Poverty & Equity Global Practice Working Paper 122 FISCAL INCIDENCE IN BELARUS: A COMMITMENT TO EQUITY ANALYSIS Kateryna Bornukova Gleb Shymanovich Alexander Chubrik October 2017 Poverty & Equity Global Practice Working Paper 122 ABSTRACT The paper employs the Commitment to Equity framework to present a first attempt at a comprehensive fiscal incidence analysis for Belarus, encompassing the revenue and expenditures components of the fiscal system, including direct and indirect taxes, as well as direct, indirect, and in-kind transfers. The analysis reveals that fiscal policies in Belarus effectively redistribute income from the top to the bottom of the income distribution. Direct transfers, especially pensions, are the most equalizing and pro-poor of the fiscal interventions—direct transfers and direct taxes lower the national poverty headcount by 17 percentage points and lower the Gini index of inequality from 0.407 to 0.267. Some of the indirect taxes, by contrast, are regressive and indirect transfers -- poorly targeted, such that the effect of these components of the fiscal system is not equalizing. Finally, the cost-efficiency of different parts of the fiscal system in Belarus varies considerably. Unemployment benefits, pensions, and child benefits are found to be cost -- efficient, while indirect subsidies are highly cost- inefficient. The analysis points toward possible reforms that would allow reducing poverty and inequality more efficiently. This paper is a product of the Poverty and Equity Global Practice Group. It is part of a larger effort by the World Bank to provide open access to its research and contribute to development policy discussions around the world. The authors may be contacted at Kateryna Bornukova, bornukova@beroc.by; Gleb Shymanovich, shymanovich@research.by; Alexander Chubrik, chubrik@research.by. The Poverty & Equity Global Practice Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. ‒ Poverty & Equity Global Practice Knowledge Management & Learning Team This paper is co-published with the World Bank Policy Research Working Papers. Fiscal Incidence in Belarus: A Commitment to Equity Analysis Kateryna Bornukova, Gleb Shymanovich, Alexander Chubrik*,** JEL Codes: H22, H5, D31, I3 Keywords: Fiscal incidence, social spending, inequality, poverty, Belarus * Kateryna Bornukova is academic director of BEROC Center for Economic Research, e-mail: bornukova@beroc.by; Gleb Shymanovich is economist at the IPM Research Center, e-mail: shymanovich@research.by; Alexander Chubrik is director of the IPM Research Center and fellow at CASE – Center for Social and Economic Research, e-mail: chubrik@research.by. ** We thank Alexandru Cojocaru, Mikhail Matytsin, participants of the World Bank CEQ workshop and referees for the excellent inputs, useful comments and suggestions. All errors are ours. We acknowledge the financial support from the World Bank. Contents List of the tables.................................................................................................................................................................................. 3 List of the figures ................................................................................................................................................................................ 3 1. Introduction ............................................................................................................................................................................... 4 2. Belarusian “welfare state”: Overall principles...................................................................................................................... 5 2.1. Poverty, growth incidence, and trends in inequality ....................................................................................5 2.2. Revenues and expenditures of the general government ..............................................................................6 Government revenues.......................................................................................................................................6 Government expenditures................................................................................................................................8 2.3. State social insurance system design ...............................................................................................................9 3. Methodology of the analysis ................................................................................................................................................. 10 3.1. Data available and CEQ approach to income concepts construction ................................................... 10 3.2. Direct taxes ...................................................................................................................................................... 12 Personal income tax and taxes on entrepreneurial income ...................................................................... 12 Social Protection Fund contributions.......................................................................................................... 13 3.3. Indirect taxes ................................................................................................................................................... 14 VAT .................................................................................................................................................................. 14 Excises .............................................................................................................................................................. 16 Import duties ................................................................................................................................................... 17 3.4. Direct transfers................................................................................................................................................ 19 3.5. Indirect Subsidies ............................................................................................................................................ 20 3.6. In-kind transfers (health care and education)............................................................................................. 22 4. Results and Discussion .......................................................................................................................................................... 23 4.1. Main results...................................................................................................................................................... 23 4.2. Distributional impact and marginal contributions of fiscal interventions ............................................. 29 4.3. Efficiency ......................................................................................................................................................... 31 4.4. Targeting and vulnerable groups .................................................................................................................. 33 4.5. Cross-country differences.............................................................................................................................. 34 5. Conclusions.............................................................................................................................................................................. 34 References .......................................................................................................................................................................................... 35 Appendix ............................................................................................................................................................................................ 37 A. Estimation of personal income tax .............................................................................................................. 37 B. Concentration curves ..................................................................................................................................... 38 C. Distribution of gains and losses at the level of final income ................................................................... 41 2 List of tables Table 1. General government revenue ........................................................................................................................................... 7 Table 2. General government expenditure and balance.............................................................................................................. 8 Table 3. Share of VAT in consumer prices by expenditure line .............................................................................................. 15 Table 4. Assumptions and inputs for estimating household payments of alcohol excises ................................................. 16 Table 5. Estimated parameters of alcohol consumption underreporting and share of excises in household expenditures by quintile................................................................................................................................................................... 17 Table 6. Assumptions and inputs for estimating import duties ............................................................................................... 17 Table 7. Main poverty and inequality indicators by income concepts .................................................................................... 24 Table 8. Decompositions of inequality changes into vertical and horizontal equity components.................................... 25 Table 9. Incidence of net effects from fiscal interventions in relation to market income by deciles ............................... 25 Table 10. Fiscal gains to the poor and fiscal impoverishment in relation to market income............................................. 28 Table 11. Fiscal impoverishment in relation to market income by social vulnerable groups, headcount, % of group 29 Table 12. Progressivity of taxes and transfers in relation to income concepts ..................................................................... 30 Table 13. Marginal contributions to inequality and poverty..................................................................................................... 31 Table 14. Efficiency measures for social fiscal interventions ................................................................................................... 32 Table 15. Sizes of transfers/taxes by vulnerable groups, relative to the rest of the population ........................................ 33 List of figures Figure 1. Poverty and economic growth, 2000–2015 .................................................................................................................. 5 Figure 2. Growth incidence curves in Belarus, 2000–2015 ........................................................................................................ 6 Figure 3. Redistribution and inequality in Belarus, 2000–2015.................................................................................................. 6 Figure 4. General government balance, 2000–2015 .................................................................................................................... 7 Figure 5. Ratio of HBS data and data of national accounts, % ............................................................................................... 10 Figure 6. Construction of income concepts ................................................................................................................................ 12 Figure 7. Incidence of direct taxes by deciles .............................................................................................................................. 14 Figure 8. Incidence of indirect taxes by deciles........................................................................................................................... 15 Figure 9. Incidence of import duties by deciles .......................................................................................................................... 18 Figure 10. Incidence of cash and in-kind benefits by deciles ................................................................................................... 20 Figure 11. Indirect subsidies by disposable income deciles ...................................................................................................... 21 Figure 12. Health and education expenditure by disposable income deciles ........................................................................ 22 Figure 13. Lorenz curves for basic income concepts ................................................................................................................ 24 Figure 14. Distribution of individuals by market and disposable income ............................................................................. 26 Figure 15. Distribution of gains and losses at the level of disposable income with respect to market income by income bins set in USD PPP ......................................................................................................................................................... 27 Figure 16. Structure of net beneficiaries and payers of fiscal system by socio-economic status ....................................... 29 Figure 17. Cross-country comparisons of redistribution and poverty reduction effect of direct transfers and taxes 34 Figure 18. Concentration curves for direct taxes........................................................................................................................ 38 Figure 19. Concentration curves for indirect taxes .................................................................................................................... 38 Figure 20. Concentration curves for direct transfers ................................................................................................................. 39 Figure 21. Concentration curves for indirect subsidies ............................................................................................................. 39 Figure 22. Concentration curves for in-kind expenditure......................................................................................................... 40 Figure 23. Distribution of gains and losses at the level of final income with respect to market income by income bins set in USD PPP................................................................................................................................................................................. 41 3 1. Introduction Belarus is often positioned as a country that has a “socially oriented economy,” as stated by the authorities. This statement is supported by low income inequality (one of the lowest in the region) and low incidence of poverty (the poverty headcount based on the international poverty line of USD 4 PPP is effectively zero). Different studies (e.g. Chubrik, 2007; Chubrik and Shymanovich, 2016) reveal the pro-poor nature of Belarusian economic growth, but there is no clear evidence whether low inequality and poverty have resulted from tax and subsidies systems design or are due to other factors. This paper seeks to fill this knowledge gap by analyzing the impact of fiscal policy on poverty and inequality in Belarus. Methodologically, the analysis follows the Commitment to Equity (CEQ) analysis, which has already been applied in more than 30 low- and middle-income countries (Lustig, 2016). This fiscal incidence analysis reveals who are the beneficiaries of public social expenditures and the contributors to the public finances, who bear the major tax burden. The assessment of the impact of fiscal policies is timely for Belarus. Currently, the country is struggling with a prolonged recession and has to optimize its budget expenses. So far, the reform debate has been centered around the pension reform (Lisenkova and Bornukova, 2017, Shymanovich, 2016), and the elimination of utility subsidies (IMF, 2016; Chubrik, Shymanovich, 2016; Zhang and Hankinson, 2015). This paper seeks to inform the current debate with information on the poverty and inequality impacts of social programs and their cost efficiency. To the best of our knowledge, this is the first attempt at a comprehensive fiscal incidence analysis for Belarus. The fiscal incidence approach captures only the effects of government policies in the form of taxes, subsidies and benefits collected from and provided to households. However, part of the social support is provided implicitly through the subsidization of state-owned enterprises (SOEs). It is partially reflected in budget expenditures (in 2015, 4.3% of GDP was spent on subsidies to the SOEs), and partially comes through quasi-fiscal operations not captured by fiscal data. On the one hand, this support helps SOEs to preserve excessive employment, on the other – it leads to inefficient resource allocation, thus reducing overall welfare. Thus, in addition to the general problems with imputation of the effects of SOEs subsidization at the household level, there is an open issue of the overall “sign” of its impact on poverty and inequality: it could be that lower subsidies to SOEs would result in faster job creation in the private sector and better job opportunities for the poor. That is why in this study we do not consider the social roles of SOEs, focusing only on the taxes paid by households and subsidies provided to them directly from the budget. Other limitations of the CEQ approach are that it does not evaluate the quality of government services, does not take into account the behavioral/rational responses to changes in the fiscal policy, and assumes an equal distribution of income and consumption within the household. Our results suggest that fiscal policy in Belarus is very effective in lowering both poverty and inequality. Direct transfers (including pensions) and direct taxes lower the national poverty measure by 17 percentage points. They also decrease the Gini index from 0.407 to 0.267. The impressive magnitude of positive fiscal effects puts Belarus among the equalization leaders in the group of developing countries. Most of the effect could be attributed to pensions. When we adopt the pensions-as-deferred-income (PDI) approach, the poverty reduction amounts only to 2.5 percentage points, and the Gini coefficient decreases only by 0.02. The results also point in the direction of possible reforms. As the government seeks to minimize expenditure, it is important to focus on the most efficient interventions. Indirect subsidies are highly cost-inefficient. A 1% of GDP spent on the utility subsidies delivers 3 times less poverty and inequality reduction compared to the same 1% spent on pensions. Indirect utility subsidies and transport tariffs are not targeted, available to everybody and regressive. Replacing indirect subsidies with well-targeted benefits programs would allow reducing poverty and inequality more efficiently. Unemployment benefits (currently at very low levels) are the most cost-efficient benefit program, suggesting that the plans to increase benefits will have significant impacts in reducing poverty and inequality. The paper proceeds as follows. We describe the welfare state in Belarus in Section 2, also discussing the role of quasi-fiscal policies for social welfare. Section 3 describes the CEQ methodology and peculiarities of its application to the Belarusian data. In Section 4 we present and discuss the results of the CEQ assessment and the fiscal impact on poverty and inequality in Belarus. Section 5 concludes. 