TECHNICAL SUPPORT FOR UNIVERSAL HEALTH COVERAGE IN ARMENIA MACROECONOMIC EFFECTS OF FINANCING UNIVERSAL HEALTH COVERAGE IN ARMENIA Hasan Dudu Adanna Chukwuma Armineh Manookian Anastas Aghazaryan Muhammad Zeshan © 2021 The World Bank Group, 1818 H Street NW, Washington, DC 20433. This report was prepared by World Bank staff 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. This report was originally published in English by the World Bank (Macroeconomic Effects of Financing Universal Health Coverage in Armenia) in 2021. Where there are discrepancies, the English version will prevail. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning any territory's legal status or the endorsement or acceptance of such boundaries. The material in this report is subject to copyright. The World Bank encourages the dissemination of its knowledge. Hence, this work may be reproduced, in whole or in part, for noncommercial purposes if full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. Design by Veronica Elena Gadea, GCSDE, The World Bank Group. MACROECONOMIC EFFECTS OF FINANCING UNIVERSAL HEALTH COVERAGE IN ARMENIA Hasan Dudu Adanna Chukwuma Armineh Manookian Anastas Aghazaryan Muhammad Zeshan M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A iv ABOUT THIS REPORT T his report, Macroeconomic Effects of Financing Universal Health Coverage in Armenia, is part of the World Bank’s technical support toward universal health coverage in Armenia, which includes advisory services and analytics aimed at facilitating the government’s efforts to expand access to high-quality health care. The report draws on computable general equilibrium modeling and an updated social accounting matrix to examine the potential impacts of different fiscal policy options for financing universal coverage on macroeconomic outcomes. The analysis was co-financed by Gavi, The Vaccine Alliance. Modeling the Actuarial costing Projecting impact of tax of a unified revenues from options on growth, benefits package alternative tax poverty, financial that meets and non-tax protection, health population sources and employment healthcare needs Informing policies to Support for increase public Modeling strategic plan for financing for allocations of primary health healthcare public financing care financing, in the benefits organization, package to and regulation maximize health  Support for Facilitating the Technical Reforms to Assessment of strategic plan alignment of support towards align public public financial for continuity service Universal Health financing for management in of care across delivery with Coverage in health with the health providers better health Armenia value sector Support for Assessment of regulating, strategic monitoring and Knowledge purchasing in paying providers exchanges on the health for better quality investing in sector Universal Health Coverage Convening Harvard-World policy and Study tours to Bank Global technical selected Flagship Course discussions on countries  on Health reform options Reform v TABLE OF CONTENTS About this Report iv Table of Contents v Acknowledgments vi About the Authors vii Acronyms ix Executive Summary 1 Chapter 1. Background and Rationale 4 1.1. Political and Economic Context 14 1.2. The Case for Universal Health Coverage (UHC) Reforms 17 1.3. Purpose of this Report 111 Chapter 2. Model and Data 15  2.1. Computable General Equilibrium (CGE) modeling 115  2.2. Construction Of The 2018 Armenian Social Accounting Matrix (SAM) 116  2.3. Descriptive Analysis of the 2018 Armenian SAM 117  2.4. The Mitigation, Adaptation, and New Technologies Applied General Equilibrium (MANAGE) Model 120 Chapter 3. Simulations and Results 22  3.1. Scenarios 22  3.2. Results 25 Chapter 4. Conclusions 50 Annex 1: The 2018 Armenian SAM Statistics 53 Annex 2: The MANAGE Model 55 Endnotes 57 M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A vi ACKNOWLEDGMENTS T his report was supervised by Sylvie Bossoutrot (Country Manager, Armenia), Tania Dmytraczenko (Practice Manager, Health, Nutrition, and Population Global Practice, Europe and Central Asia Region), and Andrew Burns (Global Lead, Macroeconomic Modeling, Macroeconomics, Trade, and Investment Global Practice). The analysis benefited from the close engagement of the Ministry of Health and American University of Armenia. We thank Artur Khachatryan and Nairuhi Jrbashyan for their support towards data collection for the assessment. The team appreciates the Ministry of Finance, Ministry of Health, and Ministry of Economy for participating in the discussion of preliminary findings of the analysis in February 2021, and in the technical workshop on computable general equilibrium modeling of fiscal policy options for health system reform. We also thank Hugo Alexander Rojas Romagosa and Lulit Mitik Beyene for their expert facilitation of the technical workshop. The team is grateful to Arvind Nair (Senior Economist, Macroeconomics, Trade, and Investment Global Practice), Evgenij Nadov (Senior Economist and Program Leader, Macroeconomics, Trade, and Investment Global Practice), Israel Osorio-Rodarte (Economist, Macroeconomics, Trade, and Investment Global Practice), and Owen K. Smith (Senior Economist, Health, Nutrition, and Population Global Practice) for their insightful feedback on initial drafts of the report. We acknowledge the excellent editorial and operational assistance from Marianna Koshkakaryan and Arpine Azaryan. All errors and omissions are the authors. vii ABOUT THE AUTHORS Hasan Dudu is a Senior Economist at the World Bank Group. He leads computable general equilibrium modeling in the macroeconomic modeling team of the Global Macroeconomics and Debt unit. He also supports World Bank teams and clients with the quantitative analysis of policy issues related to sustainable development, climate change, and macroeconomic stability. He worked as a Scientific Project Officer at the European Commission before joining the World Bank. Hasan holds a Doctor of Philosophy in Economics from the Middle East Technical University. Adanna Chukwuma is a Senior Health Specialist at the World Bank Group, where she leads the health team in Armenia, implementation support for a primary health care reform project in Romania, and the COVID-19 vaccine procurement and deployment team in Moldova. She has led and supported the design, implementation, and evaluation of reforms to improve access to high-quality health care, through service delivery organization, strategic purchasing, revenue mobilization, and demand generation, in Sri Lanka, Sierra Leone, India, Moldova, Tajikistan, the South Caucasus Countries, and Romania. Adanna obtained a medical degree from the University of Nigeria, a Master of Science in Global Health from the University of Oxford, and a Doctor of Science in Health Systems from Harvard University. Armineh Manookian is the World Bank Country Economist in Armenia, covering macroeconomic and fiscal issues, regular economic reports, and macroeconomic projections. She is engaged in macroeconomic policy dialogue with the client. Armineh joined Bank in 2017 and prior to that, worked for more than 10 years in the International Monetary Fund's Resident Representative office in Armenia as a Macroeconomist. Before moving to Armenia in 2005, Armineh worked with the Central Bank of Iran as a Senior Economist in the Research and Policy Department. She holds a Master of Public Administration in Economic Policy Management from Columbia University. Anastas Aghazaryan is the Head of the National Health Accounts Center at the National Health Institute of the Ministry of Health in the Republic of Armenia and a Consultant at the World Bank Group. He leads health expenditure data collection, analysis, and the development of the National Health Accounts of Armenia. He also leads statistical and analytical activities related to health economics, the global burden of disease, health financing, and Universal Health Coverage. Anastas is a Professor at the French University in Armenia and works with development partners on socioeconomic analysis and climate change, including the United Nations Development Programme, Asian Development Bank, United Nations Children’s Fund, M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A viii and World Health Organization. He has served as a Board Member of the State Council on Statistics in the Statistical Committee of the Republic of Armenia. Anastas holds a Doctor of Philosophy in Economics from the Yerevan State University. Muhammed Zeshan is a Postdoctoral Fellow at the Norwegian University of Science and Technology, Norway, and a Consultant at the World Bank Group. Prior to his current position, he worked at the Pakistan Institute of Development Economics as a Senior Research Fellow. Muhammad is also the main contributor to the input-output table for Pakistan in the Global Trade Analysis Project. He has research experience with government and non-government institutions and teaching experience with multiple universities. Muhammed holds a Doctor of Philosophy in Environmental Economics from Pukyong National University. ix ACRONYMS AMD Armenian Dram BBP Basic Benefits Package CDE Constant differences in elasticities CE Cross-entropy CGE Computable general equilibrium CIT Corporate income tax DALY Disability-adjusted life year ECA Europe and Central Asia GDP Gross Domestic Product GTAP Global Trade Analysis Project HPD Higher posterior density IOT Input-output table LMIC Low-and-middle-income country MANAGE Mitigation, Adaptation, and New Technologies Applied General Equilibrium MIC Middle-income-country MoF Ministry of Finance MoH Ministry of Health NCD Non-communicable disease OOP Out-of-pocket PIT Personal income tax SCRA Statistical Committee of the Republic of Armenia SAM Social Accounting Matrix UHC Universal Health Coverage UMI Upper-middle-income USD United States Dollars VAT Value-added tax EXECUTIVE SUMMARY Armenia has made significant progress in improving population health outcomes over the past two decades. In 2019, the life expectancy at birth was 76.5 years, rising from 68 years in 1990. However, non-communicable diseases (NCDs) now cause 75% of deaths and significant disability. Up to 9,834 lost years of life per 100,000 people from NCDs annually can be prevented through effective public policies, such as tobacco exposure control, and access to high-quality health care. However, essential health care for NCDs is underutilized in part due to the cost of access. In 2019, over 84% of total health expenditure in Armenia was paid by households, out-of-pocket, well above the average in Europe of 30%. While the government has committed to financing primary and emergency care, out-of-pocket payments are required for most outpatient medicines and inpatient care for majority of the population. The Armenia Transformation Strategy highlights the critical role of investing in healthy and safe citizens, for growth and poverty reduction by the year 2050. Armenia has also committed as a signatory to the Sustainable Development Goals, to making progress towards Universal Health Coverage (UHC). This commitment involves guaranteeing access to essential health care for all its citizens. Armenia will need to undertake critical health financing reforms, including increasing public financing for health through general revenue or compulsory contributions, to reduce financial barriers to accessing health care. The Ministry of Health (MoH) has developed a Concept Note for the Introduction for Universal Health Insurance that proposes to mobilize additional revenue through payroll taxes or higher budgetary allocations to the sector. However, the Ministry of Finance (MoF) has noted that revenue mobilization options should ideally demonstrate positive returns in terms of economic growth and employment. Therefore, at the request of the MoH, the World Bank has modeled the macroeconomic impacts of options to increase domestic resource mobilization to finance universal access to essential health services in the Basic Benefits Package. A computable general equilibrium 1 EXECUTIVE SUMMARY 2 model was calibrated to a 2018 Armenia social accounting matrix to analyse the impacts of taxation options, including corporate income, payroll, value-added tax, and excise taxes, on gross domestic product, employment growth, income equality, and household welfare. The analysis assumes that through UHC reforms that mobilize additional public spending, the Government would cover the cost of 95% of household needs for health care from 2021 to 2050, and that the increase in the demand for care will be supported by improvements in supply-side efficiency. The simulation results suggest that the productivity benefits of expanding