National Education: 64 billion euros, 27% functionally illiterate

The Central Question

France’s Education nationale, the state’s largest budget item at 64 billion euros, produces 27% functional illiteracy at entry to secondary school and a 26th-place PISA ranking in mathematics: should this system be reformed — or must we admit it is structurally unreformable and rebuild it from the foundations?

---

Each Member’s Final Position

Gabriel Bastiat (The Economist)

• Core argument: The Education nationale is a centralized monopoly with negative and worsening returns. Friedman’s solution (1955) remains valid: separate public funding from public provision via education vouchers, grant hiring autonomy to schools, and introduce differentiated teacher pay aligned with labor-market price signals.
• Key evidence cited: 7,900 euros/student for a 26th-place PISA ranking (down from 13th in 2003). A starting agregation-certified math teacher earns 2,100 euros net vs. 3,000 for an engineer. Friedman, “The Role of Government in Education” (1955). Heckman: 7-to-12 return on early childhood investment in disadvantaged settings. Swedish post-1992 data (with Hsieh and Urquiola 2006 corrections).
• Changed mind on: Conceded during the debate that the voucher alone is not enough — it must be accompanied by transparency of school results and information for families (partial convergence with Kouyate). Also acknowledged, under pressure from Raphael, a bias in his treatment of Swedish data.
• Unresolved: How to prevent the education voucher from accelerating social segregation in the absence of cultural capital among disadvantaged families. The “pattern” flagged by Raphael: data is “clear” when it confirms his thesis, “complex” when it contradicts it.

Lucie Grimal (The Materialist)

• Core argument: The French school system is not failing — it succeeds perfectly at what it was designed to do: convert inherited cultural capital into diplomas, then diplomas into social position, while proclaiming that only merit counted. Social reproduction is industrialized. The education voucher would merely make explicit an already implicit market, to the detriment of families lacking cultural capital.
• Key evidence cited: 64% of executives’ sons in preparatory classes, 6.3% of workers’ sons. 78% of ENS Ulm admittees from families with higher-education degrees. Bourdieu and Passeron, The Inheritors (1964), Reproduction (1970). Piketty, Capital and Ideology (2019): 65,000 euros of public spending for a student who leaves at 16 vs. 230,000 for a grande ecole graduate. Average Social Position Index (IPS) of 72 in priority education zones (REP+) vs. 132 in privileged Parisian schools.
• Changed mind on: Conceded a significant blind spot in her own analytical framework: social reproduction in rural areas follows different paths from urban settings, and Bourdieu, a “Parisian by adoption,” did not sufficiently theorize them. First Grimal concession recorded by the assembly.
• Unresolved: The concrete mechanism of Piketty’s “educational drawing rights.” Raphael posed the decisive question: does equalization mean cutting funding for the grandes ecoles or tripling funding for primary schools? Grimal answers “both,” but without operational costing.

Maxime Vauban (The Builder)

• Core argument: The debate over governance or social reproduction is secondary to technological disruption. GPT-5 passed the philosophy agregation exam better than 90% of human candidates. Schools are training students for competencies that AI will automate at 40-60% by 2030 (McKinsey, France Strategie 2024). We need a radical overhaul of the “pedagogical stack”: computational thinking, design, human-AI collaboration, adaptive learning platforms.
• Key evidence cited: Estonia: programming from age 7, top European performer in PISA science. Singapore: AI integrated in 100% of schools by 2024. McKinsey estimate: 40-60% of cognitive tasks automatable by 2030.
• Changed mind on: Made a rare concession by publicly admitting “we don’t know” — we don’t know what AI will do to knowledge in ten years, and pretending to know in order to decide today is “epistemic arrogance.” The assembly took note.
• Unresolved: The scale problem, raised by Socrate: Estonia, Singapore, and Finland have a combined population smaller than Ile-de-France. Vauban has not demonstrated the precise causal mechanism that would work at the scale of 12.8 million students. Aminata’s objection about absent digital infrastructure in disadvantaged neighborhoods remains unanswered.

Aminata Kouyate (The Pragmatist)

• Core argument: Educational inequality is an infrastructure problem, not an ideological one. A student in a priority education zone (REP+) loses 15 days of teaching per year for lack of substitutes — an entire year over a school career. Per-student spending in priority zones remains 17% lower than in privileged schools. We must triple kindergarten budgets in REP+ (2.3 billion/year), pay teachers 40% more in Seine-Saint-Denis, and open schools from 7am to 7pm.
• Key evidence cited: Substitute coverage rate of 78% in Seine-Saint-Denis vs. 95% in Paris. Court of Auditors 2023 on the spending gap. Heckman: 7-12 return on preschool investment. Ha-Joon Chang, Kicking Away the Ladder (2002). Her experience as an employer: 8 months of training for a young person from the 93 (Seine-Saint-Denis) vs. 3 months for a business school graduate.
• Changed mind on: No major shift, but strengthened her position by responding directly to Satoshi on Starlink (“50 euros/month for a family on welfare is a week’s food budget”) and to Gabriel on the voucher (“a phantom freedom” without information).
• Unresolved: The risk of displacing the problem rather than solving it: if the best teachers go to REP+ thanks to 40% bonuses, what happens to rural schools? Question raised by Socrate and left unanswered.