4 2. Belarusian “welfare state”: Overall principles 2.1. Poverty, growth incidence, and trends in inequality In its recent history, Belarus demonstrated an impressive reduction in poverty. The poverty headcount based on the official poverty line1 fell from 46% in 1999 to 4.8% in 2014; it has stayed well below 10% of population since 2007 (see Figure 1). Poverty measured with the international poverty line of USD 4 PPP was below 1% since 2011, and since 2013 it is close to 0. Economic growth (quite impressive between 2000–2010, when Belarus was among top-25 fastest growing countries in the world) was the key factor behind poverty reduction – the correlation be- tween real GDP and the national poverty headcount is -0.91. The outliers are explained by extremely fast growth of housing and utility tariffs (2002) and a hyperinflation episode (2011). However, Belarusian economic growth was not sustainable. It was driven by fast capital accumulation financed initially from the budget and later via directed lending at preferential interest rates. As a result, returns on invest- ment fell, as well as total factor productivity (Kruk and Bornukova, 2014; 2015). On the demand side, GDP growth was driven by domestic demand; fast growth of investment and household consumption led to growing external imbalances that were financed via growing government borrowing. Altogether, the unsustainability of these factors caused the economic recession, which started in Belarus at the end of 2014, and growth prospects look gloomy: recent IMF and World Bank outlooks forecast very modest growth,2 while statistical filters give real GDP long term trend growth rate below zero (Chubrik and Shymanovich, 2016). The recent GDP decline was associated with a slight increase in poverty. The poverty rate according to the national definition grew from 4.8% in 2014 to 5.7% in 2016. Moreover, the national definition of poverty does not properly take into account the significant increases in utility tariffs, which have been taking place since 2014. Regional inequality has also been increasing, with the population outside the capital, and in particular in small cities and in rural areas lagging behind the large urban centers in terms of wages and other types of income (Chubrik, 2016b; Mazol, 2016). Figure 1. Poverty and economic growth, 2000–2015 95 44 90 40 85 36 80 32 75 28 70 24 65 20 60 16 55 12 50 8 45 4 40 0 2000 2001 2002 20032004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Real GDP, BYN bn (2015=100) Poverty headcount, national poverty line, % of population (RHS) Poverty headcount, $4 PPP, % of population (RHS) Source: Belstat, World Bank POVCAL (USD 4 PPP headcount). Such a strong correlation between poverty and economic growth should mean that Belarusian economic growth on average had a pro-poor nature. Indeed, in general, the higher the initial income, the lower the rate at which it grew between 2000 and 2015. Income of the poorest decile grew by 0.65 percentage points a year faster than income of the richest decile (see Figure 2a). However, over time the real income growth rate was falling, following the real GDP growth rate, and profiles of the incidence curves changed as well. The rich benefited the most between 2005 and 2010, while the poorest – between 2010 and 2015, and the lower middle class – between 2000 and 2005 (Figure 2b). In the end, Belarus succeeded in delivering benefits of economic growth to all household groups, including 1 Absolute poverty line (“minimum subsistence basket”, or “subsistence minimum”) is calories-based poverty line; before 3Q2014, it in- cluded administratively defined set of food and non-food goods and basic services, since 3Q2014 it is calculated as the value of administra- tively defined food basket times 1.77. 2 See IMF WEO database, April 2017 (https://www.imf.org/external/pubs/ft/weo/2017/01/weodata/index.aspx), and World Bank’s Belarus Economic Update, May 2017 (http://pubdocs.worldbank.org/en/819391494832531504/Eng-EcUpdate-May14-17.pdf). 5 (and especially) the poor, and the existing system of income redistribution could be one of the important reasons behind this achievement. Figure 2. Growth incidence curves in Belarus, 2000–2015 10.4 16 15 10.3 14 10.2 13 12 10.1 11 10.0 10 9 9.9 8 7 9.8 6 9.7 5 4 9.6 dec01 dec02 dec03 dec04 dec05 dec06 dec07 dec08 dec09 dec10 dec01dec02dec03dec04dec05dec06dec07dec08dec09dec10 2000-2005 2005-2010 2000-2015 mean median 2010-2015 mean, 2000-2005 (a) Real disposable resources growth rates, (b) Real disposable resources growth rates, 15-year annual average, % 5-year annual averages, % Source: own estimates based on HBS data (disposable resources) and Belstat data (inflation). Historically, redistribution played an important role in Belarus. The share of general government expenditures in GDP stayed at the level of European welfare states (47.2% on average between 2000 and 2010). Even after impres- sive fiscal consolidation, when the average government expenditures dropped by 6.7% of GDP (see Figure 3), that share remained higher than in the upper-middle-income countries from Central and Eastern Europe and CIS. The fiscal consolidation of the last five years resulted only in a very moderate increase of the Gini index: from 0.266 between 2000 and 2010 to 0.281 between 2011 and 2015. Figure 3. Redistribution and inequality in Belarus, 2000–2015 51 0.290 49 0.285 47 0.280 45 0.275 43 0.270 41 0.265 39 0.260 37 0.255 35 0.250 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 General government expenditure, % of GDP Average, 2000-2010 Average, 2011-2015 Gini (RHS) Source: GFS, Belstat. 2.2. Revenues and expenditures of the general government Government revenues The need for fiscal consolidation was driven by a drastic reduction of general government revenues during the currency crisis of 2011 that had not been restored completely. As fiscal policy was quite conservative (the budget had a surplus of 0.6% of GDP on average between 2000 and 2010 and 1.4% of GDP between 2011 and 2015, see Figure 4a), general government expenditures followed revenues. That “conservative” policy was imposed by the size of operations “below the line”: deep interventions of the state into the economy required regular recapitaliza- tion of the largest state-owned banks and other types of support to the state-owned companies. For instance, 6 during the severe currency crisis of 2011 the government spent about 12% of GDP on net acquisitions of financial assets – far above the fiscal surplus of 3% of GDP, which required a substantial debt increase and assets sales. After the crisis of 2011, the size of these operations became smaller (see Figure 4b), but because of the debt accumulated between 2007 and 2011, the government needed to keep the fiscal surplus in order to make payments on the principal, which limited its capacity to redistribute. Figure 4. General government balance, 2000–2015 53 4.00 12 51 3.33 10 49 2.67 47 2.00 8 45 1.33 6 43 0.67 41 0.00 4 39 -0.67 2 37 -1.33 35 -2.00 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 -2 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Balance (RHS) Revenues Expenditures Net acquisition of financial assets Average, 2000-2010 (RHS) Net incurrence of liabilities Average, 2011-2015 (RHS) Balance (a) General government revenues, expenditures, and balance, (b) General government balance* and its financing, 2000–2015, % of GDP 2000–2015, % of GDP * Net lending (+)/net borrowing (-). Source: GFS. The most stable sources of government revenue are contributions to social insurance and personal income tax (see Table 1), as their tax base is mainly wage income, whose share in GDP is quite stable. VAT and excise taxes are also quite stable, relying mainly on household consumption, which is even less volatile than household incomes. In 2015, these four sources together generated 60.1% of all general government revenues, while social insurance contributions and VAT – 45.1%. Table 1. General government revenue BYR mln* % of GDP Total Revenue & Grants 37 666.540 43.3 Tax Revenue 31 991.929 36.8 Direct taxes of which 7 319.485 8.4 Personal Income Tax 3 700.907 4.3 Corporate Income Tax 2 384.990 2.7 Taxes on Property 1 233.588 1.4 Contributions to Social Insurance 9 715.236 11.2 Indirect Taxes of which 14 853.340 17.1 VAT 7 267.080 8.4 Turnover & other general taxes on goods and services 567.897 0.7 Excise Taxes 1 944.165 2.2 Customs Duties 864.359 1.0 Taxes on Exports 2 992.432 3.4 Other indirect taxes 1 217.408 1.4 Other taxes 103.867 0.1 Nontax Revenue 5 599.073 6.4 Grants 75.539 0.1 * Taking into account denomination in 10 000 times of 2016. Source: GFS. The corporate income tax, taxes on exports, and non-tax revenue together brought 29.1% of the general govern- ment revenues, but they are far more volatile for a variety of reasons. The corporate income tax and a large portion of non-tax revenue depend on the financial status of the SOEs. Thus, on the one hand, the government subsidizes them, on the other – withdraws profit and collects corporate income tax. Ceteris paribus, the lower the subsidies, the lower the SOEs’ profit and the related government revenues. Fiscal challenges force the government to reduce subsidies, and, hence, the tax/revenue base. The size of the revenue from taxes on exports depends on the current 7 design of the agreements between Belarus and the Russian Federation concerning crude oil and oil products trade. Currently, Belarus gets all export duties on oil products produced by Belarusian refineries from Russian oil (which was not the case between 2007 and 2014), but Russia may cut the oil supply to Belarus, reducing its exports and related budget revenues. Government expenditures A quick look at the structure of general government expenditures supports the “social orientation” of the fiscal policy: social spending amounts to 64.9% of the general government expenditures (see Table 2). The biggest por- tion (10.4% of GDP) is spent on old age pensions (from the Social Protection Fund and directly from the budget). Another 2.6% of GDP is spent on social allowances on temporary disability, childbirth allowance, family allow- ances, maternity pay, and disability. The IMF (2016) estimated government expenditures on housing and utilities subsidies to households of 1% of GDP. Table 2. General government expenditure and balance BYR mln* % of GDP Total Expenditure & Grants 35 629.917 41.0 Social Spending 23 131.264 26.6 Social Protection 12 991.387 14.9 Social Assistance of which 2 172.121 2.5 Noncontributory Pensions (Expenditure on social 684.931 0.8 protection, line “expenditure on old age”) Expenditure on family & children 168.322 0.2 Expenditure on housing 646.521 0.7 Other 672.346 0.8 Social Insurance (Social Protection Fund)** of which 10 819.266 12.4 Old-Age Pensions** 8 359.478 9.6 old age carer’s allowance, and funeral assis- 4 649.904 5.3 tance (financed by the SPF). Public expendi- tures on health care and education in Belarus are at the level of advanced economies (as a percentage of GDP). Indirect social spending also quite substantialEducation1 of which Pre-primary and primary 977.513 1.1 Secondary 1 918.113 2.2 Post-secondary non-tertiary 476.392 0.5 Tertiary 793.725 0.9 Health2 3 872.301 4.5 Expenditure on housing & community amenities of which 1 617.671 1.9 Expenditure on community development 1 044.949 1.2 Spending on Defense, Public Order and Safety 2 622.543 3.0 Expenditure on public debt transactions 1 504.896 1.7 Grants 71.551 0.1 Other Government Expenditure 8 299.667 9.5 Fiscal Balance Primary net lending (+) / borrowing (–) 3 541.516 4.1 Net lending (+) / borrowing (–) 2 036.620 2.3 * Taking into account denomination of 2016. 1, 2 Ministry of Finance of Belarus provides different figures for education and health care – BYR 4 186.4 mln and BYR 3 497.7 mln (de- nominated) respectively. These figures do not include investment financed within government investment programs. National classification by functions of government puts all financing of government investment programs under the line “Expenditure on general public services”, while GFS distributes these expenditures between the respective functional lines. Further imputation of health care and education subsidies is based on the Ministry of Finance data. Source: GFS, except ** – Social Protection Fund of Belarus. Although a large portion of the government expenditures was directed to subsidies to SOEs (4.7% of GDP in 2015), public investment (2.7% of GDP), and debt service (1.7% of GDP), the overall design of the redistribution system allows for inequality to remain relatively low. Not only government’s social programs, but also subsidies to the SOEs contribute to income redistribution. First, at least a part of SOEs have excessive employment (see Favaro et al., 2012; World Bank, 2012), such that government subsidies enable them to pay salaries to potentially unem- ployed people. Second, most of the SOEs bear the costs of cross-subsidization of utility tariffs, and, once again, government subsidies help them to pay higher tariffs for electricity, gas and utilities, while households pay below the level of cost recovery. Finally, SOEs apply the so called “wage grid” that sets different markups to some basic wages for different types of employees and puts limits on the difference between maximum and minimum wages 8 at any particular enterprise. Wage setting at private companies is not regulated through this mechanism, i.e. wage regulation at SOEs is inequality reducing. However, one should not overestimate the role of SOEs in the overall system of social support. First, between 1995 and 2016, their share in total employment fell from 60% to 40%, and excessive employment fell accordingly (see Chubrik, 2016a). Second, comparing to the pre-crisis level, the government cut its direct and indirect support to SOEs: subsidies to SOEs fell from 9% of GDP in 2008–2009 to 4.3% of GDP in 2015; and the directed lending portfolio – from 25.2% of GDP in 2010 to 21.4% of GDP in 2015 (IMF, 2016). These cuts leave fewer resources for the “social” functions of SOEs. 2.3. State social insurance system design The Belarusian pension system preserved the main features of the PAYG system formed in the Soviet Union. There is no mandatory funded pillar with defined contribution design, and there is a rudimentary third pillar (mainly in the form of life insurance). It is organized in the form of the state social insurance system (hereafter – SSI), which is operated by the Social Protection Fund (hereafter – SPF) and funded mainly by a payroll tax (called “insurance contribution”). The tax rate is 35% of the total wage fund, of which 34% is funded by employers, and an additional 1% is paid from employees’ wages. In 2015, revenues collected via the employer contributions amounted to 91.8% of the SPF revenues. The second biggest source of the SPF revenues (5%) is subsidies from the central budget, i.e. the current design of the SSI system cannot ensure complete funding of its obligations. Out of the 35% rate, 29 percentage points are directed to paying pensions, while the remaining 6 percentage points go to social allowances on temporary disability, childbirth allowance, family allowances, maternity pay, disabil- ity/old age care (attendance) allowance, and funeral assistance.3 Within the pension system, the majority of employees are subject to a 29% tax rate, of which 28% is paid by the employer and 1% is paid from employees’ wages (although it is also accounted as employer contribution). Physical persons (e.g. individual entrepreneurs) usually pay 29% of the minimum wage for the accounting period. Employ- ers – agricultural producers pay 25% (24+1). A very small group or employers (e.g. public associations of people with disabilities, pensioners, etc.) are subject to a 6% tax rate (5+1). Individuals employed in the High-Tech Park have a celling for this tax base amounting to one average wage in the economy (others have a celling of five average wages). The self-employed and those who get paid according to civil law contracts with foreign organizations do not pay contributions to the SPF (but they can do so at will). Several categories of employees are not subject to the state social insurance – military servants and the command and private personnel of the interior, and several state controlling, investigation, and emergency agencies. They do not pay contributions and receive pensions di- rectly from the central budget. In addition, retired government officials with a state service record in excess of 20 years receive an additional pension also paid from the central budget.4 Taking into account the high share of formal employment, the coverage is quite high: of the employed population of 4.5 million, about 3.4 million employees, 0.3 million individual entrepreneurs, and 0.3 million in other categories paid contributions in 2016.5 However, the working age population is shrinking, while the number of pensioners is growing, such that the dependency ratio increased from 44.1% in 2000 to 48.2% in 2016. Between 1956 and 2016, the pension age in Belarus remained constant: 55 years for women and 60 years for men. Starting with 2017, the pension age will increase by 6 months a year until it reaches 58 and 63 years, respectively. In addition, to be eligible for an old-age pension, in 2015 a person should have an “insurance record” (i.e. the period of paying contributions to the SPF) of no less than 15 years. In was increased from 5 years to 10 years (since 3 In addition, the SPF finances targeted social assistance and employment promotion, professional pensions, sanatorium-resort rehabilita- tion, etc. 4 The pension system of Belarus is regulated by laws (“On Basic Provision for State Social Insurance”, “On Pension Provision”, “On Civil Service in the Republic of Belarus”, “On Pension Provision for Military Servants, Command and Private Personnel of the Interior, Investi- gation Committee of the Republic of Belarus…”, etc.), Presidential edicts (“On the issues of Social Assistance”, “On the Soci al Protection Fund of the Ministry of Labor and Social Protection”, etc.), and other legislative acts. 5 High formal employment and its coverage with social security contributions are partially inherited from the times when state-owned enterprises dominated as employers. However, although their share in total employment fell from 60% to 40% between 1995 and 2016, the coverage of the employed population with social security contributions remained stable. Thus, SOEs are not the main “donor” o f the Belarusian social protection system anymore, while private companies have similar payment discipline. In addition, despite the subsidies cut, the discipline of payments to the SPF remained very high: as of the beginning of 2010, overdue arrears for taxes, duties, and social contri- butions amounted to 0.22% of GDP, in 2016 – to 0.18% of GDP. In other words, SOEs’ payments to the SPF are their own burden, not those of the state budget, at least not anymore. 9 2014) and to 15 years (since 2015). Starting with 2016, this record is increasing by 6 months a year until it reaches 20 years. The design of the pension system ensures substantial income redistribution. First, an old age pension is equal to 55% of the “wage base”, but not less than the minimum old age pension.6. Second, one year of work record above 25 years for men and 20 years for women adds 1 percentage point to the 55%, but no more than 20% of wage base (or minimum old age pension). And the most serious redistribution comes from the method of the wage base calculation: wage base  w1  w 2 , where w1  0.13  w  ( w  0.13  w )  0.45, if ( w / w )  1.3, w 2  ( w  1.3  w )  0.1, if 1.3  ( w / w )  4, where w is average wage of an individual accounted for pension calculation and w is the average wage in the economy in the same period. The redistributive effect is provided in the table below. Actual wage, % of average wage in the economy 50 100 130 400 500 Wage base, % of average wage in the economy 29.65 52.15 65.65 92.65 92.65 Wage base, % of actual wage 59.30 52.15 50.50 23.16 18.53 As a result, the Gini index for old-age pensioners7 in 2015 was 0.126, while for employees8 – 0.269. 3. Methodology of the analysis 3.1. Data available and CEQ approach to income concepts construction The analysis is based on the Household Budget Survey (HBS) data. This survey is conducted each year starting from 1995. It covers all oblasts and Minsk city, and includes observations from around 50 towns and rural councils. The sample of the survey is expected to be 6,000 households (0.2% of general population). In 2015 the actual sample included 6,269 households, including 313 households with zero sampling weights, as they did not provide basic information about their income and expenditures. The remaining households represent 9.1 million people or 96.3% of total population. The sample does not cover collective households, i.e. care homes, students’ dormitories, spe- cialized institutions, etc. As any other survey of this kind, it does not properly represent the richest households and the most marginalized households, which refuse to participate in the survey. The sample is structured to be repre- sentative at the country level for key population groups and for total population at the oblast level. Still it inevitably has some distortions. For instance, the sample overestimates rural population by 10.1% and underestimates urban population by 7.8%. Figure 5. Ratio of HBS data and data of national accounts, % Consumption Disposable Income Rural population Urban population Total population 0 20 40 60 80 100 120 6 The minimum old age pension is equal to 25% of the subsistence minimum plus 20% of the average wage. As of 2017, it is 5.5% below the subsistence minimum. 