## Fiscal Decentralization Improves Social Outcomes When Countries Have Good Governance

## Source details

**Canonical URL:** [Fiscal Decentralization Improves Social Outcomes When Countries Have Good Governance](https://www.imf.org/-/media/files/publications/wp/2022/english/wpiea2022111-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2022/english/wpiea2022111-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2022/english/wpiea2022111-print-pdf.pdf.json)

---

### Overview and Research Question
- Research question: How does governance affect social outcomes when many countries provide public services at the local level?
- Context:
  - About half of public health services have been decentralized on average across 75 countries from 1972 to 2019 (IMF’s Fiscal Decentralization Dataset 2020).
  - COVID-19 increased attention to local government roles in healthcare and to governance/corruption issues.
- Focus: Effects of fiscal decentralization and governance on health and education outcomes.

### Hypothesis and Conceptual Framework
- Core hypothesis: The effects of fiscal decentralization on social outcomes are positive only if countries have better governance; social gains from decentralizing public health/education services depend on governance quality.
- Assumption: Governance quality at the local government is postulated to be the same as, or better than that at the central government.
- Triangular relationships (fiscal decentralization ↔ governance → social outcomes):
  - Positive channels: preference matching; yardstick competition; minimizing Leviathan government.
  - Negative channels: common pool problem; flypaper effect; diseconomies of scale; increased rent-seeking in decentralized procurement.
  - Governance can enhance positive effects and mitigate negatives via procurement accountability and prevention of inefficient populist policies.

### Data and Empirical Method
- Dataset:
  - Annual country panel from 1996 to 2018, merging four data sources.
  - Sample countries included: Afghanistan, Albania, Australia, Austria, Azerbaijan, Belarus, Belgium, Bulgaria, China, Croatia, Cyprus, Czech Republic, Denmark, El Salvador, Estonia, Finland, France, Georgia, Germany, Greece, Hungary, Iceland, Indonesia, Iran, Ireland, Israel, Italy, Kazakhstan, Korea, Latvia, Lithuania, Luxembourg, Malta, Mauritius, Moldova, Mongolia, Netherlands, New Zealand, Norway, Poland, Portugal, Romania, Russia, Serbia, Slovak Republic, Slovenia, Spain, Sweden, Switzerland, and Ukraine.
- Governance indicators: government effectiveness; control of corruption; regulatory quality; accountability (Worldwide Governance Indicators 2020).
- Fiscal decentralization measures: health and education spending decentralization (IMF’s Fiscal Decentralization Dataset 2020).
- Outcome and controls:
  - Health outcomes: infant mortality rate per 1,000 live births; life expectancy at birth.
  - Health controls: hospital beds; number of physicians per 1,000 people; health expenditures; domestic private health expenditure; tertiary school enrollment.
  - Education outcome: PISA score (average of science, reading, mathematics for 15-year-olds).
  - Education control: government education expenditure.
- Econometric approach:
  - Panel instrumental variable (IV) Tobit model to address endogeneity and truncated dependent variables.
  - Instrument: lagged variable (current year decentralization highly correlated with previous year).
  - Tobit censoring points:
    - life expectancy: τ = 0;
    - infant mortality rates: τ = 0 and θ = 100;
    - PISA scores: τ = 300 and θ = 600.

### Empirical Results — Health Outcomes
- Main findings:
  - Fiscal decentralization by itself could increase infant mortality rates and shorten life expectancy (statistically significant coefficients reported in Tables 1 and 2).
  - Decentralization of health spending to local governments lowers infant mortality rate and lengthens life expectancy in countries with stronger governance (interaction terms highly statistically significant in Tables 1–2).
  - For subnational (regional) decentralization, infant mortality results are similar in sign and significance (Table 3); for life expectancy both local and subnational coefficients were statistically insignificant in Table 4.
- Control variable associations:
  - Larger number of hospital beds is associated with better health outcomes (Table 1).
  - Higher share of private health expenditure is associated with worse health outcomes.
  - Higher number of physicians per capita associated with longer life expectancy (Table 2).
  - Real GDP per capita positively and statistically significantly associated with longer life expectancy.
- Interpretation:
  - Governance indicators (government effectiveness; control of corruption; regulatory quality; accountability) are statistically significant for health outcomes, indicating governance is particularly important for health.
  - Information asymmetry in health (e.g., physician-induced demand) increases the scope for governance problems relative to education.

