## wpiea2020147-print-pdf

## Source details

**Canonical URL:** [wpiea2020147-print-pdf](https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020147-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2020/english/wpiea2020147-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2020/english/wpiea2020147-print-pdf.pdf.json)

---

### I. Introduction
- Timeframe: Over the past decade (2007-18).
- Key developments in Bulgaria:
  - Absolute poverty has dramatically declined; the severe material deprivation rate has fallen to almost 1/3 of its pre-Global Financial Crisis (GFC) level.
  - The share of the population at persistent risk of poverty has increased, particularly for the elderly and, to a lesser extent, children.
  - Income inequality increased significantly and is now the highest in the EU.
  - Drivers of increased disposable income inequality:
    - Increase in market income inequality.
    - Reduction in fiscal redistribution.
  - Fiscal policy characteristics:
    - Redistributive role of direct taxation became very limited after the introduction of the flat tax in 2008.
    - Social protection expenditure recently declined (as a share of GDP) and is now comparatively low.
    - Social protection expenditure is concentrated on a few social risks and has experienced a decline in redistributive efficiency.
- Definitions and data notes:
  - See Appendix I for definitions of concepts used.
  - EU27 includes the United Kingdom but excludes, due to data availability, Croatia.
  - Children defined as individuals 18 and below.

### II. Inequality and Poverty in Bulgaria and Europe
A. Income Inequality
- Findings:
  - Disposable income inequality in Bulgaria has been persistently higher than the EU27 and NMS averages.
  - Inequality declined in the late 2000s, then increased noticeably in the following decade; since 2016 Bulgaria has the highest measure of inequality in the EU.
  - Increase driven by a rise in market income inequality compounded by a reduction in fiscal redistribution.
- Comparative observations:
  - Market income inequality in Bulgaria was above EU27 average in 2007 and rose more than the EU27 average thereafter.
  - In 2018, market income inequality in Bulgaria is comparable to Greece, Ireland, Italy, Lithuania, Portugal, Romania, and Spain.
  - Unlike most EU countries, Bulgaria reduced fiscal redistribution, so disposable income inequality increased more than market income inequality.
- Policy note:
  - Other policies that can reduce market income inequality include adequate spending on education and health and appropriate labor-market regulations.

B. Poverty
- Absolute poverty:
  - Severe material deprivation fell by almost 2/3 over the past decade.
  - All age groups experienced declines in severe material deprivation, but large regional and age-group differences persist.
  - Example: Despite recent decline, half the population of the Pazardzhik region faced severe material deprivation in 2016.
  - Severe material deprivation is much higher for the elderly than other age groups.
- Relative and persistent poverty:
  - Relative poverty increased during the past decade; the at-persistent-risk-of-poverty rate is higher in Bulgaria than the EU27 average and the gap has grown since 2007.
  - Elderly and children face particularly high at-persistent-risk-of-poverty rates:
    - In 2018, more than one elderly out of five was at persistent risk of poverty compared to about one out of ten on average in the EU27.
    - At 19 percent, the at-persistent-risk-of-poverty rate of children is 6 percentage points above the EU27 average.
  - Child poverty:
    - A larger share of children is at-risk of poverty in Bulgaria than in the EU27 average (about 1 out of 4 children vs. 1 out of 5 on average in the EU27).
    - Risk of poverty for single parent households recently declined and is now comparable to the EU27 average, but at about 50 percent it remains high and much higher than for households with two (or more) adults with children.
  - Policy suggestions to address child poverty:
    - Provide adequate and affordable child care infrastructure to increase single parents’ ability to work.
    - Provide adequate protection against income losses associated with unemployment due to economic shocks, disease, or disability.
    - Note: In Bulgaria, policy focus should be more on availability of childcare than on its cost.

C. Wealth Inequality
- Contrasts with income:
  - Wealth inequality in Bulgaria is low by European standards; wealth levels are relatively low.
- Implications:
  - Given relatively equal and low wealth, capital income likely did not play a major role in high and rising income inequality—though this may change if rapid wealth rise in 2018 and 2019 continues.
  - An adequate social safety net is crucial because most of the population cannot rely on financial wealth to absorb negative income shocks.
- Credit Suisse estimates (noted in text):
  - Mean wealth increased from USD 17,394 in 2017 to USD 42,686 in 2019 and median wealth from USD 11,782 to USD 18,948.

D. Poverty and Inequality in Time of Crisis
- Historical pattern:
  - Poverty increased substantially during the GFC; national severe material deprivation fell from 57 ½ percent in 2005–06 to 41.2 percent in 2007, rebounded to 45.7 percent in 2010, then declined to 20.9 percent in 2018.
- COVID-19 impact expectations:
  - Poverty is likely to increase due to revenue loss from higher unemployment and underemployment.
  - Bulgaria experienced an increase of 2.2 percentage points in the unemployment rate between February and April 2020 (EU average increase was 0.2 percentage point).
  - Revenue loss will differ across population segments (working age vs. retirees, younger vs. older workers, unskilled vs. skilled), likely deepening income inequality.
  - Containment measures constrain capacity to work differently across groups (e.g., single parents when schools are closed).
- Mitigating measures:
  - Fiscal measures can mitigate poverty and inequality impacts.
  - Bulgarian government measures during the pandemic noted in text:
    - BGN 800,000 to provide food to people affected by the crisis.
    - One-off means-tested cash transfer of BGN 375 to parents forced to take unpaid leave to care for children during the state of emergency.
    - Scheme where the state covers 60 percent of the wages and social insurance for a three-month period.
    - Support scheme for freelancers.

