From Policy to Impact: Fiscal Incidence Analysis in Armenia © 2025 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington, DC 20433 Telephone: 202-473-1000; Website: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, links/footnotes and other information shown 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. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions The material in this publication is subject to copyright. Because the World Bank Group encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Attribution — Please cite the work as follows: World Bank. 2025. From Policy to Impact: Fiscal Incidence Analysis in Armenia. Washington, DC: World Bank Group. Translations — If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations — If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content — The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third-party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to reuse a component of the work, it is your responsibility to determine whether permission is needed for that reuse and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. Any queries on rights and licenses, including subsidiary rights, should be addressed to the Publishing and Knowledge Division, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; email: pubrights@worldbank.org Cover and other pictures: Designed and generated by Team Design at Midjourney. Report and Cover design: Team Design & A.K. ii Contents Abbreviations vii Acknowledgments viii EXECUTIVE SUMMARY 2 CHAPTER ONE: The Impact of Fiscal Policy on Poverty and Inequality 6 1.1. Fiscal Policy Context 7 1.1.1. Tax System: Overview 8 1.1.2. Public Expenditure: Overview 9 1.2. Redistributive Impact of Fiscal Policy 10 1.2.1. Impact of Fiscal Policy on Poverty 12 1.2.2. Impact of Fiscal Policy on Inequality 14 1.2.3. Net Payers and Receivers of the Fiscal System 16 1.2.4. The Progressivity of the Fiscal System 18 1.2.5. Marginal Contributions to Inequality Reduction 19 1.2.6. Marginal Contributions to Poverty Reduction 20 1.2.7. The Incidence of Fiscal Interventions 21 1.2.8. Conclusion 24 1.2.9. Limitations of the Analysis 24 CHAPTER TWO: The Distributional Impact of Universal Health Insurance 26 2.1. The Health System in Armenia 26 2.2. The Universal Health Care System 27 2.3. Simulating the Impact of a UHI System 27 2.4. Estimating the Distributional Impact of the UHI System 29 2.5. Results 32 2.6. Conclusion 33 2.7. Limitations 34 References 35 APPENDIX A: Data and Methodology 42 A.1. Concepts and Methodology 42 A.2. Data Sources 42 A.3. The Commitment to Equity Framework 43 A.4. Limitations 46 APPENDIX B: Sensitivity Analysis: Pensions as Deferred Income 48 APPENDIX C: Current Health System and Expenditure in Armenia 53 iii APPENDIX D: Survey of Approaches for Incorporating the Healthcare Benefits in Poverty Estimates 55 APPENDIX E: Inclusion of Health Benefits in Armenia Poverty Estimates 61 APPENDIX F: Simulating the Distributional Impact of Health Benefits 64 F.1. Monetization and Inclusion of Health Benefits in the Baseline Commitment to Equity (CEQ) Analysis 64 F.2. Simulating Insurance Premium Payments and the Identification of Payers 66 F.3. Estimation of the Benefits 67 F.4. Measuring the Distributional Impact 71 iv Figures Figure 1.1 Armenia’s tax-to-GDP ratio compares favorably with the average among peers, but is behind HICs 6 Figure 1.2 Armenia’s tax-to-GDP ratio has been on an upward trend, stabilizing at around 23 percent 6 Figure 1.3 General government expenditure (% of GDP) 7 Figure 1.4 CEQ income concepts in the pensions as government transfers scenario 11 Figure 1.5 Poverty headcount from market to consumable income 12 Figure 1.6 Poverty headcount from market to consumable income, by marz 13 Figure 1.7 The poverty impact of fiscal policy in upper- and lower- middle-income countries: from market to consumable income 13 Figure 1.8 The impact of fiscal policy on inequality, measured by the Gini coefficient 14 Figure 1.9 Impact of fiscal policy on inequality (measured by the Gini coefficient), by marz 15 Figure 1.10 The redistributive impact of fiscal policy, SDG indicator 10.4.2 16 Figure 1.11 The net fiscal position of households after taxes and transfers 17 Figure 1.12 The progressivity of selected fiscal policy instruments 18 Figure 1.13 Marginal contribution to inequality reduction of each fiscal instrument, in Gini points, measured at consumable income 19 Figure 1.14 Marginal Contributions of fiscal instruments to poverty reduction, consumable income measure 20 Figure 1.15 Absolute incidence: shares of total taxes paid by households, by market income 21 Figure 1.16 Relative incidence: (% of market income) 22 Figure 1.17 Absolute incidence: total transfers, by household quintile and market income 22 Figure 1.18 The incidence of education transfers, by household quintile and market income 23 Figure 2.1 Simulating UHI benefits using the CEQ framework 30 Figure 2.2 The concentration of UHI benefits and insurance premiums, by decile 31 Figure 2.3 The incidence of universal health insurance benefits and insurance premiums 31 Figure 2.4 The impact of universal health insurance on inequality 32 Figure 2.5 The impact of universal health insurance on poverty 33 Figure A.1 CEQ income concepts in the pensions as deferred income scenario, Armenia 44 Figure B.1 Poverty impact, PDI approach, 2017 and 2021 48 Figure B.2 Inequality impact, pensions as deferred income 49 Figure F.1 Simulating UHI benefits using the CEQ framework 65 v Figure F.2 Insurance premium payment and number of payers 67 Figure F.3 Armenia health expenditures as reported in ILCS 2021 by type of spending 68 Figure F.4 Health expenditures, average by decile (AMD per month) 68 Figure F.5 Smoothed average benefits by age and sex, individuals in Group B 69 Figure F.6 Concentration of UHI benefits and insurance premiums, by decile 70 Figure F.7 Incidence of UHI benefits and premiums 70 Tables Table 1.1 Government revenue, Armenia, 2021 8 Table 1.2 Government expenditures, Armenia, 2021 9 Table 2.1 Draft implementation schedule for the UHI system, Armenia 28 Table B.1 Marginal contributions of each fiscal intervention to inequality and poverty (for disposable, consumable, and final income) 49 Table C.1 Basic Benefits Package: coverage 53 Table D.1 Deriving SPM unit resources 56 Table D.2 Differences in SPM and HIPM methods 57 Table D.3 Overview of poverty measures: official, supplemental and health inclusive 59 Table F.1 CEQ income concept after incorporation of the UHI benefits 65 Table F.2 Payment of insurance premiums 66 Boxes Box A.1 Tax and transfer allocation methods 43 vi Abbreviations BBP Basic Benefits Package OOP Out of Pocket CEQ Commitment to Equity PDI Pensions as Deferred Income CIT Corporate Income Tax PGT Pensions as Government Transfers EUROMOD Tax-Benefit Microsimulation PIT Personal Income Tax Model for the European Union PPP Purchasing Power Parity FIA Fiscal Incidence Analysis SDG Sustainable Development Goal GDP Gross Domestic Product SPM Supplemental Poverty Measure HIPM Health Inclusive Poverty Measure UHI Universal Health Insurance KI Kakwani Index UMIC Upper-Middle-Income Country ILCS Integrated Living Conditions UN United Nations Survey (Armenia) VAT Value Added Tax MC Marginal Contribution vii Acknowledgments This report has been prepared under the South Caucasus Poverty Global Department team led by Saida Ismailakhunova (Senior Economist, EECPV) and Natsuko Kiso Nozaki (Economist, EECPV) and comprising Beenish Amjad (ETC, EECPV) and Maynor Cabrera (Consultant, EECPV). The team appreciates the insightful advice and guidance provided by the peer reviewers, including Nora Lustig (Member of the Advisory Group for the Prosperity Practice Group, Samuel Z. Stone Professor of Latin American Economics and Director of the Commitment to Equity Institute, Tulane University), Alan Fuchs (Lead Economist, EMNPV), Gabriela Inchauste (Lead Economist, EAWPV), and Stephan Younger (Scholar at Ithaca College and Expert Africa, Asia and Europe at CEQ Institute). The report was prepared under the guidance of Carolin Geginat (Country Manager, Armenia) and Ambar Narayan (Practice Manager, EECPV). The team would like to express special gratitude for the advice and guidance provided by Obert Pimhidzai (Lead Economist, EECPV) and Miguel Eduardo Sanchez Martin (Program Leader, EECDR). The team is grateful to the Ministry of Finance, Ministry of Health, and the Statistical Committee of the Republic of Armenia for data that were drawn upon for the analysis. Any errors and omissions remain the responsibility of the authors. viii Executive Summary 1 Executive Summary Over the past two decades, Armenia's economy While Armenia’s record of growth and poverty has undergone significant transformation reduction over the years is commendable, marked by periods of rapid growth, external substantial challenges remain. As of 2023, more shocks, and resilience. From 2000 to 2008, than half the population (52 percent) was living Armenia transformed into one of the Europe and below the international poverty line among Central Asia region's most rapidly growing upper-middle-income countries of US$6.85 a day economies, with average annual growth in gross (2017 purchasing power parity), and 8.4 percent domestic product (GDP) exceeding 10 percent, were living below the lower-middle-income- mainly driven by remittances, a construction country poverty line of US$3.65 a day. Income sector boom, and investments. The economy inequality, as measured by the Gini coefficient contracted sharply during the 2007–09 global based on income, improved from 0.38 in 2019 to 0.35 in 2023. However, regional disparities in financial crisis, when GDP declined by 14.1 welfare remain significant, with the national percent, but the postcrisis recovery brought poverty rate ranging from 7 percent in Syunik to steady growth, averaging 4 percent to 5 percent 43.1 percent in 2023. annually. Armenia's accession to the Eurasian Economic Union (EAEU) in 2015 and Fiscal policy is essential in advancing the strengthened trade relations with Russia, government's primary economic strategy, Kazakhstan, and other member states contributed addressing equity challenges, and maintaining to steady recovery and growth. the sustainability of public finances. Over the past two decades, fiscal performance has Armenia's economy faced unprecedented significantly improved. Tax policy and challenges in 2020 leading the government to administrative reforms have bolstered revenue launch a comprehensive sustainable collection, increasing the five-year moving average development program. The economy grappled from 21.2 percent (2014–19) to 23.5 percent in three with a double-headed crisis in 2020: the COVID- years (2021–23), aligning with levels observed 19 pandemic and a military conflict with among regional and income peer countries. Unlike Azerbaijan. These events resulted in one of the many similar economies, Armenia generates a sharpest economic contractions in the region. In larger share of revenue from indirect taxes than response, the government launched a direct taxes. Spending levels have remained comprehensive five-year program in 2021 designed prudent and below regional and income peer to drive recovery and foster sustainable averages, contributing to fiscal discipline, with development. The program sets ambitious goals, deficits averaging 3 percent of GDP in 2010–20, including achieving an average annual growth rate thereby also contributing to debt sustainability. of at least 7 percent from 2021 to 2026 and Meanwhile, the government allocates a larger share of public expenditures to social protection eliminating extreme poverty. Key initiatives and education. involve the phased introduction of universal health coverage, increased capital investment, expanded and more well targeted social assistance programs, and enhanced pension allocations. 2 This fiscal incidence analysis (FIA) establishes a Fiscal policy has reduced inequality, and baseline for assessing current fiscal policies and pensions and direct transfers are the major identifying reforms to enhance fiscal equity and drivers. Measured by the Gini index, inequality sustainability. Utilizing the Commitment to was reduced by 13.6 Gini points (from 35.9 at Equity (CEQ) methodology, it analyzes how market income to 22.3 at final income). Direct public expenditures and taxes affect various transfers and pensions were the main drivers of population groups, tracking changes in household inequality reduction. Benefits in kind, particularly income distribution from market income (before education, also contributed to inequality taxes and transfers) to final income. The report reduction, while indirect subsidies had a highlights the impact of fiscal policies, such as negligible redistributive effect. Direct taxes, such taxes, transfers, and social spending, on income as the personal income tax, were progressive while indirect taxes, such as the VAT, were regressive. inequality and poverty. It also informs the The net cash position reveals that the poorest 40 Sustainable Development Goals indicator 10.4.2, percent of the welfare distribution across the which measures the redistributive effects of fiscal population (the bottom 40) are net cash recipients policy. By identifying spending priorities that of the fiscal system. Pensions and direct transfers reduce poverty and inequality, the FIA explores are the primary sources of this support. strategies to create fiscal space without disproportionately impacting poor and vulnerable The analysis offers valuable guidance for crafting households. a unified set of fiscal policies to help the government and stakeholders address the The FIA results indicate that Armenia's fiscal country’s demographic and fiscal challenges, policy is effective in reducing poverty, but laying the foundation for more prosperous, requires further reforms to maximize the equitable, and sustainable growth. A key focus potential, including measures to address should be increasing investment in human capital, regional disparities in its impact on poverty and particularly by boosting funding for education inequality reduction. The net impact of fiscal and health care, to promote greater equality of policy on poverty is positive, reducing the opportunity. Enhancing the quality of these national poverty headcount by 2.4 percentage services and addressing spatial disparities in access points when comparing market income to are crucial to ensuring that development gains are consumable income. The reduction in poverty is equitably distributed across all regions. driven by the combined impacts of pensions and other direct transfers, decreasing the poverty rate The analysis points to an opportunity for the by 9.8 percentage points going from market government to reassess certain policies, income to disposable income. However, these including VAT, excise taxes, and cash transfers gains are partially offset by the adverse effects of like free-rent apartments, which are regressive. indirect taxes and subsidies, which increased the Some in-kind benefits, such as health and higher poverty rate by 7.4 percentage points. The impact education benefits, could also be made more on poverty and inequality also varied across progressive and pro-poor if designed regions, warranting further research to explain the appropriately. Further research is needed to gap. explore the fiscal implications of these potential reforms and their impact on poverty and equity. For example, while transitioning from regressive to progressive policies may improve equity, it is also critical to understand the financial and social consequences of such changes to ensure they are both sustainable and effective. 3 Given that UHI represents the most significant To enhance the poverty-reducing and equity- policy reforms in recent years, the welfare enhancing impacts of UHI, the government could impact of UHI was also simulated in the consider several policy adjustments. First, a assessment using the existing FIA model, which significant number of lower-income households is expected to have impacts on poverty reduction report zero out-ot-pocket (OOP) spending, often and equality. The results suggest that because they forgo medical care due to financial implementation of the UHI is likely to reduce constraints. To address this, targeted awareness poverty by protecting the poor households from campaigns and simplified enrollment procedures catastrophic healthcare costs. The analysis also can be implemented to ensure that all eligible suggests that UHI may increase inequality, which households, particularly those in lower-income is however an estimate with a low degree of groups, are enrolled and able to access the confidence due to data and model limitations. benefits. Second, the impact of UHI depends not These include a key assumption of the model that only on affordability but also on access to healthcare-seeking behavior remains unchanged healthcare services. Policies should take into under the reform, including among the large account the availability of healthcare facilities, number of households who do not report any travel costs, and waiting times to ensure that all health spending, even though lower costs could income groups can fully benefit from the reform. increase usage of the healthcare services among Third, the analysis does not account for potential the poor. 1 variations in copayments under UHI. To prevent financial burdens on lower-income households, the government could consider introducing a progressive copayment system that would minimize the costs for the most vulnerable while ensuring that higher-income households contribute in greater proportion to the system. 1 Since benefits or net savings are calculated as the difference overestimate the impact on (increasing) inequality, by making it between actual spending and simulated post-UHI reform spending, appear that wealthier households benefit more as they reported the assumption of unchanging health-seeking behavior is likely to higher actual healthcare spending. 4 C H A P T E R O N E The Impact of Fiscal Policy on Poverty and Inequality Fiscal Policy Context | Tax System: Overview | Public Expenditure: Overview | Redistributive Impact of Fiscal Policy | Impact of Fiscal Policy on Poverty | Impact of Fiscal Policy on Inequality | Net Payers and Receivers of the Fiscal System | The Progressivity of the Fiscal System | Marginal Contributions to Inequality Reduction | Marginal Contributions to Poverty Reduction | The Incidence of Fiscal Interventions | Conclusion | Limitations of the Analysis | 5 C H A P T E R O N E The Impact of Fiscal Policy on Poverty and Inequality Armenia has been implementing fiscal policies customs policy, administrative reforms, and efficiently and prudently. The level of tax improvements in the formalization of the labor collection in Armenia — a five-year moving average market. On average, the government of Armenia of around 23.5 percent over the last three years collects less revenue though indirect taxes relative (2021–23) — is comparable with peers. The observed to all its peers (12.5 percent of GDP in 2021 against improvement in the tax-to-GDP ratio, 21.2 percent an average of 15.8 percent among peers) and more of GDP in 2014–19, was supported by a tax and through direct taxes (refer to Figures 1.1 and 1.2). Figure 1.1: Armenia’s tax-to-GDP ratio compares Figure 1.2: Armenia’s tax-to-GDP ratio has favorably with the average among peers, but is been on an upward trend, stabilizing at around behind HICs 23 percent tax-to-GDP ratio, % Left axis: percentage points; right axis: % of GDP 28 4 25 3 20 24 2 15 20 1 10 0 16 -1 5 12 -2 0 Armenia Peers ECA YoY Change in Tax-to-GDP Ratio (LHS) UMIC HIC 5-Y Moving Average (RHS) Source: World Bank 2024b Note: ECA = Europe and Central Asia region. 