### Empirical Results — Education Outcomes
- Main findings:
  - Decentralization of educational expenditure to local governments per se is not associated with better educational outcomes (coefficients statistically insignificant in Table 5).
  - Interaction terms of decentralization with government effectiveness and control of corruption are positive and statistically significant for PISA scores (Table 5), implying decentralization can improve education if governance is strong.
  - Using subnational decentralization, interaction terms remain positive and significant; one specification shows the direct decentralization coefficient positive and statistically significant at the five percent level (Table 6).
- Interpretation:
  - Only some governance indicators matter for education (government effectiveness and control of corruption).
  - Subnational (state/cantonal/provincial) decentralization yields higher PISA scores than local decentralization, consistent with economies of scale and higher administrative capacity at subnational level.

### Simulations and Marginal Effects
- Simulation approach: Use coefficients from first two rows of Tables 1–6 and observed distributions of governance and decentralization to compute marginal impacts.
- Infant mortality marginal effects (Figure 2):
  - When governance varies (decentralization at average), marginal effects of governance on infant mortality are larger for countries with fairly low governance; effect fades at higher governance.
  - When decentralization varies (governance at average), marginal positive effect of decentralization on health outcomes becomes larger as countries decentralize more.
  - Health expenditure decentralization to subnational governments can improve health outcomes more than to local governments (red line steeper than blue).
- Life expectancy marginal effects (Figure 3):
  - Countries with very low governance can increase life expectancy by improving governance.
  - Control of corruption has the steepest marginal impact on life expectancy among governance indicators.
  - Positive effects of fiscal decentralization materialize only when countries decentralize to local governments to a certain degree — "approximately above 40 percent of decentralization distribution."
  - Magnitude of effects on life expectancy is relatively small.
- Education marginal effects (Figure 4):
  - Using government effectiveness and control of corruption:
    - Education decentralization can increase PISA scores at average governance; improving governance can increase PISA scores at average decentralization.
    - No clear concave or convex patterns.
  - Subnational decentralization yields higher PISA scores than local decentralization.

### Key Quantitative Results
- About half of public health services decentralized on average across 75 countries from 1972 to 2019.
- Panel period: 1996 to 2018.
- Simulation: Increasing medical supplies from the sample mean to the 75th percentile:
  - Infant mortality rates can be reduced by 0.4 percentage points.
  - Life expectancy can be lengthened by 1.7 months.
- Decentralization threshold for positive effects on life expectancy: approximately above 40 percent of decentralization distribution.
- Governance index ranges documented for variables such as Government Effectiveness and Control of Corruption: -2.5 to 2.5.

### Policy Implications and Recommendations
- Sequencing matters: If governance quality is low, improve governance before further decentralizing.
  - Priorities: control corruption; improve government effectiveness; strengthen accountability frameworks (e.g., procurement oversight, financial reporting of public hospitals).
- When governance is stronger:
  - Fiscal decentralization to local governments can improve infant mortality rates and life expectancy through better matching of local medical needs.
  - Fiscal decentralization to subnational governments can improve educational outcomes via economies of scale and higher administrative capacity.
- Health-specific recommendations:
  - Increase medical supplies where needed: hospital beds and physicians unambiguously improve infant mortality and life expectancy.
  - Low-income countries seeking to reduce infant mortality should prioritize building more hospital beds.
  - Aging societies need more physicians (positive association between share of population above 65 years old and physicians per thousand people).
- Accountability measures such as municipal audits and reporting of financial statements of public hospitals can improve performance in health and sanitation.

*Source: wpiea2022111-print-pdf - Introduction (IMF Working Paper No. WP/2022/111).*

### Introduction ...........................................................................................................

### wpiea2022111-print-pdf - Introduction

### Document structure
- Introduction (page 3)
- Empirical Method (page 5)
- Results (page 6)
- Conclusion (page 19)
- Annex I. Data Source and Description (page 21)
- References (page 24)

### Major analytical components listed
- Empirical Method
  - Indicates a dedicated methods section beginning on page 5.
- Results
  - Main results section begins on page 6 and is supported by multiple figures and tables.
- Conclusion
  - Concluding material begins on page 19.
- Annex I. Data Source and Description
  - Data sources and descriptions presented starting on page 21.
- References
  - Bibliographic references begin on page 24.