E. Regional Inequality
- Regional income disparity increased over time.
- Three phases of regional dispersion in average annual total income per household member (richest vs. poorest region ratios):
  - Pre-GFC (2001–2006): ratio on average 1.7 (annual range 1.5 to 1.8).
  - During the GFC (2007–2012): ratio averaged 2.1 (annual range 2.0 to 2.2).
  - Post-GFC (2013–2016): ratio averaged 2.3 (annual range 2.1 to 2.5).
- Despite differences in income across regions, legislative acts governing social benefits are applied uniformly; municipalities’ role is limited largely to in-kind benefits (including social housing).
- Suggestion: Investigate introduction of regional differentiation across regions or municipalities to better target social benefits given income dispersion.

### III. Fiscal Redistribution
A. Demand for and Supply of Fiscal Redistribution
- Public sentiment and demand:
  - 91 percent of Bulgarians perceive income differences as too high (Eurobarometer, 2018) vs. EU average of 84 percent.
  - Bulgaria has the highest demand for redistribution in the EU since 2006 (except 2012 and 2008).
- Actual fiscal redistribution:
  - Fiscal redistribution is comparatively low.
  - Fiscal redistribution declined from near NMS average in 2007 to the second lowest in the EU.
- Decomposition approach:
  - Uses the Euromod tax-benefit microsimulation model (Appendix I; Sutherland and Figari, 2013).
  - Instruments considered include social contributions, direct taxes, means-tested social spending, and non-means-tested social spending.
- Tax policy impact:
  - Introduction of the flat tax in 2008 significantly reduced the redistributive role of direct taxation.
  - Direct taxation is estimated to reduce the Gini coefficient (measured from 0 to 1) by 0.07 to 0.08 since 2008, about four time less than in 2007 (an impact of 0.029).

B. Fiscal instruments and taxation
- Options to reinforce redistributive role of taxation:
  - Make direct taxation more progressive by introducing an exemption threshold to the flat tax to exclude the poorest or by introducing progressive tax rates for personal income tax.
  - If progressivity increases are not desirable, raise revenue to finance larger social protection spending (for example by increasing the flat tax rate; current flat tax rate is 10 percent).
  - Use social contributions for redistributive purposes (e.g., reduce contributions from lower wages; increase or eliminate cap on contributory income).
- Specific parameters and impacts:
  - Flat tax rate is 10 percent.
  - Cap on contributory income: BGR 3,000 per month or about USD 1,710.
  - Revenue context: social contributions exceeded 7 percent of GDP in 2019; revenue from personal income tax accounted for about 3 ½ percent of GDP.
  - Cap effects: cap on contributory income results in lower effective contributory rate for high-income earners and caps several benefits such as maximum pension.

C. Social transfers, pensions, and their impact on inequality
- Pensions:
  - Pensions reduce income inequality significantly in both EU27 and Bulgaria but less in Bulgaria than in EU27.
- Non-pension benefits:
  - In 2018, about 82 percent of the reduction in income inequality achieved through non-pension benefits in Bulgaria is due to means-tested benefits (EU average: 58 percent).
- Redistributive impact trends:
  - Redistributive impact of social transfers in Bulgaria was the smallest in the EU in 2018 and has declined since 2007.
  - Before social transfers, Bulgaria’s income inequality was lower than or equal to EU27 average up to the mid-2010s, rose since 2009, and surpassed EU27 average in the second half of the 2010s.

D. Size and design of social protection
- Level of social benefits (2018, percent of GDP):
  - Bulgaria: Social benefits 13.0; o/w in cash 10.9; o/w in kind 2.1.
  - Other NMS: Social benefits 14.6; o/w in cash 12.4; o/w in kind 2.2.
  - EU28: Social benefits 20.4; o/w in cash 15.6; o/w in kind 4.8.
- Recent trends and consequences:
  - Social protection spending peaked in 2013 in both Bulgaria and the EU (in part due to GFC response).
  - Decline since 2013 is associated with increased inequality in Bulgaria but not in EU27 or NMS averages.
  - Magnitude of decline in social protection spending (change 2013 to 2018):
    - Bulgaria: 1.5 percent of GDP lower in 2018 than in 2013.
    - EU average: 0.9 percent of GDP lower.
    - Other NMS average: 0.8 percent of GDP lower.
  - In Bulgaria the decline was mostly structural (absence of indexation and reduction in old-age spending) rather than cyclical; unemployment benefits account for only 0.1 percent of GDP in Bulgaria.
- Indexation and discretionary adjustments:
  - With the exception of pensions (and to some extent the heating allowance), benefit levels are adjusted by discretionary decisions rather than systematic statutory indexation.
  - Many key transfers (notably the guaranteed minimum income) were kept unchanged in nominal terms for several years, contributing to falling social protection spending as a share of GDP and to higher relative poverty.