5-Y = five-year. HIC = high-income countries. UMIC = upper-middle-income countries. YoY = year over year. The levels of public spending in Armenia have strong commitment to prudent fiscal policy and an remained below regional and upper-middle- updated fiscal rule in 2017. The government’s fiscal income-country averages (refer to Figure 1.3). policy has also demonstrated reasonable Fiscal deficits have remained stable at around 3.5 countercyclicality and progressivity in the last percent of GDP over the last decade as a result of a decade. 6 Figure 1.3: General government expenditure (% of GDP) inequality. Fiscal Incidence Analysis (FIA) involves a 20 method to allocate the burden of taxes and the monetized value of government expenditures to 15 estimate the incidence of taxes and benefits and their impact on inequality and poverty. This FIA is based 10 on the CEQ methodology, which is a well-established 5 framework for analyzing fiscal incidence. 2 The analysis examines the impact of government taxes, 0 public social spending programs, and subsidies on 2014 2015 2016 2017 2018 2019 2020 2021 2022 various income groups using data from the Integrated Armenia Upper middle income Europe & Central Asia Living Conditions Survey (ILCS) 2021 3 and macrofiscal administrative records for fiscal year Source: WDI (World Development Indicators) (dashboard), World Bank, Washington, DC, https://datatopics.worldbank.org/world-development- 2021. The CEQ approach serves as a foundation for indicators developing more complex microsimulation models that can be used to analyze the potential This chapter examines the government of Armenia’s distributional effects of future policy reforms. These fiscal policy and the impact of the policy on poverty models are important in designing evidence-based and inequality. Besides ensuring the availability of fiscal policies that prioritize equity throughout the sustainable public finances, fiscal policy plays a policy cycle, from the planning stage to critical role in addressing equity challenges. Fiscal implementation. For more details on the data and policy can be instrumental in reducing poverty and methodology used in this analysis, see Appendix A. 1.1. Fiscal Policy Context Although significant fiscal stabilization has been achieved at low debt and deficit levels. The strong commitment to prudent fiscal policy and an updated fiscal rule in 2017 has contributed to the sustainability of public debt. Central government debt declined by around 14 percentage points, to 46.7 percent of GDP, because of strong growth, exchange rate appreciation, and improvement in the primary balance. The fiscal deficit narrowed to 2.1 percent of GDP in 2022, a reduction by 2.5 percentage points of GDP relative to 2021, largely because of lower current expenditure (3.7 percentage points of GDP) and some revenue gains (0.2 percentage points), while capital expenditures exhibited a welcome surge (IMF 2023). Realizing the ambitious government program requires boosting revenue collection and enhancing spending efficiency and service delivery outcomes. It is important also to improve the quality of public finances (World Bank 2024b). 2 The CEQ methodology aims to address four key questions: How for over 85 low- and middle-income countries over the past much income redistribution and poverty reduction are being decade. For methodological details, see Lustig (2018, 2022a, accomplished through fiscal policy? How equalizing and pro- 2022b). poor are specific taxes and government spending? How 3 See the 2021 round, ILCS (Integrated Living Conditions Survey, effective are taxes and government spending in reducing Armenia) (anonymized microdata database), Statistical inequality and poverty? What is the impact of fiscal reforms that Committee of Armenia, Yerevan, Armenia, change the size and/or progressivity of a particular tax or https://armstat.am/en/?nid=205 benefit? The methodology has been extended to and adapted 7 1.1.1. Tax system: Overview Tax collection in Armenia is stable and comparable of GDP). Armenia was the first country in the Eastern with peer countries; a major share is contributed by Europe and Central Asia region to introduce indirect taxes. The average tax-to-GDP ratio in 2021– environmental taxes and payments for the use of 23 stood at 23.5 percent, which represents a rise from natural resources. Collections of existing the 21.2 percent average in 2014–19. The government environmental protection taxes represented 0.9 collects more from indirect taxes (52.6 percent of percent of GDP in 2018–22, and the share is growing. 4 total government revenue) than from direct taxes Details on each tax are provided in Table 1.1. This (35.4 percent of total government tax revenue). FIA is based on the CEQ methodology and captures Among direct taxes, personal income tax collection the equivalent of 69 percent of direct taxes (personal is, on average, higher in Armenia. Among indirect income tax and property tax) and 82.2 percent of taxes, Armenia collects VAT (around 7 percent–8 indirect taxes. percent of GDP) and excise (1.5 percent–2.0 percent Table 1.1: Government revenue, Armenia, 2021 Fiscal accounts Portion of fiscal accounts analyzed Included AMD, Share of % of AMD, % of billions government GDP billions total revenue Total revenue and grants 1,744 100.0 24.9 1,223 70.1 Taxes 1,625 93.2 23.2 1,223 75.3 Direct taxes 618 35.4 8.8 426 69.0 Personal income tax Yes 426 24.4 6.1 426 100.0 Other (corporate income tax, property tax) No 191 11.0 2.7 Indirect taxes 917 52.6 13.1 754 82.2 VAT Yes 556 31.9 8.0 556 100.0 Excise Yes 113 6.5 1.6 113 100.0 Custom duties Yes 85 4.9 1.2 85 100.0 Other indirect taxes No 163 9.4 2.3 Social contributions Yes 43 2.5 0.6 43 100.0 Other taxes No 47 2.6 0.7 Nontax revenue No 119 6.8 1.7 Source: World Bank elaboration based on the macrofiscal administrative records for fiscal year 2021 4 World Bank, 2024b. 8 1.1.2. Public Expenditure: Overview Public expenditures on education and social to Table 1.2). Transfers in kind, including in health protection, particularly pensions, are prioritized in care and education, account for 16.8 percent of total Armenia. In fiscal year 2021, total general government expenditures (2.3 and 2.7 percent of GDP, expenditures reached 29.5 percent of GDP. Social respectively). The government’s expenditures on protection represented 28.1 percent of total subsidies amounted to 0.7 percent of total spending, government expenditures (8.3 percent of GDP) (refer which is less than 0.2 percent of GDP. Table 1.2: Government expenditures, Armenia, 2021 Fiscal accounts Portion of fiscal accounts analyzed Included AMD, Share of total % of AMD, % of billions government GDP billions total expenditure Total expenditures 2,062 100.0 29.5 670 32.5 Social spending 926 44.9 13.2 656 70.8 Social protection 580 28.1 8.3 384 66.3 Contributory pensions Yes 272 13.2 3.9 272 100.0 Conditional and unconditional cash 308 14.9 4.4 113 36.5 transfers Family and children Yes 78 3.8 1.1 72 91.4 Social pension Yes 29 1.4 0.4 29 100.0 Social exclusion Yes 14 0.7 0.2 11 79.4 Other (survivors, social exclusion, housing) Yes 187 9.1 2.7 1 0.4 Education 186 9.0 2.7 165 88.5 Preschool and primary school Yes 67 3.3 1.0 67 100.0 Secondary Yes 68 3.3 1.0 68 100.0 Postsecondary nontertiary Yes 12 0.6 0.2 12 100.0 Tertiary Yes 17 0.8 0.2 17 100.0 Other No 21 1.0 0.3 Health 160 7.8 2.3 107 66.9 Contributory Yes 38 1.8 0.5 38 100.0 Noncontributory Yes 69 3.3 1.0 69 100.0 Other No 53 2.6 0.8 9 Subsidies 14 0.7 0.2 14 100.0 Interest rate Yes 14 0.7 0.2 14 100.0 Other expenses No 1,122 54.4 16.0 Source: World Bank elaboration based on the macrofiscal administrative records for fiscal year 2021 1.2. Redistributive Impact of Fiscal Policy Recent fiscal challenges highlight the need to that is, before the effects of taxes and transfers — evaluate the redistributive impact of fiscal policies for the baseline scenario of pensions as government through FIA. Fiscal policies influence households transfers (PGT). The income concepts result from and individuals through various mechanisms, subtracting taxes and adding transfers to pre-fiscal enabling governments to generate revenue and income. Disposable income includes direct tax and finance public expenditures. Fiscal interventions, transfer effects, while consumable income considers both on the revenue side (for example, direct and direct transfers, direct taxes, indirect taxes, and indirect taxes) and the expenditure side (such as subsidies. direct transfers, contributory pensions, subsidies, and the provision of health and education services), In the CEQ methodology, pensions may be play a crucial role in income redistribution, living analyzed either as PGT or pensions as deferred standards, and poverty reduction. Fiscal policies income (PDI). The PGT approach treats pensions as significantly shape income equality and poverty social welfare payments from the government, while outcomes among households and individuals. FIA the PDI approach views them as earned income that provides valuable insights into the overall was taxed earlier and paid back to beneficiaries later effectiveness of fiscal policies in addressing poverty in life. The research here relies on the PGT and inequality in Armenia. approach for the baseline scenario (because of characteristics of the pension system in the This analysis estimates the impact of fiscal policy, country), while conducting sensitivity analysis both taxes and transfers, on household welfare. based on the PDI approach. The results presented in The analysis measured the impacts of taxes and Figure 1.5 reflect the outcomes using the baseline transfers on poverty and inequality at different PGT approach, whereas the PDI results are CEQ income concepts (refer to Figure 1.4). 5 presented in Appendix B. 6 The PDI results show Consumption (recorded in the household survey) that pensions reduce poverty significantly, by 14.4 was equated with the CEQ disposable income percentage points, indicating that individuals in the concept, and the rest of the CEQ income concepts lower income deciles largely depend on old-age were calculated based on disposable income pensions. Details on the sensitivity analysis under (consumption), minus or plus the relevant taxes or the PDI scenario are provided in Appendix B. transfers. Market income is pre-fiscal income — 5 The poverty headcount is measured as the share of population not measured at final income since in-kind health and below the national poverty line. The impacts of taxes and education benefits are not part of monetary poverty transfers on poverty were measured as the difference between measurement. the poverty headcount at market income (pre-fiscal income) 6 For the Armenia Public Expenditure Review (World Bank, 2024), and the poverty headcount at consumable income (post-fiscal the results based on the PDI approach were reported. income excluding in-kind benefits). The poverty headcount was 10 Figure 1.4: CEQ income concepts in the pensions as government transfers scenario Market Income Labor, capital, and other income, private transfers, self- consumption and imputed value of own dwelling Direct Taxes Personal income tax (wages, passive income), contribution to pension Net Market Income Direct Transfers Family benefit, child benefits, contributory and non-contributory pension, other benefits Disposable Income = Consumption Indirect Subsidies Indirect Taxes Interest rate subsidy VAT, excises, custom on agricultural loans taxes Consumable Income In-kind Transfers Education and health Final Income Source: Adapted from Lustig 2022a, 2022b The results of the analysis are presented below. 11 1.2.1. Impact of Fiscal Policy on Poverty The analysis revealed that fiscal policy has of poverty reduction from market to disposable reduced the national poverty headcount by 2.4 income. Both indirect taxes and subsidies percentage points. The combined effect of contributed to a poverty increase of 7.4 percentage pensions and direct transfers, minus the payment of points, from disposable to consumable income direct taxes, remained the most significant drivers (refer to Figure 1.5). Figure 1.5: Poverty headcount from market to consumable income Poverty Headcount at the National Poverty Line 40 35.6 33.2 31.5 30 25.8 Percent 20 10 0 Market income Net market Disposable Consumable income income income Source: World Bank calculations based on fiscal administrative data; the CEQ methodology described by Lustig 2022a, 2022b; data of the 2021 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan, Armenia The impact of fiscal policies on poverty differs from the pre-fiscal to the post-fiscal income significantly across regions, warranting deeper situation. Poverty reduction in the other eight exploration into the underlying causes (refer to regions occurred at varying degrees with the largest Figure 1.6). This analysis compared poverty impact observed in Syunik (20 percentage points headcount ratios based on market income (pre- from pre-fiscal to post-fiscal) and lowest in Tavush fiscal income) and consumable income for each (0.2 percentage points from pre-fiscal to post-fiscal). marz (region). The findings revealed that fiscal This disparity may be attributed to factors such as policy reduced poverty in all but three regions economic concentration, urbanization, migration, (Ararat, Armavir, and Kotayk). These three regions and better infrastructure, but further research is experienced a slight increase in the poverty rate needed to explain the gap. 12 Figure 1.6: Poverty headcount from market to consumable income, by marz Poverty reduction, by marz (from MI to CI) 60% 50% 40% 30% 20% 10% 0% Market income Consumable income Source: World Bank calculations based on fiscal administrative data; the CEQ methodology described by Lustig 2022a, 2022b; data of the 2021 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan, Armenia. Note: Regions are ordered based on the magnitude of the gap between MI and FI, from the largest to the smallest. Overall, the impact of fiscal policy on poverty shown in the Figure 1.7, Armenia falls below the reduction is smaller in Armenia than in some average of 5.1 percentage points, indicating that upper-middle-income countries and countries in fiscal policies have a relatively limited impact on the Europe and Central Asia region. Although poverty. In contrast, countries with higher positive Armenia has a relatively high international poverty values in the figure experience stronger poverty- rate for an upper-middle-income country (UMIC), reducing effects from their fiscal policies, the poverty decreasing impact of fiscal policies is highlighting potential areas for improvement in smaller compared to many other countries. As Armenia’s tax and transfer system. Figure 1.7: The poverty impact of fiscal policy in upper- and lower- middle-income countries: from market to consumable income 20 15 Percentage Points 10 Average: 5.1 5 0 -5 -10 Source: World Bank calculations based on fiscal administrative data; the CEQ methodology described by Lustig 2022a, 2022b; CEQ Standard Indicators (database), Commitment to Equity, Inter-American Dialogue, Washington, DC; Center for Inter-American Policy and Research and Department of Economics, Tulane University, New Orleans; data of the 2021 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan, Armenia 13 1.2.2. Impact of Fiscal Policy on Inequality Fiscal policy in Armenia is effective in achieving The largest effect is associated with primary redistribution and reducing inequality. 7 As education. Indirect subsidies (interest rate subsidy measured by the Gini index, inequality was reduced on agricultural loans) had a small impact on by 13.6 Gini points (from 35.9 at market income to inequality reduction (0.04 Gini points) despite 22.3 at final income) (refer to Figure 1.8). Most of comparable spending levels to post-secondary the inequality reduction occurred from market education (0.2 percent of GDP). The effect of in- income to net market income, suggesting that direct kind transfers on inequality reduction is much transfers — particularly pensions — are the main greater than the effect of indirect subsidies as shown drivers of inequality reduction. Direct taxes have by greater reduction in poverty from consumable also contributed towards inequality reduction income to final income compared to the reduction (moving from market income to net market from disposable income to consumable income. income). Education, at all levels, reduces inequality. Figure 1.8: The impact of fiscal policy on inequality, measured by the Gini coefficient Inequality, measured by the Gini Coefficient (per capita) 0.4 0.359 0.284 0.3 0.259 0.257 Gini coefficient 0.223 0.2 0.1 0.0 Market income Net market Disposable Consumable Final Income income income income Income concepts Source: World Bank calculations based on fiscal administrative data; the CEQ methodology described by Lustig 2022a, 2022b; data of the 2022 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan, Armenia The sensitivity analysis was also conducted to scenario, pension transfers were considered as estimate the impact of fiscal policy on inequality deferred income in the PDI scenario (refer to the PDI scenario. The results revealed a Appendix B for the details on PDI). The main significantly smaller inequality reduction drivers of inequality reduction in the PDI scenario associated with fiscal policy. The decline was were in-kind transfers in the form of health and smaller (7.6 Gini points) relative to the PGT education and social assistance transfers. scenario because, unlike in the case of the PGT 7 Equity impact is measured using the Gini index on a scale from between market income (pre-fiscal income) and the Gini at final 0 to 100, where 0 indicates perfect equality, and 100 indicates income (post-fiscal income after including all taxes, transfers, perfect inequality. For the impacts of taxes and transfers on and in-kind benefits in health care and education). inequality were measured as the difference in the Gini index 14 Armenia’s pension system consists of three pillars: Fiscal policy has played a crucial role in reducing 1). Non-contributive pensions, categorized as non- inequality, particularly in regions where pre-fiscal contributory and fully funded by taxes 2). Pay-as- income disparities were pronounced. Market you-go pensions for individuals born before 1974 income inequality is notably higher in certain marz, with at least 10 years of service. These include a basic such as Tavush and Vayots Dzor (refer to Figure pension plus additional remuneration based on 1.9). A disaggregated analysis across the marz years of service, rather than contributions. Funding reveals that the fiscal system has been more effective comes from contributions and income tax in reducing inequality in regions with relatively 3). Contributory pensions for those born after 1974, higher pre-fiscal inequality (from market income to directly linked to individual contributions. For final income), than in regions with relatively lower CEQ analysis, all benefits under Pillar 1 are pre-fiscal inequality such as in Aragatsotn. There classified as non-contributory pensions. are also slight differences between rural and urban Contributory pensions fall primarily under Pillar 3, areas in terms of inequality reduction. In rural areas, with some also in Pillar 2. While the system is the reduction is primarily driven by pensions, while designed to be contributory (consistent with the the impacts of direct transfers and in-kind transfers PDI scenario), low contribution levels relative to are roughly equivalent. In urban areas, pensions benefits make it function more like a transfer have the greatest impact, with in-kind transfers system, aligning it with the PGT scenario. being more effective than direct transfers. Figure 1.9: Impact of fiscal policy on inequality (measured by the Gini coefficient), by marz Impact of fiscal policy on inequality (by marz) (from MI to FI) 0.5 0.4 0.3 Gini 0.2 0.1 0.0 Market income Consumable income Final income Source: World Bank calculations based on fiscal administrative data; the CEQ methodology described by Lustig 2022a, 2022b; data of the 2021 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan, Armenia. Note: Regions are ordered based on the magnitude of the gap between MI and FI, from the largest to the smallest. Overall, the analysis showed that the impact of ranks 11th among 99 countries on Sustainable fiscal policy in reducing inequality was greater in Development Goal (SDG) indicator 10.4.2, which Armenia than in other upper-middle-income measures the redistributive impact of government countries in the region. The government’s fiscal fiscal policy (refer to Figure 1.10). 