Report No: 105723 FISCAL POLICY AND REDISTRIBUTION IN THE DOMINICAN REPUBLIC An analysis based on the "Commitment to Equity� methodology, for 2013 April 20, 2016 Macroeconomic and Fiscal Management Global Practice Caribbean Countries Management Unit Latin America and the Caribbean Region Document of the World Bank For official use only WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Standard Disclaimer: This volume is a product of the staff of the International Bank for Reconstruction and Development/ The World Bank. The findings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Copyright Statement: The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development/ The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470, http://www.copyright.com/. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax 202-522-2422, e- mail pubrights@worldbank.org. Contents Acknowledgements ....................................................................................................................................................... v Abbreviations................................................................................................................................................................ vi Executive Summary ..................................................................................................................................................... vii 1. Introduction ..........................................................................................................................................................1 2. Context: fiscal tools in the Dominican Republic ...................................................................................................4 3. Methodology and sources of information ..........................................................................................................13 3.1. CEQ methodology ......................................................................................................................................13 3.2. Data sources ...............................................................................................................................................14 3.3. Main assumptions ......................................................................................................................................15 4. Main results ........................................................................................................................................................19 4.1. The re-distributional impact of taxes .........................................................................................................19 4.1.1. Direct taxes ........................................................................................................................................20 4.1.2. Indirect taxes .....................................................................................................................................23 4.2. Social spending in the Dominican Republic................................................................................................28 4.2.1. Direct transfers ..................................................................................................................................28 4.2.2. Indirect subsidies ...............................................................................................................................31 4.2.3. In kind-transfers: education and health ............................................................................................33 5. Net impact of the fiscal system on income redistribution in the Dominican Republic .......................................41 5.1. Fiscal policy instruments, poverty, and inequality in the Dominican Republic ..........................................41 5.2. Is fiscal policy more or less redistributive and pro-poor than in other countries? ....................................43 5.3. Income redistribution: vertical and horizontal equity, effectiveness indicators. ......................................46 5.4. Resource needs to fill in coverage gaps .....................................................................................................47 6. Options to enhance the equity outcomes of fiscal policy in the Dominican Republic ........................................51 6.1. Alternative VAT scenarios for a Fiscal Impact Pact ....................................................................................51 6.2. Policy options .............................................................................................................................................54 7. Bibliography ........................................................................................................................................................57 8. Annexes...............................................................................................................................................................61 8.1. Main taxes in the Dominican Republic .......................................................................................................61 8.2. Recent tax reforms in the Dominican Republic ..........................................................................................62 8.3. Assumptions ...............................................................................................................................................64 Personal income taxes ........................................................................................................................................64 Indirect taxes ......................................................................................................................................................64 Direct transfers ...................................................................................................................................................66 Subsidies .............................................................................................................................................................66 i WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Health..................................................................................................................................................................67 Education ............................................................................................................................................................67 8.4. Basic characteristics of individuals and households in the bottom decile .................................................69 Income-based incidence analysis ........................................................................................................................69 8.5. Tables containing additional results...........................................................................................................71 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only List of figures Figure I. The distribution of direct and indirect taxes: concentration curves and Lorenz curve for market income ... ix Figure II. The distribution of direct transfer spending by income level, US$ in 2015 PPP terms ................................. xi Figure III. The distribution of health and education spending: concentration curves and Lorenz curve for market income ......................................................................................................................................................................... xii Figure IV. Inequality in disposable and final income versus market income (in Gini points) ..................................... xiii Figure V. Summary of options for improving the equity impact of fiscal policy in the Dominican Republic ............. xvi Figure 6. Composition of taxes and GNI per capita: International comparison ............................................................5 Figure 7. Social spending as a share of GDP: International comparison .......................................................................7 Figure 8. Income concepts used in fiscal incidence analysis .......................................................................................13 Figure 9. Progressivity of direct and indirect taxes: concentration curves and Lorenz curve for market income ......20 Figure 10. Progressivity of direct taxes: Concentration curves and Lorenz curve for market income ........................21 Figure 11. Direct taxes concentration shares per socioeconomic groups ...................................................................21 Figure 12. Direct taxes concentration shares per decile, country comparison ...........................................................22 Figure 13. Progressivity of indirect taxes: Concentration curves and Lorenz curve for market income .....................24 Figure 14. Indirect taxes concentration shares per socioeconomic groups ................................................................25 Figure 15. Indirect taxes, concentration shares per decile ..........................................................................................26 Figure 16. Distribution of direct transfer spending by level (percentages) .................................................................29 Figure 17. Concentration shares of direct transfers, by deciles, country comparison ................................................30 Figure 18. Distribution of indirect subsidies spending (left) and incidence on market income by level (right) ..........32 Figure 19. Concentration shares (left) and incidence of indirect subsidies (right) in comparable countries ..............33 Figure 20. Progressivity of health and education spending: concentration curves and Lorenz curve for market income .........................................................................................................................................................................34 Figure 21. Distribution of education spending by level (percentages) ........................................................................35 Figure 22. Enrollment in public education by level for school aged children (percentages) .......................................36 Figure 23. Incidence of education expenditures by level for school aged children (percentages) .............................37 Figure 24. Incidence of education expenditure by level for school aged children (percentages) ...............................37 Figure 25. Distribution of health spending by level (percentages) ..............................................................................38 Figure 26. Individuals who live in beneficiary households by health program and socioeconomic ranking (percentages) ...............................................................................................................................................................39 Figure 27. Incidence of health expenditures by coverage regime ...............................................................................39 Figure 28. Concentration coefficients with respect to market income, by fiscal instrument .....................................40 Figure 29. Change in inequality: Disposable and final income versus market income (in Gini points) .......................43 Figure 30. Percentage of population by socioeconomic class in the Dominican Republic ..........................................45 Figure 31. Post fiscal (left) and final income (right) as a share of market income ......................................................45 Figure 32. Fiscal incidence curves and fiscal mobility profiles by deciles ....................................................................47 Figure 33. Percentage of individuals benefiting from health (left) and public education (right) services, by daily income .........................................................................................................................................................................48 Figure 34. Beneficiaries of VAT tax expenditure for different product categories ......................................................51 Figure 35. Effects on inequality and poverty of alternative ITBIS exemption scenarios .............................................53 Figure 36. Effects on revenue increase in scenarios of ITBIS (as a percentage of total disposable income) ..............53 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only List of tables Table I. The poverty headcount rate at US$2.50 per day in PPP terms for each income concept ............................. xiv Table 2. Composition of taxes in the Dominican Republic (2013) .................................................................................4 Table 3. Composition of expenditures in Dominican Republic (2011 and 2013)...........................................................6 Table 4. Dominican Republic: Composition of public education expenditure (2013) ...................................................8 Table 5. Dominican Republic: Composition of public health expenditure (2013) .........................................................9 Table 6. Direct transfers programs in Dominican Republic in 2013 ............................................................................11 Table 7. Benchmark scenario: Population and Income shares of market income ......................................................19 Table 8. Benchmark scenario: Incidence of direct and indirect taxes by socioeconomic group (% of market income) .....................................................................................................................................................................................20 Table 9. Benchmark scenario: Incidence of personal income, interest, and dividend taxes by socioeconomic group (% of Market income) ..................................................................................................................................................22 Table 10. Benchmark scenario: Incidence of ITBIS and excises taxes by socioeconomic group (% of market income) .....................................................................................................................................................................................24 Table 11. Progressivity indices for direct and indirect taxes, country comparison .....................................................27 Table 12. Incidence of direct transfer programs on socioeconomic class income (percentages) ...............................29 Table 13. Distribution of health and education spending by socioeconomic group (% of Market income) ...............34 Table 14. Dominican Republic: Poverty and inequality indicators at each income concept .......................................42 Table 15. Average per capita income in each market income decile, in Dominican pesos a year ..............................42 Table 16. Poverty headcount rate for the US$2.50 PPP a day for each income concept ............................................44 Table 17. Taxes, transfers and subsidies: Overall redistributive effect* (Decline in Gini Points; shown as positive) .46 Table 18. Beckerman and Immervoll et al. effectiveness indicators ...........................................................................47 Table 19. Estimated resource needs to close existing social gaps in the Dominican Republic....................................50 Table 20 Personal Income Tax, rates and thresholds ..................................................................................................61 Table 21. Personal Income Tax, rates and thresholds, adjusted to 2007 prices..........................................................64 Table 22. VAT incidence with and without the evasion assumption ...........................................................................65 Table 23. VAT concentration shares with and without the evasion assumption ........................................................65 Table 24. Adjustments on direct transfer, ADESS data................................................................................................66 Table 25. Dominican Republic: Features of households at bottom and top of the income distribution ....................69 Table 26. Reduction in inequality across income concepts .........................................................................................71 Table 27. Reduction in inequality across income concepts .........................................................................................72 Table 28. Incidence for taxes and transfers (share of market income and socioeconomic group) .............................75 Table 29. Concentration coefficients and Budget shares by program ........................................................................78 Table 30. VAT tax expenditures by category of goods and beneficiaries by income group ........................................80 Table 31. Breakdown of social spending (2013) ..........................................................................................................81 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Acknowledgements Fiscal Policy and Redistribution in the Dominican Republic is a World Bank report based on the Commitment to Equity Assessment©—a tool developed by the Commitment to Equity Institute at Tulane University. The exercise was prepared under the leadership of Blanca Moreno Dodson (Lead Economist, GMFDR) and Miguel Eduardo Sánchez Martín (Country Economist, GMFDR), in collaboration with Maynor Cabrera and Jaime Aristy-Escuder (Consultants). Substantive contributions were made by Javier Eduardo Báez (Senior Economist, GPVDR), Alan Fuchs (Economist, GPVDR), Juan Carlos Parra (Economist, GPVDR), Gianluca Mele (Senior Economist, GMFDR), Onur Erdem (Public Sector Specialist, GGODR), Matías José Arnal (Research Analyst, GMFDR), Patricia Chacón-Holt (Team Assistant, GMFDR), and Richard Alm (Consultant and editor). Special thanks to Nora Lustig (Tulane University), Samantha Greenspun (Tulane University), and Gabriela Inchauste (Lead Economist, GPVDR) for the exceptional help granted to validate the accuracy of results when applying the Commitment to Equity methodology and preparing comparisons against benchmarks. The team is also obliged to peer reviewers Omar Arias, Luis Felipe López-Calva, and Sudarshan Gooptu for their comments, feedback and advice. The report has been prepared under the overall guidance and strategic direction of Sophie Sirtaine, (Country Director, LCC3C), McDonald Benjamin (Country Manager, Dominican Republic), Francisco Galrao Carneiro, (Program Leader, LCC3C), Auguste Tano Kouame (Practice Manager, GMFDR), and Miria Pigato (Practice Manager, GMFDR). The team is grateful for the collaboration of the following government and donor counterparts: Magdalena Lizardo, Antonio Morillo, Alexis Cruz, Martín Francos (Ministry of Economy, Planning and Development); Luis Madera, Augusto de los Santos, Mabely Díaz (National Office for Statistics); José Luis Actis (Ministry of Finance); Guarocuya Félix, Marvin Cardoza, Hamlet Gutiérrez (General Directorate for Internal Taxation, Ministry of Finance), Matilde Chávez (Social Cabinet), Tirsis Quezada, Rafael Montero (Ministry of Public Health); Chanel Rosa (SENASA); Pedro Castellanos, Ayacx Mercedes (DIGEPEP, Presidency); Rafael Pérez (National Council for the Social Security); Ramón González Hernández (Central Bank of the Dominican Republic); Rita Mena (UNDP); Javier Casasnovas (European Union). WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Abbreviations ADESS Administrator for Social INAPA National Institute of Water and Subsidies Sanitation ARS Health Risk Administrators ISR Income Tax BEEP Bonus for Progress in Studying ISC Selective Tax on Consumption BGC Gas Bonus to Drivers Program ITBIS Tax on the Transfer of Industrialized Goods and Services CCT Conditional Cash Transfer MEPyD Ministry of Economy, Planning and Development CDEEE Electricity Distribution Holding LPG Liquefied Petroleum Gas CEQ Commitment to Equity ONE National Office of Statistics CERS Health reform commission PAHO Pan American Health Organization CESDEM Center for Social and PIAMG Marine Officials Incentive Demographic Studies Program CIT Corporative Income Tax PIPP Incentive to Preventive Police CNSS National Council for Social PIT Personal Income Tax Security CONAVIHSIDA National VIH Commission PRA Blackout Reduction Program CPI Consumer Price Index PROMESE-CAL Program of Essential Drugs, logistics support DGII General Directorate of Internal SENASA National Health Insurance Taxation Authority DIDA Bureau of Consumer SISALRIL Bureau of Occupational Health Information and Protection and Safety DIGEPEP Special Programs Directorate of SIUBEN Single Beneficiary Selection Presidency System DR-CAFTA Dominican Republic - Central RD$ Dominican Republic Pesos American Free Trade Agreement ENDESA Demographic and Health Survey UNESCO United Nations Educational, Scientific and Cultural Organization ENFT Labor Force National Household UNICEF United Nations Children's Emergency Fund ENIGH National Survey of Household SNS National Service of Health Income and Expenditure 2006-07 GDP Gross Domestic Product TAE School based transfer program GNI Gross National Income UNDP United Nations Development Programme IADB Inter-American Development VAT Value Added Tax Bank IDEC Dominican Initiative for a Countries Quality Education IDSS Dominican Institute of Social DR Dominican Republic Security IES Incentive to Tertiary Education Regions Program ILAE Incentive for School Attendance LAC Latin American Countries Program WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Executive Summary Motivation and Context: Economic Growth, Social Inclusion and the Role of the State in the Dominican Economy Over the past three decades the Dominican Republic has ranked among the fastest growing economies in Latin America and the Caribbean (LAC), yet relatively high rates of poverty and inequality persist. Annual economic growth averaged 5.7 percent over 1991-2013, one of the highest rates in the region. Meanwhile, GNI per capita rose from 52 percent of the LAC average to 78 percent, or US$5,520 in 2012. According to official statistics the moderate poverty rate shot from 32 percent in 2000 to almost 50 percent in 2004, as the country suffered a severe banking crisis, then declined gradually to around 41 percent in 2013 before reaching 35 percent in October 2014. Income inequality improved slightly between 2000 and 2013, with the Gini index falling from 0.549 to 0.514. However, despite many years of relatively robust and broad-based growth, the country’s poverty and inequality indicators remain high by regional standards. The government’s limited revenue capacity narrows the scope for progressive fiscal policies. The government attempted to compensate for the lower tariff rates mandated by the CAFTA-DR regional free trade agreement by increasing the value-added tax (VAT) rate from 12 percent to 16 percent in 2004 (Law 288-04) and then to 18 percent in 2012 (Law 253-12). Indirect taxes represent a large share of total revenue, but remain at around 8.7 percent of GDP, a modest amount relative to other lower-middle-income countries (Garza et al., 2012). High levels of informality and significant tax exemptions contribute to low revenue mobilization, with tax expenditures amounting to an estimated 5.9 percent of GDP in 2013, including 3.4 percent of GDP in VAT exemptions alone. Meanwhile, 15-year corporate income tax holidays for companies established in Special Economic Zones further increased tax expenditures, and there is a relatively high minimum threshold for personal income tax liability. While the government has made considerable efforts to increase social spending in recent years, weaknesses in public service delivery have diminished the impact of social spending on poverty and inequality. Public education spending rose from around 2.2 percent of GDP in 2011 to close to 4 percent in 2013. Meanwhile, certain health services were privatized, and lower-income households began to receive insurance under a subsidized scheme, though a large share of the population remains uninsured. Moreover, the uneven quality of public education and health services encourages households to seek out private service providers, even households in the lower income quintiles (Sánchez-Martín and Senderowitsch, 2012). Finally, electricity subsidies and the energy sector’s large technical and commercial losses impose a fiscal cost of close 2 percent of GDP, further limiting the resources available for pro-poor spending. The banking and economic crisis of 2004 gave rise to a number of social programs designed to shield poor and vulnerable households from economic shocks. The crisis caused an abrupt economic slowdown, which pushed about 1.3 million Dominicans into poverty at a time when social security and pro- poor transfer programs were underdeveloped. In 2005 the government launched a series of reforms that created a Single Beneficiary Selection System (Sistema Único de Beneficiarios, SIUBEN) based on a quality-of-life index, and an independent Administrator for Social Subsidies (Administradora de Subsidios Sociales, ADESS). The authorities also established a cash transfer program, Solidaridad, which uses a debit card to transfer funds to beneficiaries. This system has improved the targeting of both conditional transfers for nutritional and educational support and unconditional transfers for electricity and gas consumption. Together, transfer payments reached 0.5 percent of GDP in 2013. In this context, the following report attempts to assess the progressivity of fiscal policy in the Dominican Republic and its impact on poverty and inequality. Specifically, it evaluates the hypothesis that the limited redistributive effect of fiscal policy slowed improvements in poverty and inequality during a period of strong economic growth. A fiscal-incidence analysis based on the Commitment to Equity (CEQ) WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only methodology (Lustig and Higgins, 2013)1 is used to isolate the poverty and equity impact of fiscal policies, including taxes, transfers and subsidies, as well as public spending in key social sectors. The analysis has three overarching objectives. The first is to understand how various taxes and expenditure items affect income distribution in the Dominican Republic. The second is to compare the Dominican Republic’s experience to those of similar countries, such as Costa Rica and Peru. The third is to define and evaluate alternative policy arrangements that could help to enhance the redistributive impact of Dominican fiscal policy. Data Limitations and Analytical Assumptions In order to estimate the impact of taxes and transfers on income inequality and poverty in the Dominican Republic, the Commitment to Equity methodology is applied. This methodology, applied in more than 20 countries and described in detail in Lustig and Higgins (2013), employs incidence analysis structured along successive steps. The CEQ methodology calculates net market income (after direct taxes), disposable income (after direct transfers), post-fiscal income (after indirect taxes and subsidies), and final income (by monetizing the value of public education and health services). The redistributive effects of transportation, energy and telecommunications infrastructure are not included in the analysis due to limited data and methodological considerations. Similarly, the analysis only examines personal income and sales taxes paid by households. A number of official data sources have been accessed in the case of the Domincian Republic . This exercise draws on data from the 2006-07 National Survey of Household Income and Expenditures (Encuesta Nacional de Ingresos y Gastos de los Hogares, ENIGH) and the 2013 Demographic and Health Survey (Encuesta Demográfica y de Salud, ENDESA), as well as fiscal data from the Ministry of Finance and administrative records from the Ministry of Education, the Ministry of Health, the National Health Insurance Authority (Seguro Nacional de Salud, SENASA) and ADESS. Due to a lack of updated household survey data, a set of assumptions was used to estimate the impact of recent policies. The latest household income and expenditure survey, ENIGH, was conducted in 2007, and thus the available data do not capture the important policy decisions made between 2007 and 2013. These considerations were incorporated into the CEQ methodology by modifying the major tax rates and bases, and by expanding the coverage of direct transfers. The application of the 2013 tax and social program structure to the 2007 survey data enabled a simulation of income and poverty impacts, and 2013 public revenue and spending data were deflated to 2007 prices. Statutory tax rates and income brackets were applied in the estimation of direct tax revenue, similar to other applications of the CEQ methodology (e.g. Lustig et al., 2013). Tax evasion assumptions, which were based on discussions with the authorities, were applied only to VAT, not direct or other taxes. This analysis only evaluates the equity effects of the tax system, not its buoyancy or efficiency. International benchmarking exercises and resulting policy recommendations should be interpreted with caution. Cross-country comparisons are tempered by idiosyncratic variations in taxes systems, transfer programs, and fiscal arrangements more generally, as these variations cannot be fully accounted for in an analytical model. In addition, similar policies may have different effects on poverty and inequality due to intrinsic differences in the nature of the public sector and the broader country context that are not captured by the analysis. The CEQ methodology also assumes that the government has the capacity to effectively manage and allocate rising revenues, though in practice institutional capacity limitations may reduce the marginal efficiency of public financial management. Finally, estimations are based on static incidence analysis, which assumes other variables will remain constant and will not affect the behavior of economic agents. 