coverage would compensate for its economic costs in the long run, eventually increasing the gross domestic product between 0.08 and 1.36%, depending on the fiscal measure used to finance these reforms. The reforms also increase the employment between 0.25 and 1.34%. Health sector output increases by at least 10%, regardless of fiscal measure. A higher increase in demand is precluded by supply-side constraints. In the long run, payroll taxes led to fall in household welfare generally, but more so among poor households; value-added tax was associated with welfare improvements in rich households, and welfare reductions in poor households; while corporate income tax increases were associated with welfare improvements generally, except for households in the richest decile. ES TABLE 1 • Summary Table in 2050 under different tax policy regimes, percentage change from Business-as-Usual DIRECT TAX ON CORPORATE VALUE NON-WAGE PAYROLL EXCISE ALL INCOME ADDED TAX HOUSEHOLD INCOME Gross domestic product 1.36 0.50 0.05 0.39 0.08 0.70 Total employment 1.34 1.29 0.25 0.85 0.86 1.00 Health sector output 11.53 12.02 10.28 10.73 11.41 11.10 Source: Authors The results suggest that increasing direct taxes is better than increasing indirect taxes as the former are less distortionary and cause smaller allocative inefficiencies. Secondly, the broader the tax base is, the higher the positive impact on gross domestic product. That is, when the burden of financing UHC is spread throughout the economy, such as through value-added tax increases for all commodities, this is relatively better for economic growth. Finally, a payroll tax that is paid by employers on behalf of employees appeared to significantly harm economic growth, total employment, and household welfare. These findings are consistent with experiences in other emerging economies seeking to raise public revenue to support social spending through broad-based consumption taxes. M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 3 UHC reforms and the fiscal policy measures used to fund them will create a trade-off between economic growth and income equality. In the model, funding UHC by increasing corporate income taxes, increased both gross domestic product and income inequality. On the other hand, increasing taxes on household non-wage income, deteriorated equality while yielding the highest growth effect. Indirect taxes performed worse in terms of growth and equality. Higher payroll tax, excise taxes, and VAT reduced gross domestic product and harmed equality, although payroll taxes had the most negative impacts. Adjusting all taxes appeared to be the second best for inequality after corporate income tax and had a non-negative growth effect. Hence, although financing UHC reforms with direct taxes is better than indirect taxes, spreading the tax burden as much as possible balances the trade-off between growth and equality better. To make progress towards UHC, guaranteeing access to high quality and essential health care, Armenia will need to go beyond tax policy changes. Additional revenue may also be mobilized through increasing priority for health in the national budget and health sector efficiency gains. Furthermore, successful UHC reforms often involve a suite of policy changes beyond mobilizing pre-paid revenue for health that should be considered in Armenia. These include pooling financial risk across social groups to address individual uncertainty in health spending; allocating pooled resources strategically through a competent, politically-independent third-party payer to providers, benefits, and payment mechanisms that facilitate quality and efficiency; and strengthening service delivery organization and governance, particularly at the primary health care level, to ensure the highest standards of care. CHAPTER 1. BACKGROUND AND RATIONALE 1.1. POLITICAL AND ECONOMIC CONTEXT The Republic of Armenia, a former Soviet Socialist Republic, is in the South Caucasus. The country is bordered in the North by Georgia, in the South by Iran, in the East by Azerbaijan, and in the West by Turkey. The country occupies a land mass of 29,743 km2 and is divided into ten provinces and Yerevan, the capital city. The country has experienced significant political transitions over the past 20 years. The first decade following the dissolution of the Soviet Union was characterized by fiscal constraints and socioeconomic polarization, with public protests over the economic challenges and the lack of political transparency. However, in 2018, the Velvet Revolution led to another political transition, and the ascendancy of Nikol Pashinyan, a member of the parliamentary opposition, as the prime minister. The new government has committed to the ambitious Armenia Transformation Strategy 2050, focused on poverty reduction and an increase in real wages, through the establishment of a green, knowledge-driven economy.1 Among the 16 development objectives under the Transformation Strategy, the government identifies the importance of investing in a healthy and safe citizenry as the primary focus for the health sector over the next three decades. 4 M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 5 FIGURE 1 • Per capita Gross Domestic Product (GDP) trends in Armenia, 2000-17 2000 (constant 2018, in thousand AMD) 1500 Per Capita GDP 1000 500 0 2000 2003 2006 2009 2012 2015 2018 Year Source: World Economic Outlook (2019) In 2018, Armenia also transitioned from being a lower-middle-income to an upper-middle- income (UMI) economy.2 Between 2000 and 2008, the annual per capita economic growth rates averaged at 12.6% (Figure 1). However, from 2009 to 2017, growth rates fell to 2.0% annually due to the global financial crisis in 2008 to 2009 and the Russian financial crisis in 2014 to 2015. By 2019, the gross national income per capita was United States Dollars (USD) 4,680 (Table 1).3 Poverty rates have also reduced over the past 20 years. Between 2001 and 2018, the proportion of the population living below the UMI poverty line of USD 5.50 fell from 81.0 to 42.5%.4 The COVID-19 pandemic has put these economic gains at risk. In 2020, the economy contracted by about 8%. Furthermore, the poverty headcount ratio could increase by up to 12.8 percentage points, relative to the UMI poverty line of USD 5.50.5 Pre-pandemic, the general government expenditure as a proportion of GDP in Armenia was below the UMI average, varying between 20 and 29% of GDP from 2000 to 2017.6 However, in response to the COVID-19 pandemic, the government increased the general government expenditure by an estimated 2.3% of GDP to finance policy measures to support vulnerable households and firms.7 In 2017, government revenue (excluding grants) as a proportion of GDP, at 22.5%, exceeded the UMI average of 16.3%. Taxes constitute almost 90% of general government revenue.8 Broad-based taxes on goods and services brought in 51% of tax revenue in 2019, above the 40% accruing from income, profits, and capital gains, or the 2% from property tax.9 C H A P T E R 1 . B A C KG R O U N D A N D R AT I O N A L E 6 TABLE 1 • Comparing Armenia and selected countries, indicators as of 2018-19 TOTAL TAXES PER CAPITA PUBLIC TAX REVENUE ON MOST SOLD LIFE GROSS SPENDING ON COUNTRY AS A SHARE CIGARETTES AS EXPECTANCY IN NATIONAL HEALTH PER OF GDP A % OF RETAIL YEARS INCOME IN CAPITA IN USD PRICES CURRENT USD Armenia 22.2 38.1 75 4,680 52 Belarus 13.6 50.9 74 6,290 251 Croatia 22.2 78.8 78 14,980 844 Estonia 21.0 79.4 78 23,260 1143 Georgia 19.6 71.2 74 4,780 123 Kazakhstan 11.8 52.4 73 8,820 168 Kyrgyzstan 17.7 48.6 71 1,240 37 Russia 10.9 57.7 73 11,260 362 Tajikistan N/A 42.3 71 1,030 16 Turkey 16.5 81.4 77 9,690 302 Turkmenistan N/A 32.4 68 6,740 83 Ukraine 19.2 74.7 72 3,370 109 Uzbekistan 13.1 44.7 72 1,800 31 Source: World Development Indicators; World Health Organization Tobacco Free Initiative Population and employment trends have implications for revenue mobilization in Armenia. Since 1990, the total Armenian population has reduced by 16%, primarily due to economic emigration and below-replacement fertility rates.10 Of the current population of 2.9 million, 36% live in rural communities, 33% in the capital, and the remainder in other urban areas. However, downward pressures on population growth have contributed to a fall in the urban share of the population from 69 to 64% since 1990.11 In 2018, the proportion of the population above 15 years that was employed in Armenia was 46%, below the average in UMI countries of 57%.12 The proportion of non-agricultural employment in the informal sector is also high at 33%.13 The population aged 65 years and above increased as a proportion of the total population, from 6% in 1990 to 11% in 2019, with an equivalent rise in the old-age dependency ratio.14 Strategies for sustainable revenue mobilization in Armenia to finance social sector reforms under the Transformation Strategy will need to account for secular trends in population growth and current employment patterns. M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 7 1.2. THE CASE FOR UNIVERSAL HEALTH COVERAGE (UHC) REFORMS Armenia has made significant progress in improving population health outcomes over the past two decades. In 2018, the life expectancy at birth was 75 years, equivalent to the average in UMI countries, rising from 68 years in 1990 (Table 1).15 This rise was driven by improvements in child, maternal and adult health. Between 2000 and 2019, the infant mortality rate decreased from 15.6 to 6.1 deaths per 1,000 live births, due to lower fertility and improvements in child survival.16,17 At the same time, the probability of dying between 15 and 60 years fell from 14.3 to 11.6%.18 Since 2000, maternal mortality per 100,000 live births has fallen from 52.5 to 33.3 deaths.19 Non-communicable diseases (NCDs), including cardiovascular diseases and cancers, now cause 75% of deaths and significant disability.20,21 The World Health Organization estimated that the annual economic losses from NCDs are Armenian Dram (AMD) 362.7 billion, equivalent to the 6.5% of GDP in 2017.22 Of this amount, AMD 55.6 billion is the annual public spending on health care, while lost productivity is estimated at AMD 294.9 billion. The burden of NCDs and the COVID-19 pandemic are linked. Since March 1, 2020, when Armenia reported the first case of COVID-19, the cumulative case incidence per million people has risen above the European average (Figure 2). Aging and NCDs predict severe COVID-19, providing a further rationale to reduce the burden of NCDs in Armenia.23 FIGURE 2 • Cumulative confirmed COVID-19 cases per million people in Armenia and selected comparators Estonia 80,000 Croatia Georgia Armenia 60,000 Ukraine 40,000 Belarus Russia Azerbaijan 20,000 Kazakhstan 0 Jan 31, 2020 Apr 30, 2020 Aug 8, 2020 Nov 16, 2020 Feb 24, 2021 Apr 27, 2021 Source: John Hopkins University. Note: The number of confirmed cases is lower than the number of actual cases; the main reason for that is limited testing. C H A P T E R 1 . B A C KG R O U N D A N D R AT I O N A L E 8 Access to high-quality health care is essential to the prevention, detection, and appropriate management of NCDs. Hence, it is unsurprising that underutilization and poor quality of essential care contributes to significant preventable complications and mortality, globally and in Armenia. A recent analysis revealed that 9,008 years of life per 100,000 people that were lost globally due to NCDs in 2017 could have been prevented through effective public policies, such as tobacco exposure control, and access to high-quality health care. In Armenia, the premature avertable mortality from NCDs in 2017 was 9,834 years of life per 100,000 people, exceeding the global average, and pointing to gaps in access to care.24 Armenia has committed politically to addressing this challenge. Having adopted the Sustainable Development Goals, the Republic of Armenia has committed to achieving UHC by ensuring “financial risk protection, access to quality essential healthcare services and access to safe, effective, quality, and affordable essential medicines and vaccines for all.” However, much progress remains to be made. The UHC health coverage index monitors national progress towards ensuring access to health care, measured as the mean across 14 health services, including for NCDs. In 2017, Armenia’s score on the UHC health coverage index was 69 (out of 100), below the average in Europe and Central Asia (ECA) of 75. For reproductive, maternal, and child health services, Armenia scored 67, with performance falling to 55 for services for NCDs.25 These findings on summary indexes of coverage are reflected in indicators of health care utilization. In 2018, the proportion of