E. Efficiency of social benefits and means-testing design
- Efficiency definition and evolution:
  - Efficiency = reduction in the Gini coefficient achieved by 1 percent of GDP in social benefits (Hallaert and Queyranne, 2016).
  - Bulgaria was above NMS average in 2007 but fell below NMS average by 2018.
  - Bulgaria moved further away from the efficiency frontier in 2018 compared to 2007.
- Means-tested benefits performance and problems:
  - Means-tested benefits reach a small proportion of poor households, have high non-take-up rates, and large leakage (recipients who are neither poor nor entitled).
  - Eligibility criteria are complex, considering income and several other factors (property ownership, ability of relatives including in-laws to provide support).
  - Child allowance specifics (Tasseva, 2016):
    - Only 1/4 of recipients are households in the two lowest income deciles.
    - Child allowance does not reach about 30 percent of poor households with children.
    - About 19 percent of recipients estimated to be non-eligible.
  - World Bank (2009) alternative estimates: 60.6 percent of the poor do not receive a child compensation and 69.9 percent of the individuals receiving a child compensation are not poor.
- Potential gains from reforms to means-testing:
  - Simplifying procedures and screening criteria could reduce non-take-up and leakage and lower administrative costs.
  - Adjusting income-test levels combined with improved allocation could increase efficiency and reduce child poverty.
  - Package of reforms could have limited fiscal cost: increased costs from reduced non-take-up could be offset by reduced leakage, adjusted income-test levels, and lower administrative costs.

### IV. Pension system, unemployment benefits, and social protection concentration
- Pension system: financial sustainability, access, and adequacy
  - Bulgaria prioritized financial sustainability of pensions through a deindexation of pensions in the early 2010s and a pension reform implemented starting in 2016.
  - Consequences:
    - Access to pension has been reduced. The pension reform increased the retirement age and the required retirement contribution. The number of pensioners has been declining since the introduction of the pension reform despite the ageing of the population.
    - Pensions are low compared to wages. Pensions are the only social benefit subject to automatic indexation. The indexation (currently based half on past inflation and half on projections of contributory income) was suspended for several years at the beginning of the 2010s and did not prevent average pension growth from lagging wage growth.
    - Average pension dynamics: in the past decade, average pension declined from being close to the minimum wage to about 70 percent of minimum wage.
    - Minimum and maximum pension rules: minimum pension received by 36 percent of pensioners in 2018. The minimum pension was equivalent to 57 percent of minimum wage in 2010 but only 38 percent in 2018 and is persistently below the official poverty line.
  - Behavioral effects:
    - Many pensioners continue to work to supplement their pension (but are not eligible to unemployment benefits when they lose their job). This reduces the impact of low pensions on old-age poverty and income inequality but did not prevent the real disposable income of the elderly from growing less than the income of other age groups.

- Concentration of social protection spending and coverage gaps
  - “Old age” and “family and children” account for 92 percent of total social protection spending.
  - NMS average: 75 percent (old age and family and children).
  - EU27 average: 70 percent (old age and family and children).
  - Implication: narrow concentration implies limited protection against other social risks; individuals experiencing shocks could easily fall into poverty given relatively low wealth levels.

- Unemployment benefits: eligibility and duration
  - Eligibility criteria are stringent: registered with employment agency, not entitled to old-age pension, and contributed for at least 12 months during the last 18 months.
  - Only 32 percent of the registered unemployed received unemployment benefits (2007–19 average).
  - Unemployment benefit equals 60 percent of the average contributory income over the last 24 months (subject to a minimum and a maximum daily rate).
  - Duration: 4 months to 12 months depending on contributory history (4 months if contributory history is 3 years or less; 12 months if contributory history is over 15 years).
  - Consequence for youth: insurance-based unemployment system whose benefits depend on employment history provides little protection to younger workers and contributed to the rise of youth poverty; youth poverty increased more than for other working age workers and is now above the national average.

### V. Policy implications and recommendations
- To reduce market income inequality:
  - Adequate spending on education and health and appropriate labor-market regulations.
- To increase fiscal redistribution:
  - Increase redistributive role of taxation:
    - Introduce exemption threshold to the flat tax or progressive tax rates for personal income tax.
    - Alternatively, increase the flat tax rate (currently 10 percent) to finance larger social protection spending.
    - Use social contributions for redistributive purposes (reduce contributions on lower wages; increase or remove cap on contributory income—current cap BGR 3,000 per month or about USD 1,710).
- To improve social protection:
  - Reinforce means-testing efficiency: simplify procedures, tighten screening, reduce leakage and non-take-up, and lower administrative costs.
  - Adjust income-test levels for benefits (e.g., child allowance) to better target the poor.
  - Reintroduce or strengthen systematic indexation mechanisms for key social transfers to prevent erosion of benefit value relative to wages and GDP.
  - Reverse the recent decline in social protection spending; aligning Bulgaria with NMS average would increase spending by 1 percent of GDP.
- Fiscal trade-offs and revenue options:
  - A package improving take-up and targeting could have limited net fiscal cost as reduced leakage and lower administrative costs offset higher take-up.
  - Options to improve revenue mobilization include increasing property tax revenue (property tax revenue is 1.7 percent of GDP lower than EU average) and reducing the sizable VAT compliance gap.

