France’s educational system spends heavily but delivers unequal outcomes for pupils, mainly driven by differences in their social background. A targeted, low-cost policy framework, combining tutors and adaptive technology, could help to change that without new legislation or major fiscal commitment.
France's educational system has a misdirection problem. Despite spending above the OECD average, it produces some of the steepest socio-economic attainment gaps, gaps that open before primary school and widen from there onwards (see Figure 1).
Previous reforms have relied on the wrong tools: more staff, blanket technology rollouts and centralised mandates. They have not managed to solve the issue.
Evidence from research points to two cheaper, alternative, targeted interventions: high-dosage tutoring for primary school pupils; and adaptive digital learning for middle school pupils (those aged 11-15). These schemes could be deployed within the existing infrastructure of schools – and they would cost only a fraction of the current education budget.
Figure 1. Socio-economic score gaps in mathematics and education spending (percentage of GDP)
Above: the gap in average mathematics scores between the most and least socio-economically advantaged 25% of pupils within each country (Programme for International Student Assessment, PISA, 2022). A gap of 39 points roughly equals one year of schooling.
Below: total public education spending as a share of GDP (OECD, Education at a Glance 2024, reference year 2021). Both tables rank countries among 37 OECD members.
Sources: Table 1: OECD, PISA 2022 results, volume I (2023), Table I.B1.4.3; OECD PISA 2022 country notes. Table 2: OECD, Education at a Glance 2024, Table C1.1; UNESCO Institute for Statistics.
Why is France failing its poorest pupils?
France has poured resources into its priority educational network, REP and REP+, allocating additional funding to 1.7 million pupils (20% of the pupil population). But this investment has not paid off.
By the time that children start the first year of primary school, the gap between REP+ pupils and those in private schools is as high as 31 percentage points in mathematical problem-solving and 41 points in French oral comprehension (INSEE, 2024). Rather than closing over time, research by the Observatoire des inégalités shows that these gaps widen.
France's fiscal position leaves little room for expensive or structurally complex reform. Instead, what is needed is a policy response that is affordable, evidence-based and politically manageable. What I’ve coined as the e-DevoirsFaits+ framework is designed to be all three.
What is e-DevoirsFaits+ and how does it work?
The programme builds on Devoirs Faits, a national homework-support programme launched in 2017. The original scheme offers supervised study time to middle school pupils outside classroom hours. e-DevoirsFaits+ extends and deepens this existing infrastructure, targeting solely the priority educational networks.
The programme has two components. The first is a digital adaptive learning platform, deployed within existing Devoirs Faits sessions in REP and REP+ middle schools, targeting pupils aged 11-13. The platform offers exercises that adjust in difficulty based on pupil responses, teacher dashboards showing individual and class-wide progress, collaborative group exercises and a bank of recorded lessons that is accessible throughout the year.
Since Devoirs Faits was made compulsory for all sixth-year pupils (those aged 10-11) in 2023, nine in ten children across the national year group have enrolled in the programme. This makes these sessions a reliable delivery channel for the platform (Ministère de l'Éducation Nationale, 2024).
For fifth-year pupils, participation remains voluntary and rates are lower, which is part of the rationale for the incentive and monitoring structure that e-DevoirsFaits+ puts in place. Existing Devoirs Faits supervisors would receive two hours of paid training to facilitate the digital platform’s use.
The second component targets the years immediately before this transition: CM1 and CM2, the final two years of primary school. University students deliver high-dosage tutoring to 25% of pupils in these year groups, in twice-weekly sessions in pairs. This adds up to 72 hours per year.
Teachers select pupils using a standardised grid based on national assessment results and identified learning gaps. Tutors are paid at the minimum wage, earn university credits and receive a Ministry of Education certificate. They receive initial and continuous training to ensure quality and uniformity. The attendance is monitored via digital progress reports, which are also made available to teachers.
What does the evidence say?