people who consulted a health care provider when ill was only 32.7%, varying from 29.1% in Yerevan to 39.4% in rural areas.26 There is a relatively high level of utilization of health services for maternal and child health, and lower use of essential care for NCDs. For example, in 2019, over 98.8% of eligible children received vaccinations against tuberculosis, while 97.2% were vaccinated against hepatitis B. Similarly, almost 100% of childbirths are attended by a skilled provider, an indicator that monitors access to maternal health care.27 In contrast, only 24% of people above 15 years have been screened for type 2 diabetes mellitus, while 43.5% have been screened for hypertension.28 These patterns of health care use may partially explain the improvements in maternal and child health as well as the growing burden of NCDs. The predominant drivers of underutilization of essential health care are self-management and the lack of finance to cover health care costs. In over 50% of cases where the respondent in household surveys chose to forgo skilled health care despite being sick, the condition was self-managed. However, in approximately one in five cases, a lack of finance was the reported reason for not consulting a health care provider.29 Proximity to health facilities is not a reported constraint to health care access in rural areas, Yerevan, or other urban areas. The reported adequacy of physical access to service inputs is consistent with Armenia’s score of 98 (out 100) on the UHC index of service capacity and access, which assesses among other things, hospital beds per capita, health professionals per capita, and the International Health Regulations core capacity index.30 M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 9 Challenges in health financing, including mobilizing, pooling, and purchasing health services in Armenia, may contribute to the underutilization of essential care. A review of global experience indicates that countries make progress towards universal access to essential care by raising funds predominantly through general revenue or compulsory contributions, pooling the financial risk across groups, and allocating pooled funds to providers in a manner consistent with improvements in access.31,32 In contrast, in 2019, over 84% of total health expenditure in Armenia was paid by households, out-of-pocket (OOP), well above the average in Europe of 30% (Figure 3).33 This is because public expenditure on health as a percentage of GDP in Armenia, at less than 1.5%, is one of the lowest in ECA (Table 1).34 With its UMI status, external assistance as a proportion of current health spending has fallen from a peak of 18.0% in 2001 to 1.2% in 2018, and is projected to fall further. FIGURE 3 • OOP expenditure versus public expenditure on health in Armenia and comparator countries in 2018 90 Armenia 80 Turkmenistan Azerbaijan 70 Tajikistan OOP as % of Current Health Expenditure 60 Uzbekistan Kyrgystan 50 Ukraine Georgia 40 Kazakhstan 30 20 Sweden 10 Germany R² = 0.7618 France 0 0 2 4 6 8 10 Domestic General Government Health Expenditure as % GDP Source: WHO Global Health Expenditure Database The high levels of OOP payments increase the financial risk facing Armenian households. This includes catastrophic and impoverishing health expenditures, in addition to foregone health care when the costs are prohibitively high. The most recent globally comparable data shows that in 2013, 16% of Armenian households allocated over 10% of total household consumption expenditure to health care, which is more than 200% of the average in ECA of 7%.35 Between 2010 and 2013, the average annual change in catastrophic health spending at the 10% level C H A P T E R 1 . B A C KG R O U N D A N D R AT I O N A L E 10 was 3.3%, the highest increase in the world.36 Furthermore, about 4.1% of Armenian households are pushed below the UMI poverty line annually due to health care expenditures.37 Financing health care predominantly through OOPs is inconsistent with global evidence on revenue raising and pooling for UHC. With respect to purchasing, or the allocation of health financing to service delivery, state funding for health care within the basic benefits package (BBP) significantly influences access to care. In Armenia, the government commits legally to covering access to primary health care and emergency services for 100% of the population. However, OOP payments are required for most outpatient medicines, including for NCDs, outside a few conditions of interest that are covered by the state, such as tuberculosis, cancers, and mental health disorders.38 In addition, retail medicine prices in Armenia are weakly regulated and, as a result, these prices are among the highest in the Commonwealth of Independent States. For about 30% of the population, including some state employees, low-income earners, and social groups of interest, the state covers hospital, and selected expensive diagnostic services.39 In 2020, the government increased the scope of groups prioritized for expanded coverage even further. Nevertheless, for majority of households, OOP payments remain the predominant means of financing access to care, driven by limited state coverage and the high cost of medicines. Higher coverage under the BBP, including the services included, prioritized groups, and funding levels, predict increased health care utilization in Armenia. A 2006 analysis showed that groups eligible for expanded benefits coverage had 36% higher rates of using outpatient care than other groups.40 Even among beneficiaries of expanded coverage under the BBP, there are anecdotal indications that the cost of excluded outpatient medicines and care introduces financial barriers to outpatient health care use. Further, some services in the package are reimbursed at levels that are below the cost of delivery, owing in part to the low public health financing, contributing to the demand for informal payments among providers.41 In 2018, the monthly health expenditure in the richest quintile per adult (AMD 21,784) was about 21 times the level in the poorest quintiles (AMD 1,055), which may be in part an indication of the extent to which the ability to pay, determines access to health care in Armenia.42 From the foregoing, the relatively low public financing for health care is a key constraint to achieving UHC in Armenia and addressing the high NCD burden with its attendant economic costs. While there are other countries in the region with high levels of OOP expenditure as a proportion of total health spending, Armenia is an outlier both regionally and globally, with rates comparable to those seen in fragile and conflict-affected states. To expand coverage of essential services under the BBP, including for essential outpatient medicines medicines, Armenia will need to undertake health financing reforms that mobilize additional domestic revenue for health and improve financial protection. M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 11 1.3. PURPOSE OF THIS REPORT The Ministry of Health (MoH) has developed the Concept Note for the Introduction of Universal Health Insurance and published a proposal that includes public revenue mobilization, pooling, and purchasing reforms to facilitate progress towards UHC. The proposed reforms include an increase in prepaid, pooled public expenditure for health, mobilized through payroll taxes or higher budgetary allocations to the sector; purchasing of essential services in a BBP that is uniform across the population and includes coverage for outpatient medicines and for NCD care; and establishing an accountable and independent third-party agency to undertake purchasing decisions for the state. An actuarial costing exercise is ongoing to support policy discussions on additional public financing needed for increased health care coverage. Significant public debate on the reform proposal within the MoH Concept Note has centred around revenue mobilization options to finance these UHC reforms. In 2019, the Ministry of Finance (MoF) developed a Strategy for the Implementation of Tax Reforms, with the objective of improving tax compliance and ensuring investment attractiveness, relative to the status quo (Box 1).43 These reforms broadly included a range of changes to reduce direct taxation and increase indirect taxes. In part reflective of the relatively low total taxes in Armenia, the proposal also included an increase in excise rates on tobacco and alcohol (Table 1). Hence, in contrast to the proposal in the MoH Concept Note, the MoF Strategy proposes reductions in payroll taxes and increases in taxation on consumption. C H A P T E R 1 . B A C KG R O U N D A N D R AT I O N A L E 12 BOX 1 • A short primer on recent developments in tax policy in Armenia Armenia consolidated its tax laws and regulations and adopted a unified Tax Code at the end of 2016, which became effective in January 2018. This was followed by several amendments to streamline the tax policy and reduce the tax burden on businesses. The amendments became effective in 2020 and were expected to shift the burden of taxation from direct to indirect taxes. The profit tax rate was reduced from 20 to 18%. A flat personal income tax rate of 23% was introduced, with the intention to lower it by one percentage point annually, to 20% by 2023. An annual increase in excise rates for the main tobacco products and alcoholic beverages was proposed to compensate for the loss of income taxes. Improvements in tax administration were also envisaged to compensate for these losses. The number of tax regimes was lowered from five to three regimes, including regular, turnover tax, and microentrepreneur regimes. The amendments also include generous tax exemptions for small businesses with less than AMD 24 million annual revenue and kept the threshold above which the state collects value-added tax (VAT) high at AMD 115 million annual revenue. Since 2010, the tax revenue as a percentage of GDP in Armenia has risen from 17.1 to 22.2% in 2019 (Table 1). While tax rates are comparable to those in peer countries, legislative loopholes, exemptions, and weak tax administration significantly reduce tax revenue. These losses have been assessed by the MoF to be about 7% of GDP. Hence, tax policy reforms in Armenia should prioritize addressing these loopholes and exemptions and strengthening administrative mechanisms for ensuring compliance. Unlike high-income countries, low-and-middle-income countries (LMICs) tend to document higher levels of tax exemption and evasion, relatively higher informal employment, and lower levels of employment. As noted above, Armenia has a relatively high levels of unemployment, informal employment, and aging. The trend in LMICs seeking to modernize their tax systems is to expand consumption taxes, that are less vulnerable to tax evasion and have a broader base, over increases in payroll taxes (Figure 4).44 Options include VAT and excise taxes on carbon, alcohol, and tobacco, which are administratively feasible and have positive population health impacts. The MoF has also noted that revenue mobilization options to support social spending, should ideally also demonstrate positive returns in terms of economic growth and total employment. M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 13 FIGURE 4 • How have LMICs raised tax revenue? Tax levels and composition The composition of taxes in richer countries di ers from that of poorer countries, with greater emphasis on broad-based consumption and excise taxes. 