## 1fraea2019003

## Source details

**Canonical URL:** [1fraea2019003](https://www.imf.org/-/media/files/publications/cr/2019/1fraea2019003.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2019/1fraea2019003.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2019/1fraea2019003.pdf.json)

---

### Model framework and key features
- Uses a structural stochastic model with a forward-looking benevolent government choosing the fiscal stance defined as a change in the structural primary balance (discretionary policy).
- Model features:
  - Fiscal multiplier with feedback: the primary balance affects output and output affects the primary balance via automatic stabilizers; the fiscal multiplier is cycle dependent (larger during recessions).
  - Persistent exogenous shocks and hysteresis: recessions reduce potential output through human and physical capital losses.
  - Debt-dependent interest rate and market-access risk: the interest rate is a rising function of debt and, at high debt levels, the government faces a stochastic risk of losing market access.
  - Implementation frictions: fiscal policy decided one year ahead (implementation lag), imperfect real-time knowledge of the cycle, and costs to change fiscal policy (implementation costs, tax uncertainty).

### Theoretical prescriptions and implications
- Optimal fiscal policy implications:
  - Governments should smooth the cycle with counter-cyclical fiscal policy.
  - Governments facing macro shocks should react to rising debt to preserve buffers; a low debt level functions as a buffer to accommodate shocks without excessive interest-rate or market-access costs.
  - Highly indebted governments should react less to shocks because the insurance value of remaining debt buffer is large; when the buffer is small, marginal value of additional buffer is large, inducing surpluses to restore buffers.

### Historical assessment of France’s fiscal stance (1980–2018)
- Public debt trajectory:
  - Public debt rose from 20 percent of GDP in 1980 to almost 100 percent by 2018.
  - Since 1980, public debt increased by almost 80 percent of GDP.
- Drivers: steep increase in general government expenditures not matched by higher revenues.
- Bohn (1998) sustainability test applied to France:
  - Results: higher public debt is associated with lower primary balances in France over the sample; France does not pass Bohn’s sustainability test.
  - Table 1 selected coefficients:
    - Lagged gross public debt: -0.004 (col. 1), 0.020 (col. 2), -0.005 (col. 3), 0.027 (col. 4).
    - Gap coefficient: 0.651*** (in one specification) and 0.346* (in another).
    - Sample periods: 1981–2017 and 1981–2007; Observations: 37 and 27.
  - Conclusion: over 1980–2007 (pre-GFC) the link between debt and structural primary balance was not significant.
- Other empirical findings cited:
  - Mauro et al. (2015) find a significantly negative reaction of primary balance to debt with pre-crisis data after WWII (1950–2007).
  - Boissinot and others (2004) find weak sustainability between 1978 and 2002: revenue increases offset only half of an expenditure shock and debt increases are persistent due to snowball interest effects.
- Countercyclicality:
  - Fiscal policy did not respond to cyclical conditions overall. Correlation between fiscal stance (change in structural primary balance) and current change in output gap is slightly positive, driven by the large 2009 stimulus; excluding 2009 correlation turns negative.
  - No evidence of delayed reaction: fiscal stance vs lagged change in output gap shows similar lack of response.
- Retrospective model comparison:
  - Observed structural primary balance was generally looser than model recommendation.
  - Exception: fiscal tightening starting in 1996 ahead of joining the euro was tighter than recommended.
  - 2009 stimulus was broadly appropriate per model; subsequent consolidation was slower than model recommendation.
- Output-gap measurement bias:
  - Real-time output gap measures for France were persistently negative over the retrospective period, contributing to a deficit bias.
  - Using real-time output gap in the model reduces the measured deficit bias: the deficit bias using a real-time output-gap measure is about one-third of that using an ex-post output gap measure over 2000–15.
  - Even accounting for real-time measures, since 2000 realized fiscal stance is on average looser than recommended.

### Model calibration and 2020–24 policy simulations
- Calibration: calibrated to France following Fournier (2019), taking 2019 as given and computing optimal path over 2020–24; adjustments for ongoing low interest rate environment and staff baseline projections over 2020–24. Beyond 2024, (r-g) is assumed to increase linearly over ten years to reach a long-run historical average.
- Model-recommended fiscal path (central simulation):
  - Recommend a relatively-frontloaded fiscal consolidation in context of high public debt and a closed output gap.
  - Consolidation amounts:
    - About 1 percent of GDP tightening in 2020 relative to 2019.
    - Cumulative tightening of slightly less than 2 percent of GDP by 2024.
  - Debt implications:
    - This consolidation would help put debt on a downward path starting from around 2019 levels.
- Robustness / scenario notes:
  - Alternative scenarios considered:
    - Different elasticity of interest rate to debt: high elasticity scenario: interest rate increases by 2 basis point with one GDP point increase in debt; low elasticity scenario: interest rate increases by 1 basis point with one GDP point increase in debt.
    - Different potential growth: high (low) growth scenario assumes potential growth 0.5% higher (lower).
  - Under alternative calibrations, recommended consolidation path and debt projections vary.
- Caveats:
  - Model is illustrative and does not take into account recently legislated and planned fiscal relaxation measures built into staff’s medium-term baseline fiscal scenario.
  - Results depend on calibration choices and assumptions and should be interpreted with caution.