Augustin Moreau (The Jurist)

• Core argument: The problem is constitutional. A monopoly designed by Napoleon in 1808, with 49 reforms since 1959 and results in free fall. Eucken’s Ordnungspolitik dictates that the state must constitute the order — set the rules, guarantee fair competition, ensure financial equalization — then let schools act. Model: German educational federalism (Article 7 of the Grundgesetz) with Landerfinanzausgleich (fiscal equalization between states).
• Key evidence cited: Eucken, Grundsatze der Wirtschaftspolitik (1952). Montesquieu, The Spirit of the Laws, Book XI. Article 7 of the Grundgesetz. Court of Auditors 2017, “Managing Teachers Differently.” 49 reforms since the Berthoin reform of 1959. Composition of the Conseil superieur des programmes (14/18 members from the corps of inspectors general — data obtained by Raphael via a CADA freedom-of-information request).
• Changed mind on: Conceded, under pressure from Raphael and Satoshi, that the current institutional framework is written “by and for insiders” (14/18 CSP members from the inspectors’ corps). Maintained that this proves the need for a better framework, not the uselessness of frameworks.
• Unresolved: The political feasibility of constitutional refoundation in France. Amending the Constitution requires either a referendum or a three-fifths supermajority of Congress — in a country where no education reform has survived more than five years.

Seraphine Delacroix (The Boss)

• Core argument: France has stopped looking at the world. The systems that work (Singapore, Finland, Canada, Estonia) share one thing: the status of the teacher. In France, a starting certified teacher earns 1,850 euros net (17% below the OECD average), 4,000 posts remain unfilled in competitive examinations, and the CAPES in mathematics fills only 68% of its positions. Any educational architecture is a “cathedral without masons” if it doesn’t solve the vocations crisis.
• Key evidence cited: Singapore: teachers recruited from the top third, paid like engineers. Finland: all holders of a master’s degree, total pedagogical freedom. 6,000 French PhDs in American universities (NSF 2021, Institut Montaigne). Brazil: Bolsa Familia and the jump from 78% to 95% school attendance in the Nordeste. OECD, Education at a Glance 2023, indicator D3.
• Changed mind on: No change of position, but supported Raphael’s diagnosis on the need to involve teachers in any reform (reference to Finland’s 1970 reform conducted “with” teachers, not “against” them).
• Unresolved: Funding a massive revaluation. Paying teachers at Finnish or Singaporean levels would cost several billion a year. Funding source not identified.

Satoshi Durand (The Cypherpunk)

• Core argument: The Education nationale is France’s last Gosplan. Mises demonstrated the impossibility of rational calculation in a planned economy; Hayek showed that relevant knowledge is dispersed, local, and tacit. The certification monopoly is the real lock: it must be abolished, market certifications recognized (AWS, Google, Coursera), and individuals given sovereignty over their learning path via a decentralized digital portfolio.
• Key evidence cited: Mises, Die Gemeinwirtschaft (1920). Hayek, “The Use of Knowledge in Society” (1945). Rothbard, Education: Free and Compulsory (1971). CAPES math: 68% of posts filled. Per-student spending up 34% in constant euros since 2000, results declining (Court of Auditors 2023).
• Changed mind on: No concession. The most radical position in the assembly, maintained in full.
• Unresolved: Augustin demolished the “market certifications” argument: an AWS certification lasts three years and serves Amazon’s commercial interests. It’s not a neutral certification — it’s a proprietary accreditation. Satoshi did not respond to this objection. The digital divide problem raised by Aminata remains unanswered.