### VI. Key statistics and figures (preserved exactly)
- Severe material deprivation fell by almost 2/3 over the past decade.
- Bulgaria experienced an increase of 2.2 percentage points in the unemployment rate between February and April 2020 (EU average increase was 0.2 percentage point).
- BGN 800,000 to provide food to people affected by the crisis.
- One-off means-tested cash transfer of BGN 375 to parents forced to take unpaid leave.
- Scheme where the state covers 60 percent of the wages and social insurance for a three-month period.
- Mean wealth increased from USD 17,394 in 2017 to USD 42,686 in 2019; median wealth from USD 11,782 to USD 18,948.
- Direct taxation is estimated to reduce the Gini coefficient by 0.07 to 0.08 since 2008, about four time less than in 2007 (an impact of 0.029).
- Flat tax rate is 10 percent.
- Cap on contributory income: BGR 3,000 per month or about USD 1,710.
- Social contributions exceeded 7 percent of GDP in 2019.
- Revenue from personal income tax accounted for about 3 ½ percent of GDP.
- Social benefits (2018, percent of GDP): Bulgaria: Social benefits 13.0; o/w in cash 10.9; o/w in kind 2.1. Other NMS: Social benefits 14.6; o/w in cash 12.4; o/w in kind 2.2. EU28: Social benefits 20.4; o/w in cash 15.6; o/w in kind 4.8.
- Decline in social protection spending (change 2013 to 2018): Bulgaria: 1.5 percent of GDP lower in 2018 than in 2013. EU average: 0.9 percent of GDP lower. Other NMS average: 0.8 percent of GDP lower.
- Unemployment benefits account for only 0.1 percent of GDP in Bulgaria.
- Only 32 percent of the registered unemployed received unemployment benefits (2007–19 average).
- Unemployment benefit equal to 60 percent of the average contributory income over the last 24 months.
- Unemployment benefit duration: 4 months to 12 months depending on contributory history.
- Severe material deprivation share: still 3.6 times higher than the EU average.
- Severe material deprivation for the elderly: 32.7 percent.
- Minimum pension received by 36 percent of pensioners in 2018.
- Minimum pension equivalent to 57 percent of minimum wage in 2010 and 38 percent in 2018.
- Average pension declined to about 70 percent of minimum wage in the past decade.
- “Old age” and “family and children” account for 92 percent of total social protection spending.
- NMS average: 75 percent (old age and family and children).
- EU27 average: 70 percent (old age and family and children).
- Reversing the decline in social protection spending would increase spending by 1 percent of GDP.
- Property tax revenue is 1.7 percent of GDP lower than EU average.
- 8.9 percent of GDP in 2018 contributing to about 79 percent of the decline in social protection spending over the period.

*Excerpt from wpiea2020147-print-pdf (IMF staff analysis and figures as provided in the source content).*

### REFERENCES ___________________________________________________________26

### REFERENCES ___________________________________________________________26

### I. Introduction
- Timeframe: Over the past decade (2007-18).
- Key developments in Bulgaria:
  - Absolute poverty has dramatically declined; the severe material deprivation rate has fallen to almost 1/3 of its pre-Global Financial Crisis (GFC) level.
  - The share of the population at persistent risk of poverty has increased, particularly for the elderly and, to a lesser extent, children.
  - Income inequality increased significantly and is now the highest in the EU.
  - Drivers of increased disposable income inequality:
    - Increase in market income inequality.
    - Reduction in fiscal redistribution (Bulgaria did not offset market income inequality with increased redistribution).
  - Fiscal policy characteristics:
    - Redistributive role of direct taxation became very limited after the introduction of the flat tax in 2008.
    - Social protection expenditure recently declined (as a share of GDP) and is now comparatively low.
    - Social protection expenditure is concentrated on a few social risks and has experienced a decline in redistributive efficiency.
- Definitions and data notes:
  - See Appendix I for definitions of concepts used.
  - EU27 includes the United Kingdom but excludes, due to data availability, Croatia.
  - Children defined as individuals 18 and below.

### II. Inequality and Poverty in Bulgaria and Europe
A. Income Inequality
- Overall findings:
  - Disposable income inequality in Bulgaria has been persistently higher than the EU27 and NMS averages.
  - Inequality declined in the late 2000s, then increased noticeably in the following decade; since 2016 Bulgaria has the highest measure of inequality in the EU.
  - Increase driven by a rise in market income inequality compounded by a reduction in fiscal redistribution.
- Comparative observations:
  - Market income inequality in Bulgaria was above EU27 average in 2007 and rose more than the EU27 average thereafter.
  - In 2018, market income inequality in Bulgaria is comparable to Greece, Ireland, Italy, Lithuania, Portugal, Romania, and Spain.
  - Unlike most EU countries, Bulgaria reduced fiscal redistribution, so disposable income inequality increased more than market income inequality.
- Policy note:
  - Other policies that can reduce market income inequality include adequate spending on education and health and appropriate labor-market regulations.

B. Poverty
- Absolute poverty:
  - Severe material deprivation fell by almost 2/3 over the past decade.
  - All age groups experienced declines in severe material deprivation, but large regional and age-group differences persist.
  - Example: Despite recent decline, half the population of the Pazardzhik region faced severe material deprivation in 2016.
  - Severe material deprivation is much higher for the elderly than other age groups.
- Relative and persistent poverty:
  - Relative poverty increased during the past decade; the at-persistent-risk-of-poverty rate is higher in Bulgaria than the EU27 average and the gap has grown since 2007.
  - Elderly and children face particularly high at-persistent-risk-of-poverty rates.
    - In 2018, more than one elderly out of five was at persistent risk of poverty compared to about one out of ten on average in the EU27.
    - At 19 percent, the at-persistent-risk-of-poverty rate of children is 6 percentage points above the EU27 average.
  - Child poverty:
    - A larger share of children is at-risk of poverty in Bulgaria than in the EU27 average (about 1 out of 4 children vs. 1 out of 5 on average in the EU27).
    - Risk of poverty for single parent households recently declined and is now comparable to the EU27 average, but at about 50 percent it remains high and much higher than for households with two (or more) adults with children.
  - Policy suggestions to address child poverty:
    - Provide adequate and affordable child care infrastructure to increase single parents’ ability to work.
    - Provide adequate protection against income losses associated with unemployment due to economic shocks, disease, or disability.
  - Note: In Bulgaria, policy focus should be more on availability of childcare than on its cost.

