## 16. Robustness Check: Advanced Versus Developing Countries ## Source details **Canonical URL:** [16. Robustness Check: Advanced Versus Developing Countries](https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019188-print-pdf.pdf) ## Other formats - [Markdown version](/-/media/files/publications/wp/2019/wpiea2019188-print-pdf.pdf.md) - [Structured JSON version](/-/media/files/publications/wp/2019/wpiea2019188-print-pdf.pdf.json) --- ### Introduction: research question and key innovation - Focus: effects of public spending composition on income inequality, assessing whether reallocating public spending within a fixed overall envelope can reduce income inequality. - Innovation: analyzes benefits of spending reallocation (composition) rather than only effects of levels of individual spending components. - Sample and data: newly assembled disaggregated public spending data for 83 countries across all income groups; functional classification focused on five components. - High-level finding summary: - Reallocating spending toward social protection and infrastructure is associated with lower income inequality, particularly when financed through cuts in defense spending and in countries with low risks of conflict and strong institutions. - When accounting for long-term impacts, financing higher health and education spending through cuts in defense spending is also associated with lower income inequality. - Reallocating social protection and infrastructure spending toward other types of spending tends to increase income inequality. - Robustness checks confirm results for: (1) alternative measures of income inequality; (2) spending of local governments; (3) efficiency of public spending; (4) impact of debt-financed public spending; and (5) potential differences between advanced and emerging economies. ### Data: composition of public spending and inequality measures - Source of spending data: Acosta-Ormaechea and Morozumi (2017) based on IMF’s Government Finance Statistics Yearbook; bridges GFSM1986 and GFSM2001 to create consistent series 1970–2011. - Spending components used (five): defense; transport and communication (proxy for infrastructure); health; education; social protection. - Coverage: consolidated spending at the central government level. - Inequality measure: net Gini (disposable income, post-tax and post-transfer) from the Standardized World Income Inequality Database (SWIID); alternative checks use market (pre-tax, pre-transfer) Gini and other measures. - Stylized patterns: over 2000–10 advanced and developing economies experienced a decline in the share of infrastructure and defense spending and an increase in the share of health spending; social protection spending flat in advanced economies and increased in developing countries. ### Econometric approach - Baseline specification (log Gini as dependent variable): - Log(Iit) = ∑(δj)Si,j,t-1 + βE i,t-1 + γZit-1 + νi + ηt + εi - Uses one-year lags for explanatory variables; omits one spending component (c) to identify marginal effect of reallocations (effect of moving spending from component c to component j equals δj − δc). - Estimation: fixed-effects panel regression with Driscoll-Kraay standard errors (robust to heteroskedasticity, autocorrelation up to one lag, and cross-sectional dependence). - Control variables Zit-1 include: - Income per capita (log GDP and squared term) - Human capital index (average years of schooling, Barro and Lee 2013, adjusted) - Trade openness (exports + imports over GDP) - Technological progress (share of IT capital in total capital stock) - Unemployment rate - Inflation (CPI change) - Total public spending (percent of GDP) ### Baseline results: magnitudes and direction - Total public spending: - A 1 percentage point increase in the total-spending-to-GDP ratio is associated with a decline of net income inequality of 0.8 to 1 percent (Table 1, columns 1–10). - Spending-component marginal effects (percent changes reported as percent changes in the Gini): - Social protection: a 1 percentage point increase → reduction of disposable income Gini by about 0.3 percent (Table 1, columns 4 and 9). - Infrastructure (transport and communication): a 1 percentage point increase → decline of disposable income Gini by about 0.8 percent (Table 1, columns 3 and 8). - Defense: a 1 percent increase → associated with an increase in income inequality of 0.6 percent (column 2). - Health: a 1 percent increase → associated with an increase in income inequality of 0.5 percent (column 5). - Reallocation effects (financing component = defense): - 1 percentage point increase in social spending financed by a corresponding cut in defense spending → 0.5 percent reduction in inequality. - 1 percentage point increase in infrastructure spending financed by a corresponding cut in defense spending → 0.9 percent reduction in inequality. - Contextual statistic: - The Gini coefficient increased by 5.6 percent per year on average between 1970 and 2010 in the sample; spending reallocation from defense toward social protection and infrastructure could have offset about one-fourth of that increase. ### Role of conflict and institutions - Model extension includes governance/conflict indicators and interaction terms: marginal effect of reallocating spending depends on G(i,t-1). - Six ICRG indicators used to capture political risk: bureaucracy