From Containment to Rationalization
Increasing Public Expenditure Efficiency in France

Achieving France’s medium-term fiscal targets will require significant expenditure efforts. This paper identifies areas where there is scope for increasing expenditure efficiency, with a view to achieving higher quality and more sustainable fiscal consolidation. The methodology is based on a triple benchmarking. First, the level of public expenditure in different categories is compared to other European countries. Second, the impact of spending is assessed against other European countries. Third, the input mix is analyzed to understand what components are responsible for the level of spending and for the quality of outcomes This is done for various categories of spending and policies. Based on these results, the paper then provides policy options for expenditure reform in each of these areas, drawing on successful reform episodes in other countries.

Abstract

Achieving France’s medium-term fiscal targets will require significant expenditure efforts. This paper identifies areas where there is scope for increasing expenditure efficiency, with a view to achieving higher quality and more sustainable fiscal consolidation. The methodology is based on a triple benchmarking. First, the level of public expenditure in different categories is compared to other European countries. Second, the impact of spending is assessed against other European countries. Third, the input mix is analyzed to understand what components are responsible for the level of spending and for the quality of outcomes This is done for various categories of spending and policies. Based on these results, the paper then provides policy options for expenditure reform in each of these areas, drawing on successful reform episodes in other countries.

I. Executive Summary

At 57½ percent of GDP, public expenditure in France is among the highest in the world. Spending has outpaced GDP growth for over three decades. Notwithstanding successive tax increases, France experienced chronically large fiscal deficit and a growing debt burden, approaching 100 percent of GDP. The high levels of government spending and debt are limiting the fiscal room for maneuver and imposing a substantial tax burden on the economy.

The fiscal consolidation that started in 2011 was initially supported by revenue-raising measures but is now intended to be fully expenditure-based. It aims to bring the overall deficit below 3 percent of GDP by 2017, turn around the growth in public debt, and achieve structural fiscal balance over the medium term.

Identifying areas for savings has proved difficult, and there is no clearly articulated consensus on where spending is too high or inefficient. This is in part because of concerns about the social and economic impact of specific spending cuts, in particular the impact on inequality. Spending measures have thus mainly relied on across-the-board measures to limit nominal spending growth. They have focused on central government and health, while local governments and social security funds spending -including health- have continued to grow faster than GDP.

A shift from this containment approach to broader and deeper efficiency-oriented reforms would increase the chance of success and the sustainability of the ongoing fiscal consolidation, while protecting the French social model. France has recently initiated some steps for structural savings e.g., family allowances, health, and pension.

This paper identifies areas where there is scope for greater expenditure efficiency in France, while maintaining or even improving social and economic outcomes consistent with social preferences. By nature, this requires an assessment not only of fiscal costs but also the intended results, such as the achievement of social objectives, and the provision of high quality public goods and services.

We rely on a three-step benchmarking. First, the level of public expenditure is compared to other European countries with a focus on Germany, Italy, and the United Kingdom. They are large economies with comparable income levels per capita with France. Second, social and economic outcomes in each spending area are assessed against the performance in European peers. Third, the input mix is analyzed to understand what components are responsible for the level of spending and for the quality of outcomes.

This exercise leads to several conclusions:

  • Shifting from containment to deeper efficiency-oriented reforms could yield significant fiscal savings. Most could be achieved by rationalizing social benefits and the wage bill, which explain about 90 percent of the difference in the expenditure ratio between France and the EU average as well as the average for Germany, Italy, and the United Kingdom.

  • The wage bill accounts for 13 percent of GDP and almost one quarter of public spending. Recent efforts have focused on a wage-scale freeze, but low inflation has limited the effectiveness of this approach. Employment reduction (notably at the local level) and measures to limit the wage drift would promise greater scope for efficiency gains.

  • At over 8 percent of GDP, health spending is high by EU standards. While health outcomes are good, they are similar to comparator countries whose health spending is lower. Building on the National Health Strategy of 2014, France could consider reforms implemented in other countries such as further improving generics market penetration, rationalizing hospital services and streamline costs, and strengthening cost-effectiveness evaluations to decide which services should be covered by public insurance.

  • There is significant scope to improve the impact of fiscal redistribution on inequalities and poverty through reforms of the welfare and pension systems. France has the largest social spending in Europe and the second highest tax-to-GDP ratio, but the reduction in inequality due to transfers is only slightly above the EU average. If the redistributive power of social benefits was at EU average, France could achieve the same reduction in income inequality at a fiscal cost lower by 3.5 points of GDP. Moreover, the social outcomes and poverty impact are uneven. Social protection benefits mostly the elderly due to a generous pension regime. While long-term demographic trends are more favorable than in many European countries, additional pension reforms would support consolidation and, together with a further increase in means-testing of family-related spending, make room for more resources to address child and youth poverty.

  • The unemployment benefits system, which accounts for about two-third of labor market policy spending, is comparatively generous.

  • The allocation of resources in education is less efficient than in many European countries, particularly at the secondary level, and has failed to address deteriorating test scores and rising educational inequalities. Organizational reforms could help improve both education quality and social outcomes, for instance by better allocating teaching resources to the neediest, rationalizing inefficient spending (especially in secondary education), and improving the targeting of vocational education and training for those who have difficulties getting a job.

  • Public investment spending, which is at the European average, should focus more on maintenance rather than expansion given France’s high quality and quantity of infrastructure. Rationalizing local and state-owned enterprises investment would avoid duplication.

  • Spending on housing is higher in France than in other European countries but outcomes do not appear much better than in other EU countries. This suggests potential for higher means-testing and lower institutional fragmentation and duplication.