### Sensitivity analysis — robustness of the recommendation
- Recommendation to improve the structural primary balance holds under a broad range of assumptions.
- Results not very sensitive to:
  - the elasticity of interest rates to the debt level;
  - parameters governing market-access risk;
  - the persistence of output growth shocks;
  - the magnitude of automatic stabilizers;
  - the extent of hysteresis effects.
- Results sensitive to assumptions on:
  - the average interest rate;
  - potential growth;
  - to a lesser extent, fiscal multipliers.
- Directional implications:
  - If growth were permanently higher (lower), the optimal fiscal stance would be easier (tighter).
  - If interest rates were higher, the model would recommend a tighter fiscal consolidation to counter the risk of debt snowball effects.
  - Higher fiscal multipliers imply a somewhat larger recommended consolidation (because consolidation entails larger output costs and hence the government is more debt-adverse).

### Policy response in case of shocks — calibrated stimulus for a severe recession
- Severe recession scenario: a one-year shock triggering a decline in annual GDP by 1 percent.
- Shock dynamics (appendix): one-off exogenous shock of about 4 percent of GDP in 2019; dampened by automatic stabilizers so that growth is 2.3 percent below the baseline; about 40 percent of this shock dissipates each year (autoregressive term in v_t); government reacts in 2020 (implementation delays).
- Discretionary stimulus indicated: around 1/4–1 percent of GDP could be considered to reduce short-term and long-term costs (hysteresis) of the recession.
- Size of the stimulus varies with:
  - the fiscal multiplier of the instruments used (higher multiplier → smaller stimulus required);
  - the hysteresis associated with the downturn;
  - the cost of financing.
- Permanent effects and trade-offs:
  - The adverse shock and the stimulus have a permanent effect on debt; there is a cost in terms of future consolidation as adjustment needs to be sustained longer after the shock dissipates.
  - Fiscal tightening is unambiguously desirable when debt is high and the economy is booming: it reduces debt and avoids overheating.

### Aggregate recommendation and interpretation
- Historical assessment: fiscal policy in France exhibited a deficit bias over the past four decades, being unable to react to either rising debt levels or cyclical conditions.
- Model-based conclusion: fiscal policy was generally looser than warranted by cyclical and debt sustainability considerations (partly due to real-time output gap uncertainty).
- Forward-looking recommendation: given high public debt and a closed output gap, the model recommends a fiscal consolidation of slightly less than 2 percent of GDP in the medium run, relatively frontloaded, to increase the capacity of fiscal policy to offset future shocks.
- Caveat: precise yearly amounts should be interpreted with caution because of model uncertainty; however, the recommendation to restore fiscal buffers sooner rather than later is robust to alternative calibrations.

### Key calibrated parameters and statistics (Table 2: France — Baseline Calibration)
- Welfare function:
  - Discount factor β 0.99
  - Risk aversion σ 2
  - Labor elasticity η 1/0.3
  - Weight of labor ξ 1
  - Adjustment cost χ 3
- Fiscal parameters:
  - Fiscal multiplier when the gap is null m1 0.98
  - Fiscal multiplier sensitivity to shocks m2 3
  - Automatic stabilizers (primary balance semi-elasticity to the gap) a 0.61
- Interest rate and debt parameters:
  - Growth-adjusted interest rate at current debt level 1.04%
  - Effect of debt level on the risk premium α1 1.5%
  - Effect of debt change on the risk premium α2 0.5%
  - Debt level at which the risk to lose market access is 50% d̄ 150%
  - Debt limit accuracy d1 3
  - Effect of debt change on the risk to lose market access d2 1
  - Adverse scenario coefficient in case of loss of market access d3 -1%
- Economy parameters:
  - Potential GDP per capita growth 1.0%
  - Shock persistence ρ 0.59
  - Shock size σ 2.4%
  - Hysteresis 10%
  - Hysteresis threshold -1%