7 HBS data: women 55+ and men 60+ who have pension income (total – 2,448,272 individuals). 8 HBS data: for those who received wages during 12 months, sum of all wage-related incomes (total – 3,186,139 individuals). 10 Source: Belstat. HBS data are used by the Statistical Committee while computing national accounts. However, HBS data perma- nently underestimate household consumption if compared to retail statistics. In 2015 the scale of underestimation was extraordinarily high (37.4%, see Figure 5). First, it is related to the increased volume of consumer lending. A purchase of goods on terms of consumer loans is reflected in the survey as loan servicing expenditures by house- holds instead of consumption expenditures, which creates differences between HBS and retail statistics. Therefore, disposable income of households calculated based on total expenditures within HBS is substantially lower (22.3%) than estimated within national accounts. Second, households traditionally underreport expenditures on alcohol consumption, which is a common problem of the surveys. Third, households tend not to report purchases of tobacco and fuel (and alcohol as well) for the purpose of further resale abroad, which is a widespread coping strategy in western regions of Belarus. Hence, these expenditures, accounted in retail statistics, are not actually household consumption, but rather costs associated with their entrepreneurial activity, so they should not be taken into account while calculating households’ consumption or disposable income. HBS data can be used to estimate household welfare and overall macroeconomic effects without additional adjust- ments as it represents almost the whole population and covers household expenditures in full, with the exception of alcohol consmption. Belstat uses these data for poverty and living standard analysis. Poverty analysis is based on comparison of disposable household income with the absolute poverty line. Disposable income is officially calculated as a sum of total household expenditures, net in-kind income and privileges (in-kind benefits). Hence, it is calculated based on reported expenditures rather than reported income, as it is believed to be underestimated (total reported cash income was 3.5% less than total reported expenditures in HBS 2015). The absolute poverty line is set at the level of subsistence minimum for a member of a household containing two adults and two children. Following the official approach and assuming underreporting of households’ income, we also conduct the CEQ analysis based on disposable income data, assessed through the expenditure side. We also apply the same national poverty line for analysis of fiscal effects on poverty. However, we are not able to match the official estimate of poverty for 2015, as Belstat uses quarterly data for its estimation, while we work with the annual file. In addition to the national poverty line, we also calculate moderate poverty based on the annual average minimum consumer budget set for a member of a household containing two adults and two children. Nowadays, this budget is not widely used for the purposes of social policy. Still it is believed to serve a threshold for determining households with a risk of vulnerability. For instance, it is used as an eligibility criterion for privileged loans. The core element of the CEQ analysis is the calculation of income concepts. Based on available data, we take dispos- able income as a starting point (see Figure 6). Subtracting reported direct transfers from disposable income and adding estimated direct taxes, we calculate market income. There are two approaches of assigning direct taxes and transfers based on the pension system of a country. Pensions can be viewed either as a government transfer (PGT) or a deferred income (PDI). In the first case, it implies that social security contributions are accounted as direct taxes while pensions are added to direct transfers. In the second case, pensions and related contributions are not taken into account while estimating market income – pensions are considered as a part of both market and disposable income concepts. The pension system in Belarus is purely pay-as-you-go. The link between contributions and actual pension payments in Belarus is quite weak (see Section 2.3). From this point of view, is it more natural to consider pensions as transfers similarly to other contributory programs (like unemployment benefits). The pension system in Belarus is redistributive, effectively weakening the link between the market income and pension income after retirement. Hence, even if we agree that the effects of the pension system on poverty are debatable, the redistribution effect is the direct consequence of the government fiscal policy and should be taken into account when the fiscal effects are analyzed. Finally, two important benchmark cases for the Belarus CEQ study – Russia (Lopez-Calva et al., 2017) and EU (based on EUROMOD) consider pensions as public transfers in the main scenario (in case of Russia) or as the only scenario (EU). Similar treatment of pensions in Belarus will allow for proper comparison with these countries. For the above reasons, we chose to model pensions as government transfers (PGT) in our primary scenario. We also consider the alternative approach. Pensions are often viewed not as a handout from the state, but rather as something earned in the working age. Hence it might make sense to view pensions as deferred income (PDI). Methodologically it means that now we include pensions into the definition of market income, or, since we go from consumption, we do not subtract pensions when going from disposable to the market income. Direct taxes 11 now also do not include social contributions tax, and only the personal income tax is added to the disposable income to get market income. We calculate the consumable income as disposable income plus imputed indirect subsidies minus estimated indirect taxes. Further adding imputed in-kind transfers leads to final income. Detailed principles of estimation and impu- tation of related transfers and taxes are discussed in the next section. Figure 6. Construction of income concepts Final income + in kind transfers – direct transfers (cash and (education and healthcare + indirect subsidies in kind benefits, public expenditures) (utility subsidy and transport subsidy) pensions*) Disposable income Consumable income (total cash expenditures + net in Market income kind income + in kind benefits) – indirect taxes + direct taxes (PIT and (VAT, excises, and import duties) SPF contributions*) Note. * Pensions and SPF contributions are included into direct transfers and taxes respectively only within PGT approach. Source: own elaboration. 3.2. Direct taxes Taxed paid directly by people are personal income tax (PIT), property taxes, and taxes paid by entrepreneurs. The HBS data do not contain explicit information on these taxes. Household expenditures on property taxes are included in the line “taxes and insurances” that feature expenditures on property tax payments, as well as on medical, life, auto insurance, stamp duties, fines, membership fees, and other. There is no feasible way to separate property taxes from other payments. Moreover, the role of these expenditures in the households’ welfare is marginal. On average, “taxes and insurances” constituted only 0.8% of the disposable income of households in 2015. The role of these expenditures increases with income. The richest decile spent 0.9% of their disposable income on “taxes and insur- ances” in 2015, the poorest decile - 0.6%. Hence, it may signal something about the progressivity of property taxes if one assumes that they are distributed the same way as total expenditures on “taxes and insurances”. However, the situation may have changed in 2016, as some privileges on property tax for old-age people were abolished. Still, these taxes play a limited role both in fiscal policy and households’ welfare, and ignoring them would not affect conclusions about overall impact of fiscal policy on poverty and inequality in Belarus. Personal income tax and taxes on entrepreneurial income Personal income tax (PIT) is paid from employment and related income at the flat rate of 13%. The HBS data present information on net income, implying that gross income and PIT payments should be estimated. We assumed that PIT is paid only from employment income. The HBS contains also information on income from sales of agricul- tural products, receipts from personal property and real estate sale, on dividends and rental income. However, they are fully (income from sales of agricultural products) or partly exempt from PIT. In order to simplify estimation, and due to the absence of information needed to make reliable assumptions on which part of income was taxed, we considered that all these lines of income are not subject to PIT. Employment related income presented in HBS files at the household level contains information on self-employ- ment income. This income was subtracted as entrepreneurs enjoy a special tax regime (see below). PIT legislation provides various deductions from the tax base aimed at reducing the tax burden on vulnerable groups – low income deduction, deduction for parents of children below the age of 18, on spouse in maternal/pa- ternal leave, on children above the age of 18 who are continuing education. We took into account deductions on children below 18, equaled to BYR 210,000 per month for one child and 410,000 for two or more children, and the low income deduction of 730,000 for persons with incomes less than 4,420,000 per month. As there is no information on relations between household members, we assigned deductions on children to all employed mem- bers of the household with children. 12 According to our estimates (see Appendix A) the total volume of PIT should have amounted to BYR 36.2 trillion in 2015, which fits actual data. PIT revenues in the consolidated budget in 2015 were equal to BYR 37.0 trillion.9 The ratio of estimated PIT paid by households and gross income from employment of household members is 11.3%, which can be viewed as the effective PIT rate in Belarus. The modelled PIT payments are distributed rather progressively both in absolute and relative terms (see Figure 7a). Lower deciles pay less in personal income taxes than upper deciles if measured as a share of their disposable income before taxation due to deductions provided to households with children and low-paid employees. Another factor is the lower share of employment income in disposable resources of lower deciles. Among the bottom five deciles, only the first decile relies on employment income to the same extent as the population on average (around 50% of disposable income of the first decile is generated by employment income). This stratum is comprised more of households with children and less so of households with pensioners compared to other low-income deciles10, which explains the higher role of employ- ment income in their disposable resources. Consequently, the first decile has a slightly higher PIT tax burden than other low deciles. Taxation of entrepreneurs is not uniform. It depends on the type of entrepreneurial activity, the place of residence, and its scale. In general, entrepreneurs may pay taxes on a general basis, may apply a simplified tax regime for entrepreneurs, or pay a lump-sum tax set by local authorities. Therefore, the modeling of entrepreneurial taxes was based on macroeconomic data. The total amount of taxes paid by entrepreneurs in 2015 amounted to BYR 4,212.8 billion, while total number of entrepreneurs as of the end of 2015 was 240,78111. This means that on average an entrepreneur paid BYR 1,458,000 of taxes monthly. This volume of payments was assigned to every individual who had entrepreneurial income. Information on related income is provided in HBS files together with income from side (day-to-day) jobs as a part of employment income. We assumed that entrepreneurial income was only that exceeding 2 minimal wages in annual terms. This threshold guaranteed that the number of entrepreneurs in HBS files corresponded to their total amount12. Social Protection Fund contributions The Social Protection Fund (SPF) finances public expenditures on pensions and social benefits. Contributions to SPF are payroll taxes at the rate of 35%. Employers pay the main part of the tax (34%), while employees are charged 1% of their gross wage. According to legislation, lower rates (31%, i.e. 30 and 1%) are levied on those employed in the agricultural sector (at enterprises with agricultural production exceeding 50% of total production). As most of rural population in Belarus is employed in the agricultural sector, we assumed that all rural population pays the payroll tax at the rate of 31%. Furthermore, there are upper and lower bounds for the payroll tax base. It should not be lower than the minimum wage and should not exceed 5 average wages in the case of full employment. In our sample there were only 3 observations, where gross wages exceeded 5 average wages. For these persons, SPF contributions were calculated as 35% of the upper bound. In practice, the lower bound is not applied for the employed as the minimum wage regulation applies in Belarus. It is more relevant for entrepreneurs. They are obliged to pay contributions to the SPF at the rate not less than 35% from minimal wage. Entrepreneurs can choose to pay contributions from higher wage base but they are reluctant to do it, as it does not guarantee a feasible increase in future pensions. Hence, we assumed that entrepre- neurs paid contributions to the SPF at the volume of 35% of minimum wage13. 9 In BYR before denomination of 2016 by 10,000 times. For revenues in BYR after denomination; see Table 1. 10 81% of households in the 1st decile ranked by disposable income before personal income taxation are households with children, while the average share is 47.5% (48.5% for the 4th decile). The share of households with at least one member above 60 years old is 37.3% in the first decile and 57% on average in the sample (72.1% in the 4th decile). 11 Ministry on Taxes and Duties of Belarus, www.nalog.gov.by/uploads/folderFor- Links/Ежемесячно%20на%20сайт%20по%20СМП%20и%20ССП%20на%2001.01.2017.xlsx. 12 According to this approach, the number of entrepreneurs in HBS file corresponds to the total number of 218,000. 13 SPF contributions by entrepreneurs are paid once a year in February. It means that in 2015 entrepreneurs paid contributions for 2014. So we estimated payment based on annul average minimum wage of 2014. 13 Figure 7. Incidence of direct taxes by deciles 7 USD PPP % 10 16 USD PPP % 25 9 14 6 8 20 12 5 7 6 10 15 4 5 8 3 4 10 6 2 3 4 2 5 1 2 1 0 0 0 0 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 % of disposable income net of intervention (right axis) % of disposable income net of intervention (right axis) USD PPP per day USD PPP per day (a) personal income tax (b) social security contribution Note. Deciles are ranked by disposable income before personal income tax (a) and social security contribution (b). Source: own estimates based on HBS data. This approach generated relevant estimates of payroll taxes. The total amount of estimated SPF contributions equaled to BYR 94.3 trillion, which is close to the actual amount of contributions in 2015 (BYR 95.3 trillion). Estimated contributions are distributed progressively in absolute and relative terms (see Figure 7b). A fall in the share of SPF contributions in disposable income before SPF taxation in the third decile is related to the significant share of pensioners in this stratum14. 3.3. Indirect taxes VAT VAT generates the largest volume of general government revenues. Most of the goods and services either domes- tically produced or imported are taxed at the rate of 20%. Hence, the share of VAT in consumer prices of these goods and services is 16.7%. A lower rate of 10% (corresponding to the share of 9.1% in consumer prices) is applied to agricultural and most of food products, as well as children goods. Exported goods are taxed at the rate of 0%. Furthermore, some services are exempt from VAT. The list of these services reduced substantially in the last decade, as the government strives to keep the tax base stable despite the economic recession. As of 2015, health care, education, and utilities services were exempted from VAT. VAT was similarly not applied to purchases and rent of real estate by households. The VAT exemption implies that providers of corresponding services have no VAT refunds, so they report VAT on inputs as costs, including them into basic prices. Hence, the effect of VAT taxation on consumer prices for these services depends on the share of intermediates in total production. Related estimates were accomplished within input-output tables after matching household expenditure lines with the industries from national accounts. Results of these estimations are presented in Table 3. According to these estimates, the share of VAT in household consumption is equal to 11.7%, which corresponds to the ratio of general government VAT revenue to national final consumption (11.9%).15 It reinforces the good approximation of obtained estimates of the VAT burden on households to actual VAT payments. Still, due to the discrepancy between consumption data of HBS and national accounts we may underestimate total volume of VAT payments by households. The estimated volume is BYR 37.7 trillion, which constitutes only 51.8% of total general government revenues from VAT. For instance, the share of household final consumption in total final consump- tion is 77.1%. Our estimates show that this burden is evenly distributed among the population if measured as a share of dispos- able income (see Figure 8a). It can be attributed to the fact that the structure of household expenditures does not 14 77.1% of households in the third decile ranked by disposable income before payment of SPF contributions are households with at least one member aged above 60 years. Average share is 57%. 15 The ratio of estimated VAT payments of households to their disposable income is 8.8%. According to administrative data, collected VAT revenue is equal to 8.4% of national disposable income. 