## sipea2023014

## Source details

**Canonical URL:** [sipea2023014](https://www.imf.org/-/media/files/publications/selected-issues-papers/2023/english/sipea2023014.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/selected-issues-papers/2023/english/sipea2023014.pdf.md)
- [Structured JSON version](/-/media/files/publications/selected-issues-papers/2023/english/sipea2023014.pdf.json)

---

### Introduction / Overview
- Trend increase in primary current spending over the last three decades leaves France with the highest spending ratio in Europe and requires structural spending reforms to rebuild fiscal buffers further eroded by the large fiscal response to successive shocks over 2020–22.
- Annual real primary spending growth has outpaced output growth about half of the time over the past three decades, averaging 2 and 1.5 percent over 1990-2021 respectively.
- The spending gap with peers reached 10 percentage points of GDP in 2019, driven primarily by social benefits (+6½ ppts), followed by the wage bill and subsidies (about +1¼ ppts each), together accounting for 87 percent of the gap.
- Paper’s approach: disaggregate spending by economic and functional classification; focus analysis on social protection, health, education, and economic support; assess efficiency using stochastic frontier analysis (SFA).

### Social Protection — Key Findings and Reforms
- Scale and gap:
  - Social protection spending: 23.8 percent of GDP in 2019 (second highest in Europe after Finland).
  - Gap relative to peers: about +5.5 ppts of GDP.
  - Peak spending: 24.5 percent of GDP in 2014.
- Redistributive outcomes and efficiency:
  - Poverty rate after taxes and transfers: 14.6 percent in 2019 (France 2.5 ppts below peer average per OECD data).
  - Disposable income Gini coefficient: 0.29 in 2019.
  - Reduction in Gini due to social transfers slightly higher than peer average (about ½ ppt), but redistributive efficiency is lower because France spends significantly more than peers.
  - Reduction in the Gini coefficient per one percent of GDP of social benefit spending excluding pensions: 0.55 percent for France vs. median for peers 0.65 percent (in absolute terms 0.19 vs. 0.22 percentage points).
  - Targeting: unemployment benefits means-tested only 6 percent (peer average 26 percent); share of untargeted family benefits about 13 ppts above peers.
- Pensions:
  - Public pension spending: about 14 percent of GDP in 2019.
  - System fragmented across 42 different regimes; features include low effective retirement age and high replacement rates.
  - Normal retirement age: 65; minimum retirement age: 62.
  - Average age of labor market exit: 60.6 years.
  - Life expectancy: 0.6 years above peer average.
  - Expected retirement duration: second highest in Europe (shared with Spain, after Greece).
  - Proportionality measure indicates system became relatively more generous since 2008, especially for mid-career workers.
  - COR (2022) projections: higher living standards of retirees relative to rest of population will erode over time as wages grow faster than pension benefits indexed to inflation.
- Employment of older workers:
  - Employment rate 55+ years: 55 percent.
  - Under current policies, participation rate of 65-74 cohort expected to increase by only 15 ppts by 2070, with less than half due to reforms (EC Ageing Report 2021).
- Unemployment benefits:
  - Spending: close to 2 percent of GDP.
  - Benefit duration: minimum broadly in line with peers (6 months); maximum significantly higher (24 months vs. average 18 months for those aged below 50); older cohorts: 30 months for those aged 53-54 and 36 months for those aged 55 and older.
  - Replacement rates decline less with benefit duration compared to peers, suggesting weaker labor market incentives.
- Recent reforms and expected effects:
  - 2010 and 2014 pension reforms raised minimum retirement age and age for full-rate pension by 2 years to 62 and 67; increased minimum contribution period for a full-rate pension (changes effective from 2017 and 2020).
  - Pension reform plan (presented January 10, 2023): raise minimum retirement age from 62 to 64 at a rate of 3 months per year from September 2023; increase minimum contribution period from 42 to 43 years by 2027; grandfather existing participants; align most special regimes with general regime for new participants (some special schemes excluded).
  - 2019 unemployment reform: adjusted benefit calculation and accumulation rules, tightened eligibility, introduced degressivity for higher incomes, and a modulation of employers’ unemployment contribution (bonus-malus).
  - December 2022 law: extends the 2019 reform application until end-2023 and introduces countercyclicality in unemployment benefits (effective from February via decree):
    - New rule: vary benefit duration with level and change in unemployment rate; 25 percent reduction in maximum duration when unemployment rate is below 9 percent and its quarterly rate of increase is below 0.8 ppts.
    - Duration floor: 6 months; duration extended when labor market conditions deteriorate.
    - Government considered a 40 percent reduction when unemployment rate drops below 6 percent but postponed decision to allow dialogue with social partners and to regulate in the next interprofessional agreement (early 2024).
  - Expected outcomes: reforms aim to improve employment rate of older workers, raise effective retirement age closer to peers, strengthen sustainability, generate significant savings, and improve labor market incentives.
- Social benefits: scope to streamline
  - Housing benefits:
    - Spending on housing benefits: 0.9 percent of GDP.
    - Spending on housing and community amenities: 1.1 percent of GDP (peer average 0.4 percent of GDP).
    - One in five households receive housing support.
    - Three quarters of the population eligible for social housing (Cour des Comptes, 2021c).
    - Overcrowding rate for poor households: 22.6 percent.
    - Problems: insufficient and inelastic housing supply in densely populated urban areas, limited turnover and accessibility of social housing, poor targeting of housing support, complexity of aides personelles au logement increasing administrative costs.
  - Social exclusion benefits (social minima):
    - Spending: 1.3 percent of GDP (½ percent of GDP more than peers).
    - Comprises 10 different schemes covering about 10 percent of the population.
    - Reduced the poverty rate and intensity by 3 and 8 ppts, respectively, in 2019.
    - Structural flaws: low exit rate (only 2-4 percent find durable employment); 42 percent of RSA beneficiaries remain beneficiaries after 7 years; only a third transition into employment.
    - Coverage gaps: Only 70 and 40 percent of those eligible for RSA are recipients of benefits and reintegration support, respectively.
    - Fraud/erroneous payments: CNAF estimates fraud at some €1bn in 2019 for AAH and RSA; Cour des Comptes (2020) estimates some €1.4bn, or 12 percent of spending, were wrongly paid out.
    - Reform plans: automation, digitalization, better data exchange (solidarité à la source); eligibility to RSA to be more strictly conditioned on reintegration efforts; pilot in 19 departments in 2023.
  - Family benefits:
    - Spending marginally above peer average.
    - Outcomes mixed: slightly worse than peer average for child poverty risk in vulnerable households, but better for fertility and female labor force participation.
    - Targeting of family benefits could be improved.