### Empirical analysis of fiscal consolidations and subnational roles (cross-country evidence and implications for France)
- Sample and objective:
  - Aim: assess role of subnational, institutional, and socio-economic factors in successful fiscal consolidations.
  - Sample: 22 OECD countries over 1990–2016 (listed in source).
  - Definition: successful fiscal consolidation episodes are large contractionary fiscal adjustments associated with a substantial debt reduction three years after the beginning of the episode.
- Key regression findings:
  - Supportive fiscal stances at both central and subnational levels are associated with successful fiscal consolidations.
  - Policy alignment/coordination between central and subnational levels increases probability of success.
  - Subnational contribution arises from alignment of their own fiscal stance and own spending adjustments to own resources.
  - Magnitude example: a tighter subnational fiscal stance by one increases the probability of success by 0.8; a decline in transfers from central to subnational governments by one increases the probability of success by 0.8.
- Changes since the GFC:
  - Role of subnational fiscal policy in consolidations has increased since the GFC; effect increased statistically after 2008/2009.
- Institutional and compositional determinants:
  - Greater revenue decentralization increases the role of subnational governments in successful consolidations.
  - Composition matters: larger share of health or education spending executed by subnational units increases subnational impact; capital spending share is not significant.
  - Fiscal rules at both central and subnational levels raise chances of successful consolidation.
  - Greater political autonomy and cohesion strengthen the role of subnational fiscal efforts.
- Distributional context:
  - Role of fiscal policy in successful consolidations increases in more redistributive countries and in countries with lower regional income disparities.
  - Central government fiscal policy impact is larger where fiscal policy addresses market income inequality more.
  - Subnational fiscal policy role is higher where net regional inequality is lower.
- Implications for France:
  - France registered large positive adjustments in 1996, 2011, and 2012, but none were “successful” per study definition.
  - Regression results:
    - Central fiscal policy impact in France similar to other OECD countries.
    - Subnational governments in France have played a significantly lower role relative to other OECD countries, though their role increased since 2009.
    - Fiscal rules, decentralization, political cohesion, degree of redistribution, regional inequality are not found statistically significant for France relative to OECD in this analysis.
- Selected quantitative results for France and OECD:
  - Degree of expenditure decentralization: France = 36 percent; OECD average = 38 percent.
  - Selected coefficients (Table excerpts):
    - Central Fiscal Stance: 1.305***; 1.589 **; 0.041**; 3.020***; 1.090*** (columns I–V).
    - Subnational Fiscal Stance: 6.484***; 5.625*.
    - Subnational Fiscal Stance (own spending/own revenue): 37.691**; 0.817**; 15.755***.
    - Transfers (+: net increase at the central level): 51.815***; 0.813***; 3.337*.
    - GG Debt/GDP (lagged): 0.004; 0.024; 0.001; 0.011*; 0.000.
    - GDP growth (lagged): 0.145*; 0.740***; 0.010***; 0.055; 0.003.
    - Long-term interest rate (change): -0.856***; -2.132***; -0.054***; -0.568**; -0.320**.
    - Observations: 235; 235; 235 [173]; 235.
  - Table 2 (France vs OECD) selected coefficients:
    - Central Fiscal Stance: 6.974 *; 0.989**; 0.819*; 0.109***.
    - Subnational Fiscal Stance (own spending/own revenue): 12.224*; 24.371***; 0.486; 0.687***.
    - Political cohesion (lagged): 2.536; 11.269*; 0.177; 0.787**.
    - Political autonomy (lagged): 10.525; 19.99*; 0.495; 0.941***.
    - Income inequality (lagged): 3.600; 0.375; 0.245; 0.025.

### Policy implications and recommendations (for France)
- Align the fiscal policy stance at both central and subnational levels, including through effective rules, to achieve a sustained decrease in public debt.
- Sustain the recent contractual approach with local governments limiting their spending growth to 1.2 percent; it resulted in a notable reduction in local current spending in 2018 and will need to be sustained to generate lasting benefits for public finances.
- Consider more decentralization of revenue and spending decisions from the central to local governments, provided decentralization is carefully designed to:
  - Foster fiscal accountability and responsibility of local governments.
  - Ensure fiscal objectives at the local level remain in line with overall general government fiscal targets.
- Maintain the redistributive character of fiscal policies to reduce inequality.

*Italic: Source: 1fraea2019003 - References ____________________________________________________________________________ 16 (PDF chapter/section).*

### References ____________________________________________________________________________ 16

### 1fraea2019003 - References ____________________________________________________________________________ 16

### Model framework and key features
- The note uses a structural stochastic model with a forward-looking benevolent government that chooses the fiscal stance defined as a change in the structural primary balance (discretionary policy).
- Model features:
  - Fiscal multiplier with feedback: the primary balance affects output and output affects the primary balance via automatic stabilizers; the fiscal multiplier is cycle dependent (larger during recessions).
  - Persistent exogenous shocks and hysteresis: recessions reduce potential output through human and physical capital losses.
  - Debt-dependent interest rate and market-access risk: the interest rate is a rising function of debt and, at high debt levels, the government faces a stochastic risk of losing market access.
  - Implementation frictions: fiscal policy decided one year ahead (implementation lag), imperfect real-time knowledge of the cycle, and costs to change fiscal policy (implementation costs, tax uncertainty).

### Theoretical prescriptions and implications
- Optimal fiscal policy implications from the model:
  - Governments should smooth the cycle with counter-cyclical fiscal policy.
  - Governments facing macro shocks should react to rising debt to preserve buffers; a low debt level functions as a buffer to accommodate shocks without excessive interest-rate or market-access costs.
  - Highly indebted governments should react less to shocks because the insurance value of remaining debt buffer is large; when the buffer is small, marginal value of additional buffer is large, inducing surpluses to restore buffers.

### Historical assessment of France’s fiscal stance (1980–2018)
- Public debt trajectory:
  - Public debt rose from 20 percent of GDP in 1980 to almost 100 percent by 2018.
  - Since 1980, public debt increased by almost 80 percent of GDP.
- Drivers: steep increase in general government expenditures not matched by higher revenues.
- Bohn (1998) sustainability test applied to France:
  - Results: higher public debt is associated with lower primary balances in France over the sample; France does not pass Bohn’s sustainability test.
  - Table 1 estimates (selected reported coefficients):
    - Lagged gross public debt: -0.004 (col. 1), 0.020 (col. 2), -0.005 (col. 3), 0.027 (col. 4). Standard errors and significance reported in source.
    - Gap coefficient: 0.651*** (in one specification) and 0.346* (in another).
    - Sample periods: 1981–2017 and 1981–2007; Observations: 37 and 27.
  - Conclusion: over 1980–2007 (pre-GFC) the link between debt and structural primary balance was not significant.
- Other empirical findings cited:
  - Mauro et al. (2015) find a significantly negative reaction of primary balance to debt with pre-crisis data after WWII (1950–2007).
  - Boissinot and others (2004) find weak sustainability between 1978 and 2002: revenue increases offset only half of an expenditure shock and debt increases are persistent due to snowball interest effects.
- Countercyclicality:
  - Fiscal policy did not respond to cyclical conditions overall. Correlation between fiscal stance (change in structural primary balance) and current change in output gap is slightly positive, driven by the large 2009 stimulus; excluding 2009 correlation turns negative.
  - No evidence of delayed reaction: fiscal stance vs lagged change in output gap shows similar lack of response.
- Retrospective model comparison:
  - Comparing observed structural primary balance to model-recommended balance shows France’s fiscal stance was generally looser than model recommendation.
  - Exception: fiscal tightening starting in 1996 ahead of joining the euro was tighter than recommended.
  - 2009 stimulus was broadly appropriate per model; subsequent consolidation was slower than model recommendation.
- Output-gap measurement bias:
  - Real-time output gap measures for France were persistently negative over the retrospective period, contributing to a deficit bias.
  - Using real-time output gap in the model reduces the measured deficit bias: the deficit bias using a real-time output-gap measure is about one-third of that using an ex-post output gap measure over 2000–15.
  - Even accounting for real-time measures, since 2000 realized fiscal stance is on average looser than recommended.