1Led by Nora Lustig since 2008, the Commitment to Equity (CEQ) project is an initiative of the Center for Inter- American Policy and Research (CIPR) and the Department of Economics, both at Tulane University. See www.commitmentoequity.org. WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only These caveats should be borne in mind when considering the predicted impact of the proposed policy options. The active assistance of the authorities has helped to overcome data limitations and ensure that results are consistent with all existing evidence. Discussions with the General Directorate for Internal Taxation (Dirección General de Impuestos Internos, DGII) regarding VAT compliance informed the assumptions for tax evasion used in the analysis, and VAT revenue estimates were compared against actual collections in 2013. Consultations with the Ministry of Finance, the Electricity Distribution Holding Company (Corporación Dominicana de Empresas Eléctricas Estatales, CDEEE), the Social Cabinet, the ADESS, the Ministry of Education, the Ministry of Health, and SENASA also helped ensure the accuracy of the information and results of the incidence analysis for indirect electricity transfers, direct transfers, and health, education and social security spending. In addition, the estimations used in the analysis were reviewed by faculty at Tulane University to verify the results and ensure both their consistency with the CEQ methodology (Lustig and Higgins, 2013) and their comparability with the findings of similar country analyses. The Redistribute Impact of Fiscal Policy in the Dominican Republic The analysis found that direct income taxes were progressive across all income levels in 2013. Direct taxes, which include taxes on wages, nonwage personal income, interest income and dividends, were found to be especially progressive at the upper end of the income distribution: taxpayers in the top decile, who account for around 40 percent of total market income, were responsible for 92 percent of total collections. Personal income tax was both the largest and most progressive of the direct taxes, due in part to its exemption threshold of US$615 per month in 2013, which is far above the minimum salary of roughly US$130 per month. Taxes on dividends were also found to be highly progressive, and their impact on households with per capita incomes under US$10 per day was negligible. Figure I. The distribution of direct and indirect taxes: concentration curves and Lorenz curve for market income 100% % of income or tax collection 80% 60% 40% 20% 0% 0% 20% 40% 60% 80% 100% % of population Market Income Indirect taxes Population Direct taxes Source: Authors’ estimates based on ENIGH 2007. Indirect taxes, however, were found to be just barely progressive. In 2013 the general rate for the Tax on transferring Industrial Goods and Services (Impuestos de Transferencias Bienes Industrializados y Servicios, ITBIS), the Dominican Republic’s principal VAT, was 18 percent. A reduced tax rate of 8 percent was applied to a set of basic goods, and certain staple food products, as well as education, health services and electricity, were exempt. The ITBIS represented 58.6 percent of total indirect taxes considered in the analysis, but despite the reduced rate and exemptions it was only slightly progressive. Excise taxes on oil represented about 20 percent of total indirect taxes, alcoholic beverages and tobacco accounted for 11 percent, and international trade taxes made up most of the remaining 20 percent. Dominicans in the top income decile contributed 41.2 percent of total ITBIS revenue, just above their 40.5 percent share of market income. Overall, the ITBIS reduced WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only average market income by 4.4 percent. The market income of Dominicans living on less than US$1.25 per day in purchasing-power party (PPP) terms was reduced by 3.5 percent, while the income of those living on between US$1.25 and US$2.50 per day was reduced by 4 percent. Excise taxes on consumption2 were found to be more progressive than ITBIS. They reduced the market income of those living on more than US$50 per day by 5.9 percent, a significantly larger share than the 1.2 percent reduction experienced by the poorest taxpayers. These findings suggest that the Dominican Republic’s tax structure is relatively progressive compared to those of similar countries. Direct taxes were especially progressive by international standards, although the Dominican Republic’s high personal income tax threshold caused total direct tax revenues to represent the smallest share of GDP among comparator countries.3 The Dominican Republic’s large informal sector, which employs an estimated 56 percent of its labor force, also contributes to relatively low direct tax collection capacity. Meanwhile, relatively large excise tax collections, which are primarily borne by higher- income taxpayers, increase the progressivity of indirect taxes. Indeed, the Dominican Republic was one of only a few countries analyzed according to the CEQ methodology in which indirect taxes were found to be progressive. The distribution of direct transfers is also progressive, though some reach the poor more effectively than others. Conditional cash transfer programs designed to support nutrition (Comer es Primero) and basic education (Incentivo a la Asistencia Scholar) are the country’s most progressive direct transfers. Around 52 percent of public expenditures on Comer es Primero reached households with per capita incomes below US$4 per day, 38 percent reached households with per capita incomes between US$4 and US$10 per day, and less than 10 percent went to those with incomes above US$10 per day. By contrast, more than 60 percent of total spending on gas and electricity transfers (Bonogas Hogar and Bono Luz) went to households with per capita incomes of more than US$4 per day. This is due to differences in program targeting, as many households that are defined as non-poor according to the SIUBEN index can still qualify as beneficiaries of Bonogas Hogar and Bono Luz. 2 These include single-stage sales taxes on alcoholic beverages, cigarettes, oil products, telecommunications and insurance services, and several types of imported goods. 3 Comparisons are based on Lustig et al. (2013), CEQ Standard Indicators , drawing from Commitment to Equity methodology analysis by a number of authors for the following countries: Armenia (2011), Brazil (2009), El Salvador (2011), Ethiopia (2011), Guatemala (2010), Jordan (2010), Mexico (2010), South Africa (2010), Sri Lanka (2009), and Uruguay (2009). WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Figure II. The distribution of direct transfer spending by income level, US$ in 2015 PPP terms 100.0% 100.0% 90.0% 80.0% 80.0% 70.0% 60.0% 60.0% 50.0% 40.0% 40.0% 30.0% 20.0% 20.0% 10.0% 0.0% 0.0% 0.0% 20.0% 40.0% 60.0% 80.0% 100.0% Comer es Bono Gas Bono Luz Quisqueya Other primero Hogares aprende Direct Population Market Income contigo Transfers Comer es primero Bono Gas Hogares Bono Luz Quisqueya aprende contigo Extreme Poor (y<2.5) Moderate Poor (2.54) Source: Authors’ elaboration using the CEQ methodology. Although direct transfers appear to have been progressive overall, their relationship to poverty and inequality was less pronounced than in other countries. Direct transfers represented 13.4 percent of the market income of the first decile in the Dominican Republic, similar to the rates observed in Brazil and Bolivia but well below those of Argentina, Armenia, Ethiopia, Mexico, Peru, Sri Lanka and Uruguay, all of which range from 24 to 30 percent. The relatively modest amounts allocated under the different transfer programs, as well as the less progressive targeting mechanisms of Bono Luz and Bonogas Hogar, contribute to this result. Nevertheless, the dramatic expansion in the coverage of direct transfer programs over the past ten years has greatly increased their redistributive impact. Some untargeted electricity subsidies were found to be regressive. Explicit subsidies mandated by public policy (i.e. below-cost tariffs) and indirect subsidies achieved via commercial losses (i.e. illegal connections, fraud and nonpayment) both affect the distribution of electricity costs, and the incidence analysis is modeled accordingly. The results confirm that around 81 percent of total spending on explicit electricity subsidies in 2013 benefited non-poor individuals, similar to previous estimates by Actis (2012). However, indirect electricity subsidies represented around 2.5 percent of the market income of those earning less than US$4 per day. Any effort to reduce commercial losses in the electricity sector should include an increase in explicit subsidies (e.g. Bono Luz) or other mechanisms designed to protect the poor from the regressive impact of curbing implicit subsidies. The rapid increase in education spending observed between 2011 and 2013 appears to have been progressive. Rising education expenditures from 1.9 percent of GDP in 2011 to 3.8 percent of GDP in 2013 reduced the Gini coefficient by an estimated 1.1 points over the counterfactual. The Gini coefficient fell by an estimated 5.6 points over the period, whereas it would have fallen by an estimated 4.5 points had education spending remained at its 2011 level. WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Figure III. The distribution of health and education spending: concentration curves and Lorenz curve for market income 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 0% 20% 40% 60% 80% 100% Market Income Education Population Health Source: Authors’ estimates based on ENIGH 2007. As in other countries, the overall distribution of public education spending is progressive, but it is absolutely progressive only at the preschool, primary, and lower secondary levels. Students from households in the bottom 40 percent of the population receive close to two-thirds of all preschool, primary and lower secondary spending. Upper secondary spending is progressive in relative terms: the overall distribution is broadly proportional to the population, and its benefits decrease with income level. However, tertiary education is highly regressive, with more that 80 percent of public spending going to non-poor students. This a common feature of tertiary education systems, which combine high marginal costs with a tendency to benefit students from wealthier backgrounds. While the Dominican Republic compares favorably with other countries in terms of the distribution of education spending among the poorest deciles, the observed progressivity may be accentuated by a tendency among middle- and upper-income families to opt out of public education, which is often perceived as being of mediocre quality. For example, more than 90 percent of primary school children (ages 7 to 12) from extremely poor households were enrolled in public schools in 2013, compared to just one-third of middle-income children (Sánchez-Martin and Senderowitsch, 2012). The distribution of health spending was more progressive than that of education, but limited resources blunted its redistributive impact. Both the subsidized health insurance and other noncontributory programs (e.g. hospital and outpatient care) are markedly progressive, and amounts allocated under noncontributory programs were six times larger than those allocated via subsidies. By contrast, 60 percent of spending on subsidized care provided through the Essential Medicines Program (Programa de Medicamentos Esenciales, PROMESE) goes to non-poor beneficiaries. Despite notable progress in recent years, the government faces substantial challenges in increasing health insurance coverage. According to the 2013 ENDESA survey subsidized health insurance covered less than 25 percent of the two poorest quintiles, and noncontributory programs reached less than 21 percent (CESDEM, 2014). Out-of-pocket healthcare expenditures and related costs also continue to represent a significant burden for poor households. The Redistributive Impact of Dominican Fiscal Policy in International Perspective The Dominican Republic’s fiscal policies have helped to mitigate inequality. Direct taxes and transfers reduced the Gini coefficient for disposable income by an estimated 0.012 points in 2013, similar to the effect observed in Bolivia, Peru, and Sri Lanka and somewhat higher than that of Guatemala and Indonesia (Figure IV). If in-kind education and health spending are monetized, Dominican fiscal policy causes the Gini coefficient to fall by a full 0.056 points, as public spending far outweighs direct transfers, and the poor are the WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only most likely to use public services. Brazil, Costa Rica, and South Africa, which have the most redistributive fiscal policies in the comparator group, reduce inequality through significantly higher levels social spending than prevail in the Dominican Republic. The Dominican Republic’s fiscal policies are also pro-poor, but less so than those of comparable countries. In 2013 Dominican households in the poorest decile received transfers and indirect subsidies equal to 9.2 percent of their market income, a relatively modest share by international standards. This may be due to the fact that Dominican households in the lowest decile have a higher market income per capita than their counterparts in Brazil, South Africa and Uruguay, and as a result the amounts they receive are smaller relative to their income. Including the monetized value of public health and education spending, fiscal policies boost the market income of Dominican households in the poorest decile by 68 percent, about half the average for comparator countries, excluding South Africa. The poverty rate after taxes and transfers is not significantly altered. In 2013 direct taxes had no effect on the poverty rate at the extreme poverty line of US$2.50 per day in PPP terms, which stood at 19.5 percent. While direct transfers could have reduced the poverty rate to 18.2 percent, this effect was offset by the impact of indirect taxes. Nevertheless, indirect taxes have a smaller negative impact on poverty incidence in the Dominican Republic than in countries such as Brazil or Bolivia, where cash transfers appear to have significantly reduced poverty rates, yet the incidence of extreme poverty has actually increased once indirect taxes are accounted for ). While the Dominican Republic’s direct and indirect taxes, and transfers and subsidies have modestly reduced vertical inequality, good news is that they have had a limited effect on horizontal inequality. The country’s re-ranking as a proportion of vertical inequality is by far the lowest among the five countries analyzed (Bolivia, Brazil, Indonesia, South Africa), meaning that, not many non-poor households fall into poverty as a result of fiscal policies (while others are lifted out of poverty by transfers). On the other hand, effectiveness indicators (Beckerman, 1979; Immervol, 2009) suggest the Dominican Republic has scope to increase the impact of direct transfers and refine their targeting efficiency. Figure IV. Inequality in disposable and final income versus market income (in Gini points) South Africa (2010) Guatemala ( 2010) El Salvador (2011) Costa Rica (2010) Indonesia (2012) Sri Lanka (2009) Armenia (2011) Uruguay (2009) Ethiopia (2011) Mexico (2010) Jordan (2010) Bolivia (2009) Brazil (2009) Peru (2009) DR (2013) 0.00 -0.02 -0.04 Change in Gini points -0.06 -0.08 -0.10 -0.12 -0.14 -0.16 -0.18 -0.20 Disp. vs market income Final vs market income Source: CEQ working papers (http://www.commitmentoequity.org), Tulane University and World Bank staff calculations. WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Table I. The poverty headcount rate at US$2.50 per day in PPP terms for each income concept Net Net variatio Net Disposab variation Market n (post Market le Post-fiscal Income (disposab Income fiscal Income Income le to to market) market) (1) (2) (3) (4) 2= 1- 3=2 Direct +Cash 4=3-Indirect Taxes =4-1 =3-1 Taxes Transfers Armenia (2011) 31.3% 32.0% 28.9% 34.9% 3.6% -2.4% Bolivia (2009) 19.6% 19.6% 17.6% 20.2% 0.6% -2.0% Brazil (2009) 15.1% 15.7% 11.2% 16.3% 1.2% -3.9% Costa Rica (2010) 5.4% 5.7% 3.9% 4.2% -1.2% -1.5% Dominican R. (2013) 19.5% 19.5% 18.2% 19.5% 0.0% -1.3% El Salvador (2011) 14.7% 15.1% 12.9% 14.4% -0.2% -1.8% Ethiopia (2011) 81.7% 82.7% 82.4% 84.2% 2.6% 0.7% Guatemala (2010) 35.9% 36.2% 34.6% 36.5% 0.6% -1.3% Indonesia (2012) 56.4% 56.4% 55.9% 54.8% -1.6% -0.5% Jordan (2010) 4.2% 4.2% 2.4% 1.8% -2.4% -1.8% Mexico (2010) 12.6% 12.6% 10.7% 10.7% -1.9% -1.9% Peru (2009) 15.2% 15.2% 14.0% 14.5% -0.7% -1.1% South Africa (2010) 46.2% 46.4% 33.4% 39.0% -7.2% -12.8% Source: CEQ working papers (http://www.commitmentoequity.org), Tulane University and World Bank staff calculations. Notes: Year of the survey in parenthesis. Bolivia and Indonesia include indirect taxes only. An additional 1.3 percent of GDP would be required to lift all Dominicans above the extreme poverty line and achieve full education and health coverage among that segment of the population. Closing the extreme poverty gap would require an additional RD$18.3 billion in cash transfers, equivalent to 4.9 percent of government revenue and 0.7 percent of GDP in 2013. This would imply doubling direct transfers from their current level. Achieving full public education and health coverage among extremely poor households would require a further RD$14.6 billion, or 0.6 percent of 2013 GDP. Enhancing the Progressivity and Pro-Poor Impact of Fiscal Policy Increasing fiscal revenue while maintaining the tax system’s progressivity poses a significant challenge for Dominican policymakers. While the country’s tax structure is highly progressive by international standards, total revenue levels remain relatively low. Boosting revenues by the 1.3 percent of GDP necessary to close the extreme poverty gap and expand social service coverage would require substantial changes to current tax policies. While the personal income tax represents the lion’s share of direct tax revenue, the effective tax rate among individuals earning more than US$40 per day is just 3.5 percent, far below the official rate of 15 percent. Measures to reduce tax evasion by upper-income households, especially for taxes on dividends and interest, could help bolster revenue collection, and efforts to expand the size of the formal sector could help broaden the tax base. The government could raise additional revenue by paring back ITBIS exemptions. Exemptions represented close to 3 percent of GDP in 2013 (Ministerio de Hacienda, 2015), and most of their benefits accrued to middle- and upper-income households. However, phasing out certain exemptions could have negative impacts on poverty and inequality. Exempting goods included in the basic consumption basket used in the national poverty measurement methodology, along with health and education services, while eliminating less progressive exemptions, could increase revenue by as much as 0.5 percent of GDP. Any negative effects resulting from the elimination of ITBIS electricity exemptions could be offset through the Bono Luz program, while Solidaridad could compensate for similar effects on other goods and services. Modifying the targeting criteria for Bono Luz and Bonogas Hogar could greatly improve their progressivity. Phasing out beneficiaries in the SIUBEN index category 3, who are non-poor, could yield a savings totaling around 0.1 percent of GDP. This savings could then be used to expand both programs’ coverage among the poor. Since Bono Luz and Bonogas Hogar both essentially function as universal transfers, an alternative would be to maintain non-poor beneficiaries but focus future coverage expansions on the poor. According to ADESS, out of 2.4 million potential beneficiaries in 2013, only 843,000 were enrolled in Bonogas Hogar and 533,000 in Bono Luz. Conditional cash transfers have proven effective in reaching the poor, and strengthening these programs could increase both the progressivity and pro-poor impact of fiscal policy. The authorities should consider increasing the amounts transferred through well-targeted instruments such as Solidaridad, and these transfers should be indexed to inflation. The government should continue refining the SIUBEN-based targeting mechanism and strengthening the ADESS administration, both of which have been highly successful over the past decade. However, the recent proliferation of small incentive programs could diffuse the impact of progressive transfers, and these programs may need to be consolidated over time. The establishment of support systems and programs (Progresando con Solidaridad) to facilitate labor-market integration among households that have escaped poverty will enable greater resources to be focused on the remaining poor households. xv WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Figure V. Summary of options for improving the equity impact of fiscal policy in the Dominican Republic •Maintain VAT •Eliminate implicit •Improve the quality exemptons for items electricity subsidies of public education Taxes Transfers Social Services included in the basic & compensate the by implementing the consumption basket, poor through Bono Education Pact as well as education Luz •Expand access to and health services •Prioritize the pre-primary and •Eliminate other VAT enrolment of the secondary education exemptions and use poor as Bono Luz among the poor cash transfers to and Bonogas Hogar •Increase health compensate the expand spending by 0.3 poor •Continue refining percent of GDP to •Increase compliance and updating the achieve universal with taxes on SIUBEN-based coverage of the interest and targeting system extremely poor dividends •Reduce the share of non-poor PROMESE beneficiaries Source: World Bank staff elaboration. Policymakers should strive to improve education quality by implementing measures included in the Education Pact. The recent increase in public education spending is expected to further diminish inequality, yet a number of important quality challenges persist, which calls for full implementation of the Education Pact. In addition, expanding education access should be a top priority, especially at the pre-primary and secondary levels, where enrollment rates remain low among the extremely poor. Finally, introducing a series of grants to support high-performing students from poor households could help reduce dropout rates and improve the equity of tertiary education spending. Achieving universal public health coverage among the extremely poor would require a significant increase in spending. Public health expenditures in the Dominican Republic remain low by international standards at around 1.7 percent of GDP, and while there have been notable improvements in coverage over the past decade, the bottom 40 percent of the population still has limited access to subsidized insurance and noncontributory assistance programs. All analyzed health programs and spending components were highly progressive, except for PROMESE, which is barely progressive. Refocusing PROMESE on poor households would significantly increase the progressivity of health spending. In addition, greater investment in public health facilities could expand the subsidized regime’s coverage, improve service quality, and attract more non- poor individuals to the contributory regime. Overall, the Dominican Republic’s fiscal policies are already solidly progressive. Going forward, the authorities will face important challenges in improving the quality and comprehensiveness of public service delivery while shielding poor households from the corresponding increase in revenue collection. While the Dominican Republic has successfully reduced vertical inequality through taxes, transfers and subsidies, these policies have had little impact on horizontal inequality. Some comparable countries have been able to achieve greater reductions in inequality by collecting large amounts of revenue and reinvesting it in social programs and public services. This suggests that enhancing the quality of public services could have a significant positive impact on the equity of Dominican fiscal policy, which would not only help to advance social development objectives, but would also build trust in public institutions and encourage formalization. WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only 1. Introduction In spite of sustained economic growth over the past two decades, the population in the Dominican Republic did not achieve significant welfare improvements until recently. Economic growth averaged 5.7 percent a year in 1991-2013, among the highest rates in the region. This performance enabled country’s GNI per capita (US$5,520 in 2012) to rise from 52 percent to 78 percent of the Latin America and the Caribbean (LAC) region’s average. From 2000 to 2013, a slight improvement in income inequality occurred, with the Gini index falling from 0.549 to 0.514. Disaggregation by area suggests that most of the inequality reduction took place in the rural parts of the country; inequality in urban areas did not decline significantly (World Bank, 2014a). After a sharp rise in the early 2000s, poverty rates shave been falling in recent years, and one possible explanation is that fiscal policy may not be redistributive enough. Based on the official poverty measurement methodology for the Dominican Republic (ONE and MEPyD, 2012), moderate poverty incidence soared from 32 percent in 2000 to almost 50 percent in 2004, a period that included a severe banking crisis. It then declined gradually to around 41 percent in 2013 and to about 35 percent by October 2014. Rapid poverty reduction in 20144, a year of 7.3 percent economic growth, has been attributed to rising wages, increased employment in school construction, public support to agriculture, credit to small and medium enterprises, and allocating more public investment to disadvantaged areas. At least until recently, the pace of poverty reduction has been slower in the Dominican Republic than in other countries with similar growth rates. Several studies have tried to explain the pre-2014 puzzle of slow poverty reduction at a time of rapid growth. Aristy (2016) analyzes whether the typical consumption basket for the poor differs significantly from that used to calculate the general consumer price index and the GDP deflator, but it does not find statistical distortions in the measure of poverty headcount. Other hypotheses include: (i) stagnant real wages (real earnings per hour of both self-employed and private-sector wage workers were about 27 percent lower in 2011 than in 2000) despite rising labor productivity (around 30 percent increase between 2000 and 2010, see Abdullaev and Estevao, 2013); (ii) the enclave nature of the economy, with activity in Special Economic Zones and tourist poles relatively isolated from the rest of the country; and (iii) the lack of redistributive capacity of the public sector (Carneiro et al., 2015). To explore the latter hypothesis, this study uses the Commitment to Equity (CEQ) methodology (Lustig and Higgins, 2013)5 to perform a fiscal-incidence analysis on the poverty and equity implications of the Dominican Republic’s fiscal system, including current taxes, subsidies, and overall public spending. The Dominican Republic’s tax policy has become more reliant on indirect taxes. Public revenues averaged 14.3 percent of GDP in 2004-14, with tax collections at 13.4 percent of GDP, below the LAC average.6 It is worth noting that the Government responded to a fall in fiscal revenues (partly related to declining trade taxes in the context of DR-CAFTA implementation) by adopting a total of six tax reforms between 2004 and 2012. Annex 8.2 describes in detail the main changes introduced by these different tax reforms. A country heavily dependent on indirect taxation, the Dominican Republic repeatedly increased VAT rates—from 12 percent to 16 percent (Law 288-04) and then to 18 percent (Law 253-12). This, together with the introduction of selective taxes on telecommunication services, have been the most far-reaching reforms. However, the tax bases have remained narrow, and extensive tax exemptions have persisted to erode the effective revenue base, since a large portion of the population (including both individuals and Special 4 According to ONE and MEPyD, poverty headcount index fell from 41.2% in 2013 to 35.8% in 2014. 5 Led by Nora Lustig since 2008, the Commitment to Equity (CEQ) project is an initiative of the Center for Inter- American Policy and Research (CIPR) and the Department of Economics, both at Tulane University, along with the Center for Global Development and the Inter-American Dialogue. The CEQ project is housed in the Commitment to Equity Institute at Tulane. For more details visit www.commitmentoequity.org. 6 When both tax and non-tax revenue are considered, the Dominican Republic trails only Guatemala for the lowest revenue level in Latin America, according to ECLAC Statistics. When social security contributions are excluded, Dominican Republic tax revenue is similar to the regional average. 1 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Economic Zones) have so far opposed an integral fiscal reform (World Bank, 2014b). Despite recent improvement, at 15.1 percent of GDP in 2014, fiscal revenues remain below their level in 2007 (16.6 percent). Revenue collection capacity is partly hampered by high levels of informality and existing tax exemptions, with tax expenditure amounting an estimate of 6.6 percent of GDP in 2014, including 3.2 percent of GDP in VAT exemptions (DGII, 2014). The Dominican Republic has made notable efforts to increase social spending. As mandated by law and demanded by the citizenry, public outlays for education doubled in recent years—from around 2.2 percent of GDP in 2011 to close to 4 percent in 2013. In a social security reform, some health services were privatized and lower income households began to receive insurance under a subsidized scheme. However, a large part of the population remains uninsured. In addition, subsidies on electricity (and technical and commercial losses) take a big toll on the public budget, equaling about 2 percent of GDP. Finally, a relatively large number of social assistance programs represent around 0.5 percent of GDP. A few existing fiscal incidence studies are relevant to the Dominican Republic: Santana and Rathe (1992),7 Lindert, Skoufias, and Shapiro (2006), and Barreix, Bès, and Roca (2009). Lindert et al. (2006) find low levels of social spending in the Dominication Republic. Their paper measures the extent to which social assistance and social security spending, consumption subsidies, and education and health spending favor the poor in eight LAC countries. For the Dominican Republic, the paper uses the National Survey on Living Conditions (ENCOVI) for 2004. At that time, the country had the lowest levels of social spending in the sample, and social insurance had negligible poverty impacts. The results reflect a combination of factors: (i) some programs had relatively low (net) unit subsides and weak targeting and coverage of the poor and vulnerable and (ii) social assistance programs like the school-based TAE transfer and school feeding ranked fairly high in terms of social welfare impact per dollar spent but were quite small in terms of budget and subsidy per person. The paper by Barreix et al. (2009) examines the impact of fiscal policy (social spending and taxation) on inequality, finding Dominican fiscal policy progressive in 2004. It is based on a collection of studies for Central America and the Dominican Republic written by various authors who followed a common methodology.8 The analysis uses ENCOVI 2004 and covers direct and indirect taxes; spending on education, health, and social assistance programs; and subsidies on electricity and gas. The study finds that fiscal policy in 2004 was progressive, and inequality was overall reduced thanks to a progressive social spending despite regressive tax system at that time. In addition, health and education spending was pro-poor, i.e. progressive in absolute terms.9 Some social assistance programs, like the general subsidies on electricity (Programa de Reducción de Apagones) and LPG gas that were in place prior to the shift to targeted subsidies in 2008 (Gallina et al, 2015), were progressive.10 In January 2013, a series of microsimulation exercises looked at the impact of selected fiscal policy tools on poverty and inequality; the results were mixed. The analysis found: (i) the tax reform of November 2012 (Law 253-12) had a neutral impact on poverty and inequality; (ii) the freezing of the lower exemption threshold on individual income taxes had a positive impact in terms of redistribution; and (iii) the VAT rate increases were regressive (MEPyD, 2013). A parallel microsimulation exercise showed that an RD$125 increase in the amount allocated to beneficiaries under the Comer es Primero conditional cash transfer (CCT) program would result in a 0.22 percent reduction in moderate poverty and a 0.0013 reduction in inequality (Gini index). Similarly, the expansion in the number of beneficiaries of the subsidized health regime would contribute to better equity outcomes. 7 This study used 1989 household income data and found “a degree of progressivity� in direct and indirect taxation (Chu, Davoodi, and Gupta, page 38). 8 For the Dominican Republic’s analysis, the background study was prepared by Díaz (2008). 9 A transfer will be progressive in absolute terms if the per capita amount received decreases as income rises (Lustig & Higgins, 2013). 10 Progressive in relative terms: subsidy increases as a percentage of income but per capita subsidy decline as income rises. 