30 revenue, percent of GDP 25 20 15 10 5 0 1990-99 2010-16 1990-99 2010-16 1990-99 2010-16 1990-99 2010-16 High income Upper middle income Lower middle income Low income Trade Corporate income tax (CIT) Personal income tax Excise tax Consumption Source: International Monetary Fund Beyond introducing or expanding sector-specific taxes, domestic resource mobilization to support additional health spending can be supported through conductive macroeconomic conditions that increase general government revenue, a rise in the share of the health sector in the state budget, and improvements in the efficiency of spending in the health sector. Given the constraints that LMICs face in improving revenue collection, the main source of additional budgetary space for health tends to be the national budget. In this regard, advocacy for a higher share of the state budget to be allocated to health may be facilitated by addressing potential sources of efficiency.45 To support the ongoing dialogue on financing UHC, at the request of the MoH, the World Bank has modeled the macroeconomic impacts of options to increase domestic resource mobilization to finance universal access to a uniform BBP. This analysis complements a separate assessment of options for mobilizing additional revenue for health, including through reprioritization of the state budget and efficiency gains. In this report, a computable general equilibrium model is calibrated to a 2018 Armenia social accounting matrix to analyse the impacts of taxation options, including corporate income, payroll, VAT, and excise taxes, on GDP, employment, income equality, and household welfare. The rest of the report is structured as follows. Chapter 2 describes the model and data. Chapter 3 presents the scenarios and results. Chapter 4 concludes. CHAPTER 2. MODEL AND DATA  2.1. COMPUTABLE GENERAL EQUILIBRIUM (CGE) MODELING UHC reforms to facilitate universal access to a uniform BBP would affect the Armenian economy through direct and indirect channels, where the latter may be as important as the former. The economy has a complex structure that is also influenced by social and political constraints. CGE models can take into account for all these effects on the economy and explicitly consider the interactions between different economic agents, using a set of standard assumptions.46 CGE models rely on behavioural assumptions regarding how economic agents react to changes in the economy, including prices, income, and taxes, under well-defined constraints based on the availability of resources. The standard economic agents in a CGE model are households, production activities, the government and the rest of the world. More than one type can be introduced for each of economic agent. Generally, multiple types are introduced for households, defined based on socioeconomic status, and for production activities, defined based on the sectors. The economic agents in the CGE model are representative, in the sense that an agent is constituted by aggregating actual individual units. For example, the household type for the first income decile represents all households in the first income decile. Thus, it owns all the endowments, including labor and capital, of the individual households it represents; its consumption equals the sum of consumption of these households; it receives all the transfers they receive from the government and rest of the world; and it pays all the taxes they pay. 14 M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 15 This representative household behaves like an individual household in making decisions about supplying endowments or consuming goods. This is equivalent to assuming that the behaviour of a household in the model is a weighted average of all households it represents. For example, while the share of food consumption in total household consumption is different for each household in the first income decile, for the representative household, it is the average among households in the group. These assumptions generally make CGE models suitable for long-term analysis. CGE models do not incorporate the costs of short-term adjustments, such as price inflation or unemployment, that are incurred as economic agents adjust to new market signals. In exploring fiscal policy options, the model does not capture tax avoidance but models the effective tax rate. CGE models assume that agents adjust almost immediately without facing any frictions in the price formation for goods or factors of production such as labor or capital. This assessment uses CGE modeling to assess the potential impacts of using different fiscal policy options to finance the UHC reform on the Armenian economy.  2.2. CONSTRUCTION OF THE 2018 ARMENIAN SOCIAL ACCOUNTING MATRIX (SAM) To analyse the effects of policy changes on the economy, we draw on a snapshot of the economy in the form of an input-output table (IOT) to develop a SAM that accounts for all transactions in the economy as defined in the System of National Accounts. The latest IOT available from the Statistical Committee of the Republic of Armenia (SCRA) is for 2006 and reports data for 17 sectors. However, this IOT neither accounts for the changes in economic structure since 2006 nor includes the health care and basic pharmaceuticals sector. Hence, the 2006 IOT was updated. We estimated an updated IOT by using the data available from SCRA, including national accounts, household surveys, international trade, government accounts, and sector data. To develop an updated SAM, our starting point was the Global Trade Analysis Project’s (GTAP) SAM in the GTAP version 11 database, which underlies most applied global general equilibrium models. The GTAP 11 SAM relies on an IOT compiled based on 2002 data.47 While constructing the GTAP 11 SAM, the macroeconomic totals, including GDP, consumption, investment, government savings, and trade data, were updated to 2017.48 We estimated the 2018 Armenia SAM using a hybrid cross-entropy (CE) higher posterior density (HPD) algorithm.49,50 Starting with the GTAP 11 SAM, we updated the macroeconomic totals to 2018 to match the GDP, consumption, investment, government aggregates and trade aggregates (An. Table 1). We introduced the intermediate consumption of aggregate sectors reported by the SCRA as a constraint to the CE-HPD algorithm.51 We then used C H A P T E R 2 . M O D E L A N D DATA 16 household surveys to split the single household type in the GTAP 11 SAM to ten groups based on expenditure. Finally, we split the labor to four groups based on the skill level (skilled and unskilled) and formality (formal and informal). The intermediate consumption of individual sectors was estimated by the CE-HPD algorithm consistent with the rest of the SAM which matched the 2018 national accounts, trade, government and household data. Following the estimation of a balanced SAM, we undertook adjustments based on the data available from the SCRA. These adjustments ensured that OOP spending on health, public spending on health, the share of health care industry in total gross value added, tax revenues, and tax rates matched the observed data as much as possible. The final Armenian SAM was different from the GTAP 11 SAM with over 10% of the cell entries changing more than 10%.  2.3. DESCRIPTIVE ANALYSIS OF THE 2018 ARMENIAN SAM We aggregated the 2018 Armenian SAM to include 27 activities and commodities (An. Table 2). The primary sectors, food processing, beverages, and tobacco, and construction had the highest contributions to value added (Figure 5). Health services constituted approximately 4% of the economy. The primary sectors’ value added was mostly sourced from unskilled labor with higher shares of informal labor in agriculture and forestry. On the other hand, the capital value added formed a significant part of the construction sector. Health services were mostly labor intensive with high share of formal skilled labor. FIGURE 5 •Sectoral value-added and distribution across factors of production 1200 1000 billion AMD 800 600 400 200 0 Trade Finance Other Services Textiles Water Transport Hotels and Restaurants Road Transport Communications Insurance Public Administration Chemicals Basic Pharmeceuticals Water Construction Agriculture Forestry Food Processing Beverages and Tabacco Paper and Pulp Metals and MetalProducts Other Manufacturing Electricity Gas Distribution Education Air Transport Mining Health Services Capital Land Natural Resources Skilled Formal Skilled Informal Unskilled formal Unskilled Informal Source: Authors’ calculations from the 2018 Armenian SAM M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 17 In the 2018 Armenian SAM, the richest decile household’s share in consumption, at 27%, was significantly higher than other deciles. The consumption shares for households in other deciles increased steadily, from the poorest decile, at 4.5%, to 11% for households in the second richest decile. The composition of consumption was also significantly different for the richest households. Most of their consumption was on services while expenditure in other households was mostly on food. Expenditure on health services was also distributed unequally (Figure 6 and Figure 7). However, its share in household total expenditures was almost the same for all households indicating an inelastic demand. Households in the poorest three deciles had negative savings, amounting to almost -10% of total consumption for the poorest households. Savings were a significant share of expenditure in households in the richest three deciles. The distribution of household income across sources was as expected. Richer households received most of their income from skilled labor and capital, while poorer households relied on unskilled labor, remittances, and government transfers (An. Figure 1). FIGURE 6 • Household consumption patterns by income decile 1,600 1,400 1,200 1,000 billion AMD 800 600 400 200 - ) t) 4 9 6 8 3 5 2 7 st (200) es ile ile ile ile ile ile ile ile he or ec ec ec ec ec ec ec ec ic po (r D D D D D D D D 1( 10 ile ile ec ec D D Food Manufacturing Services Health Saving Source: Authors’ calculations from the 2018 Armenian SAM C H A P T E R 2 . M O D E L A N D DATA 18 FIGURE 7 • Household consumption patterns by income decile, as percentage of total consumption 100% 6.0 5.8 5.6 5.2 4.6 5.0 7.1 5.2 6.7 7.2 80% 60% 40% 20% 0% -20% Decile Decile Decile Decile Decile Decile Decile Decile Decile Decile 1 2 3 4 5 6 7 8 9 10 (poorest) (richest) Food Manufacturing Services Health Saving Source: Authors’ calculations from the 2018 Armenian SAM The government’s revenue in the 2018 Armenian SAM was mostly from VAT and payroll tax. Production taxes, that is taxes on production activities, was the next highest at 22% of government revenue. The share of CIT was relatively low, at 13%, while excise tax and tariffs were 7% and 5% of revenue, respectively. Hence, 37% of government revenue was from direct taxes and 63% was from indirect taxes (Figure 8). M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 19 FIGURE 8 • Composition of government revenue 1% 13% 29% Other Household Direct Tax CIT 23% Payroll Production 7% Tariffs Excise 5% 21% VAT Source: Authors’ calculations from the 2018 Armenian SAM  2.4. THE MITIGATION, ADAPTATION, AND NEW TECHNOLOGIES APPLIED GENERAL EQUILIBRIUM (MANAGE) MODEL The World Bank’s MANAGE model is a recursive-dynamic single-country CGE model originally designed for analysis focused on energy, emissions, and climate change. In addition to the standard features of a single-country CGE model, the MANAGE model includes a detailed energy specification that allows for capital, labor, and energy substitution in production, intra- fuel energy substitution across all demand agents, and a multi-output, multi-input production structure. MANAGE is a dynamic model, using the neo-classical growth specification. Labor growth is exogenous. The model tracks population by age based on United Nations projections. The labor force is calculated as the number of working-age people, that is aged 15 to 64 years, multiplied by the labor force participation rate for 2018. Labor supply was calculated as the total labor force multiplied by the employment rate. In the model, capital accumulation derives from savings and investment decisions. The model allows for a wide range of productivity assumptions that include autonomous improvements C H A P T E R 2 . M O D E L A N D DATA 20 in energy efficiency that can differ across agents and energy carriers. Household demand is modeled with a two-level utility nest where aggregate consumption such as food, energy, and manufactured goods enter the upper nest with a constant difference in elasticities (CDE) utility function. Demand for aggregate bundles are distributed to individual commodities with a constant elasticity of substitution utility function at the lower level. In the model, we assume that health services consumption is at the upper nest of the utility function with non-linear demand in household income and prices, i.e. a non-linear Engel curve. Finally, the model has a vintage structure for capital that allows for putty or semi-putty assumptions with sluggish mobility of installed capital. We describe the MANAGE model in detail in Annex 2. For this analysis, the MANAGE model was calibrated to replicate the base year of the 2018 Armenian SAM. As described above, the 2018 Armenian SAM was based on 2018 macroeconomic aggregates, an IOT compiled based on 2002 data, which was updated to 2018 to match the data from SCRA, and 2018 household surveys. Like any model, the CGE model has limitations. First, the model is a simplified representation of reality. For example, increasing taxes on income or commodities may cause tax evasion and erosion to increase. Also, increasing taxes affects tax compliance to the extent that substitution of informal with formal labor allow. However, such effects are not captured in the model. Further, the