By Enguerrand Artaz, Strategist, La Financière de l’Échiquier (LFDE)
The first results from the latest PISA survey, which measures 15-year-olds’ ability to apply what they have learned in reading, mathematics and science, caused a stir in France. And for good reason: France’s ranking, already unflattering in the 2022 survey, has deteriorated further. But France is far from being the only country affected by the decline in mastery of academic and functional skills. Already in steep decline in 2022 compared with the 2018 study, especially mathematics and reading skills have fallen sharply across the OECD. Between 2015 and 2025, the decline amounts to 28 points in reading and 22 points in mathematics, equivalent, according to the OECD’s convention, to between one year and one and a half years of lost learning.
The reasons cited are numerous: the impact of the Covid crisis, the ubiquity of screens, lower expectations in the transmission of core knowledge… The causes of this collapse will need to be diagnosed precisely. But it is also worth examining the potential long-term consequences of such a phenomenon, particularly from an economic standpoint. The work of researchers Eric Hanushek and Ludger Woessmann[1], as well as the 2023 update of their research published by Gabriel Heller-Sahlgren and Henrik Jordahl[2], has demonstrated the long-term correlation between a population’s cognitive skills and economic growth.
They identify two distinct mechanisms. On the one hand, there is the minimum skills base of the majority of students, which influences employability, the ability to adopt new technologies and, ultimately, average productivity. On the other, there is the upper end of the distribution, in other words the skills of the most gifted students, which matters more for innovation, research and technology entrepreneurship. The 2023 update of the research also concludes on this point: the proportion of very high-performing students is more strongly correlated with long-term growth than the share of the population reaching only a minimum competency level.
From this perspective, the good news in the PISA 2025 study lies in the fact that the share of very high-performing students has changed little compared with 2015. Even so, some skills have clearly deteriorated even among the top deciles, particularly reading comprehension. Moreover, the OECD points to a decline in the skills – in reading comprehension – most essential in the age of AI: assessing information, the ability to connect multiple sources, critical thinking. The risk appears twofold: first, lower employability and average productivity for the population, especially if the skills that are deteriorating are precisely those that would make people complementary to AI rather than replaceable by it; second, a deterioration in skills even within the highest-performing deciles, with the risk that the share of very high performers will fall, despite the fact that this group has the strongest correlation with long-term growth.
That said, a few caveats are needed. First, the PISA survey measures only part of human capital and therefore cannot be considered a comprehensive measure of future productivity. Second, the state of skills measured at age 15 is not fixed and can be made up later, during the following years and through vocational learning. Finally, a trend over a decade is not enough to produce precise calculations of long-term impact. Nevertheless, educational performance cannot be viewed solely through a social or societal lens. Its long-term economic impact is real, and recent trends in many countries call for urgent awareness.
[1] Do better schools lead to more growth? Cognitive skills, economic outcomes, and causation, Eric A. Hanushek, Ludger Woessmann, 14.07.2012.
[2] Test scores and economic growth: update and extension, Gabriel Heller-Shalgren, Henrik Jordahl, 19.01.2023
Find all our Inside articles