2 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only This World Bank study goes beyond previous exercises. It analyzes the impact of fiscal policy in 2013, using the CEQ methodology that includes several fiscal instruments and social programs targeting the poor (direct and indirect taxes, transfers, CCTs, public services in educations and health). Some taxes (like the CIT) and public spending categories (like some infrastructure and rural development items) are not included due to the difficulty of assessing their effects on the disposable income of citizens, specially the poor. The paper’s main contributions are: First, understanding how selected taxes and transfers programs affect income distribution in the Dominican Republic. Second, comparing the Dominican Republic’s results with a number of countries in which the Commitment to Equity methodology has been applied, including some with similar incomes per capita such as Costa Rica (Sauma & Trejos, 2014) and Peru (Jaramillo, 2013).11 Third, discussing a series of alternative fiscal scenarios that would help enhance the redistributive capacity of the Dominican state. 11 The common methodology is described in Lustig and Higgins (2013). 3 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only 2. Context: fiscal tools in the Dominican Republic This section discusses the fiscal instruments and public-spending programs of the Dominican Republic’s fiscal system in 2013 that are included in the analysis. On the revenue side, the Dominican Republic relies heavily on indirect taxation,12 although the two latest tax reforms have increased the importance of direct taxation. Direct taxation increased from 23.2 percent of total tax revenue (2.8 percent of GDP) in 2004 to 37 percent (5.1 percent of GDP) in 2013. Indirect taxes accounted for 63 percent of total tax revenues in 2013 (Table 2). The VAT tax’s productivity, measured as the percentage points of GDP collected for each percentage point of VAT rate, remains low at 0.26, especially considering the high consumption to GDP ratio, averaging 93.3 percent in 2008-12. VAT productivity in DR is almost half of average 0.626 in 2010 of Latin America (Pecho, et. al., 2012). Table 2. Composition of taxes in the Dominican Republic (2013) % of Included in Estimation % of Revenue total Analysis Method GDP taxes Total Revenue 14.5 Taxes 100.0 13.8 Direct Taxes 37.0 5.1 Direct Taxes on Individuals 9.4 1.3 On Wages & income on personal income Yes Simulation 5.7 0.8 On Dividends Yes Simulation 1.7 0.2 On Interest Yes Simulation 0.6 0.1 Other personal income tax No 1.3 0.2 Corporate Income Tax No 16.5 2.3 Other Direct Taxes No 11.1 1.5 VAT and Other Indirect Taxes 63.0 8.7 Simulation with assumptions of tax ITBIS (VAT) Yes 32.0 4.4 evasion and tax expenditures Excises on Alcoholic Beverages Yes Simulation 2.4 0.3 Excises on Beer Yes Simulation 2.7 0.4 Excises on Tobacco Yes Simulation 1.2 0.2 Excises on Oil Derivates Yes Simulation 12.2 1.7 Other Indirect Taxes No 12.4 1.7 Other Taxes No 0.0 0.0 Included in % of Contributions Analysis GDP Contributions to social security No 0.4 0.1 TOTAL 100.0 13.9 Source: Authors’ calculations based on Ministry of Finance data. Reliance on indirect taxation stems from the structure of the Dominican Republic’s tax regime. In part, indirect taxes reflect the modest roles played by personal income taxes (featuring a relatively high exemption threshold) and corporate income taxes (with many companies in Special Economic Zones benefiting from exemptions). Large existing exemptions result in low effective tax rates. For example, estimated VAT effective rate (VAT collection as a percentage of private consumption of National accounts) 12 See section 8.1 in the Annex. 4 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only is 2.7 percent compared with a general statutory rate of 18 percent. Tax revenues oscillate around 14 percent of GDP. The tax effort is considered low relative to lower-middle-income countries (Garza et al., 2012: 122- 123).13 The Dominican Republic’s revenue gap, measured as the difference between the current level of tax collections and the level that would result from achieving the tax effort prevailing in countries with the same income, moves around 7 percent of GDP (Garza et al., 2012). As depicted in Figure 6, a simple comparison with other countries shows that the Dominican Republic has lower indirect and direct tax revenues than some other countries with lower GNI per cap. Figure 6. Composition of taxes and GNI per capita: International comparison 30% 20,000 18,000 25% 16,000 GNI per capita PPP 14,000 Share of GDP 20% 12,000 15% 10,000 8,000 10% 6,000 5% 4,000 2,000 0% 0 Direct Tax 1/ Indirect Tax GNI per capita (2011 PPP, right axis) Source: Authors’ estimates and Lustig et al. (2013), CEQ Standard Indicators by a number of authors. Note: Direct Taxes are personal income tax and contributions to social security; indirect taxes include VAT, excises and other indirect taxes. Level of Government: Central Government for Ethiopia, El Salvador, Guatemala, Dominican Republic, Armenia, Sri Lanka; General Government: Bolivia, Peru, Costa Rica, Mexico, Uruguay, Brazil, South Africa, Indonesia. In 2013, four categories dominated the structure of expenditures. Social programs (8 percent of GDP) were the largest, followed by economic sectors14 (3.7 percent of GDP), central administration (2.7 percent of GDP), and debt servicing (4.9 percent of GDP). The Government significantly expanded social spending between 2011 and 2013, an increase mainly driven by education spending Table 3. On the other hand, economic spending (especially for transportation infrastructure) declined in 2013, partly reversing an upsurge in public spending that expanded the central government deficit to 6.6 percent of GDP in 2012, prompting fiscal consolidation (World Bank, 2015). It is worth mentioning again that this paper’s fiscal incidence analysis focuses on expenditure policies in 2013 and, more precisely, on social spending, where the Dominican Republic trails some other countries with lower GNI per capita. Expenditures on roads or telecommunications infrastructure (economic functions) also benefit the poor, but the incidence is harder to sort out. 13 Tax effort is defined by Pessino and Fenochietto (2010) as the ratio of the actual revenue that a country collects to its tax capacity. 14 Economic sectors, according to Functional Structure of expenditures, include Farming, Irrigation, Industry and Commerce, Transportation, Rural Roads, Urban Planning and Development, Energy and others. 5 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Table 3. Composition of expenditures in Dominican Republic (2011 and 2013) 2011 2013 Included In Analysis Estimation Method (SA2) (Benchmark & SA1) % of total Gov. % of total Gov. % of GDP % of GDP Total Government Spending (A+B) Spending Spending 100.0 18.0 100.0 20.2 A. Primary Government Spending (a+b+c) 74.2 13.3 76.0 15.3 a. Social Spending (excludes contrib 32.7 5.9 39.9 8.0 pensions) (1+2+3+4) 1.Total Cash Transfers 2.3 0.4 4.0 0.8 Cash Transfers (excluding all Yes Imputation 2.3 0.4 4.0 0.8 Pensions) Noncontributory Pensions N.A. N.A. 0.0 0.0 0.0 0.0 2.Total In-kind Transfers 20.9 3.7 27.8 5.6 Education Yes Imputation 10.5 1.9 18.7 3.8 of which Tertiary Yes Imputation 1.2 0.2 1.3 0.3 Health 10.4 1.9 9.1 1.8 Imputation using alternate survey Social Security Partially (only subsidized) 1.7 0.3 1.5 0.3 DHS Endesa 2013 Imputation using alternate survey Ministry of Health Yes 7.2 1.3 6.2 1.3 DHS Endesa 2013 Other (PROMESE and Imputation using alternate survey Partial 1.5 0.3 1.4 0.3 other ncp) DHS Endesa 2013 3.Housing and Urban 1/ No 4.3 0.8 3.4 0.7 4.Other Social Spending No 5.3 1.0 4.7 1.0 b. Contributory Pensions Yes Direct Identification 4.9 0.9 4.2 0.8 c. Non-Social Spending (1+2) 36.6 6.6 31.9 6.4 1.Indirect Subsidies 7.0 1.3 6.7 1.3 On Final Goods Yes Simulation 7.0 1.3 6.7 1.3 On Inputs No 0.0 0.0 0.0 0.0 2.Other Non-Social Spending No 29.6 5.3 25.2 5.1 Memo: 0.0 0.0 0.0 Debt Servicing No 25.8 4.6 24.0 4.8 Social Spending plus contributory pensions 38.2 6.8 44.1 8.9 Interest payments No 11.2 2.0 11.5 2.3 B. Amortization payments No 14.5 2.6 12.5 2.5 Source: Authors’ calculations based on Ministry of Finance data. Note: See Details in Annex 8.4. 6 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Figure 7. Social spending as a share of GDP: International comparison 16% 20,000 14% 18,000 16,000 12% 14,000 GNI per capita 10% Share of GDP 12,000 8% 10,000 6% 8,000 6,000 4% 4,000 2% 2,000 0% - Direct Transfer Education Health Subsidies GNI per capita (2011 PPP, right axis) Source: CEQ working papers (http://www.commitmentoequity.org), Tulane University and World Bank staff calculations. In the context of the Dominican Republic’s relatively limited capacity to collect revenue, social spending has traditionally been low by regional standards. In 2007-11, public expenditures averaged 2.3 percent of GDP on education and 1.6 percent of GDP on health, compared to regional averages of around 4.5 percent in each of these sectors.15 More than 40 percent of total budget of Ministry of Education and Health in 2014 was for wages and salaries and social contributions, according to statistics by the Ministry of Finance16. Starting in 2013, the Dominican Republic raised its budget allocations for education in response to popular demand to achieve the spending levels mandated by Education Law 66-97. Public expenditure in education increased from 2 percent of GDP in 2011 to 3.8 percent in 2013, a year in which around 60 percent of the additional resources allocated to education (around 1.1 percent of GDP compared to the previoues year) were used to improve and expand infrastructure. Most of the funding went to the primary education (Table 4).17 15 Dominican Republic SISDOM indicators and World Development Indicators. 16 40.2 percent for Ministry of Health and 45.2 percent for Ministry of Education. See: http://www.digepres.gob.do/wp-content/uploads/2015/04/Institucional-del-Gasto-2014.xlsx 17 The education system in the DR is divided into three levels: Inicial (Pre-School), Básica (Primary) and Media (Secondary). The Education Law granted that education in Inicial (from 3 to 6 years) and Básica (from 6 to 14 years) is compulsory and free. In the Dominican Republic, Básica has a duration of eight years, organized in two cycles of four years. The first six years of Básica are equivalent to primary school (from 6 to 11 years old), and the last two years are equivalent to lower secondary school. Additionally, Nivel Medio is equivalent to upper secondary school, and the last four years (from 14 to 17 years). Higher education in the DR includes Institutes of Specialties, Technical Institutes of Superior Studies, and Universities. In this analysis we include Universidad Autónoma de Santo Domingo, which concentrates most of the public spending. The UNESCO International Standard Classification of Education supports the division of Básica into primary and lower secondary levels. See: http://www.uis.unesco.org/Education/ISCEDMappings/Documents/Latin%20America%20and%20the%20Caribbean /Rep_Dominicana_ISCED_mapping.xls. 7 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Table 4. Dominican Republic: Composition of public education expenditure (2013) % GDP % of GDP Spending Component In analysis Education 3.8% 3.5% Pre-School (From 3 to 5 years old) 0.2% 0.2% Primary (From 6 to 11 years old, 1st to 6th Básico) 1.8% 1.8% Lower Secondary (12 to 13 years, 7th and 8th Básico) 0.5% 0.5% Upper Secondary (14 to 17 years, 1st to 4th Medio) 0.8% 0.8% Tertiary 0.3% 0.3% Other expenses in education 0.2% 0.0% Source: Ministry of Education and Ministry of Finance. Note: Levels of education in this table are equivalent to CINE categories. Looking at education, the country saw some improvements in terms of coverage in initial education, whereas high school dropout is observed. According to the Ministry of Education, net enrollment in initial education increased from 35.5 percent in 2006-2007 to 44 percent in 2012-2013, whereas net enrollment in basic education has been oscillating around 92 percent in recent years. The low level of resources for public education in previous years has led to an inadequate supply of education services, including a lack of classrooms and teachers. In this setting, high repetition and low completion rates plague the Dominican Republic’s education system. As of 2013, only 25.8 percent and 13.9 percent of the population had completed secondary and tertiary education, respectively. For example, according to Minerd (2014), in 2012-2013 over-aged students represented 12.2 percent and 18.4 percent for Básico and Medio school level; promotion rates in Básico level were 88.8 percent for public schools compared with 96.1 percent in private schools. Adult literacy has been progressively improving, from 88 percent of total population in 2007 to 91 percent in 2013, and the implementation of a new national literacy plan (Quisqueya Aprende Contigo) started in 2014. A multi-stakeholder initiative to improve the quality of education and an Education Pact have been launched. On the quality side, international LERCE and SERCE scores suggest that the Dominican Republic is performing worse than countries at a similar stage of development (Luque et al., 2010). The quality challenges in public education have probably fueled a tendency to opt out and choose private education, even among households in the lower income quintiles (Sánchez-Martín and Senderowitsch, 2012). In 2013, 73 percent of children were enrolled in public schools, 24 percent in private schools, and 2.5 percent in semi-official private schools that operate with government funding. This will be taken into account when assessing the impact of the increase in public education spending on poverty and inequality. To improve the outcomes of public education, public and private actors, including some from civil society, engaged in a 2012 dialogue in the context of the Dominican Initiative for a Quality Education (IDEC). In addition, a society- wide Education Pact was signed in spring 2014. On health, human-development indicators point to gaps in access and coverage as well as some challenges in outcomes. For example, under age 5 mortality rates remain relatively high for a middle- income country—at 31 per 1,000 people as of 2015. Although the percentage of births attended by skilled personnel is near the regional average, maternal mortality rates are worrisome, stuck at around 100 per 100,000 live births since 2005. The health system has been subject to many changes in the past decade and a half. The Law of Health (42-01) and the Social Security Law (87-01), both enacted in 2001, created the National Health System and Dominican social security system. This legal framework divided the functions of the National Health 8 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only System—steering, service provision, assurance, and financing—among the different institutions that comprise it. These include the Ministry of Public Health, National Service of Health (SNS), Bureau of Occupational Health and Safety (SISALRIL), the Social Security Treasury (TSS), the National Health Insurance Authority (SENASA), and the Bureau of Consumer Information and Protection (DIDA) (PAHO, 2007). At 1.8 percent of GDP in 2013, public expenditures in health were lower in the Dominican Republic than in other countries with similar income levels (Figure 7). According to our calculations, Ministry of Health expenditures are close to 70 percent of public outlays on health.18 We classify this expenditure in five categories: social security institutions, Ministry of Health, Program of Essential Drugs, logistics support (PROMESE-CAL), and others. Table 5. Dominican Republic: Composition of public health expenditure (2013) % of % GDP Spending Component GDP in analysis Health 1.8% 1.6% Ministry of Public Health 1.3% 1.3% Outpatient services /b 0.3% 0.3% Hospitals /b 0.9% 0.9% Social Security System 0.3% 0.3% Subsidized Regime Social Security /c 0.2% 0.2% Dominican Institute for Social Security (IDSS) /d 0.1% 0.1% Retired (SENASA) /c 0.0% 0.0% Others 0.2% 0.1% PROMESE 2012 /a,e 0.1% 0.1% Others: Military and Police Hospital, National VIH Commision 0.2% 0.0% (CONAVIHSIDA), Health reform commission (CERS) /a,e Sources: a/Informe Nacional de Gasto en Salud 2013; b/Authors’ calculations based on Informe Nacional de Gasto en Salud 2011, 2012 y 2013; c/CNSS, Informe a Diciembre 2013; d/Senasa (2014), “Reconversión del IDSS y Red Pública única,� Mimeo; e/Ministry of Finance. Despite low health expenditures by international standards, the Dominican Republic has made important progress in expanding health-insurance coverage. According to CESDEM (2014; p.361), 55 percent of the population was covered by health insurance in 2013, a doubling since 2007 (27 percent). This expansion likely benefited poor households because the number of enrolled in the subsidized regime tripled from nearly 1 million in 2007 to 2.8 million in 2013. Despite the growth of government programs, out-of-pocket spending still accounts for more than half of Dominican health expenditures. The Dominican social security system has three insurance regimes: (i) Familiar Health Insurance, (ii) Elder, Incapacity, and Survival Insurance, and (iii) Occupational Risk Insurance. The first two regimes include three schemes: contributive, subsidized, and contributive-subsidized (still not operational).19 Although according to official figures the subsidized regime (targeting the poor) covered in 2013 28.1 percent of the total population, compared to just 10.6 percent in 2007, beneficiaries still 18 These calculations are based on data from the National Accounts on Health (Ministry of Health, 2014) and Ministry of Finance. 19 In the case of health insurance, contributive coverage is contracted through Health Risk Administrations (ARS). One of the ARS is the public insurer SENASA, which also is responsible for covering all civil servants, employees of autonomous or decentralized institutions, and their family members. SENASA is also responsible for managing the health risks for the low-income population covered by the subsidized regime, identified through the National System of Beneficiaries (SIUBEN). 9 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only face large out-of-pocket spending (see IADB and World Bank, 2014). Private out-of-pocket spending (for medicines, up-front payments for medical analysis, check-ups, etc.) amounts to 56 percent of total health expenditures, representing a heavy burden on the poorer strata of the population. Finally, people who lack any type of health insurance use the Ministry of Health facilities (hospitals, clinics, and health centers) in emergency situations, resulting in health spending not initially budgeted. Workers, employers and government share the burden of health insurance in the Dominican Republic. The health-insurance schemes are linked to the three regimes foreseen in Law 87-01 on the social security system. The subsidized regime, financed by the Dominican state, covers the self-employed earning incomes below the national minimum wage, the unemployed, disabled, and the extreme poor. The contributive regime covers formal workers and features contributions by employees and employers. Finally, the contributive-subsidized regime is designed to cover professionals and self-employed with incomes above the minimum salary, financed with worker contributions and a state subsidy. For health insurance, both the subsidized and contributive regimes are fully operational; for pensions, implementation of the contributive regime has commenced recently. The new pension system is based on individual capitalization accounts paid by employees and employers, and its subsidized regime (for poor and unemployed) is not operational yet. Since payments under the new pension schemes were not operational in 2013, this does not have significant fiscal implications that year and, thus, has not been simulated in this CEQ analysis. It is also worth noting that the Ministry of Finance has assumed the cost of pensions under a previous system, the Dominican Institute of Social Security (IDSS). As of December 2013, this scheme still had 100,927 beneficiaries, with disbursements of RD$11,079 million (0.45 percent of GDP). In the past decade or so, reforms have improved targeting of the Dominican Republic’s social safety net. The banking crisis of 2004 resulted in a sharp economic slowdown that plunged around 1.3 million Dominicans into poverty. At that time, most existing social-transfer programs were untargeted, and they were not sufficient to shield the vulnerable population from major shocks. The episode triggered a social safety net reform. The so-called Solidaridad program was established as a targeted cash-transfer scheme in 2005. The main components are Comer es Primero, a transfer for purchasing food and other basic goods, provided the household complies with health-care obligations, and Incentivo a la Asistencia Escolar (ILAE), contingent on school attendance. In addition, with the support of the international community, conditionality was introduced to this program in 2009. The Government also introduced the Single Beneficiary Selection System (SIUBEN) as a targeting mechanism based on a life-quality index, created an independent Administrator for Social Subsidies (ADESS), and eliminated some existing untargeted programs. Reforms have also improved the focus of the country’s subsidy programs for household utilities. Facing a deteriorating fiscal situation in the wake of the global increase in oil prices in 2007-08, Dominican authorities decided to phase out the generalized LPG subsidies and a geographically based electricity subsidy, introducing targeted programs using the SIUBEN system.20 The new programs, Bonogas Hogar and Bono Luz, would cover not only the poor population but also part of the middle class. They resulted in fiscal savings amounting to an estimated 0.5 percent of GDP, compared with the previous generalized subsidy schemes for gas and electricity (Gallina et al., 2015). As a result, a large number of social programs now feature direct transfers to beneficiaries. The direct transfers can be classified in four categories (Table 6). First, Comer es Primero supports the consumption of a In 2009, presidential Decree No. 421-09 eliminated the geographically based electricity subsidy (Programa de Reducción de 20 Apagones) and introduced Bono Luz, a targeted electricity subsidy. 10 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only basic basket of food products.21 Second, a series of CCT programs encourages education: Incentive for School Attendance22 (Incentivo a la Asistencia Escolar, ILAE) to promote in Básica attendance and reduce dropouts; Bonus for Progress in Studying (Bono Escolar Estudiando Progreso, BEEP); and Incentive to Tertiary Education (Incentivo a la Educación Superior, IES), a transfer to poor university students. Third, a series of targeted transfers subsidize certain utilities for the poor and middle class: LPG (Bono Gas Hogar), electricity (Bono Luz), and gas bonus to drivers (Bono Gas Choferes, BGC). Fourth, others transfers are designed to support police officers (Programa de Incentivo a la Policía Preventiva, PIPP) and marine officials (Programa de Incentivo a los Alistados de la Marina de Guerra, PIAMG); in addition, some transfer schemes from the old social security system remain in place. Table 6. Direct transfers programs in Dominican Republic in 2013 Categories of Direct Transfers Programs #Beneficiaries % of 2013 GDP CCT food program Comer es Primero (CEP) 698,196 0.24% Education CCT programs Incentivo a la Asistencia Escolar (ILAE) 299,111 0.03% Bono Estudiando Progreso (BEEP) 45,982 0.01% Incentivo a la Educación Superior (IES) 25,795 0.01% Targeted Non-CCT transfers on Gas bonus to households: Bono Gas Hogar (BGH) 843,439 0.08% utilities and commodities Electricity: Bonus Bono Luz (BL) 533,766 0.09% Gasoline bonus to public transport drivers: Bono Gas 15,726 0.03% Choferes (BGC) Other transfers Incentive to preventive police (PIPP) 22,493 0.01% Marine officials (PIAMG) - 0.00% Contributive pensions from old regime 99,802 0.01% Source: ADESS and authors’ calculations. Finally, it is worth noting a series of implicit subsidies on electricity, some informal in nature. Subsidies on electricity are in the form of below-cost tariffs, non-invoiced provision, and payments that are not enforced. In areas with high incidence of poverty, irregular connections are common. Annual losses are financed by the central government. There are similar indirect subsidies in the case of water supply, but they are not subject to analysis in this report. The public sector still pays for the remnants of defunct programs. For electricity, an explicit subsidy in the tariff occurs for consumption below 700 kilowatt hours (kwh) per month23. This was initiated in the context of the Blackout Reduction Program (PRA), which charged a minimum fee for the provision of electricity to those households located in urban areas where fraud and non-payment was once pervasive and programmed blackouts were frequent. Blackouts have been reduced, and the program has been formally phased-out, but some users are often still charged the minimum fee because of a lack of meters in some neighborhoods. The Corporación Dominicana de Empresas Eléctricas Estatales, the public electricity transmission and distribution holding, estimates that technical losses and fraud account for 35 percent of total of electricity consumption on the public grid. Commercial losses in distribution companies are compensated through hefty central government transfers, which amounted to around US$850 million, or 1.6 percent of GDP, in 2013. Finally, electricity-sector inefficiencies have been a major drag on other public spending that could benefit the poor. For example, the electricity sector deficit in 2012 (1.8 percent of GDP) represented three times the amount spent on social subsidies and CCTs, the entire health budget (excluding social security), and two-thirds of the amount devoted to education. 21 To each family that has fulfilled the following co-responsibilities: attendance by the head of the household or spouse at training events related to better nutrition and health care and attendance at a health unit by children between birth and 5 years of age to undergo growth and development assessments and immunization. 22 This monthly transfer is granted to the household head or spouse for up to four children, with ages between 6 and 16 years old, enrolled from first through eighth grade at the Básica level. 23 Also there is an explicit subsidy for consumption above 700 kwh/month, as the tariff is not adjusted to the ‘technical tariff’ and the gap is covered by the national budget. 11 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only 12 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only 3. Methodology and sources of information 3.1. CEQ methodology This study’s goal is to estimate the impact of taxes and transfers on income inequality and poverty in the Dominican Republic. We use the CEQ methodology, applying the fiscal incidence analysis described in Lustig and Higgins (2013). This starts with the individual’s market income and adds transfers and subtracts taxes in different stages (Figure 8). Figure 8. Income concepts used in fiscal incidence analysis Market Income Wages and salaries, income from capital, private transfers; TRANSFERS before government taxes, social security contributions and TAXES transfers; benchmark (sensitivity analysis 1) includes (does not include) contributory pensions - Direct taxes (personal income taxes) Net Market Income Direct transfers (CCT + transfers, food transfers, scholarships, etc.) Disposable Income Indirect subsidies (electricity) + - Indirect taxes (ITBIS, excise taxes on beverages, tobacco, oil) Post-Fiscal Income Co-payments, user fees In-kind transfers (free or subsidized government + - services in education and health) Final Income Source: Lustig & Higgins (2013). Market income is a measure of pre-tax income that does not include the effects of government policies. It is composed of pre-tax wages, salaries, self-employed income, income from capital (dividends, 13 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only interest, and rent), and pensions. It is worth mentioning that the question asked in household survey ENIGH 2007 is about labor income gross of taxes. We estimate three scenarios. The difference between the Benchmark and Sensitivity Analysis 2 scenarios is that, in order to estimate the impact of the significant increase in public education expenditures in 2013, an alternative Sensitivity Analysis 2 featuring the lower expenditure level of 2011 is built. Since there is no theoretical consensus on whether contributory pensions are part of the market income or a government transfer, in the scenario Sensitivity Analysis 1 does not include public pensions in market income, making them instead a transfer contained in disposable income, in contrast with Benchmark and Sensitivity Analysis 2, in which contributory pensions are consider to be part of market income. Net market income subtracts direct taxes. Personal income taxes on wages, dividends, and interest are included in the analysis. The Dominican Republic’s old public-pension system was privatized, so social security contributions are not included as direct taxes. Disposable income adds direct cash and food transfers to net market income. As explained in the previous section, we include CCTs for nutrition and education, non-conditional cash transfers, goods transfers like food, shoes, uniforms, and backpacks, and the alphabetization program (Quisqueya Aprende Contigo). Post-fiscal income adds implicit subsidies on electricity and subtracts indirect taxes. These levies include the Tax on the Transfer of Industrialized Goods and Services (ITBIS), a value-added tax applied on domestic and imported good and services, or VAT, and excises on alcoholic beverages, beer, tobacco, and oil derivatives. Final income includes in-kind transfers. These are measured by the monetized value of public expenditures in health (Ministry of Health, social security and others) and education (pre-school, primary, lower secondary, upper secondary, and tertiary). It is important to take into consideration that contributive health insurance is not included in the analysis, since it works de facto as a private insurance. 3.2. Data sources This fiscal-incidence analysis uses several sources of information. The main one is the National Survey of Household Income and Expenditure 2006-07 (ENIGH). This survey was collected by the National Office of Statistics (ONE) between January 2007 and January 2008 for 22,000 households and 80,131 individuals. It is representative at the national level and for four main domains: Metropolitan or Ozama, North or Cibao, South and East. ENIGH contains data on income, expenditures, auto-consumption, remittances, and use of educational services. There have been major changes in health coverage. To account for them, we complement ENIGH with the Demographic and Health Survey (ENDESA 2013). This survey has a nationally representative sample of 11,464 households, 9,372 women ages 15-49, and 10,306 men ages 15-59. The study relied largely on official sources of information. Data on government revenues were obtained from the General Directorate for Internal Taxation and the Ministry of Finance. Data on direct transfers come from ADESS, the Ministry of Finance, and the Ministry of Education. Information on electricity subsidies was facilitated by the Ministry of Finance. Finally, data on public health expenditures were obtained from the Ministry of Finance, the Ministry of Health, and SENASA. 