14 differ much across the population (with the exception of the 1st and 10th deciles), and most of the goods and services are taxed at the same VAT rate. Table 3. Share of VAT in consumer prices by expenditure line Expenditure line (COICOP) Industry (ISIC) VAT rate Food* Food products 9.1 Alcohol and tobacco* 16.7 Clothing Textiles, and textile products 16.7 Footwear Leather and footwear 16.7 Fabrics Textiles, and textile products 16.7 Housing, fuel for heating dwellings** Electricity, gas and water supply and forestry 7.6 Housing, utilities*** Electricity, gas and water supply and other community, social and 7.3 personal services Housing, other Real estate activities and renting 0.0 Household appliances Computer, electronic and optical equipment, 16.7 Furniture Wood and products of wood and cork 16.7 Health care Health and social work 3.9 Public transportation Transport 16.7 Maintenance of private vehicles Coke, refined petroleum products and nuclear fuel 16.7 Purchase of cars and other vehicles Motor vehicles, trailers and semi-trailers 16.7 Communication services Post and telecommunications 16.7 Culture, recreation and sports Hotels and restaurants 16.7 Secondary and higher education Education 3.0 Preschool education Education 3.0 Eating out and restorans Hotels and restaurants 16.7 Personal care Chemicals and chemical products 16.7 Other goods and services Manufacturing nec 16.7 Food purchased for animals and for cultivation of Agriculture 9.1 land plot Construction and purchase of real estate Construction 8.1 Notes. * VAT rate for alcohol and tobacco is 20%, while for majority of food products it is set at 10%. ** Structure of fuel used for heating in houses with autonomic heating is following: 40% wood fuel, 60% gas16. Based on these weights we estimated VAT rate for expenditures on fuel for heating as weighted average VAT rate for “electricity, gas and water supply” and “forestry”. *** VAT rate is estimated as average for “electricity, gas and water supply” and “other community, social and personal services” weighted on sectors’ total output. Source: own estimates based on the Tax Code of Belarus and Input-Output Tables for 2014. Figure 8. Incidence of indirect taxes by deciles 6 USD PPP % 10 1.0 USD PPP % 6 9 0.9 5 5 8 0.8 7 0.7 4 4 6 0.6 3 5 0.5 3 4 0.4 2 2 3 0.3 2 0.2 1 1 1 0.1 0 0 0.0 0 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 % of disposable income (right axis) % of disposable income (right axis) USD PPP per day USD PPP per day (a) VAT (b) excises Note. Deciles are ranked by disposable income. Source: own estimates based on HBS data. 16 http://www.belstat.gov.by/upload-belstat/upload-belstat-pdf/oficial_statistika/Potreblenie_energii_v_dom_hoz.pdf. 15 Excises Excises are levied on alcohol, tobacco, and fuel for motor vehicles. Excise rates are set to physical units and the government tends to review them regularly due to high inflation rates. Thus, we calculated excise payments of households based on quarterly files of HBS. HBS files contain information on household expenditures on excise goods. However, households tend to underre- port related consumption. The gap between expenditures on alcohol reported by households and its actual con- sumption based on the retail trade turnover is especially high: alcohol expenditures reported in the HBS are only about 25% of retail sales of alcohol. It implies that HBS data needs an adjustment to retail trade statistics in order to receive reliable volume of excises paid by households. We have HBS data on the total amount of alcohol expenditures on the one hand, and sales/average prices/excise rates by types on the other. Based on data on alcohol retail sales, its average prices, and respective excise rates, we estimated shares of excise taxes in retail prices by alcohol type. Next, we made an assumption about the structure of alcohol consumption by different types of households. We assumed that the cheapest alcoholic beverages are con- sumed by the poorer households, and shares of more expensive alcohol are growing with household income, distin- guishing between quintiles of households by their expenditure (see Table 4). Having the retail trade data on sales of different types of alcohol and given assumptions about the structure of alcohol consumption by quintiles, we esti- mated total alcohol expenditures by household quintiles. Dividing imputed alcohol expenditures by HBS alcohol expenditures for every quintile, we got quintile-specific “alcohol expenditure underreporting coefficients” (see Table 5). Assuming that every household in a particular quintile that reported alcohol expenditures has the same “bias”, we imputed alcohol expenditures as EXPi alc  miq , where EXPi alc is initially reported expenditures on alcohol of house- hold i from quintile q, and miq is quintile-specific alcohol expenditure underreporting coefficient from the Table 5. Consequently, the amount of alcohol excises paid by household i was calculated as EXialc  EXPi alc  miq  dEX q , where dEX q is the share of excises in expenditures on alcohol by quintile from the Table 4. Table 4. Assumptions and inputs for estimating household payments of alcohol excises Vodka Liquers Wine Fruit wine Cognac Sparkling Low alcohol Beer wine beverages Share of excise tax in retail prices of alcoholic beverages, % 1st quarter 43.1 32.6 7.6 36.9 20.5 9.6 28.0 17.5 2nd quarter 44.1 32.9 7.3 36.9 19.8 8.4 27.5 16.6 3rd quarter 44.2 29.2 7.8 39.0 18.2 8.8 27.2 16.1 4th quarter 43.9 30.0 7.8 39.0 17.9 8.5 27.2 15.6 Structure of alcoholic beverages consumption by quintiles of population, % 1st quintile 0 0 0 100 0 0 0 0 2nd quintile 30 0 0 0 0 0 0 30 3rd quintile 30 0 25 0 0 50 0 25 4th quintile 25 40 25 0 0 25 100 30 5th quintile 15 60 50 0 100 25 0 15 Total 100 100 100 100 100 100 100 100 Source: own estimates based on Belstat (retail prices for alcoholic beverages) and Ministry of Taxes and Duties (excise tax rates for 2015); own assumptions. Expenditures on tobacco and fuel for motor vehicles reported in HBS files are also below levels in retail sales data. This discrepancy can be attributed to the widespread cross border trade in cigarettes and gasoline/diesel with EU coun- tries. Hence, unlike in the case of alcohol products, there is no need to adjust consumption expenditures of house- holds on tobacco and fuel for motor vehicles to the retail statistics. In order to estimate tobacco excise payments, we calculated the average weighted share of excises in consumer prices of tobacco products. The weights for tobacco products (cigarettes with filter, cigarettes without filter, im- ported cigarettes with filter) were taken proportionally to their share in the consumer price index. According to our estimates, the share of excises in expenditures on tobacco equaled to 31.9% in 2015. Estimates of fuel excises were based on the structure of fuel expenditures reported in HBS files and shares of excises in gasoline and diesel prices (13.4 and 7.6%). 16 Table 5. Estimated parameters of alcohol consumption underreporting and share of excises in house- hold expenditures by quintile 1st quintile 2nd quintile 3rd quintile 4th quintile 5th quintile Scale of underreporting, times 7.9 6.5 5.6 4.5 2.8 Share of excises in expenditures for 38.0 32.4 28.7 27.7 23.3 alcohol, % Source: own estimates based on Belstat (retail prices for alcoholic beverages) and Ministry of Taxes and Duties (excise tax rates for 2015); own assumptions. According to our estimates, the total volume of excises paid by households in 2015 amounted to BYR 11.4 trillion. The total volume of excises collected by the general government was BYR 19.4 trillion. The difference is consti- tuted by fuel excises paid by legal entities, as well purchases of tobacco and fuel by households for further sale abroad rather than for individual consumption. According to our estimates, excises are regressive in relative term (see Figure 8b), as lower deciles of population tend to spend a bigger share of their disposable resources on alcohol and tobacco products. Import duties The impact of import duties on welfare of socially vulnerable groups was partly analyzed in research related to the consequences of Russian WTO accession to Belarus social policy (Shymanovich, 2013). These results showed that the reduction of import tariffs due to Russian WTO accession should have had a minor impact on population welfare, with a uniform distribution across the population. One of the reasons behind this was the small scale of the reduction of tariffs that is scheduled in the Russian WTO accession agreement. In this analysis, we will apply the same methodology as in Shymanovich (2013) in order to see whether conclusions on the neutral influence of import duties on inequality hold within a full-fledged abolishment of import tariffs. The volume of import duties paid by households was estimate based on data on household consumption, the share of import products in retail sales and import tariffs according to the following formula: n ti D m   Ci  Sim  , i 1 1  ti where Ci is the consumption of the product i from HBS data, n is equal to 37 product groups (see table), S im is the share of imports in the consumption of the product i, and ti is a level of import tariff for the product i.17 Table 6 presents data required for these estimations. Import tariffs for expenditure lines were taken for corre- sponding product groups within the HS classification and weighted by import volumes. The share of import prod- ucts in household consumption was supposed to be equal to the share of related goods in retail sales. Table 6. Assumptions and inputs for estimating import duties Weighted Corresponding Import share Import price Expenditure line (COICOP) average import HS code in retail** elacticity*** tariffs Expenditures for bread 1905 13.2 6.1 -0.8 Expenditures for pastry 1905 12.5 6.1 -0.8 Expenditures for flour 11 11.2 4.0 -1.2 Expenditures for cereals and beans 10 5.5 43.3 -0.9 Expenditures for macaroni food 1902 14.0 35.4 -0.7 Expenditures for milk 4 15.8 5.7 -1.4 Expenditures for sour cream and cream 4 15.8 5.7 -1.1 Expenditures for butter 405 18.2 1.3 -0.8 Expenditures for cheese 406 18.5 12.6 -0.6 Expenditures for other dairy products 4 15.8 5.7 -1.1 Expenditures for beef and veal 201 23.8 0.1 -1.0 Expenditures for pork 203 32.5 0.3 -2.9 Expenditures for sausages and smoked meat 16 14.8 0.6 -1.1 Expenditures for poultry 207 52.5 0.8 -2.9 17 This formula allows estimating first order welfare effect of import duties abolishment on welfare of population, if one assumes that reduction of duties results in proportional reduction of prices. For estimating second order effect one should take into account import price elasticities εi, n t t estimated in Shymanovich (2013) based on data from Kee, Nicita, Olarreaga (2009). Related formula is: W   Ci  i  Si  (1  i   i ). m i 1 1  ti 1  ti 17 Weighted Corresponding Import share Import price Expenditure line (COICOP) average import HS code in retail** elacticity*** tariffs Expenditures for fat 209 15.0 0.5 -1.8 Expenditures for other meats 2 34.2 0.7 -1.4 Expenditures for fish and seafood 3 8.0 42.1 -1.3 Expenditures for vegetable oil, margarine and other grease 15 14.3 74.2 -1.0 Expenditures for eggs 407 0.2 0.2 -1.0 Expenditures for potatoes 701 13.8 6.8 -1.2 Expenditures for vegetables and melons 7 13.8 32.1 -0.8 Expenditures for fruits and berries 8 3.4 89.8 -1.0 Expenditures for sugar and confectionery 17 6.1 22.0 -1.0 Expenditures for tea, coffee, cocoa 9 12.9 76.6 -1.0 Expenditures for non-alcoholic drinks 2201, 2202* 10.9 17.2 -0.9 Expenditures for other food food average 15.9 26.3 -1.2 Expenditures for alcohol and tobacco 22, 24* 7.7 11.7 -1.5 Expenditures for clothing 61, 62* 10.1 41.5 -1.9 Expenditures for footwear 64 2.3 52.4 -1.0 Expenditures for fabrics 59, 60* 5.8 30.9 -1.0 Expenditures for household appliances 85 5.3 74.7 -1.2 Expenditures for furniture 94 12.4 10.4 -0.9 Expenditures for health care 30 7.7 64.6 -0.9 Expenditures for maintenance of private vehicles 2710, 8708* 1.2 18.5 -1.4 25% for 19.2% Expenditures for purchase of cars and other vehicles of import 4.8 98.9 -1.2 Expenditures for personal care 34 10.6 77.7 -0.8 Expenditures for for food purchased for animals and for cultivation of land plot 23 4.2 47.9 -1.0 Notes. * Average import tariffs were weighted by import volume (from world in 2015). Data of tariff rates is taken from TRAINS database (as of 2014, partner – world). ** Data on share of import goods in retail sales is obtained from Belstat yearbook on retail trade 18 and monthly bulletins on retail trade. *** Elasticities are from Shymanovich G. (2013). Absent elasticities were assumed as follo wing: for “bread”, “other dairy products”, and “poultry” equaled to “pastry”, “sour cream and cream”, and “pork” respectively; for “eggs” equaled to -1; for “other food” as average of elasticities for food products; for “alcohol and tobacco” as average ela sticity of alcohol and tobacco weighted by import; for “maintenance of private vehicles” as average elasticity for HS2710 and HS8708 weighted by import volume. Sources: TRAINS database, Shymanovich G. (2013), Belstat. Figure 9. Incidence of import duties by deciles 0.8 USD PPP % 1.4 0.7 1.2 0.6 1.0 0.5 0.8 0.4 0.6 0.3 0.4 0.2 0.1 0.2 0.0 0.0 1 2 3 4 5 6 7 8 9 10 % of disposable income (right axis) USD PPP per day Note. Deciles are ranked by disposable income. Source: own estimates based on HBS data. 18 http://www.belstat.gov.by/ofitsialnaya-statistika/realny-sector-ekonomiki/vnytrennia-torgovlya/roznichnaya-torgovlya/pub- likatsii_6/index_702/. 18 According to the estimates, the total volume of import duties paid by households amounted to BYR 5.0 trillion in 2015. The Ministry of Finance reported BYR 8.6 trillion revenue from import duties. These numbers look reason- able as import of investment goods is largely exempted from import duties, while intermediate goods are mainly imported within the customs union with Russia and other CIS countries. According to our estimates import duties are progressive in absolute terms, but neutral in relative terms (see Figure 9). The payment of import duties ac- counts for similar shares of disposable income in lower and upper deciles. The only exception is the 10th decile, in which people tend to save more and spend more on real estate and services. 3.4. Direct transfers Data on the majority of direct transfers received by households is available in HBS files. They feature revenues from the following benefits and privileges:  Benefits o pregnancy registration and child birth benefit o maternity benefit o children allowances for children aged below 3, o children allowances for children aged above 3, o attendance allowance o funeral benefit o pension for death of a breadwinner o pension for disabled children o unemployment benefit o severance pay o student grants o social assistance and other  Privileges o Food privilege o Passenger transportation privilege o Hosing and utilities privileges o Fuel privileges o Electricity privileges o Communication service privileges o Health resort privileges o Privileges for pharmaceuticals o Privileges for social rehabilitation appliances o Preschool education privileges o Other privileges The total amount of benefits received by households corresponds to public expenditures of BYR 17.6 trillion. It matches the official figure of BYR 17.5 trillion that SPF spent on financing social benefits in 2015. The volume of privileges, according to HBS data, equaled to public expenditures of BYR 3.2 trillion. The distribution of cash and in-kind direct transfers differs significantly. Benefits are progressive in absolute and relative terms (see Figure 10a). The first decile ranked by disposable income net of transfer receives on average almost 2 times more benefits in absolute terms than the second decile. Benefits increase the disposable income of the first decile by 47.6%, while the average effect over the whole sample is 8.6%. Privileges are less targeted. The absolute volume of privileges received by lower and upper deciles is almost the same (see Figure 10b). Still, the disposable income of lower deciles is more vulnerable to the revision of privileges than the income of upper deciles. Belarus has a PAYG pension system, which implies that pensions have the nature of transfer rather than deferred income. Pensions reported by households are equal to total public expenditures of 96.2 trillion. Related expendi- tures of SPF were equal to 83.4 trillion. In addition, non-contributory pensions (provided to retired civil servants and persons retired from national security, defense and law enforcement agencies) were financed from the central government budget at the level of BYR 6.8 trillion. Still, HBS data overshoots actual public expenditures on pen- sions by 7.3%. This overshooting is related to the structure of the survey sample, which overestimates the actual number of population above working age by 9.5%. 19 Figure 10. Incidence of cash and in-kind benefits by deciles 3.5 USD PPP % 50 0.30 USD PPP % 3.0 45 3.0 0.25 2.5 40 2.5 35 0.20 2.0 2.0 30 25 0.15 1.5 1.5 20 0.10 1.0 1.0 15 10 0.5 0.05 0.5 5 0.0 0 0.00 0.0 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 % of disposable income net of intervention (right axis) % of disposable income net of intervention (right axis) USD PPP per day USD PPP per day (a) benefits (b) privileges 20 USD PPP % 250 18 16 200 14 12 150 10 8 100 6 4 50 2 0 0 1 2 3 4 5 6 7 8 9 10 % of disposable income net of intervention (right axis) USD PPP per day (c) pensions Note. The ratio of pensions to the disposable income net of pensions for the first decile is not calculated as related net income for this decile is close to 0. Source: own estimates based on HBS data. 3.5. Indirect subsidies Subsidies for utilities and urban public transportation are two major types of indirect subsidies in Belarus. Utilities subsidies amount to 2% of GDP, with around 1% covered by the cross-subsidization by the enterprises, and the other 1% coming directly from the budget (IMF, 2016). As of 2015, subsidies on utilities were available to everyone automatically in the form of subsidized tariffs in the utility bill (since September 2016 direct subsidies to the utilities are also available). Some households and apartments, however, are not eligible for the subsidy. These are the households where the household head owns more than one apartment or house, or if no one is registered in the apartment (usually the case when the apartment is rented out). According to IMF estimates, households with access to the subsidy covered 48.5% of the actual costs. Expenditures on utilities are reported in the HBS, but the households do not report if they get the subsidy or not. To identify the households without access to subsidized tariffs we establish a cut-off in utilities cost per square meter. If the household is paying above the cut-off of BYR 15,000 per m2 (two times higher than the average), or above BYR 1,000,000 in total per month (three time higher than the average), we assume that the household does not obtain the subsidy. The rest of the households are assigned a 51.2% subsidy on top of their actual utilities expenditure. To check if we have allocated the utilities subsidy correctly, we gross up the allocated utilities, and find that they sum up to 1.96% of GDP, which coincides with the IMF estimate of 2% of GDP. 20 Figure 11. Indirect subsidies by disposable income deciles 1.8 daily USD Share of 7% 1.4 daily USD Share of 2.2% disposable PPP disposable PPP 1.6 income income 6% 1.2 2.1% 1.4 5% 1.0 1.2 2.0% 1.0 4% 0.8 1.9% 0.8 3% 0.6 0.6 1.8% 2% 0.4 0.4 1% 0.2 1.7% 0.2 0.0 0% 0.0 1.6% 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 utility subsidies in $ PPP % of income transport subsidies in $ PPP % of income (a) Incidence of indirect utility subsidy, (b) Incidence of indirect transport subsidy, by disposable income decile, USD PPP and % by disposable income decile, USD PPP and % Source: own estimates As we see from Figure 11a, utility subsidies are regressive in absolute value: the top decile obtains twice the amount the bottom decile obtains through the subsidy. This result is not unexpected: the subsidies are equally available to the rich and the poor, but since higher income usually implies more spacious housing, higher income households face higher utility costs and receive a higher subsidy. However, in relative value the utility subsidy is progressive: lower deciles obtain a higher proportion of their income in the form of subsidy. Many types of public transportation are indirectly subsidized in Belarus by budget support to the state-owned transport companies. However, we focus on the urban public transportation as the major source of transport subsidies, and also the only one on which the data on cost coverage is available. As with utility subsidies, the transport subsidies are built-in into the tariffs, but unlike utility subsidies, transport subsidies are available to every user without exemptions. Transport expenditures are reported in HBS as a total for all kinds of expenditure, including but not limited to the urban public transport. To impute urban public transport expenditure, we follow the next steps: 1. We assume that all household transport expenditure below BYR 100,000 (a sum close to the average cost of round trip ticket between the regional centers) are expenditure on urban public transportation. A monthly pass cost above BYR 200,000 in 2015 in Minsk, hence this cut-off is not too high. 2. We build a truncated regression model for transport expenditure below the threshold of BYR 100,000. The explanatory variables are the number of working-age and retirement-age adults in the household, region, residence type (large cities or small cities) and car ownership. Income level was excluded as it turned out to be insignificant. 3. Using the estimated model, we imputed urban public transport expenditure for the rest of the households. If the imputed level was higher than the actual reported expenditure on transportation, we replaced it with the reported value. 