Between 2010 and 2024, France’s share of global scientific publications fell from 3.6% to 1.9%, according to the Ministry of Higher Education, Research and Innovation. Over 2020-2024, 63% of its publications bear the signature of at least one foreign researcher, a level comparable to Germany. This internationalization of co-publications contrasts with the decline in its citation audience and its global share of publications.
The Essentials
- France’s share of global publications fell from 3.6% to 1.9% between 2010 and 2024, a reduction of nearly half (MESRI 2026).
- The index of French scientific audience declined 33% over the same period: French research represents a smaller share of global citations than in 2010, although this indicator alone is insufficient to establish its role in structuring research agendas.
- 63% of French publications are co-signed internationally, a level close to Germany’s. Strong international cooperation coexists with a decline in audience and global share of publications, without causal evidence concerning research objects or standards.
- Carl Benedikt Frey, in his work on the institutional conditions for progress, shows that expanded collaboration and the capacity to orient the scientific agenda are two distinct dynamics that can diverge.
- Between now and 2030-2040, France must decide whether to convert its international integration into a lever of influence or continue to contribute to research whose questions are defined by others.
A Country That Collaborates More but Weighs Less on Agendas
France signs a high proportion of articles with foreign partners—63% over 2020-2024. This figure marks an increase from 2010, when international co-publications accounted for 49%. The 63% figure for international co-publications is often presented as a sign of vitality. It is, partly. It measures the openness of French laboratories, their attractiveness for partnerships, their ability to integrate into large consortiums.
But it does not measure influence.
The scientific audience index is another revealing indicator. It measures the extent to which French articles are cited by other researchers—that is, the extent to which they nourish global research. This index declined 33% between 2010 and 2024. Such a decline says something other than a simple loss of market share: it indicates that French research is less often the reference point from which other work is built.
These two sets of data tell the same story with two faces. France is very present in international co-publications, with a progression of this integration between 2010 and 2020. However, this increase in cooperation does not establish an increased capacity to orient networks. It signs articles with the world in positions of variable responsibility within consortiums. The data show a decline in global share and audience, without allowing precise characterization of the roles occupied.
Economists of Progress Facing This Kind of Setback
Carl Benedikt Frey, in How Progress Ends, argues that scientific progress depends on the capacity of institutions to orient and deploy innovation at scale, not on a general property of science. A nation can maintain active networks, mobile researchers, and frequent collaborations while losing the capacity to define what its own laboratories explore.
International cooperation does not substitute for this capacity for orientation. In certain configurations, it can even erode it: researchers align with the priorities of better-funded partners rather than pursue their own research objects.
This analytical framework illuminates the French case differently than a simple decline. France is not disconnecting from global science. It is strongly integrated into international co-publications, with its share rising from 49% in 2010 to approximately 65% in 2021, then settling at 63% over 2020-2024 according to published series. France maintains and strengthens a high rate of international co-publications while its aggregate audience indicators and global share decline. The 33% decline reflects a decrease in France’s share of global citations, without allowing inference that research would follow rather than precede.
A different reading, put forward by economists attentive to the effects of international specialization, such as those who work on knowledge value chains, would nuance this picture. In this perspective, the international division of scientific labor is not necessarily unfavorable to a country that concentrates on high-quality niches and accepts losing shares in the most standardized segments. The question then becomes whether French decline is strategic or imposed. Available data do not allow categorical judgment, but the simultaneous decline of the audience index in domains where France historically had strong presence suggests erosion rather than chosen specialization.
The Rise of Other Actors Changes the Environment
The global context in which this relative decline occurs has changed profoundly since 2010. China has multiplied its volume of publications by a considerable factor over the period, and its researchers now occupy a central position in many disciplines, notably in artificial intelligence, materials science, and chemistry. The United States maintains its domination in life sciences and computer science. India is rising. In this environment, maintaining one’s share requires growing effort, and budgetary data may constitute contextual elements but alone do not causally explain the observed decline.
But budget volume is not the only variable. The DIGITAfrica program, for example, shows that an explicit strategy for orienting research objects can compensate for a more modest institutional base, provided that researchers pursue their own questions and build consortiums around them. This counter-example, at continental rather than state scale, illustrates that scientific dependence can be countered by deliberate choices: the capacity for orientation is not reserved for the largest powers.