C. Wealth Inequality
- Contrasts with income:
  - Wealth inequality in Bulgaria is low by European standards; wealth levels are relatively low.
- Implications:
  - Given relatively equal and low wealth, capital income likely did not play a major role in high and rising income inequality—though this may change if rapid wealth rise in 2018 and 2019 continues.
  - An adequate social safety net is crucial because most of the population cannot rely on financial wealth to absorb negative income shocks.
- Credit Suisse estimates (noted in text):
  - After relative stability since 2010, mean wealth increased from USD 17,394 in 2017 to USD 42,686 in 2019 and median wealth from USD 11,782 to USD 18,948 (Credit Suisse (2019)).

D. Poverty and Inequality in Time of Crisis
- Historical pattern:
  - Poverty increased substantially during the GFC; national severe material deprivation fell from 57 ½ percent in 2005–06 to 41.2 percent in 2007, rebounded to 45.7 percent in 2010, then declined to 20.9 percent in 2018.
- COVID-19 impact expectations:
  - Poverty is likely to increase due to revenue loss from higher unemployment and underemployment.
  - Bulgaria experienced an increase of 2.2 percentage points in the unemployment rate between February and April 2020 (EU average increase was 0.2 percentage point).
  - Revenue loss will differ across population segments (working age vs. retirees, younger vs. older workers, unskilled vs. skilled), likely deepening income inequality.
  - Containment measures constrain capacity to work differently across groups (e.g., single parents when schools are closed).
- Mitigating measures:
  - Fiscal measures can mitigate poverty and inequality impacts.
  - Bulgarian government measures during the pandemic noted in text:
    - BGN 800,000 to provide food to people affected by the crisis.
    - One-off means-tested cash transfer of BGN 375 to parents forced to take unpaid leave to care for children during the state of emergency.
    - Scheme where the state covers 60 percent of the wages and social insurance for a three-month period.
    - Support scheme for freelancers.

E. Regional Inequality
- Regional income disparity increased over time.
- Three phases of regional dispersion in average annual total income per household member (richest vs. poorest region ratios):
  - Pre-GFC (2001–2006): ratio on average 1.7 (annual range 1.5 to 1.8).
  - During the GFC (2007–2012): ratio averaged 2.1 (annual range 2.0 to 2.2).
  - Post-GFC (2013–2016): ratio averaged 2.3 (annual range 2.1 to 2.5).
- Despite differences in income across regions, legislative acts governing social benefits are applied uniformly; municipalities’ role is limited largely to in-kind benefits (including social housing).
- Suggestion: There may be merit in investigating introduction of regional differentiation across regions or municipalities to better target social benefits given income dispersion.

### III. Fiscal Redistribution
A. Demand for and Supply of Fiscal Redistribution
- Public sentiment and demand:
  - High demand for redistribution: 91 percent of Bulgarians perceive income differences as too high (Eurobarometer, 2018) vs. EU average of 84 percent.
  - Bulgaria has the highest demand for redistribution in the EU since 2006 (except in 2012 and 2008 as noted).
- Actual fiscal redistribution:
  - Fiscal redistribution is comparatively low.
  - Fiscal redistribution declined from near NMS average in 2007 to the second lowest in the EU.
- Decomposition approach:
  - The decomposition of fiscal redistribution by instrument uses the Euromod tax-benefit microsimulation model (Appendix I; Sutherland and Figari, 2013).
  - Instruments considered include social contributions, direct taxes, means-tested social spending, and non-means-tested social spending.
- Tax policy impact:
  - Introduction of the flat tax in 2008 significantly reduced the redistributive role of direct taxation.
  - Direct taxation is estimated to reduce the Gini coefficient (measured from 0 to 1) by 0.07 to [text truncates here in source].

*Source: wpiea2020147-print-pdf - REFERENCES ___________________________________________________________26*

### 0.08 since 2008, about four time less than in 2007 (an impact of 0.029; Figure 11). This

### wpiea2020147-print-pdf - 0.08 since 2008, about four time less than in 2007 (an impact of 0.029; Figure 11). This

### Inequality trends and fiscal redistribution
- Gini coefficient developments:
  - Bulgaria: increase from 25 in 1990 to 37 in 1995 (Cornia et al., 2004).
  - In 1989, Gini before payroll tax was 24.5 vs. 25.6 after payroll tax (Milanovic, 1994).
- Demand for redistribution:
  - Empirical literature supports that inequality triggers more demand for redistribution but demand does not automatically produce actual redistribution (Olivera, 2015; Schmidt-Catran, 2016).
- Recent change in redistributive impact:
  - Redistributive impact of social transfers in Bulgaria was the smallest in the EU in 2018 and has declined since 2007 (Figure 11).
  - Before social transfers, Bulgaria’s income inequality was lower than or equal to EU27 average up to the mid-2010s, rose since 2009 (shortly after flat tax introduction), and surpassed EU27 average in the second half of the 2010s (Figure 12).