8 policy has achieved redistribution, and the country 8 Indicator 4.2 of SDG 10 assesses government effectiveness in social spending and transfers observed during and after the tackling inequality within the SDG framework. It is calculated pandemic. Refer to SDG Indicator 10.4.2: Redistributive Impact using the CEQ methodology as the difference between pre- of Fiscal Policy (dashboard), Sustainable Development Goals, fiscal and post-fiscal income inequality (as measured by the Office for National Statistics, Newport, South Wales UK, Gini coefficient). Estimates for most countries are based on pre- https://sdgdata.gov.uk/10-4-2/ pandemic data, which may not fully capture the increased 15 Figure 1.10: The redistributive impact of fiscal policy, SDG indicator 10.4.2 18 16 14 12 Armenia SDG 10.4.2 10 8.4 8 6 4 2 0 Source: World Bank calculations based on fiscal administrative data; the CEQ methodology described by Lustig 2022a, 2022b; data of the 2021 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan, Armenia; SDG Indicator 10.4.2: Redistributive Impact of Fiscal Policy (dashboard), Sustainable Development Goals, CEQ Institute, Commitment to Equity, Tulane University, New Orleans 1.2.3. Net Payers and Receivers of the Fiscal System The net cash position reveals that the bottom 40 aggregate sum of all cash-based interventions (all percent of the income distribution (the bottom 40) taxes, direct transfers, and indirect subsidies) for each are net cash recipients of the fiscal system, 9 owing decile. The net cash position provides a more accurate to large support from pensions and direct transfers. measure of household purchasing power, making the To determine which segments of Armenia's measure more effective in reflecting the impact of population experience cash gains or losses from taxes fiscal policy on poverty. In simpler terms, position and transfers, the analysis calculated the net cash shows whether the government has enabled position of households based on their market income individuals to afford goods and services beyond their decile (refer to Figure 1.11). The population was initial market income (Lustig 2022a, 2022b). The divided into 10 income groups (deciles) ranked by results show that the lower income groups (deciles 1- market income (pre-fiscal income). 10 For each decile, 4) benefit from the fiscal system, which helps to the stacked bars in Figure 1.11 show the incidence of reduce poverty, and that the more favorable fiscal the fiscal intervention with respect to market income. position of these groups contributes to reducing The net cash position (red dashed line) shows the inequality. 9 10 The net cash position by decile is calculated by aggregating the The size of the effects represents the overall picture by decile, relative incidence of taxes and direct transfers (excluding in-kind but there may be additional heterogeneity in the effects within benefits) with respect to pre-fiscal income (market income). If the deciles (Lustig 2022a, 2022b). net cash position is positive (negative), this indicates that the decile is a net receiver (payer) in the fiscal system. 16 Figure 1.11: The net fiscal position of households after taxes and transfers 600 500 400 300 Percent 200 100 0 -100 1 2 3 4 5 6 7 8 9 10 Deciles by Market Income plus pensions Direct Transfers Direct Taxes Contributions Indirect Taxes Subsidies Education Health Contributory pensions Net Cash Fiscal Position Source: World Bank calculations based on fiscal administrative data; the CEQ methodology described by Lustig 2022a, 2022b; data of the 2021 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan, Armenia, https://armstat.am/en/?nid=205 Estimated based on the fiscal position of The limited impact of fiscal policy on poverty households, the results show that the bottom 60 primarily derives from the low per capita income are net recipients when including in-kind benefits, of the poorest households and the limited while only the top 40 are net contributors to the generosity of transfers to households in the second fiscal system. The total fiscal position (gray dashed and third deciles. As Figure 1.11 shows, despite the line) shown in Figure 1.11 includes all cash-based exceptionally high relative net cash position among interventions, plus in-kind benefits (such as households in the first decile, their per capita education and health care benefits) valued at the income is so low that the transfers fail to lift them government’s cost of provision, which includes in- above the poverty line. This suggests that these kind benefits from health and education. The first households have a limited capacity to generate decile experiences significantly higher net benefits income, likely due to a high share of elderly relative to market income because of their reliance members which relies on pensions and government on pensions and direct transfers. The top 40 have a direct transfers. Out of 2.17 million population, negative fiscal position. about 363,332 individuals are supported by pensions. Among households in the second and third deciles, the generosity of benefits is quite low (total cash benefits accounts for, respectively, only 14.3 percent and 7.9 percent of pre-fiscal household income in these deciles), which explains the small impact of fiscal policy on poverty reduction. 17 1.2.4. The Progressivity of the Fiscal System A progressivity analysis of the fiscal system is, value of per capita benefits is greater for the indicates that direct transfers and education poorest households, and the Kakwani is higher than benefits are reaching lower-income households. 11 the market income Gini coefficient). A direct cash To assess progressivity, the Kakwani index (KI) — an transfer in the form of free-rent apartment is aggregate indicator of relative progressivity — was regressive. In addition, in-kind health benefits, both calculated for each fiscal intervention. The findings inpatient and outpatient, are only progressive and show that indirect taxes (such as the VAT and excise not pro-poor (a higher ratio of benefits to pre-fiscal on tobacco and alcohol) are regressive with a negative income among poorer households indicated by a Kakwani (refer to Figure 1.12), which indicates that positive Kakwani which is lower than the Gini lower-income households are paying a larger share of coefficient). In-kind education benefits are pro-poor their per-fiscal income compared to higher-income except for higher education and secondary households. Most direct transfers are pro-poor (that vocational, which are only progressive. Figure 1.12: The progressivity of selected fiscal policy instruments Progressivity as measured by Kakwani Tobacco Excise Free rent apartment Alcohol excise VAT Import tax Regressive Progressive Fuel tax Personal income tax Social security contributions Passive income tax Health inpatient Secondary vocational Education Tertiary Health outpatient Agricultural loan subsidy Stipendium Education Secondary Education Pre-school Education initial vocational Pro-poor Education middle school Child care Education Primary Health childbirth One off child birth Non-contributory pension Contributory pensions Family benefit Compensation for privileges Other child benefit Other benefits -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 Source: World Bank calculations based on fiscal administrative data; the CEQ methodology described by Lustig 2022a, 2022b; data of the 2021 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan, Armenia. Note: The threshold that distinguishes a program as merely progressive versus both progressive and pro-poor is the Gini coefficient of pre-fiscal (market) income. 11 Progressivity is measured by the Kakwani index (KI), which the Gini of the pre-fiscal income concept (market income). The compares the concentration coefficient of a tax or transfer (with KI for a transfer is calculated as the difference between the Gini respect to a reference income) to a Gini coefficient measured of the pre-fiscal income concept and the concentration over the reference income. A positive KI means that the fiscal coefficient of the transfer. In both cases, a positive KI means that intervention is more equally distributed than is the reference a tax or transfer is progressive, while a negative KI means that it income. A negative KI means that the fiscal intervention is less is regressive, and a zero KI means it is neutral (Lustig 2018, 2022a, equally distributed than is the reference income. A KI close to 2022b). If a concentration coefficient is negative for transfers, the zero indicates that the distribution of the fiscal intervention is Kakwani is higher than the pre-fiscal Gini, and the transfers are approximately equal to the distribution of the reference income. pro-poor. So, in this case, the per capita benefit is higher among Results are typically considered significant if the KI is higher than lower-income households. Refer to Enami, Larroulet, and (2022); 0.10 in absolute value. The KI for taxes is calculated as the Kakwani (1977), (1980); Kakwani and Pernia (2000). difference between the concentration coefficient of the tax and 18 1.2.5. Marginal Contributions to Inequality Reduction The personal income tax, pensions, and some middle school levels, contributed significantly to education transfers were the most significant inequality reduction (1.36, 1.28, and 0.79 Gini points, contributors to inequality reduction. The study respectively) (refer to Figure 1.13). The impact of estimated the marginal contributions (MCs) of direct taxes, especially the personal income tax, was fiscal interventions to assess which intervention has also significant in inequality reduction, at 1.39 Gini a greater impact in reducing inequality. 12 MCs points. This is because, despite the flat rate making consider the progressivity, size, and coverage of an PIT structurally regressive, poorer households instrument or group of instruments. The results derive a larger share of their income from informal indicate that government transfers, particularly sources, which are not subject to PIT. Health pensions, family benefits (direct transfers), and benefits, both inpatient and outpatient, do not education spending especially at primary and contribute to the reduction in inequality. Figure 1.13: Marginal contribution to inequality reduction of each fiscal instrument, in Gini points, measured at consumable income Health Inpatient Higher education Health outpatient Tobacco excise Social Security Contributions Free rent apartment transfer VAT Import taxes Initial Vocational Stipendium transfer Compensation for privileges transfer Alcohol excise Subsidy Agricultural Loans Passive income tax Other child transfers One off child birth transfer Fuel taxes, direct effect Other transfers Secondary vocational school Child care transfers Health Child Birth INEQUALITY REDUCING Secondary school Preschool Non-contributory pensions Middle school Education Primary School Family benefit PIT Contributory pensions -1 0 1 2 3 4 5 6 7 Gini points reduction (1 to 100) Source: World Bank calculations based on fiscal administrative data; the CEQ methodology described by Lustig 2022a, 2022b; data of the 2021 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan, Armenia, https://armstat.am/en/?nid=205 12 Marginal contributions (MCs) summarize the individual impact negative MC means that the fiscal intervention contributes to of a tax or transfer in reducing poverty and inequality when the an increase in the poverty or inequality indicator. MCs depend impact of all other fiscal interventions in the model is held on the size, progressivity, and coverage of fiscal interventions. constant. For example, the MC to inequality reduction is For the calculation of the MC to inequality reduction, the calculated as the difference between the Gini coefficient of the following formula was used: the Gini of consumable income, reference income (consumable income in this case) relative to minus the Gini of consumable income, including the specific tax the Gini coefficient of the reference income, minus (plus) the or transfer (minus for a tax and plus for a transfer). A similar specific tax (transfer). The procedure is similar for calculating formula is used for the calculation of MCs to poverty reduction, the MC to poverty reduction. A positive MC means that the fiscal based on the poverty difference with and without the tax or intervention contributes to poverty or inequality reduction; a transfer. 19 1.2.6. Marginal Contributions to Poverty Reduction The personal income tax and the VAT had adverse respectively) (refer to Figure 1.14). Government impact on poverty, whereas direct transfers and transfers, particularly family benefits and pensions, pensions reduced poverty. The analysis estimated helped reduce the poverty impact by 2.5 and 2.3 the MCs of the interventions to assess which has a percentage points, respectively. The impacts of greater impact in reducing poverty. The results others were minimal, although some taxes showed indicate that the personal income tax and the VAT adverse effect. increased poverty (by 12.6 and 5.7 percentage points, Figure 1.14: Marginal contributions of fiscal instruments to poverty reduction, consumable income measure PIT -12.6 VAT -5.7 Social Security Contributions -3.0 Tobacco excise -1.0 Fuel taxes, direct effect -0.9 Import taxes -0.4 Alcohol excise -0.3 Passive income tax 0.0 Free rent apartment transfer 0.0 Stipendium transfer 0.0 Other child transfers 0.1 One off child birth transfer 0.1 Direct Transfer Other 0.2 Subsidy Agricultural Loans 0.2 Compensation for privileges transfer 0.5 Child care transfers 0.8 Non-contributory pensions 2.3 Family benefit 2.5 Contributory pensions 18.9 -15 -10 -5 0 5 10 15 20 25 Poverty reduction in p.p. Source: World Bank calculations based on fiscal administrative data; the CEQ methodology described by Lustig 2022a, 2022b; data of the 2021 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan, Armenia 20 1.2.7. The Incidence of Fiscal Interventions The incidence of fiscal policy indicates that direct poor is partly driven by the tendency of the poor to taxes are paid mostly by the top quintile, whereas spend a larger share of their incomes on the the incidence of indirect taxes falls on the bottom consumption of goods and services subject to the quintile. 13 The top 20 pays about 65 percent of the VAT. However, the issue may not necessarily lie total passive income tax among all taxes collected. with the tax system itself – it also stems from the About 49.6 percent of the personal income tax and sources of income. Specifically, many households in 35.7 percent of the total VAT are also collected from the bottom quintile receives large share of income the top 20 (refer to Figure 1.15). This is because from fiscal transfers (especially from pensions). This wealthier households receive higher taxable suggests that the poor experience a incomes and consume more. However, the relative disproportionately high relative tax burden, not incidence indicates that the burden of the VAT is because the tax itself is regressive, but due to their disproportionately high among the bottom decile limited income base, which is largely derived from (refer to Figure 1.16). This higher tax burden on the post-transfer earnings. Figure 1.15: Absolute incidence: shares of total taxes paid by households, by market income 100 90 80 70 60 Percent 50 40 30 20 10 0 Personal Passive Social Alcohol Tobacco Fuel taxes VAT Import taxes Income tax Income tax security Excise Excise direct contributions Top 20% of the population ranked by market income 40-80% of the population ranked by market income 20-40% of the population ranked by market income Bottom 20% of the population ranked by market income Source: World Bank calculations based on fiscal administrative data; the CEQ methodology described by Lustig 2022a, 2022b; data of the 2021 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan 13 Absolute incidence is calculated as the total taxes (transfers) since absolute incidence is based on the aggregate transfers of that each decile pays (receives) as a share of total taxes each decile, it does not allow for distinguishing whether (transfers). In the case of social protection direct transfers, coverage or generosity is driving the results. 21 Figure 1.16: Relative incidence: (% of market income) 25 Relative incoence (% of pre-fiscal income) 20 15 10 5 0 1 2 3 4 5 6 7 8 9 10 Per capita market income deciles Personal Income tax VAT Source: World Bank calculations based on fiscal administrative data; the CEQ methodology described by Lustig 2022a, 2022b; data of the 2021 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan, Armenia The results also indicate that direct transfers and the form of free rent on apartments (69.8 percent) pensions provide crucial support to vulnerable (refer to Figure 1.17). In-kind education benefits populations. Of all government transfers, are distributed more progressively, especially compensation for the privileges, the family benefit primary and middle school education, which is and child benefit are mostly targeted on the consistent with the findings from Kakwani index bottom 20. In comparison, the two richest deciles (refer to Figure 1.18). receive a larger share of government transfers in Figure 1.17: Absolute incidence: total transfers, by household quintile and market income 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Bottom 20% of the population ranked by market income 20- 40% of the population ranked by market income 40-80% of the population ranked by market income Top 20% of the population ranked by market income Source: World Bank calculations based on fiscal administrative data; the CEQ methodology described by Lustig 2022a, 2022b; data of the 2021 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan, Armenia 22 Figure 1.18: The incidence of education transfers, by household quintile and market income a. Absolute incidence Concentration (share of benefits by decile) 40 30 20 10 0 Pre-school Primary Middle-school Secondary Initial vocational Secondary Higher vocational education Poorest 2 3 4 Richest Source: World Bank calculations based on fiscal administrative data; the CEQ methodology described by Lustig 2022a, 2022b; data of the 2021 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan, Armenia b. Relative incidence of educational benefits Incidence (benefits as % of pre-fiscal income) 16 14 12 10 8 6 4 2 0 Pre School Primary School Middle School Secondary Initial Vocational Secondary Tertiary School Vocational Poorest 2 3 4 Richest Source: World Bank calculations based on fiscal administrative data; the CEQ methodology described by Lustig 2022a, 2022b; data of the 2021 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan, Armenia 23 1.2.8. Conclusion In Armenia, the combination of taxes and primary contributors to inequality reduction were transfers significantly reduced inequality by 13.6 pensions, PIT, family benefits, along with in-kind Gini points and lowered national poverty by 2.4 benefits from education at primary and middle percentage points. Analysis showed that the school levels. In terms of poverty, most of the bottom 40 percent are the net cash beneficiaries of reduction resulted from direct transfers, namely the fiscal system, and bottom 60 percent are the pensions and Family benefits, which outweighed beneficiaries of the total fiscal system, including the poverty-increasing impact of PIT and VAT the in-kind transfers of education and health. The payments. 1.2.9. Limitations of the Analysis The CEQ assessment framework for FIA is a static first approximation of a true counterfactual. 