14 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only 3.3. Main assumptions It will be important to make some adjustments to account for the policy changes of recent years. Compared to other countries studies with the CEQ methodology, the Dominican Republic is especially challenging because the “departure point,� the most recent household income and expenditure survey, dates to 2007. It is necessary to consider that numerous policy decisions were adopted between 2007 and 2013, including the modification of the rates and bases of the main taxes (e.g., ITBIS, ISR, ISC). Furthermore, there has been a notable expansion in the coverage of direct transfers (e.g., CEP, BGH, BGCH), and the value of certain in-kind transfers, such as education, has been expanded. In the light of these changes, the methodology applied the tax and public expenditure structures of 2013 to ENIGH 2007. On the tax side, rates and definitions of the 2013 tax base were used. On the expenditure side, the value of the 2013 peso was deflated by the change in the consumer price index (CPI) between 2007 and 2013. In other words, the public revenues and spending vectors of 2013 were used to calculate income poverty—but in 2007 prices. Expenditures were adjusted only for inflation and not by GDP growth. This is because the majority of the recorded public-spending variations were below the growth rate during the period. Overall, the objective was to adapt the CEQ methodology’s various definitions of income using the ENIGH 2007 and the public revenue and expenditure structure of 2013, expressed in 2007 prices. We opted for this alternative (instead of inflating to 2013 the variables of the ENIGH 2007) because, besides inflation between 2007 and 2013, relative prices of production factors, structure of employment and size of households in Dominican Republic could have experimented important changes in income distribution, that we otherwise would not have been able to replicate with available information. The adjustment factor was 42.5 percent, i.e. inflation between June 2007, date of the survey, and December 2013. It is worth noting that the following analysis only evaluates the tax system along one dimension—its impact on equity. It does not assess other important features of a tax system, such as its efficiency—which measures the amount collected given the rate— buoyancy (i.e. response of tax collections to economic growth), simplicity, and ease of administration. An estimation of direct taxes was made by applying statutory rates and income brackets from 2012 (in 2007 prices) to the salaries and wages declared in ENIGH 2007. Individuals have to pay direct taxes out of market income. Because income tax payments in 2013 were made taking into consideration income from 2012, we deflate from 2012 to 2007 prices. Due to the fact that income brackets were adjusted by inflation from 2008 to 2012, mismatch between effective income brackets is expected to be minimal. As pointed out by Dominican authorities, tax evasion among the self-employed is considered significant, while we were unable to access to profiles of payments of independent business or official estimations of evasion; thus, so we do not calculate personal income taxes for those groups. In addition, we do not use assumptions on informality of wage earners or other assumptions on tax evasion on personal income tax. In order to ensure incidence analysis is not detached from reality due to assumptions, we contrasted simulated collections applying statutory tax rates and actual collections, and discussed results with the tax authority in the Dominican Republic to ensure consistency. The personal income tax is levied on individuals with income above the exemption threshold. The system uses three rates that rise with tax brackets: 15 percent, 20 percent, and 25 percent. Dividends and interest income are taxed at 10 percent. It is assumed that informal self-employed workers do not pay income taxes. The corporate income tax is also not included in the analysis. Two caveats apply: (i) using statutory rates does not measure taxes actually paid and (ii) even if the survey’s simulated total income tax payment is similar to actual collection, the incidence by quintile could be over or under the estimated values. We assume 15 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only the household survey includes labor income gross of taxes, because ENIGH 2007 survey asks for gross salary without deductions (see details in annex 8.3). Indirect taxes were estimated using the simulation method. We include ITBIS, excises, a tax on telecommunications, and the insurance tax. ENIGH 2007 has a detailed list of household purchases of goods and services, categorized according to the Classification of Individual Consumption According to Purpose (COICOP).24 We separate each good or service into one of three groups: (i) those exempt in 2007 and 2013, (ii) those exempt in 2007 but not in 2013, and (iii) those taxable by both ITBIS and excises. Within ITBIS, it was necessary to distinguish between goods that were and were not exempt. To avoid overestimating the taxes paid by low income earners, we decided, after discussion with authorities, to include tax evasion in all scenarios—a practice that follows previous CEQ papers. We incorporated the assumption of tax evasion by creating four groups of goods and services: (i) high propensity for evasion; (ii) high propensity to pay ITBIS; (iii) products with estimated compliance rates, according to the General Directorate for Internal Taxation;25 and (iv) products on which the VAT was paid as a condition of purchase.26 Indirect taxes were down-scaled to prevent overestimation, using the method in Lustig and Higgins (2013). For example, we adjust VAT payments to equalize the ratio of total VAT to disposable income in the survey to the ratio of VAT collection to private consumption in the national accounts in 2013. Also, we take into account exemptions and reduced rates on each kind of good and services according to statutory rates. Direct transfers received were assigned if the household fell into a SIUBEN category that indicates eligibility for each program—e.g., categories “poor� 1 and “poor2� in the case of Comer es Primero. Ultimately, beneficiaries were randomly selected as a sub-group of the household, based on coverage statistics. A series of steps were taken: (i) adjust the population of ADESS beneficiaries in 2013, taking into consideration the variation in the population between 2007 and 2013; (ii) calculate transfers at 2007 prices; (iii) adjust the coverage in terms of SIUBEN categories to reproduce the number of beneficiaries and coverage as a percent of the population. When the household survey and the national accounts differed on the ratio of direct transfers to national income, we down-scaled the value of the transfer to make the ratios comparable. Other transfers, like those on shoes, uniforms, and backpacks, plus the alphabetization program, were imputed using average costs estimated by the Ministry of Education and UNICEF—once again, 2013 values adjusted to 2007 prices. Implicit electricity transfers were calculated by applying existing tariffs. Using 2007 prices, we estimated the implicit kwh consumed by each household and applied the subsidy to users consuming less than 700 kwh a month. For those in the ENIGH survey who consume electricity but declare not to pay the bill, an implicitly standard subsidy is calculated. Education benefits depend on the number of students and the average cost of education. The survey identifies individuals who attend school, their levels of education, and whether the schools are private or public. The education benefit is based on the cost per student by level, estimated by UNESCO and the Dominican Republic Ministry of Education. We adjust these figures to 2007 prices. Following Lustig and Higgins (2013), we prevent overestimation by adjusting the ratio of education expenditures to disposable income, making it equal the ratio calculated using national accounts. An alternative analysis examines the impact of larger budget for public education. To account for the significant increase in public education expenditures in 2013, from 1.9 percent of GDP in 2011 to 3.8 of 24 See http://unstats.un.org/unsd/cr/registry/regcst.asp?Cl=5. 25 DGII provided a list with estimated compliance rates for VAT payments. 26 We estimated a detailed list of goods and services according to these assumptions. 16 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only GDP in 2013, we estimated the alternative Sensitivity Analysis 2, featuring the lower expenditure level of 2011. Because gross coverage rates did not significantly change in primary schools and changed little in elementary and secondary schools between 2007 through 2013, the different scenarios assume coverage did not change.27 Finally, we account for in-kind health transfers by estimating the impact of the subsidized social security regime only, which is free for the poor and vulnerable, and not the contributory regime, which works as a private insurance.28 We use the Demographics and Health Survey (ENDESA 2013) to determine whether individuals with health insurance belong in social security's subsidized regime. For the uninsured, we identify only those who use the services of public hospitals or ambulatory centers. It is also possible to identify those who are insured by the Dominican Institute of Social Security (IDSS). Finally, public spending under the Essential Medicines Program (PROMESE) is also computed; this includes spending to purchase medicines and medical supplies for public health institutions as well as the distribution of subsidized medicines to the population. Drawing from information in the ENDESA 2013 survey, we use matching-score analysis to identify beneficiaries in the ENIGH 2007 survey. For beneficiaries of the subsidized regime, we impute an insurance value based on the average transfer by insured (per capita) from the government to SENASA. For IDSS affiliates, we estimated an average insurance value by dividing the government transfer by the total number of insured. For the uninsured who report using public facilities, we impute an average cost per user at hospital and ambulatory centers. It is estimated by dividing total expenditure on each level of health services from National Health Accounts (Ministry of Health, 2013) by users of health public services in the survey, identified using matching-score analysis from ENDESA 2013. For PROMESE, once we selected the beneficiaries of this program, we estimate an average benefit by dividing the program’s expenditures in 2013 by the number of users reported in ENDESA 2013. As with education, the ratio of health expenditure to disposable income under the survey is adjusted to match the ratio calculated using national accounts. In sum, counting with a dated household survey in the Dominican Republic implied a number of additional assumptions when applying the CEQ methodology. Overall, the validity of results depends on the fact that changes in income distribution between 2007 and 2013 have been observed but are not dramatic (e.g. a decline in GINI from 0.487 to 0.471, according to World Development Indicators); this is the most relevant caveat in our analysis. In the case of education, since no significant change in enrollment is observed between 2007 and 2013 (except for pre-primary education), and given that the team accessed official data detailing the cost of delivery of education services, we are confident that incidence analysis for this sector is relatively precise. In the case of health services, having counted with ENDESA 2013, a specialized survey collected during the year of analysis that details information on the insurance beneficiaries and effective use of health services by income level, helps ensuring the robustness of results. In addition, a matching scores technique has been applied, and results should be thus as robust as those in other CEQ exercises using a specialized health survey. With respect to conditional cash transfers, a careful revision of the indicators was performed to ensure consistency with actual population coverage, transfers per capita, and budget for the different programs in 2013. In the case of indirect electricity subsidies, results should be interpreted with caution, since administrative registries do not adequately identify beneficiaries, and the analysis was performed on the basis of a profile of beneficiaries described by authorities of the sector. 27 The rise in spending mostly went for construction and repairing classrooms, extension of school hours from five to eight, higher salaries for teachers, and hiring new teachers. 28 The contributive system is actuarially fair. In the case of the subsidized regime, workers do not make contributions. This regime, financed by the Dominican state, covers the self-employed, disabled, and the extreme poor (as defined by the national poverty line). 17 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Some mitigation measures on potential caveats include the use of additional sources of information to the household survey, discussions with authorities, and revision of results by the developers of the CEQ methodology. Discussions have been hold with authorities to ensure results are consistent with existing evidence and knowledge. This includes discussions with the General Directorate for Internal Taxation, the Ministry of Finance and the Electricity Distribution Holding (CDEEE), the Social Cabinet and the ADESS, the Ministry of Education, the Ministry of Health, and SENASA, to ensure the accuracy of the information and results relating the incidence analysis for direct and indirect taxes, indirect electricity transfers, direct transfers, and health spending and social security, respectively. Finally, estimations have gone through two thorough review rounds by Tulane University, to verify results, correct for mistakes, and ensure the consistency with CEQ methodology (Lustig and Higgins, 2013) and the comparability to similar analyses. 18 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only 4. Main results As a departure point for the fiscal incidence analysis, population and income shares in total market income by socioeconomic group are presented. As illustrated in the table, the 5.7 percent of total population lives below US$1.25 ppp a day, and has a share of only 0.5 percent of total market income. Around 19.5 percent of the population in 2013 lived below US$ 2.5 ppp at 2005 prices. The poor totals about 37 percent of the population, whereas 40 percent of the population remains vulnerable according to the World Bank definition used in the Middle Class flagship for Latin America of 2013. Table 7. Benchmark scenario: Population and Income shares of market income Group % Population % Income Ultra Poor (y < 1.25) 5.7% 0.5% Extreme Poor (1.25 < = y < 2.50) 13.8% 3.1% Moderate Poor (2.50 <= y < 4.00) 17.4% 6.6% Vulnerable Poor (4.00 <= y < 10.00) 40.0% 29.6% Middle Class (10.00 <= y < 50.00) 21.6% 46.6% Upper Class (50.00 <= y) 1.4% 13.6% Total 100.0% 100.0% Source: Authors’ estimates based on ENIGH 2007. Note: income definition is USD PPP at 2005 prices. 4.1. The re-distributional impact of taxes The Dominican Republic imposes a variety of taxes that affect final income under the CEQ analysis. As previously mentioned, the country depended on indirect taxes for 63 percent of total tax revenues (8.8 percent of GDP) in 2013.29 The most important sources were the ITBIS (4.4 percent of GDP), a value-added tax on the transfer of industrialized goods and services, and the excise tax on oil derivatives (1.7 percent of GDP). Excise taxes on alcoholic beverages, beer, and tobacco added to 0.9 percent of GDP. Direct taxes only amounted to 5.2 percent of GDP. Corporate income taxes (2.4 percent of GDP) were the principal direct tax. Taxes on wages and personal income represented 1.3 percent of GDP and other direct taxes, including property taxes and taxes on lottery, accounted for 1.5 percent of GDP. According to the results of the CEQ analysis, and using the Lorenz curves estimates, both direct and indirect taxes appear to be progressive.30 As shown in Figure 9, the concentration curves for direct and indirect taxes lie below the Lorenz curve for market income. As expected, direct taxes are much more progressive than indirect taxes. 29 This figure includes taxes on imported goods, which are not included in the incidence analysis on poverty and income distribution. 30 A tax is everywhere progressive (regressive) if its concentration curve lies everywhere below (above) the market income Lorenz curve. 19 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Figure 9. Progressivity of direct and indirect taxes: concentration curves and Lorenz curve for market income 100% 80% 60% 40% 20% 0% 0% 20% 40% 60% 80% 100% Market Income Indirect taxes Population Direct taxes Source: Authors’ estimates based on ENIGH 2007. Direct taxes only have a significant average incidence on the market income of individuals in the middle and upper classes, although it is perhaps smaller than what might be expected (Table 8).31 Direct taxes reduce the market income of the upper class (per capita income above US$50 PPP a day) by 4.1 percent. Indirect taxes reduce the market income of the total population, but the incidence is progressive in absolute terms. The market income of the ultra poor is reduced 4.7 percent, while the upper classes’ income is reduced by 10.4 percent. This is explained by the higher levels of consumption by the upper class, especially on goods that are outside the basic consumption basket (currently exempt). Table 8. Benchmark scenario: Incidence of direct and indirect taxes by socioeconomic group (% of market income) Direct taxes Indirect taxes Ultra Poor (<1.25 USD PPP) 0.0% 4.7% Extreme Poor (1.25-2.5 USD PPP) 0.0% 5.4% Moderate Poor (2.5-4 USD PPP) 0.0% 5.4% Vulnerable (4-10 USD PPP) 0.0% 6.3% Middle Class (10-50 USD PPP) 1.6% 7.8% Upper Class (>50 USD PPP) 4.1% 10.4% 1.3% 7.5% Source: Authors’ estimates based on ENIGH 2007. Note:_income definition is USD PPP at 2005 prices 4.1.1. Direct taxes Direct taxes (i.e., taxes on wages and personal income, interest income, and dividends) are found to be progressive (Figure 10). They represent 1.3 percent of total market income. Concentration shares show 31For Benchmark and Sensitivity Analysis 2, the results are the same and for Sensitivity Analysis 1 are very similar. For this reason, we include only the Benchmark results. 20 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only that the top decile of the population pays 92 percent of direct taxes32, while it receives 40.5 percent of total market income. Direct taxes decrease market income 3 percent for the top decile; they only decrease the market income of the seventh decile by 0.1 percent. In terms of socioeconomic groups, middle-class households (per capita income between US$10 and US$50 a day) pay 56.3 percent of direct taxes, and the richest (above US$50 a day per capita income) pay 42.5 percent. It is important to take into account that the middle class accounts for 21.6 percent of total population and 46.6 percent of market income. Meanwhile, the richest group represents 1.4 percent of population and 13.6 percent of market income. This means that the relative tax burden is much higher among the rich. Figure 10. Progressivity of direct taxes: Concentration curves and Lorenz curve for market income 100.0% 80.0% 60.0% 40.0% 20.0% 0.0% 0.0% 20.0% 40.0% 60.0% 80.0% 100.0% Market Income Personal Income Population Interest Dividends Source: Authors’ estimates based on ENIGH 2007. Personal income taxes—which account for 90.6 percent of the direct taxes in the analysis—are highly progressive in the Dominican Republic. These taxes. Personal income taxes reduce the market income of the top decile by 2.75 percent and the ninth decile by 0.46 percent. In terms of socioeconomic groups, personal income taxes reduce the average market income of the middle class by 1.5 percent and the richest segment of the population by 3.6 percent. The middle class represent 58.3 percent of total personal income tax payments and the highest-income group 41.6 percent (Figure 11). It is worth noting that the mean dividend tax in upper class is higher than middle class but, since the second group has more individuals, share of tax paid by the middle class over total collections is larger. In addition there could be some under reporting of income dividends in the household survey by high income individuals. Figure 11. Direct taxes concentration shares per socioeconomic groups 32Refers exclusively to direct taxes analyzed in the study paid by individuals (i.e. taxes on personal income, interests, and dividends) 21 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only 100% 80% 60% 40% 20% 0% Personal Income tax Interest tax Dividend tax Extreme Poor (y <2.5) Moderate Poor (2.550) Source: Authors’ estimates based on ENIGH 2007. Note:_y means income; for example, y<2.5 means income lower than 2.5 USD PPP at 2005 prices. The tax on interest income affects the middle and upper socioeconomic groups. Established by the November 2012 tax reform, this tax represents 7.8 percent of total direct tax revenues. It reduces the market income of the population by 0.09 percent. The top decile’s income is reduced by 0.2 percent due to the 10 percent tax on interest earnings. The middle class pays 27.6 percent and of the total interest tax and the upper class 65.9 percent. In terms of socioeconomic groups, the data show that some people within the vulnerable population are paying tax on interest, resulting in a 0.02 percent reduction of their market income. Table 9. Benchmark scenario: Incidence of personal income, interest, and dividend taxes by socioeconomic group (% of Market income) Personal Income tax Interest tax Dividend tax Ultra Poor (<1.25 USD PPP) 0.00% 0.00% 0.00% Extreme Poor (1.25-2.5 USD PPP) 0.00% 0.01% 0.00% Moderate Poor (2.5-4 USD PPP) 0.00% 0.01% 0.01% Vulnerable (4-10 USD PPP) 0.01% 0.02% 0.02% Middle Class (10-50 USD PPP) 1.49% 0.06% 0.04% Upper Class (>50 USD PPP) 3.65% 0.45% 0.01% 1.19% 0.09% 0.03% Source: Authors’ estimates based on ENIGH 2007. Note: income definition is USD PPP at 2005 prices. Dividend-tax payments reduce the average Dominican’s market income by 0.03 percent. The top three deciles account for 84.8 percent of total dividend tax payments. In terms of socioeconomic groups, the middle class pays 67.3 percent of dividend taxes, a much higher proportion than the richest population (6.3 percent). Those taxes reduce the market income of the middle class by 0.04 percent, while the toll on the richest population was only 0.01 percent (Table 9). Figure 12. Incidence of direct taxes per decile, as share of market income, country comparison 22 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only 20% 15% 18% 13% Share of GDP (right axis) Share of market income 16% 11% 14% 12% 9% 10% 7% 8% 5% 6% 3% 4% 2% 1% 0% -1% Poorest decile Richest decile Share of GDP (right axis) Source: Authors’ estimates and Lustig et al. (2013), CEQ Standard Indicators by a number of authors. Figure 12 suggests that direct taxes could be more progressive in the Dominican Republic than in other countries. Of the selected cases, Jordan, and Peru have similar or higher progressivity. Low-income households in other countries, such as Armenia, Brazil, and Uruguay, pay much higher percentages of their market income as direct taxes. At the same time, it is worth noting that the Dominican Republic’s high exemption threshold results in the lowest share of direct taxes to GDP among surveyed countries. A decrease in informality, which currently accounts for 56 percent of labor activity, could also have a positive effect on personal income tax revenues. Nonetheless, the high amounts of foregone revenue can probably be explained by evasion among the richest. All these cross-country comparisons are based on a same estimation methodology (Lustig, 2013); nonetheless, since the taxes, rates, and exemptions may differ across countries, results should be interpreted with caution. 4.1.2. Indirect taxes The analysis includes the ITBIS and several excises paid by Dominican Republic residents. The indirect taxes are subtracted from disposable income (i.e., net market income plus direct government transfers) to calculate post-fiscal incomes (once indirect subsidies are also added). The indirect taxes considered in the analysis are: the ITBIS; excise taxes on alcoholic beverages, beer, and cigarettes; and excise taxes on oil products, telecommunications, insurance services, and several other imported goods. Rates vary on the Dominican Republic’s indirect taxes. The ITBIS is a value-added tax, which had two tax rates in 2013. The general tax rate was 18 percent and the reduced tax rate, levied on a group of primary goods, was 8 percent.33 The excise taxes on consumption are a single stage sales tax. The excise taxes on 33Law No. 253-12 of November 2012 states that the reduced tax rates would be increasing annually until 16 percent in 2016. It was also stated that the general tax rates would be reduced to 16 percent if the tax income achieve 16 percent of GDP in 2015. 23 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only alcoholic beverages, beer, and cigarettes include specific taxes and ad valorem taxes.34 Telecommunications services are taxed at 10 percent and insurance services at 16 percent. In terms of concentration, the share of indirect tax payments of the first eight deciles (35.3 percent) is below their share of market income (43.5 percent). By socioeconomic groups, the concentration share of those living on less than US$4 a day is lower for indirect taxes (7.3 percent) than for market income (10.2 percent). The middle class (per capita income between US$10 and US$50 a day)35 has a higher share in indirect taxes (48.9 percent) than market income (46.6 percent). Indirect taxes have reduced the market income across all deciles; at the same time, their incidence is higher on the richer deciles, which makes these taxes progressive. Indirect taxes reduce the market income of the poorest decile by 5.1 percent, compared to 9.0 percent in the top decile. In terms of the socioeconomic groups, indirect taxes reduce middle class market income (per capita income between US$10 and US$50 a day) by 7.8 percent. Figure 13. Progressivity of indirect taxes: Concentration curves and Lorenz curve for market income 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 0% 20% 40% 60% 80% 100% Market Income ITBIS Population Excises taxes Source: Authors’ estimates based on ENIGH 2007. In terms of tax revenue, the ITBIS is the most important indirect tax, representing 58.6 percent of total indirect taxes included in this study. ITBIS is just slightly progressive, as depicted in Figure 13, where the concentration curves and Lorenz curve for market income are almost on top of each other. The top decile income population accounts for 41.2 percent of total ITBIS paid, just above its share of market income (40.5 percent). Average market income is reduced 4.4 percent by ITBIS incidence (Table 10). Among population segments, the ultra-poor suffer a 3.5 percent reduction of market income and the extreme poor a 4.0 percent reduction. The tax reduces market incomes by 4.5 percent for both the middle and upper classes. Table 10. Benchmark scenario: Incidence of ITBIS and excises taxes by socioeconomic group (% of market income) 34 See Title IV of the Law No. 11-92 Tax Code of the Dominican Republic. Ad valorem taxes are 10 percent on alcoholic beverages and beer and 20 percent on cigarettes. In this study, only ad valorem excise taxes are included because there is not enough information to map fixed excise taxes onto consumption of alcoholic beverages and cigarettes. 35 For a definition of middle class specific to the Dominican Republic, please see Guzmán (2011). 24 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Source: Authors’ estimates based on ENIGH 2007. Box 1. Including VAT evasion assumptions in the Dominican Republic Value-added tax (VAT) evasion is a problem in the Dominican Republic. According to General Directorate of Internal Taxation (DGII) estimates for 2010, about 29.7 percent of this tax was evaded. Therefore, it was important to include an adjustment for evasion in estimating the CEQ. In consultation with DGII experts, estimates of actual tax payments for a limited group of products were obtained. It was necessary to make assumptions of tax evasion for the products not covered by DGII data. The evidence suggests that taxes on some goods are either regularly evaded or paid in full, while evasion or payment depends on place of purchase for another group of goods. With this in mind, goods were clustered in the following four groups: 1. Highly probable that no tax is paid (100 percent evasion on the purchases of these goods). 2. Highly probably that taxes are paid (0 percent evasion on the purchases of these goods). 3. On those which the DGII has information on the proportion of tax paid, the effective tax rate was applied. 4. On those which it is assumed that tax payments are conditional on place of purchase, a different evasion rate was applied to urban and rural consumers. To make these adjustments, we created two auxiliary files. The first includes each of the goods contained in the ENIGH 2007 that were classified in one of the four categories described above (product code and product group). The second defines whether the tax on the product is evaded or paid according to the place of purchase for those cases where evasion is conditional. With the information on tax evasion, and taking into account the nominal tax rate for 2007 (16 percent), we calculated the VAT tax base for each household, given the level of consumption for each good in 2007. Then we applied the nominal tax rates for 2013 (18 percent and a reduced rate of 8 percent for some goods) for each type of good, adjusted by evasion levels. This allowed us to estimate the VAT payment for each good consumed by households in the survey. Note: income definition is USD PPP at 2005 prices. Figure 14. Indirect taxes concentration shares per socioeconomic groups 25 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Source: Authors’ estimates based on ENIGH 2007. Note: Socio-economic income groups are defined in USD PPP at 2005 prices. Excise taxes account for 41.4 percent of the indirect taxes included in this study. These taxes are more progressive than ITBIS. Almost 60 percent of excise taxes are paid by the top decile of the population. In terms of socioeconomic groups, the middle class receives 46.6 percent of total market income and pays 51.1 percent of excise taxes (Figure 14). The 1.4 percent richest population (per capita income above US$50 PPP a day) accounts for 13.6 percent of total market income and pays 26 percent of excise taxes. Excise taxes reduce the market income received by the upper class by 5.9 percent, which is significantly higher than the reduction for the ultra-poor (1.2 percent). As a percentage of GDP, the Dominican Republic receives a relatively high level of revenue through indirect taxes. Compared with selected countries, indirect-tax revenues are higher in the Dominican Republic than in Mexico, Indonesia, Guatemala, Sri Lanka, Peru, and Ethiopia. At the same time, it is worth noting that the Dominican Republic’s VAT tax rate is also high (18 percent) by international standards.36 In addition, the Dominican Republic is one of the few countries (for example, Peru) with progressive indirect taxes. This is mostly due to the previously discussed progressivity of excise taxes. Figure 15. Incidence of indirect taxes per decile, as share of market income, country comparison 36 The average nominal VAT rate in Latin America is around 15.6 percent. 