### Health — Findings and Recommendations
- Outcomes and risk factors:
  - France has one of the highest life expectancies compared to peers.
  - Mortality rates from treatable causes lower than peers.
  - Gaps in life expectancy by education similar to peers (and lower for women).
  - High smoking rates and alcohol consumption among disadvantaged socio-economic groups; chronic conditions and cancer incidence much higher than in peers.
  - Quality of care: effective primary care with lower avoidable hospital admissions; below peers in safety of primary care (high antibiotics prescriptions) and preventive care.
- Spending and composition:
  - Total health spending in 2019: 11 percent of GDP (peer average 9.6 percent of GDP).
  - Public health spending: about 77 percent of total (peers: 70 percent).
  - Public health spending to GDP increased by 1 ppt over past two decades.
  - Compensation of employees in health: 2.2 percent of GDP (peers: 2 percent of GDP).
  - Social transfers in health: 4.3 percent of GDP (peers: 2.7 percent of GDP).
  - Inpatient and outpatient care spending: 4.3 percent of GDP (peers: about 3.9 percent of GDP); these categories account for about 50 percent of all health spending in France.
  - Administrative costs: 0.5 percent of GDP (peers: 0.2 percent of GDP).
- Medical resources and workforce:
  - Number of hospital beds in France far exceeds peers.
  - Number of doctors below peers; density of general practitioners reduced from 1.5 per 1 000 population in 2012 to 1.4 in 2021, with increased regional disparities.
  - Recommendation: increase by an additional 20 percent the number of students admitted to medical schools in France during 2021-25 compared with the previous five-year period (ONDPS, 2021).
- Efficiency and reform recommendations:
  - Reduce spending on inpatient care, medical goods, and administrative costs by restructuring the public hospital network and developing primary care and ambulatory centers.
  - Improve management of human resources and equipment purchases.
  - Invest in prevention; increase incentives for preventive action (e.g., higher remuneration for doctors for preventive action) and promote tobacco and alcohol control policies.
  - Shift remuneration away from fee-for-service or per admission toward lump-sum payments; consider an annual individualized per-patient budget for chronic diseases.
  - Encourage greater use of generics: current share in France is 16 percent in value and 30 percent in volume (much lower than Germany or the UK: 35 percent in value and more than 80 percent in volume).

### Education — Findings and Recommendations
- Performance and attainment:
  - PISA test scores in secondary education below Germany, the UK, and other advanced economies.
  - Worse than G5 in math, science, and reading for grades 4 and 8.
  - Only 30 percent of the population have completed a short post-secondary education cycle (UK: 35 percent; Germany: 33 percent).
  - One of the lowest shares of students graduating with a master’s degree or higher among the G5; one of the highest shares of graduates for upper secondary education.
- Spending and composition:
  - Public education spending in 2019: 5.2 percent of GDP (peer average 5 percent of GDP).
  - Compensation of non-teaching staff: 22 percent in non-tertiary and 38 percent in tertiary education (peers: 12 and 28 percent, respectively).
  - Expenditure per student:
    - Secondary education: 5 percent higher than peers.
    - Upper secondary level: over 30 percent higher than peers.
    - Primary education: 20 percent lower than peers.
    - Tertiary education: 8 percent lower than peers.
  - Teacher/pupil ratios lower than peers at all levels, slightly higher in upper secondary education.
- Reform suggestions:
  - Rationalize excessive resources in secondary and tertiary education and rebalance spending toward pre-primary and primary levels.
  - Strengthen selectivity of teacher hiring, particularly in scientific subjects, and use assessments to steer the system.
  - Strengthen training of teachers (initial and lifelong); address lower preparedness on pedagogical aspects and less in-class pedagogy training and collaboration (TALIS, 2018).
  - Align teacher compensation with performance and challenges:
    - The 2023 budget envisages an increase of teachers’ salaries by 10 percent and the guarantee that no teacher would have a starting salary of below Euros 2000.
    - This unconditional increase would be accompanied by another 20 percent increase in salaries if supplementary educational missions are performed.
  - Give more autonomy to school administrations and improve early access for children from low socio-economic backgrounds.

### Public Administration and Local Governments
- Wage bill and employment:
  - Wage bill in 2019: 12.2 percent of GDP and over one fifth of public spending.
  - Government employment about 15 percent of the working age population since early 2000s.
  - Share of local government employment about 12 percent of total employment—the highest among peers.
- Containment challenges:
  - Wage bill about 1.5 ppt of GDP larger in France than in peers.
  - Rapid increase in local governments’ wage bill and employment driven by weak hiring practices and rapid promotions at the municipal level.
  - Large number of municipalities per 100K inhabitants compared to peers.
  - Procurement purchases sub-optimal at local level; transfers from the center have supported additional and frequently inefficient spending.
- Recommendations:
  - Contain local wage bill and employment growth; standardize procurement processes via centralized electronic procurement.
  - Clarify responsibilities across levels of government; increase tax autonomy and consider introducing new own-source local taxes or broaden local control over rates.
  - Update property valuations to strengthen local property tax revenues.
  - Prioritize fiscal capacity of regions to support vocational training and clarify roles of national and regional councils.