15 In and retrospective accounting exercise without practice, there are no standard errors calculated behavioral, life-cycle, or general equilibrium that would allow a statistical assessment of the effects. Major challenges include the failure to allocations made to individuals and households. 16 capture top-income households and the exclusion Nonetheless, the CEQ assessment framework for of some interventions from the model — such as FIA provides a standard methodology (which also the CIT and public expenditure on infrastructure, enables international comparisons) for estimates of defense, or debt interest payments — because of the impact of fiscal policy on poverty, inequality, methodological limitations. 14 This means the and social welfare more generally. incidence results represent an accounting-based 14 16 Household surveys also do not capture some low-income There are, however, CEQ assessment procedures for assessing households well, such as institutional populations (prisons, the statistical significance of the estimated impact of a fiscal old-age care facilities, youth care facilities) and households policy (or set of fiscal policies) on poverty and inequality without a domicile address (informal housing). indicators and other indicators. 15 Meaning the results show distributional (static) impacts, not causal impacts. 24 C H A P T E R T W O The Distributional Impact of Universal Health Insurance The Health System in Armenia | The Universal Health Care System | Simulating the Impact of a UHI System | Estimating the Distributional Impact of the UHI System | Results | Conclusion | Limitations | 25 C H A P T E R T W O The Distributional Impact of Universal Health Insurance 2.1. The Health System in Armenia Health services are underfunded in Armenia, and co- expenditure in 2019 (World Bank 2023). As a result of payments are needed (NIH 2022). The health sector the high share of OOP, the burden of financing health receives low public financing, which raises equity care lies primarily with Armenian households, which concerns. Public health spending as a share of GDP has negative implications for welfare and equity. was less than 2 percent between 2000 and 2019. The According to a recent analysis, catastrophic health share rose to 2.3 percent in 2021, but remained lower spending — that is household health expenditure at than the average of 3.4 percent in upper-middle- or more than 10 percent of total household income countries and peers (such as Albania, Estonia, consumption — exceeded 19.9 percent in 2021 and was Georgia, and Tunisia). 17 Current health expenditure particularly common in households in which per capita, adjusted for purchasing power parity in members were living with chronic diseases or were Armenia, was US$1,924 in 2021, an increase from the unemployed, households in urban areas, and lower- US$135 in 2001. 18 The burden of financing health care income households. 19 The share of such households lies primarily with Armenian households. This is exceeds the regional average of 7.4 percent. 20 The reflected in exceptionally high out-of-pocket (OOP) share of total OOP expenditures in Armenia expenditure compared with the Europe and Central accounted for 78.7 percent of total health expenditure Asia regional average and the upper-middle-income- in 2021, which is the highest share among the selected country average. peer countries and double the upper-middle-income- country average (44.1 percent) (WHO and European Voluntary health insurance schemes (prepaid, Observatory on Health Systems and Policies 2022). private health spending) play a marginal role in The high OOP levels have repercussions for financial Armenia, which leads to high share of OOP. Multiple protection and accessibility, particularly among voluntary health insurance schemes exist, but vulnerable groups. accounted for only 1.1 percent of total health 17 19 Refer to GHED (Global Health Expenditure Database), World Health The indicator on catastrophic health spending is measured as the Organization, Geneva, https://apps.who.int/nha/database percentage share of the population that has household health expenditures exceeding 10 percent of total household expenditure 18 Refer to WDI (World Development Indicators) (dashboard), World or income. Bank, Washington, DC, https://datatopics.worldbank.org/world- 20 development-indicators Refer to GHED (Global Health Expenditure Database), World Health Organization, Geneva, https://apps.who.int/nha/database 26 2.2. The Universal Health Care System Armenia does not have UHI, whether contributory (World Bank 2024a). Armenia faces a growing or noncontributory. The population is entitled to burden of noncommunicable disease, and the aging the Basic Benefits Package (BBP), a subsidized population is also a critical issue. The share in the health care program that provides varying levels of total population of the population ages 65 or more coverage (refer to Appendix C). The BBP is funded rose from 6.5 percent in 1993 to 13.7 percent in in through general government revenues and includes 2023. 21 emergency care and maternal and child health services. It also covers certain medications and Although the country has been undertaking diagnostic services for specific groups (for instance, important reforms in governance, service delivery, social communities and vulnerable populations) and health sector financing, substantial gaps and addresses certain conditions (such as diabetes, persist. 22 To counter challenges in the health care mental health disorders, and epilepsy), regardless of system and health care delivery, the government is beneficiary social status. working toward the implementation of a UHI system. If implemented, the system would be Health care access in Armenia is hospital-centric designed to ensure affordable access to basic health and covers a growing share of the aging care services by all population groups, enhance the population. One Armenian in five forgoes health quality of medical care, and improve health care because of inability to pay the cost of services outcomes. 2.3. Simulating the Impact of a UHI System Using the CEQ methodology, the analysis the terms of civil law (including public sector estimated the distributional impact of the UHI. For employees), the self-employed, hired employees in the purpose of this simulation, it is assumed that the the other sectors, agricultural workers, notaries, and UHI system will be rolled out in phases, with individuals with incomes from other sources, such coverage progressively established within four years. as lease agreements, dividends, loans, and royalties. The parameters of the implementation plan, Depending on the employment category, the initial including the determination of user fees and year of the insurance premium payment will vary premium payments, is presented in Table 2.1. 23 It is during the phased implementation of the UHI. All assumed that the insurance premium per insured categories mentioned in Table 2.1 will pay the fixed person will be AMD 164,400 per calendar year and amount by the full launch of the program. will be paid by government employees hired under 21 Refer to WDI (World Development Indicators) (dashboard), World strategy, which aims to improve the quality of primary health Bank, Washington, DC, https://datatopics.worldbank.org/world- care and ensure accessible health care services across the development-indicators country. 23 22 The government of Armenia has adopted a five-year program These assumptions are based on the draft UHI law that was to improve inclusive growth and eliminate extreme poverty prepared by the government in February 2024. (Office of the Prime Minister 2023). The 2021–26 program is a subset of the national health care system development 27 Table 2.1: Draft implementation schedule for the UHI system, Armenia Beneficiaries Insurance premium payers Test Test Test Full Test Test Test Full phase phase phase launch phase phase phase launch year 1 year 2 year 3 year 1 year 2 year 3 Children (people aged 0-18) Half year Yes Yes Yes No No No No Persons without parents (18-23 years) Half year Yes Yes Yes No No No No Students under the age of 23 Half year Yes Yes Yes No No No No Full time students (23-26 years) Half year Yes Yes Yes No No No No Pensioners (63 and older) No Yes Yes Yes No No No No Disabled people aged 18 to 63 Half year Yes Yes Yes No No No No Persons from 18 to 63 years of age included in socially disadvantaged and special groups Military or with family in active service Half year Yes Yes Yes No No No No Women with children under the age of 2 No No No Yes No No No No Guardianship (care) of 3 or more minor children No No No Yes No No No No Population aged 18 to 63 and self-employed Social protection program beneficiaries Half year Yes Yes Yes No No No No Government employees No Yes Yes Yes No Yes Yes Yes Hired workers No No Yes Yes No No Yes Yes Self-employed, notaries No No Yes Yes No No Yes Yes Agricultural workers No No Yes Yes No No Yes Yes Others (relatives of hired workers and self- employed) No No No Yes No No No Yes Source: Draft Universal Health Insurance Law, February 14, 2024 28 2.4. Estimating the Distributional Impact of the UHI System To assess the distributional impact of health where: reforms on inequality and poverty, the analysis = disposable income (income after direct taxes, focuses on the distribution of disposable income. plus government cash transfers) 25 This refers to income after direct taxes, including = income modified (after including the SPM transfer payments and refundable tax credits. The definition and Barofsky and Younger method) definition of disposable income does not account (refer to Appendix D) for indirect taxes or the value of public spending = public health (the monetized value of per capita benefits, such as health care and education benefits public health benefits received by households) or most housing subsidies (Smeeding and = medical OOP (as reported in the survey) Rainwater 2002). To assess how publicly funded in- = education OOP kind health care transfers would affect the income = childcare expenses. distribution, the monetary value to beneficiaries must be estimated (Barofsky and Younger 2019; The distributional impact of the UHI is estimated Lustig 2022a, 2022b). 24 using the CEQ methodology. In the CEQ methodology, government spending in the form of For the estimation of the impact of government education and health are treated as in-kind transfers health spending on poverty, both the and there are two approaches, government cost and supplemental poverty measure (SPM) insurance value for the monetization and valuation of methodology and the Barofsky and Younger (2019) the healthcare services provided by the government method are applied (refer to the prevailing to the households. These approaches include the literature cited in Appendix D; for the methodology, “actual consumption approach” and the “insurance refer to Appendix E). Applying the Barofsky and value approach”. The first approach allocates the Younger (2019) method, household disposable value of public services to the individuals who are income was taken from the survey, and the using the service. The second approach assigns the monetary value of the benefits provided by the same per capita spending to everybody sharing the government for health care was included in same characteristic such as age, state, type of care, household resources. The modified income gender etc. One special case of the “insurance value definition adopted help us estimate the impact of approach” is using eligibility to a specific health UHI at “final income” level. The modified income system as the shared characteristic. The reliance on definition in the case of Armenia is as follows: one approach over the other depends, mainly, on data availability such as identification of beneficiaries of public health services and the type of services = + − − − (2.1) received, and total government spending on each of the different types of health services. 24 25 While monetizing provides a sense of the financial value of in- Disposable income is the conventional concept of the amount kind transfers, estimating the impact of such transfers on of money remaining in the pockets of household members to poverty is challenging. This is because (a) in-kind transfers do purchase goods, give away, or save after the government takes not always translate directly into increased cash income, and some away in direct taxes and gives some back in direct (2) estimating the impact against a fixed poverty line based on transfers. income is not straightforward. As a convention, the CEQ approach does not generally involve the estimation of the impact of in-kind transfers on a poverty rate. 29 Estimating the impact of proposed UHI reforms in UHI system; and comparison with the baseline Armenia consisted of several steps. This included CEQ results. (Details on the methodology adopted monetization and the inclusion of existing health for each of these steps are provided in Appendix F). benefits in the baseline CEQ analysis; simulating Poverty was measured at final income after health spending (including identification of the accounting for the impact of OOP and public beneficiaries and estimation of insurance premium health spending. Figure 2.1 illustrates the approach payments); estimating poverty through health within the CEQ framework based on the above benefits; measuring the distributional impact of the method. Figure 2.1: Simulating UHI benefits using the CEQ framework a. Household income before UHI b. Household income after UHI Disposable income = Consumption Disposable income = Consumption + PH - EOOP - CC - IP + PH - MOOP - EOOP - CC Reported health expenditures in ILCS Yes No - MOOP* + BMuhi = Modified Income baseline scenario = Modified Income UHI simulated scenario Note: The figure was constructed based on the SPM and the Barofsky and Younger (2019) methodologies. BMuhi = the estimated medical benefits under UHI (health care to be received under UHI). CC = childcare expenses. EOOP = education OOP. IP = insurance premium. MOOP = medical OOP (as reported in the survey). PH = public health (monetized value of per capita public health benefits received by households). Using the methodology outlined above, poverty The benefits or net savings are measured as the was estimated both ex ante and ex post, and the difference between actual spending and spending two were compared. Ex post poverty was calculated after the UHI. A significant share of poor as disposable income, minus the change in expected households either did not report OOP expenses or OOP expenses. A reduction in poverty indicates a reported low amounts. As a result, the net savings positive effect. Among individuals who reported or benefits among these households are minimal. health expenditures in the survey, the change in The low OOP spending among poor households expected OOP was calculated as the difference may be attributed either to the unaffordability or to between actual OOP and UHI OOP. Among those the unavailability of health care. The distribution of who did not report health expenditures, the change UHI benefits and insurance premiums is shown in in expected OOP was determined by subtracting Figure 2.2. The incidence of UHI benefits and UHI OOP from the estimated OOP. insurance premiums as a share of disposable income reveals a noticeable concentration of benefits in the higher income deciles (refer to Figure 2.3). 30 Figure 2.2: The concentration of UHI benefits and insurance premiums, by decile a. Insurance premiums b. Benefits (savings, plus health care) Insurance premium (% of total by per capita disposable 60 income deciles) (% of total benefits by decile) 30 50 25 40 20 30 Pcrcent 15 20 10 5 10 0 0 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Per capita disposable income deciles Per capita disposable income deciles Test phase year 1 Test phase year 2 Test phase year 1 Test phase year 2 Test phase year 3 Full launch Test phase year 3 Full launch Source: World Banks’ staff calculation based on ILCS 2021 and Draft Universal Health Insurance Law, February 14, 2024 Figure 2.3: The incidence of universal health insurance benefits and insurance premiums a. Insurance premiums b. Benefits (savings, plus health care) 25 ( paymes as % of disposablwe income) ( paymes as % of disposablwe income) 20 10 15 10 5 5 0 0 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Per capita disposable income deciles Per capita disposable income deciles Test phase year 1 Test phase year 2 Test phase year 1 Test phase year 2 Test phase year 3 Full launch Test phase year 3 Full launch Source: World Banks’ staff calculation based on ILCS 2021 and Draft Universal Health Insurance Law, February 14, 2024. Notes: Premium amount is very low during the first year of the test phase therefore blue bars are appearing low in Figure 2.3-a. Calculation was made based on the CEQ, SPM and the Barofsky and Younger (2019) methodologies. 31 2.5. Results The distributional impact results indicate that UHI results therefore indicate that the reforms may implementation will reduce poverty and slightly benefit these households more substantially. Health increases inequality. The impact on poverty is a decisions involve trade-offs between various factors, reduction by 0.6 percentage points from pre-reform such as provider choice, cost (including both to full launch (refer to Figures 2.4 and 2.5). The monetary costs and nonmonetary costs, such as reduction in poverty derives from the protection of transport and waiting time), health status, and poor households from catastrophic expenditures. income. Large numbers of the poor reported zero The results also indicate that inequality will be OOP spending, mainly because they cannot afford increased by 2.2 Gini index points, from pre-reform medical visits. A UHI system would enable these to full launch. This may be attributed to data individuals to access health services they previously limitations because a large number of the poor did could not afford. Assuming UHI benefits are not report OOP spending. In this analysis, benefits distributed similarly across age cohorts, relative or net savings are calculated as the difference savings are expected to decline as income rises, between actual spending and spending after UHI. which would have an equalizing effect on income Wealthier households tend to spend more on health inequality. 