tax rates in the model are effective rates rather than statutory rates and hence an increase in a tax rate in the model includes both tax base expansion and rate increase simultaneously. That is, the same effects can be obtained by increasing the tax rate or expanding the tax base and the model cannot distinguish between them. Third, the model relies on assumptions that may not hold due to the social, economic, and political environment. Lastly, the model is based on economic theory and its stylized findings and does not aim to predict the future. Therefore, the results presented in the report aim to highlight the main impact channels, constraints, and synergies related to different policies. CHAPTER 3. SIMULATIONS AND RESULTS  3.1. SCENARIOS UHC reforms to introduce universal access to a uniform BBP will make health care more accessible and require additional funds from the budget. Improved health care access will increase the average labor productivity, life expectancy, and quality of life. To meet the surge in demand, health care services will need to scale up including by improving efficiency, such that supply-side constraints do not limit the benefits of UHC reforms. We assumed that under the UHC reforms the state would pay for 95% of household’s health care needs throughout the simulation period, starting in 2021 and ending in 2050. We introduced this coverage as a subsidy on health care. Hence, we include the ambitious assumption that post-reform households would pay for only 5% of the cost of health care, which corresponds to co-payments or other small contributions from the households. These payments from the budget would significantly increase the government’s overall spending. We then assumed that the government would increase other taxes in a budget-neutral way, that is keeping the budget deficit constant. The modeling scenarios were defined following discussions with the MoH on the proposal in the Concept Note, the MoF’s Tax Reform Strategy, and trends in LMICs seeking to modernize their tax systems. We considered impacts from raising (a) payroll tax; (b) CIT; (c) direct taxes on other or non-wage household income (which excludes payroll tax but includes all other direct taxes paid by households); (d) VAT on commodities; (e) excise taxes on beverages and tobacco; and (f) a proportionally-equal increase in all of the above taxes. 21 C H A P T E R 3 . S I M U L AT I O N S A N D R E S U LT S 22 To fix terms, we define payroll taxes as taxes paid by the production activities for employing labor and which are part of social contributions and income taxes paid by the employer on behalf of the employee. CIT is tax paid by firms on capital income. By construction, capital income is paid to the firms in the model. Firms pay corporate taxes, make savings and transfers to rest of the world, and redistribute the remaining income to the households according to their share in capital ownership. The tax increases were endogenously calculated by the model by keeping the government budget balance constant at the baseline levels. The analysis is not aimed at proposing optimal tax rates for the reform. By endogenizing the tax increases to pay for the cost of the UHC reforms, we accounted for the endogenous change in the demand for health services, with changes in the relative prices of other commodities or disposable income of households, as the tax rates changed. This allowed us to estimate the costs and benefits simultaneously. The endogenously calculated tax rates also captured pass-through as a function of demand elasticities, such as for VAT to households. The UHC reforms were assumed to lead to reductions in the frequency of illness and absenteeism in the working age population and a reduction in the mortality rates across all age groups.52 This assumption is consistent with a recent analysis, undertaken with World Bank support, in which an hypothetical increase in public spending on health by AMD 63.2 billion averted an additional 46,000 disability-adjusted life years (DALYs), where each DALY is a summary metric that accounts for disability and death. The 2018 working-age population in Armenia was estimated at 2 million people, of which 1.5 million were employed. The above estimate of averted DALYs imply a 3% increase in time spent at work among the working-age population. Hence, we introduced a 3% growth in the labor force exogenously to the model. The benefits of the UHC reforms for labor productivity were endogenously introduced to the model. We used a simple law of motion equation for that purpose: where is labor productivity at time t, Ht is private health spending, and is the elasticity of labor productivity to health spending, which is assumed to be 0.5. The equation linked health spending with labor productivity, assuming decreasing returns. That is, the higher the health spending, the lower the labor productivity gains per unit of health spending. Our analysis does not account for the additional health benefits in terms of DALYs due to the reduced consumption of beverages and tobacco, for excise taxes. M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 23 We performed a sensitivity analysis for the assumed value of and the main findings of the model do not change significantly when the value of does not deviate from the diminishing returns assumption. The sensitivity analysis suggests that as the value of gets smaller, the benefits disappear, and the costs of UHC surpass its benefits. However, the ordering of the impacts of different fiscal options to fund the UHC reforms and the main impact channels do not change. In our model, we account for health risks using a human capital approach in which the value of a DALY is equated to the present value of the loss in income associated with mortality. However, alternative approaches to valuing health gains may consider an individual’s marginal rate of substitution between money and the risk of mortality in each period, consistent with the value of a statistical life literature. Value of a statistical life estimates tend to exceed the present value of future income, as they reflect the value of living which would include more than the effects of health on productivity.53 Hence, in using a human capital approach, we aim to be conservative and underestimate willingness to pay for improved health in our model. To ensure supply-side constraints accommodate the increase in demand following UHC reforms, efficiency improvements are necessary. A recent data envelopment analysis estimated that there is potential for up to 8% improvement in efficiency of public spending in the health care sector.54 The data envelopment analysis examines the extent to which public spending on health in Armenia would decrease if the country achieved the health outcomes obtainable in the most efficient comparator counties. As the efficiency improvements will require investments over time, such as in pooled procurement and integrated health care, we assume that these benefits will be spread until 2030. We introduce a 0.7% annual improvement in the health sector’s total factor productivity that adds up to a 20% increase in 2050. In each scenario, we introduce the revenue mobilization option, defined in terms of the fiscal measure, and accounting for the benefits of the UHC reforms described above. We examine the economic impact of each revenue mobilization option, in terms of GDP, employment growth, and changes in household welfare. The analysis accounts for the economic harms due to tax increases and the compensatory benefits of UHC reforms. We also highlight the main impact mechanisms and the structural constraints that are underlying these results. An important caveat is that we do not account for potential re-prioritization of health in government budget and assume that the status quo in terms of budgetary allocations holds throughout the time period considered. That is, we ignore the fact that government can reallocate some of the budget from education, defence, public administration and other spending to cover at least some of the costs of the UHC. Introducing such measures would reduce the tax increases needed to finance UHC and reduce the distortionary impacts. As this is true for all the tax scenarios considered, the relative impact would still be the same and the direction of our main findings would not change. C H A P T E R 3 . S I M U L AT I O N S A N D R E S U LT S 24  3.2. RESULTS 3.2.1 OVERALL RESULTS The impact of UHC reforms on the GDP was significant under all scenarios, although the sign and size varied considerably (Figure 9). The initial impact was negative for all scenarios. The most negative impact was observed with the excise taxes, followed by the payroll tax, with over 4.8% and 3% declines in the GDP relative to the business-as-usual (BaU) scenario, respectively. The GDP remained below the baseline levels under an increased payroll tax even when the longer-term benefits of UHC reforms materialized. However, the negative impact of excise taxes reversed over time and almost disappeared by 2050. An economy-wide increase in VAT, (that is for all commodities), also kept the GDP below the baseline level throughout the simulation period, although its short-term impacts were less severe than payroll and excise taxes with an almost 1.8% decline. Increasing direct taxes on household income, that broadens the taxation of income to capital, self-employed, and informal income, had the most favourable impact on the GDP in the long run. The decline in GDP under increasing CIT was small and the least harmful option in the short-term. FIGURE 9• GDP by revenue mobilization scenario 2 1 % change from BaU 0 -1 -2 -3 -4 -5 -6 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 Other Household Income CIT Payroll VAT Excise All Taxes Source: Authors; Note: HH - household M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 25 While GDP fell through lower consumption under all scenarios, there were benefits through higher investments, except for CIT (Figure 10). Funds available for investment declined with the introduction of higher corporate taxes and the investment level fell. Investment contributed more under the excise tax scenario where the tax burden was focused on one commodity. In addition, a favourable trade account added positively to the GDP, particularly in the direct taxes on other household income and excise tax scenarios. The increase in trade surplus was mostly due to the productivity benefits of the UHC reforms which increased the competitiveness of Armenian exports and reduced the need for imports by boosting domestic production. On the other hand, the decrease in consumption had adverse implications for the GDP in all scenarios. As the government reduced the cost of health care and increased the price of other commodities by imposing new taxes, households reduced their consumption. Consumption was reduced the most under the excise tax scenario. FIGURE 10 • GDP components in 2050 by revenue mobilization scenario 10 8 % change from BaU 6 4 2 0 -2 -4 -6 Other CIT Payroll VAT Excise All Taxes Household Income Consumption Investment Exports Imports Source: Authors Mobilizing the additional revenue for UHC reforms required a sustained subsidy throughout the period of analysis under all scenarios (Figure 11). The volume of subsidy varied slightly across the scenarios. Between 2020 and 2050, the subsidy rises from AMD 270 billion to AMD 390 billion, depending on the scenario. The highest increase was witnessed in excise taxes where the subsidy grew by 42% while the CIT scenario showed the minimum rise of 36.5% by 2050. The initial burden of UHC reforms was compensated with the more productive labor, better health outcomes, and a sustained economic growth in the long run. In the short run, the cost of UHC is high and reaches up to 4.8% of the GDP. C H A P T E R 3 . S I M U L AT I O N S A N D R E S U LT S 26 FIGURE 11A • Cost of UHC by revenue mobilization scenario in billion AMD 410 390 370 Billion AMD Billion AMD 350 410 330 390 310 370 290 350 270 330 2020 2022 2024 2026 2028 2030 2032 2034 2036 2038 2040 2042 2044 2046 2048 2050 310 Other Household Income CIT Payroll 290 VAT Excise All Taxes 270 2020 2022 2024 2026 2028 2030 2032 2034 2036 2038 2040 2042 2044 2046 2048 2050 Source: Authors 5.0% 4.8%• Cost of UHC FIGURE 11B Other by revenue Household mobilization scenario Income CIT as percentage Payroll of GDP VAT Excise All Taxes 4.6% 5.0% 4.4% % of GDP 4.8% 4.2% 4.6% 4.0% 4.4% % of GDP 3.8% 4.2% 3.6% 4.0% 3.4% 2020 2030 2040 2050 3.8% Other Household Income CIT Payroll VAT Excise All Taxes 3.6% 3.4% 2020 2030 2040 2050 Other Household Income CIT Payroll VAT Excise All Taxes Source: Authors