### Tax Expenditures — Findings and Recommendations
- Overall:
  - Total spending on economic support via tax expenditures and subsidies: close to 6 percent of GDP (1½ ppts higher than peer average).
  - Two thirds of spending support general affairs and transportation.
  - More than 40 percent channeled through subsidies and tax expenditures, accounting for nearly 90 percent of the spending gap with peers.
  - Including the CICE, tax expenditures amounted to 4 percent of GDP in 2019.
  - Number of tax expenditures: 394 (second only to Italy at 490).
- Key items and concerns:
  - R&D (crédit d’impôt recherche, CIR):
    - CIR size: ¼ percent of GDP (single largest tax expenditure; more than 80 percent of R&D-related and 8 percent of total tax expenditures).
    - Disproportionate use by large firms: in 2018 the 100 largest beneficiaries accounted for about a third of total spending; firms with more than 5k employees accounted for 36 percent.
    - In 2019 large firms accounted for only 14 percent of R&D tax relief recipients but 70 percent of benefits.
    - Return on CIR spending (patents per euro spent) 2.5 times larger for micro firms than for large firms (2.9 times when considering patents covering multiple jurisdictions).
    - Design feature: reduced CIR rate of 5 percent for R&D spending above €100mn (vs. 30 percent below).
    - Estimates by Le Gall et al (2021): CIR reform generates 0.8pp additional growth and 60k jobs in the long term, implying an average annual cost of more than 83k per job created.
  - Housing tax expenditures:
    - Size: ½ percent of GDP in 2019 (peer average for available data: 0.2 percent).
    - Account for some 16 percent of tax expenditures (excl. CICE) and equal more than 80 percent of the budget allocation for territorial cohesion.
    - Effectiveness assessments inconclusive or point to limited effectiveness, significant cost, and/or undesirable effects.
  - Fossil fuels:
    - Fuel tax expenditures: ⅓ percent of GDP in 2019 (more than double the peer average).
    - Mostly reduced rates for fossil fuel use in transport and agriculture; contribute to pollution, congestion, and undermining climate goals.
  - Savings incentives:
    - Household savings incentives: about ¼ percent of GDP.
    - Given saving surplus accumulated during the pandemic, reduction or elimination of savings incentives could be considered.
- Recommendations:
  - Rationalize or redesign costly, distortive, regressive, or ineffective tax expenditures (notably R&D/CIR, housing-related, fossil fuels, and household savings incentives).
  - Enhance CIR targeting: lower ceiling, raise CIR rate for small firms or differentiate by firm size, and reduce overlap between R&D incentives.
  - Improved transparency, systematic evaluation, and use of sunset clauses (new fiscal and social tax expenditures subject to 5 and 3 year sunset clauses, respectively).
  - Potential fiscal impact: coupled with rationalizing other tax expenditures, this could yield savings ranging from ½ to 1 percent of GDP.

### Efficiency Assessment (SFA Methodology and Results)
- Methodology:
  - Stochastic Frontier Analysis (SFA) with inefficiency term assumed to follow an exponential distribution.
  - Production form: 푦푖∗ = c ∗ ∏(푥푖푗)훼푗; log-linear estimation ln(푦푖) = 훼0 + ∑훼푗 ln(푥푖푗) + 푣푖 − 푢푖.
  - Inputs: public spending on health, education, and social protection; controls for private spending, level of development, old-age dependency ratio, unemployment rate, and other determinants.
- Results — Social Protection:
  - France can improve efficiency of social protection spending.
  - Social protection index (working- and old-age poverty rates, disposable income Gini, long-term unemployment rate, school enrollment rate, housing overburden rate) is the second-highest after Finland.
  - France’s efficiency score is above the peer average but well-below best performers: Finland, Portugal, Ireland, and Denmark.
  - Efficiency score axis range shown: 0.70, 0.75, 0.80, 0.85, 0.90, 0.95, 1.00.
- Results — Health:
  - Some room for efficiency savings in health spending.
  - Health outcome indicator: life expectancy at birth, healthy life years, infant mortality, health satisfaction level.
  - Best performers (lower spending for a given outcome): Spain, Sweden, Belgium, Norway, the Netherlands; Germany also performs better than France.
  - Efficiency score axis range shown: 0.8, 0.85, 0.9, 0.95, 1.
- Results — Education:
  - France appears to have some room for efficiency savings, mostly in secondary and tertiary education spending.
  - Education outcome indicator: PISA scores, education satisfaction, population speaking a second language, secondary and tertiary attainment.
  - Best performers: Ireland, Germany, and Finland.
  - Efficiency score axis ranges shown:
    - Overall education: 0.70 to 1.00.
    - Primary: 0.75 to 1.00.
    - Secondary: 0.85 to 1.00.
    - Tertiary: 0.0, 0.2, 0.4, 0.6, 0.8, 1.0.

### Aging and Long-Term Care Pressures
- Demographic and LTC projections:
  - 24 percent of the French population will be aged 65 or over in 2030.
  - Number of long-term care recipients will increase by 76 percent from 2015-2050 (based on the French long-term care indicator).
  - 2021 Ageing Report projections: public expenditure on LTC is projected to increase in relation to GDP by 0.7 ppt (or 37 percent) between 2019 and 2050 (from 1.9 percent in 2016 to 2.6 percent in 2050).
- Implication: achieving efficiency savings is important given these build-ups from aging and medical progress.

### Conclusions and Policy Recommendations (Summary)
- Reform priorities span social protection, health, education, and the civil service to preserve or improve outcomes while generating savings.
- Key policy actions recommended:
  - Pension reform: increase effective retirement age (as envisaged), introduce automatic adjustment indexed to life expectancy, complement with proportional reduction in accrual rates or extension of minimum contribution period, unify and simplify 42 schemes, accompany with measures to improve working conditions and skills of older workers.
  - Unemployment benefits: entrench recent reforms and revisit eligibility and generosity (ceilings and degressivity) to enhance labor market incentives.
  - Streamline social benefits: simplify and better target housing and social minima schemes; automate, digitalize, and improve data exchange (solidarité à la source); condition RSA eligibility on reintegration efforts; pilot reforms in 19 departments in 2023.
  - Health sector: improve efficiency—review compensation structures, administrative costs, hospital bed utilization, preventive care; scale up medical school admissions (additional 20 percent during 2021-25 vs previous five years).
  - Education: rationalize spending to achieve efficiency savings; rebalance from secondary/tertiary to pre-primary/primary; strengthen teacher recruitment, training, and performance incentives.
  - Local public administration: contain local wage bill and employment growth; standardize procurement; better align devolution with own-revenue sources; restore tax autonomy.
  - Tax expenditures: rationalize or redesign costly/distortive/regressive/excessive measures (notably CIR, housing-related measures, fossil fuel tax expenditures, and savings incentives); strengthen transparency, systematic evaluation, and use of sunset clauses.
- Fiscal potential:
  - Rationalizing tax expenditures coupled with other measures could yield savings ranging from ½ to 1 percent of GDP.