II. Introduction

1. Public expenditure reached 57½ percent of GDP in 2014; about 11 percentage points (ppts) above the EU average2 and the second largest after Finland. The reason is high current spending (notably social benefits and the wage bill), while public investment is in line with the EU average.

2. As a result, and despite a high tax ratio, France has experienced persistent fiscal deficits and rising public debt. Therefore, France has little fiscal space to cope with unexpected needs or economic shocks, including a possible rebound of interest rates from their currently historically low level.3 It also limits the scope for reducing the tax burden on firms and households, as planned by the government and as may be necessary to raise potential growth, estimated by IMF staff at 1 ½ percent over the medium term.

3. The fiscal consolidation, initiated in 2011, was initially revenue-based. Revenue, which had been stable at about 49 ½ percent of GDP in the 2000s, increased by about 4 points between 2010 and 2014 to reach 53.6 percent of GDP (8 ppts higher than EU average). France, who had the fourth highest tax ratio in the EU in 2009, now ranks second.

4. In 2014, the government announced a shift to expenditure-based consolidation while simultaneously reducing taxes. Identifying areas for savings proved difficult however, and spending measures have so far relied on across-the-board measures to limit nominal spending growth, especially for the central government and health. Social security funds have continued to grow faster than GDP. A shift from this partial containment approach to broader and deeper efficiency-oriented reforms would increase the chance of success of the planned fiscal consolidation, while protecting the French social model. However, at this point, there is still no clearly articulated consensus on where spending is too high or inefficient.

5. This paper aims to identify areas where significant efficiency gains may be achieved and thus where spending reform could have a large pay off. By nature, it will focus on areas where improvements are possible or desirable but little will be said of areas where public expenditure achieves good results by international standards. To estimates efficiency gains various methods are possible. Rather than relying on econometric measures, our approach is to focus on a more in-depth analysis to identify spending drivers in each sector. More precisely, our analysis relies on a benchmarking for both spending and outcomes leading to policy recommendations that draw on lessons from successful expenditure reforms in other advanced countries. We compare France to other EU countries with a focus on three peers: Germany, Italy, and the United Kingdom. Then, the input mix is analyzed to understand what components are responsible for the level of spending and for the quality of outcomes. To the best of our knowledge, this paper represents the first comprehensive and in-depth comparative study on public expenditure efficiency in France. As a result, it does not present a literature review section but each section reviews sector-specific literature, drawing from a large array of both country-specific and international sources.

6. The outline of the paper is as follows. The first section examines France’s recent expenditure containment approach comparing its results with the fiscal consolidation carried in the EU. The next section focuses on the composition of public expenditure in economic and functional terms, and identifies areas where there is scope for reforms to improve efficiency and equity. The final section discusses reform priorities and offers policy conclusions.

III. Limitation of the Expenditure Containment Approach

A. Expenditures Trends across Europe

7. General government expenditure in France is significantly above the European average and comparators (Germany, Italy, and the United Kingdom).4 In 2014, France spent about 11 ppts of GDP more than the European average. At about 13 ppts, the gap is wider for Germany and the United Kingdom. Apart from France, only Finland and Denmark’s public expenditure is above 55 percent of GDP (Figure 1).

Figure 1.
Figure 1.

Evolution of General Government Expenditure

Citation: IMF Working Papers 2016, 007; 10.5089/9781513540139.001.A001

Sources: IMF World Economic Outlook and Staff calculations.

8. Spending has outpaced GDP growth for decades (Figure 1). This trend has become more pronounced since the early 2000s. Spending growth remained well above EU rates even as it slowed down in the post-crisis consolidation period (Table 1). By contrast, spending growth became negative in comparator countries such as the United Kingdom and Italy, as well as the euro area on average.

Table 1.

Real Primary Expenditure and GDP Growth 1/

article image
Sources: Eurostat and IMF staff calculation.

Deflated by GDP deflators. Spending by levels of government includes transfers across levels of government, which are netted in general government data.

9. Spending growth was primarily driven by local governments and social security funds (Table 1). In the 2000s, spending containment has been largely limited to the central government in France. In fact, until 2010, France’s central government spending grew in real terms more slowly than in euro area comparators. In recent years, central government spending continued to grow, albeit more slowly, while other countries were cutting spending to support large fiscal consolidation. By contrast, at the local government and social security levels, spending growth has continued to outpace GDP and most comparators.

  • For local governments, the substantial increase in spending has been a long term trend, which is only partially related to successive waves of fiscal decentralization. Local governments’ spending increased by 3.2 ppts of GDP from 1983-2014, of which 1.65 ppt is due to fiscal decentralization (Draft 2016 budget law, 2015). Spending growth at the local governments level decelerated during 2010-14, while it was reduced in other European countries (except Germany).

  • Social security funds’ spending has been growing at around twice the rate of GDP, and well above euro area average and comparators since the early 2000s.

10. While public spending increased less in France in the immediate response to the crisis, it continued to grow afterwards as consolidation was initially revenue-based. Many European countries saw a jump in spending during the crisis as a result of discretionary stimulus measures, the impact of automatic stabilizers, and in some cases the cost of bank bailouts. However, they reverted to cutting primary spending in the post-crisis consolidation period (Figure 2 and Table 1). Although structural adjustment was similar to the EU average, expenditures in France increased during the whole period as the consolidation strategy relied initially on raising revenue (Figure 2). The expenditure containment started in earnest in 2013 and fiscal consolidation is now expected to be fully expenditure-based.

Figure 2.
Figure 2.

Fiscal Developments During the Crisis

Citation: IMF Working Papers 2016, 007; 10.5089/9781513540139.001.A001

Sources: Eurostat and IMF Staff calculations.