The impacts of all tax policy changes on the main production sectors were positive, with relatively better outcomes in agriculture and services. Higher labor productivity and better health outcomes from the health subsidy added positively to the production (Figure 12). The agriculture sector displayed a higher growth rate than all other sectors mostly indicating a M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 27 resource shift from the manufacturing and services sectors to the agriculture sector. Direct taxes on other household income and excise tax scenarios displayed better economic recovery than other scenarios. The agriculture and services sectors grew by 6.9% and 5.4% in the direct taxes on non-wage household income scenario and by 6.8% and 9.8% in the excise tax scenario. FIGURE 12 • Production in 2050 by revenue mobilization scenario 12 10 % change from BaU 8 6 4 2 0 Other CIT Payroll VAT Excise All Taxes Household Income Agriculture Manufacturing Services Source: Authors The UHC reforms increased total employment (Figure 13). The positive impact was highest for the direct taxes that did not distort the labor markets as much. In contrast, payroll tax increases reduced employment in the short and longer terms. The VAT and excise taxes had a negative impact in the short run until the benefits of UHC reforms were realized. The impact became positive after 2030s. The scenario in which all taxes were adjusted had small effects at the beginning but increased overall employment by as much as 1% by 2050. C H A P T E R 3 . S I M U L AT I O N S A N D R E S U LT S 28 FIGURE 13 • Employment in 2050 by revenue mobilization scenario 1.5 1 % change from BaU 0.5 0 -0.5 -1 -1.5 2040 2044 2030 2050 2046 2020 2048 2034 2024 2042 2036 2026 2038 2028 2032 2022 Other Household Income CIT Payroll VAT Excise All Taxes Source: Authors Financing UHC reforms boosted welfare both for richest and poorest households independent of the financing measure. The welfare increase was significantly higher for the poorest households if UHC was financed with a CIT and lowest if payroll taxes were the means of financing the reforms. Since the increases in welfare in 2020 occurred before the labor productivity benefits of UHC kicked in, they resulted from increased consumption of health care services. This demonstrates the importance of the consumption value of health spending even before the human capital investment benefits are realized. The welfare gains in the model results were wiped out by 2050 due to the changing income and prices. Poorest households were better off only under the CIT scenario, while richest households improved their welfare if other household income taxes or excise taxes increased. The loss of welfare by 2050 was mostly driven by the decline in household income caused by the effects of increased taxes. Direct taxes, which are less distortive, still increased the welfare of the households whose expenditure was affected less from an increase in those taxes. M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 29 FIGURE 14 • Household welfare gains at baseline and 2050 7 6 % difference from BaU 5 4 3 2 1 0 -1 -2 -3 Poorest Richest Poorest Richest 2020 2020 2050 2050 Other Household Income CIT Payroll VAT Excise All Taxes UHC and the fiscal policy measures used to fund it created a trade-off between economic growth and income equality. Figure 15 presents the trade-off between GDP growth and inequality measured as the deterioration in the Gini coefficient. Funding UHC by increasing the CIT was the only measure that increased GDP while reducing income inequality. On the other hand, direct taxes on other household income yielded the highest growth benefit but reduced income equality. However, as this tax is 1% of government tax revenue, the increase in the effective tax rate relied on tax base extension more than the increase in tax rates. Indirect taxes performed the worst in terms of growth or equality dimensions. Payroll tax reduced both GDP and equality and was the least preferable option. This finding is expected as payroll taxes push employment towards informal sector. Excise taxes on beverages and tobacco also had a negative impact on growth and equality but to a lesser degree. However, the model did not fully account for the health benefits of reduced beverage and tobacco consumption and may have underestimated the growth effect. The effect of VAT tax was better than excise tax in terms of growth, but almost the same in terms of income equality, as the VAT spread the burden to a larger tax base by a proportionally equal increase in the tax rate for all commodities and services. Adjusting all taxes appeared to be the second-best option for inequality and had a non-negative growth effect. Hence, although financing UHC reforms with direct taxes appeared to be better than indirect taxes, spreading the tax burden as much as possible balanced the trade-off between growth and equality better. C H A P T E R 3 . S I M U L AT I O N S A N D R E S U LT S 30 FIGURE 15 • Growth and equality impacts of different fiscal policy measures in 2050 1 Equality: % difference from baseline 0.5 CIT 0 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 -0.5 All -1 Excise Other Household Income -1.5 Payroll VAT -2 Growth: % Difference from baseline Source: Authors 3.2.2 SCENARIO 1: DIRECT TAXES ON OTHER (NON-WAGE) HOUSEHOLD INCOME The GDP decreased initially under this scenario. GDP declined by around 1% in 2020 and then increased by nearly 1% till 2050 (Figure 16). The major drivers of the changes in GDP growth were personal consumption, government consumption, and public investment. Reductions in private consumption were the key driver of the initial GDP losses. Private consumption reduced by 1.4% in 2020 and by 2.16% in 2050. In the last decade before 2050, moderate growth in government consumption, private investment, and public investment supported the positive GDP growth rate. The revenues collected from the new taxes increased government consumption and public investment, while economic recovery in the later years attracted more private investment. In addition, the better health outcomes and higher labor productivity originating from the health subsidy increased trade competitiveness and led to an increase in exports by about 6.9% in the long run. M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 31 FIGURE 16 • GDP components in direct taxes on non-wage household income scenario 8 7 6 % change from BaU 5 4 3 2 1 0 -1 -2 es n t n t s en t en s io io ic en rt -3 rt pt pt pr m po m po m um st um st nt Ex st Im ve ve ve ta ns ns In In ns In co co ic e co at bl e ic at iv at Pu bl iv Pr Pu P Pr D G 2020 2035 2050 Source: Authors Production increased in all the sectors. However, the agriculture sector grew at a higher rate than the manufacturing and services sectors (Figure 17). In 2020, agriculture, manufacturing, and services sectors grew by around 5.8%, 2.7% and 2.9%, respectively. This trend was similar throughout the period of analysis. The rising production was reinforced by higher labor productivity, a growing health sector, and more demand for unskilled labor. The health sector grew significantly throughout the period of analysis and attained a 24.5% growth rate in 2050. FIGURE 17 • Production in direct taxes on non-wage household income scenario 30 25 % change from BaU 20 15 10 5 0 Agriculture Manufacturing Health Other Services 2020 2035 2050 Source: Authors C H A P T E R 3 . S I M U L AT I O N S A N D R E S U LT S 32 Increasing direct tax on non-wage household income to finance UHC reforms boosted employment by more than 1.4% in the long run (Figure 18). Employment of skilled informal labor declined in the short-run but grew in the long-run. This is because the increased tax levels predominantly affected richer households, that spend a higher share of their income on services, where skilled informal labor is mostly employed. Employment of unskilled labor increased steadily throughout the simulation period. The agriculture sector is the main employer of unskilled labor, and the increasing demand for unskilled labor by the booming agriculture sector seemed to be the key factor for the rising employment of unskilled labor. The manufacturing and services sectors mainly hired skilled labor, and these sectors performed moderately well initially. As a result, skilled labor performed poorly at the beginning, but improved in the long run. FIGURE 18 • Employment in direct taxes on non-wage household income scenario 1.6 1.4 1.2 % change from BaU 1 0.8 0.6 0.4 0.2 0 -0.2 -0.4 Total Skilled Skilled Unskilled Unskilled Formal Informal Formal Informal 2020 2035 2050 Source: Authors The welfare levels of all the households increased initially, but more so for the rich households (Table 2). Rich households spend more on the health-related expenditures than the poor households and secured higher welfare levels from the UHC reforms. Besides, the new health subsidy would make health commodities relatively cheaper for the rich than the poor. The welfare level of the richest household increased by around 3.9% in 2020. The share of health expenditures to total income was the lowest for households in Decile 6. As a result, the welfare levels of the average household in this Decile were the lowest in 2020, and turned negative later, decreasing by 1.9% in 2050. M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 33 TABLE 2 • Household welfare in direct taxes on non-wage household income scenario, percentage change from BaU 2020 2035 2050 Decile 1 (poorest) 2.4 0.8 -0.9 Decile 2 3.1 1.6 -0.1 Decile 3 3.5 1.7 0.1 Decile 4 2.9 1.1 -0.5 Decile 5 2.5 0.8 -0.8 Decile 6 1.2 -0.3 -1.9 Decile 7 4.4 2.6 1.1 Decile 8 3.6 1.7 -0.1 Decile 9 4.7 2.5 0.5 Decile 10 (richest) 3.9 2.0 0.7 Source: Authors 3.2.3 SCENARIO 2: CIT Increasing CIT to finance UHC reforms led to a fall in GDP at the beginning, but it became stable in later years (Figure 19). A trade surplus favoured positive GDP growth, but other components of the GDP deteriorated significantly. Overall, the GDP reduced by 0.58% in 2020, but increased by 0.18% in 2050. A fall in government consumption, private investment, and public investment were the key elements driving the initial decline in GDP. By 2050, increasing government consumption, public investment, and improved trade surplus supported GDP growth marginally. Through new CIT, the government raised more tax revenues and increased her consumption and investment. C H A P T E R 3 . S I M U L AT I O N S A N D R E S U LT S 34 FIGURE 19 • GDP components in CIT Scenario 6 5 % change from BaU 4 3 2 1 0 -1 -2 -3 s s s t n t n e en t rt rt en io io en ic po po pt m pt m pr m st Ex Im um st um st nt ve ve ve ta ns ns In In In ns co co ic e co at bl ic e at iv Pu bl at Pr iv Pu P Pr D G 2020 2035 2050 Source: Authors Production in all the sectors positively contributed to the economy. However, production in the agriculture and services sectors grew more than the manufacturing sector (Figure 20). The value of production in the agriculture and services sectors increased by 7% and 3.2% in 2020 and by about 7.4% and 4.5% in 2050. The manufacturing sector showed a lower growth rate in later years indicating a shift of resources from manufacturing to agriculture and services sectors in the long run. The rising economic activity was supported by a growing health sector, which showed a remarkable growth rate of 13.3% in 2020 and 25.4% in 2050. M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 35 FIGURE 20 • Production in CIT Scenario 30 25 % change from BaU 20 15 10 5 0 Agriculture Manufacturing Health Other Services 2020 2035 2050 Source: Authors The impact of increased CIT on employment was similar to the scenario involving direct taxes on non-wage household income. Skilled labor in informal sector was affected negatively in the short run and benefited less in the long run. However, employment of all labor types increased in the long run. The underlying reason for this was the declining income of the richer households who received most of the capital income through firms. The decline in the richer household’s income reduced the demand for services where most skilled labor is employed (Figure 21). C H A P T E R 3 . S I M U L AT I O N S A N D R E S U LT S 36 FIGURE 21 • Employment in CIT Scenario 1.6 1.4 1.2 % change from BaU 1 0.8 0.6 0.4 0.2 0 -0.2 -0.4 Total Skilled Skilled Unskilled Unskilled Formal Informal Formal Informal 2020 2035 2050 Source: Authors All the households enjoyed welfare gains initially, but the welfare levels of the poor households increased more than the rich households (Table 3). The rich households attained higher shares of corporate income compared to the poor