*Source: sipea2023014 chapter excerpt (IMF).*

### References _________________________________________________________________________________ 22

### sipea2023014 - References _________________________________________________________________________________ 22

### Introduction
- Trend increase in primary current spending over the last three decades leaves France with the highest spending ratio in Europe and requires structural spending reforms to rebuild fiscal buffers further eroded by the large fiscal response to successive shocks over 2020–22.
- Annual real primary spending growth has outpaced output growth about half of the time over the past three decades, averaging 2 and 1.5 percent over 1990-2021 respectively.
- The spending gap with peers reached 10 percentage points of GDP in 2019, driven primarily by social benefits (+6½ ppts), followed by the wage bill and subsidies (about +1¼ ppts each), together accounting for 87 percent of the gap.
- Paper’s approach: disaggregate spending by economic and functional classification; focus analysis on social protection, health, education, and economic support; assess efficiency using stochastic frontier analysis (SFA).

### Social Protection — Key Findings
- Scale and gap:
  - Social protection spending: 23.8 percent of GDP in 2019 (second highest in Europe after Finland).
  - Gap relative to peers: about +5.5 ppts of GDP.
  - Peak spending: 24.5 percent of GDP in 2014, later declined due to lower pension, unemployment, and family benefit spending.
- Redistributive outcomes and efficiency:
  - Poverty rate after taxes and transfers: 14.6 percent in 2019 (France 2.5 ppts below peer average per OECD data).
  - Disposable income Gini coefficient: 0.29 in 2019 (in line with simple average of peers).
  - Reduction in Gini due to social transfers slightly higher than peer average (about ½ ppt), but redistributive efficiency is lower because France spends significantly more than peers.
  - Reduction in the Gini coefficient per one percent of GDP of social benefit spending excluding pensions: 0.55 percent for France vs. median for peers 0.65 percent (in absolute terms 0.19 vs. 0.22 percentage points).
  - Targeting: unemployment benefits means-tested only 6 percent (peer average 26 percent); share of untargeted family benefits about 13 ppts above peers.
- Pensions:
  - Public pension spending: about 14 percent of GDP in 2019 (among the highest in Europe).
  - System features: relatively generous, fragmented (42 different regimes), low effective retirement age, high replacement rates.
  - Normal retirement age: 65 (in line with peer average); minimum retirement age: 62 (one of the lowest in Europe).
  - Average age of labor market exit: 60.6 years (2 years below peers).
  - Life expectancy: 0.6 years above peer average.
  - Expected retirement duration: second highest in Europe (shared with Spain, after Greece).
  - Proportionality measure indicates system became relatively more generous since 2008, especially for mid-career workers (Fouejieu et al. 2021).
  - Conseil d’Orientation des Retraites (COR, 2022) projections: higher living standards of retirees relative to rest of population will erode over time as wages grow faster than pension benefits indexed to inflation.
- Employment of older workers:
  - Employment rate 55+ years: 55 percent (well below peer average).
  - Under current policies, participation rate of 65-74 cohort expected to increase by only 15 ppts by 2070, with less than half due to reforms (EC Ageing Report 2021).
- Unemployment benefits:
  - Spending: close to 2 percent of GDP (almost twice peers).
  - Long-term and youth unemployment in France above peer average.
  - Benefit design: minimum duration broadly in line with peers (6 months); maximum duration significantly higher (24 months vs. average 18 months for those aged below 50); older cohorts: 30 months for those aged 53-54 and 36 months for those aged 55 and older (highest among peers except Belgium, which has no ceiling).
  - Replacement rates decline less with benefit duration compared to peers, suggesting weaker labor market incentives.
- Recent reforms and expected effects:
  - 2010 and 2014 pension reforms: raised minimum retirement age and age for full-rate pension by 2 years to 62 and 67; increased minimum contribution period for a full-rate pension (changes effective from 2017 and 2020 respectively).
  - Pension reform plan (presented January 10, 2023): raise minimum retirement age from 62 to 64 at a rate of 3 months per year from September 2023; increase minimum contribution period from 42 to 43 years by 2027; grandfather existing participants; align most special regimes with general regime for new participants (some special schemes excluded).
  - 2019 unemployment reform: adjusted benefit calculation and accumulation rules, tightened eligibility, introduced degressivity for higher incomes, and a modulation of employers’ unemployment contribution (bonus-malus).
  - December 2022 law: extends the 2019 reform application until end-2023 and introduces countercyclicality in unemployment benefits (effective from February via decree).
    - New rule: vary benefit duration with level and change in unemployment rate; 25 percent reduction in maximum duration when unemployment rate is below 9 percent and its quarterly rate of increase is below 0.8 ppts.
    - Duration floor: 6 months; duration extended when labor market conditions deteriorate.
    - Government considered a 40 percent reduction when unemployment rate drops below 6 percent but postponed decision to allow dialogue with social partners and to regulate in the next interprofessional agreement (early 2024).
  - Expected outcomes: reforms aim to improve employment rate of older workers, raise effective retirement age closer to peers, strengthen sustainability, generate significant savings, and improve labor market incentives.

*Source: sipea2023014 (IMF), January 10, 2023.*

### 2024. While the  reform will reduce labor market disincentives by shortening the  duration when  the

### sipea2023014 - 2024. While the  reform will reduce labor market disincentives by shortening the  duration when  the

### Social benefits: scope to streamline
- Housing benefits:
  - Spending on housing benefits: 0.9 percent of GDP.
  - Spending on housing and community amenities: 1.1 percent of GDP (highest among peers; peer average 0.4 percent of GDP).
  - One in five households receive housing support.
  - Three quarters of the population eligible for social housing (Cour des Comptes, 2021c).
  - Overcrowding rate for poor households (below 60 percent of equivalized median income): 22.6 percent (equal to peer average).
  - Problems identified: insufficient and inelastic housing supply in densely populated urban areas, limited turnover and accessibility of social housing despite a large stock, poor targeting of housing support, complexity of housing subsidies (aides personelles au logement) increasing administrative costs.
  - Housing-related tax expenditures described as costly and regressive and could be rationalized (¶18).