### Fiscal instruments and taxation
- Options to reinforce redistributive role of taxation:
  - Make direct taxation more progressive by introducing an exemption threshold to the flat tax to exclude the poorest or by introducing progressive tax rates for personal income tax.
  - If progressivity increases are not desirable, raise revenue to finance larger social protection spending (for example by increasing the flat tax rate; current flat tax rate is 10 percent).
  - Use social contributions for redistributive purposes (e.g., reduce contributions from lower wages; increase or eliminate cap on contributory income).
- Specific parameters and impacts:
  - Flat tax rate is 10 percent.
  - Cap on contributory income: BGR 3,000 per month or about USD 1,710.
  - Revenue context: social contributions exceeded 7 percent of GDP in 2019; revenue from personal income tax accounted for about 3 ½ percent of GDP.
  - Cap effects: cap on contributory income results in lower effective contributory rate for high-income earners and caps several benefits such as maximum pension.

### Social transfers, pensions, and their impact on inequality
- Pensions:
  - Pensions reduce income inequality significantly in both EU27 and Bulgaria but less in Bulgaria than in EU27.
- Non-pension benefits:
  - In 2018, about 82 percent of the reduction in income inequality achieved through non-pension benefits in Bulgaria is due to means-tested benefits (EU average: 58 percent).
- Gini coefficient dynamics (Eurostat series, scales 0 to 100; Figure 12):
  - Time series points shown: 2007, 2009, 2011, 2013, 2015, 2017, 2019 for EU27 and Bulgaria (specific plotted values in source).

### Size and design of social protection
- Level of social benefits (2018, percent of GDP; Table 2):
  - Bulgaria: Social benefits 13.0; o/w in cash 10.9; o/w in kind 2.1.
  - Other NMS: Social benefits 14.6; o/w in cash 12.4; o/w in kind 2.2.
  - EU28: Social benefits 20.4; o/w in cash 15.6; o/w in kind 4.8.
- Recent trends and their consequences:
  - Social protection spending peaked in 2013 in both Bulgaria and the EU (in part due to GFC response).
  - Decline since 2013 is associated with increased inequality in Bulgaria but not in EU27 or NMS averages.
  - Magnitude of decline in social protection spending (change 2013 to 2018):
    - Bulgaria: 1.5 percent of GDP lower in 2018 than in 2013.
    - EU average: 0.9 percent of GDP lower.
    - Other NMS average: 0.8 percent of GDP lower.
  - In Bulgaria the decline was mostly structural (absence of indexation and reduction in old-age spending) rather than cyclical; unemployment benefits account for only 0.1 percent of GDP in Bulgaria.
- Indexation and discretionary adjustments:
  - With the exception of pensions (and to some extent the heating allowance), benefit levels are adjusted by discretionary decisions rather than systematic statutory indexation.
  - Many key transfers (notably the guaranteed minimum income) were kept unchanged in nominal terms for several years, contributing to falling social protection spending as a share of GDP and to higher relative poverty.

### Efficiency of social benefits and means-testing design
- Efficiency measure and evolution:
  - Efficiency defined as reduction in the Gini coefficient achieved by 1 percent of GDP in social benefits (Hallaert and Queyranne, 2016).
  - Bulgaria was above NMS average in 2007 but fell below NMS average by 2018 (Figures 13–14).
  - Bulgaria moved further away from the efficiency frontier in 2018 compared to 2007.
- Means-tested benefits performance and problems:
  - Means-tested benefits reach a small proportion of poor households, have high non-take-up rates, and large leakage (recipients who are neither poor nor entitled).
  - Eligibility criteria are complex, considering income and several other factors (property ownership, ability of relatives including in-laws to provide support).
  - Child allowance specifics (Tasseva, 2016):
    - Only 1/4 of recipients are households in the two lowest income deciles.
    - Child allowance does not reach about 30 percent of poor households with children.
    - About 19 percent of recipients estimated to be non-eligible.
  - World Bank (2009) alternative estimates reported higher non-receipt: estimates that 60.6 percent of the poor do not receive a child compensation and that 69.9 percent of the individuals receiving a child compensation are not poor.
- Potential gains from reforms to means-testing:
  - Simplifying procedures and screening criteria could:
    - Reduce disincentives to apply and reduce inappropriate disqualification.
    - Close loopholes allowing non-eligible households to receive benefits.
    - Reduce administrative costs.
  - Adjusting income-test levels (for example for child allowance) combined with improved allocation could increase efficiency and reduce child poverty.
  - Package of reforms could have limited fiscal cost: increased costs from reduced non-take-up could be offset by reduced leakage, adjusted income-test levels, and lower administrative costs.

### Other relevant quantitative indicators and dynamics
- Historical tax-to-GDP decline:
  - In the 1980s and 1990s, large increase in Gini accompanied by a 12.6 percentage-point decline in the tax-to-GDP ratio; decline primarily explained by a 9.6 percentage points drop in direct-taxes-to-GDP ratio (Chu et al., 2004).
  - Comparative declines: Hungary -5.1 percentage points; Poland -4.4 percentage points (periods referenced in source).
- Social benefits relative level and trend visualizations referenced:
  - Efficiency and frontier analyses displayed in Figures 13–14 and referenced charts showing reduction in Gini achieved by social benefits and social benefits as percent of GDP.
  - Figure 15 links Gini coefficient of equivalized disposable income and social benefits in cash (index series and inverted vertical axis for social benefits).