### Figures included
- Figure 1. Triangular Relationship among Fiscal Decentralization, Governance, and Social Outcomes
- Figure 2. Marginal Effects on Infant Mortality Rate
- Figure 3. Marginal Effects on Life Expectancy
- Figure 4. Marginal Effects on PISA Score
- Figure 5. Flowchart of Policy Actions
- Figure 6. Effects of Medical Supplies
- Figure 7. Effects of Hospital Beds on Infant Mortality Rates
- Figure 8. Demographic Aging and Numbers of Physicians

### Tables included
- Table 1. Results for Infant Mortality Rate, Local Governments
- Table 2. Results for Life Expectancy, Local Governments
- Table 3. Results for Infant Mortality Rate, Subnational Governments
- Table 4. Results for Life Expectancy, Subnational Governments
- Table 5. Results for PISA Score, Local Governments
- Table 6. Results for PISA Score, Subnational Governments

### Key metadata and title
- IMF WORKING PAPERS
- Title (as presented on page 3): Fiscal Decentralization Improves Social Outcomes When Countries Have Good Governance
- Publisher: INTERNATIONAL MONETARY FUND

*Source: wpiea2022111-print-pdf - Introduction (canonical URL provided with the content).*

### Introduction

### Introduction

### Overview
- Research question: How does governance affect social outcomes when many countries provide public services at the local level?
- Context: About half of public health services have been decentralized on average across 75 countries from 1972 to 2019 (IMF’s Fiscal Decentralization Dataset 2020). The COVID-19 pandemic increased attention to local government roles in healthcare and to governance/corruption issues.
- Focus: Effects of fiscal decentralization and governance on health and education outcomes.

### Hypothesis and Conceptual Framework
- Core hypothesis: The effects of fiscal decentralization on social outcomes are positive only if countries have better governance; social gains from decentralizing public health/education services depend on governance quality.
- Assumption: Governance quality at the local government is postulated to be the same as, or better than that at the central government (supported by Azfar et al findings).
- Triangular relationships (fiscal decentralization ↔ governance → social outcomes):
  - Potential positive channels of fiscal decentralization: preference matching (Oates 1972), yardstick competition (Besley and Case 1995), minimizing Leviathan government (Brennan and Buchanan 1977; Edwards and Keen 1996).
  - Potential negative channels: common pool problem, flypaper effect, diseconomies of scale, increased rent-seeking in decentralized procurement (Kyriacou and Roca-Sagalés 2021).
  - Governance can enhance positive effects and mitigate negatives (examples: procurement accountability, prevention of inefficient populist policies).

### Data and Empirical Method
- Dataset: Annual country panel from 1996 to 2018, merging four data sources (annex table referenced).
- Governance indicators: government effectiveness; control of corruption; regulatory quality; and accountability (Worldwide Governance Indicators 2020).
- Fiscal decentralization measures: health and education spending decentralization (IMF’s Fiscal Decentralization Dataset 2020).
- Outcome and control variables:
  - Health outcomes (two dependent variables): infant mortality rate per thousand live births; life expectancy at birth.
  - Health controls: hospital beds; number of physicians per 1000 people; health expenditures; domestic private health expenditure; tertiary school enrollment.
  - Education outcome: PISA score (average of science, reading, mathematics for 15-year-olds).
  - Education control: government education expenditure.
- Econometric approach:
  - Panel instrumental variable (IV) Tobit model to address endogeneity and truncated dependent variables.
  - Instrument: lagged variable (current year decentralization highly correlated with previous year).
  - Tobit censoring points used:
    - life expectancy: τ = 0;
    - infant mortality rates: τ = 0 and θ = 100;
    - PISA scores: τ = 300 and θ = 600.
- Sample countries included (as reported): Afghanistan, Albania, Australia, Austria, Azerbaijan, Belarus, Belgium, Bulgaria, China, Croatia, Cyprus, Czech Republic, Denmark, El Salvador, Estonia, Finland, France, Georgia, Germany, Greece, Hungary, Iceland, Indonesia, Iran, Ireland, Israel, Italy, Kazakhstan, Korea, Latvia, Lithuania, Luxembourg, Malta, Mauritius, Moldova, Mongolia, Netherlands, New Zealand, Norway, Poland, Portugal, Romania, Russia, Serbia, Slovak Republic, Slovenia, Spain, Sweden, Switzerland, and Ukraine.