quality; corruption; democratic accountability; external conflict; internal conflict; law and order. (Higher value = lower risk.) - Key result: - Reducing defense spending is associated with lower inequality only when institutions are strong and risk of political instability is low. - Financing higher social and infrastructure spending through cuts in defense spending is associated with lower income inequality only in countries with the lowest risks of conflict and strong institutions. - In countries with weak institutions and higher risks of conflict, no evidence is found that lowering defense spending to finance infrastructure and social outlays is equalizing. - Median-split robustness: splitting sample by median institutional quality confirms that financing higher social and infrastructure spending through cuts in defense spending is associated with lower income inequality only in countries with stronger institutions (Table 7). ### Further robustness checks - Long-run impact (10-year framework): - Replace Log(Iit) with average level of inequality between t and t+9. - Advantages: captures long-term effects (notably education), reduces reverse causality concerns, helps with missing inequality data. - Findings: confirm that increasing social and infrastructure spending through cuts in defense spending is associated with reduced income inequality. Accounting for long-term impact, higher education spending financed through cuts in defense spending is associated with lower income inequality (Table 8). Health shows similar but not always significant patterns. - Sensitivity: results robust to 8-year or 5-year frameworks; 10-year framework better captures long-run education effects. - Alternative inequality indicators: - Alternative measures are highly correlated (Appendix 2). - Using change in disposable income Gini confirms main results (Table 11). - Market income Gini: fiscal policy has limited impact on market-income inequality, except for health spending (Table 12). - Difference between market and disposable Gini (extent of redistribution) confirms main results. - Accounting for local government spending: - Using consolidated general government data (including local governments) confirms main results; education, health, and social spending associated with lower income inequality when financed through cuts in defense spending (Table 13). - Inclusion of local spending helps capture the equalizing effect of health outlays. - Efficiency of public spending: - Uses PIMA overall index (Public Investment Management Assessment) to assess public investment management efficiency. - Equalizing impact of infrastructure spending increases with the efficiency of capital spending (Table 14). - Debt financing: - While public debt slightly reduces the equalizing effect of public spending on infrastructure, social assistance, health, and education, main results remain valid—higher public spending on infrastructure and social assistance is associated with lower income inequality (Table 15). ### Case study: Indonesia — trends and policy implications - Historical trends: - Indonesia grew on average 5 percent between early 1990s and 2017; poverty declined, Gini increased by about 7 points between early 1990s and 2013, then declined by 2 points through 2017. - In 2013, top decile accounted for 28 percent of total consumption (up from <25 percent in 1990). - Wealth concentration: wealthiest 1 percent owned nearly 45 percent of total wealth. - Inequality of opportunities: - Education: gap in attainment modest vs. peers, but poorest quintile has higher share with <4 years schooling; gaps in upper secondary and tertiary enrollment remain. - Health: broad health insurance coverage across income groups, but unequal access to health services across regions (e.g., only 28 percent of villages in Maluku and Papua have health centers vs. national average 38 percent). - Pensions: poorest households essentially no access to pensions; richest households >60 percent access. - Financial services: lower overall access compared to peers; bank account ownership higher in top 60 percent than bottom 40 percent. - Labor markets: nonagricultural informality >70 percent; female labor force participation 51 percent in 2017 vs. 83 percent for men. ### Scenario and assumed spending changes (Indonesia) - Increase in total reallocation considered: 1.5 percentage points of GDP. - Assumed split: health and social programs each will increase by 0.75 percent of GDP. - Education spending as a share of total spending is already close to the statutory 20 percent. ### Estimated distributional impacts (Indonesia partial-equilibrium exercise) - Increase in infrastructure under the comprehensive scenario could lower the Gini coefficient by close to 0.4 points annually. - Taken together, higher infrastructure, health, and social protection spending could reduce the Gini coefficient by 0.5 points annually. - Cumulative impact over 10 years: 5.4 points reduction in the Gini coefficient. - Contextual benchmark: this would reverse the net increase in income inequality (5 points of the Gini coefficient) observed in Indonesia between 2002 and 2017. ### Fiscal financing requirements and effects of financing choices - Critical revenue objective: implementing a medium-term revenue strategy that could raise at least 3 percent of GDP over the medium term to finance equity-enhancing