### Model calibration and 2020–24 policy simulations
- Calibration: model calibrated to France following Fournier (2019), taking 2019 as given and computing optimal path over 2020–24; adjustments made for ongoing low interest rate environment and staff baseline projections over 2020–24. Beyond 2024, (r-g) is assumed to increase linearly over ten years to reach a long-run historical average (see model calibration table in appendix).
- Model-recommended fiscal path (central simulation):
  - Recommend a relatively-frontloaded fiscal consolidation in context of high public debt and a closed output gap.
  - Consolidation amounts:
    - About 1 percent of GDP tightening in 2020 relative to 2019.
    - Cumulative tightening of slightly less than 2 percent of GDP by 2024.
  - Debt implications:
    - Model simulations show this consolidation would help put debt on a downward path starting from around 2019 levels (figures show debt in percent of GDP trending down under model recommendation).
- Robustness / scenario notes reported in figures:
  - Alternative scenarios considered:
    - Different elasticity of interest rate to debt: in the high elasticity scenario interest rate increases by 2 basis point with one GDP point increase in debt; in the low elasticity scenario interest rate increases by 1 basis point with one GDP point increase in debt.
    - Different potential growth: high (low) growth scenario assumes potential growth 0.5% higher (lower).
  - Under these alternative calibrations, the model-recommended consolidation path and debt projections vary (figure lines reported for each scenario).
- Caveats:
  - Model is illustrative and does not take into account recently legislated and planned fiscal relaxation measures built into staff’s medium-term baseline fiscal scenario.
  - Results depend on calibration choices and assumptions and should be interpreted with caution.

### Policy implication headline
- Given historically insufficient fiscal response to rising debt and measurement biases, and given high debt with a closed output gap in 2019, the model suggests that a relatively-frontloaded fiscal tightening of about 1 percent of GDP in 2020 and nearly 2 percent cumulatively by 2024 is appropriate to restore fiscal buffers and put debt on a downward path, while limiting macroeconomic risks because the output gap is close to equilibrium.

*Italic: Source: 1fraea2019003 - References ____________________________________________________________________________ 16 (PDF chapter/section).*

### 12. These results are fairly robust to

### 12. These results are fairly robust to sensitivity checks

### Sensitivity analysis — robustness of the recommendation to improve the structural primary balance
- The model recommendation to improve the structural primary balance holds under a broad range of assumptions.
- Results are not very sensitive to:
  - the elasticity of interest rates to the debt level (see footnote 6 explanation of offsetting effects);
  - parameters governing market-access risk (because the optimal policy reacts preemptively to contain the interest rate burden, before being too constrained by the debt limit);
  - the persistence of output growth shocks;
  - the magnitude of automatic stabilizers;
  - the extent of hysteresis effects.
- Results are sensitive to assumptions on:
  - the average interest rate;
  - potential growth;
  - to a lesser extent, fiscal multipliers.
- Directional implications:
  - If growth were permanently higher (lower), the optimal fiscal stance would be easier (tighter) as debt dynamics would become more (less) favorable.
  - If interest rates were higher (for instance, faster normalization of monetary policy rates), the model would recommend a tighter fiscal consolidation to counter the risk of debt snowball effects.
  - Higher fiscal multipliers imply a somewhat larger recommended consolidation (because consolidation entails larger output costs and hence the government is more debt-adverse).

### Comparison with Carnot’s rule of thumb
- Carnot’s rule averages a primary gap indicator and a macroeconomic score.
- With the same initial conditions used in the model exercise (high debt, closed output gap, negative structural primary balance), Carnot’s output is similar: both indicate the government should tighten fiscal policy by a comparable amount in 2020.
- Difference: Carnot’s rule does not internalize the record-low interest rate environment and therefore recommends a sharper medium-term consolidation than the model.