26 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only 35% 16.0% Incidence as a share of Market Income 30% 14.0% 12.0% 25% Share of GDP 10.0% 20% 8.0% 15% 6.0% 10% 4.0% 5% 2.0% 0% 0.0% Poorest decile Richest decile Share of GDP (right axis) Source: Authors’ calculations and Lustig et al. (2013). Tax progressivity in the Dominican Republic is high compared to other developing countries. Table 11 shows the Kakwani indexes for direct and indirect taxes in selected countries, allowing us to compare the progressivity of taxes. This index is equal to the difference between the concentration coefficients of a particular tax and the Gini coefficient of the reference income. When the Kakwani index is above zero, the tax is progressive. If it is below zero, the tax is regressive. And if it is equal to zero, the tax is neutral. The Reynolds-Smolensky (RS) Index shows the difference in value of Gini coefficient after Direct or Indirect Taxes. Among the selected countries, the Dominican Republic has one of most progressive direct taxes, with a Kakwani index of 0.42. Only Jordan, Sri Lanka, and Peru have more progressive direct-tax systems. In the Dominican Republic, indirect taxes are slightly progressive, with a Kakwani index of 0.05. International practice dictates that a Kakwani index between -0.1 and 0.1 could be considered neutral; however, looking at this group of countries, we conclude that the Dominican Republic has the second most progressive indirect tax system, just behind Ethiopia. Table 11. Progressivity indices for direct and indirect taxes, country comparison Kakwani Direct Kakwani Indirect index for taxes as a index for taxes as a RS index RS index direct share of indirect share of taxes GDP taxes GDP (3) = (3) = (1) (2) (1) (2) (1)*(2)*100 (1)*(2)*100 Armenia (2011) 0.23 5.2% 1.19 -0.04 12% -0.48 Bolivia (2009) -0.13 11% -1.46 Brazil (2009) 0.27 4.2% 1.13 -0.03 14% -0.46 Costa Rica( 2010) 0.00 0.00 Dominican Republic (2013) 0.42 1.3% 0.54 0.05 7% 0.37 El Salvador ( 2011) 0.00 0.00 27 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Ethiopia (2011) 0.28 3.9% 1.11 0.06 8% 0.50 Indonesia (2012) -0.05 4% -0.22 Jordan (2010) 0.63 3.3% 2.09 -0.06 11% -0.60 Mexico (2010) 0.30 3.9% 1.14 0.01 4% 0.05 Peru (2009) 0.43 1.5% 0.65 0.02 7% 0.14 South Africa (2010) 0.13 14.3% 1.79 -0.08 10% -0.86 Sri Lanka (2009) 0.53 2.9% 1.52 0.00 7% 0.02 Uruguay (2009) 0.25 4.7% 1.18 -0.05 7% -0.37 Sources: Armenia (Younger et al., 2014), Bolivia (Paz et al., 2014), Brazil (Higgins and Pereira, 2014), Ethiopia (Hill et al., 2014), Indonesia (Jellema et al., 2014), Jordan (Serajuddin et al., 2014), Mexico (Scott, 2014), Peru (Jaramillo, 2014), Sri Lanka (Arunatilake et al., 2014), Uruguay (Bucheli et al., 2014), and authors’ estimates for Dominican Republic. 4.2. Social spending in the Dominican Republic This section assesses the incidence of direct transfers. These include the conditional cash transfer (CCT) food program Comer es Primero, CCT programs related to education, targeted transfers for LPG and electricity consumption, transfers to policemen and marines, indirect subsidies (mainly on electricity), and health and education services. The aim is to gain a better understanding of the extent to which Dominican social spending is progressive, using other countries as a benchmark for comparison. 4.2.1. Direct transfers Total concentration shares from the fiscal-incidence analysis show that some of the Dominican Republic’s direct transfer do better than others in reaching the poor. Around 52 percent of the public expenditures under Comer es Primero reaches poor households (per capita income below US$4 a day), 38 percent goes to the vulnerable (between US$4 and US$10 a day), and less than 10 percent benefits middle- class households (above US$10 a day per capita). For Bonogas Hogar and Bono Luz, more than 60 percent of total spending goes to the non-poor (earning more than US$4 a day); as previously explained, this relates to the fact that, unlike the CCTs, a group of the non-poor according to the SIUBEN life quality index can be beneficiaries of these programs. This makes Bonogas Hogar and Bono Luz the only programs progressive in relative terms (Figure 16, left panel). In contrast, Comer es Primero and the aggregate of other direct transfers are progressive in both relative and absolute terms, since, apart from representing a larger share of market income for poor households than for non-poor households, total transferred amount in aggregate terms are also larger for the former group. The CCT incentivizing school attendance, ILAE, would be the most progressive direct transfer program in the Dominican Republic. 28 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Figure 16. Distribution of direct transfer spending by level (percentages) 100.0% 100.0% 90.0% 80.0% 80.0% 70.0% 60.0% 60.0% 50.0% 40.0% 40.0% 30.0% 20.0% 20.0% 10.0% 0.0% 0.0% 0.0% 20.0% 40.0% 60.0% 80.0% 100.0% Comer es Bono Gas Bono Luz Quisqueya Other primero Hogares aprende Direct Population Market Income contigo Transfers Comer es primero Bono Gas Hogares Bono Luz Quisqueya aprende contigo Extreme Poor (y<2.5) Moderate Poor (2.54) Source: authors´ elaboration using the CEQ methodology. Note: Socio-economic income groups are defined in USD PPP at 2005 prices. In terms of incidence, Comer es Primero would be the program with the largest impact. These direct transfers represent 5.5 percent of market income among the ultra poor (less than US$1.25 a day) and 2.1 percent for the extremely poor (below US$2.50 a day) (Table 12). This has to do with the amount of the transfer, which is significantly larger for Comer es Primero than for ILAE; the latter is included in the Other Direct Transfers category. The incidence of Bonogas Hogar, Bono Luz, and Quisqueya Aprende Contigo is more limited due to the relatively modest amount transferred and the fact that some the funds go to the non-poor population. Table 12. Incidence of direct transfer programs on socioeconomic class income (percentages) Quisqueya Comer es Bono Gas Other Direct Bono Luz Aprende Primero Hogares Transfers Contigo Ultra poor (<1.25 USD PPP) 5.55% 1.14% 1.15% 1.18% 5.92% Extreme Poor (1.25-2.5 USD PPP) 2.15% 0.51% 0.57% 0.52% 2.29% Moderate Poor (2.5-4 USD PPP) 1.00% 0.28% 0.31% 0.27% 1.15% Vulnerable (4-10 USD PPP) 0.39% 0.16% 0.17% 0.11% 0.32% Middle Class (10-50 USD PPP) 0.06% 0.05% 0.05% 0.03% 0.04% Upper Class (>50 USD PPP) 0.00% 0.00% 0.01% 0.00% 0.00% 0.31% 0.11% 0.12% 0.09% 0.29% Source: Authors’ estimates based on ENIGH 2007. Note:_income definition is USD PPP at 2005 prices. 29 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Although the Dominican Republic’s direct transfers are progressive, international comparisons suggest more could be done to help the poor. The Dominican Republic exhibits declining concentration shares for direct transfers by deciles, indicating that public spending in this category was progressive in relative terms in 2013 (unlike in Bolivia or Brazil in 2009). Nonetheless, as observed in Figure 17, the decline in shares from the poorest to the richest decile is less steep than in the rest of the countries.37 This suggests that there would be room for a more pronounced income redistribution strategy using this fiscal policy tool. Figure 17. Concentration shares of direct transfers, by deciles, country comparison 60% 4.5% 4.0% 50% 3.5% 40% 3.0% Share of Benefits Share of GDP 2.5% 30% 2.0% 20% 1.5% 1.0% 10% 0.5% 0% 0.0% Poorest decile Richest decile Share of GDP (right axis) Source: CEQ working papers (http://www.commitmentoequity.org), Tulane University and World Bank staff calculations. The Dominican Republic is less able to reduce inequality through direct transfer programs than most of these other countries. The incidence of direct transfers as a share of market income for individuals in the first decile (11 percent) is similar in the Dominican Republic and Peru, although the Andean country invests only a third of the Dominican Republic’s budget. Incidence is much smaller in the Dominican Republic than in Argentina (247 percent), Brazil (107.3 percent), Uruguay (61.9 percent), Bolivia (33.2 percent), or México (31.4 percent). The main explanation is that half of the Dominican Republic’s spending on direct transfers is benefiting the non-poor. Overall, the amounts granted under CCTs and other targeted and untargeted programs in the Dominican Republic are relatively modest. On one hand, this would help limit discouraging job search. On the other hand, small CCT amounts may be insufficient to mitigate a sharp economic shock. In a microsimulation exercise, Valderrama et al. (2013) assessed ex-ante the planned increase in monthly Solidaridad grants from RD$700 to RD$830 (around US$3 more). According to the results, this would have resulted in a decrease of 0.22 percent in moderate poverty and 0.65 percent in extreme poverty. Summarizing, cash transfers in the Dominican Republic are generally well targeted and benefit the poor and vulnerable more than proportionately. Most direct transfer programs are built on three transparent mechanisms or institutions: the Solidaridad debit card, the SIUBEN census of beneficiaries, and 37These cross-country comparisons are based on the same estimation methodology (Lustig, 2013); nonetheless, results need to be interpreted with caution because taxes, rates, and exemptions may differ across countries. 30 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only ADESS as independent administrator for transferring funds. Comer es Primero and Incentivo a la Asistencia Escolar are highly progressive programs. On the other hand, 60 percent of public spending on Bono Luz, and Bonogas Hogar goes to the non-poor (vulnerable and middle class), making them barely progressive. Compared to other countries, the impact of direct transfers on poverty and equity is modest due to the fact that, while coverage has noticeably expanded over the past eight years, the amount of individual transfers is relatively small, and part of public spending is directed to the non-poor. 4.2.2. Indirect subsidies In addition to targeted direct transfer mechanisms, generalized subsidies remain in place—for electricity. As previously mentioned, both subsidies have in common a structure of explicit (tariffs below costs) and implicit (irregular connections, fraud, non-payment) components. Given this partly informal nature, few studies have analyzed the distributional impact of utility subsidies in the Dominican Republic. In what is probably the most comprehensive of them, Actis (2012) estimated that 83 percent of electricity subsidies were directed to non-poor households. Following a similar approach, an analysis consistent with the CEQ methodology has been prepared (Box 2). Results confirm that around 81 percent of total spending on electricity in 2013 benefited non-poor individuals. As in many countries, indirect subsidies were only progressive in relative terms (improving the distribution relative to market income), but are regressive in absolute terms (remain below the 45 degree line in Figure 18, left panel). Most spending on indirect subsidies is concentrated on the vulnerable and middle class. Nonetheless, indirect subsidies represent 4.4 percent of the market income of the ultra-poor and around 2.5 percent of the market income of the extreme poor (Figure 18, right panel). So eliminating these subsidies, if feasible, would need compensatory mechanisms to shield the poor from a deterioration in their purchasing power. This could be done through well-targeted and formally established mechanisms, such as Bono Luz. 31 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Box 2. Electricity subsidy estimation approach The Dominican Republic has a fixed electricity fee, applied to households that have not been yet provided with a meter, and a electricity tariff for metered households. The official reference table of the Dominican Superintendence for Electricity established different tariffs by energy consumption intervals, and it is used to determine consumption. The ENIGH 2007 survey distinguishes between these two groups of households. However, it does not provide information on the consumption of those declaring to be subject to the variable tariff. For this analysis, the following method was developed to estimate energy consumption: (i) depart from the official reference table of the Dominican Superintendence for Electricity containing consumption intervals and tariffs to be applied; (ii) take the value of the electricity invoice of the household (data in ENIGH 2007); (iii) apply a multi-tier algorithm that divides the value of the invoice paid by the household by the tariff in each of the different consumption intervals (the tariff varies as kwh consumption increases); (iv) make calculations for both the fixed and variable tariffs set by the Superintendence for Electricity. Given that not all households report paying for electricity, energy consumption was applied to households that have not paid for service. The average consumption of households paying for electricity was applied to these individuals, depending on their SIUBEN life conditions category. Once consumption estimates were computed for all households, the electricity subsidy was estimated as the energy cost per kwh minus the average tariff according to the consumption interval. The assigned energy cost was RD$8.75 per kwh in 2013, or RD$6.16 per kwh in 2007 prices. Finally, to monetize the subsidy at the household level, the subsidy per kwh was multiplied by the energy consumption of the household. Figure 18. Distribution of indirect subsidies spending (left) and incidence on market income by level (right) 6% 45% 100.0% 40% 90.0% 5% 35% 80.0% 70.0% 4% 30% 60.0% 25% 3% 50.0% 20% 40.0% 2% 15% 30.0% 10% 20.0% 1% 5% 10.0% 0% 0% 0.0% y < 1.25 1.25 < y 2.50 < y 4.00 50 USD PPP) 0.0% 0.1% 0.0% Note: * Sensitivity Analysis 2 1.7% 3.3% 1.7% Source: Authors’ estimates based on ENIGH 2007. Note: income definition is USD PPP at 2005 prices. 34 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Education Total public education expenditures are progressive in absolute terms, according the CEQ analysis, but only pre-school, primary, and lower secondary levels achieve this standard of progressivity. For these levels, the bottom 40 percent of the population receives close to two-thirds of spending (Figure 21, left). Upper secondary income is progressive in relative terms and almost proportional to population, which means that the proportion received in relation to market income decreases with income. As in other countries, tertiary education is the least progressive, with more that 20 percent of public spending going to non-poor students. Educational failure and opt-out reduce participation of the poor in higher levels of education. In lower levels, like pre-school and primary, almost 60 percent of total expenditures go to poor households. The share shrinks to 40 percent for secondary levels and less than 20 percent for tertiary levels (Figure 21, right panel). This may be caused by quality concerns about public education, which leads to those who can afford it opting out form the public system and into private schools. Sánchez-Martín and Senderowitsch (2012, p. 13) explained that “the education sector in the DR presents faulty public service delivery, which originates a private offer that is more of a reactive upshot to deficiencies in state education than a high quality alternative (at least not in every case).� Figure 21. Distribution of education spending by level (percentages) Concentration Curves Distribution by socioeconomic group 100% 100% 90% 90% 80% 80% 70% 70% 60% 50% 60% 40% 50% 30% 40% 20% 30% 10% 20% 0% 10% 0% 20% 40% 60% 80% 100% 0% Market Income Population Education: Education: Education: Education: preschool primary secondary tertiary Pre-School Primary Lower Secondary Upper Secondary Extreme Poor (y < 2.5) Moderate Poor (2.5 < y < 4) University Non Poor (y > 4) Source: Authors’ estimates based on ENIGH 2007. Note: Socio-economic income groups are defined in USD PPP at 2005 prices. For the poor, the benefits of education are high for primary schooling but not at other levels. First, Figure 22 shows that almost all children from extremely poor households are enrolled in primary education. This declines to two-thirds in secondary education, less than a quarter in pre-school, and only 6 percent in university.39 Second, public primary-school enrollment declines as income increases; in increases for secondary school and university. For the lower levels, it could be the result of opt-out to private schools for quality concerns. Finally, pre-school enrollment is low in public schools. Around three quarters of students go 39According to the Ministry of Education, using a different classification, net enrollment rates in 2012-13 were 44.0 percent for Inicial, 92.6 percent for Básico, and 54.1 percent for Medio. 35 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only to public schools;40 however close to 90 percent of students of first quintile go to public schools, compared to 34 percent and 42 percent of fifth quintile students in Basico and Medio, respectively. Figure 22. Enrollment in public education by level for school aged children (percentages) 100% 91% 90% 80% 83% 77% 80% 67% 70% 60% 55% 50% 40% 30% 23% 26% 20% 14% 13% 10% 6% 7% 00% Education: preschool Education: primary Education: secondary Education: tertiary Extreme Poor (y < 2.5) Moderate Poor (2.5 < y < 4) Non Poor (y > 4) Source: Authors’ estimates based on ENIGH 2007. Note: Socio-economic income groups are defined in USD PPP at 2005 prices. At more than 30 percent, the monetized value of primary education is large compared to market income for the ultra-poor (Figure 23, left panel). It is smaller for the extreme poor and moderate poor but still important. However, it is almost negligible for the vulnerable non-poor, middle, and upper classes for two reasons: they attend less primary and lower-secondary public education, and the impact of public spending per capita is low relative to their income level. Tertiary education has only a small impact on income, and it is almost proportional or neutral in relation to income. Because pre-school has low coverage, it has a lower impact than secondary education, even though both are progressive (Figure 23, right panel). In particular, upper-secondary incidence is significant for the vulnerable non-poor population, even more important than lower secondary and pre-school. The middle and upper classes make up around 23 percent of the population, and they hardly use the public education services, with the exception of higher education and upper secondary. However, education reform introduced extended school days. This program not only increases school hours but also provides breakfast, lunch, and snacks. Education reform also includes improvements in education infrastructure, postgraduate programs for teachers, innovative teaching practices, foreign languages, and technology (OECD, 2015). As a result, public-education use probably will increase in non-poor households, especially among vulnerable and middle class in the near future. 40According to administrative records, this figure is 75 percent of Básica students and 77 percent of Medio students in 2012-201. In our analysis, 74 percent of students of Básica and 70 percent of Medio go to public schools. 36 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Figure 23. Incidence of education expenditures by level for school aged children (percentages) 35% 32% 8% 30% 7% 6% 25% 5% 20% 4% 15% 12% 3% 2% 10% 6% 1% 5% 2% 0% 0% 0% 0% Pre-School Lower Secondary Primary Upper Secondary University Source: Authors’ estimates based on ENIGH 2007. Note: Socio-economic income groups are defined in USD PPP at 2005 prices. Figure 24. Incidence of education expenditure by level for school aged children (percentages) 190% 6.0% Incidence as a share of Market Income 5.0% 140% 4.0% Share of GDP 90% 3.0% 40% 2.0% -10% Armenia Brazil DR (2013) Indonesia Mexico Peru Uruguay 1.0% (2011) (2009) (2012) (2010) (2009) (2009) -60% 0.0% Poorest decile Richest decile Share of GDP (right axis) Source: CEQ working papers (http://www.commitmentoequity.org), Tulane University and World Bank staff calculations. The Dominican Republic compares favorably with other countries in education spending’s incidence on the income of the poorest deciles. For example, countries with similar levels of education spending, like Indonesia and Armenia, have smaller income impacts on the poorest decile (Figure 24). In contrast, education expenditures have a higher incidence on the poorest deciles in Uruguay than in the Dominican Republic. Peru spends less on education, but it has almost the same spending incidence as the Dominican Republic.41 41 These cross-country comparisons are based on the same estimation methodology (Lustig, 2013); nonetheless, the results need to be interpreted with caution because taxes, rates, and exemptions may differ across countries. 37 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Health Health expenditures are even more progressive than education, according to the CEQ results. Due to the limited resources devoted to health, however, the redistributive effect is lower. All components of public health in the analysis are progressive in absolute terms. Subsidized health insurance covers a large portion of the extreme poor, and non-contributive programs (hospital and outpatient care) reach a big portion of the moderate poor. In contrast, the Essential Medicines Program (PROMESE), which includes spending to purchase medicines and medical supplies for public health institutions as well as the distribution of subsidized medicines, is just barely progressive (Figure 25, left panel). Figure 25. Distribution of health spending by level (percentages) Concentration Curves Distribution by socioeconomic group 100% 100% 90% 80% 80% 45% 46% 70% 60% 60% 60% 50% 40% 40% 23% 25% 30% 20% 20% 20% 32% 30% 10% 20% 0% 00% 0% 20% 40% 60% 80% 100% Subsidized Health Non contributive Promese Market Income Population Insurance Subsidized Health Insurance Promese - Medicines Extreme Poor (y < 2.5) Moderate Poor (2.5 < y < 4) Non contributive Non Poor (y > 4) Source: Authors’ estimates based on ENIGH 2007. Note: Socio-economic income groups are defined in USD PPP at 2005 prices. Despite the progressivity, many people in the low-income strata are still not covered by subsidized or non-contributive health insurance. Figure 26 shows coverage is low in poor households. The finding is consistent with information from ENDESA 2013 (CESDEM, 2014), where the poorest two quintiles had coverage of less than 25 percent in the subsidized regime and less than 21 percent in the non-contributive regime. In the lowest quintile, two-thirds of the population does not report having health insurance. Hence, substantial challenges remain in terms of increasing health-insurance coverage. Despite the progress already made, further increases could benefit poor households. Valderrama, et al. (2012) analyze the impact of the projected increase in SENASA coverage to 4 million in 2016. Using the ENFT household survey to simulate the impact on income, they conclude that this policy could reduce extreme poverty 0.78 percent to 1.18 percent. 38 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Figure 26. Individuals who live in beneficiary households by health program and socioeconomic ranking (percentages) 100% 90% 80% 66% 70% 60% 53% 49% 50% 40% 26% 28% 30% 21% 23% 16% 18% 20% 10% 0% Subsidized Health Insurance Non contributive Promese Extreme Poor (y < 2.5) Moderate Poor (2.5 < y < 4) Non Poor (y > 4) Source: Authors’ estimates based on ENIGH 2007. Note: Socio-economic income groups are defined in USD PPP at 2005 prices. The incidence of non-contributive health is the most important of this category. This is because the amount of the health insurance granted under the non-contributive health regime is six times larger than the subsidized scheme. As designed, the subsidized regime does not benefit the non-poor and moderate poor, only the extreme poor and ultra-poor (Figure 27). Finally, PROMESE expenditures—related to cheaper medicines that can be acquired by poor and non-poor at the so-called Boticas Populares—is small compared to market income. However, pharmaceutical products are very important, accounting for 2.6 percent of household budget (CPI basket). Figure 27. Incidence of health expenditures by coverage regime 25% 20% 15% 10% 5% 0% Ultra Poor Extreme Poor Moderate Vulnerable Middle Class Upper Class Poor Subsidized Health Insurance Non Contributive Promese - Medicines Source: Authors’ estimates based on ENIGH 2007. Note: Socio-economic income groups are defined in USD PPP at 2005 prices. In the Dominican Republic, spending policies vary greatly in their impact on the poor. To better understand the effects of the different lines of social spending on equity, Figure 28 adds to the previously 39 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only presented concentration curves by presenting concentration coefficients for each fiscal instrument.42 Most social programs are progressive in absolute terms, with a coefficient below -0.1. This includes most components of education expenditures—except for tertiary education, which is regressive, as in most countries. All health-spending components are also progressive in absolute terms. The most progressive cash transfer is the Incentivo a la Asistencia Escolar (-0.5), followed by Bonogas Chofer and Comer es Primero. Bonogas Hogar and Bono Luz are practically neutral in terms of redistribution; Incentivo a la Marina is regressive. Both the indirect electricity subsidy and the tax expenditure are highly regressive in the sense that they contribute to increasing the disposable income per capita of the wealthier proportionately more than they benefit the poor. We include also contributory pensions (analyzed in Sensitivity Analysis 2), whose incidence is almost neutral (very close to Gini of Market Income), and analysis of VAT tax expenditure, which is detailed in section 6.1, Alternative VAT scenarios for a Fiscal Impact Pact. Figure 28. Concentration coefficients with respect to market income, by fiscal instrument Incentivo a la Asistencia Escolar -0.50 Primary Education Spending -0.33 School supplies -0.33 School Food Program -0.33 Bono Gas Choferes -0.33 Pre-school Education Spending -0.30 Salud no contributiva Ambulatoria -0.27 Salud no contributiva Hospitales -0.26 Health Spending -0.26 Regimen Subsidiado Senasa -0.25 Comer es primero -0.24 All Cash Transfers -0.19 Lower Secondary Education Spending -0.19 Suplemento alimienticio envejecientes -0.19 Education (all levels) -0.17 Quisqueya Aprende -0.12 Total CEQ Social Spending -0.12 Promese -0.04 Bono Gas Hogares -0.02 Bono Luz -0.01 Upper Secondary Education Spending 0.05 Electricity Subsidy 0.29 Tertiary Education Spending 0.30 Incentivo a la Marina 0.44 Total Contributory Pensions 0.46 Tax Expenditures 0.47 Gini (Market Income) 0.51 -0.60 -0.40 -0.20 0.00 0.20 0.40 0.60 Source: Authors’ estimates based in ENIGH 2007. 42Concentration coefficients are calculated in the same manner as the Gini; when the concentration coefficient is above the diagonal, the difference between the triangle of perfect equality and the area under the curve is negative and spending is progressive in absolute terms (i.e., the size of the transfer per capita falls with per capita income). 40 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only 5. Net impact of the fiscal system on income redistribution in the Dominican Republic This section builds on the earlier analysis to take a more comprehensive look at the Dominican Republic’s fiscal system. It assesses the overall capacity of the system to redistribute income, in as well as such related aspects as vertical and horizontal equity, efficiency, and coverage of public spending. 5.1. Fiscal policy instruments, poverty, and inequality in the Dominican Republic Dominican Republic fiscal policy contributes to reducing market income inequality. Using income per capita as the welfare indicator, fiscal policy in 2013 reduced the market income Gini coefficient from 0.514 to 0.458—a decline of 5 Gini points—when all taxes and transfers examined in the previous section are taken into account (including CCTs, indirect subsidies, and the monetized value of education and health). Excluding the monetized value of education and health services, the improvement in inequality is still significant, with the Gini falling from 0.514 to 0.492. The incidence of extreme poverty declines, whereas moderate poverty would remain slightly higher after indirect taxes, both under the national and international definitions. The headcount poverty rate for the ultra-poor (below $1.25 per day) drops from 5.7 percent to 4.9 percent, whereas the rate for the moderately poor (below $4 per day) increases to 37.6 percent (Table 14). This is partly explained by the ultra- poor benefiting more in relative terms from indirect subsidies, and consuming mainly basic food products that are exempt from VAT. The analysis includes the combined effect of all taxes and transfers but not in- kind services such as education and health. It is also more common to see the incidence of poverty calculated with disposable income (before ITBIS); in this case, direct taxes and transfers reduce moderate poverty incidence by about 1 percentage point. 41 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Table 14. Dominican Republic: Poverty and inequality indicators at each income concept Market Net market Disposable Post-fiscal Final income income income income income -1 -2 -3 -4 -5 (2) =(1) -- (3)=(2)+Cash (4)= (3)-- 5=4 + In- Direct taxes transfers Indirect taxes kind transfers Inequality indicators Gini coefficient 0.514 0.509 0.502 0.492 0.458 Theil index 0.521 0.506 0.495 0.468 0.413 90/10 10.41 10.34 9.69 9.28 7.13 Headcount poverty indicators National extreme poverty line* 13.8% 13.8% 12.5% 13.1% – National moderate poverty line* 41.2% 41.2% 40.1% 42.3% – US$1.25 PPP per day 5.7% 5.7% 4.7% 4.9% – US$2.50 PPP per day 19.5% 19.5% 18.2% 19.5% – US$4.0 PPP per day 37.0% 37.0% 35.9% 37.6% – Source: Authors’ estimates based in ENIGH 2007. * Official poverty estimates based in ONE and MEPyD (2012). The lower bound poverty line was set at RD$1,397 per month in 2005/06 using March 2006 prices for urban areas and RD$1,458 for urban areas. The upper bound poverty line was set at RD$2,883 per month in 2005/06 using March 2006 prices for rural areas and RD$3,238 for urban areas. Socio-economic income groups are defined in USD PPP at 2005 prices. Table 15. Average per capita income in each market income decile, in Dominican pesos a year Net market Disposable Post-fiscal Market income Decile income income income (1) (2) (3) (4) Poorest 9,456 9,456 10,454 10,251 2 17,977 17,972 18,924 18,361 3 25,507 25,503 26,339 25,429 4 32,515 32,512 33,282 32,066 5 40,341 40,334 41,033 39,387 6 49,635 49,628 50,251 47,934 7 62,468 62,447 63,047 60,021 8 80,991 80,941 81,466 77,422 9 117,220 116,510 116,953 109,930 Richest 296,428 287,676 287,939 263,070 Source: Authors’ estimates based in ENIGH 2007. The analysis allows us to measure the post-fiscal income on income. In monetary terms, people in the first decile see their per capita incomes increase from RD$9,456 to RD$10,251 a year (an 8.4 percent increase), still far from the average market income per capita of the second decile. Netting out the impact of 42 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only indirect taxes would take post-fiscal income to RD$10,454 (Table 15). Fiscal policy reduces incomes for 8 deciles because the burden of progressive direct and indirect taxes rises with income, and direct transfers are concentrated in lower deciles. It modestly raises incomes for only two deciles because of the limited amounts granted under direct transfers. 