4. We applied the subsidy of 62%19 to the imputed urban transportation costs. According to our imputation, transport subsidies amount to almost 1% of GDP. Unfortunately, we do not have the aggregate data on the value of transport subsidies, so we cannot check the validity of our imputation of transport subsidies by grossing up. Hence, all the results concerning those subsidies should be interpreted with caution. Transport subsidies are regressive in both absolute and relative value: the top decile obtains twice the amount the bottom decile obtains through the subsidy (see Figure 11b). The result is mainly driven by the fact that only the urban population (with higher incomes) has access to the transport subsidy. Moreover, employed working-age individuals are more likely to use public transportation and enjoy the subsidy, and they also happen to be the ones with higher incomes. 19 See https://news.tut.by/society/505674.html. 21 3.6. In-kind transfers (health care and education) Health care in Belarus is dominated by the government, which remains the main provider of health services. The public health system is a Soviet-style centralized Semashko system, with an extensive network of state-owned poly- clinics and hospitals providing comprehensive health care. Health care in the public system is free to every citizen of Belarus, independent of income, employment or any other socioeconomic characteristics. No contributions are necessary to gain access to health care. According to the official data (Ministry of Finance, 2016), total health expenditure in the government budget amounted to BYR 34,977 billion in 2015 (4.0% of GDP). Around 40% of the total public health expenditure is spent on primary and secondary care through polyclinics, and the rest covers tertiary care through hospitals (World Bank, 2013). While health care expenditure is universal, and it would be tempting to distribute health expenditure (and benefits) equally, we prefer to assign health expenditure to the actual beneficiaries. In this case we would attribute expendi- ture to actual receivers of services. This approach would allow us to capture differences in needs, for example, by gender and age. Moreover, despite the universal system, de-facto access to health services is very different for rural and urban residents. HBS stopped reporting doctor’s visits and hospital stays since 2008. We use the 2008 data set to model the number of doctor’s visits and the probability of a hospital stay, and then use the model to predict those variables in the 2015 data. We use the Poisson model for the number of doctor visits, and a probit to model the probability of a hospital stay. In both cases the explanatory variables are age, age squared, gender, a child dummy, smoker status, self-reported health evaluation, region, residence type, body mass index, and level of education. Figure 12. Health and education expenditure by disposable income deciles 1.8 Share of 14% 3.5 USD PPP Share of 35% USD PPP disposable disposable 1.6 income 12% income 3.0 30% 1.4 10% 2.5 25% 1.2 1.0 8% 2.0 20% 0.8 6% 1.5 15% 0.6 4% 1.0 10% 0.4 2% 0.5 5% 0.2 0.0 0% 0.0 0% 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Primary and secondary healthcare Terciary Healthcare Preschool, primary school Secondary school Higher school (a) Incidence of health expenditure, (b) Incidence of education expenditure, by disposable income decile, USD PPP (bars) and % (lines) by disposable income decile, USD PPP (bars) and % (lines) Source: own estimates. After using the estimated models to predict the number of doctor’s visits and the probability of a hospital stay in 2015, we allocate primary and tertiary care expenditure on health care proportionately to them. In absolute values health expenditures are allocated rather flatly across different disposable income deciles, although there is a slight upward slope. Ceteris paribus one might expect a negative relationship between health expenditure and income (with poorer people usually having lower health). However, in Belarus the lower income deciles are largely represented by rural non-retiree households, which have lower access to health care. Hence, the expenditure schedule across deciles looks flatter than expected. In relative terms, however, health expenditures are clearly progressive, reflecting the free universal access to health care (see Figure 12a). Public education expenditure amounted to 4.8% of GDP in 2015: 1.1% of GDP was spent on pre-school and primary school education; 2.24% of GDP on general secondary (school) education; 0.56% on continued secondary education (vocational and specialized non-college education) and 0.90% of GDP on higher (college) education (Ministry of Finance, 2016). 22 Public school education is free, although households pay for textbooks and food out of pocket; private schools and colleges exist, but they are few and negligible in their coverage. At the tertiary level (vocational, specialized and college education) fees are widespread, and access to free education is conditional on performance. Usually the fees are below the total education cost. As the primary and secondary school enrollment rates are 100% or higher in Belarus (World Bank, 2013), we allocate the preschool and primary and general secondary education expenditures to all children of ages 3-10 and 11-16 correspondingly. At 16 children graduate from the obligatory school, and are free to either continue in high school to enter college, go to the labor market or enroll in the vocational or specialized educational institutions. Since 2015 HBS data lacks information on the socioeconomic status, and we do not observe whether a young person is a student or not. To impute the probability of being a student, we use 2014 data which still had the socioeconomic status variable. We build a probit probability model for people aged 16-40, with age, gender, region, residence type, household size and education level being the explanatory variables. After imputing the probability of being a student, we assign expenditure on continued secondary and college education according to age. We do not scale down health and education expenditure. The main reasoning behind scaling down is that normally all the taxes and transfers in the CEQ analysis are not forced to be equal to their counterparts in the national accounts (see Higgins, Lustig, 2016). However, in our exercise the allocation of most of the taxes and transfers are quite close to their counterparts in the national accounts. Major education expenditure categories, primary and secondary education, are highly progressive both in relative and absolute terms, as seen in Figure 12b. Households with children, the main recipients of educational expendi- ture, are on usually poorer (in per capita terms) than households without children; poorer households also tend to have more children. The expenditure on college (higher school) education, in contrast, is regressive in absolute terms, as individuals from higher-income households are more likely to enroll in colleges. 4. Results and discussion 4.1. Main results The results of the CEQ analysis show that fiscal interventions contribute much to the reduction of inequality in Belarus if the pension system is modeled as a part of fiscal policy. Most of the effect on inequality comes from direct transfers and privileges in the case when pensions are treated as a government transfer (PGT) (see Figure 13). The decrease in inequality indicators calculated for market and disposable income, when pensions are viewed as a government transfer, is massive, from 0.407 to 0.267 for the Gini index. Particularly large improvements are observed for the ratio of the average income of the richest 10% to the poorest 10% - from 14.82 to 3.25 (see Table 7). A significant number of people without market income relying on direct transfers, on pensions in particular, explains this result20. When pensions are alternatively modelled as deferred income (PDI), the reduction of ine- quality related to fiscal interventions determining the difference between market and disposable income is only marginal. Furthermore, moving from disposable to consumable income does not influence the overall level of inequality, implying that the burden of indirect taxes and gains from indirect transfers is distributed among popu- lation proportionally to income. In-kind transfers are clearly more progressive, as inequality indicators fall substan- tially from consumable to final income. Fiscal interventions also determine the level of absolute poverty. According to the international poverty lines of 2.5 and 5 USD PPP per day the risk of poverty is eliminated in Belarus at the level of disposable income. Moreover, the risk of absolute poverty at these lines is also negligible according to the market income concept when pensions are treated as deferred income (PDI). Only when excluding pensions from market income (PGT approach), pov- erty lines of 2.5 and 5 USD PPP per day can reveal some vulnerable population, as some households rely heavily on pensions. The risk of poverty is much higher if one considers the line of 10 USD PPP per day (which is often used for defining the middle class in international studies) or the national poverty line. In fact, the subsistence minimum in Belarus – national absolute poverty line – exceeds 10 USD PPP line, implying relatively high overall level of income in the country if measured in USD PPP terms. Hence, we will use national absolute poverty line as a main benchmark for analyzing the influence of fiscal policy on poverty in Belarus. For a more detailed analysis of vulnerable groups of population one can also apply the minimum consumer budget as a national moderate poverty line. 20 3.5% of the population lives in households without market income (modeled by the PGT approach). 23 Figure 13. Lorenz curves for basic income concepts 1.0 market income (PGT) market income (PDI) 0.9 final income consumable income 0.8 disposable income 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0.0 0.2 0.4 0.6 0.8 1.0 Source: own estimates based on CEQ methodology. According to the national poverty line, 20.2% of the population would be poor if there were only market income without any fiscal interventions including the pension system21. In the case when pensions are treated as deferred income, the level of poverty is much lower (5.9%), stressing the significant role of the pension system in poverty reduction. On the one hand, it stems from the size of pensions that exceed the national poverty line. On the other hand, it stresses that pensions are a sole or dominant income for many households in Belarus. Fiscal redistribution related to other direct transfers and taxes offsets the risk of poverty down to 3.4%22. Moreover, the poverty gap, based on the disposable income concept, is also rather low, highlighting the absence of extreme poverty within the disposable income concept. On the contrary, the system of indirect taxes and subsidies increases the risk of poverty up to 5.2%, not affecting much its depth (the poverty gap remains low). Table 7. Main poverty and inequality indicators by income concepts Market Income Disposable Consumable Final PGT PDI Income Income Income Gini 0.407 0.292 0.267 0.270 0.227 Theil Index 0.259 0.151 0.129 0.131 0.094 90/10 14.82 3.70 3.25 3.32 2.61 USD 2.5 PPP Headcount Index 8.3 0.2 0.0 0.0 0.0 Poverty Gap 5.9 0.1 0.0 0.0 0.0 USD 5 PPP Headcount Index 11.6 0.7 0.0 0.1 0.0 Poverty Gap 8.0 0.2 0.0 0.0 0.0 USD 10 PPP Headcount Index 19.0 5.0 2.5 3.9 0.7 Poverty Gap 11.6 1.3 0.4 0.7 0.1 National poverty line (USD Headcount Index 20.2 5.9 3.4 5.2 0.8 10.62 PPP) Poverty Gap 12.1 1.6 0.5 0.9 0.1 National moderate poverty Headcount Index 32.7 21.6 19.8 24.8 7.5 line (USD 16.69 PPP) Poverty Gap 17.2 5.7 4.2 5.6 1.3 Note. Minimum consumer budget represents national moderate poverty line. Source: own estimates based on CEQ methodology. The scale of redistribution caused by fiscal interventions is higher than the change in inequality indicators may suggest. The decomposition of the change in the Gini indicator reveals that vertical equity generated by fiscal policy is accompanied by significant horizontal effects (see Table 8). In particular, the reduction in Gini related to the influence of direct taxes and transfers, including pensions, would be 0.20 pp if there were no re-ranking effects between market (PGT) and disposable income. However, changes in the relative welfare of the population across these income concepts is natural as it is attributed to the pension system, which stands for a major part of direct 21 Application of national absolute poverty line to other income concepts than disposable income is for illustration purposes only, as it is constructed based on actual retail, i.e. post-fisc prices. 22 Poverty estimates based on disposable income and national poverty line should correspond to the official share of low-income people in Belarus. In practice there is significant difference, as official estimates are done based on quarterly data. 24 taxes and transfers. Modeling the pension system as a deferred income results in minor horizontal equity effects of direct taxes and transfers. Less desirable is horizontal equity (reranking) caused by indirect taxation and subsidies, as well as in-kind transfers. The reranking effect of consumable income is higher than of disposable income, while vertical equity remains unchanged, irrespectively of the market income concept applied. Hence, indirect taxes and transfers do not reduce inequality, but lead to households switching places in the distribution by income. On the contrary, in-kind transfers are associated with a significant vertical equity effect. In the case when market income is modeled according to the assumption of pensions as deferred income, in-kind transfers generate most of the vertical equity effect of the fiscal policy. However, they also cause some horizontal equity effects if households are initially ranked by market income that includes pensions (PDI). On the contrary, the comparison of the household distribution by final income and market income modeled according to PGT approach reveals that in-kind transfers reduce the scale of reranking. This contradiction is rooted in the nature of modeled in-kind transfers. Namely, education transfers are inevitably lower for households with elderly members. Therefore, they make pensioners less wealthy relative to youth. When we compare the final income distribution to the market income distribution prior to pension system effects (PGT), we see that in-kind transfers limit the reranking effect associated with the pension system. Vice versa, the initial ranking of households by market income including pensions (PDI) results in deteriorated relative welfare of pensioners and increasing horizontal equity effects after accounting for in-kind transfers. Table 8. Decompositions of inequality changes into vertical and horizontal equity components Change to market income (PGT) Change to market income (PDI) Disposable Consumable Final Disposable Consumable Final income income income income income income Gini change with respect to market income 0.139 0.137 0.180 0.025 0.022 0.065 Vertical equity (Reynolds-Smolensky Index) 0.200 0.202 0.236 0.031 0.030 0.087 Reranking (Atkinson-Plotnick Index of 0.060 0.065 0.056 0.006 0.008 0.022 horizontal equity) Note. Vertical equity implies reduction of gap in welfare between rich and poor due to fiscal intervention. Horizontal equity implies that fiscal intervention does not influence ranking position of an individual, see Kakwani (1984). Source: own estimates based on CEQ methodology. The incidence analysis of income changes caused by fiscal interventions shows that benefits are concentrated within the lowest deciles. The bottom two deciles, ranked by market income that does not include pensions (PGT approach), enjoy a substantial average increase of income, which is partly related to the low base of market income (see Figure 14). The positive effect disappears by the 4–5 deciles, while 7–10 deciles suffer income reduction (see Table 9). If market income is modeled within the PDI approach, the positive effect of fiscal interventions holds only for the first deciles (with the exception of in-kind transfers) and its scale is much lower. Consequently, the scale of losses by upper deciles is also lower. Moreover, losses are distributed evenly among relatively wealthy deciles, implying a similar tax burden for upper deciles. The incidence of net effects changes substantially after accounting for in-kind transfers. At the level of final income, the effect from fiscal interventions steadily diminishes from lower deciles to the upper deciles (both compared to the market income modeled by PDI and PGT ap- proach), implying progressivity of in-kind transfers. Moreover, the effect at the level of final income becomes negative only for the wealthiest deciles. Table 9. Incidence of net effects from fiscal interventions in relation to market income by deciles Market income by PGT approach Market income by PDI approach Disposable Consumable Final Disposable Consumable Final Income Income Income Income Income Income 1 890.6 805.1 1214.1 18.2 8.0 56.5 2 111.4 95.2 169.7 4.9 -3.1 33.8 3 37.5 27.7 68.1 1.3 -6.0 23.6 4 13.8 5.9 36.2 -1.6 -8.4 18.0 5 1.2 -4.9 18.6 -2.4 -8.2 14.0 6 -7.1 -13.1 6.8 -3.6 -9.8 10.7 7 -12.7 -18.2 -0.8 -4.8 -10.8 8.4 8 -14.9 -20.1 -6.1 -5.5 -11.4 4.1 9 -17.6 -23.7 -11.4 -6.7 -13.7 -0.1 10 -19.5 -23.9 -17.3 -6.9 -12.1 -4.0 Total -3.2 -9.5 9.3 -3.5 -9.8 9.0 Source: own estimates based on CEQ methodology. 25 Differences in the incidence of effects generated by moving from market income to disposable, consumable and final income are related not only to the progressivity of fiscal interventions, but also to the scale of redistribution, caused by them and the modeling of separate income concepts. The large size of the pension system compared to other social expenses determines the higher scale of effects if post-fiscal income is compared to the base of market income by the PGT approach, rather than by the PDI approach. A more positive effect at the level of final income compared to consumable income is rooted in the modeling, as total modeled final income is 9.3% higher than total market income by the PGT approach (9.0% by the PDI approach, see Table 9), while consumable income is 9.5% lower (9.8% lower by PDI approach). Figure 14. Distribution of individuals by market and disposable income 150 150 disposable income, USD PPP 100 100 50 50 0 0 0 50 100 150 0 50 100 150 market income (PGT), USD PPP market income (PDI), USD PPP Source: own estimates based on CEQ methodology. Overall, the progressivity of fiscal interventions and the average positive effect observed for lower deciles, does not necessary imply that everybody from low income groups automatically benefits from fiscal interventions (see dots below 45-degree line). Differences in the incidence of effects generated by moving from market income to disposable, consumable and final income are related not only to the progressivity of fiscal interventions, but also to the scale of redistribution, caused by them and the modeling of separate income concepts. The large size of the pension system compared to other social expenses determines the higher scale of effects if post-fiscal income is compared to the base of market income by the PGT approach, rather than by the PDI approach. A more positive effect at the level of final income compared to consumable income is rooted in the modeling, as total modeled final income is 9.3% higher than total market income by the PGT approach (9.0% by the PDI approach, see Table 9), while consumable income is 9.5% lower (9.8% lower by PDI approach). Figure 14). On the one hand, distribution plots of net recipients and donors overlap to a rather small extent if they are ranked by market income modeled according to the PGT approach (see Figure 15a). Around 45% of all net beneficiaries within the system of direct taxes and transfers (that includes the pension system) have market incomes below 10 USD PPP (which is a proxy for the national poverty line). Moreover, 97.5% of population with market incomes below 10 USD PPP benefit from this system. The donors, in turn, are largely people with market incomes above 20 USDD PPP, which may be interpreted as a good targeting of the system. On the other hand, ranking people according to disposable (post-fiscal) income changes the situation dramatically. Distribution plots of net recipients and donors within the system of direct taxes and transfers largely coincide (see Figure 15b). As the result, 60% of population with disposable incomes below 10 USD PPP per day faces a reduction of income due to the system of direct taxes and transfers that includes