### Policy response in case of shocks — calibrated stimulus for a severe recession
- Severe recession scenario simulated: a one-year shock triggering a decline in annual GDP by 1 percent.
- Shock dynamics in the appendix: one-off exogenous shock of about 4 percent of GDP in 2019; dampened by automatic stabilizers so that growth is 2.3 percent below the baseline; about 40 percent of this shock dissipates each year (autoregressive term in v_t); government reacts in 2020 (implementation delays).
- Discretionary stimulus indicated: around ¼–1 percent of GDP could be considered to reduce short-term and long-term costs (hysteresis) of the recession.
- Size of the stimulus varies with:
  - the fiscal multiplier of the instruments used (higher multiplier → smaller stimulus required);
  - the hysteresis associated with the downturn;
  - the cost of financing.
- Permanent effects and trade-offs:
  - The adverse shock and the stimulus have a permanent effect on debt; thus, there is a cost in terms of future consolidation as the adjustment effort needs to be sustained longer after the shock dissipates.
  - Conversely, fiscal tightening is unambiguously desirable when debt is high and the economy is booming: it reduces debt and avoids overheating.

### Aggregate recommendation and interpretation
- Historical assessment: fiscal policy in France exhibited a deficit bias over the past four decades, being unable to react to either rising debt levels or cyclical conditions.
- Model-based conclusion: fiscal policy was generally looser than warranted by cyclical and debt sustainability considerations (partly due to real-time output gap uncertainty).
- Forward-looking recommendation: given high public debt and a closed output gap, the model recommends a fiscal consolidation of slightly less than 2 percent of GDP in the medium run, relatively frontloaded, to increase the capacity of fiscal policy to offset future shocks.
- Caveat: precise yearly amounts should be interpreted with caution because of model uncertainty; however, the recommendation to restore fiscal buffers sooner rather than later is robust to alternative calibrations.

### Key calibrated parameters and statistics (Table 2: France — Baseline Calibration)
- Welfare function
  - Discount factor β 0.99
  - Risk aversion σ 2
  - Labor elasticity η 1/0.3
  - Weight of labor ξ 1
  - Adjustment cost χ 3
- Fiscal parameters
  - Fiscal multiplier when the gap is null m1 0.98
  - Fiscal multiplier sensitivity to shocks m2 3
  - Automatic stabilizers (primary balance semi-elasticity to the gap) a 0.61
- Interest rate and debt parameters
  - Growth-adjusted interest rate at current debt level 1.04%
  - Effect of debt level on the risk premium α1 1.5%
  - Effect of debt change on the risk premium α2 0.5%
  - Debt level at which the risk to lose market access is 50% d̄ 150%
  - Debt limit accuracy d1 3
  - Effect of debt change on the risk to lose market access d2 1
  - Adverse scenario coefficient in case of loss of market access d3 -1%
- Economy parameters
  - Potential GDP per capita growth 1.0%
  - Shock persistence ρ 0.59
  - Shock size σ 2.4%
  - Hysteresis 10%
  - Hysteresis threshold -1%

### Model mechanics — concise points from Appendix I
- Government maximizes household utility by choosing changes in the structural primary balance to stabilize output intertemporally; state variables: government debt d_t, output gap gap_t, structural primary balance pb^st_t.
- Per-period utility: u(c_t, L_t) = c_t^(1−σ)/(1−σ) − ξ y*_t^(1−σ) L_t^(1+η)/(1+η), calibrated so utility peaks at potential output.
- Rising market pressure when debt rises: interest rate increases in public debt and in the change in debt; probability to lose market access P(lost market access) modeled as a logistic function with parameters d1, d2, and d̄.
- Government budget constraint: d_t = [ (1 + r_0 + α_1 d_{t−1} + α_2 Δd_t) / (1 + g_t^*) ] d_{t−1} − pb_t + sf_t (stock-flow adjustment sf_t captures planned one-offs).
- Output: Y_t = A_t L_t; potential output Ȳ_t = A_t L̄; productivity affected by permanent hysteresis when output below threshold.
- Fiscal multiplier specification: ∂gap_t/∂pb^st_t = −m1 (1 − m2 gap_t(pb^st_t, v_t)); multiplier larger in downturns.
- Primary balance: pb_t = pb^st_t + a·gap_t; automatic stabilizers a reduce effects of shocks and changes in primary balance.
- Aggregate resource constraint includes adjustment costs: c_t = y_t (1 − χ (Δpb^st_t)^2).

*Source: IMF staff calculations and model appendix.*

### 1.      The global financial crisis (GFC) has affected the public finances of OECD economies at

### 1.      The global financial crisis (GFC) has affected the public finances of OECD economies at

### Overview of GFC impact
- Central government finances were the most affected due to the loss in revenue and increased expenditure associated with financial and real sector support.
- Subnational governments were also impacted.
- In advanced OECD economies, the general government (GG) debt increased on average by about 30 percentage points of GDP between 2007 and 2017.
- France experienced a somewhat similar trend, although the contribution of subnational governments has been historically lower.

### Fiscal consolidation and outcomes after the GFC
- Fiscal consolidation efforts started in 2009, leading to a decline in central government deficits and, to a lesser extent, subnational government deficits.
- Implementation of consolidation measures was hindered in several cases by fragile economic recovery, increasing inequality, and political fragmentation.
- In France: the central government deficit was already larger than the OECD median before the crisis; the rate of post-crisis improvement slowed after 2013, creating a gap relative to peers in recent years.
- The recovery in subnational government balances in France, while somewhat slower than peers, has caught up with them in recent years.