households. The imposition of new CIT affected the budget constraint of rich households more than the poor households. In 2050, the welfare level of the richest household reduced significantly by 1.9%, while the poorest household enjoyed an increase in welfare level by 0.7%. TABLE 3 • Household welfare in CIT Scenario, percentage change from BaU 2020 2035 2050 Decile 1 (poorest) 5.2 2.8 0.7 Decile 2 7.5 4.7 2.6 Decile 3 6.4 3.7 1.5 Decile 4 5.7 3.0 0.8 Decile 5 5.2 2.6 0.4 Decile 6 5.2 2.6 0.4 Decile 7 6.9 4.2 2.2 Decile 8 6.6 3.6 1.3 Decile 9 6.8 3.7 1.1 Decile 10 (richest) 2.2 -0.1 -1.9 Source: Authors M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 37 3.2.4 SCENARIO 3: PAYROLL TAX For the scenario in which payroll tax increases subsidised the cost of UHC reforms, this adversely affected economic growth initially, but it had a slightly favourable impact on the GDP later (Figure 22). The imposition of new taxes reduced private consumption throughout the period of analysis, it reduced by 3.04% in 2020 and by 0.34% in 2050. The falling private consumption did not allow the GDP to grow substantially. This effect was exacerbated by decreasing government consumption, public investment, and an unfavourable trade balance in 2020. However, all the indicators except private consumption became favourable in 2050 and contributed positively to the GDP. The major contributor was the rising private investment and the favourable trade balance. FIGURE 22 • GDP components in payroll tax scenario 6 % change from BaU 5 4 3 2 1 0 -1 -2 -3 -4 es t s s t n t n en en rt rt en io io ic po po m pt m pt m pr st st Ex Im um st um nt ve ve ve ta ns ns In In In ns co co ic e co at bl ic e at iv Pu bl at Pr iv Pu P Pr D G 2020 2035 2050 Source: Authors In this scenario, the UHC reforms had the most favourable impact on the manufacturing and services sectors throughout the period of analysis (Figure 23). However, the production in agriculture sector shrinked slightly initially. The agriculture sector is the biggest employer of the labor, and the new payroll taxes offered the labor in agriculture less incentive to work. Although most of the unskilled labor in agriculture belongs to informal sector, a significant part of total unskilled labor is employed by the formal sector in agriculture. Hence, the reduction in employment of formal unskilled labor in the agriculture adversely affected its production. C H A P T E R 3 . S I M U L AT I O N S A N D R E S U LT S 38 FIGURE 23 • Production in payroll tax scenario 25 20 % change from BaU 15 10 5 0 Agriculture Manufacturing Health Other Services -5 2020 2035 2050 Source: Authors An increase in payroll tax negatively affected the employment of all labor types in the short run. The losses were highest for formal labor types with a change of between -1.8 and -2.4% in 2020. The regressive effect of the tax on economy also reduced informal employment, albeit to a lesser degree. In the long run, informal employment recovered and caused total employment to increase. This implies a significant shift from formal employment to informal employment (Figure 24). M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 39 FIGURE 24 • Employment in payroll tax scenario, percentage change from BaU 1 0.5 % change from BaU 0 -0.5 -1 -1.5 -2 -2.5 -3 Total Skilled Skilled Unskilled Unskilled Formal Informal Formal Informal 2020 2035 2050 Source: Authors The welfare levels of all the households increased initially, but the impact was higher for the rich households than the poor households (Table 4). Poor households are mainly employed in the agriculture sector. The decreasing production of agricultural sector and higher payroll taxes both adversely affected their real income and consumption. Therefore, poor households enjoyed a smaller increase in welfare levels than the rich households initially. Due to the new taxes, most of the households showed decreasing welfare levels in 2050, and the impact was higher for the poor households than the rich households. TABLE 4 • Household welfare in payroll tax scenario, percentage change from BaU 2020 2035 2050 Decile 1 (poorest) 0.3 -1.0 -2.5 Decile 2 0.6 -0.5 -1.9 Decile 3 1.3 -0.1 -1.5 Decile 4 1.5 0.0 -1.5 Decile 5 1.0 -0.4 -1.8 Decile 6 0.0 -1.2 -2.6 Decile 7 2.2 0.9 -0.4 Decile 8 1.9 0.3 -1.3 Decile 9 3.2 1.2 -0.5 Decile 10 (richest) 1.4 0.0 -1.0 Source: Authors C H A P T E R 3 . S I M U L AT I O N S A N D R E S U LT S 40 3.2.5 SCENARIO 4: VAT Financing UHC reforms through an increase in VAT unfavourably affected GDP initially. However, GDP converged to a positive growth trajectory in the long run after materializing the effects of the health subsidy (Figure 25). At the beginning, most of the indicators negatively responded in this scenario, including private consumption, government consumption, public investment, and trade accounts. However, by 2050, all the indicators showed promising increases, except private consumption. Government consumption, private and public investments along with a favourable trade balance were the key drivers of GDP growth, while the deteriorating private consumption was the main element reducing GDP growth throughout the period of analysis. FIGURE 25 • GDP components in VAT scenario 3 % change from BaU 2 1 0 -1 -2 -3 -4 s es s t t n t n rt en rt en en io io ic po po pt m m pt pr m Ex Im st st um st um nt ve ve ve ta ns ns In In In ns co co e ic co at bl e ic at iv at Pu bl Pr iv Pu P Pr D G 2020 2035 2050 Source: Authors The agriculture sector showed the most promising growth, whereas the manufacturing and services sectors also grew, but moderately (Figure 25). The relatively lower tax burden on the agriculture compared to manufacturing and services sectors resulted in higher agricultural production. Overall, the agriculture contributed 13% in the total VAT, whereas manufacturing and services sectors contributed nearly 59% and 28%, respectively. In 2050, the production in agriculture sector rose by 5%, whereas it rose by 0.4% and 3.1% in manufacturing and services sectors, respectively. At the micro level, the health subsidy lifted the production in health sector enormously, with growth from 10.7% in 2020 to 22.9% in 2050. M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 41 FIGURE 26 • Production in VAT scenario 25 20 % change from BaU 15 10 5 0 Agriculture Manufacturing Health Other Services -5 2020 2035 2050 Source: Authors An economy-wide increase in the VAT rates adversely affected employment in the short-run, but increased employment in the long-run. The only exceptions were informal skilled labor who were mostly employed in manufacturing sector and which were the most affected from the tax change. Employment of all kinds of labor turned positive in 2035, except skilled informal labor which was affected negatively until 2035, but recovered in the longer term. Overall, employment was better off by 2050 (Figure 27). C H A P T E R 3 . S I M U L AT I O N S A N D R E S U LT S 42 FIGURE 27 • Employment in VAT scenario 1.5 1 % change from BaU 0.5 0 -0.5 -1 -1.5 Total Skilled Skilled Unskilled Unskilled Formal Informal Formal Informal 2020 2035 2050 Source: Authors The welfare levels of all the households were positive initially, with a higher impact for the richer households (Table 5). By 2050, the change in welfare levels for all the poor households became negative, but most of the rich households enjoyed increasing welfare levels. Although the health subsidy made the health commodities cheaper, the imposition of new VAT increased the cost of many other commodities. The real income and consumption of the poor households reduced more than the rich households, adversely affecting their welfare levels. TABLE 5 • Household welfare in VAT scenario, percentage change from BaU 2020 2035 2050 Decile 1 (poorest) 1.4 -0.1 -1.8 Decile 2 3.0 1.3 -0.4 Decile 3 2.7 0.9 -0.8 Decile 4 2.6 0.7 -1.0 Decile 5 2.3 0.5 -1.3 Decile 6 2.0 0.2 -1.5 Decile 7 4.1 2.2 0.6 Decile 8 4.3 2.0 0.1 Decile 9 4.8 2.4 0.3 Decile 10 (richest) 2.1 0.2 -1.1 Source: Authors M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 43 3.2.6 SCENARIO 5: EXCISE TAXES ON BEVERAGES AND TOBACCO Financing UHC reforms through excise taxes had a huge negative affect on the GDP initially, but the GDP performed relatively better in the long run (Figure 28). At the beginning, the decreasing private consumption, government consumption and public investment were the key elements for the GDP loss. Among them, the private consumption, declined the most. In the long run, all the elements showed positive growth rates, except for private consumption. The rising private investment and improved trade balance were the key elements for the improved GDP growth in later years. Although these factors pushed the GDP towards a better recovery, the reductions in private consumption did not allow the GDP to grow significantly. The new excise taxes permanently reduced the real incomes of the households and the private consumption declined. However, these findings do not account for the health benefits of reduced consumption of tobacco and beverages. Hence, the benefits are likely to be underestimated. FIGURE 28 • GDP components in excise tax scenario 10 % change from BaU 8 6 4 2 0 -2 -4 -6 -8 t es t n t n ts s en en en io rt io ic or tm po pt tm pt m pr p m st um s Ex Im s nt ve ve ve su ta ns In In n In ns co co ic e co at bl ic e at iv Pu bl at Pr iv Pu P Pr D G 2020 2035 2050 Source: Authors After introducing the health subsidy, higher labor productivity and better health conditions contributed positively to the production in all the sectors (Figure 29). The agriculture sector grew more than the manufacturing and services sectors initially. But, in the long run, the services sectors displayed the highest growth rate of 9.8%. At the micro level, the production in health sector significantly increased. It grew by 10% initially, increasing to 22.8% in the long run. C H A P T E R 3 . S I M U L AT I O N S A N D R E S U LT S 44 FIGURE 29 • Production in excise tax scenario 25 20 % change from BaU 15 10 5 0 Agriculture Manufacturing Health Other Services 2020 2035 2050 Source: Authors Increased excise taxes on beverages and tobacco had a strong negative impact on employment in the short run, but a positive impact in the long run as the benefits of UHC materialized and sectors adjusted their factor use. Skilled informal labor was affected the most in the short run. Note that almost 9% of skilled informal labor is employed in beverages and tobacco sector. Thus, an increase in the excise tax caused a significant layout of skilled informal workers. This drove wages down and caused other manufacturing sectors to substitute other factors of production with skilled informal labor in the long run. In the end, overall employment increased as UHC benefits kicked in (Figure 30). M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 45 FIGURE 30 • Employment in excise tax scenario, percentage change from BaU 1.5 1 % change from BaU 0.5 0 -0.5 -1 Total Skilled Skilled Unskilled Unskilled Formal Informal Formal Informal 2020 2035 2050 Source: Authors The welfare levels of all the households increased except the average households in Decile 6 (Table 6). Further, it increased more for rich households compared to the poor households. The share of beverages and tobacco consumption in total household income was higher for the poor households compared to the rich households. Hence, the burden of the new excise taxes fell more on the poor households as indicated by their welfare levels. The highest burden was faced by the households in Decile 6, who faced the welfare losses of 0.2% initially, that increased over time. The richest households enjoyed the highest rise in welfare levels, with an increase of 1.5% in the long run. TABLE 6 • Household welfare in excise tax scenario, percentage change from BaU 2020 2035 2050 Decile 1 (poorest) 0.8 0.2 -1.2 Decile 2 1.8 1.4 0.1 Decile 3 1.0 0.5 -0.6 Decile 4 1.4 0.8 -0.3 Decile 5 0.2 -0.1 -1.2 Decile 6 -0.2 -0.4 -1.5 Decile 7 1.3 1.0 0.1 Decile 8 2.1 1.5 0.3 Decile 9 2.6 1.8 0.4 Decile 10 (richest) 3.0 2.2 1.5 Source: Authors C H A P T E R 3 . S I M U L AT I O N S A N D R E S U LT S 46 3.2.7 SCENARIO 6: ALL TAXES In this scenario, the UHC reforms had an undesirable impact on the GDP initially. However, GDP grew moderately in the long run (Figure 31). Decreasing private consumption, government consumption, and public investment were responsible for the diminishing GDP. Among them, the deteriorating private consumption had a key role in reducing GDP. All the indicators turned positive by 2050, but the private consumption kept reducing over time. In 2050, the moderate growth in GDP was due to the trade balance. Although the