the pension system. Hence, direct transfers and taxes, and the pension system in particular, substantially reduce poverty and inequality, but generate significant reranking effect as well. If the pension system is excluded from the analysis, the number of net beneficiaries becomes much lower. Their distribution by income bins coincides to a significant degree with the distribution of net donors even if people are ranked by market income (see Figure 15c). The share of people benefiting from the system of direct transfers and taxes among the population with pre-fisc market income (PDI approach) below 10 USD PPP is 82.3%. However, they constitute only 14.6% of all net beneficiaries from the system of direct taxes and transfers. Furthermore, 26 15.8% of population with market income (PDI approach) below 10 USD PPP suffer a fall in their welfare due to direct taxes. The tax system limits the size of their losses, providing deductions for low income households (see PIT tax). The tax burden, if measured by net losses from the system of direct taxes and benefits (excluding pension system) is the same for all income bins starting from 10 USD PPP per day. Hence, the system of direct taxes and transfers (excluding the pension system) plays an important role in mitigating poverty (especially extreme poverty, as net benefits from direct transfers constitute 64% of disposable income of households with market income below 5 USD PPP), but at the same time a significant part of fiscal support is targeted to relatively wealthy households. The distribution of net benefits and losses by income bins related to the difference between consumable income and disposable income has a similar profile. It implies that indirect taxes and subsidies do not have significant redistributive effects. The influence of fiscal interventions determining the difference between market and final income is more pronounced (see Appendix C). It increases the number of net beneficiaries within the relatively low-income population and shifts the distribution of net donors towards high-income bins. This is due to account- ing for in-kind transfers that make the total final income of the population significantly higher than market income. However, there are still net fiscal donors among the population with disposable incomes below 10 USD PPP. The tax burden at the level of final income is more progressive than at the level of disposable income, but the difference between high- and low-income bins is not large (especially if the bins are formed based on disposable income). Figure 15. Distribution of gains and losses at the level of disposable income with respect to market income by income bins set in USD PPP Gains and losses compared to the level of market income (PGT approach) distributed by bins, a) ranked by market income b) ranked by disposable income 500 100 300 90 thsd. pers. % thsd. pers. % 250 75 400 80 200 60 300 60 150 45 200 40 100 30 100 20 50 15 0 0 0 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 recipients, thsd recipients, thsd donors, thsd donors, thsd average reciept, % of disposable income (right axis) average reciept, % of disposable income (right axis) average donation, % of disposable income (right axis) average donation, % of disposable income (right axis) c) Gains and losses compared to the level of market income by PDI approach distributed by bins, ranked by market income 500 100 thsd. pers. % 400 80 300 60 200 40 100 20 0 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 recipients, thsd donors, thsd average reciept, % of disposable income (right axis) average donation, % of disposable income (right axis) Source: own estimates based on CEQ methodology. 27 Reshuffling of the population by income ranking caused by the fiscal intervention leads to impoverishments for some households. According to the disposable income concept, 1.9% of population became poor or fall into deeper poverty due to the system of direct taxes and transfers (irrespective of the approach to pension system modelling, see Table 10). It means that more than half of the poverty (56.3%) is caused by fiscal redistribution. At the level of consumable income, the scale of fiscal impoverishment is even higher: 3.3% of population turned poor due to direct and indirect taxes and subsidies, including the pension system, which constituted 67% of the total poverty headcount (if fiscal interventions do not include the pension system, the scale of impoverishment is even larger). Nevertheless, fiscal gains to the poor are much higher than fiscal impoverishment. Around 20% of the population, poor at the level of market income by the PGT approach (i.e. 97.3% of pre-fiscal poverty headcount) enjoy an increase in their welfare after receiving direct transfers and pensions, and paying direct taxes. Broader concepts of fiscal interventions create benefits for similar shares of pre-fiscal poor population. The scale of these benefits is around 60% of the national absolute poverty line, which is also higher than the impoverishment effect. If the pension system is not treated as a fiscal intervention, the scale of poverty reduction generated by direct transfers, as well as indirect subsidies, is much lower (around 4% of population). Nevertheless, around three-quarters of pre- fisc poor population benefits from fiscal interventions, and the size of the net benefit is around 25% of the poverty line. The fact that fiscal gains to the poor surpass fiscal impoverishment in absolute terms is also illustrated by poverty gap dynamics, which is actually the difference between fiscal gains to the poor per capita and fiscal impoverishment per capita. Estimates show, that the poverty gap reduces from 12.1% at the level of market income by PGT ap- proach to less than 1% at the level of disposable and consumable income. If the pension system is excluded from the analysis, the scale of the poverty gap reduction is much lower, as initially depth of poverty is rather low. Table 10. Fiscal gains to the poor and fiscal impoverishment in relation to market income Pensions as government transfer Pensions as deferred income National extreme National moderate National extreme National moderate poverty line poverty line poverty line poverty line Disposable Consumable Disposable Consumable Disposable Consumable Disposable Consumable Income Income Income Income Income Income Income Income Fiscal impoverishment headcount, % of population 1.89 3.25 10.52 15.39 1.89 3.56 11.00 16.96 headcount, % of post-fiscal poor 56.30 62.96 53.03 62.01 56.12 68.98 55.47 68.33 fiscal impoverishment per capita, % of market income 0.24 0.45 1.46 2.32 0.12 0.36 0.63 1.64 fiscal impoverishment per capita among fiscally impoverished, % of poverty line 12.49 13.73 13.85 15.06 6.59 10.00 5.76 9.66 Fiscal gains to the poor headcount, % of population 19.67 19.29 28.38 26.89 4.44 4.13 12.22 10.86 headcount, % of pre-fiscal poor 97.30 95.42 86.88 82.30 75.26 70.04 56.60 50.33 fiscal gains to the poor per capita, % of market income 11.77 11.65 14.51 13.92 1.14 1.04 2.14 1.69 fiscal gains to the poor per capita among fiscal gainers, % of poverty line 59.87 60.39 51.14 51.77 25.71 25.07 17.49 15.53 Poverty gap reduction, pp (fiscal gains minus fiscal impoverishment per capita) 11.54 11.20 13.06 11.60 1.02 0.68 1.50 0.05 Note. Fiscal impoverishment is a considered a situation when i) somebody who is non-poor according to the pre-fiscal income (market income) appears to be poor according to the post-fiscal income (disposable or consumable), ii) somebody poor according to the pre-fiscal income suffers further income reduction due to fiscal interventions. Fiscal gains to the poor take place when somebody poor according to the pre-fiscal income enjoys income increase due to fiscal interventions. Source: own estimates based on CEQ methodology and Higgins, Lustig (2016). According to the socioeconomic status, the pension age population enjoys most of the gains from fiscal interven- tions if the pension system is treated as a fiscal intervention, while the employed population of working age bear most of the costs (see Figure 16). If the pension system is excluded from the analysis, beneficiaries of the system of direct taxes and transfers are children, the non-employed population of working and above working age, while the employed population (both working and above working age) are net payers within the system. Indirect taxes and subsidies contribute to further increases of welfare of the unemployed population of working age and of 28 children, while the positive effect for non-employed people of working age reduces. In addition, in-kind transfers result in increasing redistribution within the employed population of working age. Figure 16. Structure of net beneficiaries and payers of fiscal system by socio-economic status payers payers income income Final Final beneficiaries beneficiaries Disposable Consumable Disposable Consumable payers payers income income beneficiaries beneficiaries payers payers income income beneficiaries beneficiaries Population Population 0% 50% 100% 0% 50% 100% children children working age, employed working age, employed working age, partly employed working age, partly employed working age, non-employed working age, non-employed above working age, employed above working age, employed above working age, non-employed above working age, non-employed Effect with relation to the market income according to the PGT Effect with relation to the market income according to the PDI approach approach Note. HBS does not contain information on social economic status of respondents. Employed where considered those receiving income employment for more than 6 months. Those receiving employment related income for less than 6 months and those reported income from self-employment where considered as partly employed. Those who did not report any employment related income were considered ono- employed. This group comprises unemployed and economically inactive, including housewives, students, disabled persons. Source: own estimates based on CEQ methodology. Despite generally benefiting from fiscal policy, children still have a higher risk of fiscal impoverishment than the population on average (see Table 11). An especially high risk is observed for large families with three or more children. The existing tax burden appears to be too high for households with several dependents and it is not fully mitigated by the system of child allowances. Other social vulnerable groups also face a high risk of fiscal impover- ishment. In particular, fiscal policy may affect the welfare of partly employed or self-employed people. Being low paid and bearing tax obligations related to labor market participation determine the high risk of poverty for this group. The same applies to those living in rural areas, where employment opportunities are limited to low paid jobs in agriculture. Table 11. Fiscal impoverishment in relation to market income by social vulnerable groups, headcount, % of group Pensions as government transfer Pensions as deferred income disposable income consumable income disposable income consumable income children 3.96 6.41 3.40 5.85 working age partly employed 3.98 6.56 4.28 7.37 working age non-employed 2.71 3.74 2.88 4.58 rural area household member 3.52 5.89 2.70 6.53 lone parent household member 3.54 5.37 2.81 4.59 large family (3+ children) member 10.68 12.89 6.43 10.65 average 1.89 3.25 1.89 3.56 Source: own estimates based on CEQ methodology. 4.2. Distributional impact and marginal contributions of fiscal interventions Concentration curves of fiscal interventions help understand whether a particular intervention is equalizing or not. If the curve of the intervention is above the 45-degree line (i.e. concentration coefficient is negative), it is progressive and equalizing. However, the intervention can be equalizing even if it is regressive and below the 45- degree line. To be equalizing, the concentration curve of a transfer only needs to be above the selected income 29 concept (the concentration coefficient of a transfer exceeds the Gini coefficient for that income). Taxes are equal- izing when their concentration curves are below the Lorenz curve of related income concept (the concentration coefficient of a tax is lower than the Gini coefficient). To preserve space, the concentration curves are presented in Appendix B, while concertation coefficients, as well as Kakwani index, representing differences between the Gini coefficient of income and the concentration coefficient of an intervention, are presented in Table 12. Table 12. Progressivity of taxes and transfers in relation to income concepts Market income (PGT) Market income (PDI) Disposable income concentration Kakwani concentration Kakwani concentration Kakwani coefficient index coefficient index coefficient index All direct transfers incl contributory pensions -0.285 0.691 -0.265 0.557 0.098 0.169 benefits -0.121 0.527 -0.304 0.596 -0.091 0.358 privileges -0.194 0.601 -0.051 0.343 0.065 0.202 pensions -0.350 0.757 -- -- 0.169 0.098 All direct taxes 0.425 0.018 0.337 0.045 0.292 0.024 Personal income tax 0.434 0.027 0.337 0.045 0.296 0.029 Social Contributions 0.422 0.015 -- -- 0.290 0.022 All indirect subsidies 0.127 0.280 0.188 0.104 0.186 0.081 Utilities subsidy 0.059 0.348 0.133 0.159 0.135 0.133 Public transport subsidy 0.268 0.139 0.300 -0.008 0.293 -0.026 All indirect taxes 0.186 -0.221 0.220 -0.072 0.224 -0.044 VAT 0.225 -0.182 0.260 -0.032 0.265 -0.002 Tobacco excise 0.145 -0.262 0.073 -0.219 0.058 -0.209 Alcohol excise 0.010 -0.397 0.049 -0.244 0.048 -0.219 Fuel excise 0.349 -0.058 0.326 0.033 0.327 0.060 Import tariffs 0.186 -0.221 0.240 -0.052 0.249 -0.018 All gross in-kind transfers 0.006 0.401 -0.070 0.362 -0.075 0.342 Gross health transfers -0.052 0.459 0.025 0.267 0.038 0.230 Primary and secondary health expenditure -0.055 0.462 0.034 0.258 0.049 0.219 Tertiary (hospital) health expenditure -0.050 0.456 0.019 0.273 0.030 0.237 Gross education transfers 0.055 0.352 -0.150 0.442 -0.170 0.437 Preschool and primary school education 0.018 0.389 -0.238 0.530 -0.249 0.516 Secondary education -0.032 0.439 -0.248 0.540 -0.268 0.535 Continued secondary education 0.139 0.268 -0.035 0.327 -0.064 0.332 Higher (college) education 0.253 0.154 0.113 0.179 0.089 0.178 All taxes 0.355 -0.052 0.267 -0.025 0.272 0.004 All net transfers and subsidies excl contributory pensions -0.143 0.550 -0.079 0.371 0.049 0.218 Gini 0.407 0.292 0.267 Note. Concentration coefficient of a tax exceeding Gini coefficient of related income implies positive Kakwani index and progressivity of tax. Concentration coefficient of a transfer exceeding Gini coefficient of related income implies negative Kakwani index and regressivity of a transfer. Bold characters reflect income concept influenced by the intervention. Source: own estimates based on CEQ methodology. Direct taxes levied on market income are minimally equalizing. The concentration curves of both the personal income tax and the social contributions tax are slightly below the income schedules (the Kakwaini index is positive, but close to zero). It is not surprising, given the flat schedule of taxes with some exemptions for low-income individuals. Indirect taxes, paid from disposable income, in turn, are disequalizing. This is particularly true for the alcohol and tobacco excises, which put most of the burden on poor households. However, since the purpose of these taxes is not to be equalizers, and, ideally, not even to deliver revenues to the budget, but to penalize unhealthy behaviors, we cannot judge these excises on their redistribution properties. Import tariffs and VAT taxes have virtually no redistributive effects with respect to disposable income. A higher share of consumption within lower deciles results in a higher burden of VAT and import duties for them, but lower VAT rates for food products23 mitigate this effect. The fuel excise, however, is equalizing as the fuel excise applies mainly to car owners, such that the burden falls mainly on the upper deciles. The concentration curves for direct transfers show that most of transfer interventions are equalizing and progres- sive when ranked by market income (see Appendix B, Figure 20). Pensions play a major role in redistribution from market to disposable income. Benefits and privileges interventions are also equalizing. The scale of their progres- sivity depends on the analyzed income concept. While privileges are more progressive if households are ranked by market income before the pension system intervention, benefits play a more important redistributive role when 23 Difference in expenditure structure of the 1st and 10th deciles is mainly related to food purchase and savings (see Shymanovich, 2013). 30 pensions are considered as market income (see Table 12). Moreover, benefits are progressive with respect to dis- posable income. It implies that benefits, contrary to privileges, target groups of population that are not covered by the pension system. Furthermore, benefits do not fully employ their equalizing potential. On the contrary, pensions are not progressive and only slightly equalizing with respect to disposable income. This reinforces their significant role in redistribution and their notable contribution to the disposable income of population. Indirect subsidies received at the level of disposable income are regressive, but equalizing. However, it is achieved only on account of utilities subsidy, while the transport subsidy is disequalizing with respect to disposable income. The transport subsidy has an equalizing effect only in the absence of the pension system (see Appendix B, Figure 21). Health expenditures are distributed quite equally, and we can see at the Figure 22 (Appendix B), that health trans- fers are equalizing with respect to disposable income and all the curves lie very close to the 45-degree line. Conse- quently, concentration coefficients are close to zero. Education transfers are also equalizing. Furthermore, all types of education transfers except for the college education are progressive. Primary and basic secondary school have the most pronounced equalizing effect. The concentration curves and related indices allow us to see the direction of the redistribution impact of fiscal interventions, but they do not allow us to estimate the size of this impact, or to see the impact on poverty. Marginal contributions in Table 13 show the change to inequality and local poverty measures after the application of the intervention. Table 13. Marginal contributions to inequality and poverty Gini National Poverty Moderate Poverty Benefits 2.0 5.1 3.9 Privileges 0.2 0.7 0.3 Pensions 11.1 23.5 19.2 Personal income tax 0.4 4.1 0.9 Social Contributions 1.1 8.9 2.1 VAT 0.0 5.2 1.2 Excises on tobacco -0.1 0.4 0.0 Excises on alcohol -0.4 2.1 0.3 Excises on fuel 0.0 0.1 0.0 Import duties 0.0 0.9 0.1 Utilities subsidy 0.6 3.6 1.2 Transport subsidy -0.1 1.2 0.3 Primary health care 0.9 3.9 1.5 Tertiary health care 1.4 5.7 2.5 Primary education 1.4 3.2 2.8 Secondary education 3.6 5.8 6.6 Cont. secondary education 0.6 1.6 1.2 Higher education 0.7 2.2 1.3 Note. Changes to inequality and poverty for the direct benefits and taxes are measured in comparison to the disposable income prior to the related intervention (i.e. market income (PGT) plus direct taxes and transfers net of analysed intervention). For indirect subsidies and in-kind transfers changes to inequality and poverty are measured in comparison to the disposable income. Negative values for Gini mean increase of inequality due to fiscal intervention. Positive values for poverty mean increase of poverty headcount in case of abolishment of direct transfers, need to cover in full costs of utilities and public transportation, as well as pay for education and health care services, and increase of poverty headcount caused by presence of taxes. Source: own estimates. Pensions are the most effective fiscal intervention, lowering the Gini index by 11 points, and extreme poverty by over 23 percentage points. Benefits also have positive effects, but they are much smaller. The most effective of all benefits is the childcare benefit (for children aged 0-3), contributing 1.3 points to the Gini decrease and 3 points to poverty decrease. Excises on tobacco and alcohol increase inequality, although only modestly. Same is true of the transport subsidy. Utilities subsidy decreases inequality and poverty. Primary and secondary education and tertiary health care have sizable equalizing effects and