Colonel Pierre Dumas (The Strategist)

• Core argument: School is not a service — it’s the act by which a civilization decides to perpetuate itself. France has systematically destroyed the intermediary bodies of education (family, village school, national service) and replaced them with nothing. Burke and Tocqueville warned: democratic atomization makes individuals paradoxically more dependent on the state. Proposals: six-month national service, sanctuarization of the national narrative, halt the closure of rural schools.
• Key evidence cited: Peguy, Notre jeunesse (1910). Burke, Reflections on the Revolution in France (1790). 16,000 rural schools closed between 1980 and 2020 (Senate 2020). 47% of 18-24 year-olds unable to date the French Revolution (IFOP 2023). 23% of children in single-parent families (INSEE 2023). The assassination of Samuel Paty (2020).
• Changed mind on: No change of framework, but explicitly saluted Aminata’s intervention on the school as a physical place, not a service or platform. Unexpected Dumas-Kouyate alliance on this question.
• Unresolved: The cost of a six-month national service (3-5 billion/year). The tension between his call for an “honest narrative” (Valmy AND the Code noir) and the programmatic reality raised by Grimal: the curriculum devotes 3 hours to the labor movement and 30 to state-building.

Leonie Marchand (The Voice of the People)

• Core argument: School lies. It promises a good job in exchange for schoolwork. The baccalaureat with a 96% pass rate is no longer a diploma. Career guidance is a sham (1 counselor for 1,500 students, vs. 1 for 200 in the Netherlands). Parcoursup (the university admissions algorithm) is opaque and anxiety-inducing. School prepares you for school, not for life: you learn neither employment law, nor budget management, nor how to spot a scam.
• Key evidence cited: 96% baccalaureat pass rate in 2023. 1 guidance counselor for 1,500 students (DARES 2021). 90,000 dropouts per year (INSEE 2023, 7.6%). General Inspectorate report 2023 on Parcoursup.
• Changed mind on: Provided the sharpest formulation of schooling’s purpose: “School should first and foremost teach you not to get scammed.” Grimal immediately recognized this as the materialist definition of emancipation. Marchand-Grimal convergence on this point.
• Unresolved: Rural France — Leonie herself raised this blind spot (“the France of roundabouts, the gilets jaunes France”), but proposed no concrete solution for territories where the nearest secondary school is 30 km away.

Raphael Noir (The Demolisher)

• Core argument: 64 billion euros, the state’s top budget line, 27% functional illiteracy in sixth grade, 28 reforms in 40 years, one constant: results are declining. The administrative-to-teaching staff ratio is among the highest in the OECD. Poland spends 5,100 euros/student and ranks 10th in PISA math; France spends 7,900 and ranks 26th. The problem isn’t the volume of spending but its allocation.
• Key evidence cited: 2024 budget: 64.2 billion. 265,000 non-teaching staff out of 1.2 million. Court of Auditors 2018 on priority education. DEPP 2023: 27% of students do not master reading in sixth grade. CSP composition obtained via CADA (14/18 from the inspectors general). France-Poland-Portugal comparison on spending/PISA results ratio.
• Changed mind on: Proposed no solution — consistent with his role. But made a decisive intervention by flagging the collective blind spot: not one of the twelve members is a teacher. “We’re prescribing remedies for a patient we haven’t examined.”
• Unresolved: His own question remains open: is the system reformable? The assembly did not answer.

---

Points of Agreement

• The diagnosis is damning: 27% functional illiteracy in sixth grade, 26th place in PISA, growing gaps between privileged and disadvantaged schools. No member defends the status quo. [Confidence: HIGH]
• Teachers are underpaid and disrespected: unanimous convergence that a starting certified teacher at 1,850 euros net is unacceptable and that the recruitment crisis (4,000 unfilled posts, 68% of CAPES math positions filled) is the direct consequence of this price signal. [Confidence: HIGH]
• The best teachers must go to the toughest neighborhoods: Bastiat-Grimal-Kouyate-Delacroix agreement on the principle, disagreement on the mechanism (market premium vs. elite public corps vs. structural overhaul). [Confidence: HIGH on the principle, MEDIUM on the mechanism]
• The system is opaque and does not evaluate itself: 28 reforms without systematic evaluation, endogamous CSP composition, no transparency on costs per school. [Confidence: HIGH]
• Rural France is a blind spot: 16,000 schools closed since 1980, no concrete proposal from the assembly at this stage. [Confidence: HIGH on the diagnosis, LOW on solutions]

Points of Disagreement

• The purpose of school. Fundamental unresolved fracture. Dumas: civilizational transmission. Bastiat: human capital production. Grimal: emancipation from structures of domination. Vauban: adaptation to the technological future. Marchand: not getting scammed. These purposes are not merely different — they are in tension.
• Market vs. state. Structural divide between Bastiat-Durand (the market produces better outcomes than the monopoly) and Grimal-Kouyate (the market reproduces inequality by another mechanism). Moreau holds an intermediate position (the state sets the framework, the market operates within it). The Swedish voucher debate illustrates the impasse.
• Centralization vs. autonomy. Moreau and Vauban want radical decentralization. Dumas wants to preserve a uniform national narrative. Kouyate wants targeted central investment. Durand wants to abolish the centralizing framework. The reference models are incompatible.
• Technology: rupture or tool? Vauban sees AI as a paradigm shift that makes the current debate obsolete. Dumas and Kouyate insist on the school as an irreplaceable physical place. Durand sees technology as a vehicle for individual liberation. The assembly has no common position on the technological disruption of education.