B. The Expenditure Containment Approach

11. France relied on spending containment rather than on spending reforms. Spending containment was largely implemented across the board. These measures include the freezing of the public wage scale and of some social benefits. Among European countries, France was the least selective both in implementing its fiscal stimulus in response to the crisis and in carrying its spending containment (Table 2).

Table 2.

Selectivity in Spending Cuts/Increases

article image
Sources: OECD and IMF staff calculation.Note: Countries are ranked according to the value of the selectivity in 2010-13. Spending is in national currency deflated by the GDP deflator. The closer to 0 the indicator the more spending change was similar across classifications. For details on the methodology, see Lorach and Sode (2015).

For Spain up to 2010-12.

For Belgium, Italy, and Sweden: 2001-07.

Out of a maximum of 66 COFOG categories as three categories are excluded (0107, 0401, 0407).

12. In 2014, France announced an expenditure saving plan that would however maintain its public spending above comparator countries. The planned savings of €50 billion relative to trend over 2015-17 would imply an effort of 2.2 percent of GDP distributed at all levels of government (Figure 3). Cumulatively, the cuts represent about 4 percent of spending for each level of government (central, sub-national, and social security). Over 2014-18, the authorities project the spending-to-GDP ratio to decline by 2.4 ppts of GDP to 54.8 percent of GDP. However, spending would remain high compared to other countries, and the gap with Italy and the United Kingdom would further increase, as these countries are planning to reduce spending by 3.1 and 8.1 ppts of GDP over the same period.

Figure 3.
Figure 3.

Distribution of the Expenditure Containment Plan by Level of Government

(In billions of Euros)

Citation: IMF Working Papers 2016, 007; 10.5089/9781513540139.001.A001

Source: Draft 2016 budget law.

13. The proposed savings are at different stages of identification. First, the wage freeze, de-indexation of social benefits, and impact of recent reforms in social benefits are well identified. Second, central government savings (outside of wage-scale freeze) are announced on a yearly basis as part of the budget. Third, savings in health spending are identified and rationing mechanisms are in place to ensure execution, although health spending is subject to several factors that fall outside the control of the authorities. Fourth, of the social spending saving, about €4 billion are predicated upon the recent reform of supplementary pensions and the forthcoming reform of unemployment. Fifth, indicative saving targets are set for local governments. However, the degree to which the cut in the state transfers will translate in spending cuts or will be offset by an increase in taxes or debt is uncertain.

14. Deeper spending reforms are needed to underpin fiscal consolidation and eventually create space for alleviating France’s high tax burden (IMF, 2015a). To achieve the significant reduction in public expenditure needed to reach the fiscal targets, the containment strategy may prove insufficient. Without deeper reforms, it will be difficult to make the most efficient use of public resources and there are significant risks that savings fall short of the intended targets.

IV. Benchmarking Public Spending Efficiency

15. Benchmarking public expenditure helps identify areas where France diverges from comparators. Based on both economic and functional classifications, this paper benchmarks France’s public spending against the European average and comparators. These areas constitute either sources of possible savings or risks when further rationalization may unnecessarily lower the quality of services provided. In economic terms, we focus on social benefits, goods and services, the wage bill, and capital spending. In functional terms, we benchmark pensions, health, unemployment, education and vocational training, and housing.

16. Evaluating public expenditure efficiency requires benchmarking both spending and related outcomes. Comparing the levels of public spending across countries is insufficient to determine areas in which expenditure rationalization is desirable. Public spending levels and the role of the public sector are determined by social preferences, and reflect countries economic and demographic characteristics. For some policies, such as education, pensions, and health, countries may have different approaches to involving the private sector. Looking at public spending-related outcomes allows a more informed judgment on their efficiency and realistic approaches to spending rationalization. A particular attention will be given to equity considerations and the redistributive impact of public spending.

A. Public Expenditure from an Economic Classification Perspective

17. Public expenditure is much larger in France than on average in the EU because of current spending High public expenditure in France is not driven by public investment, which is at the EU average, though higher than comparator countries. The key to understanding France’s public expenditure is current spending (Figure 4), which is in the vicinity of Nordic levels and exceeds the EU median by about 11 ½ percentage points of GDP. Moreover, this gap has widened: the share of primary current spending in GDP has increased by 1.4 ppt since 2010 while it declined in the EU by 0.8 ppt (Table 3).

Figure 4.
Figure 4.

Public Capital and Current Spending in EU Countries, 2014

Citation: IMF Working Papers 2016, 007; 10.5089/9781513540139.001.A001

Source: Eurostat.Note: Dashed lines indicate medians.
Table 3.

France and EU General Government Expenditure by Economic Classification 1/

article image
Source: Eurostat.

When the analysis was undertaken, Eurostat data differed slightly from INSEE in 2014. According to INSEE, expenditures reached 57.5 percent in 2014.

18. Spending on social benefits and on civil servants accounts for 90 percent of the difference between France and comparators. In 2014, the share of social benefits in GDP was 43 percent higher than the EU average and the wage bill ratio was 20 percent higher. While the comparators and the EU as a whole reduced spending in these two categories during the fiscal consolidation, France made no savings on the wage bill and spending on social benefits increased by 1 ppt. This suggests significant efficiency gains, which we will explore below.

Social Benefits and the Reduction in Inequalities and Poverty

19. Social benefits constitute half of current spending.5 Social benefits are 7.8 ppts of GDP larger in France than on average in the EU, explaining almost three quarters of the difference in total expenditure-to-GDP ratio (Table 3).6 France spending in this area is above countries with similar income per capita (Figure 5) and even countries with higher income, including the Nordic countries.

Figure 5.
Figure 5.

Social Benefits in European Countries, 2014

Citation: IMF Working Papers 2016, 007; 10.5089/9781513540139.001.A001

Sources: Eurostat and IMF staff calculations.