26 care, and they reported higher expenditures. The Figure 2.4: The impact of universal health insurance on inequality Inequality, measured by the Gini coefficient (modified income) 0.250 0.231 0.230 0.232 0.236 0.214 0.200 Gini coefficient 0.150 0.100 0.050 0.000 Pre-reform Test phase year 1 Test phase year 2 Test phase year 3 Full launch Source: World Banks’ staff calculation based on ILCS 2021 and Draft Universal Health Insurance Law, February 14, 2024 26 The poverty reduction impact seems to be greater if only the launch year, it would 16.5 percent (because of the impact of the households that reported OOP health spending in the ILCS are payment of the insurance premium). So, poverty reduction is included in the sample and analyzed. As an alternate scenario, greater among those households reporting health expenditures households that only reported health expenditures are included, in the survey. On the other hand, among those not reporting and the distributional impact of UHI was analyzed. In this case, health expenditures, poverty as a result of the savings in OOP the baseline poverty rate according to the SPM income and insurance premium payments would increase. However, methodology among this sample was 19.4 percent. In the second this estimate does not consider the effects of the benefits year, the poverty rate would drop to 14.8 percent, and, in the full derived from access to health care. 32 Figure 2.5: The impact of universal health insurance on poverty Poverty (headcount) modified income 28 26.6 26.0 26 23.9 24 22.0 21.7 22 Percent 20 18 16 14 12 10 Pre-reform Test phase year 1 Test phase year 2 Test phase year 3 Full launch Source: World Banks’ staff calculation based on ILCS 2021 and Draft Universal Health Insurance Law, February 14, 2024 2.6. Conclusion The implementation of the UHI system is expected by wealthy households under the current system, to reduce poverty, improve health equity, and and the analysis does not account for the potential render health care more accessible and affordable, behavioral change. particularly among the vulnerable. The analysis highlights the interplay of socioeconomic factors in The welfare benefits of the UHI system include shaping health care expenditure and evaluates the improved health outcomes. Health insurance is impact of UHI reforms on OOP spending. The UHI designed to cover necessary care, regardless of an system would enhance health equity and reduce individual or family's health condition. The goal of poverty by shielding vulnerable groups from universal coverage is to encourage the use of health financial shocks. If the take-up of health services services, ensuring that those who cannot afford care increased more among the poor and the vulnerable, are not forced to suffer or face serious health the impact on poverty reduction is expected to be consequences. The underlying rationale is that larger because the UHI system would offer financial health expenditures are not discretionary. By protection against catastrophic health expenditures expanding universal health care, it is expected that and promote equitable health care access, OOP expenses will decline, allowing low-income contributing to inclusive growth. The rise in individuals to allocate more resources to other inequality reported in the analysis largely derives essential needs. from the disproportionate reliance on health care 33 2.7. Limitations The major limitations of the estimates are twofold: that individuals reporting health expenditures in data limitations and methodological limitations. the ILCS would continue to spend similar amounts The household survey (ILCS) seems to underreport under the proposed system. The analysis assumed health expenditures. Only one-third of the no differences between the copayments currently population reported health-related spending and applied across the insured population and those provided detailed expenditure data. This affects anticipated under the UHI once operational. The both the baseline model and the reform simulations. specific copayment structure will be determined Because of the small share of respondents reporting during the implementation of the UHI. health spending, several assumptions were made to Additionally, it was assumed that all households estimate unreported expenditures. 27 Among enrolled in the UHI would have full access to health respondents who reported expenditures, care services. The implementation of the reform significantly larger health spending was recorded by will require substantial fiscal allocations from the higher-income deciles than by lower-income government because insurance premium collection groups. A notable number of poorer households is not sufficient to ensure the success of the system. reported zero OOP spending, and it could not be ascertained whether this reflected measurement The implementation of the UHI system would error, underreporting, or actual spending patterns require increased fund allocations to the health reflecting the inaccessibility to services because of sector. The UHI is the most comprehensive health financial constraints. Additionally, the survey does care reform proposed for implementation by the not specify whether health expenditures were government and therefore would require significant incurred within Armenia or abroad. Furthermore, funding. The UHI system would require wide- by age-group, reported health spending on behalf of ranging coverage in terms both of types of services children was particularly low. The methodological and the population. The need for investments in limitation of the CEQ model applies in the personnel, health infrastructure, and other inputs to simulations as well. 28 ensure and improve service quality would also increase financial needs. It is thus critical that the The analysis assumed consistency in the health UHI rollout be preceded by an effort to understand spending patterns of individuals, copayment clearly the cost implications and the financing plan structures, and health care access. It was assumed (World Bank 2023). 27 28 The assumptions were made after consultations with the World Major limitations include the static, retrospective analysis that Bank Armenia health team and the World Bank does not incorporate consideration of behavioral change, life- Macroeconomics, Trade, and Investment Global Practice team. cycle, or general equilibrium effects (refer to Chapter 1). 34 References 1. Adusah-Poku, Frank, and Kenji Takeuchi. 2019. “Household Energy Expenditure in Ghana: A Double- Hurdle Model Approach.” World Development 117 (May): 266–77. https://doi.org/10.1016/j.worlddev.2019.01.018 2. Akazili, James, Diane McIntyre, Edmund W. Kanmiki, John Gyapong, Abraham Oduro, Osman Sankoh, and John E. Ataguba. 2017. “Assessing the Catastrophic Effects of Out-of-Pocket Healthcare Payments Prior to the Uptake of a Nationwide Health Insurance Scheme in Ghana.” Global Health Action 10 (1), 1289735. https://doi.org/10.1080/16549716.2017.1289735 3. Amaya-Lara, Jeannette Liliana. 2016. “Catastrophic Expenditure Due to Out-of-Pocket Health Payments and Its Determinants in Colombian Households.” International Journal for Equity in Health 15 (1): 1–11. https://doi.org/10.1186/s12939-016-0472-z 4. Ampaw, Samuel, Edward Nketiah-Amponsah, Frank Agyire-Tettey, and Bernardin Senadza. 2020. “Distributional Analysis of Rural-Urban Household Healthcare Expenditure Differentials in Developing Countries: Evidence from Ghana.” International Journal of Development Issues 19 (3): 359–75. https://doi.org/10.1108/ijdi-07-2019-0126 5. Ampaw, Samuel, Edward Nketiah-Amponsah, Nkechi Srodah Owoo, and Bernardin Senadza. 2019. “Nonfarm Enterprise Participation and Healthcare Expenditure among Farm Households in Rural Ghana.” International Journal of Social Economics 46 (1): 18–30. https://doi.org/10.1108/ijse-06-2017-0248 6. Antošová, Irena, Naďa Hazuchova, and Jana Stávková. 2019. “Market Segmentation in Healthcare.” Marketing and Management of Innovations 10 (3): 151–66. https://mmi.sumdu.edu.ua/volume-10-issue-3 7. Asgary, Ali, Ken Willis, Ali Akbar Taghvaei, and Mojtaba Rafeian. 2004. “Estimating Rural Households Willingness to Pay for Health Insurance.” European Journal of Health Economics 5 (3): 209–15. https://doi.org/10.1007/s10198-004-0233-6 8. Barofsky, Jeremy T., and Stephen D. Younger. 2019. “The Effect of Government Health Expenditure on the Income Distribution: A Comparison of Valuation Methods in Ghana.” CEQ Working Paper 66 (June), Commitment to Equity, CEQ Institute, Tulane University, New Orleans. 9. Blank, Rebecca M. 2008. “Presidential Address: How to Improve Poverty Measurement in the United States.” Journal of Policy Analysis and Management 27 (2): 233–54. 10. Borde, Moges Tadesse, Eskindir Loha, Kjell Arne Johansson, and Bernt Lindtjørn. 2020. “Financial Risk of Seeking Maternal and Neonatal Healthcare in Southern Ethiopia: A Cohort Study of Rural Households.” International Journal for Equity in Health 19 (1), 69. https://doi.org/10.1186/s12939-020-01183-7 11. Burkhauser, Richard V., Kevin Corinth, James Elwell, and Jeff Larrimore. 2019. “Evaluating the Success of President Johnson’s War on Poverty: Revisiting the Historical Record Using an Absolute Full-Income Poverty Measure.” NBER Working Paper 26532 (December), National Bureau of Economic Research, Cambridge, MA. 35 12. Callander, Emily J., Haylee Fox, and Daniel Lindsay. 2019. “Out-of-Pocket Healthcare Expenditure in Australia: Trends, Inequalities and the Impact on Household Living Standards in a High-Income Country with a Universal Health Care System.” Health Economics Review 9 (1), 10. https://doi.org/10.1186/s13561-019-0227-9 13. CEA (Council of Economic Advisers, United States). 2019. Economic Report of the President Together with the Annual Report of the Council of Economic Advisers. March. Washington, DC: CEA, White House. 14. Chukwuma, Adanna, Bruno Meessen, Hratchia Lylozian, Estelle Gong, and Emma Ghazaryan. 2020. Strategic Purchasing for Better Health in Armenia. Washington, DC: World Bank. 15. Corbett, Thomas. 1999. “Poverty: Improving the Measure after Thirty Years, A Conference.” Focus 20 (2): 51–55. 16. Creamer, John. 2024. “Health Inclusive Poverty Measure in the United States: 2022.” Working Paper SEHSD-WP-2024-05 (March), Social, Economic, and Housing Statistics Division, US Census Bureau, Washington, DC. 17. de Broucker, Gatien, Sayem Ahmed, Md. Zahid Hasan, Gazi Golam Mehdi, Jorge Martin del Campo, Md. Wazed Ali, Md. Jasim Uddin, Dagna Constenla, and Bryan Patenaude. 2020. “The Economic Burden of Measles in Children under Five in Bangladesh.” BMC Health Services Research 20 (November), 1026. https://doi.org/10.21203/rs.3.rs-45001/v3 18. Ellwood, David T., and Lawrence H. Summers. 1986. “Poverty in America: Is Welfare the Answer or the Problem?” In Fighting Poverty: What Works and What Doesn’t, edited by Sheldon H. Danziger and Daniel H. Weinberg, 78–105. Cambridge, MA: Harvard University Press. 19. Enami, Ali, Patricio Larroulet, and Nora Claudia Lustig. 2022. “How Accurate Is the Kakwani Index in Predicting Whether a Tax or a Transfer Is Equalizing? An Empirical Analysis.” CEQ Working Paper 116 (January), Commitment to Equity, CEQ Institute, Tulane University, New Orleans. 20. Friedman, Jerome H. 1984. “A Variable Span Smoother.” Technical Report LCS 5 (November), Laboratory for Computational Statistics, Department of Statistics, Stanford University, Stanford, CA. 21. Garfinkel, Irwin, Lee Rainwater, and Timothy M. Smeeding. 2005. “Welfare State Expenditures and the Redistribution of Well-Being: Children, Elders, and Others in Comparative Perspective.” LIS Working Paper 387 (September), Luxembourg Income Study, LIS Cross-National Data Center, Luxembourg. 22. Gertler, Paul J., Luis Locay, and Warren Sanderson. 1987. “Are User Fees Regressive? The Welfare Implications of Health Care Financing Proposals in Peru.” Journal of Econometrics 36 (1–2): 67–88. https://doi.org/10.1016/0304-4076(87)90044-3 23. Ichwan, Mohamad, Firmansyah Firmansyah, and Eko Jokolelono. 2021. “Health Demand: Empirical Study of Effective Urban Households Demand in Indonesia.” Macedonian Journal of Medical Sciences 9 (E): 812–16. https://doi.org/10.3889/oamjms.2021.6532 24. IMF (International Monetary Fund). 2022. “Republic of Armenia, 2023 Article IV Consultation and Second Review under the Stand-By Arrangement and Request for Modifications of Performance Criteria and Monetary Policy Consultations Clause: Press Release, Staff Report, and Statement by the Executive Director.” IMF Country Report 23/416 (December), IMF, Washington, DC. 36 https://www.imf.org/en/Publications/CR/Issues/2023/12/14/Republic-of-Armenia-2023-Article-IV- Consultation-and-Second-Review-Under-the-Stand-By-542533 25. Kakwani, Nanak C. 1977. “Measurement of Tax Progressivity: An International Comparison.” Economic Journal 87 (345): 71-80. 26. Kakwani, Nanak C. 1980. Income Inequality and Poverty: Methods of Estimation and Policy Applications. New York: Oxford University Press. 27. Kakwani, Nanak C., and Ernesto M. Pernia. 2000. “What Is Pro-Poor Growth?” Asian Development Review 18 (1): 1–16. 28. Khan, Jahangir A. M., Sayem Ahmed, and Timothy G. Evans. 2017. “Catastrophic Healthcare Expenditure and Poverty Related to Out-of-Pocket Payments for Healthcare in Bangladesh: An Estimation of Financial Risk Protection of Universal Health Coverage.” Health Policy and Planning 32 (8): 1102–10. https://doi.org/10.1093/heapol/czx048 29. Korenman, Sanders D., and Dahlia K. Remler. 2016. “Including Health Insurance in Poverty Measurement: The Impact of Massachusetts Health Reform on Poverty.” Journal of Health Economics 50 (December): 27–35. 30. Lavado, Rouselle F., Susanna Hayrapetyan, and Samvel Kharazyan. 2018. “Expansion of the Benefits Package: The Experience of Armenia.” Universal Health Coverage Study 27, World Bank, Washington, DC. 31. Lee, Munjae, and Kichan Yoon. 2019. “Catastrophic Health Expenditures and Its Inequality in Households with Cancer Patients: A Panel Study.” Processes 7 (1), 39. https://doi.org/10.3390/pr7010039 32. Li, Chengming, Daming Li, Si He, Shiqi Sun, Yuan Tian, and Zeyu Wang. 2022. “The Effect of Big Data- Based Digital Payments on Household Healthcare Expenditure.” Frontiers in Public Health 10 (June), 922574. https://doi.org/10.3389/fpubh.2022.922574 33. Lustig, Nora Claudia, ed. 2018. Commitment to Equity Handbook: Estimating the Impact of Fiscal Policy on Inequality and Poverty. New Orleans: CEQ Institute, Tulane University; Washington, DC: Brookings Institution Press. 34. Lustig, Nora Claudia, ed. 2022a. Fiscal Incidence Analysis: Methodology, Implementation, and Applications. Vol. 1 of Commitment to Equity Handbook: Estimating the Impact of Fiscal Policy on Inequality and Poverty, 2nd ed. New Orleans: CEQ Institute, Tulane University; Washington, DC: Brookings Institution Press. 35. Lustig, Nora Claudia, ed. 2022b. Methodological Frontiers in Fiscal Incidence Analysis. Vol. 2 of Commitment to Equity Handbook: Estimating the Impact of Fiscal Policy on Inequality and Poverty, 2nd ed. New Orleans: CEQ Institute, Tulane University; Washington, DC: Brookings Institution Press. 36. Mahumud, Rashidul Alam, Abdur Razzaque Sarker, Marufa Sultana, Ziaul Islam, Jahangir A. M. Khan, and Alec Morton. 2017. “Distribution and Determinants of Out-of-Pocket Healthcare Expenditures in Bangladesh.” Journal of Preventive Medicine and Public Health 50 (2): 91–99. https://doi.org/10.3961/jpmph.16.089 37 37. Mitchell, Emily M. 2019. “Concentration of Health Expenditures and Selected Characteristics of High Spenders, U.S. Civilian Noninstitutionalized Population, 2016.” Statistical Brief 521 (February), Medical Expenditure Panel Survey, Agency for Healthcare Research and Quality, US Department of Health and Human Services, Rockville, MD. 38. Molla, Azaher Ali, Chunhuei Chi, and Alicia Lorena Núñez Mondaca. 2017. “Predictors of High Out-of- Pocket Healthcare Expenditure: An Analysis Using Bangladesh Household Income and Expenditure Survey, 2010.” BMC Health Services Research 17 (January), 94. https://doi.org/10.1186/s12913-017-2047-0 39. Moon, Marilyn. 1993. “Incorporating Health Issues in the Measurement of Poverty.” July, Urban Institute, Washington, DC. 40. NIH (National Institute of Health, Armenia). 2022. National Health Accounts of Armenia, 2021. Yerevan, Armenia: National Institute of Health, Ministry of Health. https://new.nih.am/national-health-accounts-annual-reports/national-health-accounts-of-armenia-2021 41. Nübler, Laura, Reinhard Busse, and Martin Siegel. 2023. “The Role of Consumer Choice in Out-of-Pocket Spending on Health.” International Journal for Equity in Health 22 (January), 24. https://doi.org/10.1186/s12939-023-01838-1 42. Office of the Prime Minister, Armenia. 2023. “Five-Year Action Program.” Office of the Prime Minister, Yerevan, Republic of Armenia. https://www.gov.am/en/Five-Year-Action-Program 43. Ogundari, Kolawole, and Awudu Abdulai. 2014. “Determinants of Household's Education and Healthcare Spending in Nigeria: Evidence from Survey Data.” African Development Review 26 (1): 1–14. https://doi.org/10.1111/1467-8268.12060 44. Olasehinde, Noah, and Olanrewaju Olaniyan. 2017. “Determinants of Household Health Expenditure in Nigeria.” International Journal of Social Economics 44 (12): 1694–1709. https://doi.org/10.1108/ijse-12-2015-0324 45. Panel on Poverty and Family Assistance, Committee on National Statistics, Commission on Behavioral and Social Sciences and Education, and National Research Council. 1995. Measuring Poverty: A New Approach. Edited by Constance F. Citro and Robert T. Michaels. Washington, DC: National Academy Press. 46. Rahman, Md. Mizanur, Stuart Gilmour, Eiko Saito, Papia Sultana, and Kenji Shibuya. 2013. “Health- Related Financial Catastrophe, Inequality and Chronic Illness in Bangladesh.” PLoS One 8 (2), e56873. https://doi.org/10.1371/journal.pone.0056873 47. Remler, Dahlia K. 2019. “‘Free Care’, Health Insurance Reforms, and Health-Inclusive Poverty in New York.” Presentation at the 8th Conference of the American Society of Health Economists, “The Crossroads of Public Policy and Health Economics,” Washington, DC, June 23–26, 2019. https://ashecon.confex.com/ashecon/2019/webprogram/Paper8284.html 48. Remler, Dahlia K., and Sanders D. Korenman. 2023. “On the Importance and Intrinsic Difficulties of Incorporating Health Insurance Benefits in Absolute Poverty Trends.” Journal of Health Politics, Policy and Law 48 (5): 761–98. 38 49. Remler, Dahlia K., Sanders D. Korenman, and Rosemary T. Hyson. 2017. “Estimating the Effects of Health Insurance and Other Social Programs on Poverty under the Affordable Care Act.” Health Affairs 36 (10): 1828–37. 50. Rezaei, Satar, and Mohammad Hajizadeh. 2019. “Measuring and Decomposing Socioeconomic Inequality in Catastrophic Healthcare Expenditures in Iran.” Journal of Preventive Medicine and Public Health 52 (4): 214–23. https://doi.org/10.3961/jpmph.19.046 51. Ruggles, Patricia. 1990. Drawing the Line: Alternative Poverty Measures and Their Implications for Public Policy. Washington, DC: Urban Institute Press. 52. Ruggles, Patricia. 2008. “Poverty Measurement and Politics: Why Is It So Hard to Change the Way We Measure Poverty?” Review of Agricultural Economics 30 (3): 591–99. 53. Sanoussi, Yacobou, Alexandre Yedjannavo Zounmenou, and Muriel Ametoglo. 2023. “Catastrophic Health Expenditure and Household Impoverishment in Togo.” Journal of Public Health Research 12 (3), 22799036231197196. 54. Sassi, Franco, Annalisa Belloni, Andrew J. Mirelman, Marc Suhrcke, Alastair Thomas, Nisreen Salti, Sukumar Vellakkal, Chonlathan Visaruthvong, Barry M. Popkin, and Rachel Nugent. 2018. “Equity Impacts of Price Policies to Promote Healthy Behaviours.” Lancet 391 (10134): 2059–70. https://doi.org/10.1016/s0140-6736(18)30531-2 55. Shafie, Asrul Akmal, and Mohamed Azmi Hassali. 2013. “Willingness to Pay for Voluntary Community- Based Health Insurance: Findings from an Exploratory Study in the State of Penang, Malaysia.” Social Science and Medicine 96 (November): 272–76. https://doi.org/10.1016/j.socscimed.2013.02.045 56. Short, Kathleen. 2011. “The Research Supplemental Poverty Measure: 2010.” Current Population Report P60-241 (November), US Census Bureau, Economics and Statistics Administration, US Department of Commerce, Suitland, MD. 57. Smeeding, Timothy M., and Lee Rainwater. 2002. “Comparing Living Standards across Nations: Real Incomes at the Top, the Bottom and the Middle.” LIS Working Paper 266 (February), Luxembourg Income Study, LIS Cross-National Data Center, Luxembourg. 58. Sommers, Benjamin D., and Donald Oellerich. 2013. “The Poverty-Reducing Effect of Medicaid.” Journal of Health Economics 32 (5): 816–32. https://doi.org/10.1016/j.jhealeco.2013.06.005 59. Tur-Sinai, Aviad, Racheli Magnezi, and Haya Grinvald-Fogel. 2018. “Assessing the Determinants of Healthcare Expenditures in Single-Person Households.” Israel Journal of Health Policy Research 7 (October), 48. https://doi.org/10.1186/s13584-018-0246-8 60. United Nations. 1948. “Universal Declaration of Human Rights.” General Assembly Resolution 217 A (December 10), United Nations, New York. https://www.un.org/en/about-us/universal-declaration-of-human-rights 61. Venkataraman, Surendran, Suguna Anbazhagan, and Surekha Anbazhagan. 2019. “Expenditure on Health Care, Tobacco, and Alcohol: Evidence from Household Surveys in Rural Puducherry.” Journal of Family Medicine and Primary Care 8 (3): 909–13. https://doi.org/10.4103/jfmpc.jfmpc_91_19 39 62. Wherry, Laura R., Genevieve M. Kenney, and Benjamin D. Sommers. 2016. “The Role of Public Health Insurance in Reducing Child Poverty.” Academic Pediatrics 16 (3, Supplement): S98–S104. https://doi.org/10.1016/j.acap.2015.12.011 63. WHO (World Health Organization) and European Observatory on Health Systems and Policies. 