outlays. - Financing channels and likely distributional consequences: - Higher value-added taxes: potentially dampen the estimated equalizing impact. - Higher income taxes or excise taxes on luxury goods: tend to be more progressive and would magnify equalizing effects. - Budget-neutral reallocation: shifting defense spending towards infrastructure and other social spending would strengthen the equalizing impact. ### Institutional and security conditionalities - Reallocating from defense to social protection and infrastructure is associated with lower income inequality primarily in countries with: - The lowest risks of conflict. - Strong institutions. - In countries with higher risk of conflict, there is no evidence that lowering defense spending to finance infrastructure and social outlays improves income distribution. - Reallocating social protection and infrastructure spending toward other types of spending tends to increase income inequality. ### Long-term considerations - Accounting for long-term effects of health and particularly education spending is important to capture the full impact on income inequality. - Higher education spending, and to some extent higher health spending, financed through cuts in defense spending are associated with lower income inequality in the long run. ### Policy recommendations and complementary measures - Where political and security conditions permit, reallocate public spending away from defense and toward social protection, infrastructure, and education to reduce income inequality. - Improve the progressivity and efficiency of tax systems by reducing exemptions and preferential rates to increase revenue and equity. - Improve education sector efficiency and quality by: - Strengthening the link between teacher compensation and performance. - Improving access to education, especially in rural areas. - Tailoring education to labor market needs (Shin 2018). - Enhance targeting and coverage of social benefits; as administrative capacity improves, expand conditional cash transfers to avoid costly universal price subsidy schemes. - Rely on technology solutions to advance financial inclusion. - Improve the overall business environment, simplify business registration, and reduce red tape to reduce informality and support inclusive growth. *Source: "16. Robustness Check: Advanced Versus Developing Countries" (chapter/section) from the provided PDF content.* ### References ............................................................................................................. ### wpiea2019188-print-pdf - References ................................................................................................................................46 ### Figures - 1. Composition of Spending: Functional Classification .........................................................7 - 2. Progressive Taxation, Public Spending, and Income Inequality ......................................24 - 3. Marginal Effect of Defense Spending Across the Level of Institutional Quality (No distinction of financing source) ..........................................................................25 - 4. Marginal Effect of Social Spending Across the Level of Institutional Quality (Financed by cuts in defense spending) .....................................................................26 - 5. Marginal Effect of Infrastructure Spending Across the Level of Institutional Quality (Financed by cuts in defense spending) .....................................................................27 - 6. Selected Asia: Income Inequality, Pre-1990 ....................................................................17 - 7. Selected Asia: Income Inequality, 1990-Latest ................................................................17 - 8. Indonesia: Net Gini Index .................................................................................................18 - 9. Indonesia: Net Gini Index (National data) ........................................................................18 - 10. Selected Asia: Top 10 Income/Consumption Share .........................................................18 - 11. Selected Asia: Wealth Gini Index .....................................................................................18 - 12. Education by Wealth Quintile ..........................................................................................19 - 13. Indonesia: Coverage of Health Insurance and Pension ....................................................19 - 14. Financial Services by Income Share, 2014 .......................................................................20 - 15. Gender Inequality Index, 2017 .........................................................................................20 - 16. Indonesia: Estimated Marginal Impact of Spending Reallocation on GINI Coefficient ..22 - 17. Indonesia: Annual Impact of the Fiscal Structural Reform Package on Inequality .........22 ### Tables - 1. Impact of Total Spending and Spending Components on Income Inequality ..................28 - 2. Impact of Spending Components on Income Inequality (Financed by cut in defense spending) ......................................................................................................29 - 3. Impact of Spending Components on Income Inequality (Financed by cut in education