20. High spending on social benefits translates into sizable levels of fiscal redistribution. Fiscal redistribution increases the income of the bottom quintile, does not affect income of the second quintile, and reduces the income of the three highest quintiles. As reported in Table 4, the benefit of fiscal redistribution concentrates on the lowest quintile, whose income increases by 60 percent (by 164 percent for the lowest decile). The cost of fiscal redistribution increases with income, with the revenue of the two top quintiles being reduced by more than 10 percent.

Table 4.

Impact of Fiscal Redistribution by Income Level, 2014

(In Euros by consumption unit and by quintile before redistribution)

article image
Source: INSEE (2015).Excludes transfers in kind and contributions to and transfers from pensions and unemployment benefits.

21. Though France has the highest level of social expenditure and the second largest tax-to-GDP ratio in Europe, it does not achieve the largest reduction in inequality (Figures 6, 7, and 8).7 Taxes and transfers reduce inequality (Gini coefficient) by almost 45 percent. In the EU, France appears in an intermediate position both for market (i.e., pre-tax-and-transfer) and disposable (i.e., post-tax-and-transfer) income inequalities. While market income inequality is below the three comparators, they are above other EU countries with similar level of income per capita (Belgium, Denmark, Finland, the Netherlands, and Sweden). After fiscal redistribution, disposable income inequality remains higher in France, despite a higher level of public spending than in Belgium, the Netherlands, and Sweden.

Figure 6.
Figure 6.

Tax Revenue

(2009-2013, in percent of GDP)

Citation: IMF Working Papers 2016, 007; 10.5089/9781513540139.001.A001

Sources: OECD.
Figure 7.
Figure 7.

Market and Disposable Income Inequality in Europe, 2013

Citation: IMF Working Papers 2016, 007; 10.5089/9781513540139.001.A001

Sources: Eurostat and IMF staff calculations.
Figure 8.
Figure 8.

Contributions of Fiscal Policy to the Reduction of Income Inequality in Europe, 2013

Citation: IMF Working Papers 2016, 007; 10.5089/9781513540139.001.A001

Sources: Eurostat and IMF Staff calculations.Note: SC= Social Contributions; DT=Direct Taxes; MT=Means-tested social spending; NMT=Non-means-tested social spending.

22. The redistributive power of social spending (i.e., the reduction in income inequality due to 1 percent of GDP in social spending) is low by European standards (Figure 8). Fiscal redistribution achieves a smaller reduction in inequality than in Ireland and Germany, where total public spending is significantly below France. If France could raise the redistributive power of social spending to the EU (weighted) average, it could achieve the same reduction in inequality with 3.5 points of GDP lower spending. This would require deeper reforms in a number of programs (see next section).

23. Particular emphasis should be placed on means-testing of benefits and on better targeting the poor. The low redistributive impact is explained by a high share of social spending benefits the richer households: 27 percent of social cash benefits go to the highest income quintile, and less than 17 percent to the lowest income quintile. Among EU (21),8 only Portugal, Greece, and Italy have a higher share of social benefits going to the highest income quintile (OECD, 2014b). The large share of social benefits received by richer household is mostly accounted for by the large pension payments (see below), and by relatively low means–testing. About 11 percent of total social expenditures (in kind or in cash) is means-tested. This is slightly more than EU average but less than in Germany and the United Kingdom (Figure 9). The reason is that a large share of family-related spending is not targeted, but accounts for a significant share of fiscal redistribution because of the sheer size of the amount involved (Table 5).9 The 2014-15 reform increased the means-testing of family transfers and should improve its efficiency though there remains scope for increased targeting.

Figure 9.
Figure 9.

Share of Means-tested Social Expenditures in Europe, 2012

Citation: IMF Working Papers 2016, 007; 10.5089/9781513540139.001.A001

Sources: Eurostat and IMF staff calculations.Note: Dashed lines represent EU medians.
Table 5.

Contribution of Taxes and Transfers to the Reduction of Inequalities 1/

(In percent)

article image
Source: INSEE, “France, Portrait social,” various years.

Excludes transfers in kind and contributions to and transfers from pensions and unemployment benefits.

24. Much of France’s social spending benefits the elderly. The OECD (2012) calculates that, while the elderly (65 year old and more) accounts for slightly less than 17 percent of the population, they receive 48 ½ percent of public social and education expenditures (or above 18 percent of GDP) in 2009 (Table 6). This is 10 ppts higher than the OECD average (and the second largest share in the EU after Italy) despite the share of elderly in the population being only 3 ppts higher in France. In addition, the tax wedge on the elderly is below other age groups, even when excluding social contributions, which are mostly paid by in-work population (Conseil des prélèvements obligatoires, 2008).

Table 6.

The Focus of Social and Education Spending on the Elderly, 2009

(In percent of GDP)

article image
Source: OECD (2012).

Italy is the OECD country that dedicates the largest share of GDP in public social and education spending to elderly, and is second after Japan for the share of public social and education spending dedicated to elderly.

Japan has both the largest share of elderly in the population and the largest share of public spending dedicated to elderly.

25. This leads to some intergenerational issues. Through fiscal redistribution, France achieves a uniform rate of inequalities between working age population and retirees. As market inequality is the stronger among the senior, this implies that fiscal redistribution focuses on the elderly (Figure 10). This is mainly achieved through public pension spending, which have a larger redistributive impact than non-pensions transfers (Figure 8).

Figure 10.
Figure 10.

Impact of Fiscal Redistribution on Income Inequality by Age Group, 2012

Citation: IMF Working Papers 2016, 007; 10.5089/9781513540139.001.A001

Source: OECD.