- Social exclusion benefits (social minima):
  - Spending: 1.3 percent of GDP (½ percent of GDP more than peers).
  - Comprises 10 different schemes covering about 10 percent of the population.
  - Reduced the poverty rate and intensity by 3 and 8 ppts, respectively, in 2019 (DREES, 2022).
  - Spending concentrated on the four largest schemes (accounting for 95 percent of spending): revenu de solidarité active (RSA), allocation aux adultes handicapés (AAH), allocation de solidarité spécifique (ASS), allocation de solidarité aux personnes âgées (ASPA).
  - RSA (including prime d’activité and administration cost): 0.6 percent of GDP.
  - Real spending growth over 2000-19: 3.4 percent (exceeded peers).
  - Post-GFC recovery phase (2012-18) averaging 3.5 percent (similar to peers).
  - 2019-20 sharp acceleration due to cyclical/discretionary factors and structural flaws.
  - Structural flaws:
    - Low exit rate: only 2-4 percent of beneficiaries find durable employment; 42 percent of RSA beneficiaries remain beneficiaries after 7 years; only a third transition into employment.
    - Widening of eligibility (especially for disability scheme) and limited controls due to fragmentation of roles and responsibilities, including administration and financing.
    - Limited effectiveness of some schemes in preventing poverty (Cour des Comptes, 2021c, 2022b).
  - Reform efforts:
    - 2018 plans to introduce revenu universel d’activité (RUA) were interrupted by the pandemic.
    - Current reform plans (solidarité à la source) focus on automation, digitalization, and better data exchange.
    - Eligibility to RSA to be more strictly conditioned on reintegration efforts through training or work, with rights and obligations modeled after contrat d’engagement jeune (CEJ).
    - Pilot of the reform in 19 (out of 43) departments in 2023.
  - Coverage gaps: Only 70 and 40 percent of those eligible for RSA are recipients of benefits and reintegration support, respectively (Cour des Comptes, 2022b).
  - Fraud/erroneous payments:
    - CNAF estimates fraud at some €1bn in 2019 for AAH and RSA.
    - Cour des Comptes (2020) estimates some €1.4bn, or 12 percent of spending, were wrongly paid out, mostly due to fraud.

- Family benefits:
  - Spending marginally above peer average.
  - Outcomes mixed: slightly worse than peer average for child poverty risk in vulnerable households (proxied by parental educational attainment), but better for fertility and female labor force participation rates.
  - Targeting of family benefits could be improved (¶4).

### Health
- Outcomes and risk factors:
  - France has one of the highest life expectancies compared to peers.
  - Mortality rates from treatable causes lower than peers.
  - Gaps in life expectancy by education similar to peers (and lower for women).
  - High smoking rates and alcohol consumption among disadvantaged socio-economic groups contribute to life expectancy gaps.
  - Chronic conditions and cancer incidence much higher than in peers.
  - Quality of care: effective primary care with lower avoidable hospital admissions; below peers in safety of primary care (high antibiotics prescriptions) and preventive care.

- Spending and composition:
  - Total health spending in 2019: 11 percent of GDP (peer average 9.6 percent of GDP).
  - Public health spending accounts for about 77 percent of total (peers: 70 percent).
  - Public health spending to GDP increased by 1 ppt over past two decades (in line with peers).
  - Compensation of employees in health: 2.2 percent of GDP (peers: 2 percent of GDP).
  - Social transfers in health: 4.3 percent of GDP (peers: 2.7 percent of GDP).
  - Inpatient and outpatient care spending: 4.3 percent of GDP (peers: about 3.9 percent of GDP); these categories account for about 50 percent of all health spending in France.
  - Administrative costs: 0.5 percent of GDP (peers: 0.2 percent of GDP).

- Medical resources and workforce:
  - Number of hospital beds in France far exceeds peers—potential overuse and cost implications.
  - Number of doctors below peers; barely kept up with population growth since 2000 and projected to fall in next few years.
  - Recommendation: increase by an additional 20 percent the number of students admitted to medical schools in France during 2021-25 compared with the previous five-year period (ONDPS, 2021).
  - Density of general practitioners reduced from 1.5 per 1 000 population in 2012 to 1.4 in 2021, with increased regional disparities.

### Education
- Performance and attainment:
  - PISA test scores in secondary education below Germany, the UK, and other advanced economies.
  - Worse than G5 in test scores in math, science, and reading for grades 4 and 8.
  - Only 30 percent of the population have completed a short post-secondary education cycle (UK: 35 percent; Germany: 33 percent).
  - One of the lowest shares of students graduating with a master’s degree or higher among the G5.
  - One of the highest shares of graduates for upper secondary education.
  - Attainment trends likely to continue into 2025 absent structural reforms.

- Spending and composition:
  - Public education spending in 2019: 5.2 percent of GDP (peer average 5 percent of GDP).
  - Spending notably above Germany, Spain, and Italy but lower than the UK.
  - Compensation of non-teaching staff represents a larger share of current expenditure from primary to tertiary education: 22 percent in non-tertiary and 38 percent in tertiary education (peers: 12 and 28 percent, respectively).
  - Expenditure per student:
    - Secondary education: 5 percent higher than peers.
    - Upper secondary level: over 30 percent higher than peers.
    - Primary education: 20 percent lower than peers.
    - Tertiary education: 8 percent lower than peers.
  - Teacher/pupil ratios lower than peers at all levels, slightly higher in upper secondary education.
  - Higher average class sizes could reduce time devoted to learning and teaching.
  - Scope identified for efficiency savings from rationalizing education spending.

### Public Administration
- Wage bill and employment:
  - Wage bill in 2019: 12.2 percent of GDP and over one fifth of public spending.
  - Wage bill about 1.5 ppt of GDP larger in France than in peers; exceeds peers in almost all sectors, with social sectors contributing the most.
  - Government employment about 15 percent of the working age population since early 2000s—much higher than peers and other G4 economies.
  - Share of local government employment about 12 percent of total employment—the highest among peers.
  - Rapid increase in local governments’ wage bill and employment driven by weak hiring practices and rapid promotions at the municipal level (Cour des Comptes, 2009 and 2012).
  - France has a much larger number of municipalities per 100K inhabitants compared to peers.