spending) ...................................................................................................30 - 4. Impact of Spending Components on Income Inequality (Financed by cut in health spending) .........................................................................................................31 - 5. Impact of Spending Components on Income Inequality (Financed by cut in social spending) .........................................................................................................32 - 6. Impact of Spending Components on Income Inequality (Financed by cut in infrastructure spending) .............................................................................................33 - 7. Robustness: The Role of Institutions Using Median Value..............................................34 - 8. Robustness: Long-Term Impact of Spending (Ten-year framework) ..............................35 - 9. Robustness: Long-Term Impact of Spending (Eight-year framework) ............................36 - 10. Robustness: Long-Term Impact of Spending (Five-year framework) .............................37 - 11. Robustness: Change in Net Gini as the Dependent Variable ...........................................38 - 12. Robustness Check: Market Gini .......................................................................................39 - 13. Robustness Check: Use of General Government Data .....................................................40 - 14. Robustness Check: Controlling for the Efficiency of Capital Spending ..........................41 - 15. Robustness Check: Controlling for Debt Financing of Public Outlays ............................42 *Source: wpiea2019188-print-pdf - References* ### 16. Robustness Check: Advanced Versus Developing Countries ..........................................43 ### 16. Robustness Check: Advanced Versus Developing Countries ### Introduction: research question and key innovation - Focus: effects of public spending composition on income inequality, assessing whether reallocating public spending within a fixed overall envelope can reduce income inequality. - Innovation: analyzes benefits of spending reallocation (composition) rather than only effects of levels of individual spending components. - Sample and data: newly assembled disaggregated public spending data for 83 countries across all income groups; functional classification focused on five components. - High-level finding summary: - Reallocating spending toward social protection and infrastructure is associated with lower income inequality, particularly when financed through cuts in defense spending and in countries with low risks of conflict and strong institutions. - When accounting for long-term impacts, financing higher health and education spending through cuts in defense spending is also associated with lower income inequality. - Reallocating social protection and infrastructure spending toward other types of spending tends to increase income inequality. - Robustness checks confirm results for: (1) alternative measures of income inequality; (2) spending of local governments; (3) efficiency of public spending; (4) impact of debt-financed public spending; and (5) potential differences between advanced and emerging economies. ### Data: composition of public spending and inequality measures - Source of spending data: Acosta-Ormaechea and Morozumi (2017) based on IMF’s Government Finance Statistics Yearbook; bridges GFSM1986 and GFSM2001 to create consistent series 1970–2011. - Spending components used (five): defense; transport and communication (proxy for infrastructure); health; education; social protection. - Coverage: consolidated spending at the central government level. - Inequality measure: net Gini (disposable income, post-tax and post-transfer) from the Standardized World Income Inequality Database (SWIID); alternative checks use market (pre-tax, pre-transfer) Gini and other measures. - Stylized patterns: over 2000–10 advanced and developing economies experienced a decline in the share of infrastructure and defense spending and an increase in the share of health spending; social protection spending flat in advanced economies and increased in developing countries. ### Econometric approach - Baseline specification (log Gini as dependent variable): - Log(Iit) = ∑(δj)Si,j,t-1 + βE i,t-1 + γZit-1 + νi + ηt + εi - Uses one-year lags for explanatory variables; omits one spending component (c) to identify marginal effect of reallocations (effect of moving spending from component c to component j equals δj − δc). - Estimation: fixed-effects panel regression with Driscoll-Kraay standard errors (robust to heteroskedasticity, autocorrelation up to one lag, and cross-sectional dependence). - Control variables Zit-1 include: - Income per capita (log GDP and squared term) - Human capital index (average years of schooling, Barro and Lee 2013, adjusted) - Trade openness (exports + imports over GDP) - Technological progress (share of IT capital in total capital stock) - Unemployment rate - Inflation (CPI change) - Total public spending (percent of GDP) ### Baseline results: magnitudes and direction - Total public spending: - A 1 percentage point increase in the total-spending-to-GDP ratio is associated with a decline of net income inequality of 0.8 to 1 percent (Table 1, columns 1–10). - Spending-component marginal effects (percent changes reported as