2022. “Health Systems in Action: Armenia.” WHO European Center for Health Policy, Brussels. 64. World Bank. 2023. Armenia Public Expenditure Review: Improving Spending Efficiency. October. Washington, DC: World Bank. 65. World Bank. 2024a. “Armenia. Primary Healthcare Assessment Report: Strengthening Primary Healthcare.” October. Washington, DC: World Bank. 66. World Bank. 2024b. Armenia Public Finance Review: Enhancing Fiscal Policy Efficiency. December. Washington, DC: World Bank. 67. Yu, Weiyu, Peggy Wanza, Emmah Kwoba, Thumbi Mwangi, Joseph Okotto-Okotto, Diogo Trajano Gomes da Silva, and James A. Wright. 2023. “Seasonal Household Variation in Harvested Rainwater Availability in Siaya County, Kenya: A Modelling Analysis.” Npj Clean Water 6 (1), 32. 40 Appendix A Concepts and Methodology | Data Sources | The Commitment to Equity Framework | Limitations | 41 A P P E N D I X A Data and Methodology estimate official poverty statistics in Armenia. A.1. Concepts and Methodology The ILCS was implemented by the Statistical Committee of the Republic of Armenia. The A fiscal incidence analysis (FIA) measures how ILCS 2021 was a 12-month survey covering 5,184 government taxes and social expenditures affect households. It has information on income, household welfare. The goal of such an analysis is to employment, consumption, health, and answer questions about the impact of taxes and social education, among others. expenditures on poverty and inequality, the progressivity of taxes and social expenditures, and the • Tax legislation and comprehensive households that are net payers or net receivers in the administrative data on direct and indirect taxes, fiscal system. By developing detailed and tax rates, tax collection, and government budget parametrized microsimulation models, the FIA execution data from the Ministry of Finance and creates a platform that may be used to assess the World Bank teams. distributional impacts of existing fiscal policies, but • Social protection data from the main social also to simulate the ex-ante distributional impact of programs on budget, beneficiaries, and potential policy reforms. These models may therefore allocation rules from the Ministry of Labor and be particularly useful for informing evidence-based Social Affairs and the World Bank social fiscal policy design and investigating equity protection team. considerations both before and after policy • Data from the Ministry of Economy and implementation. Sustainable Development and the World Bank energy team to calculate the value of subsidies for domestic electricity distribution and fuel. A.2. Data Sources • Data on education expenditures and student enrollment from the Ministry of Ministry of The FIA in Armenia relied on the following data Education, Science, Culture and Sport and the sources: World Bank education team. • The most recent nationally representative • Data from the Ministry of Health and the World household survey at the time of analysis, the 2021 Bank health team on health expenditure and Integrated Living Conditions Survey (ILCS), a health visits. multipurpose survey on household budgets and living standards used to 42 2022b). Other fiscal incidence methodologies, such A.3. Commitment to Equity as the Tax-Benefit Microsimulation Platform (CEQ) Framework (EUROMOD), cover fewer fiscal interventions; for instance, in-kind benefits are typically excluded. 30 The 2021 FIA of Armenia was developed using the The CEQ methodology has been implemented in Commitment to Equity (CEQ) methodology, which over 70 countries, which facilitates the production provides a systematized framework for of results that are internationally comparable. determining the impacts of the fiscal system on poverty and inequality. 29 A CEQ assessment is a The first step in the FIA is the allocation of the main rigorous and standardized fiscal incidence taxes and transfers to households in the survey. methodology that permits systematic analysis of The main modeling techniques in this report several fiscal interventions that affect household include direct identification, inference, imputation, welfare (direct taxes, indirect taxes, direct transfers, and simulation. These concepts are described in Box indirect subsidies, and in-kind benefits in health A.1. Typically, fiscal interventions, such as direct care and education) and simulates how fiscal taxes, indirect taxes, and indirect subsidies, are systems work in practice. CEQ uses a common modeled based on simulation, combining tax rules framework developed by the CEQ Institute and and the characteristics of employment and presented in the CEQ Handbook (Lustig 2022a, consumption patterns among individuals. Box A.1: Tax and transfer allocation methods The CEQ methodology allocates each tax and transfer by drawing from a set of methods described in detail in Chapter 3 of the CEQ Handbook 2022, namely: • Direct identification: The survey asks whether a tax (transfer) was paid (received), and the amount. • Inference: Whether a tax (transfer) was paid (received), or the amount, is determined based on administrative data on the taxes (transfers) and other relevant secondary information self-reported by households. • Imputation: Taxpayers (beneficiaries) are directly identified in the survey, but the amount of a tax (subsidy) paid (received) is imputed based on program rules. • Simulation: Both the identification of taxpayers (beneficiaries) and the amount of a tax (subsidy) paid (received) are estimated based on program rules. • Prediction: A regression model is applied to predict who receives benefits or pays taxes, and the amount. Source: Adapted from Lustig 2022a, 2022b 29 30 The Commitment to Equity project (CEQ) is led by Nora Lustig at Refer to EUROMOD (Tax-Benefit Microsimulation Platform) Tulane University, New Orleans. For more information, refer to (dashboard), Centre for Microsimulation and Policy Analysis, Lustig (2022a, 2022b). Institute for Social and Economic Research, University of Essex, Colchester, UK, https://www.microsimulation.ac.uk/euromod 43 The building block of fiscal incidence analysis is the (after government taxes and transfers). After all the construction of income concepts. Conceptually, for taxes and transfers are modeled, the CEQ each household, the analysis starts with market methodology calculates various income concepts for income or market income, plus pensions, as the pre- each household to assess how fiscal policy affects fiscal income (any private income that excludes any household income at various stages of redistribution. government intervention). 31 Then taxes are Once completed, the distributional impacts of any subtracted and transfers are added to obtain final tax or transfer decisions on the welfare of a country’s household income or household post-fiscal income households can be estimated (refer to Figure A.1). Figure A.1: CEQ income concepts in the pensions as deferred income scenario, Armenia − Direct Transfers + Indirect Family benefit, child Subsidies + In-Kind Transfers benefits, non-contributory Interest rate subsidy Education and health pension, other benefits agricultural loans Market Income plus Pensions Labor, capital, and other income, private Disposable transfers, self-consumption and imputed Consumable Income = Final Income value of own dwelling Income Consumption PLUS contributory pensions MINUS Contributions to Old Age Pensions + Direct Taxes − Indirect Taxes Personal income tax VAT, excises, custom (wages, passive taxes income) Source: Adapted from Lustig 2022a, 2022b Note: For CEQ Armenia, consumption was equated with disposable income and analyzed backwards (adding back direct taxes and deducting direct transfers) to reach at Net market income and Market income levels. For the rest of the interventions, the concept was applied as is. Building a fiscal incidence model based on the CEQ Given that disposable income already includes methodology requires understanding how a certain interventions of the fiscal system, the analysis country’s fiscal system works and allocating the can move backward to remove fiscal interventions taxes and social expenditure across households and until reaching pre-fiscal income, a measure of total individuals, using microdata from a representative household income from private sources only. The socioeconomic household survey. In countries in analysis also works forward from disposable income, which consumption is a more reliable measure of adding additional fiscal interventions until reaching welfare than income, the analysis starts by setting post-fiscal income (income once all taxes and disposable income equal to the consumption welfare transfers modeled here have been included). aggregate and adds or removes certain fiscal Figure A.1 displays the steps of the income interventions to calculate the other income concepts. construction process. 31 Pre-fiscal income could be equivalent to market income or market income plus pensions depending on the pension scenario chosen for the baseline. 44 In the case of Armenia, this FIA relies on the This analysis uses PGT as the main scenario as pensions as government transfers (PGT) scenario Armenia’s old-age pensions, though designed to in the CEQ framework. In this scenario, pensions be contributory (consistent with the PDI scenario), are treated as government transfers, that is, a tool low contribution levels relative to benefits make it for redistribution from one segment of society to function more like a transfer system, aligning it another. In the PGT scenario, the pre-fiscal income with the PGT scenario. Actual pension transfers are concept is market income. 32 The alternative to the funded by tax collection, and the Armenian pension PGT scenario is the pensions as deferred income system redistributes from tax-paying individuals to (PDI) scenario, in which pensions are treated as pension recipients. Nonetheless, treating pensions saved private income. In the PDI scenario, the pre- solely as a government transfer means that those fiscal income concept is market income, plus relying solely on a pension will appear to have zero pensions (Refer to the sensitivity analysis under the market income, thus overestimating the proportion PDI scenario in Appendix B). of the poor (Lustig 2022a, 2022b). In a CEQ assessment, two scenarios may be used to Different post fiscal income concepts are used to treat pensions. The baseline scenario treats calculate the impact of fiscal interventions on pensions as government transfers, a tool for poverty and inequality. In calculating the total redistribution from one segment of society to effect on inequality reduction, the post fiscal another by taxing individuals through pension income is the final income concept, that is, the total contributions and transferring benefits through effect is the difference in inequality levels between pension income. An alternative scenario treats market income and final income. However, in pensions as saved private income (as a result of calculating total poverty reduction, the post fiscal previous savings) spent later in the form of PDI. income is the consumable income concept, that is, Both scenarios present extremes, and reality is the total effect is the difference in the poverty rate always some combination of the two. between market income and consumable income. One reason for this is that in-kind health care and education benefits are allocated at the average government cost of provision, which provides little information about how the transfers improve household purchasing power or household monetary poverty status. Another reason is that the publicly financed portions of health care and education services consumed by households are generally not included in either the consumption aggregate or the consumption-based poverty line. 32 As a result, pension income is subtracted from disposable income along with the direct transfers to get to net market income, and pension contributions are added with the direct taxes to reach market income. 45 A.4. Limitations The CEQ assessment framework for FIA is a static accounting-based first approximation of a true and retrospective accounting exercise without counterfactual. 36 In practice, there are no standard behavioral, life-cycle, or general equilibrium errors calculated that would allow a statistical effects. Major challenges include the failure to assessment of the allocations made to individuals capture top-income households 33 and the exclusion and households. 37 Nonetheless, the CEQ assessment of some interventions from the model — such as the framework for FIA provides a standard corporate income tax (CIT), 34 public expenditure methodology (which also enables international on infrastructure, 35 defense, or debt interest comparisons) for estimates of the impact of fiscal payments — because of methodological limitations. policy on poverty, inequality, and social welfare This means the incidence results represent an more generally. 33 Household surveys also do not capture some low-income infrastructure (publicly provided water), and local roads. The households very well either, such as institutional populations methodology for including infrastructure expenditure in fiscal (prisons, old-age care facilities, youth care facilities) and incidence models is currently under development by the CEQ households without a domicile address (informal housing). Institute. 36 34 CIT is better modelled with tax administrative data. Meaning distributional (static) impacts, not causal impacts. Furthermore, data on CIT cannot help in the allocation of CIT 37 There are, however, CEQ Assessment procedures for assessing burdens to households in the standard CEQ FIA framework. the statistical significance of the estimated impact of a fiscal 35 Examples of the relevant infrastructure investments that are policy (or set of fiscal policies) on poverty, inequality, and other missing and that could affect households differently in the indicators. income distribution could be connectivity infrastructure, water 46 Appendix B 47 A P P E N D I X B Sensitivity Analysis: Pensions as Deferred Income In the CEQ methodology, pensions are treated in two The CEQ Armenia results indicate the increase in different ways: pensions as government transfers poverty if pensions are treated as deferred income. (PGT) and pensions as deferred income (PDI). The Three CEQs have been conducted in Armenia over analysis presented in the report uses the PGT the years. The first used the 2011 ILCS; the second approach for the baseline scenario, while this used the 2017 ILCS; and the third used the 2021 ILCS. appendix presents the sensitivity analysis of the In all three CEQs, fiscal policies increased poverty results based on the PDI approach. The PDI scenario rates in the comparison of market income, plus implies that (1) the contributions to private social pensions. The increase was 10.4 percentage points security schemes are considered mandatory savings (2011), 11.0 percentage points (2017), and 12.0 (of income) rather than as a tax on income and (2) percentage points (2021). Figure B.1 compares the pension income among pension recipients is treated impact of the fiscal system on poverty in 2017 and in as current market income, plus pensions. Hence, 2021 and shows a consistent pattern. Figure B.2 and public pensions in this analysis are treated as deferred Table B.1 supply additional information. income (embedded in the pre-fiscal income concept of market income, plus pensions). Figure B.1: Poverty impact, PDI approach, 2017 and 2021 Poverty headcount at national poverty llne 33.2 31.5 31.5 29.6 24.9 25.8 20.5 21.2 Market Income plus pensions Net Market Income Disposable Income Consumable Income 2017 2021 Source: World Bank calculations based on fiscal administrative data; the CEQ methodology described by Lustig 2022a, 2022b; data of the 2021 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan, Armenia Note: Market incomes, plus pensions, serve as the baseline for the figure 48 Figure B.2: Inequality impact, pensions as deferred income Inequality, measured by the Gini coefficient (per capita) 0.4 0.342 0.328 0.299 0.305 0.305 0.284 0.289 0.3 0.259 0.257 Gini coefficient 0.223 0.2 0.1 0.0 Market Income plus Net Market Income Disposable Income Consumable Income Final Income pensions Income concepts 2017 2021 Source: World Bank calculations based on fiscal administrative data; the CEQ methodology described by Lustig 2022a, 2022b; data of the 2017 and 2021 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan, Armenia Table B.1: Marginal contributions of each fiscal intervention to inequality and poverty (for disposable, consumable, and final income) Size (with Concentration Kakwani Redistributive Poverty respect to Coefficient Coefficient Effect Reduction Effect Original Income)1 Marginal Marginal Contribution Contribution (National Moderate Poverty Line) Disposable Income 0.8594 Direct transfer: Non-contributory pension 0.0136 -0.3582 0.6574 0.007 0.020 Direct transfer: Free rent apartment 0.0001 0.4532 -0.1540 0.000 0.000 Direct transfer: Compensation for privileges 0.0052 -0.4697 0.7689 0.000 0.003 Direct transfer: Family benefit 0.0150 -0.6235 0.9227 0.012 0.025 Direct transfer: One off child birth 0.0009 -0.2903 0.5895 0.001 0.001 Direct transfer: Child care 0.0035 -0.3012 0.6004 0.002 0.007 Direct transfer: Other child benefit 0.0008 -0.6870 0.9862 0.000 0.001 Direct transfer: Other benefits 0.0020 -0.7629 1.0621 0.001 0.002 Direct transfer: Stipendium 0.0004 -0.1007 0.3999 0.000 0.000 All direct transfers excl contributory pensions 0.0415 -0.4830 0.7822 0.026 0.057 All direct transfers incl contributory pensions 0.0415 -0.4830 0.7822 0.026 0.057 Direct taxes: Personal Income tax -0.1558 0.4508 0.1517 0.012 -0.092 49 Direct taxes: Passive Income tax -0.0009 0.6674 0.3682 0.000 -0.001 Social security contributions -0.0254 0.4548 0.1556 0.000 -0.023 All direct taxes -0.1567 0.4521 0.1529 0.013 -0.092 All contributions -0.0254 0.4548 0.1556 0.000 -0.023 All direct taxes and contributions -0.1821 0.4525 0.1533 0.017 -0.099 Consumable Income 0.7902 Direct transfer: Non-contributory pension 0.0136 -0.3582 0.6574 0.008 0.023 Direct transfer: Free rent apartment 0.0001 0.4532 -0.1540 0.000 0.000 Direct transfer: Compensation for privileges 0.0052 -0.4697 0.7689 0.000 0.005 Direct transfer: Family benefit 0.0150 -0.6235 0.9227 0.014 0.025 Direct transfer: One off child birth 0.0009 -0.2903 0.5895 0.001 0.001 Direct transfer: Child care 0.0035 -0.3012 0.6004 0.003 0.008 Direct transfer: Other child benefit 0.0008 -0.6870 0.9862 0.001 0.001 Direct transfer: Other benefits 0.0020 -0.7629 1.0621 0.001 0.002 Direct transfer: Stipendium 0.0004 -0.1007 0.3999 0.000 0.000 All direct transfers excl contributory pensions 0.0415 -0.4830 0.7822 0.029 0.060 Direct taxes: Personal Income tax -0.1558 0.4508 0.1517 0.014 -0.126 Direct taxes: Passive Income tax -0.0009 0.6674 0.3682 0.000 0.000 Social security contributions -0.0254 0.4548 0.1556 -0.001 -0.030 All direct taxes -0.1567 0.4521 0.1529 0.014 -0.127 Subsidy: Agricultural loans 0.0022 -0.0719 0.3711 0.000 0.002 All indirect subsidies 0.0022 -0.0719 0.3711 0.000 0.002 Indirect tax: Alcohol Excise -0.0031 0.2267 -0.0725 0.000 -0.003 Indirect tax: Tobacco Excise -0.0082 0.2003 -0.0988 -0.001 -0.010 Indirect tax: Fuel taxes direct -0.0076 0.3317 0.0325 0.001 -0.009 Indirect tax: VAT -0.0492 0.2584 -0.0408 0.000 -0.057 Indirect tax: Import taxes -0.0033 0.2801 -0.0191 0.000 -0.004 All indirect taxes -0.0714 0.2591 -0.0401 0.001 -0.076 All taxes -0.2282 0.3917 0.0925 0.013 -0.167 Final Income 0.9101 Direct transfer: Non-contributory pension 0.0136 -0.3582 0.6574 0.006 Direct transfer: Free rent apartment 0.0001 0.4532 -0.1540 0.000 Direct transfer: Compensation for privileges 0.0052 -0.4697 0.7689 0.000 Direct transfer: Family benefit 0.0150 -0.6235 0.9227 0.009 50 Direct transfer: One off child birth 0.0009 -0.2903 0.5895 0.000 Direct transfer: Child care 0.0035 -0.3012 0.6004 0.002 Direct transfer: Other child benefit 0.0008 -0.6870 0.9862 0.000 Direct transfer: Other benefits 0.0020 -0.7629 1.0621 0.001 Direct transfer: Stipendium 0.0004 -0.1007 0.3999 0.000 All direct transfers excl contributory pensions 0.0415 -0.4830 0.7822 0.020 Direct taxes: Personal Income tax -0.1558 0.4508 0.1517 0.013 Direct taxes: Passive Income tax -0.0009 0.6674 0.3682 0.000 Social security contributions -0.0254 0.4548 0.1556 -0.001 All direct taxes -0.1567 0.4521 0.1529 0.014 Subsidy: Agricultural loans 0.0022 -0.0719 0.3711 0.000 All indirect subsidies 0.0022 -0.0719 0.3711 0.000 Indirect tax: Alcohol Excise -0.0031 0.2267 -0.0725 0.000 Indirect tax: Tobacco Excise -0.0082 0.2003 -0.0988 -0.001 Indirect tax: Fuel taxes direct -0.0076 0.3317 0.0325 0.001 Indirect tax: VAT -0.0492 0.2584 -0.0408 0.001 Indirect tax: Import taxes -0.0033 0.2801 -0.0191 0.000 All indirect taxes -0.0714 0.2591 -0.0401 0.002 All taxes -0.2282 0.3917 0.0925 0.014 All taxes and contributions -0.2536 0.3980 0.0988 0.018 In-kind health benefits: Outpatient 0.0198 0.1193 0.1799 -0.002 In-kind health benefits: Inpatient 0.0269 0.1542 0.1450 -0.005 In-kind health benefits: Child birth 0.0091 -0.3169 0.6161 0.004 Net health transfers 0.0558 0.0653 0.2339 -0.003 In-kind education benefits: Pre- school 0.0089 -0.2229 0.5221 0.004 In-kind education benefits: Primary school 0.0209 -0.2915 0.5906 0.012 In-kind education benefits: Middle- school 0.0256 -0.2423 0.5415 0.011 In-kind education benefits: Secondary school 0.0118 -0.1487 0.4479 0.004 In-kind education benefits: Initial vocational 0.0004 -0.0789 0.3781 0.000 In-kind education benefits: Secondary vocational 0.0034 -0.0575 0.3566 0.001 In-kind education benefits: Tertiary 0.0059 0.0668 0.2324 -0.002 Net education transfers 0.0769 -0.2064 0.5056 0.031 Source: World Bank calculations based on fiscal administrative data; the CEQ methodology described by Lustig 2022a, 2022b; data of the 2021 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan, Armenia 51 Appendix C 52 A P P E N D I X C Current Health System and Expenditure in Armenia Armenia's health system has undergone significant provider, and copayments are often required, though development in recent years. It operates under a some outpatient services may be fully or partially decentralized structure. The Ministry of Health waived among socially vulnerable and special groups. serves as the central body overseeing publicly funded For example, pensioners receive a 30 percent health services. There is no universal health coverage. discount, and children ages under 7 can access However, certain groups have access to the Basic medicines free of charge. The government also Benefits Package (BBP). This subsidized health provides a separate package for additional care, program includes a range of essential health services including certain hospital services for groups, such as that focus on increasing access to health care among people with disabilities, children ages under 18, vulnerable groups (refer to Table C.1). To access the pregnant women, recipients of the Family Benefits BBP, individuals must register with a primary health Program, and military personnel. Table C.1: Basic Benefits Package: coverage Type of coverage Description Coverage, Coverage, % millions population Noncontributory National Health BBP primary health care services 3.01 100% coverage, Service provision for enrolled but 98% are participants (does not include registered National Health Service scheme coverage for noninsured populations) Noncontributory health BBP (inpatient coverage) for poor and 1.15 38% insurance schemes vulnerable groups; Social Package for civil servants and military personnel Private health insurance Voluntary private health insurance for 0.2 6.7% schemes corporations and individuals, including beneficiaries of the Social Package* Source: World Bank calculations based on Lavado, Hayrapetyan, and Kharazyan 2018; World Bank 2023 Note: * The social package covers approximately 120,000 public sector employees. The services covered include emergency services and most outpatient services, including annual medical checkups, treatment of injuries (100% reimbursement), and expensive diagnostic examinations (50% reimbursement). Most inpatient care is covered, including therapies for endovascular, oncological, and cardiovascular diseases, diagnostic examinations in hospitals, and so on. 53 Appendix D 54 A P P E N D I X D Survey of Approaches for Incorporating the Healthcare Benefits in Poverty Estimates Although health care is widely perceived as a key underlying barrier was the inability to basic need (refer, for example, to United Nations determine the needs of a particular family in 1948, Article 23), creating a valid measure of paying for health insurance (Korenman and poverty that incorporates the need for health Remler 2016). care has “bedeviled analysts since the 1970s” (Panel on Poverty and Family Assistance et al. Despite the substantial difficulties in developing 1995, 223; refer also to Ellwood and Summers 1986; estimates, health reforms in the United States, Moon 1993; Ruggles 1990). The official poverty particularly the Patient Protection and measure in the United States is determined by Affordable Care Act, made the inclusion of health comparing pretax income with a standard in poverty measurement possible. In response to threshold historically based on the cost of food. It long-standing criticisms that the official poverty does not account for noncash benefits, such as measure does not accurately represent the needed education and health care. Critics point out that resources or expenses of the most disadvantaged, the official income or resource measure fails to the US Census Bureau introduced the account for noncash government benefits, medical supplemental poverty measure (SPM) in 2011. The out-of-pocket (OOP) expenses, and work SPM income or resource measure is cash income, expenses. Numerous research efforts have been plus in-kind government benefits (such as food undertaken to devise a valid method to include stamps and housing subsidies), minus health benefits in poverty measurement. Inclusion nondiscretionary expenses (taxes, medical OOP is important because catastrophic health expenses, and work expenses) (refer to Table D.1) expenditures can significantly add to the financial (Wherry, Kenney, and Sommers 2016). The SPM hardships of vulnerable households, pushing them thresholds are based on a broad measure of below the poverty line and eroding their life necessary expenditures — food, clothing, shelter, savings. Because a household’s health care depends and utilities — and are based on recent annually on the health of household members, determining updated expenditure data. The SPM addresses the amount to include in a poverty threshold is many shortcomings of the official measure, which difficult. Including health in poverty measures did not include health in the estimation of poverty requires assigning a dollar value to health needs and health resources, such as health insurance provided by the government or employers. The 55 (Short 2011). 38 As measured by the SPM, health needs. Despite these shortcomings, health insurance benefits can affect poverty but only to treatment within poverty measurement gained the extent they affect medical OOP expenditures broad support among poverty scholars (Corbett on care or insurance expenditures. The SPM 1999, for instance). A lengthy process ensued to cannot detect if people are going without health revise official poverty measures (Blank 2008; insurance or need care to pay for other basic Ruggles 2008). 39 Table D.1: Deriving SPM unit resources SPM resources = money income from all sources PLUS MINUS • Housing subsidies • Federal individual income taxes • Low-Income Home Energy Assistance Program • State individual income taxes (LIHEAP) • National School Lunch Program (NSLP) • Payroll taxes • Supplemental Nutrition Assistance Program (SNAP) • Child support paid • Special Supplemental Nutrition Program for Women, • Medical out-of-pocket (MOOP) expenses Infants, and Children (WIC) • Refundable tax credits (such as Earned Income Tax • Work expenses (includes childcare expenses) Credits (EITC)) Source: Short 2013, http://www.census.gov/prod/2013pubs/p60-247.pdf Note: SPM = Supplemental Poverty Measure Based on the shortcomings of SPM, a health expenditures for care. To construct a HIPM, add the inclusive poverty measure (HIPM) was devised, full price (or actuarial value) of health insurance to which incorporates an explicit need for health care the threshold of the existing material (non-health) or insurance in the poverty threshold and counts poverty measure (to avoid double-counting, the health insurance benefits as resources available to threshold should not already include health meet that need. The HIPM methodology is an insurance or care paid by insurance). For a partly adaptation of the current SPM methodology that nonmarket system (such as Medicare in the United incorporates a value for health insurance in the States) or an entirely nonmarket system (such as the poverty threshold and resources (Creamer 2024). National Health Service in the United Kingdom), The HIPM builds on the SPM by adding health the full price added to the threshold includes insurance needs to the threshold, also adding health payments from all funders, including the insurance benefits received to resources, and government. For those without government- or modifying the SPM’s deduction of medical OOP 38 The SPM has been widely applied to estimate the impact of health Medicaid. They randomly assigned medical OOP spending under spending on poverty. Wherry, Kenney, and Sommers (2016) use the this counterfactual scenario based on the distribution of spending SPM to estimate the impact of public health insurance in reducing among matched controls. child poverty. Sommers and Oellerich (2013) also assess the 39 The panel recognized that a material poverty measure “does not poverty-reducing effect of Medicaid using the SPM. The authors explicitly acknowledge a basic necessity, namely, medical care modeled the counterfactual of the medical OOP costs and poverty that is just as important as food and housing . . . and devalues the status among individuals covered by Medicaid as of 2011 as if the benefits of having health insurance, except indirectly” (Panel on program did not exist. Their methodology relies on propensity Poverty and Family Assistance et al. 1995, 236). score matching to compare Medicaid enrollees with those without 56 employer-provided health benefits, add nothing to needs equal ensures that health insurance benefits resources. Such households will be designated poor do not meet non-health needs. For households with if they do not have sufficient resources to purchase health benefits that require OOP premium both health insurance and material needs. Add to payments, add to resources a net health insurance resources the value of any subsidies available to value, that is, the value of the insurance need, minus purchase health insurance (up to the plan value in required OOP premiums. This approach the threshold). For households with health benefits incorporates how payments to obtain insurance (from government or employers) that require no reduce resources for material needs, while still OOP premium contributions, such as Medicaid, ensuring that health benefits are not assumed to add to resources an amount that equals the health meet material needs (Korenman and Remler 2016) insurance need. Making the insurance resources and (refer to Table D.2). 40 Table D.2: Differences in SPM and HIPM methods SPM HIPM Poverty Thresholds Non-Health Component Based on recent food, clothing, shelter, utilities, telephone and internet expenses Benchmark health plan: • Private Insurance and Medicaid: Second-lowest cost Silver Plan Health Component None • Medicare: Average government contribution to Medicare plus prescription drug benefit (Medicare Advantage or Part D) Resources Non-Health Component Sum of cash and non-cash income minus taxes (plus tax credits) work expenses, child support paid • Adds net health benefits (value of benchmark plan minus premium MOOP) Subtracts premium, non-premium • Zero value for uninsured and Health Component and over-the-counter expenses from unsubsidized coverage non-medical resources • Subtracts capped non-premium MOOP • No deduction of over-the-counter- expenses Source: Creamer 2024 40 While estimating the validity of HIPM as a measure of poverty needs are the unsubsidized premium of the Basic Plan. Health (Korenman and Remler 2016), a pilot HIPM for the Massachusetts insurance resources include any subsidies to, direct payments health reform was constructed directly on the SPM. The for, or direct provision of health insurance by the government or construction of the HIPM was based on two conditions: first is employers. For those with employer-provided or government- the consideration of health care as a basic need, and, second, provided insurance, Health Insurance Resources = Basic Plan the health system is to offer universal health insurance (UHI) Premium – premium medical OOP (up to the amount required with premiums unrelated to health status (community rating) for the Basic Plan). and caps on medical care OOP expenditures. Health insurance 57 Building on this framework, Remler, Korenman, the Supplemental Nutrition Program for Women, and Hyson (2017) use the same HIPM approach to Infants, and Children), and refundable federal tax estimate the impacts of private and public health credits (the Earned Income Tax Credit and the insurance, non-health means-tested benefits, Child Tax Credit). Non-health resources include social insurance, and federal refundable tax private cash income and non-health public benefits. credits on health-inclusive poverty in the United For cost-sharing needs, such as deductibles, the States under the Affordable Care Act. The official OOP spending is subtracted from resources, but poverty measure in the United States, implemented subtractions are capped at the applicable limit. A by the Census Bureau, treats cash income as the only household is considered poor if its resources fall resource available to meet needs. The authors argue below its poverty threshold. that health, as a basic need similar to food, should be included in poverty measurement, provided that In 2019, the Economic Report of the President guaranteed issues and community rating presented long-term trends using the full-income regulations are in effect, which means that anyone poverty measure (FPM). The report analysis used an can purchase insurance, and the price of insurance absolute poverty measure, the full-income poverty does not depend on health status. The authors measure, which includes health benefits examine the impacts of health insurance and other (Burkhauser et al. 2019; CEA 2019). The full-income social benefits on poverty using an HIPM. The poverty measure threshold includes health measure includes the need for health insurance in insurance and care in three ways. First, full income determining the poverty threshold and counts in 1963 included employer-provided health health insurance benefits as resources available to insurance benefits, and the poverty threshold was meet health needs. For the HIPM, the authors value defined as the amount of full income. Second, the health insurance benefits from the government and measure threshold effectively includes any medical employers at the cost of the unsubsidized premium OOP expenditures on health care and insurance of the second-lowest-cost Silver Plan. 41 This amount (nonpremium medical OOP and premium medical is added to other resources. For households that OOP, respectively), because full income includes all received health insurance benefits and had to pay income, no matter how it is spent. Third, the OOP premiums, the authors deduct from their threshold is updated over time using the personal resources the minimum required OOP spending on consumption expenditure index, a measure of premiums to measure the net value of health inflation designed to adjust personal consumption insurance resources. Total household resources are expenditures. Personal consumption expenditures the sum of net health insurance resources and non- include care purchased by insurers for those they health resources. Health insurance benefits include insure. benefits from employer-sponsored insurance and from public health insurance (Medicare, Medicaid, In their paper, Remler and Korenman (2023) and Affordable Care Act premium subsidies). Non- consider the viability and validity of incorporating health public benefits include social insurance health benefits into absolute poverty measures (unemployment insurance, Social Security, workers’ and trends in absolute poverty. For incorporating compensation, and veterans’ benefits), means-tested health benefits, even at a point in time, the authors cash and in-kind transfers (Temporary Assistance illustrate the importance and meaning of two key for Needy Families; the Supplemental Nutrition recommendations of the 1995 Panel report on Assistance Program; housing and energy subsidies; poverty. First, health insurance benefits should not 41 Even though some people may receive more valuable private or public insurance, the authors do not allow the value of health insurance resources to exceed the value of the second- lowest-cost Silver Plan. 58 be treated as perfectly fungible. Second, definitions insurance value by dividing these dollar inflows by of resources and thresholds must be consistent the number of the uninsured. Data on health (recommendation 4.1, Panel on Poverty and Family insurance premiums and cost-sharing information Assistance et al. 1995, 206). Thus, the full value of are used. For the main estimate of the implicit health insurance benefits should not be added to insurance value of free care based on fund inflows, resources if the threshold need is conceptualized the authors first compile information on the funds and operationalized as a need only to pay OOP for going to hospitals that are intended for indigent care and insurance without a need for health care. They assume that these funds are spent on the insurance or the care it pays for. uninsured. The total flow estimate is the sum of the total in-flows of money for free care for the Remler, Korenman, and Hyson (2017) also examine uninsured in New York State = Federal funding for how accounting for the implicit insurance value of Health Resources and Services Administration free care changes health-inclusive poverty rates Funding for Federally Qualified Health Centers in and gaps and the estimated impacts of health New York + Charitable Funding for Federally benefits and policies on those rates and gaps in Qualified Health Centers in New York + New York. Their first approach focuses on the dollar Disproportionate Share Hospital Funding for inflows that explicitly fund uncompensated care for hospitals in New York for the uninsured (refer to the uninsured. Later, they calculate an implicit Table D.3 for an overview). Table D.3: Overview of poverty measures: official, supplemental and health inclusive Official poverty Supplemental poverty Health inclusive poverty measure measure (SPM) measure (HIPM) Needs threshold 3 times basic food 33rd percentile of spending 33rd percentile of spending needs in the 1960s, on food, shelter, clothing, on food, shelter, clothing, updated for inflation and utilities, plus other items and utilities, plus other items with the consumer price + cost of basic health index insurance Resources Pretax cash income After-tax cash income After-tax cash income + tax credits + tax credits + in-kind benefits (non-health + in-kind benefits (non-health insurance) insurance) + net health insurance benefits Subtractions − work and childcare − work and childcare from resources expenses: expenses: − OOP expenditures on care − capped OOP expenditures (nonpremium medical OOP) on care (nonpremium − OOP expenditures on medical OOP) insurance (premium medical OOP) Source: Remler 2019 59 Appendix E 60 A P P E N D I X E Inclusion of Health Benefits in Armenia Poverty Estimates It is critical to estimate the impact of health household’s resources. Besides health expenses, such spending on poverty as a baseline. This would as medical OOP costs (including health insurance enable the impact of the proposed health reforms to premiums, physician co-pays, and over-the-counter be simulated to compare the effects on poverty. As medications), other items include child support adopted in Barofsky and Younger (2019), household paid outside the household and work expenses, such disposable income is taken from the survey and the as childcare and the cost of commuting, tools, monetary value of the benefits provided by the uniforms, or licensing fees related to a person’s government for health is included among household employment. The definition could not be entirely resources. The modified income definition is as applied to the Armenian data because of limitations follows: in the household survey. The Armenia household survey (ILCS) does not provide information on = + (E.1) work-related expenses. The income definition applied in the Armenia case is therefore estimated where: using the following formula: = disposable income (income after direct taxes, plus government cash transfers) 42 = + − − − (E.2) = income modified (after including the SPM definition and Barofsky and Younger [2019] method) where: = public health (the monetized value of per capita = disposable income (income after direct taxes, public health benefits received by households) plus government cash transfers) 43 = income modified (after including the SPM The SPM methodology also includes other definition and Barofsky and Younger [2019] method) expenses besides the medical spending of the = public health (the monetized value of per capita household. As estimated in the supplemental public health benefits received by households) poverty measure (SPM) approach, this analysis = medical OOP (as reported in the survey) subtracts the household medical out-of-pocket = education OOP (OOP) spending (on premiums and care) from each = childcare expenses 42 43 Disposable income is the conventional concept of the amount Disposable income is the conventional concept of the amount of money remaining in the pockets of household members to of money remaining in the pockets of household members to purchase goods, give away, or save after the government takes purchase goods, give away, or save after the government takes some away in direct taxes and gives some back in direct some away in direct taxes and gives some back in direct transfers. transfers. 