Trésor-Info - Publications de la direction générale du Trésor - Productivity

Artificial Intelligence and Its Effects on Employment

2026-06-30T00:00:00+02:00

Artificial intelligence (AI) capabilities have rapidly advanced with the rise of large language models (LLMs) and generative AI, offering new opportunities for automation in every industry of the economy.

As with previous technological revolutions, AI has provoked unease: according to the Baromètre du numérique 2025, a survey on the use of digital technologies, 62% of French people think it puts jobs at risk.

AI may affect the volume of employment through two opposing channels: (i) it could replace some workers whose tasks can be automated, and (ii) it could increase worker productivity and stimulate labour demand.

Empirical studies cannot determine the total effect of AI on employment, due to a lack of temporal perspective and the fact that AI adoption by firms remains modest. However, certain trends could be linked to the emergence of AI. The first rounds of redundancies attributable to AI are mainly in the most exposed industries (finance and IT), although AI adoption is sometimes used as a cover for other reasons for redundancies. Furthermore, there appears to be a slowdown in the employment of young workers.

During previous technological revolutions, the jobs created by new technologies offset those lost, and the nature of work was deeply transformed. In light of this, public policy measures must provide support, particularly in the area of training, to workers so they can retrain when their jobs are put at risk by AI, and foster AI’s diffusion across the economy.

In a globally competitive environment, investing in AI is necessary to support competitiveness – and therefore employment – as remaining competitive will be increasingly tied to the adoption of AI.

What Issues Does the Nascent Quantum Computing Sector Raise for the French Economy?

2026-05-22T00:00:00+02:00

Quantum computing is of growing interest in scientific and economic circles, with France’s Michel Devoret being awarded the Nobel Prize in Physics in 2025, while Alain Aspect received the accolade in 2023.

Quantum computers are currently still at the prototype stage, and there are considerable lingering uncertainties about the conditions of their maturity. However, the potential associated productivity gains could be significant, with foresight methodologies – that remain uncertain – indicating an eventual 8% increase in France’s GDP.

These benefits will be reaped to greater extent in some sectors than in others, particularly those in which such technology will have the most powerful applications (e.g. logistics, finance, pharmaceuticals, chemicals). The potential of quantum computing will only be realised if businesses from these sectors adopt it.

Innovation in quantum computing is currently driven by major digital tech firms and a thriving network of start-ups. The dynamic between these stakeholders spurs competitiveness and investment. However, moves towards consolidation are already underway, with several recent acquisitions. In the long term, this trend could create barriers to entry and foreclosure effects similar to those seen in the AI computer chip sector, restricting the market’s competitiveness and slowing down the technology’s dissemination.

France boasts solid assets within the quantum computing sector, including world-class laboratories, one of the strongest patent portfolios in Europe (relative to GDP) and several leading start-ups. The National Quantum Strategy, launched in January 2021 and financed under the France 2030 Investment Plan, was set up to bolster this ecosystem. There are nevertheless still major obstacles, such as a financing gap that is widening with non-European start-ups.

Beyond Health Impacts: the Heavy Economic Toll of Air Pollution

2026-02-05T00:00:00+01:00

Air pollution is a major public health issue. Fine particulate matter is associated with more than 40,000 deaths each year in France. It also interacts with other pollutants to trigger or worsen many diseases affecting organs including the lungs, heart and brain. It is estimated to account for 20% of new asthma cases in children.

These health impacts carry a high socio-economic cost, related not just to direct healthcare expenditure (e.g. hospitalisation and medicine costs) but also to impaired quality of life and lost production due to sickness absence.

In addition to these effects on mortality and morbidity, recent research papers have shown that air pollution also affects cognitive and physical abilities, resulting in lower productivity. These impacts can be seen across a broad array of occupations not just involving physically demanding work, in agriculture and industry for example, but also involving a cognitive aspect, such as financial services, call centres, and even umpiring in sport.

In addition to these direct short-term impacts, air pollution may also be undermining future productivity by reducing academic achievement and hampering innovation.

The Economic Impact of the Performance Levels of the French Education System

2025-12-02T00:00:00+01:00

The academic performance levels of students in France are declining over time and in comparison with other countries. National assessments show a slump in academic achievement in mathematics and French over the past 30 years. A survey by the Programme for International Student Assessment (PISA) confirms this decline, which is faster than in other countries, although performance levels in France remain close to the OECD average and, out of OECD countries, are among the most strongly correlated with social background and reveal significant gender gaps.

Reversing the trend in educational performance levels would be a source of long-term growth for the French economy. Indeed, a large proportion of the productivity gains in developed economies can be explained by the accumulation of human capital, mainly acquired during initial training. Reducing gender and social inequality would increase the pool of potential innovators and skilled workers.

Assessment of education policies is essential, to identify the most effective actions and optimise their socio-economic impact. In particular, the literature shows that increasing the teacher-student ratio has a significant impact on academic performance levels. Teacher training and in-service professional development, as well as the appropriate and judicious use of digital resources, which are still underutilised in France, can also have a positive effect on learning.

Additional resources have been allocated to the education system in recent years (an 11% increase in domestic education expenditure in constant euros per student in primary education between 2017 and 2023). Halving of class sizes has been gradually implemented since the start of the 2017 academic year for students in the first and second years of primary school and in the final year of nursery school in priority education networks, to target students from disadvantaged social backgrounds in the acquisition of core knowledge.