26. France achieves better distributive outcomes for the elderly poor, than for the rest of its population. Reflecting the stronger focus of redistribution on the elderly, the poverty rate of elderly as well as their at-risk-of-poverty rate after social transfers (an indicator for the most vulnerable households at the lower end of the income distribution) is lower than in other European countries. In contrast, the risk-of-poverty rate for the population under 65 is comparable to the EU median and only slightly below comparators (Figure 11).

Figure 11.
Figure 11.

Poverty in France and in Europe

Citation: IMF Working Papers 2016, 007; 10.5089/9781513540139.001.A001

Sources: Eurostat, OECD, and IMF Staff calculations.1/ Dashed lines represent EU medians. For the sake of clarity, Ireland has been excluded from the chart.2/ Poverty rate of an age group relative to the poverty rate of the entire population = 100.

27. Poverty risk is particularly acute for the young. Their relative poverty rate is increasing and is larger than in other European countries (Figure 11). In contrast with Germany and the United Kingdom (but not Italy), the poverty rate is much higher for children than for the elderly and this gap has been widening during the crisis. France is also poorly positioned in the OECD index on child well-being relative to comparators, particularly for health and safety, and quality of school life (OECD, 2009). Problems of inequalities and poverty are particularly prevalent in poor urban areas, despite significant public spending and support (Box 1).

28. Social spending has not adapted to new forms of poverty. In 2013, the poverty rate of retirees was 7.9 percent compared to 19.6 percent for the population below 18 and 37.3 percent for unemployed 18 year old or older (INSEE, 2015). Low poverty rate among the elderly has been achieved by developing a social benefits system to address old-age poverty from the mid-1950s up to now (the minimum pension scheme increased by 25 percent over 2009-12 and was raised again in 2014). Yet, social spending has not adapted to the rising poverty of children and the young that is largely explained by the increase in parents’ and youth unemployment, notably long-term unemployment and inactivity. The unemployment rate of the 15-24 year old was close to 24 percent in the first half of 2015, more than twice the national average of 10 percent, and their underemployment rate,10 at 12 percent of employment in 2014, was more than twice the level for 25–64 year old (DARES, 2015).

29. As many EU countries, France should step up its efforts to reduce child poverty. The government has recently announced a national strategy in this area (Commission Enfance et Adolescence, 2015). Some EU countries, such as Belgium and the United Kingdom, have implemented specific strategies that integrate various forms of social interventions and benefits to effectively tackle the causes of child poverty. They include improving parents’ participation in the labor market, improving living standards of poor families through better targeted social transfers, and raising educational attainment of poor children by investing in early childhood education. In France, further increasing the means-testing of family-related spending could help reduce child poverty and a reform of pension spending (see below) could provide fiscal room for additional spending in this area.

30. Rather than increasing social benefits for the young working age population, France should aim at providing them with more job opportunities. Income instability is a major source of income inequality and plays a stronger role for the young (Ceci-Renaud et al., 2015) 11 and the unemployed who are 5 times more at risk of poverty than people in work (DARES, 2015). Therefore, a crucial way to reduce the youth’s poverty is to implement labor market reforms that would improve their job situation notably reducing their comparatively higher unemployment rate.

Urban Development Spending and Social Outcomes

About 7 percent (4.4 million people) of France’s population live in 751 poor urban areas (zones urbaines sensibles) that benefit from targeted public supports. While no consolidated data is available for general government, central government spending earmarked on these areas is estimated at about €7 billion (0.3 percent of GDP) in 2014, including transfers to local governments (Draft 2015 budget law, 2014b). It finances additional services aiming at providing more education resources (zones d’éducation prioritaires), improving social outcomes, and public safety. These urban areas have also benefited from a large urban development investment plan (of about 2 percent of GDP for the period 2003–13, and projected 0.2 percent of GDP planned from 2014–20), and EU social and structural funds. Despite these efforts, social outcomes are poor (Observatoire national des zones urbaines sensibles, 2015). The rates of poverty are threefold in poor urban areas compared to average (38.4 percent at 60 percent of the national median disposable income, and 10.1 percent at 40 percent). They are even higher for the young: 51.4 percent at 60 percent of the national median income for 18 years of age and below (19.6 percent on average), and 48.0 for the 18–24 years-old (23.3 percent on average). The unemployment rate culminates at 23.2 percent in 2014, compared to 9.3 percent outside these areas. And the participation rate in the labor market is 10 ppts below average. Education outcomes are also much lower in these poor urban areas, in relation to higher education inequalities in France than in most advanced countries (see paragraph 70).

Layers of Government

31. Expenditure containment has been largely focused on consumption of goods and services at the central government level. France spends 1 percent of GDP less on goods and services than the European average (Figure 12). Containment of the consumption of goods and services has been the priority of successive governments and remains an important part of the saving plan for 2015–17. As a result, it was stabilized in percent of GDP for the general government over the last thirty years. However, this containment strategy has been focused on the central government which has halved its consumption of goods and services over the last 30 years (Figure 12). Going forward, limited additional savings can be expected in this area without reducing the quality of services provided by the central government.

Figure 12.
Figure 12.

Spending on Goods and Services

Citation: IMF Working Papers 2016, 007; 10.5089/9781513540139.001.A001

Sources: Eurostat and IMF Staff calculations.

32. By contrast, local governments’ spending on goods and services more than doubled in the last three decades (Figure 12). This is only partly due to large transfers of responsibilities from the central government as it also reflects insufficient spending constraints at this level of government. From 1980-2013, when controlling for the impact of fiscal decentralization, local governments’ current spending increased threefold.