### Empirical Results — Health Outcomes
- Main patterns:
  - Fiscal decentralization by itself could increase infant mortality rates and shorten life expectancy (statistically significant coefficients reported in Tables 1 and 2).
  - Decentralization of health spending to local governments lowers infant mortality rate and lengthens life expectancy in countries with stronger governance (interaction terms highly statistically significant in Tables 1–2).
  - For subnational (regional) decentralization, results for infant mortality are similar in sign and significance (Table 3), but for life expectancy both local and subnational coefficients were statistically insignificant in Table 4.
- Control variable findings:
  - Larger number of hospital beds is associated with better health outcomes (Table 1).
  - Higher share of private health expenditure is associated with worse health outcomes.
  - Higher number of physicians per capita associated with longer life expectancy (Table 2).
  - Real GDP per capita positively and statistically significantly associated with longer life expectancy.
- Interpretation:
  - Governance variables (government effectiveness; control of corruption; regulatory quality; accountability) are statistically significant for health outcomes, indicating governance is particularly important for health.
  - Information asymmetry in health (e.g., physician-induced demand) increases room for governance problems relative to education.

### Empirical Results — Education Outcomes
- Main patterns:
  - Decentralization of educational expenditure to local governments per se is not associated with better educational outcomes (coefficients statistically insignificant in Table 5).
  - Interaction terms of decentralization with governance effectiveness and control of corruption are positive and statistically significant for PISA scores (Table 5), implying decentralization can improve education if governance is strong.
  - Using subnational decentralization, interaction terms remain positive and significant; one specification shows the direct decentralization coefficient positive and statistically significant at the five percent level (Table 6).
- Interpretation:
  - Only some governance indicators matter for education (government effectiveness and control of corruption).
  - Subnational (state/cantonal/provincial) decentralization yields higher PISA scores than local decentralization, consistent with economies of scale and higher administrative capacity at subnational level.

### Simulations and Marginal Effects
- Simulation approach: Use coefficients from first two rows of Tables 1–6 and observed distributions of governance and decentralization to compute marginal impacts.
- Infant mortality marginal effects (Figure 2):
  - When governance varies (decentralization at average), marginal effects of governance on infant mortality are larger for countries with fairly low governance; effect fades at higher governance (green and yellow lines convex).
  - When decentralization varies (governance at average), marginal positive effect of decentralization on health outcomes becomes larger as countries decentralize more (red and blue lines concave).
  - Health expenditure decentralization to subnational governments can improve health outcomes more than to local governments (red line steeper than blue).
- Life expectancy marginal effects (Figure 3):
  - Countries with very low governance can increase life expectancy by improving governance.
  - Among governance indicators, control of corruption has the steepest marginal impact on life expectancy.
  - Positive effects of fiscal decentralization materialize only when countries decentralize to local governments to a certain degree — "approximately above 40 percent of decentralization distribution."
  - Magnitude of effects on life expectancy is relatively small.
- Education marginal effects (Figure 4):
  - Using government effectiveness and control of corruption:
    - Education decentralization can increase PISA scores at average governance; improving governance can increase PISA scores at average decentralization.
    - No clear concave or convex patterns.
  - Subnational decentralization yields higher PISA scores than local decentralization (green/red lines above yellow/blue), likely due to higher administrative capacity and economies of scale.

### Key Quantitative Results Highlighted
- About half of public health services decentralized on average across 75 countries from 1972 to 2019.
- Panel period: 1996 to 2018.
- Simulation: Increasing medical supplies from the sample mean to the 75th percentile:
  - Infant mortality rates can be reduced by 0.4 percentage points.
  - Life expectancy can be lengthened by 1.7 months.
- Decentralization threshold for positive effects on life expectancy: approximately above 40 percent of decentralization distribution.

### Policy Implications and Recommendations
- Sequencing matters: If governance quality is low, countries should improve governance before further decentralizing.
  - Priorities: control corruption; improve government effectiveness; strengthen accountability frameworks (e.g., procurement oversight, financial reporting of public hospitals).
- When governance is stronger:
  - Fiscal decentralization to local governments can improve infant mortality rates and life expectancy through better matching of local medical needs.
  - Fiscal decentralization to subnational governments can improve educational outcomes via economies of scale and higher administrative capacity.
- Health-specific policy implications:
  - Increase medical supplies where needed: hospital beds and physicians unambiguously improve infant mortality and life expectancy.
  - Low-income countries seeking to reduce infant mortality should prioritize building more hospital beds.
  - Aging societies need more physicians (positive association between share of population above 65 years old and physicians per thousand people).
- Accountability measures such as municipal audits and reporting of financial statements of public hospitals can improve performance in health and sanitation.

