In-depth analysisGrowth and productivity
France's productivity slump since 2019
Never so much employment, and an hour of work that produces less: composition, investment or measurement?
Version 2026.1 · Public data frozen on 30 September 2026 · 6 catalogue series · Lire cette analyse en français
The question
Hourly labour productivity — what one hour of work produces, on average, in France — is measured by Eurostat from the national accounts, as an index based 100 in 2015. It stood at 102.5 in 2019 and 101.7 in 2025: −0.8% in six years. Over the same period, the employment rate gained 3.1 points of percentage and the economy grew by +5.7%.
The question fits in one sentence: how can an economy employ more people than ever and produce less per hour than in 2019? Three families of explanation compete — the composition of employment, investment, measurement itself — and this page says what the series allow to be said about them.
What the data show
Fifty years of data tell a slowdown in three stages. From 1975 to 2000, hourly productivity grows by 2.61% a year on average — a France ÉcoScan calculation, average annual rate between the two endpoints. From 2000 to 2019, the pace falls to 0.89% a year. From 2019 to 2025 it turns negative: −0.14% a year. The series' peak is reached in 2020 (102.7), a year when hours worked collapsed faster than output — a lockdown artefact, not progress.
The slump proper sits between 2019 and 2022: −2.7%, from 102.5 to 99.8. Since then, productivity has been rising slowly without regaining its 2019 level. Nothing comparable happened in previous recessions: in 2009, productivity had fallen by −1.0% before rebounding the following year; in 1993, it had kept growing (+1.2%).
Chart 1 of 1 · index, 2015 = 100 · Period 1975-2025
Hourly labour productivity since 1975
Where does the slump sit in fifty years of data?
Reading
- Hourly productivity (2015 = 100)
What to see
- A steep slope until 2000, gentler after, flat since 2019.
- The 2020 peak is an artefact: hours fell faster than output.
What this does not mean
- A flat index is not a stagnant economy: output grew, but so did hours worked.
Limits · Eurostat, GDP in volume per hour worked; 2025 provisional, series revised with each campaign. Source · Eurostat — Labour productivity and unit labour costs (nama_10_lp_ulc)
What has changed
What changed after 2019 is the relationship between employment and output. The employment rate went from 66.2% to 69.3%; the employment rate of 15-24 year-olds, driven by apprenticeships, from 29.9% to 34.5% (4.6 points). The economy created jobs faster than it created value, and average productivity mechanically suffered: an apprentice or a newly hired employee produces less, per hour, than an experienced one.
Meanwhile, business investment did not weaken: 12.1% of gross domestic product (GDP) in 2019, 12.7% in 2022, 12.3% in 2024, against 11.3% on average in the 2010s. The productivity slump did not happen in an economy that under-invests — which rules out the simplest explanation.
The visible relationships
What cannot be claimed
This page does not carry hours worked, nor productivity per head, nor productivity by sector: it therefore cannot apportion the slump among its causes, only rule out those the series contradict. Nor does it carry neighbouring countries' productivity, and cannot say whether the slump is French or European.
The 2025 productivity vintage is provisional, and the whole series is revised with each national-accounts campaign: a slump of less than one per cent over six years is within the range of revisions. The sign is probably robust; the magnitude is not.
Finally, none of the three explanations excludes the others, and their sum is not measured: saying that “composition explains half” would be an invention.
The events that shed light on the path
Where we stand today
In 2025, hourly productivity is at 101.7, below its 2019 level (102.5), with growth of 0.8%, an employment rate of 69.3% and business investment at 12.3% of GDP in 2024. The scissors have not closed.
What the reader can take away
French productivity slumped because the economy hired faster than it produced — apprentices, subsidised jobs, returns to work — and because some of those hours are poorly measured. Investment is not at fault by its volume; it may be by its destination. The expected rebound from maturing cohorts is weaker than predicted, and that residual is what remains unexplained.
What the reader can keep: a record employment rate and stagnant productivity are two sides of one collective choice, that of putting more people to work at the cost of lower output per hour. An account gives access to the full cross-reading — employment, investment, measurement, chronology — and to what each of the three explanations can or cannot carry.
Sources and method
Every figure on this page is read from an identified public series, or computed from it according to the stated rule. The series are those of France ÉcoScan's catalogue, checked against their producer and frozen on the date given at the top of the page.
- Hourly labour productivity (2015=100) — Eurostat — Labour productivity and unit labour costs (nama_10_lp_ulc) · ec.europa.eu
- Employment Rate (15-64, %) — INSEE — Labour Force Survey (enquête Emploi en continu) · www.insee.fr
- Business gross fixed capital formation (% of GDP) — Eurostat — Key indicators of national accounts (nasa_10_ki) · ec.europa.eu
- GDP Growth (%) — INSEE — annual national accounts, base 2020 · www.insee.fr
- Active enterprises — Eurostat — Business demography by size class (bd_size, EBS regulation) · ec.europa.eu
- Youth Employment Rate (15-24, %) — INSEE — Labour Force Survey (enquête Emploi en continu) · www.insee.fr