### Research objective and sample
- Aim: assessing the role of subnational, institutional, and socio-economic factors in successful fiscal consolidations using a cross-country panel regression analysis.
- Sample: 22 OECD countries over the period 1990–2016. The 22 countries are: Austria, Australia, Belgium, Canada, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Japan, the Netherlands, New Zealand, Norway, Portugal, Slovak Republic, Slovenia, Spain, Sweden, the United Kingdom, and The United States.
- Definitions:
  - Successful fiscal consolidation episodes: large contractionary fiscal adjustments associated with a substantial debt reduction. Success assessed by a large reduction in the debt-to-GDP ratio three years after the beginning of fiscal consolidation episodes.
  - During 1990–2016, more than 40 percent of contractionary fiscal adjustment episodes led to a significant debt reduction.

### Empirical approach and variables
- Fiscal policy at different government levels measured by the fiscal policy stance adopted by the central and subnational governments; stance computed as a cyclically-adjusted fiscal impulse.
- Institutional factors: degree of fiscal decentralization (e.g., share of GG spending executed through or revenue collected by subnational units) and existence of fiscal rules at different government levels.
- Policy cohesion: measured by the number of votes that the government won in national (federal) and subnational elections.
- Socio-economic controls: income inequality and unemployment dispersion across regions.
- Estimation: unbalanced panel probit model with standard errors clustered at the country level. Controls include initial level of public debt, GDP growth, inflation, long-term interest rates, and time-fixed effects for global shocks.
- For the analytical analysis, a cyclically-adjusted change in the Social Security Fund balance enters as a control.

### Key regression findings
- Supportive fiscal policy stance at both central and subnational government levels is associated with successful fiscal consolidation episodes.
- Policy alignment/coordination: a contractionary fiscal policy impulse at both levels increases the probability of successful fiscal consolidation.
- Subnational contribution:
  - The supportive role of subnational government is not only a result of lower central transfers but also of alignment of their own fiscal policy stance to the central level.
  - A tighter fiscal policy stance measured by adjustment of subnational government own spending to own resources has a positive and statistically significant impact on success.
- Magnitude example reported in the analysis: a tighter fiscal policy stance at the subnational level by one increases the probability of successful fiscal consolidation by 0.8. Similarly, a decline in transfers from central to subnational governments by one increases the probability of a successful fiscal consolidation by the same magnitude.
- Income inequality across regions and political cohesion across government levels have no direct effects on fiscal consolidation success but indirectly increase the contribution of central and subnational fiscal policy.

### Changes since the GFC
- The role of subnational fiscal policy has increased since the GFC:
  - Impact of subnational fiscal policy on fiscal consolidation has been increasing over time, consistent with increasing fiscal decentralization across OECD countries.
  - The effect of subnational fiscal policy on fiscal consolidation increased statistically after the GFC with the beginning of fiscal stabilization efforts in 2009.

### Institutional and compositional determinants
- Greater revenue decentralization increases the role of subnational governments in successful fiscal consolidation (measured by the ratio of subnational government taxes to general government taxes).
- Successful consolidations are equally supported by higher expenditure decentralization.
- Composition matters: the role of subnational fiscal policy is significantly larger in countries with a larger share of health or education spending executed by subnational units; the size of capital spending does not have statistically significant effects.
- Fiscal rules raise the chances of a successful fiscal consolidation:
  - Consolidations are more successful in countries with fiscal rules at both central and subnational government levels.
  - Fiscal rules provide legal basis and enforcement mechanisms, increasing credibility, and alleviate soft budget constraints and the common pool problem.
- Greater political autonomy and cohesion strengthen the role of subnational fiscal efforts:
  - The effect of a given fiscal effort by subnational governments on the likelihood of success is increased when subnational units have greater autonomy.
  - Greater political cohesion (measured by votes won in subnational elections) also increases the effect of subnational fiscal policy on likelihood of consolidation success.

### Distributional context
- The role of fiscal policy in successful consolidations increases in more redistributive countries and in countries with lower regional income disparities.
- The impact of central government fiscal policy on the probability of success is statistically larger in countries where fiscal policy addresses market income inequality more (i.e., more redistributive).
- The role of subnational government fiscal policy in influencing likelihood of success is higher in countries with lower net regional inequality (measured by dispersion of post-tax and transfer income across regions).
- The role of central government fiscal policy does not depend on the level of gross income inequality.

### Implications for France
- France registered three episodes of large positive adjustment of the general government balance in 1996, 2011, and 2012; none of these adjustments constitute successful fiscal consolidations as defined in this study (they did not lead to significant and durable reductions in the level of general government debt).
- Regression results for France:
  - Similar impact of central fiscal policy in France relative to other OECD countries in the probability of implementing large fiscal consolidations.
  - Subnational governments in France have played a significantly lower role relative to other OECD countries.
  - The role of subnational governments in the probability of contractionary fiscal adjustments in France has significantly increased since 2009.
  - Other factors (fiscal rules, decentralization, political cohesion, degree of redistribution, regional inequality) are not found to be statistically significant in France compared to other OECD countries.

*Prepared by Aleksandra Zdzienicka (FAD).*

### 13.      What does the analysis imply for the design and conduct of fiscal policy in France

### 13.      What does the analysis imply for the design and conduct of fiscal policy in France

### Policy implications and recommendations
- Align the fiscal policy stance at both central and subnational levels, including through the use of effective rules, to achieve a sustained decrease in public debt.
- Sustain the recent contractual approach with local governments limiting their spending growth to 1.2 percent, noting it resulted in a notable reduction in local current spending in 2018 and will need to be sustained to generate lasting benefits for France’s public finances.
- Consider more decentralization of revenue and spending decisions from the central to local governments, provided decentralization is carefully designed to:
  - Foster fiscal accountability and responsibility of local governments.
  - Ensure that fiscal objectives at the local level remain in line with overall general government fiscal targets.
- Maintain the redistributive character of fiscal policies to reduce inequality.