- Containment challenges and decentralization:
  - Procurement purchases sub-optimal at local level due to large number of municipalities not following standardized processes.
  - Transfers from the center have supported additional and frequently inefficient spending.
  - Golden rule for local governments: can only borrow to finance investment.
  - Indicative expenditure-growth limit for local governments introduced in 2014 (ODEDEL); 2014-18 medium-term programming bill targets largely met by cutting investment spending.
  - Cour des Comptes publishes an annual review of local government finances and reporting (OECD, 2021).
  - Devolution of spending responsibilities could be better coordinated with assignment of own-revenue sources.
  - France categorized as medium decentralized spending and medium tax revenues (OECD, 2019).
  - 2010 local finance reform reduced share of own-source local taxes and thus tax autonomy.
  - Empirical evidence cited: revenue decentralization associated with income gains; a 10 percentage point increase in tax decentralization associated with 0.1 percentage points more economic growth (Blöchliger, Égert and Fredriksen, 2013; Blöchliger and Akgun, 2018).

### Tax expenditures
- Overall:
  - Total spending on economic support via tax expenditures and subsidies: close to 6 percent of GDP (1½ ppts higher than peer average).
  - Two thirds of spending support general affairs and transportation.
  - More than 40 percent channeled through subsidies and tax expenditures, accounting for nearly 90 percent of the spending gap with peers.
  - Including the CICE, tax expenditures amounted to 4 percent of GDP in 2019 (below peer average but above Italy, Spain, Germany).
  - Number of tax expenditures: 394 (second only to Italy at 490); more than 3.5 times higher than the peer average.
  - Costly/distortive/regressive or ineffective tax expenditures include those for fossil fuels, housing, household savings, and R&D.
  - Recommendation: rationalize or redesign these tax expenditures.

- R&D (crédit d’impôt recherche, CIR):
  - CIR size: ¼ percent of GDP (single largest tax expenditure; more than 80 percent of R&D-related and 8 percent of total tax expenditures).
  - Evidence shows positive effect on R&D spending, investment, and innovation but with a relatively low additionality ratio (CNEPI 2021, Appelt et al. 2020, Cour des Comptes 2022a).
  - Disproportionate use by large firms: in 2018 the 100 largest beneficiaries accounted for about a third of total spending; firms with more than 5k employees accounted for 36 percent (up from 32 percent in 2009).
  - In 2019 large firms accounted for only 14 percent of R&D tax relief recipients but 70 percent of benefits (OECD, 2021e).
  - Evidence of declining innovation impact with firm size: return on CIR spending (patents per euro spent) 2.5 times larger for micro firms than for large firms (2.9 times when considering patents covering multiple jurisdictions) (Aghion et al., 2022).
  - Design feature: reduced CIR rate of 5 percent for R&D spending above €100mn (vs. 30 percent below).
  - Estimates by Le Gall et al (2021): CIR reform generates 0.8pp additional growth and 60k jobs in the long term, implying an average annual cost of more than 83k per job created.

- Housing tax expenditures:
  - Size: ½ percent of GDP in 2019 (well above peer average for available data: 0.2 percent).
  - Account for some 16 percent of tax expenditures (excl. CICE) and equal more than 80 percent of the budget allocation for territorial cohesion.
  - Effectiveness assessments inconclusive or point to limited effectiveness, significant cost, and/or undesirable effects (Cour des Comptes, 2022a).
  - Examples: some tax exemptions and rebates for social housing organizations undermine objectives; reduced rates on housing maintenance, development and renovation work have limited impact on employment and tend to be regressive (CPO, 2015).

- Fossil fuels:
  - Fuel tax expenditures: ⅓ percent of GDP in 2019 (more than double the peer average).
  - Mostly reduced rates for fossil fuel use in transport and agriculture.
  - Some phased reductions (road freight transport and non-road diesel) but other fossil fuel tax expenditures remain, contributing to pollution, congestion, and undermining climate goals; costly and possibly underestimated (EC, 2022).

- Savings:
  - Household savings incentives: about ¼ percent of GDP, largely exemptions for wage and housing savings.
  - Given saving surplus accumulated during the pandemic, reduction or elimination of savings incentives could be considered (revisit favorable tax treatment for housing investment and sale of immovable property) (OECD, 2021a).

### G. Conclusions and Policy Recommendations
- Reforms underway and proposed:
  - Labor market reform will reduce disincentives by shortening benefit duration when the labor market is relatively tight; scope remains to revisit level and duration of benefits, which remain relatively high.
  - Streamline social benefits: simplify and better target housing and social minima schemes; automate, digitalize, and improve data exchange for access (solidarité à la source); condition RSA eligibility on reintegration efforts and pilot reforms in 19 departments in 2023.
  - Housing policy: address insufficient and inelastic urban housing supply, improve targeting and accessibility of social housing, simplify housing subsidies, and rationalize housing-related tax expenditures (¶18).
  - Social minima: increase effectiveness of reintegration policies, tighten controls and coordination across fragmented administration and financing.
  - Health sector: address high spending by improving efficiency—review compensation structures, administrative costs, hospital bed utilization, and preventive care; scale up medical school admissions (additional 20 percent during 2021-25 vs previous five years) and address regional disparities in general practitioner density.
  - Education: rationalize spending to achieve efficiency savings, improve targeting and outcomes given relatively lower PISA scores and post-secondary attainment.
  - Public administration and local governments: contain local wage bill and employment growth, standardize procurement processes, better align devolution of spending responsibilities with own-revenue sources, and restore tax autonomy to strengthen local budget constraints.
  - Tax expenditures: rationalize or redesign costly, distortive, regressive, or ineffective tax expenditures (notably R&D/CIR, housing-related, fossil fuels, and household savings incentives); reassess the distributional and effectiveness outcomes of major tax support programs.

*Source: IMF staff analysis from the provided chapter content.*

### 19.      The analysis    shows that higher spending in France relative to peers is not always

### 19.      The analysis    shows that higher spending in France relative to peers is not always 

### Overview
- Higher spending in France relative to peers is not always commensurate with better outcomes, leaving scope to improve the efficiency of spending.
- Reform priorities span social protection, health, education, and the civil service to preserve or improve outcomes while generating savings to help meet medium-term adjustment needs.
- Rationalizing costly, distortive, or inefficient tax expenditures would allow for base broadening and partially offset permanent revenue losses from rebalancing revenues away from labor and production taxes.
- Findings and policy recommendations are presented by spending category below.