The question is therefore not whether France can rival China or the United States volume for volume. It cannot, and attempting to would be a misallocation. The question is whether it can maintain, in the domains it chooses, a position as a definer of problems rather than a solver of problems defined elsewhere.
Co-publications: Asset or Dependence Depending on the Position Occupied
63% international co-publications is a high figure but not exceptional in European research. Germany reaches a comparable level. What differentiates research systems at this level of cooperation is the position that national researchers occupy in consortiums: coordinators, corresponding authors, protocol architects, or contributors to a single line of data among others.
Available data on the corresponding author—the researcher who bears intellectual responsibility for the article and manages the relationship with the journal—constitute a valuable signal on this point. They are rarer and less systematically published than simple co-publication volumes. MESRI data document a decline in audience and a high level of co-publications, without establishing an association with project leadership.
This shift in position has concrete effects on scientific autonomy. A researcher entering a consortium under the aegis of a better-funded American or Chinese institution often adopts, often without conscious awareness, the priorities, methods, and questions of their principal partner. Automated science amplifies this dynamic: the platforms and models that structure AI-assisted research are predominantly developed outside France, which adds another layer of dependence on tools and the biases they encode.
Available Choices for France by 2035
The two scenarios opening to France for the next ten to fifteen years are logically symmetrical, but their likelihood depends on decisions that can still be made.
In the first scenario, France orients its co-publications toward consortiums where it pursues priority research objects. It builds national programs of excellence chairs in specific domains, conditioning excellence on the researcher actually directing the co-published project rather than merely contributing to it. In this case, the decline in France’s share of global publications does not necessarily translate into a loss of influence on scientific agendas. The quality of position within networks partially compensates for lesser volume.
In the second scenario, France continues to increase its co-publication rate without explicit strategy on domains where it intends to set standards. The audience index continues to decline. Researchers integrate further into networks, but in secondary positions. French contribution gradually becomes a human and infrastructural resource serving agendas defined by others. This drift would be all the more difficult to reverse as young researchers would form their ambitions and work habits in configurations where direction comes from outside.
The most reliable signal for distinguishing these two trajectories is not publication volume. The proportion of international co-publications where a researcher affiliated with a French institution is corresponding author or project coordinator is a more precise indicator. The evolution of the audience index in domains where France has explicitly defined national priorities is a second one. These two signals, tracked over five years, would indicate whether the strategy of influence through position quality works or whether French relative decline consolidates.
Available Reforms, and What They Really Demand
Two levers are available without deep institutional reform. The first is to reform the evaluation criteria for laboratories and researchers. Today, the volume of international co-publications is valued as such. Valuing instead the capacity to orient agendas—that is, to be corresponding author, to direct projects, to define protocols—would change incentives without increasing budgets. This lever is underused.
The second lever is concentrating available resources on domains where critical mass already exists in France. Dispersing investments in hopes of maintaining a presence in all scientific fields accelerates relative decline in each. Concentration on a limited number of strategic domains, creating conditions where French researchers are numerous and well-funded enough to define questions rather than answer those of others, is the wager that comparable-sized countries have made with measurable results. The Netherlands has thus maintained scientific influence disproportionate to its size by concentrating resources on a restricted number of excellence domains.
This type of choice is politically difficult to assume: it requires explicitly accepting to forgo presence in certain fields. Yet between 2010 and 2024, aggregate indicators document a decline in French audience and global share, without allowing conclusion on all domains or on the intentionality of public decisions. A chosen and concentrated loss would be less costly in influence than an imposed and dispersed loss.
France still possesses the scientific critical mass, institutions, researchers, and networks to undertake this readjustment. The question is whether the actors who decide evaluation criteria, funding priorities, and attractiveness policies will treat this signal as a governance urgency or as one statistic among others in the following annual report.
Sources
- Ministry of Higher Education, Research and Innovation, The State of Higher Education, Research and Innovation in France 2026, publication.enseignementsup-recherche.gouv.fr
- Carl Benedikt Frey, How Progress Ends: Technology, Innovation, and the Fate of Nations, Princeton University Press, press.princeton.edu
- OECD, Main Science and Technology Indicators, comparative data on global publication shares and citation indices by country