Disparities in Farm Income

2024-10-15T00:00:00+02:00

Income disparities between farmers are significant and depend on economic parameters such as farm size and production type.

In aggregate terms, real farm income per work unit has been growing on average since 2010, despite a dip in 2023 following a record year in 2022. Farmers’ incomes are subject to acute cyclical fluctuations due to climatic conditions and global food markets. However, farm subsidies help to stabilise these incomes to some extent.

Although the median standard of living among farm households is similar to that among all working French households, the headline figure masks significant disparities between farms and reflects a higher workload:

– In 2020, the poverty rate among people in farm households was 16%, ranging from over 20% for livestock farmers to 12% for arable farmers. By comparison, the poverty rate among the French population as a whole was 14%.

– In 2022, farmers worked an average of 15 hours more per week than the general population and were more likely to report working evenings, nights and weekends.

Farmers, most of whom work on individual holdings or with a small number of partners, tend to pay themselves little, preferring instead to invest in equipment and machinery in order to build up substantial business assets and achieve productivity gains.

There have been significant productivity gains in French agriculture over the past 60 years, but the lion’s share of the benefits have flowed to the downstream sector and consumers.

The Economic Implications of Artificial Intelligence

2024-04-02T00:00:00+02:00

Artificial intelligence (AI) refers to a set of techniques which enables machines to simulate human intelligence. Its development is a technological revolution which, much like with previous revolutions of this kind, could generate profound economic changes. While research to quantify the impact of AI is still in the exploratory stage, such work provides some preliminary insights.

On a macroeconomic level, it is too early to empirically discern an impact on growth, but some initial microeconomic studies suggest that certain specific applications of AI have a significant positive impact on individual worker productivity. In a given job, these gains benefit the least productive workers the most, allowing them to catch up to their most productive peers. However, the impact of AI on business productivity has been found to be modest for the time being. This may be due to companies’ still limited and uneven adoption of AI, although there is more widespread adoption among large companies and digital firms.

The theoretical impact of AI on employment is uncertain. In the short term, this impact will depend on the speed at which AI is deployed, the shift of certain occupations towards AI-complementary tasks and the reallocation of labour towards occupations in growing demand. Furthermore, initial empirical estimates indicate that the tasks and occupations impacted by AI will not be the same as those affected by previous technological revolutions. Skilled occupations are expected to be more impacted by AI due to its ability to perform abstract, non-routine tasks, whereas the previous waves of automation and computerisation had impacted unskilled occupations and mid-level occupations, respectively.

These various findings point to the need to strengthen science curricula in primary and secondary education and AI curricula in higher education, to focus on continuing training for occupations affected by AI and to remove certain barriers to the diffusion of artificial intelligence, particularly by adapting competition policy to its particular qualities.

The EU Single Market, a Driver for Economic and Trade Integration

2024-03-05T00:00:00+01:00

The single market plays a central role in European integration, and its purpose is to ensure the political stability and economic prosperity of its Member States. As of 2023, the market included 27 countries, 23 million businesses and nearly 450 million people, making it the world’s largest developed market. Within this market, economic and financial relations between Member States are very close (see Map on this page).

The EU single market has met its initial expectations, promoting innovation, boosting productivity and enabling convergence in the European Union (EU). The strengthened internal economic relations and economic convergence made possible by the single market have helped significantly raise incomes and the standards of living in EU Member States.

Between 1984 and 2019, EU economies’ goods export structures have become more similar, most likely reflecting the development of intra-industry trade. The single market has therefore not led to increased relative specialisation within the EU, where each Member State would have specialised in different sectors.

Japan's Growth Drivers

2023-04-27T00:00:00+02:00

After a dazzling economic catch-up in the post-war years, Japan has been experiencing low growth levels since the noughties – which distinguish it within the OECD – despite the efforts of successive governments to generate fresh momentum. Although it posted the highest GDP per capita growth in the G7 until the end of the 1990s, Japan, together with Italy, is the country in which growth has fallen the most on average since then (see Chart).

Initially buttressed by a rapid rise in capital stock, Japan’s growth has only been sustained by the workforce and total factor productivity since 2010, and this leverage is now weakening. The population is ageing and declining, and hourly labour productivity is growing meekly, especially in SMEs which are predominant in the manufacturing base and are less productive than large enterprises. Too much investment is earmarked to offset capital depreciation – a sign that this factor is poorly allocated. Potential growth, which stood at around 4.0% in 1990, is now estimated at less than 0.5%.

“Abenomics”, which were unveiled in 2013, aimed to expand the workforce and improve working conditions but they were only partly successful. Subsequent governments have focused on investments for the future to boost supply and lastingly lift productivity, through the digital transformation and the low-carbon transition initiated as from 2020, followed by the plan promoting a “new capitalism” which was introduced in 2021 and puts priority on entrepreneurship and vocational training.

Japanese growth will continue to be hampered by demographic trends and this calls for continued support for the employment of older workers, immigration and the birth rate by means of bold structural reforms. Japan also has headroom for improving the allocation of capital and the rollout of new technologies, a sector with strong growth potential.