### Key empirical findings (selected quantitative results)
- Degree of expenditure decentralization: France = 36 percent (share of subnational government own spending to total general government spending) compared to OECD average = 38 percent.
- Table 1: Determinants of a Successful Fiscal Consolidation and Contractionary Fiscal Adjustment (selected coefficients and significance levels as reported)
  - Central Fiscal Stance: 1.305***; 1.589 **; 0.041**; 3.020***; 1.090*** (columns I–V respectively; robust standard errors in parentheses as in source).
  - Subnational Fiscal Stance: 6.484***; 5.625* (reported in relevant columns).
  - Subnational Fiscal Stance (own spending/own revenue): 37.691**; 0.817**; 15.755*** (reported values).
  - Transfers (+: net increase at the central level): 51.815***; 0.813***; 3.337* (reported values).
  - GG Debt/GDP (lagged): 0.004; 0 .024; 0.001; 0.011*; 0.000 (as reported across columns).
  - GDP growth (lagged): 0.145*; 0.740***; 0.010***; 0.055; 0.003.
  - Long-term interest rate (change): -0.856***; -2.132***; -0.054***; -0.568**; -0.320**.
  - CPI (lagged): 3.163; 1.665; 0.578**; 4.480; 1.375.
  - Observations: 235; 235; 235 [173]; 235 (as reported).
- Table 2: Determinants of a Contractional Fiscal Adjustment — France vs. OECD countries (selected coefficients)
  - Central Fiscal Stance: 6.974 *; 0.989**; 0.819*; 0.109*** (columns reported).
  - Subnational Fiscal Stance (own spending/own revenue): 12.224*; 24.371***; 0.486; 0.687***.
  - Political cohesion (lagged): 2.536; 11.269*; 0.177; 0.787**.
  - Political autonomy (lagged): 10.525; 19.99*; 0.495; 0.941***.
  - Income inequality (lagged): 3.600; 0.375; 0.245; 0.025.
- Figures summarize roles of subnational fiscal policy, social factors, and differences between France and OECD averages:
  - Subnational fiscal stance and the share of subnational own spending/own revenue significantly affect the probability of successful fiscal consolidation, with effects varying before/after 2008, by degree of decentralization, presence of fiscal rules, and subnational political autonomy/support.
  - Social factors (including the share of income inequality smoothed through redistributive policies) materially influence the success of fiscal consolidations.
  - France shows different coefficients for central and subnational fiscal policy impacts on contractionary fiscal adjustments relative to OECD averages (as depicted in reported coefficient charts).

### Methodological approach (concise)
- Identification of contractionary fiscal adjustment episodes:
  - The cyclically-adjusted general government (GG) fiscal impulse (FI) is computed via regressions and comparisons of GG balance with GDP growth (Equations (1)–(3) in the source).
  - Episodes of large positive fiscal adjustment are identified as FI being more than one standard deviation above the country average (Eq. (4)).
  - Episodes of successful fiscal consolidation require a large positive fiscal adjustment and a large reduction in GG debt within three years; specifically, the GG debt decline between t and t+3 must be below the 25th percentile of the country distribution over the sample period (Eq. (5)).
- Estimation framework:
  - Panel probit estimates with robust standard errors clustered at the country level (Equations (6a), (6b), (7), and (8)).
  - Dependent variable y_it is a dummy for successful fiscal consolidation (based on Eq. (5)) or for large contractionary fiscal adjustment (based on Eq. (4)) when analyzing France specifically.
  - Key regressors include central fiscal stance (CB), subnational fiscal policy stance with and without net transfers (SNB, SNV), net transfers (T), and socio-political controls (θ), with lagged controls for debt, GDP growth, inflation, and long-term interest rates (δ).
- Data and sample:
  - Unbalanced panel of 22 advanced OECD economies during 1990–2016 due to data availability of subnational fiscal variables.
  - Observations and estimation details reported alongside coefficients in tables and figures.

*Italic: Source — IMF staff analysis in "13. What does the analysis imply for the design and conduct of fiscal policy in France" (extracted from the provided chapter).*

### 4.      Data sources

### 4.      Data sources

### Overview of data sources used
- Fiscal variables at general, central, and subnational levels are taken from:
  - IMF Fiscal Decentralization and Global Debt database
  - OECD National Accounts database
  - INSEE
- Macroeconomic and financial control variables come from:
  - IMF IFS and WEO databases
- The subnational (national) factors are included in Table A1 below.