61 To determine which household is poor, the national For Group A: (E.4) poverty line is used. The national poverty line is estimated using the adult equivalent estimation of ∗ ∗ = + − ( + ) − − SPM income, which accounts for health, education, and childcare expenditures. The poor are defined as where: ∗ follows: = the income definition under UHI = disposable income (income after direct taxes, = _ < (E.3) plus government cash transfers) = public health (the monetized value of per capita where: public health benefits received by households) = the poverty line (based on food, clothing, ∗ = new medical OOP under UHI shelter, and other basic needs, but not health) = education OOP = childcare expenses To estimate the effects of universal health insurance (UHI) on poverty, households were For Group B: (E.5) divided into two groups. The benefits of UHI were estimated using the 2021 round of the ILCS, as ∗ = + + − − − follows: 44 where: ∗ Group A: Individuals who reported health = the income definition under UHI expenditures or spending in ILCS = disposable income (income after direct taxes, Group B: Individuals who did not report health plus government cash transfers) spending in ILCS = public health (the monetized value of per capita public health benefits received by households) The effect of the UHI on poverty is estimated = estimated medical benefits under UHI through the net effect of savings because of the (health care to be received under UHI) lower health OOP and the insurance premium = insurance premium payments. 45 So, the new income definition after = education OOP incorporating the impact of UHI will be as follows: = childcare expenses 44 Refer to ILCS (Integrated Living Conditions Survey, Armenia) the private insurance premium payments. However, the (anonymized microdata database), Statistical Committee of insurance premium differs according to the insurance company Armenia, Yerevan, Armenia, https://armstat.am/en/?nid=205 and the health risks assessed for each insured. So, for simplicity and lack of information, the difference is assumed to be 45 For those with private health insurance, the effect may be negligible relative to the UHI. positive if the UHI insurance premium payments are lower than 62 Appendix F Monetization and Inclusion of Health Benefits in the Baseline Commitment to Equity (CEQ) Analysis | Simulating Insurance Premium Payments and the Identification of Payers | Estimation of the Benefits | Measuring the Distributional Impact | 63 A P P E N D I X F Simulating the Distributional Impact of Health Benefits Estimating the impact of the proposed universal health insurance (UHI) reforms in Armenia involved several steps, including the following. F.1. Monetization and Inclusion of Health Benefits in the Baseline Commitment to Equity (CEQ) Analysis The health benefits provided to households are without the resources utilized on OOP. For the monetized and included in the baseline CEQ monetization of existing in-kind health benefits analysis to estimate the distributional impact of (received by individuals or households), per capita government health spending. As described above, benefits are added at the final income level to the CEQ methodology considers health and measure the cash value of health transfers under education as in-kind transfers and does not estimate government programs. The per capita health their impact on poverty. To estimate the impact of benefit is imputed based on Ministry of Finance UHI on poverty, the health benefits are included, as data, at inpatient health services (specialized a first step, in the poverty estimates on which the mother and childcare medical services), and Wherry, Kenney, and Sommers (2016) supplemental outpatient services (specialized dental services and poverty measure (SPM) method is adopted. To paramedical services). The identification of the estimate the impact of health spending on poverty beneficiaries in the ILCS is completed using the in Armenia, the expenditures considered include total monthly visits of the individual by type of medical out-of-pocket (OOP) costs (such as health health service (aggregated inpatient versus insurance premiums, physician co-pays, and over- outpatient) (refer to Figure F.1). Poverty is the-counter medications); education OOP measured at final income after accounting for the expenses; and child support paid outside the OOP and the impact of public health spending. The household (refer to Equation E.2). diagram explaining the approach under the CEQ framework based on the above method is presented It is assumed that reported consumption in the in Figure F.1. Based on the methodology adopted, household survey (the Integrated Living Conditions table 6-5 in the CEQ Handbook (Lustig 2022a, Survey [ILCS]) already accounts for (subtracted) 2022b) has been updated and presented as Table F.1. OOP. Disposable income is therefore income 64 Figure F.1: Simulating UHI benefits using the CEQ framework a. Household income before UHI b. Household income after UHI Disposable income = Consumption Disposable income = Consumption + PH - EOOP - CC - IP + PH - MOOP - EOOP - CC Reported health expenditures in ILCS Yes No - MOOP* + BMuhi = Modified Income baseline scenario = Modified Income UHI simulated scenario Note: The panels rely on the Barofsky and Younger (2019) SPM methodology Table F.1: CEQ income concept after incorporation of the UHI benefits Components of income Disposable income Income SPM before UHI Income SPM after UHI concepts Consumption + + + + + − − − − − − ∗ − + Source: Adapted from Lustig 2022a, 2022b Note: Disposable income = consumption in the Integrated Living Conditions Survey (ILCS). Adult equivalent consumption is used to measure poverty in Armenia. BMuhi = the estimated medical benefits under UHI (health care to be received under UHI). CC = childcare expenses. Consumption = the consumption aggregate. EOOP = education OOP. IP = insurance premium under UHI. MOOP = medical OOP. MOOP* = medical OOP under UHI. PH = public health benefits. 65 F.2. Simulating Insurance Premium Payments and the Identification of Payers Household survey data are used to identify • Military or individuals with family in active insurance premium payers and simulate the service payments. It is assumed that enrolled individuals, • Women with children under age 2 with varied sources of income, will make contributions in the form of insurance premiums. • Recipients of social benefits The annual amount of the insurance premium is AMD 164,400 (AMD 13,700 per month). The For nonexempt individuals, the insurance premium payment is linked to income. The income premium payers are divided into two categories: (1) reported in the ILCS is used to simulate the exempt and (2) nonexempt payers. Exempt and premium payment amounts. The schedule of the nonexempt individuals are identified in the ILCS insurance premium payment is shown in Table F.2. based on the criteria stipulated in the draft law. For the first year, all individuals, irrespective of their status or income, are considered exempt. According to the law, the following individuals are Based on the ILCS and the draft Law, the results on exempted from the payment of the premium: the number of insurance premium payers and the • Ages less than 18 years, persons without amount of the payments by year (from test year 1 to parents (ages 18–23 years), and students under the full launch of UHI) are illustrated in Figure F.2. age 26 The results indicate that government employees • Elderly people ages 63 or more and private sector individuals will be the biggest contributors to insurance premiums. • Disabled Table F.2: Payment of insurance premiums Hired workers Notaries, self-employed Agricultural workers % of insurance premium payment Less than twice the Less than 15 times the 40 minimum salary minimum salary Between 2 and 4 times Between 15 and 30 times 60 the minimum salary the minimum salary More than 4 times the More than 30 times the 100 minimum salary minimum salary If they have submitted an 50 annual declaration of income of natural persons Relatives of the above Relatives of the above Relatives of the above 80 Source: Draft Law on UHI, February 14, 2024 66 Figure F.2: Insurance premium payment and number of payers a. Insurance premium payment, AMD millions b. Insurance premium payers, population 100 000 1 000 000 90 000 900 000 80 000 800 000 70 000 700 000 AMD millions 60 000 600 000 Individuals 50 000 500 000 40 000 400 000 30 000 300 000 20 000 200 000 10 000 100 000 0 0 Test year 1 Test year 2 Test year 3 Full launch Test year 1 Test year 2 Test year 3 Full launch Year of implementation Year of implementation Government employess Private sector workers Government employess Private sector workers Self-employed, notaries Agricultural workers Agricultural Self-employed Relatives Relatives Source: World Bank estimates based on the draft law; data of the 2021 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan, Armenia, https://armstat.am/en/?nid=205 F.3. Estimation of the Benefits The UHI benefits for individuals and households are estimated as the difference between the actual spending and the expected spending on health care. The UHI benefits are received in the form of reduced household OOP spending on health care. However, in the ILCS, there are two broad categories of respondents on health care spending, as follows: • Group A: Individuals who reported health expenditures or spending in the ILCS • Group B: Individuals who did not report health spending in the ILCS For respondents falling into Group A, some were identified as having existing health insurance (from private sector providers). For these individuals, it was assumed that the net benefits are zero: Net benefits under UHI = benefits from existing private health insurance – benefits from proposed UHI (F.1) Because these individuals already benefit from private health insurance, it was assumed they would enroll in the UHI and that the UHI benefit would equate with the current benefit from private insurance. For the individuals without existing health insurance, the benefits are equivalent to the following: Bene�its = actual health spending under OOP (as reported in the survey) – expected OOP health spending if (F.2) enrolled in UHI 46 46 Expected spending is taken as the median spending value among those with insurance. 67 The actual spending reported in the survey laboratory tests, X-rays, and medication. included spending on medical services, hospitals, Medicines are purchased in the last 30 days (not at and medicines. This included medical services, hospitals or clinics). The results indicate that, on such as visits, medical assistance, medical average, uninsured individuals spend more on personnel, laboratory tests, X-rays, ambulance health care than insured individuals, and they services, and other medical services provided by experience more frequent spikes in their health family doctors, therapists, and ambulatory, spending (refer to Figure F.3). The average health polyclinic, or village health centers. Hospital expenditure is higher among the uninsured than services are payments to hospital cashiers, among the insured (per decile) (refer to Figure F.4). including medical staff (doctors, nurses, and so on), Figure F.3: Armenia health expenditures as reported in ILCS 2021 by type of spending Armenia 2021: Medical services (from a family doctor, Health expenditure in Hospital, Medical services and Medicines therapist, ambulatory, polyclinic or village health center): Hospital • Visits, Medical assistance. with Health Insurance Medical services • Medical personnel Medicines • Laboratory tests • X-rays • Ambulance services • Other medical services without Health Insurance Hospital (services received in hospitals): • Medical personnel • Laboratory tests, X-rays, medication Medicines (payments for): • Medicines in the last 30 days (Not in 0 500,000 1,000,000 1,500,000 hospitals) Monthly Expenditure (in Drams) Source: World Bank calculations based on fiscal administrative data; the CEQ methodology described by Lustig 2022a, 2022b; data of the 2021 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan, Armenia Figure F.4: Health expenditures, average by decile (AMD per month) 50,000 40,000 30,000 Insured Uninsured 20,000 10,000 0 0 2 4 6 8 10 Decile of disposable income Source: World Bank calculations based on fiscal administrative data; the CEQ methodology described by Lustig 2022a, 2022b; data of the 2021 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan, Armenia 68 Among Group B, average savings were estimated • Individuals ages 61 and above grouped as one by age, location, and sex based on the data on category reported spending by Group A. Group B consists of • Ages < 61 years individuals who neither reported any spending on health nor had any health insurance. The average For individuals who did not report any health savings (Benefits (Savings) = Actual spending − expenditures in the ILCS, the benefits were imputed Expected spending if enrolled in UHI) were using the smoothed distribution of the average estimated by age, location, and sex based on data on benefits by age and sex (adaptive bandwidth local the reported spending among Group A using linear regression based on Friedman’s [1984] super Winsorized distribution (that is, replacing smoother algorithm). The estimated smoothed percentile 90 if the value is higher than percentile distribution results, by age and sex, are presented in 90). The individuals were divided into the following Figure F.5. The smoothed average of these benefits two groups based on age, location, and sex: was then assigned, per age and sex distribution, to the nonrespondents of corresponding age and sex. Figure F.5: Smoothed average benefits by age and sex, individuals in Group B a. Males b. Females Source: World Bank calculations based on fiscal administrative data; the CEQ methodology described by Lustig 2022a, 2022b; data of the 2021 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan, Armenia The results indicate that, although the insurance demonstrated by the concentration of expenditures premium payments are modestly progressive (as a on health care from all sources: 5 percent of the share of total payment), the concentration of individuals account for half the expenditures, while benefits appears to be progressive until the ninth the bottom 50 account for only 2.4 percent of total decile only. The high concentration of benefits in health expenditures in the United States in 2014 top deciles, particularly in the 10th decile, may (Mitchell 2019). The results also indicate that the potentially be driven by the high-reporting on- relative burden of premiums is high on low-income health spending by these individuals and their groups. Because the insurance premium is fixed, the financial capacity to afford health care. The burden is lower among those in the top deciles, even literature suggests that uneven needs are with premium reductions for lower incomes. 69 The benefits or net savings are measured as the attributed to the unaffordability or the difference between actual spending and unavailability of health care. The concentration of spending after the UHI. A large number of the UHI benefits and insurance premium results are poor did not report OOP spending or reported a shown in Figure F.6. The incidence of UHI low amount. Therefore, the net savings or benefits benefits and insurance premiums (as a percentage are low among these households as well. Low of disposable income) shows the spike in top spending on OOP by poor households can be deciles (refer to Figure F.7). Figure F.6: The concentration of UHI benefits and insurance premiums, by decile a. Insurance premiums b. Benefits (savings + health care) 30 60 (% of total benefits by decile) 25 50 20 40 Pcrcent 15 30 10 20 5 10 0 0 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Per capita disposable income deciles Per capita disposable income deciles Test phase year 1 Test phase year 2 Test phase year 1 Test phase year 2 Test phase year 3 Full launch Test phase year 3 Full launch Source: World Bank calculations based on fiscal administrative data; the CEQ methodology described by Lustig 2022a, 2022b; data of the 2021 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan, Armenia Figure F.7: The incidence of universal health insurance benefits and insurance premiums a. Insurance premiums b. Benefits (savings + health care) 25 ( paymes as % of disposablwe income) ( paymes as % of disposablwe income) 20 10 15 10 5 5 0 0 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Per capita disposable income deciles Per capita disposable income deciles Test phase year 1 Test phase year 2 Test phase year 1 Test phase year 2 Test phase year 3 Full launch Test phase year 3 Full launch Source: World Bank calculations based on fiscal administrative data; the CEQ methodology described by Lustig 2022a, 2022b; data of the 2021 round, ILCS (Integrated Living Conditions Survey, Armenia) (anonymized microdata database), Statistical Committee of Armenia, Yerevan, Armenia 70 F.4. Measuring the Distributional Impact Based on the above methodology, poverty ex ante was estimated and compared with poverty ex post. Poverty ex post was estimated as disposable income, MINUS the change in expected OOP. For respondents who reported health expenditures in the survey, the change in expected OOP = Actual OOP − UHI OOP. For those who did not report health expenditures in the survey, change in expected OOP = Estimated OOP − UHI OOP. Positive change indicates improvement in household income, while negative change corresponds to deterioration of household income. 71 72