### Policy implications and recommendations (synthesized from analysis in text)
- Taxation:
  - Introduce exemption threshold to the flat tax or progressive rates for personal income tax to increase progressivity.
  - Alternatively, raise flat tax rate (currently 10 percent) to finance larger social protection spending.
  - Use social contributions to target redistribution (reduce contributions on lower wages; increase or remove cap on contributory income—current cap BGR 3,000 per month or about USD 1,710).
- Social protection design:
  - Reinforce means-testing efficiency: simplify procedures, tighten screening to reduce leakage and non-take-up, and lower administrative costs.
  - Adjust income-test levels for benefits (e.g., child allowance) to better target the poor.
  - Reintroduce or strengthen systematic indexation mechanisms for key social transfers to prevent erosion of benefit value relative to wages and GDP.
  - Consider reversing decline in social spending (share of GDP) and systematically adjust levels of existing benefits (notably minimum pension).
- Expected fiscal trade-offs:
  - A package improving take-up and targeting could have limited net fiscal cost as reduced leakage and lower administrative costs offset higher take-up.

*Italic source: Excerpt from wpiea2020147-print-pdf (IMF staff analysis and figures as provided in source content).*

### 8.9 percent of GDP in 2018 contributing to

### wpiea2020147-print-pdf - 8.9 percent of GDP in 2018 contributing to

### Pension system: financial sustainability, access, and adequacy
- Bulgaria prioritized financial sustainability of pensions through a deindexation of pensions in the early 2010s and a pension reform implemented starting in 2016.
- Consequences of reforms:
  - Access to pension has been reduced. The recent pension reform increased the retirement age and the required retirement contribution. The number of pensioners has been declining since the introduction of the pension reform despite the ageing of the population.
  - Pensions are low compared to wages. Pensions are the only social benefit subject to automatic indexation. The indexation (currently based half on past inflation and half on projections of contributory income) was suspended for several years at the beginning of the 2010s and did not prevent average pension growth from lagging wage growth.
  - Average pension dynamics: in the past decade, average pension declined from being close to the minimum wage to about 70 percent of minimum wage.
  - Minimum and maximum pension rules: pensions are subject to maximum and minimum levels which are not indexed. The absence of automatic and systematic indexation of minimum and maximum pensions results in irregular adjustments and explains why the minimum pension, which was received by 36 percent of pensioners in 2018, was equivalent to 57 percent of minimum wage in 2010 but only 38 percent in 2018 and is persistently below the official poverty line.
- Behavioral effects:
  - Many pensioners continue to work to supplement their pension (but are not eligible to unemployment benefits when they lose their job). This reduces the impact of low pensions on old-age poverty and income inequality but did not prevent the real disposable income of the elderly from growing less than the income of other age groups.

### Concentration of social protection spending and coverage gaps
- Social protection in Bulgaria is concentrated on a few social risks: “Old age” and “family and children” account for 92 percent of total social protection spending.
  - This concentration is significantly more than in the NMS (75 percent on average) and the EU27 (70 percent on average).
- Implication: narrow concentration implies limited protection against other social risks; individuals experiencing shocks could easily fall into poverty given relatively low wealth levels.

### Unemployment benefits: eligibility and duration
- Unemployment spending is very limited in Bulgaria for two main reasons:
  - Eligibility criteria are stringent. To be eligible to unemployment benefits, an unemployed must be registered to the employment agency, not be entitled to an old-age pension, and have contributed for at least 12 months during the last 18 months.
  - Only 32 percent of the registered unemployed received unemployment benefits (2007–19 average).
  - The unemployment benefit is provided for a relatively short period of time (from 4 months to 12 months) which depends on the length of contribution history.
- Benefit calculation and duration details:
  - For eligible unemployed, the unemployment benefit is equal to 60 percent of the average contributory income over the last 24 months (subject to a minimum and a maximum daily rate).
  - It is granted for 4 months if the contributory history of the unemployed is 3 years or less. To receive the benefit for the maximum 12-month period, a contributory history of over 15 years is required.
- Consequence for youth: an insurance-based unemployment system whose benefits depend on employment history provides little protection to younger workers and contributed to the rise of youth poverty; in Bulgaria youth poverty increased more than for other working age workers and is now above the national average.

### Inequality, poverty, and policy recommendations
- Trends and levels:
  - Disposable income inequality in Bulgaria is rising and has become the highest in the EU.
  - Fiscal redistribution declined and is now among the lowest in Europe; this decline (due to the introduction of the flat tax and to low and declining social spending, whose efficiency has fallen below NMS average) contributed to a significant increase in disposable income inequality.
  - Poverty is among the highest in the EU. The share of population facing severe material deprivation, despite having dropped markedly, is still 3.6 times higher than the EU average, and, at 32.7 percent, is particularly high for the elderly.
  - One elderly out of four is at persistent risk of poverty compared to about one out of ten on average in the EU.
  - Children’s poverty rate, which was similar to old-age poverty in 2009, increased less than for any other age group in the past decade but remains a concern given its high level.
- COVID-19 impact:
  - As prior pandemics, the COVID-19 will affect inequality and poverty. The expected drop in activity and increase in unemployment and underemployment is likely to increase absolute poverty despite mitigating measures put in place by the authorities.
  - The revenue loss associated with the economic impact of the pandemic and containment measures will differ across groups, likely affecting revenue distribution and thus income inequality.
- Policy options to reduce inequality and poverty:
  - Adequate spending on education and health as well as appropriate labor-market regulations to foster a less unequal distribution of market income.
  - Increase fiscal redistribution through:
    - (i) increasing the redistributive role of taxation;
    - (ii) addressing issues with the means-testing of social benefits to increase efficiency;
    - (iii) reversing the recent decline in social protection spending.
  - These measures can be fiscally neutral. Improving the efficiency of social protection schemes and increasing the redistributive role of taxation can be designed to be fiscally neutral.
  - Reversing the recent decline in social protection spending would align Bulgaria with NMS average. It would increase spending by 1 percent of GDP, an amount that can be covered by improved revenue mobilization.
    - Options for improving revenue mobilization include increasing property tax revenue (which is 1.7 percent of GDP lower than EU average) and reducing the sizable VAT compliance gap.