### Social protection: findings and reforms
- Social protection spending explains the bulk of the spending gap with peers.
- Higher spending is associated with better outcomes on poverty risk and income inequality by age, employment status, or educational attainment of parents, but efficiency can be improved.
- An estimate of France’s social protection efficiency score is above the peer average but well-below that of the best performers.
- Policy recommendations:
  - Pension reform:
    - Increasing the effective retirement age as envisaged in pension reform plans is a step in the right direction.
    - Introducing automatic adjustment by indexing the retirement age to life expectancy—as in Denmark, Finland, Greece, Italy, Netherlands, and Portugal (OECD 2021d)—would further enhance sustainability.
    - To generate savings while maintaining the same replacement rate, this should be complemented by a proportional reduction in accrual rates or extension of the minimum contribution period.
    - Unifying and simplifying the highly fragmented system—covering all 42 schemes—would improve equity, lower administrative cost, and facilitate labor mobility.
    - Accompany pension reform with measures to improve working conditions (including flexible work arrangements and partial retirement) and skills of older workers to facilitate longer careers and increase employment of older cohorts.
  - Unemployment benefit reform:
    - Entrench recent unemployment benefit reforms and further revisit eligibility and generosity, especially regarding benefit ceilings and degressivity, to enhance labor market incentives and generate additional savings.
  - Reform of other social benefits:
    - Strengthen means-testing of family, housing, and social assistance benefits to improve redistributive efficiency and generate savings.
    - Simplify and unify social assistance schemes, differentiating for special conditions, to reduce complexity and facilitate better monitoring and control of eligibility.
    - Sequence reforms: harmonize means-testing across schemes first before simplifying and unifying schemes (Cour des Comptes, 2021c).
    - Ensure automation, digitalization, and better data exchange (solidarité à la source plan) are accompanied by enhanced efforts to reintegrate beneficiaries (Cour des Comptes, 2022b).

### Health: efficiency gains and cost‑effectiveness
- There is significant room for efficiency savings in health spending without compromising outcomes.
- The Recovery and Resilience Plan envisages modernization that could achieve greater efficiency, including wider implementation of technological health and eHealth systems, encouraging R&D, training and scientific research, and greater investment in elderly care facilities.
- Additional reforms to obtain efficiency savings and improve cost-effectiveness:
  - Reduce spending on inpatient care, medical goods, and administrative costs by restructuring the public hospital network and developing primary care and ambulatory centers; improve coordination between ambulatory and hospital care.
  - Improve management of human resources and equipment purchases in hospitals.
  - Develop pluri-professional care exercises and houses by pooling resources, including generalists, to increase quality and rationalize expensive equipment and administrative costs (CdC, 2021).
  - Invest in prevention relative to curative care and discourage unhealthy behaviors to reduce costs associated with hospital admissions for cardiovascular, cancer, and psychiatric diseases; increase incentives for preventive action (e.g., higher remuneration for doctors for preventive action) and promote tobacco and alcohol consumption control policies.
  - Shift remuneration away from fee-for-service or per admission toward lump-sum payments; consider an annual individualized per-patient budget based on patients’ health condition and needs to contain expenditure linked to chronic diseases (CdC, 2021).
  - Align costs to the burden and complexity of care; current pricing favors most expensive modes and includes thousands of distinct items with complex definitions that do not reflect production costs.
  - Encourage doctors to increase the share of generics in prescriptions: the share in France is 16 percent in value and 30 percent in volume, much lower than Germany or the UK (35 percent in value and more than 80 percent in volume).
  - Develop core care services for pathologies and best practices in drug prescriptions, including generics, and align and adapt tariffs to pathologies to reduce redundant/inadequate prescriptions.

### Aging and long-term care pressures
- Demographic projections and long-term care implications:
  - 24 percent of the French population will be aged 65 or over in 2030.
  - Based on the French long-term care indicator, the number of long-term care recipients will increase by 76 percent from 2015-2050.
  - According to the 2021 Ageing Report projections, public expenditure on LTC is projected to increase in relation to GDP by 0.7 ppt (or 37 percent) between 2019 and 2050 (from 1.9 percent in 2016 to 2.6 percent in 2050).
- Achieving efficiency savings is important given these build-ups from aging and medical progress.

### Education: structural reforms to improve efficiency and equity
- More ambitious structural reforms are needed to achieve efficiency savings while addressing inequities.
- Suggested reforms:
  - Rationalize excessive education resources in secondary and tertiary education and rebalance spending from secondary and tertiary towards pre-primary and primary levels.
  - Better adapt the structure of expenditures to changing student demographics (e.g., using overtime to anticipate declining school demographics and avoid hiring) (CdC, 2021).
  - Strengthen selectivity of teacher hiring, particularly in scientific subjects, and use assessments to steer the education system.
  - Strengthen training of teachers (initial education and lifelong training) and collaborative practices; French teachers are less prepared on pedagogical aspects and receive less training related to in-class pedagogy and collaborate less with other teachers (TALIS, 2018).
  - Align teacher compensation with performance and challenges in difficult areas to reduce educational inequities:
    - The 2023 budget envisages an increase of teachers’ salaries by 10 percent and the guarantee that no teacher would have a starting salary of below Euros 2000.
    - This unconditional increase would be accompanied by another 20 percent increase in salaries if supplementary educational missions are performed.
  - Give more responsibilities and autonomy to school administrations to allocate resources according to locally developed projects (CdC, 2021).
  - Provide better access to schooling from early ages for children from low socio-economic backgrounds and improve outcomes for schools in difficult areas; consider providing internet access and leveraging online learning.