### Table A1. France: Subnational and Central Institutional and Socio-Political Variables — Factors, measures, and sources
- Fiscal spending and revenue decentralization
  - Measure: Spending executed, or revenue collected at the subnational level in percent of total general government spending or revenue
  - Source: IMF Fiscal Decentralization
- General government debt
  - Source: IMF Global Debt
- Structure of spending
  - Measure: Health, capital, and other spending executed at the subnational level in percent of total general government spending
  - Source: IMF Fiscal Decentralization
- Debt or budget deficit rules at the central and subnational level
  - Measure: 0–1 variables indicating the existence of the rule or not
  - Source: IMF Fiscal Rule Dataset
- Income inequality
  - Measure: Market and net GINI, regional income dispersion
  - Sources: World Inequality Database and OECD Regional Statistics
- Other measures of regional inequality
  - Measure: Regional unemployment dispersions
  - Source: OECD Regional Statistics
- Political cohesion at the central and subnational levels
  - Measure: the number of votes that the ruling party won in national (federal) and subnational elections
  - Sources: Global Elections Database, Database of Political Institutions, National Election Database
- Subnational autonomy in decision-making process
  - Measure: Composite indicator of the existence of the legal framework, the possibility of self-organization, independence of the spending and revenue allocation decisions, a degree of the independence from the central government control and the influence on the central government decisions by the subnational governments
  - Source: the European Commission Self-Rule Indicator for Local Government Report (2015)

### Notes on the data compilation
- Fiscal variables span general, central, and subnational government levels and combine IMF, OECD, and national (INSEE) sources.
- Institutional and socio-political factors combine fiscal datasets with political, inequality, and regional statistics datasets to capture multi-dimensional subnational characteristics.
- Specific composite indicators (e.g., subnational autonomy) rely on established external indices (European Commission Self-Rule Indicator).

*Source: International Monetary Fund — 4. Data sources.*

### Box 2. Key Regulatory Restrictions for Start-ups, Sale of Medicines, Retail Distribution, and

### Box 2. Key Regulatory Restrictions for Start-ups, Sale of Medicines, Retail Distribution, and Professional Services

### Administrative Burden on Start-Ups (distance to frontier = 1.5)
- Lack of one-stop shop for issuing all authorizations, permits and licenses that are required to open up a business (0.5).
- Lack of program to review and reduce the number of licenses and permits required by subnational governments (0.5).
- Excessive number of public and private bodies typically need to be contacted to start a Limited Liability Company (0.1) and a Personally-Owned Enterprise (0.075).
- Excessive number of mandatory procedures required to register a Limited Liability Company or allow it to be done via one-stop shop (0.1).

### Sales of medicines (distance to frontier = 4.2)
- Restrictions on the number of pharmacies that can be located in a given geographic area. (0.8)
- Restrictions on where a pharmacy can be located. (0.8)
- Restrictions for non-pharmacists to own a pharmacy (0.8)
- Restrictions to sell non-prescription medicines in non-pharmacies. (0.8)
- Restrictions on advertising of prices of non-prescription medicines. (0.8)

### Retail distribution (distance to frontier = 1.9)
- Requiring registration to establish a new retail outlet (in a specific register, beyond registration in a commercial or trade registry) for selling clothing and food and beverages. (0.4)
- Requiring authorization for establishing a new retail outlet for selling clothing and food and beverages. (0.3)
- Requiring a specific authorization to sell prescription and non-prescription medicines. (0.4)
- Restrictions on advertising of prices of non-prescription medicines. (0.4)

### Professional services

Accountants (distance to frontier = 2.7)
- Restrictions for non-accountants to have up to 100 percent of the voting rights in an accounting firm. (0.4)
- Exclusive rights for accountants on activities, such as drawing up financial statements, insolvency practice, tax advice, and on non-statutory audit. (0.3)
- Requiring membership in a professional organization in order to practice the profession. (0.3)
- Lack of Mutual Recognition Agreements (MRAs) of professionals with other countries. (0.3)
- Lack of clear process for recognizing education titles earned abroad. (0.2)

Lawyers (distance to frontier = 2.6)
- Restrictions for non-lawyers to own up to 100 percent of the capital of a law firm. (0.4)
- Restrictions for non-lawyers to have up to 100 percent of the voting rights in a law firm. (0.4)
- Territorial restrictions on the ability of professionals to practice (0.3)
- Exclusive rights for lawyers on activities, such as representation before administrative agencies, drawing up legal documents, and on advice on matters predominantly regulated by domestic, foreign, or international law (0.3)
- Requiring membership in a professional organization in order to practice the profession. (0.3)
- Requiring a professional examination in order to legally practice the profession. If existing, it should be administered by an independent regulator. (0.2)
- Lack of a clear process for recognizing education titles earned abroad. (0.3)

Architects (distance to frontier = 2.2)
- Restrictions for non-architects to own up to 100 percent of the capital of an architectural firm. (0.3)
- Restrictions for non-architects to have up to 100 percent of the voting rights in an architectural firm. (0.3)
- Exclusive rights for architects on activities, such as design and planning, and on representation for obtaining permits. (0.2)
- Requiring membership in a professional organization in order to practice the profession. (0.3)
- Requiring a professional examination in order to legally practice the profession. If existing, it should be administered by an independent regulator. (0.2)
- Lack of a clear process for recognizing education titles earned abroad. (0.3)

Notaries (distance to frontier = 1.1)
- Restrictions for non-notaries to own up to 100 percent of the capital of a notary firm. (0.4)
- Restrictions for non-notaries to have up to 100 percent of the voting rights in a notary firm. (0.4)
- Exclusive rights for notaries on activities, such as administering oaths and certificating legal documents, and transferring of real estate titles (0.4)
- Restrictions on the number of professionals allowed to practice set by law or self-regulation by professional bodies (0.4)
- Territorial restrictions on the ability of professionals to practice (0.4)
- Minimum prices or fees regulated by the government or self-regulated (0.3)

*International Monetary Fund*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2019/1fraea2019003.pdf_