gains for the poor. 4.3. Efficiency The marginal contributions of fiscal programs described in the previous section are especially useful when evalu- ating social programs, fiscal interventions designed primarily to combat poverty and inequality. In the case of Belarus, marginal contributions to poverty and inequality are the major indicators for the effectiveness of benefits, privileges, pensions and indirect subsidies. But the marginal contributions miss another dimension of the fiscal 31 interventions – their cost. While pensions have the biggest impact on poverty and inequality, they are also the costliest program of all. Table 14 lists several efficiency measures which reflect both the impact on poverty/inequality, and the cost of the intervention. The size column lists the size of the intervention as a percentage of GDP according to our allocation in micro data, not according to aggregate data. Efficiency measures are derived as the ratio of marginal contributions to the relative size of intervention . Hence, efficiency measures represent the effect (on the reduction of the Gini index or poverty) of 1% of GDP spent on the particular intervention. The composite measure reflects the effect of 1% of GDP spent on the composite impact measure, consisting of poverty and inequality effects both weighted by 0.5. Table 14. Efficiency measures for social fiscal interventions Size (relative Inequality Poverty Composite Share spent on Share spent on to GDP) Efficiency Efficiency Efficiency top 5 deciles top 2 deciles Benefits: 1.957% 1.04 2.02 1.53 42% 12% pregnancy registration 0.110% 0.82 1.28 1.05 48% 15% pregnancy and childbirth 0.069% 0.72 1.21 0.97 45% 17% child 0-3 1.039% 1.21 2.28 1.74 34% 11% child 3+ 0.171% 1.28 2.00 1.64 31% 7% Attendance 0.104% 1.06 1.19 1.12 40% 8% Funeral 0.095% -0.53 0.10 -0.21 88% 24% child support after breadwinners death 0.126% 1.03 2.26 1.65 42% 9% children-disabled 0.048% 0.83 1.30 1.07 45% 8% assistance and other 0.048% 0.63 1.67 1.15 52% 5% unemployment benefit 0.004% 0.00 5.50 2.75 16% 8% severance pay 0.076% -0.53 0.00 -0.26 88% 18% student grants 0.069% 0.43 0.24 0.33 60% 23% Privileges 0.360% 0.53 0.96 0.74 56% 15% Pensions 10.708% 1.02 1.79 1.41 62% 11% Utilities subsidy 1.956% 0.31 0.61 0.46 59% 26% Transport subsidy 0.945% -0.02 0.36 0.17 70% 37% Note. Deciles are identified based on market income. Source: own estimates. Pensions are not as efficient as total benefits (1.41 versus 1.53 composite efficiency), but nevertheless remain among the efficiency champions even after accounting for the cost. Among benefits, unemployment benefits are the most efficient program. Severance pay, funeral benefit and student stipends are among the most inefficient benefits – a direct result of the absence of means-testing. Unemployment benefits are notoriously low in Belarus (around 10 USD per month). Currently the government is considering increasing benefits at least to the subsistence minimum. Our results suggest that these plans will have significant (and efficient) impacts in reducing poverty and inequality. Most of the child-related benefits are also efficient. Interestingly, the childcare benefit which is paid for mothers of children below 3 years (child 0-3) is more efficient than the pregnancy and childbirth benefit (it is paid from the end of the pregnancy and after the childbirth, for the total of 126 days). One possible explanation is that the childcare benefit is paid in full only if one of the parents stays at home to take care of the child. Parents with higher wages are less likely to take all three years of maternity leave. On the other hand, privileges and indirect subsidies are highly inefficient. The result is not unexpected for the indirect subsidies: they are not targeted, available to everybody and usually regressive. Privileges, on the other hand, are targeted: households and individuals have to meet certain criteria to get access to privileges. The low efficiency of privileges suggests the low quality of targeting or the misuse of the privileges programs. As an additional measure of fiscal efficiency, we also compute the proportion of the program expenditure going to the upper income deciles (by market income): to the top 5 deciles (everyone above the median); and top 2 deciles. More than half of the pensions go to the individuals above the market income median, reflecting the fact that the pensions program is not means-tested. According to this measure privileges and indirect subsidies are again highly inefficient. If all benefits, privileges and indirect subsidies were not available to the top 2 deciles of market income, savings would amount to 1.4% of GDP. 32 The CEQ effectiveness measures provide a similar picture24. The highest effectiveness is assigned to pensions and direct benefits, while indirect subsidies are not effective. 4.4. Targeting and vulnerable groups To understand how the expenditures and taxes are focusing (or not) on vulnerable groups, we compare the size of the transfer/tax allocated to vulnerable groups to the average transfer/tax size for the rest of the population. The results are in Table 15. The first part (a) of the Table looks at vulnerable groups by type of individual, namely by age and employment status. HBS does not have the employment or socio-economic status variables. We assign the status of the employed to individuals who report some wage income and worked during more than 6 months over the year. Those who worked 6 months or less, and had wage or self-employment income, are marked as partially employed. The rest are classified as non-employed, which includes the traditional definition of unem- ployed, inactive (out of labor force), retired and those on childcare leave. Adults (18 and older) are divided in three age groups: youth (18-24), working age (25-55 for women and 25-60 for men) and elderly (above retirement age; over 60 for men and 55 for women). Table 15. Sizes of transfers/taxes by vulnerable groups, relative to the rest of the population (a) By individual type (age/employment) Youth Working age Elderly Partly em- Non-em- non-em- Partly em- Non-em- Employed Employed ployed ployed ployed ployed ployed Benefits (without pensions) 69 127 139 225 40 61 52 Pensions 31 36 36 90 253 261 488 Direct taxes 143 126 109 69 131 74 26 Indirect subsidies 114 116 108 91 158 133 105 Indirect taxes 108 107 96 90 131 115 89 Health 82 98 93 116 106 109 141 Education 93 139 213 81 23 29 20 (b) By household type HH with Rural HH HH of elderly HH with children Large HH lone parent Benefits (without pensions) 74 55 306 189 328 Pensions 102 414 15 13 9 Direct taxes 63 34 99 78 47 Indirect subsidies 48 154 60 104 40 Indirect taxes 78 101 79 88 57 Health 104 150 74 91 77 Education 75 7 298 297 261 Note. The number of 69% in panel a) line “Benefits”, column “Youth employed” means that the young employed receive on average only 69% of benefits the rest of the population (not young and employed) receive. Description of household types and socio-economic economic status see in Table 11, Figure 16. Source: own estimates. Despite the modest size of unemployment benefits, working age non-employed enjoy higher support through benefits programs than the population on average. Childcare benefits, which normally go to households with non- employed mothers can explain this trend. In turn, elderly people and employed youth do not enjoy benefit support. For the elderly, this lack of social support is compensated by pensions. Employed (and partially employed) carry most of the burden of direct and indirect taxes, but they also enjoy more indirect subsidies. Unsurprisingly, young people receive more in-kind transfers from education, while the elderly receive higher than average support from health care. In the panel (b) of Table 15 vulnerable groups are classified by household type. The major vulnerable types are rural households, households of elderly people (were the only adults are above the retirement age), households with children, especially households with a lone parent (household with at least one child and only one adult) and multi-child households (with three or more children). Households with children are the major target group of benefits, with the childcare benefits being the most sub- stantial program. Pensions are naturally skewed towards the elderly households. Most of the vulnerable groups pay 24 However, the CEQ package function ceqef is bugged and the CEQ effectiveness measures are hence incomplete and are not provided in this paper. 33 fewer taxes than average. Health and education are again following their corresponding demographic patterns, with health expenditure more substantial for the elderly, and education – for households with children. While the vulnerable groups in general get more than average out of fiscal transfers and pay fewer direct taxes, there is space for improvement. Currently the benefits are focused on families with children, and pensions – on the elderly, while other vulnerable groups, like people in the rural areas and youth, remain untargeted by social programs. Indirect subsidies are particularly badly targeted, often offering less than average support to the vulner- able groups. 4.5. Cross-country differences Figure 17 shows the results of CEQ analysis for several countries including Belarus. Countries are ranked by the Gini reduction effect with pensions modeled as government transfers. All the effects are going from market to disposable income. On average fiscal interventions decrease the Gini index by 0.036, and poverty by 6.05 percent- age points. Belarus is doing better than average on both counts. However, it is lagging behind the EU28 group of countries in redistribution. Figure 17. Cross-country comparisons of redistribution and poverty reduction effect of direct transfers and taxes 0.25 45 Gini reduction Poverty reduction, % 40 0.20 35 30 0.15 25 0.10 20 15 0.05 10 5 0.00 0 Dominican Republic… Bolivia (2009) Peru (2009) Jordan (2010) Ecuador (2011) Georgia (2013) South Africa (2010) Chilie (2013) Average Argentina (2012) Honduras (2011) Mexico (2010) Brazil (2009) Uruguay (2009) United States (2011) Sri Lanka (2010) Colombia (2010) Guatemala (2011) Belarus (2015) EU-28 (2011) Indonesia (2012) El Salvador (2011) Costa Rica (2010) Russia (2010) Gini PGT Gini PDI Poverty ($4 PPP) Source: Lustig (2016), own estimation based on CEQ methodology The fiscal impact in Belarus is similar to that of Russia in terms of the Gini index reduction. As in Russia, in Belarus the PGT scenario delivers more equalization than the PDI scenario (Lopez-Calva et al., 2017). Poverty reduction in Belarus is lower. However, it does not mean that the fiscal interventions reduce poverty less in Belarus than in Russia, it is merely the reflection of the fact that the 4 USD PPP poverty in Belarus is zero in disposable income, so the reduction in poverty is at its maximum possible level. The Belarusian taxation has a very different redistribution impact when compared to the EU. In most of the EU countries, the PIT tax is equalizing due to its progressive nature. The flat PIT structure in Belarus, however, delivers very little redistribution, while the redistribution task is completely delegated to the expenditure side of the fiscal policy. Given the high level of possible tax evasion in Belarus, this design of the fiscal policy is optimal. The VAT tax, in contrast, is not regressive in Belarus, unlike in many EU countries. This is achieved through multiple VAT exemptions. 5. Conclusions This paper presents the assessment of fiscal incidence in Belarus using the Commitment to Equity (CEQ) meth- odology developed in Lustig & Higgins (2016). Using the household budget survey and aggregate data we have allocated fiscal interventions across households. The allocation allows us to measure the effect of fiscal policies on redistribution and poverty. Fiscal policies in Belarus effectively redistribute income from the top to the bottom of income distribution - 97.5% of population with market income below 10 USD PPP benefit from these policies. The direct transfers (including pensions) and direct taxes lower the national poverty measure by 17 percentage points. They also decrease the Gini 34 index from 0.407 to 0.267. The impressive magnitude of positive fiscal effects puts Belarus among the equalization leaders in the group of developing countries. However, most of the effect is attributed to the pension system. If we treat pensions as a deferred market income rather than government transfer, direct taxes and transfers result in a minor Gini index reduction of 0.025 pp. Indirect taxes and subsidies do not contribute to inequality reduction. Furthermore, fiscal interventions may be a cause of poverty, as 1.9% of population becomes poor due to direct taxes and transfers. The fiscal impoverishment headcount goes up to 3.3% when we account for indirect taxes and subsidies as well. Vulnerable groups like large and lone parent households, people living in rural areas and those not being fully employed have an especially high risk of becoming poor due to fiscal interventions. These groups are in general benefitting from social policy, but high impoverishment rates suggest that related social security measures need improvement and better targeting. Direct transfers, in particular pensions, are the most equalizing and pro-poor of the fiscal interventions. Pensions, for example, are often assigned to households with zero market income, effectively pulling them out of poverty. Taxes in Belarus are not equalizing. Direct taxes are neutral in their influence over inequality. Indirect taxes are regressive. Indirect transfers and taxes increase poverty by 1.8 percentage points, mostly due to the regressive nature of the indirect taxes and poor targeting of subsidies. Our results also point towards possible reforms. Unemployment benefits (currently at very low level) are the most cost-efficient benefits program, suggesting that the plans to increase benefits will have a significant impact on reducing poverty and inequality. Pensions and child-care benefits are also cost-efficient. Indirect subsidies are highly cost-inefficient: 1% of GDP spent on utility subsidies delivers 3 times less reduction in poverty and inequality compared to the same 1% spent on pensions. The indirect subsidies to utility and transport tariffs are not targeted, but they are available to everybody and regressive. They are also offering less- than-average support to vulnerable groups. Replacing indirect subsidies with a well-targeted benefits program will reduce poverty and inequality more efficiently. Restricting access to benefits (like student grants and childcare benefits) and subsidies for households at the top of the income distribution also might be a possibility worth exploring: even the most conservative estimate suggests possible savings of 1% of GDP from the better targeting of transfers and subsidies. References Chubrik, A. (2007). GDP Growth and Income Dynamics: Who Reaps the Benefits of Economic Growth in Belarus? In Haiduk, K., Pelipas, I., Chubrik, A. (Eds.) Growth for All? Economy of Belarus: The Challenges Ahead; IPM research Center. Chubrik, A. (2016a). Uslovija obuslovlennogo finansirovanija: MVF vs. EFSR [Conditions of Conditional financ- ing: the IMF vs. the EFSD], IPM Research Center, Commentary DP/16/05. Chubrik, A. (2016b). Effects of Economic Downturn on Regional Development in Belarus, World Bank back- ground paper, mimeo. Chubrik, A., Shymanovich, G. (2016). Impact assessment of the social economic policy changes on the status of the socially vulnerable groups in Belarus, UNDP report, mimeo. Chubrik, A., Shymanovich, G. (2016). Increasing utility tariffs: social aspects, IPM Research Center, Commentary 16/01 (in Russian) Favaro, E., Smits, K., Bakanova, M. (2012). Structural challenges for SOEs in Belarus: A case study of the machine building sector, World Bank Policy Research working paper WPS 6010, Washington D.C. Higgins, S. (2017). Producing Indicators and Results, and Completing Sections D and E of CEQ Master Workbook using the CEQ Stata Package. Chapter 7 Lustig, N. (Ed.) Commitment to Equity Handbook. A Guide to Estimating the Impact of Fiscal Policy on Inequality and Poverty, Tulane University and the World Bank, Fall 2016. Higgins, S., Lustig, N. (2016). Allocating Taxes and Transfers and Constructing Income Concepts. Chapter 5 in Lustig, N. (Ed.) Commitment to Equity Handbook. A Guide to Estimating the Impact of Fiscal Policy on Inequality and Poverty, Tulane University and the World Bank, Fall 2016. IMF (2016). Staff Report For The 2016 Article IV Consultation, IMF Country Report 16/298. Kakwani, N. (1984). On the measurement of tax progressivity and redistributive effect of taxes with applications to horizontal and vertical equity, Advances in Econometrics, 3, 149-168. 35 Kee, H., Nicita, A., Olarreaga, M. (2009). Estimating Trade Restrictiveness Indices, Economic Journal, 119, 172-199. Lisenkova, K., Bornukova, K. (2017). Effects of Population Ageing on the Pension System in Belarus, Baltic Journal of Economics, forthcoming. Lustig, N. (2016) “Fiscal Redistribution in Low and Middle Income Countries.” Chapter 8 i n N. Lustig (editor) Commitment to Equity Handbook. A Guide to Estimating the Impact of Fiscal Policy on Inequality and Poverty , Tulane University and the World Bank, Fall 2016. Lopez-Calva, L. F., Lustig, N., Matytsin, M., Popova, D. (2017). Who Benefits from Fiscal Redistribution in Rus- sia?, in The Distributional Impact of Fiscal Policy: Experience from Developing Countries, edited by Gabriela Inchauste and Nora Lustig (Washington: World Bank, forthcoming). Lustig, N., Higgins, S. (2016). The CEQ Assessment: Measuring the Impact of Fiscal Policy on Inequality and Poverty. Chapter 8 in Lustig, N. (Ed.) Commitment to Equity Handbook. A Guide to Estimating the Impact of Fiscal Policy on Inequality and Poverty, Tulane University and the World Bank, Fall 2016. Mazol, A. (2016). Spatial Wage Inequality in Belarus. BEROC Working Paper Series, WP no. 35 Ministry of Finance of the Republic of Belarus (2016). On the status of government finance in January-December 2015, accessed through http://minfin.gov.by/upload/bp/doklad/2015/yd2015.pdf as of 20th of December 2016 Shymanovich G. (2013). Social Effects of Accession of Belarus to the WTO, IPM Research Center, Policy Discussion Paper PDP/13/08. Shymanovich, G. (2016). Impact of socio-economic policy changes on the position of elderly people in Belarus, IPM Research Center, Commentary 16/04 (in Russian). World Bank (2012). Belarus country economic memorandum: economic transformation for growth. Washington, DC: World Bank. World Bank (2013). Belarus – Public expenditure review : enhancing public services in times of austerity. Public Expenditure Review (PER). Washington DC; World Bank Group. Zhang, F., and Hankinson, D. (2015). Belarus Heat Tariff Reform and Social Impact Mitigation, World Bank Pub- lications, The World Bank, number 22574. 36 Appendix A. Estimation of personal income tax Based on legislation regulating PIT taxation we calculated gross income from employment by formula: GW   NW  rPIT  ( DCH  DLI  /(1  rPIT  rSSF ) Where NW – net income from employment, rPIT – PIT rate of 13%, rSPF – rate of contribution to the Social Protection Fund done by employee equaled to 1%, DCH – deduction on children aged below 18, 0, if there is no children aged below 18 in the household  DCH  210, if there is 1 child aged below 18 in the household 410, if there are 2 and more children aged below 18 in the household  DLI – deduction for persons with low income, DLI = 730 if NW  (4420  1  rPIT  rSSF   rPIT  (730  DCH )) i.e. reported personal income is below upper threshold for net income for a person who received deduction on low income. Then size of PIT paid monthly by individual is equal to: PIT  (GW  DCH  DLI )  rPIT () 37 B. Concentration curves Figure 18. Concentration curves for direct taxes (a) Concentration curves of direct taxes (b) Concentration curves of direct taxes with respect to market income with respect to disposable income Source: own estimates with use of CEQ Stata package. Figure 19. Concentration curves for indirect taxes (a) Concentration curves of indirect taxes (b) Concentration curves of indirect taxes with respect to market income with respect to disposable income Source: own estimates with use of CEQ Stata package. 