The case for combining tutoring and adaptive technology rests on a substantial body of research. A 2025 report by the French Council of Economic Analysis (Conseil d’analyse économique) identifies these two policies as generating the highest social return per euro spent in education, higher than most of the measures that France has already tried (Grenet and Landais, 2025).
The tutoring evidence is particularly strong. Studies of similar programmes consistently find meaningful improvements in attainment for disadvantaged pupils. For context, the effect sizes documented in the research are roughly equivalent to moving a child from the middle of the class ranking to somewhere in the top third (Nickow et al, 2020). The digital platform evidence is more mixed across the board, but the design of e-DevoirsFaits+ is intended to capture the conditions under which digital tools tend to work best.
Adaptive learning platforms have attracted growing interest, with several national programmes now integrating them. These tools adjust the difficulty of exercises in real time based on how a pupil responds. The evidence suggests that their effectiveness depends heavily on how they are embedded. Stand-alone digital tools tend to underperform when deployed without structured supervision (Escueta et al, 2020).
What makes e-DevoirsFaits+ different is that the platform sits inside an already institutionalised slot: Devoirs Faits sessions run for around an hour, twice a week, under adult supervision. That regularity and structure is precisely what research identifies as necessary for digital learning tools to deliver consistent gains.
The proposal deliberately uses a conservative estimate of the platform’s effect to reflect the fact that these remain homework sessions rather than dedicated instructional time, but the expectation is that sustained, repeated use within a familiar routine will produce results above what one-off or unstructured deployment typically achieves.
The analysis projects a return of around €10 for every €1 invested, translating learning gains into projected lifetime earnings using established estimates of the relationship between childhood skills and adult wages. Even under pessimistic assumptions, the projected return stays positive.
These projections should be read with appropriate caution. The underlying studies draw on evidence from across Europe and OECD countries, and the results may not transfer directly to the French context. Even so, this uncertainty can be accounted for by building in a pilot phase before any national rollout, testing whether projected effects hold in practice.
What is the political pathway of the policy?
The programme requires no new legislation. Under Article 34 of the French constitution, the organisation of educational programmes falls within the regulatory domain, exercisable through executive action alone.
The precedent is clear: Devoirs Faits itself was created in 2017 through a ministerial circular, with no primary legislation and no parliamentary vote. e-DevoirsFaits+ would enter the 2027 budget as a line item, enacted through the same toolkit of ministerial orders and operational guidance.
This approach is deliberate. There is no symbolic reform narrative against which opposition parties could mobilise. Raising the educational outcomes of disadvantaged pupils commands broad consensus across France’s political spectrum.
Spain's PROA+ programme offers a comparable case. It was framed as a technical extension of existing inclusion tools rather than a flagship initiative. PROA+ has since been extended to 2028 with €420 million in European Union funding, now reaching over 3,600 schools (Eurydice, 2025).
Teacher unions cannot block enactment, but they can make or break implementation. The programme anticipates this through paid coordination hours, meaningful school-level autonomy and formal consultation mechanisms during the pilot phase.
How could the implementation challenges be circumvented?
Political feasibility is one thing, but implementation is another. Educational reforms in France have a long history of failing at the classroom level.
Carlo Barone is a member of the French National Scientific Council for Education. He has reviewed e-DevoirsFaits+ and describes it as ‘a promising policy’. His two primary concerns are recruiting sufficient university tutors at scale and ensuring reliable equipment in schools. Teacher interviews conducted for the research added two more: headteacher autonomy; and teacher workload.
The programme addresses each of these directly. On tutor recruitment, an exploratory survey of 43 university students finds that about 75% expressed strong interest in participating, a small sample, but one pointing in an encouraging direction.
The two strongest motivations were pay and social impact, each cited by 81% of respondents. The main concern was commuting time, cited by just over 60%. The programme addresses this by building in an online tutoring option where distance makes in-person sessions impractical. The existing Cordées de la Réussite mentoring programme further demonstrates that French university students will engage in educational support roles when given proper incentives.