percent changes in the Gini): - Social protection: a 1 percentage point increase → reduction of disposable income Gini by about 0.3 percent (Table 1, columns 4 and 9). - Infrastructure (transport and communication): a 1 percentage point increase → decline of disposable income Gini by about 0.8 percent (Table 1, columns 3 and 8). - Defense: a 1 percent increase → associated with an increase in income inequality of 0.6 percent (column 2). - Health: a 1 percent increase → associated with an increase in income inequality of 0.5 percent (column 5). - Reallocation effects (financing component = defense): - 1 percentage point increase in social spending financed by a corresponding cut in defense spending → 0.5 percent reduction in inequality. - 1 percentage point increase in infrastructure spending financed by a corresponding cut in defense spending → 0.9 percent reduction in inequality. - Contextual statistic: - The Gini coefficient increased by 5.6 percent per year on average between 1970 and 2010 in the sample; spending reallocation from defense toward social protection and infrastructure could have offset about one-fourth of that increase. ### Role of conflict and institutions - Model extension includes governance/conflict indicators and interaction terms: marginal effect of reallocating spending depends on G(i,t-1). - Six ICRG indicators used to capture political risk: bureaucracy quality; corruption; democratic accountability; external conflict; internal conflict; law and order. (Higher value = lower risk.) - Key result: - Reducing defense spending is associated with lower inequality only when institutions are strong and risk of political instability is low. - Financing higher social and infrastructure spending through cuts in defense spending is associated with lower income inequality only in countries with the lowest risks of conflict and strong institutions. - In countries with weak institutions and higher risks of conflict, no evidence is found that lowering defense spending to finance infrastructure and social outlays is equalizing. - Median-split robustness: splitting sample by median institutional quality confirms that financing higher social and infrastructure spending through cuts in defense spending is associated with lower income inequality only in countries with stronger institutions (Table 7). ### Further robustness checks - Long-run impact (10-year framework): - Replace Log(Iit) with average level of inequality between t and t+9. - Advantages: captures long-term effects (notably education), reduces reverse causality concerns, helps with missing inequality data. - Findings: confirm that increasing social and infrastructure spending through cuts in defense spending is associated with reduced income inequality. Accounting for long-term impact, higher education spending financed through cuts in defense spending is associated with lower income inequality (Table 8). Health shows similar but not always significant patterns. - Sensitivity: results robust to 8-year or 5-year frameworks; 10-year framework better captures long-run education effects. - Alternative inequality indicators: - Alternative measures are highly correlated (Appendix 2). - Using change in disposable income Gini confirms main results (Table 11). - Market income Gini: fiscal policy has limited impact on market-income inequality, except for health spending (Table 12). - Difference between market and disposable Gini (extent of redistribution) confirms main results. - Accounting for local government spending: - Using consolidated general government data (including local governments) confirms main results; education, health, and social spending associated with lower income inequality when financed through cuts in defense spending (Table 13). - Inclusion of local spending helps capture the equalizing effect of health outlays. - Efficiency of public spending: - Uses PIMA overall index (Public Investment Management Assessment) to assess public investment management efficiency. - Equalizing impact of infrastructure spending increases with the efficiency of capital spending (Table 14). - Debt financing: - While public debt slightly reduces the equalizing effect of public spending on infrastructure, social assistance, health, and education, main results remain valid—higher public spending on infrastructure and social assistance is associated with lower income inequality (Table 15). ### Case study: Indonesia — trends and policy implications - Historical trends: - Indonesia grew on average 5 percent between early 1990s and 2017; poverty declined, Gini increased by about 7 points between early 1990s and 2013, then declined by 2 points through 2017. - In 2013, top decile accounted for 28 percent of total consumption (up from <25 percent in 1990). - Wealth concentration: wealthiest 1 percent owned nearly 45 percent of total wealth. - Inequality of opportunities: - Education: gap in attainment modest vs. peers, but poorest quintile has higher share with <4 years schooling; gaps in upper secondary and tertiary enrollment remain. - Health: broad health insurance coverage across income groups, but unequal access to health services across regions (e.g., only 28 percent of villages in Maluku and Papua have health centers vs. national average 38 percent). - Pensions: poorest households essentially no access to pensions; richest households >60 percent