5.2. Is fiscal policy more or less redistributive and pro-poor than in other countries? Compared to other countries, the Dominican Republic achieves a modest poverty reduction, although it performs better once education and health care are included. One of the advantages of applying the CEQ methodology is that it allows for international comparison (Lustig and Higgins, 2013). This helps to understand how the Dominican Republic compares to other middle-income countries in fiscal redistribution. Direct taxes, cash transfers, indirect taxes, and health and education spending all contribute to inequality reduction, a desirable result. Relative to its peers, when looking at disposable income, fiscal policy in the Dominican Republic attains a modest reduction in inequality—a drop of 0.012 in the Gini. The results are similar to those in Bolivia, Peru, and Sri Lanka and only higher than Guatemala and Indonesia (Figure 29). Once in-kind education and health spending are monetized, the Dominican Republic compares much more favorably in terms of inequality reduction (0.056) because public spending is much larger than the budgeted for direct transfers, and the poor are more likely to use these public services. Brazil, Costa Rica, and South Africa, the countries with the most redistributive fiscal policies, achieve their inequality reductions through significantly higher levels social spending than the Dominican Republic (return to Figure 7). In addition, South Africa has the most equitable fiscal policy in the sample.43 Figure 29. Change in inequality: Disposable and final income versus market income (in Gini points) South Africa (2010) Guatemala ( 2010) El Salvador (2011) Costa Rica (2010) Indonesia (2012) Sri Lanka (2009) Armenia (2011) Uruguay (2009) Ethiopia (2011) Mexico (2010) Jordan (2010) Bolivia (2009) Brazil (2009) Peru (2009) DR (2013) 0.00 -0.02 -0.04 Change in Gini points -0.06 -0.08 -0.10 -0.12 -0.14 -0.16 -0.18 -0.20 Disp. vs market income Final vs market income Source: CEQ working papers (http://www.commitmentoequity.org), Tulane University and World Bank staff calculations. 43These cross-country comparisons are based on the same estimation methodology (Lustig, 2013); nonetheless, results need to be interpreted with caution because taxes, rates, and exemptions may differ across countries. 43 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Poverty incidence, using the standard of $2.50 per day, does not significantly change when considering post-fiscal income in the Dominican Republic (Table 16). In other countries, even in countries where the incidence of direct taxes and cash transfers on poverty reduction is slightly below average, indirect taxes have a lower incidence on the income of the poor. For example, in Brazil or Bolivia is significantly reduce poverty incidence through cash transfers; however, when looking at post-fiscal income (after indirect taxes), extreme poverty incidence has increased in those countries. Table 16. Poverty headcount rate for the US$2.50 PPP a day for each income concept Net variation Net variation Market Net Market Disposable Post-fiscal (post fiscal to (disposable to Income Income Income Income market) market) (1) (2) (3) (4) 2= 1- Direct 3=2 +Cash 4=3-Indirect =4-1 =3-1 Taxes Transfers Taxes Armenia (2011) 31.3% 32.0% 28.9% 34.9% 3.6% -2.4% Bolivia (2009) 19.6% 19.6% 17.6% 20.2% 0.6% -2.0% Brazil (2009) 15.1% 15.7% 11.2% 16.3% 1.2% -3.9% Costa Rica (2010) 5.4% 5.7% 3.9% 4.2% -1.2% -1.5% Dominican Republic (2013) 19.5% 19.5% 18.2% 19.5% 0.0% -1.3% El Salvador (2011) 14.7% 15.1% 12.9% 14.4% -0.2% -1.8% Ethiopia (2011) 81.7% 82.7% 82.4% 84.2% 2.6% 0.7% Guatemala ( 2010) 35.9% 36.2% 34.6% 36.5% 0.6% -1.3% Indonesia (2012) 56.4% 56.4% 55.9% 54.8% -1.6% -0.5% Jordan (2010) 4.2% 4.2% 2.4% 1.8% -2.4% -1.8% Mexico (2010) 12.6% 12.6% 10.7% 10.7% -1.9% -1.9% Peru (2009) 15.2% 15.2% 14.0% 14.5% -0.7% -1.1% South Africa (2010) 46.2% 46.4% 33.4% 39.0% -7.2% -12.8% Notes: Year of the survey in parenthesis. Bolivia and Indonesia include indirect taxes only. Source: CEQ working papers (http://www.commitmentoequity.org), Tulane University and World Bank staff calculations. Fiscal policy reduces poverty in the Dominican Republic. Overall, when looking at post-fiscal income in the Dominican Republic, we observe a decline in the share of population living on less than US$1.25 a day, while the percentages of extremely poor, moderately poor, and vulnerable increase. At the same time, we see a reduction in the size of the middle and upper classes (Figure 30). Nonetheless, it is worth noting that poverty incidence figures do not give a sense of the total impact on the poor. When using the non- anonymous measure of fiscal impoverishment, 27 percent of the post-fiscal poor were impoverished using the US$1.25 line (poor made poorer and non-poor made poor). However, these results do not consider the effects the monetized value of in-kind education and health services would have on household income (final income). 44 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Figure 30. Percentage of population by socioeconomic class in the Dominican Republic 100% 1.4% 1.2% 1.2% 1.0% 90% 20.6% 21.6% 21.8% 22.0% 80% 70% Upper class (>50$) 60% 40.7% 40.0% 40.1% 41.0% Middle class (10-50$) 50% Vulnerable (4<10$) 40% Moderately poor (2.5<4$) 30% Extremely poor (1.25<2.5$) 17.4% 17.4% 17.6% 18.2% 20% Ultra poor (<1.25$) 13.8% 13.8% 13.5% 14.6% 10% 5.7% 5.7% 4.7% 4.9% 0% Market Income Net Market Disposable Post-fiscal Income Income Income Source: Authors’ estimates based in ENIGH 2007, applying the CEQ methodology. Note: Socio-economic income groups are defined in USD PPP at 2005 prices. It is also important to understand the extent to which fiscal policy boosts the income of the poor. In the Dominican Republic, households in the poorest decile receive transfers and indirect subsidies that are worth 9.2 percent of their market income, which is relatively low compared to most countries (Figure 31, left panel). This may be due to two causes: the lowest decile in terms of market income per capita is not as poor in the Dominican Republic as in other countries; and, probably, the amounts granted under CCT programs are smaller than in Brazil, South Africa, or Uruguay. Including monetized value of public spending in health and education, households in the poorest decile see an increase of 68 percent relative to market income, about half the average for the selected group of countries, excluding South Africa (Figure 31, right panel). Figure 31. Post fiscal (left) and final income (right) as a share of market income 100% 100% Armenia (2011) 80% 80% Brazil (2009) Mexico (2010) 60% 60% Peru (2009) South Africa (2010) 40% 40% Uruguay (2009) DR (2013) 20% 20% 0% 0% -20% -20% -40% -40% Source: CEQ working papers (http://www.commitmentoequity.org), Tulane University and World Bank staff calculations. 45 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Households’ net cash position after taxes and transfers is positive for the bottom 30 percent of the population, which is similar to other middle-income countries. The fact that the line is flatter for the Dominican Republic than for similar countries reflects an overall lower income per capita redistribution across deciles. Once the monetized value of in-kind spending on education and health are included, only the top 30 percent are net contributors in fiscal terms in the Dominican Republic. 5.3. Income redistribution: vertical and horizontal equity, effectiveness indicators. A fiscal system can generate horizontal inequity by generating different impacts on the disposable income of similar households (Duclos and Araar, 2006). For example, let’s imagine two poor individuals, A and B, with similar consumption patterns. The market income is just 100 Dominican pesos higher for B than that of A. Both households should be entitled to conditional cash transfers, but B does not receive these benefits due to limitations in coverage of the social programs. As a result, disposable income after intervention will be lower for B than for A. In this hypothetical case, the fiscal system would be generating horizontal inequality. Table 17. Taxes, transfers and subsidies: Overall redistributive effect* (Decline in Gini Points; shown as positive) South Africa Bolivia Brazil DR Indonesia (2010) (2009) (2009) (2013) (2012) Gini (Market income) 0.771 0.503 0.579 0.514 0.418 Gini (Post-fiscal income) 0.695 0.503 0.546 0.492 0.416 1 Redistributive Effect 0.077 0.000 0.033 0.023 0.002 2 0.025 Vertical Equity (VE) 0.083 0.003 0.048 0.007 3 Re-ranking Effect (RR) 0.006 0.003 0.014 0.001 0.005 RR/VE 0.075 1.000 0.300 0.026 0.706 44 Source: Lustig(2015). Notes: 1. Redistributive Effect calculated as the difference between market income and post-fiscal income Gini. 2. Reynolds- Smolensky Index. 3. Atkinson-Plotnick Index. Fiscal policy’s overall redistributive effect is defined as the change in inequality associated with direct and indirect taxes as well as direct transfers and subsidies. This effect can be decomposed into vertical equity and re-ranking effects. The latter postulates that the pre-fiscal policy income ranking of individuals should be preserved. If not, there is a loss of horizontal equity. Results for five middle-income countries are presented in Table 17. An extreme case of horizontal inequity induced by fiscal policy is Bolivia, where the re-ranking of individuals completely wipes out the reduction in vertical inequity. In the Dominican Republic, the fiscal system achieves intermediate levels of inequality reduction through direct and indirect taxes and transfers and subsidies, and it generates very little horizontal inequality. The country’s r e-ranking as a proportion of vertical inequality is by far the lowest among the five countries. Figure 32 shows, disposable and post-fiscal income incidence curves in the Dominican Republic hardly vary when the re-ranking effect is considered. It is worth noting that a series of geographical disparities in income distribution in the Dominican Republic are observed, while they remain beyond the scope of this analysis. 44 Based on Higgins and Pereira (2014), Jellema et al. (2014), Paz Arauco et al. (2014), Inchauste et al. (2015). Indonesia is the only country in which the analysis has followed a consumption-based definition of income. Market income includes contributory pensions in all cases except for the Dominican Republic. 46 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Figure 32. Fiscal incidence curves and fiscal mobility profiles by deciles 12.0% 12.0% 9.0% 9.0% 6.0% 6.0% 3.0% 3.0% 0.0% 0.0% -3.0% 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 -3.0% -6.0% -6.0% -9.0% -9.0% -12.0% -12.0% Disposable Income with Reranking Post-fiscal Income with Reranking Disposable Income without Reranking Post-fiscal Income without Reranking Source: Authors’ estimates based in ENIGH 2007, applying the CEQ methodology. Effectiveness indicators (Beckerman 1979; Immervol 2009) suggest the Dominican Republic has space to improve the effectiveness of direct transfers and focus them on the extreme poor. According to Table 18, the share of direct transfers that contribute to eliminating extreme poverty is low—8 percent for US$1.25 PPP, 29 percent for US$2.50 PPP, and 20.7 percent for extreme national poverty.45 The effectiveness for moderate poverty is better because vertical efficiency and poverty reduction efficiency increase with the level of the poverty line. Although direct transfers are not very good at reducing extreme poverty, the spillover index shows there are few impacts on the non-poor. In moderate poverty, only 2 percent of direct transfers received by poor raise their incomes above the poverty-line threshold. In contrast, direct transfers reduce a bigger share of the poverty gap in extreme poverty (19.2 percent for US$1.25 PPP, 10.9 percent for US$2.50 PPP, and 13.5 percent for extreme national poverty) than in moderate poverty (less than 6 percent). Table 18. Beckerman and Immervoll et al. effectiveness indicators $1.25 PPP $2.50 PPP $4.00 PPP National Extreme National Moderate per day per day per day PL PL Vertical Expenditure 0.088 0.289 0.503 0.207 0.549 Efficiency Poverty Reduction 0.059 0.243 0.469 0.162 0.515 Efficiency Spillover Index 0.128 0.049 0.026 0.063 0.020 Poverty Gap Efficiency 0.192 0.109 0.062 0.135 0.056 Source: Authors’ estimates based in ENIGH 2007, applying the CEQ methodology. Note: Socio-economic income groups are defined in USD PPP at 2005 prices. 5.4. Resource needs to fill in coverage gaps The relatively high efficiency of Dominican public education and health expenditures in reducing inequality has to do with their high levels of progressiveness in terms of coverage. The Dominican 45The extreme poverty line under the official poverty measurement methodology (ONE and MEPyD, 2012) is US$2.07 PPP a day for urban households and US$2.00 PPP a day for rural ones. 47 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Republic has a subsidized health regime targeted to the poor; it is estimated that 90 percent of the extreme poor and 83 percent of the moderately poor benefit from public health services. Compared with other countries, the Dominican middle and upper classes participate less in subsidized health care because they usually benefit from the contributory health regime or private health insurance. As a result, the percentage of beneficiaries declines markedly by socioeconomic strata as daily market income increases (Figure 33, left panel). This is a distinguishing feature of the Dominican Republic when compared with the other surveyed countries. Figure 33. Percentage of individuals benefiting from health (left) and public education (right) services, by daily income 94% 99% 100% 90% 93% 90% 83% 78% 81% 90% 80% 80% 70% 70% 65% 70% 68% 58% 60% 53% 60% 51% 50% 50% 40% 40% 27% 30% 30% 20% 20% 10% 0% 10% 0% Bolivia Brazil Dominican Mexico Peru Sri Lanka Republic y<2.5 2.550 y<2.5 2.550 Source: CEQ working papers (http://www.commitmentoequity.org), Tulane University and World Bank staff calculations. Note: income definition is USD PPP at 2005 prices. Turning to education expenditures, markedly declining percentages of beneficiaries by socioeconomic strata are more common as daily market income increases (Figure 33, right panel). Yet, only about 65 percent of the extreme poor in the Dominican Republic benefit from public education spending—a low figure compared to other middle-income countries for which results are available. This may be due to the perceived low quality of public education, which compels household heads (even in poor families) to send their children to private schools (Sánchez-Martín and Senderowitsch, 2012). It is worth noting that this opting-out behavior may have declined with the significant increases of education expenditures after 2012. This would, of course, not be reflected in the ENIGH 2007 survey used in this analysis. Using calculations from applying the CEQ methodology, it is possible to quantify the resources that would be needed to lift all Dominicans out of poverty and cover education and health coverage gaps. Closing the extreme poverty gap (below US$2.50 PPP per capita a day) would require from an additional RD$18.3 billion in cash transfers, the equivalent to 4.9 percent of government revenue and 0.7 percent of GDP in 2013 (Table 19). This would mean doubling the current level of spending on direct transfers. Closing the human-capital gap, defined by public education and health coverage needs for the moderately poor (US$4 PPP a day), would require RD$28.4 billion, or 1.1 percent of 2013 GDP. To fill in the overall poverty gap (US$4 PPP a day), additional resources equivalent to a quarter of total government revenue would be needed, other policies (e.g. taxation) equal. These results are in Dominican Pesos of 2013, and take into account population growth since 2007. One caveat: this exercise assumes that the Government has the capacity to 48 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only manage and efficiently allocate the higher funding, which may not be always the case because of administrative bottlenecks encountered when scaling-up public spending. 49 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Table 19. Estimated resource needs to close existing social gaps in the Dominican Republic Gap in millions Required increase to close gap of LCU 2013 Total Spending Primary Gov. Revenue 2013 GDP Spending Spending or Revenues in millions of -- 515,562 391,884 370,573 2,558,585 LCU Income Poverty $2.5 PPP 18,325 3.6% 4.7% 4.9% 0.7% Gap per day $4 PPP 65,941 12.8% 16.8% 17.8% 2.6% per day Education $2.5 PPP 7,757 1.5% 2.0% 2.1% 0.3% Coverage Gap per day $4 PPP 14,608 2.8% 3.7% 3.9% 0.6% per day Health Coverage $2.5 PPP 6,864 1.3% 1.8% 1.9% 0.3% Gap per day $4 PPP 13,778 2.7% 3.5% 3.7% 0.5% per day Human Capital $2.5 PPP 14,621 2.8% 3.7% 3.9% 0.6% Gap per day $4 PPP 28,386 5.5% 7.2% 7.7% 1.1% per day Overall Poverty $2.5 PPP 32,946 6.4% 8.4% 8.9% 1.3% Gap per day $4 PPP 94,327 18.3% 24.1% 25.5% 3.7% per day Source: Authors’ estimates based in ENIGH 2007, applying the CEQ methodology. Note: income definition is USD PPP at 2005 prices 50 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only 6. Options to enhance the equity outcomes of fiscal policy in the Dominican Republic 6.1. Alternative VAT scenarios for a Fiscal Impact Pact Dependence on indirect taxes remains a challenge for the Dominican Republic. As previously mentioned, tax expenditures derived from ITBIS exemptions amount to around 3 percent of GDP in the Dominican Republic (DGII, 2015). The estimations in Figure 34 suggest that the bulk of total tax expenditures (88 percent) benefits non-poor households. The share of tax expenditures hold by the poor (US$4 a day PPP definition) would be largest in the case of exemptions relating to food (around 20 percent) and household furnishings (16 percent). Figure 34. Beneficiaries of VAT tax expenditure for different product categories 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Groups: y < 1.25 1.25 < y < 2.5 2.5 < y < 4 4 < y < 10 10 < y < 50 y > 50 Source: Authors’ estimates based in ENIGH 2007 and DGII. Note: Socio-economic income groups are defined in USD PPP at 2005 prices. Taking as a starting point the analysis of the World Bank (2006), we estimate alternative ITBIS reform scenarios. The purpose of these short-term simulations is to explore the likely effects on revenue collection, poverty, and inequality that would follow total or partial elimination of ITBIS exemptions. As a caveat, it is important to note that this is based on a static incidence analysis, and simulations do not consider potential changes in the behavior of taxpayers due to the changes in ITBIS. The four scenarios simulated are: (i) total elimination of ITBIS exemptions; (ii) elimination of all exemptions except for health, education, and electricity; (iii) partial elimination of exemptions, preserving those on the basic basket of goods and services; (iv) finally, partial elimination of exemptions except for electricity, health, education, and basic goods—a combination of ii and iii. 51 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only In the first scenario, we simulate the elimination of all exemptions; i.e., all exempted goods and those with reduced rate would pay a rate of 18 percent. This exercise also takes into account ITBIS tax evasion, drawing from information by the General Directorate of Internal Taxation for 2010 by different product lines (Box 1). So we assume that tax payments on ITBIS goods that had been exempted will have an average evasion rate about 29.7 percent in 2010, equal to what was estimated by DGII (2015). The second scenario retains exemptions for some products. The World Bank (2006) warns that some goods and services are hard to tax for political and efficiency reasons, like educational, health, and electricity supply services. The second simulation is also ambitious in broadening the tax base by eliminating all exemptions except for those relating to these sectors. In the third scenario, only exemptions on the basic basket products will remain. In cooperation with public-sector institutions and international agencies, ONE drafted a report identifying the basic basket of goods (ONE 2012), and we use it to select the goods that remain exempt goods in this scenario.46 The final scenario for dealing with ITBIS combines the previous two. We estimate a more conservative scenario that maintains exemptions on politically sensible goods and the basic basket of consumption.47 The simulations show that ITBIS changes would not have a significant impact on the Gini coefficient. Elimination of all exemptions slightly increases inequality. However, the second scenario had the greatest inequality increases because of the elimination of exemptions in some basic goods and services (including food products). The third and fourth scenarios preserve basic food exemptions, and inequality remains unchanged. Eliminating all exemptions would increase poverty. In the first scenario, moderate poverty incidence would increase by 1.3 percentage points and extreme poverty by 0.7 percentage points. If only politically sensitive goods were exempt, moderate poverty increase would be lower but still significant. By contrast, extreme poverty would not increase if ITBIS when exemptions on the basic basket of goods are kept in place (Figure 35), which seems to indicate that the poor purchase almost exclusively products in this basket. This is not surprising, since the national poverty definitions are according country specific patterns of consumption and caloric requirements (ONE and MEPyD, 2012). In the first scenario, with all exemptions removed, revenue collection increases the most—around 2.2 percent of disposable income. In the second scenario, with all exemptions but those on education, health, and electricity removed, revenue collection would increase by 1.7 percent of disposable income. Finally, if basic food were also exempt, tax revenue would increase by only about 0.3 percent of disposable income (Figure 36). It is worth noting that the incidence analysis simulated using the ENIGH 2007 has been adjusted to reflect the amount of tax expenditure estimated by official sources in 2013. According to our analysis, eliminating exemptions would result in improved tax collection. In all scenarios, inequality would not suffer significantly, but income poverty would be sensitive to changes in ITBIS exemptions under simulation scenarios 1 and 2. There is an important tradeoff in terms of revenue collection (most improved under first and second scenarios) and poverty incidence (less affected under the third and fourth scenarios). Overall, it can be argued that by leaving only the basic basket of goods exempt, the Government would modestly improve revenue collection; the impact on extreme and moderate poverty incidence would be minimized (0.1 percentage points increase). This impact could potentially be compensated by direct cash transfers using existing well-targeted programs. 46See annex 5 in ONE (2012). 47The World Bank (2006) considered a fourth scenario with reduced rates for basic food. However, we do not consider this scenario because Dominican Republic has been phasing out reduced rates. 52 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Figure 35. Effects on inequality and poverty of alternative ITBIS exemption scenarios Gini coefficient Headcount ratio (national poverty lines) Sim4 National Basic Basket + 50% 43.6% 42.3% 43.4% 42.4% 42.4% 0.4922 41.2% hlth+edu+elec) 40% Sim3 (Only National Basic 30% 0.4922 Basket) 20% 13.8% 13.8% 13.7% 13.1% 13.2% 13.2% Sim2 (Sim 1 - Edu, health & 0.4932 electricity) 10% 0% Sim1 All exceptions 0.4925 Post Fiscal Income BM 0.4922 Market Income BM 0.5141 0.4750 0.4850 0.4950 0.5050 0.5150 Extreme Moderate Source: Authors’ estimations based on ENIGH 2007. Figure 36. Effects on revenue increase in scenarios of ITBIS (as a percentage of total disposable income) 2.5% 2.2% 2.0% 1.7% 1.5% 1.0% 0.5% 0.3% 0.3% 0.0% Sim1 All exceptions Sim2 (Sim 1 - Edu, Sim3 (Only National Sim4 National Basic health & electricity) Basic Basket) Basket + hlth+edu+elec) Source: Author’s estimates based on ENIGH 2007. 53 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only 6.2. Policy options Fiscal incidence analysis applying the CEQ methodology show that, as of 2013, the Dominican Republic’s fiscal policy was progressive overall. Compared to other countries subject to the same methodology, the Dominican fiscal system achieves intermediate levels of inequality reduction through direct and indirect taxes as well as transfers and subsidies, and it generates very little horizontal inequality. Re- ranking of households as a proportion of vertical inequality is by far the lowest among similar countries. Using income per capita as the welfare indicator, fiscal policy in 2013 reduced the market income Gini coefficient from 0.514 to 0.458—a decline of 5 Gini points—when all taxes and transfers (including the monetized value of education and health) are taken into account. Excluding the monetized value of education and health services, the improvement in inequality is more modest, with the Gini falling to 0.492. The incidence of extreme poverty also declines when comparing market and post-fiscal incomes (excluding education and health), whereas moderate poverty would remain slightly higher after indirect taxes, both under the national and international definitions. In terms of poverty reduction, the incidence of direct transfers is modest. This is due to the fact that households in the poorest decile receive transfers and indirect subsidies worth 10 percent of their market income, which is relatively low compared to most countries (back to Figure 31, left panel). This likely relates to the amounts granted under CCT programs being smaller than in Brazil, South Africa, or Uruguay. For the Dominican Republic, resources amounting 1.3 percent of GDP would be needed to lift the extremely poor in the Dominican Republic. Under the international poverty line of US$2.50 PPP a day, ending extreme poverty and ensuring the poor have access to public education and health would require an increase in public resources to social services equivalent to 1.3 percent of GDP, other things remaining equal. This section presents a series of policy options that could help in further improving equity outcomes using fiscal policy. On the education front, the challenge will be increasing the quality of education through measures included in the Education Pact. The Dominican Republic has already significantly boosted public spending, from 2.2 percent of GDP in 2011 to around 4 percent of GDP from 2013 onwards. This has had a significant effect in terms of inequality reduction, given that education spending is highly progressive. In the analysis, we are monetizing the value of public spending in education to estimate changes in inequality. However, if the quality of the service provided is not good, the de facto welfare improvement would be smaller. Enrollment in primary school is higher among the poor than among the non-poor; this is probably because the latter have the resources to opt out and choose private education because of the perception that the quality of public education remains mediocre. Thus, the priority in the sector at the moment should be increasing the quality of education through implementation of the measures included in the Education Pact. In addition, authorities could try to improve access and coverage among the poor, especially in pre-primary and secondary education, where enrollment remains low among the extreme poor (23 percent in pre-primary and 67 percent secondary). Finally, introducing a series of grants to support top performers among the poor could help mitigating school dropout and improve access to and equity in tertiary education. Unlike education, health will require significant increases in expenditures in the Dominican Republic. The country’s public health resources remain low by international standards at around 1.7 percent of GDP, half the amount spent by South Africa and Brazil and a third of Costa Rica’s outlays. The Dominican Republic has had noticeable improvements in terms of coverage, with the percentage of population with health insurance improving from 27 percent in 2007 to 55 percent in 2013. According to the ENDESA 2013 (CESDEM, 2014). However, the bottom 40 percent of the population has coverage of less than 25 percent in the subsidized regime and less than 21 percent in the non-contributive regime. In the first 54 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only quintile, two-thirds of the population does not report having health insurance. In fact, a number of people who do not have insurance are using the Ministry of Health’s hospitals and clinics in emergency situations. A strategy to increase the subsidized regime’s coverage while improving the quality of services would likely result in substantial equity gains, and may require also from upgrading in public facilities in order to attract non-poor individuals into the contributory regime as well. As discussed in the previous section, health spending would need to be increased by around 0.3 percent of GDP to extend coverage to the population living under US$2.50 PPP a day per capita. All the analyzed components and programs of health spending are highly progressive except for PROMESE, which is barely progressive and could be revised to focus resources and medicines on the poor and vulnerable. The non-poor could pay for these health services. A revision of tax policies could be considered to finance the 1.3 percent of GDP in additional resources needed to fill the abovementioned gaps. Personal income taxes make up the lion’s share of direct tax collections; yet, according to our simulations, effective rates of 3.5 percent among upper-class earners (more than US$40 a day PPP) are far from the 15 percent called for in the tax schedule. A positive impact on personal income tax revenue would come from tax administration measures to reduce evasion by the upper class and measures to decrease informality among independent workers, which currently accounts for 56 percent of the active working force. In the Dominican Republic, the challenge will be raising added revenue while maintaining the tax system’s progressivity. The country’s tax progressivity seems high compared to other countries. Of the selected countries, only Jordan, Sri Lanka, and Peru have more progressive direct tax systems. On income taxes, it bears repeating that we have applied statutory rates, and preliminary evidence would need to be contrasted with actual data on collections by income level. The Dominican Republic’ could raise additional revenue by reforming its system of indirect taxes, focusing on the ITBIS exemptions. The indirect taxes are slightly progressive, mostly due to the progressivity of excise taxes; ITBIS is almost neutral. The ITBIS exemptions represent close to 3 percent of GDP (Ministerio de Hacienda, 2015), and the majority of tax expenditures from these exemptions is related to the consumption of middle and upper class households. At the same time, phasing out certain exemptions would have negative impacts on poverty and inequality. With that in mind, a possible option could be for goods in the basic consumption basket (based on the national poverty measurement methodology) to remain taxed at a zero rate, along with health and education services. Other exemptions, especially those that are regressive, could be removed, potentially granting up to 0.5 percent of GDP in additional revenue collection. The impact of the removal of ITBIS exemptions on electricity for the poor could be mitigated through the Bono Luz program. Electricity subsidies could be withdrawn from the non-poor, while taking care of the poor through Bono Luz. Explicit (tariffs below costs) and implicit (irregular connections, fraud, non-payment) electricity subsidies are equalizing in absolute terms but not in relative terms. Simulations applying the CEQ methodology confirm evidence presented by Actis (2012), who estimated that 83 percent of electricity subsidies benefited non-poor households. Fostering a culture of payment by improving service quality and reducing blackouts and adjusting tariffs to market rates are among the measures that could help reduce the deficit in the electricity sector (more than 1.5 percent of GDP in 2013). At the same time, the poor and vulnerable could be shielded from decreases in purchasing power through Bono Luz. Bono Luz and Bonogas Hogar are among the programs that could be slightly reshaped because, at the moment, they are just barely progressive in relative terms. One way would be phasing out the eligibility of beneficiaries in SIUBEN quality of living index category 3 (non-poor). The savings, totaling around 0.1 percent of GDP, could be used to expand both programs’ coverage among the poor. Since these programs are pretty much functioning as universal transfers, another policy alternative would be maintaining 55 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only non-poor as beneficiaries but focusing future coverage expansions on the poor. According to ADESS, 843,000 would be beneficiaries of Bonogas Hogar in 2013 and 533,000 for Bono Luz, compared to a universe of up to 2.4 million potential beneficiaries. Finally, conditional cash transfers have been effective in reaching the poor, and could be further strenghtened. These programs, such as Comer es Primero and Incentivo a la Asistencia Escolar, are highly progressive, with less than 10 percent of public expenditures seeming to go to the middle class. Comer es Primero are fruitful in terms of reducing poverty and inequality, representing 5.5 percent of market income for the ultra poor (living on less than US$1.25 a day) and 2.1 percent for the extremely poor (below US$2.50 a day). Even so, authorities could consider increasing the individual cash amounts transferred through these well-targeted instruments, or at least make sure they are indexed to prevent an erosion of purchasing power. The past decade’s success in putting both conditional and non-conditional cash transfers under the SIUBEN single-targeting mechanism and ADESS administration should be continued. At the same time, the more recent proliferation of small incentive programs may need to be limited to attain more powerful outcomes. Some promising steps are being taken by establishing support schemes and facilitate labor-market integration to those households that have reached non-poor status and will graduate from Progresando con Solidaridad, thus facilitating other poor households to become beneficiaries of the CCT in a context of still limited coverage and resources. All in all, overall fiscal policy in the Dominican Republic is already progressive. Going forward, the challenge is rising revenue collection without affecting the poor and vulnerable, at the same time that public service delivery is improved. As abovementioned, compared to other countries, the fiscal system achieves intermediate levels of inequality reduction (5 Gini points) through direct and indirect taxes and transfers and subsidies, and it generates very little horizontal inequality. Some European States are able to reduce the Gini by more than 15 percentage points, but achieve it through reinvesting large revenue collection in social programs and public services. In this sense, enhancing the quality of public services would be a priority in the Dominican Republic, as it would not only help achieving social outcomes, but also improve citizen trust in institutions, which could ultimately lead towards formalization of economic activity and improved revenue collection. 