38 Figure 20. Concentration curves for direct transfers (a) Concentration curves of direct transfers (b) Concentration curves of direct transfers with respect to market income with respect to disposable income Source: own estimates with use of CEQ Stata package. Figure 21. Concentration curves for indirect subsidies (a) Concentration curves of indirect subsidies (b) Concentration curves of indirect subsidies with respect to market income with respect to disposable income Source: own estimates with use of CEQ Stata package. 39 Figure 22. Concentration curves for in-kind expenditure (a) Concentration curves of health expenditure (b) Concentration curves of education expenditure with respect to disposable income with respect to disposable income Source: own estimates with use of CEQ Stata package (Higgins, 2017). 40 C. Distribution of gains and losses at the level of final income Figure 23. Distribution of gains and losses at the level of final income with respect to market income by income bins set in USD PPP Gains and losses compared to the level of market income by PGT approach distributed by bins, ranked by market income ranked by disposable income 600 120 350 120 % thsd. pers. % thsd. pers. 500 100 300 100 250 400 80 80 200 300 60 60 150 200 40 40 100 100 20 20 50 0 0 0 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 recipients, thsd recipients, thsd donors, thsd donors, thsd average reciept, % of disposable income average reciept, % of disposable income average donation, % of disposable income average donation, % of disposable income Gains and losses compared to the level of market income by PDI approach distributed by bins, ranked by market income 600 100 thsd. pers. % 500 80 400 60 300 40 200 100 20 0 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 recipients, thsd donors, thsd average reciept, % of disposable income average donation, % of disposable income 41 Poverty & Equity Global Practice Working Papers (Since July 2014) The Poverty & Equity Global Practice Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. This series is co‐published with the World Bank Policy Research Working Papers (DECOS). It is part of a larger effort by the World Bank to provide open access to its research and contribute to development policy discussions around the world. For the latest paper, visit our GP’s intranet at http://POVERTY. 1 Estimating poverty in the absence of consumption data: the case of Liberia Dabalen, A. L., Graham, E., Himelein, K., Mungai, R., September 2014 2 Female labor participation in the Arab world: some evidence from panel data in Morocco Barry, A. G., Guennouni, J., Verme, P., September 2015 3 Should income inequality be reduced and who should benefit? redistributive preferences in Europe and Central Asia Cojocaru, A., Diagne, M. F., November 2014 4 Rent imputation for welfare measurement: a review of methodologies and empirical findings Balcazar Salazar, C. F., Ceriani, L., Olivieri, S., Ranzani, M., November 2014 5 Can agricultural households farm their way out of poverty? Oseni, G., McGee, K., Dabalen, A., November 2014 6 Durable goods and poverty measurement Amendola, N., Vecchi, G., November 2014 7 Inequality stagnation in Latin America in the aftermath of the global financial crisis Cord, L., Barriga Cabanillas, O., Lucchetti, L., Rodriguez‐Castelan, C., Sousa, L. D., Valderrama, D. December 2014 8 Born with a silver spoon: inequality in educational achievement across the world Balcazar Salazar, C. F., Narayan, A., Tiwari, S., January 2015 Updated on October 23, 2017 by POV GP KL Team | 1 9 Long‐run effects of democracy on income inequality: evidence from repeated cross‐sections Balcazar Salazar,C. F., January 2015 10 Living on the edge: vulnerability to poverty and public transfers in Mexico Ortiz‐Juarez, E., Rodriguez‐Castelan, C., De La Fuente, A., January 2015 11 Moldova: a story of upward economic mobility Davalos, M. E., Meyer, M., January 2015 12 Broken gears: the value added of higher education on teachers' academic achievement Balcazar Salazar, C. F., Nopo, H., January 2015 13 Can we measure resilience? a proposed method and evidence from countries in the Sahel Alfani, F., Dabalen, A. L., Fisker, P., Molini, V., January 2015 14 Vulnerability to malnutrition in the West African Sahel Alfani, F., Dabalen, A. L., Fisker, P., Molini, V., January 2015 15 Economic mobility in Europe and Central Asia: exploring patterns and uncovering puzzles Cancho, C., Davalos, M. E., Demarchi, G., Meyer, M., Sanchez Paramo, C., January 2015 16 Managing risk with insurance and savings: experimental evidence for male and female farm managers in the Sahel Delavallade, C., Dizon, F., Hill, R., Petraud, J. P., el., January 2015 17 Gone with the storm: rainfall shocks and household well‐being in Guatemala Baez, J. E., Lucchetti, L., Genoni, M. E., Salazar, M., January 2015 18 Handling the weather: insurance, savings, and credit in West Africa De Nicola, F., February 2015 19 The distributional impact of fiscal policy in South Africa Inchauste Comboni, M. G., Lustig, N., Maboshe, M., Purfield, C., Woolard, I., March 2015 20 Interviewer effects in subjective survey questions: evidence from Timor‐Leste Himelein, K., March 2015 21 No condition is permanent: middle class in Nigeria in the last decade Corral Rodas, P. A., Molini, V., Oseni, G. O., March 2015 22 An evaluation of the 2014 subsidy reforms in Morocco and a simulation of further reforms Verme, P., El Massnaoui, K., March 2015 Updated on October 23, 2017 by POV GP KL Team | 2 23 The quest for subsidy reforms in Libya Araar, A., Choueiri, N., Verme, P., March 2015 24 The (non‐) effect of violence on education: evidence from the "war on drugs" in Mexico Márquez‐Padilla, F., Pérez‐Arce, F., Rodriguez Castelan, C., April 2015 25 “Missing girls” in the south Caucasus countries: trends, possible causes, and policy options Das Gupta, M., April 2015 26 Measuring inequality from top to bottom Diaz Bazan, T. V., April 2015 27 Are we confusing poverty with preferences? Van Den Boom, B., Halsema, A., Molini, V., April 2015 28 Socioeconomic impact of the crisis in north Mali on displaced people (Available in French) Etang Ndip, A., Hoogeveen, J. G., Lendorfer, J., June 2015 29 Data deprivation: another deprivation to end Serajuddin, U., Uematsu, H., Wieser, C., Yoshida, N., Dabalen, A., April 2015 30 The local socioeconomic effects of gold mining: evidence from Ghana Chuhan-Pole, P., Dabalen, A., Kotsadam, A., Sanoh, A., Tolonen, A.K., April 2015 31 Inequality of outcomes and inequality of opportunity in Tanzania Belghith, N. B. H., Zeufack, A. G., May 2015 32 How unfair is the inequality of wage earnings in Russia? estimates from panel data Tiwari, S., Lara Ibarra, G., Narayan, A., June 2015 33 Fertility transition in Turkey—who is most at risk of deciding against child arrival? Greulich, A., Dasre, A., Inan, C., June 2015 34 The socioeconomic impacts of energy reform in Tunisia: a simulation approach Cuesta Leiva, J. A., El Lahga, A., Lara Ibarra, G., June 2015 35 Energy subsidies reform in Jordan: welfare implications of different scenarios Atamanov, A., Jellema, J. R., Serajuddin, U., June 2015 36 How costly are labor gender gaps? estimates for the Balkans and Turkey Cuberes, D., Teignier, M., June 2015 37 Subjective well‐being across the lifespan in Europe and Central Asia Bauer, J. M., Munoz Boudet, A. M., Levin, V., Nie, P., Sousa‐Poza, A., July 2015 Updated on October 23, 2017 by POV GP KL Team | 3 38 Lower bounds on inequality of opportunity and measurement error Balcazar Salazar, C. F., July 2015 39 A decade of declining earnings inequality in the Russian Federation Posadas, J., Calvo, P. A., Lopez‐Calva, L.‐F., August 2015 40 Gender gap in pay in the Russian Federation: twenty years later, still a concern Atencio, A., Posadas, J., August 2015 41 Job opportunities along the rural‐urban gradation and female labor force participation in India Chatterjee, U., Rama, M. G., Murgai, R., September 2015 42 Multidimensional poverty in Ethiopia: changes in overlapping deprivations Yigezu, B., Ambel, A. A., Mehta, P. A., September 2015 43 Are public libraries improving quality of education? when the provision of public goods is not enough Rodriguez Lesmes, P. A., Valderrama Gonzalez, D., Trujillo, J. D., September 2015 44 Understanding poverty reduction in Sri Lanka: evidence from 2002 to 2012/13 Inchauste Comboni, M. G., Ceriani, L., Olivieri, S. D., October 2015 45 A global count of the extreme poor in 2012: data issues, methodology and initial results Ferreira, F.H.G., Chen, S., Dabalen, A. L., Dikhanov, Y. M., Hamadeh, N., Jolliffe, D. M., Narayan, A., Prydz, E. B., Revenga, A. L., Sangraula, P., Serajuddin, U., Yoshida, N., October 2015 46 Exploring the sources of downward bias in measuring inequality of opportunity Lara Ibarra, G., Martinez Cruz, A. L., October 2015 47 Women’s police stations and domestic violence: evidence from Brazil Perova, E., Reynolds, S., November 2015 48 From demographic dividend to demographic burden? regional trends of population aging in Russia Matytsin, M., Moorty, L. M., Richter, K., November 2015 49 Hub‐periphery development pattern and inclusive growth: case study of Guangdong province Luo, X., Zhu, N., December 2015 50 Unpacking the MPI: a decomposition approach of changes in multidimensional poverty headcounts Rodriguez Castelan, C., Trujillo, J. D., Pérez Pérez, J. E., Valderrama, D., December 2015 51 The poverty effects of market concentration Rodriguez Castelan, C., December 2015 52 Can a small social pension promote labor force participation? evidence from the Colombia Mayor program Pfutze, T., Rodriguez Castelan, C., December 2015 Updated on October 23, 2017 by POV GP KL Team | 4 53 Why so gloomy? perceptions of economic mobility in Europe and Central Asia Davalos, M. E., Cancho, C. A., Sanchez, C., December 2015 54 Tenure security premium in informal housing markets: a spatial hedonic analysis Nakamura, S., December 2015 55 Earnings premiums and penalties for self‐employment and informal employees around the world Newhouse, D. L., Mossaad, N., Gindling, T. H., January 2016 56 How equitable is access to finance in turkey? evidence from the latest global FINDEX Yang, J., Azevedo, J. P. W. D., Inan, O. K., January 2016 57 What are the impacts of Syrian refugees on host community welfare in Turkey? a subnational poverty analysis Yang, J., Azevedo, J. P. W. D., Inan, O. K., January 2016 58 Declining wages for college‐educated workers in Mexico: are younger or older cohorts hurt the most? Lustig, N., Campos‐Vazquez, R. M., Lopez‐Calva, L.‐F., January 2016 59 Sifting through the Data: labor markets in Haiti through a turbulent decade (2001‐2012) Rodella, A.‐S., Scot, T., February 2016 60 Drought and retribution: evidence from a large‐scale rainfall‐indexed insurance program in Mexico Fuchs Tarlovsky, Alan., Wolff, H., February 2016 61 Prices and welfare Verme, P., Araar, A., February 2016 62 Losing the gains of the past: the welfare and distributional impacts of the twin crises in Iraq 2014 Olivieri, S. D., Krishnan, N., February 2016 63 Growth, urbanization, and poverty reduction in India Ravallion, M., Murgai, R., Datt, G., February 2016 64 Why did poverty decline in India? a nonparametric decomposition exercise Murgai, R., Balcazar Salazar, C. F., Narayan, A., Desai, S., March 2016 65 Robustness of shared prosperity estimates: how different methodological choices matter Uematsu, H., Atamanov, A., Dewina, R., Nguyen, M. C., Azevedo, J. P. W. D., Wieser, C., Yoshida, N., March 2016 66 Is random forest a superior methodology for predicting poverty? an empirical assessment Stender, N., Pave Sohnesen, T., March 2016 67 When do gender wage differences emerge? a study of Azerbaijan's labor market Tiongson, E. H. R., Pastore, F., Sattar, S., March 2016 Updated on October 23, 2017 by POV GP KL Team | 5 68 Second‐stage sampling for conflict areas: methods and implications Eckman, S., Murray, S., Himelein, K., Bauer, J., March 2016 69 Measuring poverty in Latin America and the Caribbean: methodological considerations when estimating an empirical regional poverty line Gasparini, L. C., April 2016 70 Looking back on two decades of poverty and well‐being in India Murgai, R., Narayan, A., April 2016 71 Is living in African cities expensive? Yamanaka, M., Dikhanov, Y. M., Rissanen, M. O., Harati, R., Nakamura, S., Lall, S. V., Hamadeh, N., Vigil Oliver, W., April 2016 72 Ageing and family solidarity in Europe: patterns and driving factors of intergenerational support Albertini, M., Sinha, N., May 2016 73 Crime and persistent punishment: a long‐run perspective on the links between violence and chronic poverty in Mexico Rodriguez Castelan, C., Martinez‐Cruz, A. L., Lucchetti, L. R., Valderrama Gonzalez, D., Castaneda Aguilar, R. A., Garriga, S., June 2016 74 Should I stay or should I go? internal migration and household welfare in Ghana Molini, V., Pavelesku, D., Ranzani, M., July 2016 75 Subsidy reforms in the Middle East and North Africa Region: a review Verme, P., July 2016 76 A comparative analysis of subsidy reforms in the Middle East and North Africa Region Verme, P., Araar, A., July 2016 77 All that glitters is not gold: polarization amid poverty reduction in Ghana Clementi, F., Molini, V., Schettino, F., July 2016 78 Vulnerability to Poverty in rural Malawi Mccarthy, N., Brubaker, J., De La Fuente, A., July 2016 79 The distributional impact of taxes and transfers in Poland Goraus Tanska, K. M., Inchauste Comboni, M. G., August 2016 80 Estimating poverty rates in target populations: an assessment of the simple poverty scorecard and alternative approaches Vinha, K., Rebolledo Dellepiane, M. A., Skoufias, E., Diamond, A., Gill, M., Xu, Y., August 2016 Updated on October 23, 2017 by POV GP KL Team | 6 81 Synergies in child nutrition: interactions of food security, health and environment, and child care Skoufias, E., August 2016 82 Understanding the dynamics of labor income inequality in Latin America Rodriguez Castelan, C., Lustig, N., Valderrama, D., Lopez‐Calva, L.‐F., August 2016 83 Mobility and pathways to the middle class in Nepal Tiwari, S., Balcazar Salazar, C. F., Shidiq, A. R., September 2016 84 Constructing robust poverty trends in the Islamic Republic of Iran: 2008‐14 Salehi Isfahani, D., Atamanov, A., Mostafavi, M.‐H., Vishwanath, T., September 2016 85 Who are the poor in the developing world? Newhouse, D. L., Uematsu, H., Doan, D. T. T., Nguyen, M. C., Azevedo, J. P. W. D., Castaneda Aguilar, R. A., October 2016 86 New estimates of extreme poverty for children Newhouse, D. L., Suarez Becerra, P., Evans, M. C., October 2016 87 Shedding light: understanding energy efficiency and electricity reliability Carranza, E., Meeks, R., November 2016 88 Heterogeneous returns to income diversification: evidence from Nigeria Siwatu, G. O., Corral Rodas, P. A., Bertoni, E., Molini, V., November 2016 89 How liberal is Nepal's liberal grade promotion policy? Sharma, D., November 2016 90 CPI bias and its implications for poverty reduction in Africa Dabalen, A. L., Gaddis, I., Nguyen, N. T. V., December 2016 91 Pro-growth equity: a policy framework for the twin goals Lopez-Calva, L. F., Rodriguez Castelan, C., November 2016 92 Building an ex ante simulation model for estimating the capacity impact, benefit incidence, and cost effectiveness of child care subsidies: an application using provider‐level data from Turkey Aran, M. A., Munoz Boudet, A., Aktakke, N., December 2016 93 Vulnerability to drought and food price shocks: evidence from Ethiopia Porter, C., Hill, R., December 2016 94 Job quality and poverty in Latin America Rodriguez Castelan, C., Mann, C. R., Brummund, P., December 2016 95 With a little help: shocks, agricultural income, and welfare in Uganda Mejia‐Mantilla, C., Hill, R., January 2017 Updated on October 23, 2017 by POV GP KL Team | 7 96 The impact of fiscal policy on inequality and poverty in Chile Martinez Aguilar, S. N., Fuchs Tarlovsky, A., Ortiz‐Juarez, E., Del Carmen Hasbun, G. E., January 2017 97 Conditionality as targeting? participation and distributional effects of conditional cash transfers Rodriguez Castelan, C., January 2017 98 How is the slowdown affecting households in Latin America and the Caribbean? Reyes, G. J., Calvo‐Gonzalez, O., Sousa, L. D. C., Castaneda Aguilar, R. A., Farfan Bertran, M. G., January 2017 99 Are tobacco taxes really regressive? evidence from Chile Fuchs Tarlovsky, A., Meneses, F. J., March 2017 100 Design of a multi‐stage stratified sample for poverty and welfare monitoring with multiple objectives: a Bangladesh case study Yanez Pagans, M., Roy, D., Yoshida, N., Ahmed, F., March 2017 101 For India's rural poor, growing towns matter more than growing cities Murgai, R., Ravallion, M., Datt, G., Gibson, J., March 2017 102 Leaving, staying, or coming back? migration decisions during the northern Mali conflict Hoogeveen, J. G., Sansone, D., Rossi, M., March 2017 103 Arithmetics and Politics of Domestic Resource Mobilization Bolch, K. B., Ceriani, L., Lopez‐Calva, L.‐F., April 2017 104 Can Public Works Programs Reduce Youth Crime? Evidence from Papua New Guinea’s Urban Youth Employment Project Oleksiy I., Darian N., David N., Sonya S., April 2017 105 Is Poverty in Africa Mostly Chronic or Transient? Evidence from Synthetic Panel Data Dang, H.‐A. H., Dabalen, A. L., April 2017 106 To Sew or Not to Sew? Assessing the Welfare Effects of the Garment Industry in Cambodia Mejía‐Mantilla, C., Woldemichael, M. T., May 2017 107 Perceptions of distributive justice in Latin America during a period of falling inequality Reyes, G. J., Gasparini, L. C., May 2017 108 How do women fare in rural non‐farm economy? Fuje, H. N., May 2017 109 Rural Non‐Farm Employment and Household Welfare: Evidence from Malawi Adjognon, G. S., Liverpool‐Tasie, S. L., De La Fuente, A., Benfica, R. M., May 2017 Updated on October 23, 2017 by POV GP KL Team | 8 110 Multidimensional Poverty in the Philippines, 2004‐13: Do Choices for Weighting, Identification and Aggregation Matter? Datt, G., June 2017 111 But … what is the poverty rate today? testing poverty nowcasting methods in Latin America and the Caribbean Caruso, G. D., Lucchetti, L. R., Malasquez, E., Scot, T., Castaneda, R. A., June 2017 112 Estimating the Welfare Costs of Reforming the Iraq Public Distribution System: A Mixed Demand Approach Krishnan, N., Olivieri, S., Ramadan, R., June 2017 113 Beyond Income Poverty: Nonmonetary Dimensions of Poverty in Uganda Etang Ndip, A., Tsimpo, C., June 2017 114 Education and Health Services in Uganda: Quality of Inputs, User Satisfaction, and Community Welfare Levels Tsimpo Nkengne, C., Etang Ndip, A., Wodon, Q. T., June 2017 115 Rental Regulation and Its Consequences on Measures of Well‐Being in the Arab Republic of Egypt Lara Ibarra, G., Mendiratta, V., Vishwanath, T., July 2017 116 The Poverty Implications of Alternative Tax Reforms: Results from a Numerical Application to Pakistan Feltenstein, A., Mejia‐Mantilla, C., Newhouse, D. L., Sedrakyan, G., August 2017 117 Tracing Back the Weather Origins of Human Welfare: Evidence from Mozambique? Baez Ramirez, J. E., Caruso, G. D., Niu, C., August 2017 118 Many Faces of Deprivation: A multidimensional approach to poverty in Armenia Martirosova, D., Inan, O. K., Meyer, M., Sinha, N., August 2017 119 Natural Disaster Damage Indices Based on Remotely Sensed Data: An Application to Indonesia Skoufias, E., Strobl, E., Tveit, T. B., September 2017 120 The Distributional Impact of Taxes and Social Spending in Croatia Inchauste Comboni, M. G., Rubil, I., October 2017 121 Regressive or Progressive? The Effect of Tobacco Taxes in Ukraine Fuchs, A., Meneses, F. September 2017 122 Fiscal Incidence in Belarus: A Commitment to Equity Analysis Bornukova, K., Shymanovich, G., Chubrik, A., October 2017 Updated on October 23, 2017 by POV GP KL Team | 9 123 Who escaped poverty and who was left behind? a non‐parametric approach to explore welfare dynamics using cross‐sections Lucchetti, L. R., October 2017 124 Learning the impact of financial education when take-up is low Lara Ibarra, G., Mckenzie, D. J., Ruiz Ortega, C., November 2017 125 The impact of fiscal policy on inequality and poverty in Zambia De La Fuente, A., Rosales, M., Jellema, J. R., November 2017 126 The Whys of Social Exclusion: Insights from Behavioral Economics Hoff, K., Walsh, J. S., December 2017 127 De La Fuente, A., Rosales, M., Jellema, J. R., November 2017 Gine, X., Mansuri, G., Shrestha, S. A., December 2017 For the latest and sortable directory, available on the Poverty & Equity GP intranet site. http://POVERTY WWW.WORLDBANK.ORG/POVERTY Updated on October 23, 2017 by POV GP KL Team | 10