*Source: wpiea2022111-print-pdf - Introduction.*

### Conclusion

### Conclusion

### Main findings on fiscal decentralization and social outcomes
- Fiscal decentralization by itself does not necessarily improve social outcomes and could even worsen them.
- Countries reap the benefits from decentralization when the quality of their governance arrangements exceeds a certain threshold.
- Sequencing and staging of decentralization matter: countries with different levels of governance quality should consider the timing and extent of decentralizing fiscal operations.
- The relationship between fiscal decentralization and social outcomes is nonlinear.
- Improving government effectiveness and controlling corruption help improve both health and education outcomes.

### Health-specific results and simulations
- The marginal effects of governance on infant mortality are larger for countries with low governance levels.
- For a given level of governance, the favorable effects of fiscal decentralization on health outcomes become larger when countries decentralize health expenditure more.
- Health expenditure decentralization to local governments can improve health outcomes more than to subnational governments, mainly manifested in life expectancy results:
  - Governance does not seem to significantly improve fiscal decentralization’s effect on increasing life expectancy for subnational governments (Table 4).
  - Life expectancy simulation charts show that lines for subnational governments are relatively flat (Figure 3).
- Better health infrastructure is associated with lower infant mortality rates.
- More medical supplies are associated with longer life expectancy.
- Policy-relevant implication: ceteris paribus, building hospital beds or increasing the number of physicians—from the average levels observed in countries in the sample—could improve health outcomes, up to a point.
  - Low-income countries with increasing population may wish to increase hospital beds to reduce infant mortality.
  - Medical supplies may not be as urgently needed for advanced countries with aging populations.

### Education-specific results
- Education expenditure decentralization to subnational governments is found to bring more benefits compared to decentralization to local governments.
  - Education decentralization to subnational governments always yields higher results for PISA scores than for local decentralization (Figure 4).

### Policy recommendations
- Ensure strong institutional arrangements and oversight mechanisms are in place prior to fiscal decentralization.
- Improve government effectiveness and reduce corruption before expanding health/educational expenditure decentralization to reap positive effects.
- Sequence decentralization reforms according to existing governance quality: countries with weaker governance should strengthen institutions before deepening fiscal decentralization.

### Annex: data sources and key variable descriptions (selection)
- Government Effectiveness — Perceptions of the quality of public services, the quality of the civil service and the degree of its independence from political pressures, the quality of policy formulation and implementation, and the credibility of the government's commitment to such policies. The index ranges from -2.5 to 2.5 (worst and best possible outcome, respectively).
- Control of Corruption — Perception of extent to which public power is exercised for private gain, including both petty and grand forms of corruption, as well as “capture” of the state by elites and private interests. The index ranges from -2.5 to 2.5 (worst and best possible outcome, respectively).
- Regulatory Quality — Perceptions of the ability of the government to formulate and implement sound policies and regulations that permit and promote private sector development. The index ranges from -2.5 to 2.5 (worst and best possible outcome, respectively).
- Health spending decentralization — Share of health spending of local governments (or subnational governments) as a proportion of general government spending.
- Education spending decentralization — Share of education spending of local governments (or subnational governments) as a proportion of general government spending.
- PISA score — Average score of 15-year-old students on the PISA scores on science, reading and mathematics.
- Life expectancy at birth, total (years) — The number of years a newborn infant would live if prevailing patterns of mortality at the time of its birth were to stay the same throughout its life.
- Infant mortality rate (per 1,000 live births) — The number of infants dying before reaching one year of age in a given year.
- Hospital beds (per 1,000 people) — Hospital beds include inpatient beds available in public, private, general, and specialized hospitals and rehabilitation centers. Beds for both acute and chronic care are included.
- Physicians (per 1,000 people) — Physicians include generalist and specialist medical practitioners.
- Current health expenditure (% of GDP) — Current health expenditures include healthcare goods and services consumed during each year. This indicator does not include capital health expenditures such as buildings, machinery, IT and stocks of vaccines for emergency or outbreaks.

*IMF Working Paper No. WP/2022/111*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2022/english/wpiea2022111-print-pdf.pdf_