3–25.

Autor, D. H., Dorn, D., & Hanson, G. H. (2013). The China syndrome: Local labor market effects of import competition in the United States. American Economic Review, 103(6), 2121–2168.

Baldwin, R. (2023). Where in the world are manufacturing jobs going? Factful Friday, 22 December 2023. (The previous vintage of this analysis; partially revised by the 2025 data.)

Baldwin, R. (2026). Manufacturing deglobalisation. Factful Friday, 24 July 2026. (China’s rising share of world manufactured exports.)

Fort, T. C., Pierce, J. R., & Schott, P. K. (2018). New perspectives on the decline of US manufacturing employment. Journal of Economic Perspectives, 32(2), 47–72.

OECD, Trade in Employment (TiM), 2025 edition – employment and employment embodied in foreign final demand, total manufacturing (ISIC C). Author’s calculations. Figures for 2020–2022 are preliminary.

International Labour Organization. (2025). ILOSTAT database: Employment by sex and economic activity, ILO modelled estimates [Data set]. Geneva: ILO.

Rodrik, D. (2016). Premature deindustrialization. Journal of Economic Growth, 21(1), 1–33.

United Nations Industrial Development Organization. (2024). Industrial development report 2024. Vienna: UNIDO.

Annex: definitions and vintage notes.

Groupings. Advanced economies follow the IMF World Economic Outlook classification; 36 of them appear in the TiM country list (including Korea, Chinese Taipei, Singapore, Israel and the euro-area Baltics and central Europeans such as Czechia, the Slovak Republic, Slovenia and Croatia, but excluding Poland, Hungary, Romania and Bulgaria, which the IMF classes as emerging). China is mainland China only. “Other emerging economies” is every remaining economy in the dataset, including India, Indonesia, Brazil, Mexico, Nigeria, Pakistan and Viet Nam. The dataset’s country sample covers the large majority of world manufacturing employment but is not the full world; totals are for the TiM sample.

Vintage. The 2023 column used the TiM edition ending in 2018, which showed China losing 22.5 million manufacturing jobs over 2013–2018 and the world losing about 20 million over the same window. The 2025 edition puts China’s 2013–2018 loss at 23 million – essentially confirming it – but adds a 5-million partial recovery to 2022 and revises world totals so that the global decline disappears. All figures in this column are from the 2025 edition only; no splicing across vintages.

Measures. “Manufacturing employment” is persons employed in ISIC Rev.4 section C. “Trade-linked” jobs are domestic employment embodied in foreign final demand (TiM measure FFD_DEM), i.e. jobs supported by exports directly and through domestic supply chains.