### Local public administration and decentralization
- Efficiency in local public administration is critical to ensure decentralization benefits.
- Key points and recommendations:
  - Ensure adequate subnational capacity and transparent multi-level governance with efficient coordination mechanisms across levels of government.
  - Clarify responsibilities assigned to different government levels to help contain local spending, including on the wage bill.
  - Increase tax autonomy to impose a hard budget constraint; consider introducing new own-source local taxes and broadening local decision-making power on rates or based on existing local taxes (e.g., property tax).
  - Update property valuations, as out of date property values have been identified as lowering local property tax revenues (OECD, 2022).
  - Prioritize fiscal capacity of regions to support vocational training while clarifying roles of national and regional vocational training councils to avoid interference (OECD, 2022).
  - Centralized electronic procurement would help standardize processes and up-skilling the local civil service would improve efficiency of local procurement.

### Tax expenditures: rationalization and redesign
- Rationale:
  - Tax expenditures that are costly, distortive, regressive, or ineffective should be rationalized (e.g., fossil fuels and housing) or redesigned (e.g., R&D).
- R&D tax credit (CIR) recommendations:
  - Enhance effectiveness by lowering the ceiling on the tax credit, raising the CIR rate, and/or differentiating the CIR rate by firm size (with a higher rate for smaller firms), and reduce overlap between R&D incentives (Aghion et al, 2022; Cour des Comptes 2021, 2022a; CPO 2022).
- Potential fiscal impact:
  - Coupled with rationalizing other tax expenditures, this could yield savings ranging from ½ to 1 percent of GDP.
- Governance of tax expenditures:
  - Transparency and reporting on tax expenditures has improved.
  - New fiscal and social tax expenditures will be subject to 5 and 3 year sunset clauses, respectively, allowing for review before possible extension.
  - Systematic evaluation of tax expenditures could be strengthened to improve effectiveness, reduce cost, and eliminate those that fail to meet objectives.

### Appendix I — Benchmarking methodology (summary)
- Benchmarking compares country performance to peers and builds an "efficiency frontier" from best performers to assess distances (inefficiency).
- Efficiency concepts:
  - Technical efficiency: ability to produce a given output with minimum inputs.
  - Allocative (cost) efficiency: cost-minimizing use of inputs given input prices.
- This paper focuses on technical efficiency due to lack of comparable multi-country data on prices.
- Estimation approaches:
  - Parametric methods (e.g., stochastic frontier analysis, SFA) require assumptions on error distributions and functional form but allow separation of inefficiency from noise.
  - Non-parametric methods (mathematical programming) do not require distributional or functional form assumptions but do not include randomness and thus are sensitive to outliers or noise.

*Italic: Source — sipea2023014 chapter excerpt (IMF).*

### 5.      This paper employs an SFA model to estimate the efficiency of social protection,

### sipea2023014 - 5.      This paper employs an SFA model to estimate the efficiency of social protection,

### Methodology
- The model is a Stochastic Frontier Analysis (SFA) with the inefficiency term assumed to follow an exponential distribution.
- Functional form (as presented):
  - 푦푦푖∗ = 푓푥푖1, 푋푖−1, 훼
  - where 푦푦푖∗ is the optimal outcome, 푥푖1 is government spending in a specific function, 푋푖−1 = {푥푖2, .... , 푥푖푘} is a vector of other determinants of 푦푖∗, and 훼 = (훼0, ..., 훼푘) is a vector of parameters.
- Inefficiency and shocks:
  - Inefficiencies: 0 < 휀푖 = exp(−푢푖) ≤ 1
  - Shocks: 푣푖
  - Observed outcome: 푦푖 = 푦푖∗ 휀푖 e(푣푖)
- Production form assumed: 푦푖∗ = c ∗ ∏(푥푖푗)훼푗
- Log-linear estimation specification used:
  - ln(푦푖) = 훼0 + ∑훼푗 ln(푥푖푗) (for j = 1 to k) + 푣푖 − 푢푖
- Inputs and controls:
  - Public spending on health, education, and social protection used as inputs.
  - Weighted average of several standardized outcome measures used as outcomes in these areas.
  - SFA models control for private spending, the level of development, and other relevant determinants of the outcomes.

### Results — Social Protection
- Main finding:
  - France can improve the efficiency of its social protection spending.
- Evidence:
  - Higher public spending on social protection in France is associated with better outcomes when controlling for private spending, the level of development, old-age dependency ratio, and unemployment rate.
  - The social protection index (comprising working- and old-age poverty rates, disposable income Gini coefficient, long-term unemployment rate, school enrollment rate, and housing overburden rate) is the second-highest after Finland.
  - France spends significantly more than any other European country on social protection.
  - France’s efficiency score is above the peer average but well-below that of the best performers: Finland, Portugal, Ireland, and Denmark.
- Efficiency score axis range (as shown): 0.70, 0.75, 0.80, 0.85, 0.90, 0.95, 1.00

### Results — Health
- Main finding:
  - France appears to have some room for efficiency savings in health spending.
- Evidence:
  - Health outcome indicator comprises life expectancy at birth, healthy life years, infant mortality, and health satisfaction level.
  - The score in France implies potential efficiency savings in health spending.
  - Best performers (lower spending for a given outcome): Spain, Sweden, Belgium, Norway, the Netherlands; Germany also performs better than France.
- Efficiency score axis range (as shown): 0.8, 0.85, 0.9, 0.95, 1

### Results — Education
- Main finding:
  - France appears to have some room for efficiency savings in education spending.
- Evidence:
  - Education outcome indicator comprises PISA scores, education satisfaction level, population speaking a second language, and secondary and tertiary education attainment.
  - The score in France implies potential efficiency savings in education spending.
  - Best performers: Ireland, Germany, and Finland; several other advanced economies also perform better than France given relatively lower spending for a certain level of the composite education outcome.
  - By education level:
    - Potential savings mostly in secondary and tertiary education spending.
- Efficiency score axis ranges (as shown):
  - Overall education: 0.70, 0.75, 0.80, 0.85, 0.90, 0.95, 1.00
  - Primary education: 0.75, 0.80, 0.85, 0.90, 0.95, 1.00
  - Secondary education: 0.85, 0.90, 0.95, 1.00
  - Tertiary education: 0.0, 0.2, 0.4, 0.6, 0.8, 1.0

*Sources: OECD, Eurostat, and IMF staff calculations.*

---


_Source: https://www.imf.org/-/media/files/publications/selected-issues-papers/2023/english/sipea2023014.pdf_