33. France is gradually stepping up efforts to contain local spending. In addition to the “golden rule” for local governments (under which they can only borrow to finance investment), France has started designing an indicative national expenditure growth target for local governments (objectif d’évolution de la dépense publique locale, ODEDEL) aiming at emulating the success of the health expenditure target (ONDAM described below). In addition, cuts in transfers from the central government (by €12.2 billion or 0.5 percent of GDP over 2014–17) are expected to reduce local governments’ spending (provided that local taxation is not increased). However, in the first year, local government chose to reduce public investment (-10 percent in nominal terms) rather than current spending (+3 percent) magnifying the political cycle of investment in 2014 (see below section on public investment). The increase in current spending was due to both the increase in consumption of goods and services (+1 percent) and to the wage bill (+4 percent).

34. Other EU countries have implemented more ambitious reforms of their fiscal decentralization and local governance frameworks. Spain has improved fiscal coordination and incentives for national governments to consolidate. Fiscal targets have been discussed among central and local governments prior to approval within intergovernmental fiscal bodies, with specific measures on the expenditure side (IMF, 2015b). A new Budget Stability Law approved in 2012 has introduced structural budget balance, expenditure, and debt rules at the regional level, with preventive and corrective mechanisms to penalize deviations from fiscal rules and targets. Italy enforces since 1999 a domestic stability pact that sets expenditure and balance targets, and sanctions including through reduction in total expenditure, and hiring (Chiades and Mengotto, 2013). And in Portugal, central government transfers are now conditional on the achievement of expenditure reduction targets by local governments. These reforms were implemented in countries where local governments have, like in France, considerable fiscal autonomy guaranteed by the Constitution.

35. Reducing duplication of spending and public interventions at different levels of governments remains a challenge. France has more layers of governments (State, Departments, Regions, and municipalities) than other EU countries. While the number of regions will be reduced, through merger, from 22 to 13 in 2016, the municipal levels remains highly fragmented in France (58 municipalities per 100 000 inhabitants), compared to the EU average (18 municipalities) (Malvy and Lambert, 2014). To limit duplication, the “general competency clause”, which allows local governments to spend in all areas in which a local interest can be justified (even if it is an area where the local government has not an explicit competency), was eliminated at regional and departmental level in 2015 but still remains in place at municipal level. Other European countries have recently more drastically rationalized their local governments. Portugal has privatized or dissolved about half of local SOEs, and reduced by nearly one third the number of parishes responsible for public services at the municipal level (IMF, 2015c). Italy has also reformed the organization of its provinces, as a first set before abolishing this administrative unit between municipalities and regions through Constitutional amendment.

36. Despite recent reforms, containing spending at the local level, particularly at the municipal level, remains challenging. Because of the constitutional fiscal autonomy of local governments, the government has relied on indirect measures to contain local government spending (reduction in transfers, rationalization of structures of regional governments, elimination of the general competency clause, and creation of a new indicative target for the evolution of local public expenditure, the ODEDEL) However, the reduction in transfers should not be offset by an increase in local taxation or results in cut in investment spending rather than in current spending. In this context, it could be useful considering (i) curtailing the general competencies of municipalities; (ii) developing further the ODEDEL to incentivize reductions in areas of inefficient spending, especially staffing, and preserving growth-friendly expenditures, such as infrastructure investment; and (iii) tightening the rules on possible variations on local taxes.

The Wage Bill

37. The wage bill has proved difficult to contain across general government (Figure 13). The wage bill is 2.2 ppts of GDP larger in France than on average in the EU, explaining almost 20 percent of the difference in total expenditure-to-GDP ratio (Table 3).12 The wage bill for the central government has declined over the last twenty years, owing to: (i) the decentralization process, (ii) employment reduction with the introduction of employment caps from 2006, and natural attrition targets (replacing only 1 of 2 retiring civil servants) from 2007-12, and (iii) the wage-scale freeze since 2010. Some of the reduction is also explained by the transfer of public employees to autonomous entities (ODAC)13 boosting the ODAC wage bill between 2008 and 2013. Local governments’ wage spending was also dynamic. From 1996 to 2008, local government employment increased by 24.6 percent, and by 64.5 percent in their autonomous entities. This surge is only partly the result of the decentralization process: between 2002 and 2009, local governments created over a quarter million jobs (about 5 percent of public employment) in addition to the job creation related to the decentralization process (Cour des comptes, 2009 and 2012a). The rapid increase in local governments’ wage bill at was also related to loose hiring practices and rapid promotions, particularly at the municipal level.

Figure 13.
Figure 13.

General Government Wages and Employment

Citation: IMF Working Papers 2016, 007; 10.5089/9781513540139.001.A001

Sources: Eurostat, INSEE, WEO database, and IMF Staff calculation.1 / 2014 data for wages, 2012 or latest data for employment. Dashed lines represent the European median.2 / Excludes the “contrats aidés.”

38. The wage bill exceeds EU averages in almost all sectors, with social sectors contributing the most. Public compensation spending is higher in eight sectors relative to the European average and only below in two sectors (public order and safety and economic affairs). Social sectors explain more than half of the overall difference of 2.1 ppts of GDP in the wage bill between France and the EU average: education (+0.5 ppts of GDP), health (+0.4 ppts of GDP), and social protection (+0.3 ppts). Both the general public services and defense sectors also contribute significantly (+0.3 ppts each).

39. France’s follows a model of high public employment and relatively modest public wages. Stable at about 20 percent of the total labor force since the early 2000’s, government employment in France is among the highest in Europe (Figure 13). In contrast, there is no wage premium for the public sector in France on average as evidenced by a lower public-private compensation gap than EU average, and comparators (Figure 14). When controlling for differences in the structure in employment between the private and the public sector, public wages are 3 ½ percent lower than in private sector. However, highly educated employees in the public sector have a negative wage gap of about 10 percent, while the less educated benefit from higher wages (of about 6 percent) than in the private sector (de Castro et al., 2013). For example, experienced teachers receive comparatively lower salaries than the European average (Figure 14), and comparators (except Italy). The combination of relatively high employment and low wages partly reflects political choices (e.g., to reduce unemployment), but also the fact that the wage evolution decided by the state applies to all levels of government while, because of constitutionally-guaranteed fiscal autonomy, local governments can freely decide on their level of employment (within the constraints of the golden rule).