access. - Financial services: lower overall access compared to peers; bank account ownership higher in top 60 percent than bottom 40 percent. - Labor markets: nonagricultural informality >70 percent; female labor force participation 51 percent in 2017 vs. 83 percent for men. - Policy implications and fiscal recommendations for Indonesia: - Authorities have prioritized inclusive growth; commitments include expanding public health insurance coverage to 100 percent by 2019 and continuing equity-enhancing programs: Program Keluarga Harapan (PKH), Beras untuk Rakyat Miskin (RASKIN), Bantuan Siswa Miskin (BSM). - Current spending on PKH, RASKIN, and BSM together represent only 0.3 percent of GDP—room to increase coverage, adequacy, and targeting. - Integrated social assistance database (Pemutakhiran Basis Data Terpadu - PBDT) covering bottom 45 percent is an important step to consolidate programs and improve targeting. - IMF fiscal-structural reform package recommendation (summarized): higher priority public spending in infrastructure and targeted transfers in education, health, and social programs combined with structural reforms could raise potential growth by 1 percentage point to 6.5 percent in the medium term (Shin 2018). - Empirical assessment (scenario cut off in source): under the reform scenario, infrastructure spending would increase by 1.3 percentage points of GDP, while spending on education, health, and social programs would increase by [text ends here]. *Italic: Source: "16. Robustness Check: Advanced Versus Developing Countries" (chapter/section) from the provided PDF content.* ### 1.5 percentage points of GDP. It is assumed that health and social programs each will ### wpiea2019188-print-pdf - 1.5 percentage points of GDP. It is assumed that health and social programs each will ### Scenario and assumed spending changes - Increase in total reallocation considered: 1.5 percentage points of GDP. - Assumed split: health and social programs each will increase by 0.75 percent of GDP. - Education spending as a share of total spending is already close to the statutory 20 percent. ### Estimated distributional impacts (Indonesia partial-equilibrium exercise) - Increase in infrastructure under the comprehensive scenario could lower the Gini coefficient by close to 0.4 points annually. - Taken together, higher infrastructure, health, and social protection spending could reduce the Gini coefficient by 0.5 points annually. - Cumulative impact over 10 years: 5.4 points reduction in the Gini coefficient. - Contextual benchmark: this would reverse the net increase in income inequality (5 points of the Gini coefficient) observed in Indonesia between 2002 and 2017. ### Fiscal financing requirements and effects of financing choices - Critical revenue objective: implementing a medium-term revenue strategy that could raise at least 3 percent of GDP over the medium term to finance equity-enhancing outlays. - Financing channels and likely distributional consequences: - Higher value-added taxes: potentially dampen the estimated equalizing impact. - Higher income taxes or excise taxes on luxury goods: tend to be more progressive and would magnify equalizing effects. - Budget-neutral reallocation: shifting defense spending towards infrastructure and other social spending would strengthen the equalizing impact. ### Institutional and security conditionalities - Reallocating from defense to social protection and infrastructure is associated with lower income inequality primarily in countries with: - The lowest risks of conflict. - Strong institutions. - In countries with higher risk of conflict, there is no evidence that lowering defense spending to finance infrastructure and social outlays improves income distribution. - Reallocating social protection and infrastructure spending toward other types of spending tends to increase income inequality. ### Long-term considerations - Accounting for long-term effects of health and particularly education spending is important to capture the full impact on income inequality. - Higher education spending, and to some extent higher health spending, financed through cuts in defense spending are associated with lower income inequality in the long run. ### Policy recommendations and complementary measures - Where political and security conditions permit, reallocate public spending away from defense and toward social protection, infrastructure, and education to reduce income inequality. - Improve the progressivity and efficiency of tax systems by reducing exemptions and preferential rates to increase revenue and equity. - Improve education sector efficiency and quality by: - Strengthening the link between teacher compensation and performance. - Improving access to education, especially in rural areas. - Tailoring education to labor market needs (Shin 2018). - Enhance targeting and coverage of social benefits; as administrative capacity improves, expand conditional cash transfers to avoid costly universal price subsidy schemes. - Rely on technology solutions to advance financial inclusion. - Improve the overall business environment, simplify business registration, and reduce red tape to reduce informality and support inclusive growth. *Source: IMF staff analysis from the provided content.* --- _Source: https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019188-print-pdf.pdf_