On equipment, the programme bypasses the patchwork of municipal budgets for information technology that has historically produced unreliable results in REP schools. Teacher interviews confirm that much existing equipment ‘no longer works’ due to lack of institutional maintenance. National procurement directly addresses this, with binding delivery timelines, a five-year replacement cycle and funding through the national budget rather than local authorities.
On headteacher authority, the programme deliberately preserves existing Devoirs Faits governance. Headteachers retain full discretion over session scheduling and supervisor assignments. No authority is removed from school leadership.
On teacher workload, primary teachers receive one paid hour per week for integrating tutor progress reports, and teachers who wish to contribute to platform design are given the opportunity to do so and will be paid as well.
Implementation follows a three-tier model. The ministry sets binding standards and provides funding. Regional education authorities coordinate university partnerships, conduct capacity audits and manage equipment delivery. Schools retain meaningful autonomy over scheduling, pupil grouping and family communication.
A 2026 pilot would test these logistics before a 2027 national launch, drawing on ten REP+ schools selected for geographic and demographic diversity.
What can England learn from this?
England faces a recognisable version of France’s problem. The attainment gap between disadvantaged pupils and their peers remains persistent despite sustained investment (Education Policy Institute, 2025). The evidence on what works points in a similar direction: the Education Endowment Foundation consistently identifies high-dosage tutoring as one of the most cost-effective interventions available for closing attainment gaps.
England's national tutoring programme (NTP) offers both encouragement and instructive lessons. Launched in 2020/21 as part of the government's Covid-19 catch-up strategy, it provides the closest available comparator.
Evaluations find that in schools where tutoring was concentrated among the most disadvantaged pupils, positive effects on attainment were consistently detected. The researchers themselves note that methodological constraints are likely to have led them to underestimate the true impact (National Foundation for Educational Research, NFER, 2024).
At the same time, the programme’s weaker overall results reflect specific design failures: a fragmented provider market with inconsistent quality, difficulty systematically targeting the most disadvantaged pupils, and the absence of an institutional anchor within school routines that could sustain delivery.
These are precisely the risks that e-DevoirsFaits+ is designed around. The NFER evaluation highlights university student tutors as a promising lower-cost model: paid for their time, earning course credits and operating within structured school settings. That is almost exactly the design that e-DevoirsFaits+ proposes, reinforced by the training structure and compliance mechanisms that the NTP largely lacked.
The broader lesson for UK policy-makers is about design logic. Tutoring works, but its effectiveness depends heavily on how it is embedded. Programmes consistently outperform those that bolt provision onto schools from the outside when they are anchored within existing infrastructure, target the right pupils and use trained tutors with monitored delivery.
Where can I find out more?
- The full e-DevoirsFaits+ policy paper, including the cost-benefit analysis, legal framework and stakeholder mapping underpinning this article, is available on request from the author.
- Education: how to better target public spending: this 2025 report by Julien Grenet and Camille Landais for the Conseil d’analyse économique introduces the social return per euro framework underpinning the programme’s design, identifying tutoring and adaptive digital learning as the two highest-return educational interventions.
- NTP: what we’ve learned – and haven’t – about tutoring’s impact: a 2024 blog post from NFER, which ran the independent evaluations of England’s national tutoring programme, drawing out the lessons for schools and policy-makers considering tutoring at scale.
- The landscape of educational inequalities in Europe – and what to do about them: this 2024 blog post by Moris Triventiplaces France’s educational gap in a broader European context, examining why some countries are far better than others at decoupling pupil outcomes from social background.
- Education’s biggest challenge is ensuring consistency in what is being done across the system, within each school, and within each classroom: this 2021 blog post by Lee Elliot Major and Andy Eyles argues that the key failure in educational reform – including tutoring programmes – is not policy design but implementation consistency, and examines what this means in practice for UK schools.