Figure 14.
Figure 14.

Public Sector Wage Level

Citation: IMF Working Papers 2016, 007; 10.5089/9781513540139.001.A001

Sources: Eurostat, OECD, and IMF Staff calculations.

40. The wage-scale freeze, implemented since 2010, has had only limited effect on the wage bill-to-GDP ratio. Since 2010, successive governments have frozen the public sector wage-scale for all levels of government, though wage increases were granted at the lower levels of the wage scale (garantie du pouvoir d’achat, GIPA, and alignment of the gross public minimum wage with the gross private minimum wage), for equity reasons. This approach has proven less effective than anticipated for generating savings, given low levels of inflation and the significant wage drift embedded in the current system.14 When progression is automatic, average wage levels rise as the public sector workforce becomes more experienced, even in the absence of wage and employment increases.15 In past years, France has been characterized by a powerful automatic progression system, which has translated into an early attainment of a high level wage in most occupational careers, and particularly for high-skilled workers. As a result of this mechanism and of recent increase in employment, the wage bill which had declined from 13.1 percent of GDP in 2009 to 12.8 percent in 2011 as since increased and reached 13 percent in 2014. This ratio may further increase in light of the 2015 decision to cancel a large share of the planned reduction in public employment for security reasons and to increase temporary hiring in the public sector to reduce unemployment.

41. Structural measures have a more durable impact in reducing the public wage bill. An analysis of recent consolidation episodes in advanced economies concludes that the reduction of the government wage bill has been larger and more durable when the adjustment included structural measures (Figure 15 and IMF, 2014). Structural reforms include rationalizing the size and structure of government, outsourcing non-core functions, and improving the efficiency of the wage formation and hiring processes. Short term measures such as wage or hiring freezes have generally expired within a few years, and generated less durable reduction in the wage bill.

Figure 15.
Figure 15.

Cumulative Change in the Public Wage Bill Ten Years after the First Year of Measures

(In percent of GDP)

Citation: IMF Working Papers 2016, 007; 10.5089/9781513540139.001.A001

Source: IMF (2014).Note: t indicates the year of introduction of the wage bill measure. Episodes with structural measures are: Austria (1996–97), Belgium (1982), Canada (1991—92), Italy (1993), Portugal (2005–07), the United Kingdom (1994). Those without are: Belgium (1992, 1994), Denmark (1983–84), Germany (1983-84), Germany (1995, 1997, 2000), Ireland (1982), Ireland (1987–88), the Netherlands (1984-86), the Netherlands (2005), Portugal (2000, 2003), Spain (1997).

42. The analysis suggests that, for France, priority could be given to reducing public employment, rather than freezing wages. Given that the public sector has no wage premium in France, the current wage-scale freeze can only help containing the wage bill in the short run, but is not sustainable in the long run. A durable and sustainable containment of the wage bill requires a reduction in public employment, which would also provide room for targeted salary increases to attract qualified staff. This will require thorough reviews of staffing in administration, underpinned by streamlining of processes. Employment practices in labor-intensive public sectors such as health and education should also be assessed. Employment should also be rationalized especially at the municipal level supported by reforms to streamline and remove overlap (intercommunalités). Other EU countries have succeeded in reducing public employment (Table 7), both at the central and local governments’ levels, particularly by setting binding entry-to-exit ratio, but also through voluntary departures schemes and mobility pools. Ensuring that working hours are effectively not lower than the 35 hour work week, particularly the impact of the reduction in employment (Cour des comptes, 2015a). This would require a more stringent legal control of local governments’ employment practices by the central government offices at the local level (préfectures).

Table 7.

Reduction in Public Employment, 2010–14

article image

2010-2013

In Employment positions

In Full time equivalent

2011-2014

Sources: Portugal: Direçao-Geral da Administraçao Publico ; Germany: Statistisches Bundesamt; Italy: Istituto Nazionale di Statistica; United Kingdom: Office for National Statistics.

43. Measures are also needed to contain the wage drift embedded in the current system. The authorities could slow down automatic progression by lengthening the maximum duration for a civil servant in each scale level. This could be anchored on the extension of the retirement age that will be progressively implemented, and would make career progression more gradual to avoid an early attainment of high level wages that are detrimental to productivity.

Public investment

45. France’s infrastructure and facilities are widely available and of excellent quality by international standards (Figure 16 and Table 8). Capital stock is above the European average in terms of quantity, and superior in terms of quality. In addition, France managed to stabilize its capital stock in percent of GDP over the last fifty years, while other advanced countries reduced it sharply. In real per capita terms, France also managed to increase its public capital stock, as in other advanced economies. Finally, France is among the best performers in terms of public investment efficiency and above advanced countries’ median.

Figure 16.
Figure 16.

Public Investment and Capital Stock

Citation: IMF Working Papers 2016, 007; 10.5089/9781513540139.001.A001

Sources: Center for International Comparisons (2013), OECD (2014a), World Economic Forum (2011–12 and 2014–15), IMF (2015d), and IMF Staff calculations.1/ The index ranges between 0 and 1. It provides an estimate of the relationship between the public capital stock and indicators of access to and the quality of infrastructure assets. Countries with the highest levels of infrastructure coverage and quality (output) for given levels of public capital stock and income (inputs) form the basis of an efficiency frontier and are given a score of 1.2/ The box shows the median as well as the 25th and 75th percentile while whiskers show the maximum and minimum values.3/ Public capital stock was constructed using the perpetual inventory method. Quality of roads index is based on a survey from the World Economic Forum’s Global Competitiveness Report (2011–12).4/ Quality of roads index is based on a survey from the World Economic Forum.
Table 8.