56 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only 7. Bibliography Abdullaev, U., and Estevao, M., 2013. "Growth and Employment in the Dominican Republic: Options for a Job-Rich Growth," IMF Working Papers 13/40, International Monetary Fund. Actis, J.L. 2012. “Una evaluación de la eficacia de los subsidios al consumo residencial de energía eléctrica en la República Dominicana.� Mimeo. Artana, D., Bour, J.L., Catena, M., and Navajas, F. 2012. “Una Revisión del Gasto Público de República Dominicana.� Banco Interamericano de Desarrollo, Fundación de Investigaciones Económicas Latinoamericanas. Junio 2012. Mimeo. Aristy, J., 2016. “Income definition, prices indices and poverty headcount in the Dominican Republic�. Forthcoming. Barreix, A., Bès, M., and Roca, J. 2009. Equidad fiscal en Centroamérica, Panamá y República Domicana. Washington, D.C.: Interamerican Development Bank/Eurosocial. Bucheli, M., Lustig, N., Rossi, M., and Amábile, F. 2014. “Social Spending, Taxes, and Income Redistribution in Uruguay.� Public Finance Review. May 2014, Volume 42, Issue 3. Cabrera, Maynor, Nora Lustig, and Hilcías Morán. 2014. Fiscal Policy, “Inequality and the Ethnic Divide in Guatemala.� CEQ Working Paper No. 20, Center for Inter-American Policy and Research and Department of Economics, Tulane University and Inter-American Dialogue. CESDEM/ICF. 2014. “Encuesta Demográfica y de Salud República Dominicana 2013,� retrieved from http://countryoffice.unfpa.org/dominicanrepublic/drive/DRDHS2013-Final02-10-2013.pdf, September 20, 2015. Chu, M. K. Y., Davoodi, M. H. R., and Gupta, M. S. 2000. Income distribution and tax and government social spending policies in developing countries (No. 0-62). International Monetary Fund. DGII. 2014. “Gastos Tributarios en República Dominicana. Estimación para el Presupuesto General del Estado del año 2015� Dirección General de Impuestos Internos, Ministerio de Hacienda. Septiembre, 2014. DGII. 2015. “Sistema tributario en República Dominicana.� Dirección General de Impuestos Internos, Ministerio de Hacienda. Abril, 2015. Duclos, J.Y., and Araar, A.. 2006. Poverty and Equity: Measurement, Policy, and Estimation with DAD. New York: Springer and International Development Research Centre. Gallina, A., Inchauste, G., Isa., P., Lee, C., and Sánchez-Martín, M.E. 2015. “Political Economy of Subsidy Reform Case Study for the Dominican Republic.� Mimeo. Galrão Carneiro, Francisco, Aleksandra Iwulska, José-Daniel Reyes, and Miguel Eduardo Sánchez-Martín. 2015. “Resilient Growth, Persisting Inequality: Identifying Potential Factors Limiting Shared Prosperity in the Dominican Republic.� World Banexpenditurek Other Operational Studies 21499. World Bank. Garza, M., P. Morra, and D. Simard. 2012. “The Fiscal Position: Prospects and Options for Adjustment.� In Piñón et al. Central America, Panama, and the Dominican Republic. Challenges Following the 2008-09 Global Crisis. IMF, Washington, DC. 57 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Guzmán, R.. 2011. “Composición Económica Dominicana. El Estrato de Ingresos Medios en el Umbral del Siglo XXI.� Mimeo. Document prepared for the Ministry of Economy, Planning, and Development of Dominican Republic. Higgins, S., and Pereira, C. 2014. “The Effects of Brazil’s Taxation and Social Spending on the Distribution of Household Income.� Public Finance Review. May 2014, Volume 42, Issue 3. IADB and World Bank. 2014. República Dominicana, Notas de Política. Versión para el Día del Diálogo. Inchauste, Gabriela, Lustig, Nora, Maboshe, Mashekwa, Purfield, Catriona, and Woolard, Ingrid. 2015. “The distributional impact of fiscal policy in South Africa.� Policy Research Working Paper Series 7194. The World Bank. Jaramillo, M. 2013. The Incidence of Social Spending and Taxes in Peru. CEQ Working Paper No. 9, Center for Inter-American Policy and Research (CIPR) and Department of Economics, Tulane University. New Orleans. Jellema, Jon, Matthew Wai-Poi, and Rythia Afkar. 2014. “The Distributional Impact of Fiscal Policy in Indonesia.� World Bank. (forthcoming). Lindert, K., Skoufias, E., and Shapiro, J. 2006. “Redistributing income to the poor and the rich: Public transfers in Latin America and the Caribbean.� Social Safety Nets Primer Series. Lustig, Nora and Carola Pessino. 2014. “Social Spending and Income Redistribution in Argentina in the 2000s: the Rising Role of Noncontributory Pensions.� Public Finance Review. May 2014, Volume 42, Issue 3. Lustig, Nora and Sean Higgins. 2013. “Commitment to Equity Assessment (CEQ): Estimating the Incidence of Social Spending, Subsidies and Taxes. Handbook.� CEQ Working Paper No. 1, July 2011; revised September 2013. Lustig, Nora, Florencia Amábile, Marisa Bucheli, George Gray Molina, Sean Higgins, Miguel Jaramillo, Wilson Jiménez Pozo, Veronica Paz Arauco, Claudiney Pereira, Carola Pessino, Máximo Rossi, John Scott, and Ernesto Yáñez Aguilar. 2013. “The Impact of Taxes and Social Spending on Inequality and Poverty in Argentina, Bolivia, Brazil, Mexico, Peru, and Uruguay: An Overview.� CEQ Working Paper No. 13, August 2013. Lustig, Nora. 2015. “Fiscal Policy, Inequality and the Poor in the Developing World. Round 1.� CEQ Working Paper No. 23, Center for Inter-American Policy and Research and Department of Economics, Tulane University and Inter-American Dialogue (forthcoming). Margarita Beneke de Sanfeliu, Nora Lustig, and José Andrés Oliva. 2015. “La incidencia de los impuestos y el gasto social sobre la pobreza y la desigualdad en El Salvador� (forthcoming). MEPyD. 2013. “El efecto de la Ley de fortalecimiento de la capacidad recaudadora del Estado (Ley 253 -12) sobre la pobreza y desigualdad.� Ministerio de Economía Planificación y Desarrollo y Oficina Nacional de Estadística. Enero de 2013. Documento de Trabajo. Minerd. 2014. “Tercer estudio regional y comparativo,� retrieved from http://www.minerd.gob.do/documentosminerd/ResultadosComparado.pdf. September 20, 2015. Ministerio de Hacienda. 2012. “Gastos Tributarios en República Dominicana, Estimación para el Presupuesto General del Estado del año 2013,� Informe de la Comisión Interinstitucional Coordinada por la Dirección General de Política y Legislación Tributaria. 58 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Ministry of Health. 2014. “Gasto Nacional en Salud,� Powerpoint presentation. National Office of Statistics (ONE) and MEPyD, 2012. Metodología para el cálculo de la medición oficial de la pobreza monetaria en la República dominicana. Dominican Republic Poverty Committee, July 2012. Accessible at www.one.gob.do/Multimedia/Download?ObjId=1936. Pablo Sauma, Juan and Diego Trejos, Gasto. 2014. “Público Social, Impuestos, Redistribución del Ingreso y Pobreza en Costa Rica.� CEQ Working Paper No. 18, January 31, 2014. Paz Arauco, Veronica, George Gray Molina, Wilson Jiménez Pozo, and Ernesto Yáñez Aguilar. 2014. “Explaining Low Redistributive Impact in Bolivia.� Public Finance Review. May 2014, Volume 42, Issue 3. Pecho, Miguel, Peláez, Fernando and Jorge Sánchez. 2012. “Estimación del incumplimiento tributario en América Latina: 2000-2010�, Dirección de Estudios e Investigaciones Tributarias, Documento de Trabajo No. 3-2012. CIAT: Panamá. Pessino, C. and R. Fenochietto. 2010). “Determining Countries’ Tax Effort.� Hacienda Pública Española/Revista de Economía Pública, Vol. 195, No. 4, pp. 65-87. Sánchez-Martín, Miguel Eduardo and Roby Senderowitsch. 2012. “The political economy of the middle class in the Dominican Republic: individualization of public goods, lack of institutional trust and weak collective action.� Policy Research Working Paper Series 6049. World Bank. Sauma, P., and Trejos, J. D. 2014. “Social Public Spending, Taxes, Redistribution of Income, and Poverty in Costa Rica.� CEQ Working Paper No. 18, Economics Department and Center por Inter-American Policy and Research (CIPR), Tulane University, New Orleans. Santana, Isidoro, and Magdalena Rathe. 1993. “The Distributive Impact of Fiscal Policy in the Dominican Republic.� in Government Spending and Income Distribution in Latin America, ed. by Ricardo Hausmann and Roberto Rigobon. Washington: Inter-American Development Bank. Scott, John. 2014. “Redistributive Impact and Efficiency of Mexico’s Fiscal System.� Public Finance Review. May 2014, Volume 42, Issue 3. SENASA. (2014). “Reconversión del IDSS y Red Pública única.� Mimeo. Valderrama, D., Francos, M., Jiménez, M., Lizardo, M., and Morillo, A. 2013. “El efecto de un programa de transferencias condicionadas sobre la pobreza y desigualdad en el corto plazo: El caso de Progresando con Solidaridad en República Dominicana.� MEPyD and ONE. Mimeo. Valderrama, D., Francos, M., Lizardo, M., and Morillo, A. 2013. “Seguro Familiar de Salud en el Régimen Subsidiado: Sus efectos de corto plazo sobre la pobreza y desigualdad.� ONE and MEPyD. Dominican Republic. Woldehanna, Tassew, Eyasu Tsehaye, Ruth Hill, Gabriela Inchauste, and Nora Lustig. 2014. “A Fiscal Incidence Analysis for Ethiopia.� Background Paper for World Bank Poverty Assessment (forthcoming). World Bank, 2006. “Achieving More Pro-Poor Growth.� Poverty Assessment, Report No. 32422-DO. The World Bank, October 30t, 2006. World Bank, 2014a. “When Prosperity is not Shared: the Puzzle of the Weak Links between Growth and Equity in Dominican Republic.� Equity Assessment, the World Bank. World Bank, 2014b. “Patronage or Reform? The Political Economy of Reform in the Dominican Republic.� An Institutional and Governance Review. Washington, DC. 59 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only World Bank. 2015. “Public Expenditure Management and Financial Accountability Review for the Dominican Republic.� Washington, DC. Mimeo. Younger, Stephen D. and Artsvi Khachatryan. 2014. “Fiscal Incidence in Armenia.� Background Paper for World Bank Armenia Public Expenditure Review (forthcoming). 60 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only 8. Annexes 8.1. Main taxes in the Dominican Republic Tax System in effect in 2012 (Source: CIAT http://www.ciat.org/index.php/en/products-and- services/ciatdata/countries/dominican-republic.html). The central government has authority over the following taxes: 1. Taxes on income, profits, and capital gains • Income taxes. - Income tax rates on wages and personal income. Table 20 Personal Income Tax, rates and thresholds Taxable Income (RD$) Rate on excess Less than 339,923 0% 339,923.01 to 599,884.00 15% of the amount over RD$339,923.01 599,884.01 to 773,173.00 RD$29,994.00 plus 20% of the amount over RD$599,884.01 Over 833,171.01 RD$76,652.00 plus 25% of the amount over RD$883,171.01 Source: Artículo 296, Codigo Tributario República Dominicana. - Income tax rate for dividends is 10% (Article 308, Tax Code). - Income tax rate for interest is 10% (Article 306b, Tax Code). • Asset taxes. 2. Taxes on properties • IPI/VSS tax on real estate, luxury homes and undeveloped urban plots, contained in Law No. 18- 88 and amendment laws. • Tax on the transfer of assets, contained in Law No. 288-04 and amendment laws. • Tax on the registry and recording of motor vehicles, contained in Law No. 557-05 and amendment laws. • Tax on successions and donations, contained in Law No. 25-69 and amendment laws. • Surtax on real-estate operations, contained in Law No. 3341 and amendment laws. 3. General consumption taxes • ITBIS – Tax on the transfer of industrial goods and services (VAT). 4. Excise taxes • ISC – excise tax. • Tax on hydrocarbons (2), contained in Law No. 112-00 and amendment laws. • Ad-valorem excise tax on internal consumption of fossil fuels and oil-based products contained in Law No. 557-05 and amendment laws. 61 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only • Departure payment (3), contained in Law No. 199-66 and amendment laws. 5. Taxes on financial transactions • Tax on checks and bank transfers. 6. Taxes on foreign trade • Tariffs, contained in Law No. 14-93 and amendment laws. 7. Others • Tax on the transit of motor vehicles, contained in Law No. 241-67 and amendment laws. • Tax on casinos, contained in Law No. 351 and amendment laws. • Single tax on sports betting, contained in Law No. 80-99 and amendment laws. The Tax Code, approved by means of Law No. 11-92 and amendment laws, rules income tax (Title II), ITBIS (Title III), ISC (Title IV), and tax on assets (Title V). Social Security payments are stipulated in Law No. 87-01 and amendment laws. 8.2. Recent tax reforms in the Dominican Republic Law 288-04; -raised ITBIS tax rate from 12% to 16%; -included advertising services on the ITBIS tax base with a rate of 6%; -introduced the payment of monthly advances in charge of the ISR by legal persons; -raised and unified the specific tax on spirits, eliminating this tax from the calculation of the base of any other tax; -raised the specific tax on tobacco products and arranged its annual adjustment for inflation; -introduced a selective tax of 10% on telecommunications services; -introduced a tax of 0.15% of the banking and financial transactions with the exception of ATM and withdrawal by “ventanilla�, and other operations; -introduced a tax on real estate property of 1% on the surplus value of ownership over RD$ 5 million. Law 557-05; -created a higher rate for the income tax of individuals with a rate of 30% temporarily, establishing a reduction schedule to reach the starting position (25% rate) within a period of three years; -introduced an ad-valorem selective tax of 13% on fuel consumption in replacement of the Exchange Commission; -introduced a tax to the registration of new motor vehicles entering the national territory of 17%, in replacement of the Exchange Commission; 62 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only -empowered the General Directorate of Internal Taxation to designate companies as agents of retention when operating with exempt persons or individuals. Law 495-06; -strengthened the coercive capacity of the General Directorate of Internal Taxation to attack evasion and enforceable efficiency; -amended taxes specific to spirits and tobacco products and introduced a selective ad-valorem tax of 15% on the value of the drinks and 100% on tobacco products; -introduced a selective tax on insurance services of 16%; -increased the rate of the tax selective ad-valorem on trade in fuels to 16%; -increased the flat tax on gambling. Act 139-11; -increased the load to the casinos, gambling in general and seats of lottery and sports betting, and froze the number of seats of betting and lottery for a period of 10 years; -temporarily raised the single rate of the tax on legal entities to 29%, for two years; -introduced a tax of 1% per year on the net financial assets of the financial institutions in the country, also as a temporary measure for two years; -introduced a tax of 2.5% on the value of the sales carried out in national territory to export zones, as part of the corporate income tax. Law 253-12; -an increase in the ITBIS rate from 16 percent to 18 percent; -coffee, sugar, oils, yoghourt and chocolates were included in the VAT tax base, experiencing a gradual increase in rates; -tax rates on alcoholic drinks increased from 7 percent to 15 percent, and tobacco increased to 20 percent; - a 10 percent rate will be applied to dividend payments (including free trade zones); -tax withholdings were extended to house renting and other activities; -a 1percent charge on the value of acquired motor vehicles; -an increase in the sales tax applied to FTZ manufactured products sold in national territory. 63 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only 8.3. Assumptions Personal income taxes We only estimate the personal income tax for wages, dividends, and interest. Tax evasion among the self-employed is considered significant and we do not have access to profiles of payments of independent business, so we do not calculate personal income taxes from these sources. In addition, we do not use assumptions on informality of wage earners. We estimate taxable income using annualized wage income from primary and secondary sources for those who report to work as employees or blue-collar workers (obrero) in the General Government, Public Institution, Private Institution or Free Trade Zone categories. We assume the household survey includes labor income gross of taxes, because ENIGH 2007 survey asks for gross salary (without deductions). Personal income taxes were estimated by applying statutory rates and income brackets from 2013 (in prices of 2007) to salaries (see table below). We annualized interest and dividend income and applied a 10 percent rate to these income sources. Table 21. Personal Income Tax, rates and thresholds, adjusted to 2007 prices Income (in 2007 Dominican pesos) Taxes paid From To 0 291,557.69 0 291,557.70 437,336.17 15% in excess of 291,557.70 437,336.18 604,710.45 21,867 plus 20% in excess of 437,336.17 604,710.46 55,34.62 plus 25% in excess of 437,336.17 Source: Authors’ estimations using Personal Income Tax Schedule 2013 and Consumer Price Index. Indirect taxes We estimate indirect taxes using the simulation method. The analysis includes ITBIS, Selective on Consumption (ISC), telecommunications taxes, and insurance taxes. ENIGH 2007 has a detailed list of goods and services purchased by households, identified according to the Classification of Individual Consumption According to Purpose (COICOP). For ITBIS, we apply a general statutory rate of 18 percent and reduced rate of 8 percent, adhering to the Tax Code of 2013. In addition, we omit products exempted by law and incorporate tax evasion assumptions. For this purpose, we create four groups of goods and services: i. Product with an assumption of a high propensity of evasion that paid zero; ii. Products with an assumption of a high propensity to pay ITBIS (VAT) that paid the general rate (18 percent) or reduced rate (8 percent) (taxable consumption (net of ITBIS at applicable rate in 2007) * applicable rate in 2013); iii. Products with estimated compliance rate, using a detailed list provided by DGII, for those goods we applied estimated compliance and general or reduced rate (taxable consumption * estimated compliance by product * applicable rate in 2013); and iv. Products that paid VAT conditional to the place of purchase (taxable consumption * estimated compliance by product by place of purchase * applicable rate in 2013). 64 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Table 22 and Table 23 present the results. We estimate excise taxes on alcoholic beverages, beer, tobacco, oil derivate products, and vehicles. Excise taxes on alcoholic beverages, tobacco, and beer were simulated. From reported consumption in household surveys, we subtracted the estimated value of ITBIS and selective taxes applicable in 2007, obtaining the tax base on which to apply rates in 2013 (10 percent for beer and alcoholic drinks; 20 percent for tobacco). For tax on oil derivate products, we apply effective rates on value net of ITBIS and excises. Indirect taxes were down-scaled to prevent overestimation using the method in Lustig and Higgins (2013). In the case of VAT, for example, we adjust VAT payments to equalize the ratio of total VAT to disposable income to VAT collection (national accounts) to private consumption (national accounts). Table 22. VAT incidence with and without the evasion assumption Evasion No evasion Deciles 1 -4.2% -4.9% 2 -4.5% -5.1% 3 -4.5% -5.2% 4 -4.5% -5.2% 5 -4.8% -5.5% 6 -5.0% -5.7% 7 -5.1% -5.8% 8 -5.1% -5.8% 9 -5.5% -6.2% 10 -6.2% -7.2% Total Population -5.5% -6.3% Source: Authors’ estimates using ENIGH 2007 and DGII data. Table 23. VAT concentration shares with and without the evasion assumption Evasion No evasion Deciles 1 1.0% 1.0% 2 2.0% 2.0% 3 2.8% 2.8% 4 3.6% 3.6% 5 4.9% 4.8% 6 6.2% 6.1% 7 7.8% 7.8% 8 10.3% 10.1% 9 16.1% 15.6% 10 45.3% 46.1% Total Population 100.0% 100.0% Source: Authors’ estimates using ENIGH 2007 and DGII data. 65 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Direct transfers For those programs, we simulate benefits according to the program rules, assuming reported coverage and targeting according to the SIUBEN poor household categories, where applicable. With reported coverage, we use a random selection of beneficiaries. For other programs that did not use SIUBEN, we randomly select the population that matched the profile of beneficiaries.48 However, we need to cross-check the results with ADESS administrative data. Because we found some differences, we adjust coverage and transfers to changes in population and relative prices between 2013 and 2007. First, we adjust beneficiaries to take into consideration changes in population between 2007 and 2013; second, we calculate transfers in 2007 pesos; third, we adjust coverage to replicate the ADESS population of beneficiaries; fourth, we estimate total benefits by multiplying estimated benefits in 2007 prices by adjusted coverage; fifth, we apply downscaling to equalize any differences between the benefits/disposable income ratio of the national accounts and household survey. Table 24. Adjustments on direct transfer, ADESS data. PROGRAMS AVERAGE ADESS SIUBEN EQUIVALEN ESTIMATED ADESS BENEFICIARI CATEGORI T POPULATION BENEFICIARI ES (ADJUSTED ES COVERAGE OF ADESS ES IN 2013 TO BENEFICIARI POPULATION ES IN 2007) IDENTIFIED IN ENIGH 2007 COMER ES PRIMERO 596,768 1, 2 60.11% 596,628 636,488 INCENTIVO A LA ASISTENCIA ESCOLAR 251,775 1, 2 72.78% 251,545 268,533 SUPLEMENTO ALIMENTICIO- 77,919 1, 2 28.20% 78,017 ENVEJECIENTES 83,105 INCENTIVO A LA EDUCACIÓN SUPERIOR 22,163 1, 2, 3 27.00% 21,794 23,639 BONOGAS HOGAR 738,401 1, 2, 3 35.64% 738,432 787,548 BONOGAS CHOFER 14,627 1, 2 19.7% 14,691 15,601 BONO LUZ 486,988 1, 2, 3 65.68% 486,920 519,402 BONO ESCOLAR ESTUDIANDO 41,543 1, 2 37.46% 43,887 PROGRESO 44,308 Source: ADESS. We imputed transfer programs linked to education—School Food Program; Shoes, Uniform, and Backpacks programs—using coverage of education by level and average costs at 2007 prices. The simulation for Quisqueya Aprende Contigo, the alphabetization program, relies on comparisons of people with illiteracy and randomly selected households to replicate reported coverage. Subsidies The subsidy on electricity works through three schemes, one of them is Bono Luz, estimated as a direct transfer. For the scheme applied to the tariff, we use 2007 prices to estimate the implicit kwh 48 Programa de Incentivo a la Policía Preventiva, Programa Incentivo a los Alistados de la Marina de Guerra. 66 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only consumed by each household and apply the subsidy to users with less than 700 kwh monthly. The subsidy is equal to the full tariff minus the estimated payment based on the subsidy scheme. The second scheme is provision non-paid electricity for those households without an electric meter who fall into the SIUBEN categories 1 to 4. Health We divide health expenditures into four categories, but we could only find information for the first three: Ministry of Health, social security institutions, PROMESE, and others.49 Our main sources of information were the Health National Accounts, estimated by the Ministry of Health for 2010 to 2013, and the ENDESA Demographic and Health Survey. Using these sources, we could perform a disaggregated estimation of public health expenditures. Ministry of Health spending is divided between hospitals and outpatient care. First, we perform a propensity match score to capture up-to-date information on use of health services from ENDESA 2013 for each category of the non-insured population.50 Second, we estimate total expenditures in 2013 and deflate them to obtain spending in 2007 pesos for both categories. Next, we impute per-household transfer by dividing the previous figure between households that report using of health services. Finally, we downscale in-kind transfer values to match the national accounts ratio. The Dominican social security system includes three health-insurance regimes: contributory, subsidized, and contributory-subsidized. Contributory works like private insurance, even when involving public insurers, and we decided not to estimate this regime. The contributory-subsidized regime was not operational in 2013. So we estimate only the subsidized regime. Using coverage information from ENDESA 2013, we impute average per-capita expenditures for this insurance, deflated to 2007 pesos.51 However, some people are included in social security institutions created before the reform, and we identify affiliates using ENDESA 2013 and performing matching-score analysis. One of these is the retirees and pensioners regime that receives a subsidy from the central government.52 Finally, Dominican Institute of Social Security, a public pay-as-you-go regime of social security, received a transfer to cover the deficit in these institutions. We estimate the per-capita transfer to affiliates in 2007 pesos, using data provided by SENASA (2014). In addition, we prevent overestimation using the Lustig and Higgins (2013) method, adjusting the ratio to education expenditures to disposable income to the same ratio calculated from national accounts. For PROMESE, after identifying beneficiaries using ENDESA 2013, we randomly select beneficiaries to match coverage in 2013 (2.78 percent) and impute average values in 2007 pesos. Education The Dominican Republic’s education system presents several challenges for CEQ evaluation. The survey reports whether individuals attend school, the level of education, and whether the school is private or public. The education benefit is based on the cost per student by level estimated by UNESCO and Minerd. 49 It is equivalent to 10 percent of public spending on health and includes the Military Hospital, Police Hospital, Commission on AIDS and others. 50 The insured population was addressed in social security spending in health. 51 Estimation used data from Consejo Nacional de la Seguridad Social from 2013, disbursements of Government divided between the number of affiliates. 52 In June 2013, there were 10,553 affiliates, based on information of Senasa. Each one received a uniform benefit equal to 15,158 Dominican pesos of 2007. 67 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only To include only expenditures in education, we exclude benefits like the school food program, uniforms, and alphabetization, all of which were estimated as direct transfers. We adjust levels of education to UNESCO International Standard Classification of Education. The first six years of Básica are equivalent to primary education and the last two years could be categorized as lower secondary level. These adjustments made it possible to use estimations of expenditure by level and make international comparisons. We adjust these figures to 2007 prices. In addition, we prevent overestimation using the Lustig and Higgins (2013) method, adjusting the ratio of education expenditures to disposable income to the ratio calculated form national accounts. Because public education expenditures increased in 2013, we estimate a Sensitivity Analysis 2 with lower expenditure level of 2011. The gross coverage rate in primary school changed little from 2007 to 2013, and the changes in elementary and secondary schools have not been very large. This allowed us to assume that coverage has not been modified in all scenarios analyzed. 