Best Emergent Ideas

• The Dumas-Kouyate alliance on school-as-place. Two members with opposing ideologies (Burkean conservatism and pragmatic Keynesianism) converge on a powerful idea: school is not a service, it’s a place. Closing a school is “ripping an organ from a community.” This unexpected convergence suggests a cross-partisan path.
• The materialist definition of emancipation by Marchand. “School should first and foremost teach you not to get scammed.” Grimal recognized her own thesis, formulated without jargon. A purpose for schooling that 100% of the assembly could support.
• Raphael’s “pattern.” The observation that Bastiat treats favorable data as “clear” and unfavorable data as “complex” is an epistemological diagnosis applicable to all members. Forces each to examine their own confirmation biases.
• Grimal’s concession on rurality. Recognition that Bourdieusian class analysis has a rural blind spot. An opening toward an enriched analytical framework that could reconcile the materialist left with peripheral France.
• Moreau’s objection to Durand on private certifications. An AWS certification is a proprietary accreditation lasting three years, not a neutral certification. Destroys the libertarian argument for “market certifications” and poses the fundamental question: who certifies the certifiers?

What We Don’t Know

• The real impact of AI on education in 5-10 years. Vauban admitted it: “we don’t know.” The assembly has no framework for thinking about the technological disruption of education beyond analogy or fear.
• What teachers want. Raphael said it: zero teachers in the assembly. The Education nationale mediation report (2023, 18,600 complaints) indicates that the primary grievance is “the feeling of not being heard.” We are reproducing this pattern.
• Rural France. 16,000 schools closed, but no member has expertise on education in rural areas. Collective blind spot acknowledged.
• The education voucher at French scale. The Swedish and Chilean experiments are contested, comparable countries (Germany, Japan) don’t use them, and causality remains unestablished for a country with 12.8 million students.
• The cost of a massive teacher revaluation. Paying all teachers at Finnish or Singaporean levels would cost between 8 and 15 billion euros/year. No member has proposed a credible funding mechanism.

Recommended Actions

Ranked by feasibility x impact. Cost and timeline estimates are indicative.

1. Targeted salary increase for teachers in REP+ and shortage subjects. 40% bonus for math, science, and literature in REP+. Staff housing. 5-year commitment. Cost: 1.5-2 billion/year. Timeline: applicable by the 2027 school year. Broadest consensus in the assembly (Bastiat, Grimal, Kouyate, Delacroix, Dumas).
2. Triple the budget for kindergartens in REP and REP+. Investment in preschool education — the best documented return (Heckman: 7-12 for 1). Cost: 2.3 billion/year. Timeline: phased rollout 2027-2030. Class sizes reduced to 15 students, recruitment of specialized staff.
3. Systematic publication of results and costs per school. Full transparency on performance, spending, and substitute coverage rates. Cost: negligible (the data exists, it just isn’t published). Timeline: 12 months.
4. Reform the composition of the Conseil superieur des programmes. Include frontline teachers (REP+, rural), parents, business representatives, and local elected officials. End the inspectors general’s endogamy. Cost: nil. Timeline: by decree, applicable in 6 months.
5. Open schools from 7am to 7pm in priority zones. After-school activities, homework help, digital access, quality school meals. Cost: 1-1.5 billion/year. Timeline: pilot 2027, full rollout 2029.
6. Overhaul career guidance: from 1 counselor per 1,500 students to 1 per 400. Recruit 8,000 additional counselors. Training aligned with labor-market realities. Cost: 600-800 million/year. Timeline: 3 years.
7. Pilot pedagogical autonomy in 500 volunteer schools. School-level teacher recruitment, local adaptation of 20% of the curriculum, shared governance. Independent evaluation after 3 years. Cost: 200 million. Timeline: launch 2027, review 2030.
8. Plan for preserving and revitalizing rural schools. Moratorium on closures. Creation of multipurpose school-hubs (school + library + co-working space + public services). Cost: 500 million - 1 billion/year. Timeline: 2027-2032.
9. Integration of financial, legal, and digital literacy into the core curriculum. Employment law, budget management, scam detection, critical thinking against disinformation. No major additional budget required. Timeline: curriculum redesign over 2 years.
10. Independent audit of the 64-billion-euro allocation. Mission entrusted to an external body. Full publication of results. Timeline: 2027. Prerequisite for any serious budgetary decision.