Availability of and Access to Infrastructure

article image
Sources: World Bank and IMF Staff calculation.

For MRI Units: EU 27

Note: Telecommunication infrastructure is not reported as they have been mostly privatized.

46. Over the medium term, public investment could be rationalized at the local level and for public corporations. Given its large and high quality public capital stock, France does not have pressing needs for large aggregate additional investment in order to fill a void or to massively upgrade infrastructure. Priority should be given to maintaining the stock of capital and its quality.16 Moreover, given duplication, there is room for rationalizing investment further, notably at the level of local governments which account for about 58 percent of general government investment.17 The elimination in July 2015 of the general competency clause for Régions and Départements will contribute to this rationalization. However, the measure does not cover the municipal level where rationalization is also needed and where anecdotal evidence of duplication and inadequate budgeting of maintenance costs suggest scope for rationalization. Rationalization of investment spending needs also to cover SOEs, which accounted for 1.4 percent of investment (net of government capital transfers) in addition to general government investment in 2013 (Figure 17 and Table 9), in particular for high speed rail transportation (Cour des comptes, 2014b). Finally, to increase public investment efficiency, there is scope to improve and expand, at least to large local governments, the ex-ante cost-benefit evaluation of public investment created by the multi-year budget law 2012-17 which is currently limited to large investments and does not cover any local investment (Draft 2015 budget law, OECD 2014a, and Cour des comptes 2015b).

Figure 17.
Figure 17.

The Role of State Owned Enterprises in Infrastructure Sectors, 20131/

Citation: IMF Working Papers 2016, 007; 10.5089/9781513540139.001.A001

Sources: IMF staff estimates based on OECD Product Market Regulation Database.1/ The indicator measures the state ownership of the largest firm in each sector.
Table 9.

Public Sector Investment, 2013

(In percent of GDP)

article image
Sources: For France: Eurostat, Agence des participations de l’Etat, and Union sociale de l’habitat; For UK: HM Treasury and ONS; and IMF staff calculations.

Net of capital transfers.

B. Public Expenditure by Sector

47. High total public spending is reflected across most functional categories, but especially for social protection and health (Table 10).18 France spends less on public order and safety, and is close to the European average for general public services sector (which includes interest payments). Spending is substantially above the EU average on defense (by 0.6 ppts of GDP), housing (by 0.7 ppts of GDP, i.e. twice the EU average), health (by 1.7 ppts of GDP), and social protection (by 7.3 ppts of GDP). To some extent, this reflects the broader issues identified in the previous section: (i) high social spending that is not always well targeted on poverty reduction; (ii) a wage bill driven by high public employment; and (iii) high spending by many layers of local government. However, to identify specific efficiency gains, it is necessary to look much deeper into individual functional spending categories and economic policies.

Table 10.

General Government Expenditure by Functional Classification 1/

article image
Sources: Eurostat and IMF staff calculations.

When the analysis was undertaken, Eurostat data differed slightly from INSEE in 2014. According to INSEE, expenditures reached 57.5 percent in 2014.

48. Social spending (health and social protection) has been a key driver behind France’s rapid growth in public expenditure over the past decades. From about 25 percent of GDP in 1990, social spending reached 30 percent before the crisis and almost 32 percent in 2014 (accounting for more than half of total public spending). As indicated in Table 10, social spending continued to grow as a share of GDP in the post-crisis fiscal consolidation period, partly reflecting age-related spending pressure. While spending was contained in most sectors from 2010-2013, the social protection sector continued to increase (+0.9 ppts of GDP) and health (+0.2 ppts of GDP), highlighting challenges associated with spending pressures in these sectors (Table 10).

49. France’s social spending is among the highest in Europe and the OECD both for the public sector, and even when accounting for the role of private sector provision (Figure 19). At about 32 percent of GDP, public social expenditures are 6 percent of GDP higher than in Germany, 4 percent of GDP higher than in Sweden and Austria, and 10.5 percent of GDP higher than the OECD average. When private spending is added, to take into account differences in financing system, France ranks second after Denmark.

Figure 18.
Figure 18.

Social Protection Spending by Functions, 2012

(In percent of GDP)

Citation: IMF Working Papers 2016, 007; 10.5089/9781513540139.001.A001

Source: Eurostat.
Figure 19.
Figure 19.

Social Expenditure in the OECD, 2011 1/

Citation: IMF Working Papers 2016, 007; 10.5089/9781513540139.001.A001

Source: OECD.1/ Total public social expenditure reached 31.9 percent of GDP in 2014. The breakdown by categories is only available up to 2011.2/ Pensions are defined as in cash spending for old age and survivors. Private expenditure includes both mandatory and voluntary schemes.

50. France is above EU average across all social risks (Figure 18). Levels of social benefits spending by risks vary across EU countries, based on historical and demographic circumstances, as well as social preferences. France’s spending is above the EU average for each social risk, namely health, disability, old age, survivors, family, unemployment, housing and social exclusion (Haut Conseil de financement de la protection sociale, 2015). Spending is also above all comparators for unemployment and social exclusion, and France is outspent by only one comparator country in all other categories (except survivors for which Germany and Italy spend more).

Pensions

51. At 13 percent of GDP, pension spending is among the highest in the world. This is true even when controlling for income and the involvement of the private sector, and whether (or not) early retirement schemes and disabilities are taken into account (Figures 19 and 20). France has the largest spending on pension in the OECD. When early retirement schemes are added, only Italy and Greece spend more on pension.

Figure 20.