### Key statistics and figures (preserved exactly as presented)
- 8.9 percent of GDP in 2018 contributing to about 79 percent of the decline in social protection spending over the period.
- Minimum pension received by 36 percent of pensioners in 2018.
- Minimum pension equivalent to 57 percent of minimum wage in 2010 and 38 percent in 2018.
- Average pension declined to about 70 percent of minimum wage in the past decade.
- “Old age” and “family and children” account for 92 percent of total social protection spending.
- NMS average: 75 percent (old age and family and children).
- EU27 average: 70 percent (old age and family and children).
- Only 32 percent of the registered unemployed received unemployment benefits (2007–19 average).
- Unemployment benefit equal to 60 percent of the average contributory income over the last 24 months (subject to a minimum and a maximum daily rate).
- Unemployment benefit duration: 4 months to 12 months depending on contributory history (4 months if contributory history is 3 years or less; 12 months if contributory history is over 15 years).
- Severe material deprivation share: still 3.6 times higher than the EU average.
- Severe material deprivation for the elderly: 32.7 percent.
- Reversing the decline in social protection spending would increase spending by 1 percent of GDP.
- Property tax revenue is 1.7 percent of GDP lower than EU average.

*Source: wpiea2020147-print-pdf - 8.9 percent of GDP in 2018 contributing to (IMF staff analysis and figures as provided in the source content).*

### APPENDIX II. AT-PERSISTENT-RISK-OF-POVERTY RATE BY AGE GROUPS IN EU

### APPENDIX II. AT-PERSISTENT-RISK-OF-POVERTY RATE BY AGE GROUPS IN EU COUNTRIES (2018)

### Overview
- Source: Eurostat.
- Charts present country-level ordering of at-persistent-risk-of-poverty rates for the following age groups: 65 years or over; 18 to 24 years; 25 to 49 years; 50 to 64 years; Less than 18 years; Total.
- Axis scale for the age-group charts: 0 to 40.

### 65 years or over — country ordering (as presented)
- HUN
- DNK
- CZE
- NLD
- FRA
- GRC
- ESP
- FIN
- SWE
- BEL
- LUX
- POL
- CYP
- EU27
- AUT
- ITA
- SVN
- DEU
- ROM
- MLT
- BGR
- LVA
- EST

### 18 to 24 years — country ordering (as presented)
- SWE
- CYP
- HUN
- SVN
- AUT
- FRA
- LUX
- FIN
- DEU
- MLT
- CZE
- BEL
- DNK
- LVA
- EST
- EU27
- GRC
- POL
- BGR
- ROM
- NLD
- ESP
- ITA

### 25 to 49 years — country ordering (as presented)
- SWE
- SVN
- FIN
- HUN
- CZE
- NLD
- MLT
- FRA
- CYP
- DNK
- LVA
- EST
- LUX
- DEU
- POL
- EU27
- BEL
- AUT
- BGR
- ESP
- GRC
- ITA
- ROM

### 50 to 64 years — country ordering (as presented)
- DNK
- FIN
- SWE
- CZE
- CYP
- FRA
- BEL
- HUN
- AUT
- NLD
- MLT
- EU27
- POL
- DEU
- GRC
- BGR
- LUX
- SVN
- EST
- ESP
- ROM
- ITA
- LVA

### Less than 18 years — country ordering (as presented)
- DNK
- SVN
- FIN
- CYP
- HUN
- SWE
- LVA
- NLD
- DEU
- CZE
- POL
- EST
- AUT
- EU27
- BEL
- FRA
- MLT
- GRC
- ESP
- ITA
- LUX
- BGR
- ROM

### Total — chart present
- Chart axis scale: 0 to 40.
- Country ordering for the Total series is presented visually in the source chart (specific sequence shown in the graphic).

---

### APPENDIX III. ALTERNATIVE ESTIMATES OF THE EFFICIENCY OF SOCIAL BENEFITS

### Key findings (text from source)
- Due to differences in methodology (Appendix I), Eurostat and Euromod provide somewhat different estimates of the reduction in the GINI coefficient achieved by social benefits. These differences affects the measurement of the efficiency of social benefits (measured as the reduction in the GINI coefficient achieved by 1 percent of GDP in social benefit).
- As shown in the charts below, Eurostat estimates result in a higher redistributive power of social spending than Euromod similations but the trends are similar.
- In both cases, the efficiency of social benefits:
  - declined in bulgaria and in NMS between 2007 and 2018;
  - was higher in Bulgaria than NMS average in 2007;
  - was lower in Bulgaria than NMS average in 2008.

### Reduction in the Gini Coefficient Achieved with 1 Percent of GDP of Social Benefits
- Chart title scale: (Scale: 0 to 1)
- Panels presented: Euromod and Eurostat.

### Chart axis tick values and labels (as presented)
- Vertical axis tick values: 0.000, 0.001, 0.002, 0.003, 0.004, 0.005, 0.006, 0.007, 0.008, 0.009, 0.010, 0.011, 0.012, 0.013, 0.014, 0.015, 0.016
- Horizontal axis years: 2007, 2018
- Series labels shown: Bulgaria, NMS, EU

*Sources: Euromod, Eurostat, and IMF Staff calculations.*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020147-print-pdf.pdf_