68 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only 8.4. Basic characteristics of individuals and households in the bottom decile Income-based incidence analysis Among individuals and households, the poorest and richest differ sharply in the Dominican Republic. The bottom market-income decile has 922,955 individuals living in 197,793 households. The mean per capita market income of RD$9,456 in 2007 compared to RD$296,428 for the top decile. Compared to the top decile, bottom-decile households are less urban, larger, younger, less educated, and more likely to include children and more pensioners. Among the bottom-decile households, 38 percent live in urban areas, compared to 87 percent for top-decile households. The average household size for the bottom decile is 5.6, while the top decile is at 3.5. Households in the bottom decile have an average age of 24.1 years, almost eight years lower than the 32.8 years in the top decile. Furthermore, the average number of years of schooling for household heads in the bottom decile is 4.1 years, compared to 11.5 years for household heads in the top decile. About 85 percent of bottom-decile households have children below 18 years of age, compared to 40 percent of the households in the top decile. Descriptive results further show that there is a higher proportion of pension-aged members in bottom-decile households (23 percent) than in top-decile households (11 percent). Table 25. Dominican Republic: Features of households at bottom and top of the income distribution Poorest Decile Richest Decile % of household living in urban areas 38.76 87.33 Average Household Size 5.6 3.5 Average Age 24.1 32.8 Years of Schooling of HH Head 4.1 11.5 % households with children <18 years 85.1 40.3 % households with pension age adult (>65 years) 22.9 11.3 Per Capita annual Income before taxes and transfers (RD$) 9,456 14,691 Per Capita annual Income after taxes and transfers (RD$ 296,428 264,648 Source: Authors’ estimates using ENIGH 2007. Consumption-based incidence analysis A consumption-based53 methodology varies somewhat from the income-based approach. In this alternative view, there are 923,561 individuals with a total of 387,541 households in the bottom decile. Under the consumption approach, 66 percent of the households in the bottom decile have zero market incomes. The mean per capita market income in the bottom decile is RD$35,375, compared to RD$222,093 in the top decile. This approach yields similar results, finding that bottom-decile households is less urban, larger, younger, less educated, and more likely to include children and more pensioners. Descriptive statistics show that 49 percent of bottom-decile households live in urban areas, compared with 90 percent for top- decile households. The average household size is 4.7 in the bottom decile and 2.7 in the top decile. The average age in the bottom decile is 24.1 years, compared to 32.8 years in the top decile. In addition, the average years of schooling for household heads is 4.1 years in the bottom decile and 11.5 years in the top decile. Further, 85.1 percent of bottom-decile households have children below 18 years of age, compared to 53 Using per capita household consumption reported in survey. 69 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only 40.3 percent for households in the top decile. In addition, 22.9 percent of the bottom decile households have at least a pension-aged adult, compared to only 11.3 percent in the top decile. 70 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only 8.5. Tables containing additional results Table 26. Reduction in inequality across income concepts Net Market Disposable Post-fiscal Final Final Dominican Republic Market Income Income Income Income* Income Income Benchmark Gini 0.514 0.509 0.502 0.492 0.470 0.458 change wrt market income -- -0.005 -0.012 -0.022 -0.044 -0.056 Significance (p-value) -- 0.000 0.000 0.000 0.000 0.000 change wrt net market income -- -- -0.006 -0.017 -0.039 -0.050 Significance (p-value) -- -- 0.000 0.000 0.000 0.000 Theil Index 0.521 0.506 0.495 0.468 0.440 0.413 % change wrt market income -- -1.5% -2.6% -5.3% -8.1% -10.9% Significance (p-value) -- 0.000 0.000 0.000 0.000 0.000 % change wrt net market income -- -- -1.2% -3.8% -6.7% -9.4% Significance (p-value) -- -- 0.000 0.000 0.000 0.000 90/10 10.408 10.342 9.692 9.289 7.512 7.138 % change wrt market income -- -0.1% -0.7% -1.1% -2.9% -3.3% Significance (p-value) -- 0.169 0.000 0.000 0.000 0.000 % change wrt net market income -- -- -0.7% -1.1% -2.8% -320.4% Significance (p-value) -- -- 0.000 0.000 0.000 0.000 Sensitivity Analysis 1 Gini 0.5141 0.5086 0.5021 0.4921 0.4699 0.4583 change wrt market income -- -0.5% -1.2% -2.2% -4.4% -5.6% Significance (p-value) -- 0.000 0.000 0.000 0.000 0.000 change wrt net market income -- -- -0.006 -0.017 -0.039 -0.050 Significance (p-value) -- -- 0.000 0.000 0.000 0.000 Theil Index 0.521 0.506 0.495 0.468 0.440 0.412 % change wrt market income -- -0.015 -0.027 -0.053 -0.082 -0.109 Significance (p-value) -- 0.000 0.000 0.000 0.000 0.000 % change wrt net market income -- -- -1.2% -3.8% -6.7% -9.4% Significance (p-value) -- -- 0.000 0.000 0.000 0.000 90/10 10.408 10.342 9.695 9.303 7.516 7.140 % change wrt market income -- -0.1% -0.7% -1.1% -2.9% -3.3% Significance (p-value) -- 0.169 0.000 0.000 0.000 0.000 % change wrt net market income -- -- -0.6% -1.0% -2.8% -3.2% Significance (p-value) -- -- 0.000 0.000 0.000 0.000 Sensitivity Analysis 2 Gini 0.514 0.509 0.503 0.493 0.480 0.469 change wrt market income -- -0.005 -0.011 -0.021 -0.034 -0.045 Significance (p-value) -- 0.000 0.000 0.000 0.000 0.000 change wrt net market income -- -- -0.006 -0.016 -0.028 -0.040 Significance (p-value) -- -- 0.000 0.000 0.000 0.000 Theil Index 0.521 0.506 0.496 0.470 0.457 0.430 % change wrt market income -- -0.015 -0.025 -0.052 -0.064 -0.091 Significance (p-value) -- 0.000 0.000 0.000 0.000 0.000 % change wrt net market income -- -- -0.011 -0.037 -0.049 -0.077 Significance (p-value) -- -- 0.000 0.000 0.000 0.000 90/10 10.408 10.342 9.755 9.357 8.090 7.626 % change wrt market income -- -0.1% -0.7% -1.1% -2.3% -2.8% Significance (p-value) -- 0.169 0.000 0.000 0.000 0.000 % change wrt net market income -- -- -0.6% -1.0% -2.3% -2.7% Significance (p-value) -- -- 0.000 0.000 0.000 0.000 Source: Authors’ estimates based in ENIGH 2007, applying the CEQ methodology. 71 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Table 27. Reduction in inequality across income concepts Market Net Market Disposable Post-fiscal Dominican Republic Income Income Income Income US$ 1.25 PPP Headcount index 5.7% 5.7% 4.7% 4.9% % change wrt market income -- 0.0% -1.0% -0.8% Significance (p-value) -- 0.000 0.000 0.000 % change wrt net market income -- -- -1.0% -0.8% Significance (p-value) -- -- 0.000 0.000 Poverty Gap 1.968% 1.968% 1.478% 1.502% % change wrt market income -- 0.0% -0.5% -0.5% Significance (p-value) -- 0.000 0.000 0.000 % change wrt net market income -- -- -0.5% -0.5% Significance (p-value) -- -- 0.000 0.000 Squared Poverty Gap 1.026% 1.026% 0.698% 0.705% % change wrt market income -- 0.0% -0.3% -0.3% Significance (p-value) -- 0.000 0.000 0.000 % change wrt net market income -- -- -0.3% -0.3% Significance (p-value) -- -- 0.000 0.000 US$2.50 PPP Headcount index 19.513% 19.527% 18.242% 19.474% % change wrt market income -- 0.0% -1.3% 0.0% Significance (p-value) -- 0.282 0.000 0.787 % change wrt net market income -- -- -1.286% -0.1% Significance (p-value) -- -- 0.000 0.711 Poverty Gap 7.094% 7.094% 6.223% 6.538% % change wrt market income -- 0.0% -0.9% -0.6% Significance (p-value) -- 0.084 0.000 0.000 % change wrt net market income -- -- -0.9% -0.6% Significance (p-value) -- -- 0.000 0.000 Squared Poverty Gap 3.692% 3.692% 3.071% 3.196% % change wrt market income -- 0.0% -0.6% -0.5% Significance (p-value) -- 0.152 0.000 0.000 % change wrt net market income -- -- -0.6% -0.5% Significance (p-value) -- -- 0.000 0.000 US$4 PPP Headcount index 36.951% 36.951% 35.877% 37.660% % change wrt market income -- 0.0% -1.1% 0.7% Significance (p-value) -- 0.000 0.000 0.000 % change wrt net market income -- -- -1.1% 0.7% Significance (p-value) -- -- 0.000 0.000 Poverty Gap 14.965% 14.969% 13.995% 14.726% % change wrt market income -- 0.0% -1.0% -0.2% Significance (p-value) -- 0.000 0.000 0.000 % change wrt net market income -- -- -1.0% -0.2% Significance (p-value) -- -- 0.000 0.000 Squared Poverty Gap 8.293% 8.294% 7.479% 7.859% % change wrt market income -- 0.0% -0.8% -0.4% Significance (p-value) -- 0.003 0.000 0.000 % change wrt net market income -- -- -0.8% -0.4% Significance (p-value) -- -- 0.000 0.000 National Extreme Headcount index 13.755% 13.755% 12.493% 13.147% Poverty Line % change wrt market income -- 0.0% -1.3% -0.6% (US$2.09 urban PPP, US$2.00 rural Significance (p-value) -- 0.000 0.000 PPP) % change wrt net market income -- -- -1.3% -0.6% Significance (p-value) -- -- 0.000 0.000 Poverty Gap 4.753% 4.754% 3.987% 4.165% % change wrt market income -- 0.0% -0.8% -0.6% Significance (p-value) -- 0.179 0.000 0.000 % change wrt net market income -- -- -0.8% -0.6% Significance (p-value) -- -- 0.000 0.000 Squared Poverty Gap 2.452% 2.452% 1.939% 2.000% % change wrt market income -- 0.0% -0.5% -0.5% Significance (p-value) -- 0.144 0.000 0.000 % change wrt net market income -- -- -0.5% -0.5% Significance (p-value) -- -- 0.000 0.000 National Moderate Headcount index 41.181% 41.200% 40.101% 42.281% Poverty Line % change wrt market income -- 0.0% -1.1% 1.1% (US$4.63 urban PPP, US$4.13 rural Significance (p-value) -- 0.090 0.000 0.000 PPP) % change wrt net market income -- -- -1.1% 1.1% Significance (p-value) -- -- 0.000 0.000 Poverty Gap 16.9% 16.9% 15.9% 16.8% % change wrt market income -- 0.0% -1.0% -0.1% Significance (p-value) -- 0.000 0.000 0.000 % change wrt net market income -- -- -1.0% -0.1% Significance (p-value) -- -- 0.000 0.000 Squared Poverty Gap 9.4% 9.4% 8.6% 9.0% % change wrt market income -- 0.0% -0.8% -0.4% Significance (p-value) -- 0.001 0.000 0.000 % change wrt net market income -- -- -0.8% -0.4% Significance (p-value) -- -- 0.000 0.000 72 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Market Net Market Disposable Post-fiscal Dominican Republic Income Income Income Income Sensitivity Analysis 1 US$1.25 PPP Headcount index 5.726% 5.726% 4.696% 4.920% % change wrt market income -- 0.0% -1.0% -0.8% Significance (p-value) -- 0.000 0.000 0.000 % change wrt net market income -- -- -1.0% -0.8% Significance (p-value) -- -- 0.000 0.000 Poverty Gap 1.968% 1.968% 1.478% 1.502% % change wrt market income -- 0.0% -0.5% -0.5% Significance (p-value) -- 0.000 0.000 0.000 % change wrt net market income -- -- -0.5% -0.5% Significance (p-value) -- -- 0.000 0.000 Squared Poverty Gap 1.026% 1.026% 0.698% 0.705% % change wrt market income -- 0.0% -0.3% -0.3% Significance (p-value) -- 0.000 0.000 0.000 % change wrt net market income -- -- -0.3% -0.3% Significance (p-value) -- -- 0.000 0.000 US$2.50 PPP Headcount index 19.513% 19.527% 18.223% 19.474% % change wrt market income -- 0.0% -1.3% 0.0% Significance (p-value) -- 0.282 0.000 0.787 % change wrt net market income -- -- -1.3% -0.1% Significance (p-value) -- -- 0.000 0.000 Poverty Gap 7.094% 7.094% 6.220% 6.535% % change wrt market income -- 0.0% -0.9% -0.6% Significance (p-value) -- 0.084 0.000 0.000 % change wrt net market income -- -- -0.9% -0.6% Significance (p-value) -- -- 0.000 0.000 Squared Poverty Gap 3.692% 3.692% 3.069% 3.195% % change wrt market income -- 0.0% -0.6% -0.5% Significance (p-value) -- 0.152 0.000 0.000 % change wrt net market income -- -- -0.6% -0.5% Significance (p-value) -- -- 0.000 0.000 US$4 PPP Headcount index 36.951% 36.951% 35.809% 37.599% % change wrt market income -- 0.0% -1.1% 0.6% Significance (p-value) -- 0.000 0.000 0.000 % change wrt net market income -- -- -1.1% 0.6% Significance (p-value) -- -- 0.000 0.000 Poverty Gap 14.965% 14.969% 13.975% 14.705% % change wrt market income -- 0.0% -1.0% -0.3% Significance (p-value) -- 0.000 0.000 0.000 % change wrt net market income -- -- -1.0% -0.3% Significance (p-value) -- -- 0.000 0.000 Squared Poverty Gap 8.293% 8.294% 7.470% 7.849% % change wrt market income -- 0.0% -0.8% -0.4% Significance (p-value) -- 0.003 0.000 0.000 % change wrt net market income -- -- -0.8% -0.4% Significance (p-value) -- -- 0.000 0.000 National Extreme Headcount index 13.755% 13.755% 12.483% 13.137% Poverty Line % change wrt market income -- 0.0% -1.3% -0.6% (US$2.09 PPP urban, US$2.00 PPP Significance (p-value) -- 0.000 0.000 0.000 rural) % change wrt net market income -- -- -1.3% -0.6% Significance (p-value) -- -- 0.000 0.000 Poverty Gap 4.8% 4.8% 4.0% 4.2% % change wrt market income -- 0.0% -0.8% -0.6% Significance (p-value) -- 0.179 0.000 0.000 % change wrt net market income -- -- -0.8% -0.6% Significance (p-value) -- -- 0.000 0.000 Squared Poverty Gap 2.5% 2.5% 1.9% 2.0% % change wrt market income -- 0.0% -0.5% -0.5% Significance (p-value) -- 0.144 0.000 0.000 % change wrt net market income -- -- -0.5% -0.5% Significance (p-value) -- -- 0.000 0.000 National Moderate Headcount index 41.2% 41.2% 40.0% 42.2% Poverty Line % change wrt market income -- 0.0% -1.1% 1.0% (US$4.63 PPP urban, US$4.13 PPP Significance (p-value) -- 0.090 0.000 0.000 rural) % change wrt net market income -- -- -1.2% 1.0% Significance (p-value) -- -- 0.000 0.000 Poverty Gap 16.901% 16.904% 15.890% 16.733% % change wrt market income -- 0.0% -1.0% -0.2% Significance (p-value) -- 0.000 0.000 0.000 % change wrt net market income -- -- -1.0% -0.2% Significance (p-value) -- -- 0.000 0.000 Squared Poverty Gap 9.413% 9.414% 8.559% 9.002% % change wrt market income -- 0.0% -0.9% -0.4% Significance (p-value) -- 0.001 0.000 0.000 % change wrt net market income -- -- -0.9% -0.4% Significance (p-value) -- -- 0.000 0.000 73 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Market Net Market Disposable Post-fiscal Dominican Republic Income Income Income Income Sensitivity Analysis 2 US$ 1.25 PPP Headcount index 5.726% 5.726% 4.762% 5.076% % change wrt market income -- 0.0% -1.0% -0.7% Significance (p-value) -- 0.000 0.000 0.000 % change wrt net market income -- -- -1.0% -0.7% Significance (p-value) -- -- 0.000 0.000 Poverty Gap 2.0% 2.0% 1.5% 1.6% % change wrt market income -- 0.0% -0.4% -0.4% Significance (p-value) -- 0.000 0.000 0.000 % change wrt net market income -- -- -0.4% -0.4% Significance (p-value) -- -- 0.000 0.000 Squared Poverty Gap 1.03% 1.03% 0.73% 0.74% % change wrt market income -- 0.0% -0.3% -0.3% Significance (p-value) -- 0.000 0.000 0.000 % change wrt net market income -- -- -0.3% -0.3% Significance (p-value) -- -- 0.000 0.000 US$ 2.50 PPP Headcount index 19.513% 19.527% 18.372% 19.586% % change wrt market income -- 0.0% -1.1% 0.1% Significance (p-value) -- 0.282 0.000 0.602 % change wrt net market income -- -- -1.2% 0.1% Significance (p-value) -- -- 0.000 0.679 Poverty Gap 7.094% 7.094% 6.307% 6.629% % change wrt market income -- 0.0% -0.8% -0.5% Significance (p-value) -- 0.084 0.000 0.000 % change wrt net market income -- -- -0.8% -0.5% Significance (p-value) -- -- 0.000 0.000 Squared Poverty Gap 3.692% 3.692% 3.130% 3.259% % change wrt market income -- 0.0% -0.6% -0.4% Significance (p-value) -- 0.152 0.000 0.000 % change wrt net market income -- -- -0.6% -0.4% Significance (p-value) -- -- 0.000 0.000 US$4 PPP Headcount index 36.951% 36.951% 35.926% 37.703% % change wrt market income -- 0.0% -1.0% 0.8% Significance (p-value) -- 0.000 0.000 0.000 % change wrt net market income -- -- -1.0% 0.8% Significance (p-value) -- -- 0.000 0.000 Poverty Gap 14.965% 14.969% 14.091% 14.825% % change wrt market income -- 0.0% -0.9% -0.1% Significance (p-value) -- 0.000 0.000 0.000 % change wrt net market income -- -- -0.9% -0.1% Significance (p-value) -- -- 0.000 0.000 Squared Poverty Gap 8.293% 8.294% 7.559% 7.942% % change wrt market income -- 0.0% -0.7% -0.4% Significance (p-value) -- 0.003 0.000 0.000 % change wrt net market income -- -- -0.7% -0.4% Significance (p-value) -- -- 0.000 0.000 National Extreme Headcount index 13.755% 13.755% 12.610% 13.232% Poverty Line % change wrt market income -- 0.0% -1.1% -0.5% (US$2.09 PPP urban, US$2.00 PPP Significance (p-value) -- 0.000 0.000 0.001 rural) % change wrt net market income -- -- -1.1% -0.5% Significance (p-value) -- -- 0.000 0.001 Poverty Gap 4.753% 4.754% 4.062% 4.243% % change wrt market income -- 0.0% -0.7% -0.5% Significance (p-value) -- 0.179 0.000 0.000 % change wrt net market income -- -- -0.7% -0.5% Significance (p-value) -- -- 0.000 0.000 Squared Poverty Gap 2.452% 2.452% 1.987% 2.051% % change wrt market income -- 0.0% -0.5% -0.4% Significance (p-value) -- 0.144 0.000 0.000 % change wrt net market income -- -- -0.5% -0.4% Significance (p-value) -- -- 0.000 0.000 National Moderate Headcount index 41.181% 41.200% 40.158% 42.301% Poverty Line % change wrt market income -- 0.0% -1.0% 1.1% (US$4.63 PPP urban, US$4.13 PPP Significance (p-value) -- 0.090 0.000 0.000 rural) % change wrt net market income -- -- -1.0% 1.1% Significance (p-value) -- -- 0.000 0.000 Poverty Gap 16.901% 16.904% 16.009% 16.859% % change wrt market income -- 0.0% -0.9% 0.0% Significance (p-value) -- 0.000 0.000 0.037 % change wrt net market income -- -- -0.9% 0.0% Significance (p-value) -- -- 0.000 0.028 Squared Poverty Gap 9.413% 9.414% 8.653% 9.100% % change wrt market income -- 0.0% -0.8% -0.3% Significance (p-value) -- 0.001 0.000 0.000 % change wrt net market income -- -- -0.8% -0.3% Significance (p-value) -- -- 0.000 0.000 74 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Table 28. Incidence for taxes and transfers (share of market income and socioeconomic group) Benchmark scenario Deciles All Other Transfers Direct Taxes All Taxes Net Comer Quisqueya Direct All Net Post- In- (excluding and Bono Disposable Indirect Indirect In-kind In-kind (Direct Final Market es BGH aprende Transfers Direct Indirect Fiscal kind all Taxes) Contributions Luz Income Subsidies Taxes Education Transfer and Income Income primero contigo (Targeted Transfers Taxes Income Health plus to SS Indirect) or Not) Indirect Subsidies Deciles 1 0.0 0.0 3.9 0.9 0.9 0.9 4.3 11.0 11.0 3.3 -5.1 -1.8 9.2 37.0 21.6 58.6 72.8 -5.1 67.7 2 0.0 0.0 2.0 0.5 0.5 0.4 2.0 5.4 5.3 2.5 -5.5 -3.0 2.3 17.5 10.3 27.8 35.6 -5.5 30.2 3 0.0 0.0 1.1 0.3 0.3 0.3 1.4 3.4 3.4 1.9 -5.3 -3.4 0.0 12.1 7.4 19.5 24.8 -5.3 19.4 4 0.0 0.0 0.9 0.3 0.3 0.2 0.8 2.4 2.3 1.8 -5.5 -3.6 -1.3 9.0 4.9 13.9 18.1 -5.5 12.6 5 0.0 0.0 0.6 0.2 0.3 0.2 0.5 1.7 1.7 1.9 -6.0 -4.1 -2.4 6.6 3.5 10.1 13.8 -6.1 7.7 6 0.0 0.0 0.4 0.2 0.2 0.1 0.4 1.3 1.2 1.6 -6.3 -4.6 -3.4 4.7 2.4 7.1 10.0 -6.3 3.7 7 -0.1 -0.1 0.3 0.1 0.2 0.1 0.3 1.0 0.9 1.6 -6.4 -4.8 -3.9 3.3 1.6 4.9 7.5 -6.4 1.1 8 -0.1 -0.1 0.2 0.1 0.1 0.1 0.1 0.6 0.5 1.6 -6.4 -4.9 -4.3 2.2 1.1 3.3 5.5 -6.5 -1.1 9 -0.5 -0.5 0.1 0.1 0.1 0.0 0.1 0.4 -0.2 1.3 -7.4 -6.1 -6.2 1.4 0.5 1.9 3.5 -7.9 -4.4 10 -3.0 -3.0 0.0 0.0 0.0 0.0 0.0 0.1 -2.9 0.6 -9.0 -8.4 -11.3 0.4 0.1 0.5 1.2 -12.0 -10.8 Total Population -1.3 -1.3 0.3 0.1 0.1 0.1 0.3 0.9 -0.4 1.2 -7.5 -6.2 -6.6 3.3 1.7 5.0 7.2 -8.791 -1.6 Socioeconomic Groups All Other Transfers All Direct Taxes Net Comer Direct All Post- In- (excludin Taxes Bon Quisquey Indirect Net In-kind In-kind Final and Market es BG Transfers Direct Disposabl Indirec Fiscal kind g all (Direct o a aprende Subsidie Indirec Educatio Transfe Incom Contribution Incom primer H (Targete Transfer e Income t Taxes Incom Healt Taxes) and Luz contigo s t Taxes n r e s to SS e o d or s e h plus Indirect Not) Indirect ) Group: Subsidies y < 1.25 0.0 0.0 5.55 1.14 1.15 1.18 5.92 14.9 14.9 4.4 -4.7 -0.3 14.6 50.9 28.6 79.5 98.8 -4.7 94.1 1.25 < = y < 2.50 0.0 0.0 2.15 0.51 0.57 0.52 2.29 6.0 6.0 2.5 -5.4 -2.9 3.1 19.9 11.9 31.8 40.3 -5.4 34.9 2.50 <= y < 4.00 0.0 0.0 1.00 0.28 0.31 0.27 1.15 3.0 3.0 1.9 -5.4 -3.5 -0.6 11.1 6.4 17.5 22.4 -5.4 17.0 4.00 <= y < 10.00 0.0 0.0 0.39 0.16 0.17 0.11 0.32 1.2 1.1 1.7 -6.3 -4.6 -3.5 4.2 2.2 6.4 9.2 -6.3 2.9 10.00 <= y < -7.8 -6.8 -8.2 50.00 -1.6 -1.6 0.06 0.05 0.05 0.03 0.04 0.2 -1.3 1.0 0.9 0.3 1.2 2.5 -9.4 -6.9 50.00 <= y -4.1 -4.1 0.00 0.00 0.01 0.00 0.00 0.0 -4.1 0.2 -10.4 -10.1 -14.2 0.1 0.0 0.1 0.4 -14.5 -14.1 Total Population -1.3 -1.3 0.31 0.11 0.12 0.09 0.29 0.9 -0.4 1.2 -7.5 -6.2 -6.6 3.3 1.7 5.0 7.2 -8.8 -1.6 Source: Authors’ estimates based in ENIGH 2007, applying the CEQ methodology. 75 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Sensitivity Analysis 1 (contributory pensions as a transfer) Deciles All Transfers Direct Taxes All Taxes Net Net Post- In- Housing (excluding all and Contributory All Direct Disposable Indirect Indirect In-kind In-kind (Direct Final Market Indirect Fiscal kind and Taxes) plus Contributions Pensions Transfers Income Subsidies Taxes Education Transfers and Income Income Taxes Income Health Urban Indirect to SS Indirect) Subsidies Deciles 1 0.0 0.0 0.0 11.0 11.0 3.3 -5.1 -1.8 9.2 37.0 21.6 58.6 72.9 -5.1 67.8 2 0.0 0.0 0.0 5.4 5.4 2.5 -5.5 -3.0 2.4 17.5 10.3 27.8 35.7 -5.5 30.2 3 0.0 0.0 0.2 3.6 3.5 1.9 -5.3 -3.4 0.1 12.1 7.4 19.5 25.0 -5.3 19.6 4 0.0 0.0 0.1 2.4 2.4 1.8 -5.5 -3.6 -1.2 9.0 4.9 13.9 18.2 -5.5 12.7 5 0.0 0.0 0.2 1.9 1.9 1.9 -6.0 -4.1 -2.3 6.6 3.5 10.1 13.9 -6.1 7.9 6 0.0 0.0 0.2 1.4 1.4 1.6 -6.3 -4.6 -3.2 4.7 2.4 7.1 10.2 -6.3 3.9 7 -0.1 -0.1 0.2 1.2 1.1 1.6 -6.4 -4.8 -3.6 3.3 1.6 4.9 7.7 -6.4 1.3 8 -0.1 -0.1 0.2 0.9 0.8 1.6 -6.4 -4.9 -4.1 2.2 1.1 3.3 5.7 -6.5 -0.8 9 -0.5 -0.5 0.2 0.6 0.0 1.3 -7.4 -6.1 -6.0 1.4 0.5 1.9 3.7 -7.9 -4.2 10 -3.0 -3.0 0.1 0.2 -2.8 0.6 -9.0 -8.4 -11.2 0.4 0.1 0.5 1.2 -12.0 -10.8 Total Population -1.3% -1.3 0.1 1.1 -0.3 1.2 -7.5 -6.2 -6.5 3.3 1.7 5.0 7.3 -8.8 -1.5 Socioeconomic Groups All In-kind Transfers All Direct Taxes Transfers Net All Net Post- In- Housing (excluding Taxes and Disposable Indirect Indirect In-kind plus Final Market Contributory Direct Indirect Fiscal kind and all Taxes) (Direct Contributions Income Subsidies Taxes Education Housing Income Income Pensions Transfers Taxes Income Health Urban plus and to SS and Indirect Indirect) Urban Group: Subsidies y < 1.25 0.0 0.0 0.0 14.9 14.9 4.4 -4.7 -0.3 14.6 50.9 28.6 79.5 98.8 -4.7 194.1 1.25 < = y < 2.50 0.0 0.0 0.0 6.1 6.1 2.5 -5.4 -2.9 3.1 19.9 11.9 31.8 40.4 -5.4 135.0 2.50 <= y < 4.00 0.0 0.0 0.1 3.1 3.1 1.9 -5.4 -3.5 -0.4 11.1 6.4 17.5 22.5 -5.4 117.1 4.00 <= y < 10.00 0.0 0.0 0.2 1.3 1.3 1.7 -6.3 -4.6 -3.3 4.2 2.2 6.4 9.4 -6.3 103.1 10.00 <= y < 50.00 -1.6 -1.6 0.1 0.4 -1.2 1.0 -7.8 -6.8 -8.0 0.9 0.3 1.2 2.6 -9.4 93.2 50.00 <= y -4.1 -4.1 0.1 0.1 -4.0 0.2 -10.4 -10.1 -14.2 0.1 0.0 0.1 0.4 -14.5 85.9 Total Population -1.3 -1.3 0.1 1.1 -0.3 1.2 -7.5 -6.2 -6.5 3.3 1.7 5.0 7.3 -8.8 98.5 Source: Authors’ estimates based in ENIGH 2007, applying the CEQ methodology. 76 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Sensitivity Analysis 2 (public expenditures education level of 2011) Deciles All Transfers Net Net Post- In-kind Transfers All Taxes Direct Taxes and All Direct Disposable In-kind In-kind Housing (excluding all Taxes) Final Market Indirect Fiscal plus Housing and (Direct and Contributions to SS Transfers Income Education Health and Urban plus Indirect Income Income Taxes Income Urban Indirect) Subsidies Deciles 1 0.0% 0.0% 9.9% 9.9% -1.8% 8.1% 18.3% 21.6% 0.0% 39.9% 53.0% -5.1% 47.9% 2 0.0% 0.0% 4.8% 4.8% -3.0% 1.8% 8.7% 10.3% 0.0% 19.0% 26.3% -5.5% 20.8% 3 0.0% 0.0% 3.0% 3.0% -3.4% -0.4% 5.9% 7.4% 0.0% 13.3% 18.3% -5.3% 12.9% 4 0.0% 0.0% 2.1% 2.1% -3.6% -1.5% 4.5% 4.9% 0.0% 9.4% 13.4% -5.5% 7.9% 5 0.0% 0.0% 1.6% 1.6% -4.1% -2.6% 3.3% 3.5% 0.0% 6.8% 10.3% -6.1% 4.2% 6 0.0% 0.0% 1.2% 1.1% -4.6% -3.5% 2.3% 2.4% 0.0% 4.7% 7.5% -6.3% 1.2% 7 -0.1% -0.1% 0.9% 0.8% -4.8% -3.9% 1.7% 1.6% 0.0% 3.3% 5.8% -6.4% -0.7% 8 -0.1% -0.1% 0.6% 0.5% -4.9% -4.4% 1.1% 1.1% 0.0% 2.2% 4.3% -6.5% -2.2% 9 -0.5% -0.5% 0.3% -0.2% -6.1% -6.3% 0.7% 0.5% 0.0% 1.2% 2.8% -7.9% -5.1% 10 -3.0% -3.0% 0.1% -2.9% -8.4% -11.3% 0.2% 0.1% 0.0% 0.3% 1.0% -12.0% -11.0% Total Population -1.3% -1.3% 0.8% -0.5% -6.2% -6.7% 1.7% 1.7% 0.0% 3.4% 5.5% -8.8% -3.3% Socioeconomic Groups In-kind All Transfers Direct Taxes and Net Net Post- All Taxes All Direct Disposable In-kind In-kind Housing Transfers plus (excluding all Final Contributions to Market Indirect Fiscal (Direct and Transfers Income Education Health and Urban Housing and Taxes) plus Indirect Income SS Income Taxes Income Indirect) Urban Subsidies Group: y < 1.25 0.0% 0.0% 13.4% 13.4% -0.3% 13.1% 25.2% 28.6% 0.0% 53.9% 71.7% -4.7% 67.0% 1.25 < = y < 2.50 0.0% 0.0% 5.4% 5.4% -2.9% 2.5% 9.9% 11.9% 0.0% 21.7% 29.7% -5.4% 24.2% 2.50 <= y < 4.00 0.0% 0.0% 2.7% 2.7% -3.5% -0.8% 5.5% 6.4% 0.0% 11.9% 16.5% -5.4% 11.1% 4.00 <= y < 10.00 0.0% 0.0% 1.1% 1.0% -4.6% -3.6% 2.1% 2.2% 0.0% 4.3% 7.0% -6.3% 0.7% 10.00 <= y < 50.00 -1.6% -1.6% 0.2% -1.4% -6.8% -8.2% 0.5% 0.3% 0.0% 0.8% 2.0% -9.4% -7.4% 50.00 <= y -4.1% -4.1% 0.0% -4.1% -10.1% -14.2% 0.0% 0.0% 0.0% 0.1% 0.3% -14.5% -14.2% Total Population -1.3% -1.3% 0.8% -0.5% -6.2% -6.7% 1.7% 1.7% 0.0% 3.4% 5.5% -8.8% -3.3% Source: Authors’ estimates based in ENIGH 2007, applying the CEQ methodology. 77 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Table 29. Concentration coefficients and Budget shares by program Concentration Coefficient Size of budget from Size of budget from Concentration Coefficient with Concentration Coefficient with with respect to National Accounts as a Household Survey as a % of Program respect to SENSITIVITY respect to SENSITIVITY BENCHMARK CASE % of GDP from Total household income from ANALYSIS 1 market income ANALYSIS 2 market income market income National Accounts Household Survey Comer es primero -0.24 -0.24 -0.24 0.24% 0.31% Standard Error 0.03 0.03 0.03 Bono Gas Hogares -0.02 -0.02 -0.02 0.08% 0.11% Standard Error 0.02 0.02 0.02 Incentivo a la Asistencia Escolar -0.50 -0.50 -0.50 0.03% 0.05% Standard Error 0.01 0.01 0.01 Bono Gas Choferes -0.33 -0.33 -0.33 0.03% 0.03% Standard Error 0.14 0.14 0.14 Incentivo a la Marina 0.44 0.44 0.44 0.00% 0.00% Standard Error 0.01 0.01 0.01 Suplemento alimienticio envejecientes -0.19 -0.19 -0.19 0.02% 0.02% Standard Error 0.10 0.10 0.10 Bono Luz -0.01 -0.01 -0.01 0.09% 0.12% Standard Error 0.03 0.03 0.03 Quisqueya Aprende -0.12 -0.12 -0.12 0.09% 0.09% Standard Error 0.02 0.02 0.02 School Food Program -0.33 -0.12 -0.12 0.20% 0.17% Standard Error 0.01 0.02 0.02 School supplies -0.33 -0.33 -0.33 0.01% 0.00% Standard Error 0.01 0.01 0.01 Pre-school Education Spending -0.30 -0.30 -0.30 0.22% 0.25% Standard Error 0.02 0.02 0.02 Primary Education Spending -0.33 -0.33 -0.33 1.75% 1.48% Standard Error 0.01 0.01 0.01 Lower Secondary Education Spending -0.19 -0.19 -0.19 0.52% 0.52% Standard Error 0.02 0.02 0.02 Upper Secondary Education Spending 0.05 0.05 0.05 0.78% 0.89% Standard Error 0.01 0.01 0.01 Tertiary Education Spending 0.30 0.30 0.30 0.27% 0.18% Standard Error 0.03 0.03 0.03 Education (all levels) -0.17 -0.17 -0.17 3.76% 3.32% Standard Error 0.01 0.01 0.01 Salud no contributiva Ambulatoria -0.27 -0.27 -0.27 0.34% 1.08% Standard Error 0.01 0.01 0.01 Salud no contributiva Hospitales -0.26 -0.26 -0.26 0.92% 0.39% Standard Error 0.03 0.03 0.03 Regimen Subsidiado Senasa -0.25 -0.25 -0.25 0.20% 0.19% Standard Error 0.01 0.01 0.01 Promese -0.04 -0.04 -0.04 0.10% 0.07% Standard Error 0.03 0.03 0.03 Health Spending -0.26 -0.26 -0.26 1.84% 1.73% Standard Error 0.01 0.01 0.01 Electricity Subsidy 0.29 0.29 0.29 1.3% 1.23% Standard Error 0.01 0.01 0.01 All Cash Transfers -0.19 -0.11 -0.19 0.78% 0.92% 78 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Concentration Coefficient Size of budget from Size of budget from Concentration Coefficient with Concentration Coefficient with with respect to National Accounts as a Household Survey as a % of Program respect to SENSITIVITY respect to SENSITIVITY BENCHMARK CASE % of GDP from Total household income from ANALYSIS 1 market income ANALYSIS 2 market income market income National Accounts Household Survey Standard Error 0.01 0.01 0.02 Total Non-contributory pensions 0.00% Standard Error Total Contributory Pensions --- 0.46 --- 0.85% 0.13% Standard Error --- 0.01 --- Total Education Spending -0.17 -0.17 -0.17 3.76% 3.32% Standard Error 0.01 0.01 0.01 Total Health Spending -0.26 -0.26 -0.26 1.84% 1.73% Standard Error 0.01 0.01 0.01 Tax Expenditures 0.47 0.47 0.47 7.78% Standard Error 0.00 0.00 0.00 Total CEQ Social Spending -0.12 -0.12 -0.08 6.38% 7.20% Standard Error 0.01 0.01 0.01 Total CEQ Social Spending plus Contrib Pensions -0.11 -0.11 -0.07 7.23% 7.33% Standard Error 0.01 0.01 0.01 Total Primary Spending N/A N/A N/A 15.33% Total Government Spending N/A N/A N/A 20.16% Source: Authors’ estimates based in ENIGH 2007, applying the CEQ methodology. 79 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Table 30. VAT tax expenditures by category of goods and beneficiaries by income group Share of benefits going to each income group Groups: y< 1.25 < y 2.5 < y 4 50 y < 2.5 y<4 y>4 Total 1.25 < 2.5 <4 10 50 Total VAT Tax Expenditure 0.6% 3.6% 7.5% 31.5% 45.5% 11.4% 4.2% 11.7% 88.3% 100.0% Food & Non Alcoholic Beverages 1.2% 6.5% 12.1% 41.4% 34.2% 4.6% 7.7% 19.8% 80.2% 100.0% Housing, water, electricity, gas and other fuels 0.2% 1.5% 3.9% 24.4% 54.8% 15.2% 1.7% 5.6% 94.4% 100.0% Furnishings, household equipment and routine household maintenance 0.8% 4.6% 10.3% 35.2% 41.3% 7.8% 5.4% 15.7% 84.3% 100.0% Health 0.5% 2.6% 6.4% 29.1% 47.9% 13.5% 3.1% 9.5% 90.5% 100.0% Transport 0.3% 2.1% 5.1% 26.2% 49.6% 16.7% 2.4% 7.6% 92.4% 100.0% Education 0.7% 3.6% 7.5% 31.4% 47.4% 9.4% 4.3% 11.9% 88.1% 100.0% Recreation and culture 0.1% 0.6% 2.8% 21.5% 62.6% 12.4% 0.7% 3.5% 96.5% 100.0% Others 0.4% 2.1% 5.1% 26.1% 48.9% 17.3% 2.5% 7.6% 92.4% 100.0% Income shares 0.5% 3.1% 6.6% 29.6% 46.6% 13.6% 3.6% 10.2% 89.8% 100.0% Population shares 5.7% 13.8% 17.4% 40.0% 21.6% 1.4% 19.5% 37.0% 63.0% 100.0% Benefits by category of goods Groups: y < 1.25 1.25 < y < 2.5 2.5 < y < 4 4 < y < 10 10 < y < 50 y > 50 y < 2.5 y<4 y>4 Total Total VAT Tax Expenditure 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% Food & Non Alcoholic Beverages 66.8% 64.5% 57.6% 46.9% 26.8% 14.4% 64.8% 60.2% 32.4% 35.6% Housing, water, electricity, gas and other fuels 6.0% 7.0% 8.6% 12.9% 20.0% 22.2% 6.8% 8.0% 17.7% 16.6% Furnishings, household equipment maintenance 0.3% 0.3% 0.3% 0.3% 0.2% 0.2% 0.3% 0.3% 0.2% 0.2% Health 8.2% 7.5% 8.8% 9.5% 10.9% 12.3% 7.6% 8.4% 10.6% 10.3% Transport 9.9% 12.3% 14.3% 17.3% 22.6% 30.6% 12.0% 13.4% 21.7% 20.8% Education 3.4% 2.9% 3.0% 2.9% 3.1% 2.4% 3.0% 3.0% 2.9% 2.9% Recreation and culture 1.0% 1.1% 2.3% 4.2% 8.4% 6.7% 1.1% 1.8% 6.7% 6.1% Others 4.5% 4.4% 5.1% 6.2% 8.0% 11.3% 4.4% 4.8% 7.8% 7.4% Source: Authors’ estimates based in ENIGH 2007, applying the CEQ methodology. 80 WB – Fiscal Policy and Redistribution in the Dominican Republic – For official use only Table 31. Breakdown of social spending (2013) Description Included In Analysis Total In Incidence Analysis Source % GDP Direct Cash and Food Transfers 0.80% 0.80% Direct Transfers 0.51% 0.51% a Comer es Primero Yes 0.24% 0.24% a Bonogas Hogar Yes 0.08% 0.08% a Bonoluz Yes 0.09% 0.09% a Incentivo a la Asistencia Escolar Yes 0.03% 0.03% a Bonogas Chofer Yes 0.03% 0.03% a Suplemento Alimenticio - Envejecientes Yes 0.02% 0.02% a Incentivo a la Policía Preventiva Yes 0.01% 0.01% a Incentivo a la Educación Superior Yes 0.01% 0.01% a Incentivo a la Marina De Yes 0.00% 0.00% a Bono Escolar Estudiando Progreso Yes 0.01% 0.01% a School Food programs 0.20% 0.20% b Pre-School (From 3 to 5 years old) Yes 0.02% 0.02% b Primary (From 6 to 11 years old, 1st to 6th Basico) Yes 0.18% 0.18% b Shoes, uniform and backpacks 0.01% 0.01% b Pre-School (From 3 to 5 years old) Yes 0.00% 0.00% b Primary (From 6 to 11 years old, 1st to 6th Basico) Yes 0.00% 0.00% b Lower Secundary (12 to 13 years, 7th and 8th Basico) Yes 0.00% 0.00% b Upper Secundary (14 to 17 years, 1st to 4th Medio) Yes 0.00% 0.00% b Alphabetization Yes 0.09% 0.09% b Education 3.76% 3.54% Pre-School (From 3 to 5 years old) Yes 0.22% 0.22% b Primary (From 6 to 11 years old, 1st to 6th Basico) Yes 1.75% 1.75% b Lower Secundary (12 to 13 years, 7th and 8th Basico) Yes 0.52% 0.52% b Upper Secundary (14 to 17 years, 1st to 4th Medio) Yes 0.78% 0.78% b Tertiary Yes 0.27% 0.27% b Other expenses in education ncp No 0.22% 0.00% b Health 1.84% 1.62% Ministerio de Salud Pública 1.26% 1.26% c Centros de Salud (estimado) Yes 0.34% 0.34% c Hospitales (estimado) Yes 0.92% 0.92% c Institutuciones de seguridad social 0.31% 0.30% c Subsidiado 2013 Yes 0.20% 0.20% c IDSS (Fuente: Senasa) Yes 0.10% 0.10% d Jubilados Senasa No 0.01% 0.00% d PROMESE 2012 Yes 0.07% 0.07% c OTROS (Hospitales Militar, Policía, CONAVIHSIDA, CERSS, No 0.21% 0.00% c otros) Social Spending Analyzed (Benchmark) 6.0% 0.0% Total Social Spending (Benchmark) 6.1% 0.0% Contributory Pensions Instituto Dominicano de Seguridad Social 0.85% 0.85% e Social Spending Analyzed (Sensitivity Analysis 1.) 6.81% 0.00% Total Social Spending (Sensitivity Analysis 2) 6.96% 0.00% Non Social Spending Indirect Subsidies (electricity) 1.3% 1.3% g Notes and sources: a ADESS http://www.adess.gov.do/v2/Reportes/ReportView2.aspx?Provincia=0 b Ministry of Education http://www.minerd.gob.do/Lists/Anlisis%20Financiero/AllItems.aspx c Informe Nacional de Gasto en Salud 2013 d Own Calculations based on Informe Naiconal de Gasto en Salud 2011, 2012 y 2013 e CNSS, Informe a Diciembre 2013 http://cnss.gob.do/serve/listfile_download.aspx?id=3734&num=1 f Senasa (2014), "Reconversión del IDSS Y Red P´ública Unica", Mimeo g Ministry of Finance 81