trade balance showed a slight improvement at the beginning, it improved considerably by 2050. FIGURE 31 • GDP components in all taxes scenario 4.0 % change from BaU 3.0 2.0 1.0 0.0 -1.0 -2.0 -3.0 s es s t n t n t rt rt en en en io io ic po po pt m pt pr m m Ex Im st um st um st nt ve ve ve ta ns ns In In In ns co co e ic co at bl e ic at iv at Pu bl Pr iv Pu P Pr D G 2020 2035 2050 Source: Authors This scenario favoured the agriculture sector the most, compared to the manufacturing and services sectors (Figure 32). This effect was driven by lower agricultural taxes, better labor productivity, and improved health conditions. The subsidy boosted the growth of health sector overtime, which grew above 11.6% in 2020 and nearly 23.8% in 2050. The firms in the booming health sector demanded more labor (Figure 33). As a result, the demand for unskilled labor in informal sector grew higher compared to other categories of labor. The demand for unskilled labor in formal and informal sectors was more stable than other types of labor overtime. M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 47 FIGURE 32 • Production in all taxes scenario 25.0 20.0 % change from BaU 15.0 10.0 5.0 0.0 Agriculture Manufacturing Health Other Services -5.0 2020 2035 2050 Source: Authors Adjusting all taxes proportionally reduced employment slightly in the short run, but increased it in the long run. Employment of skilled informal labor was affected significantly initially, and it recovered very slowly in the long run. This was expected as skilled labor employment declined under all taxes and the effect was higher for informal skilled labor for most scenarios. Unskilled labor employment was not as affected as skilled labor and increased in the long term by more than 1%. As a result, under the increased taxes, the UHC reforms led to the creation of more informal jobs (Figure 33). C H A P T E R 3 . S I M U L AT I O N S A N D R E S U LT S 48 FIGURE 33 • Employment in all taxes scenario 1.5 1 % change from BaU 0.5 0 -0.5 -1 Total Skilled Skilled Unskilled Unskilled Formal Informal Formal Informal 2020 2035 2050 Source: Authors The welfare levels of all the households rose, but the rich households had higher welfare levels than the poor households (Table 7). Financing the health subsidy through all the taxes reduced the real income of the poor households more than the rich households, which had adverse implications for their consumption. Most of the poor households faced welfare losses, while most of the rich households enjoyed moderately rising welfare levels in the long run. TABLE 7 • Household welfare in all taxes scenario, percentage change from BaU 2020 2035 2050 Decile 1 (poorest) 2.7 0.9 -1.0 Decile 2 4.5 2.4 0.6 Decile 3 4.0 1.9 0.0 Decile 4 3.7 1.5 -0.4 Decile 5 3.3 1.2 -0.7 Decile 6 3.0 1.0 -0.9 Decile 7 5.1 2.9 1.1 Decile 8 5.0 2.6 0.5 Decile 9 5.5 2.8 0.6 Decile 10 (richest) 2.2 0.2 -1.3 Source: Authors CHAPTER 4. CONCLUSIONS The Government of Armenia has committed to the Armenia Transformation Strategy and has highlighted the critical role of investing in healthy and safe citizens for growth and poverty reduction by the year 2050. The country is embarking on the implementation of this ambitious strategy at a time when the COVID-19 pandemic has resulted in economic contraction, reduced fiscal space, and reversed the recent gains in poverty reduction. At the same time, the aging population, unemployment rates and relatively high informal employment levels may limit opportunities to raise revenue to support additional social spending through direct taxes. Despite these constraints, the need for UHC reforms to improve access to high-quality health care to improve population health in Armenia is clear. In 2017, Armenia lost 9,834 years of life per 100,000 people annually due to NCDs and the overall cost to the economy exceeded AMD 300 billion. High OOP spending on health care presents financial barriers to essential health care use. Hence, to improve the health of the population, increasing public financing for health in Armenia through general revenue or compulsory contributions to converge with public spending levels in MICs is critical. Where allocated strategically, pooled public financing for health can facilitate reductions in the high level of OOPs, improving essential health care use, population health, and productivity. Therefore, at the request of the MoH, this analysis modeled the macroeconomic impacts of options for revenue mobilization to finance universal access to essential health services, drawing on an estimation of the 2018 SAM and the World Bank’s MANAGE model. The limitations of this approach have been described in earlier sections. CGE models may not capture the costs of short-term adjustments and are at best, simplifications of reality. The simulation necessarily makes assumptions about the cost of the state subsidy for health care under the proposed UHC reforms, improvements in health and labor productivity that would follow these transfers, and efficiency gains in the health sector to adjust to the increased 49 CHAPTER 4. CONCLUSIONS 50 demand. The taxation scenarios considered provide indicative conclusions on the implications of raising revenue to finance UHC in Armenia. We find that by 2050, productivity improvements that accrue to expanding access to health care universally would compensate the costs of the reform. Long run increases in GDP ranged from 0.05% where the UHC reforms were financed via payroll tax to 1.36% for direct taxes on non-wage household income. Total employment also rose, with increases ranging from 0.25% for payroll tax to 1.34% for direct taxes on non-wage household income. Health sector demand increased by at least 10% in every scenario, limited from rising further by supply-side constraints. The long run changes in household welfare varied by tax and household income decile. In the long run, payroll taxes led to falls in household welfare generally, but more so among poor households; VAT was associated with welfare improvements in rich households, and welfare reductions in poor households; while CIT increases were associated with welfare improvements generally, except for households in the richest decile. The results suggest that increasing direct taxes would be less distortionary and introduce smaller allocative inefficiencies than indirect taxes. The broader the tax base is, the higher the positive impact on GDP. Thus, spreading the burden of financing UHC throughout the economy through a VAT increase for all commodities had relatively better implications for economic growth. In addition, payroll tax paid by employers on behalf of employees appeared to significantly harm economic growth, total employment, and household welfare. These findings are consistent experiences in other emerging economies seeking to raise public revenue to support social spending through broad-based consumption taxes.55 Indirect taxes also spread the burden more and handle the trade-off between growth and equality better. Nonetheless, the share of direct taxes on non-wage income in total tax revenues was very small in the base year implying a very low effective tax rate. The low effective tax rate reflects the narrow tax base and difficulties ensuring compliance in reporting non-wage income. Hence, increasing direct tax rate alone is unlikely to raise significant tax revenues in Armenia. It would also be important to expand the tax base through improvements in administration that increase the registration of non-wage income. This highlights the non-trivial effort required to increase tax revenue from non-wage income. To make progress towards UHC, Armenia must expand protection from the negative financial impact of paying for health care OOP. Hence, mobilizing additional domestic revenue is a necessary component of UHC reforms. Beyond changes in fiscal policy, the additional revenue may be raised through increasing priority for health in the national budget and health sector efficiency gains. However, domestic resource mobilization alone is insufficient to ensure progress towards UHC. Hence, countries that have significantly expanded access to high- quality care have often undertaken a suite of policy changes beyond mobilizing pre-paid revenue for health, that can inform the next generation of reforms in Armenia. These include M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 51 pooling financial risk across social groups to address individual uncertainty in health spending; allocating pooled resources strategically through a competent, politically independent third- party payer to providers, benefits, and payment mechanisms that facilitate quality and efficiency; and strengthening service delivery organization and governance, particularly at the primary health care level, to ensure the highest standard of care.56 A N N E X 1 : T H E 2 0 1 8 A R M E N I A N S A M S TAT I S T I C S 52 ANNEX 1: THE 2018 ARMENIAN SAM STATISTICS ANNEX TABLE 1 • Macro-SAM used for the estimation of the 2018 Armenian SAM Government Households Commodity Enterprises Investment Rest of the Activity Factors World Taxes Total Activity 8,505 8,505 Commodity 3,028 528 4,799 1,613 2,261 12,229 Enterprises 2,438 240 64 85 2,827 Factors 5,113 5,113 Government 133 1,398 127 1,658 Households 2,147 2,675 425 724 5,972 Investment 531 416 378 560 1,885 Taxes 365 582 110 341 1,398 Rest of the World 3,142 38 48 454 75 3,757 Total 8,505 12,229 2,827 5,113 1,658 5,972 1,885 1,398 3,757 ANNEX TABLE 2 • Dimensions of the 2018 Armenian SAM ACTIVITY AND COMMODITY FACTORS HOUSEHOLDS Agriculture Skilled Formal Labor Household Decile 1 Forestry Skilled Informal Labor Household Decile 2 Mining Unskilled formal Labor Household Decile 3 Food Processing Unskilled Informal Labor Household Decile 4 Beverages and Tobacco Capital Household Decile 5 Textiles Land Household Decile 6 Paper and Pulp Natural Resources Household Decile 7 Chemicals Household Decile 8 Basic Pharmaceuticals Household Decile 9 Metals and Metal Products Household Decile 10 (richest) Other Manufacturing M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 53 Electricity Gas Distribution Water Construction Trade Hotels and Restaurants Road Transport Water Transport Air Transport Communications Finance Insurance Other Services Public Administration Education Health Services ANNEX FIGURE 1 • Household income and distribution across factors of production 120% 100% 17% 15% 15% 15% 15% 14% 12% 15% 16% 8% 3% 80% 10% 8% 8% 7% 5% 6% 13% 13% 11% 19% 9% 7% 8% 19% 10% 9% 60% 11% 10% 10% 11% 24% 25% 21% 19% 19% 23% 40% 24% 20% 19% 62% 20% 36% 39% 41% 40% 43% 42% 46% 28% 31% 0% 4 6 9 ) 8 3 5 t) 2 7 st es ile ile ile ile ile ile ile ile he or ec ec ec ec ec ec ec ec ic po D D D D D D D D (r 1( 10 ile ile ec ec D D Capital Formal Labor Informal Labor Land and Natural Resources Government Transfers Remittances A N N E X 2 : T H E M A N AG E M O D E L 54 ANNEX 2: THE MANAGE MODEL The MANAGE model is a recursive dynamic CGE model. 57 Each year in a scenario is solved as a static equilibrium, with dynamic equations linking exogenous factors, such as employment growth and capital accumulation, across years with update equations for productivity factors. Each static equilibrium relies on a relatively standard set of equation specifications. Production is modeled using a series of nested constant-elasticity-of-substitution functions designed to capture the substitutions and complements across the different inputs, notably capital and labor, with a focus on energy. Energy is assumed to be a near-complement with capital in the short-run, but a substitute in the long-run. The model features a vintage capital structure that captures the semi-putty or putty relations across inputs with more elastic long- run behavior as compared to the short-run. The model also allows for both multi-input and multi-output production. The former, for example, would allow for electricity supply to be produced by multiple activities—thermal, hydro, solar, and other renewable forms of electricity production. The latter allows for a single activity to produce more than one product. For example, oil seed crushing produces both vegetable oils and oil cakes for feed, or corn-based ethanol production can produce both ethanol and distillers’ dry grains soluble that can be used as a feed substitute. Labor and capital income are largely allocated to households with pass-through accounts to enterprises. Government revenue is derived from both direct and indirect taxes. Household demand is modeled using the CDE demand function that is the standard utility function used in the GTAP model. A transition matrix approach is used to convert consumer goods to supplied goods that also relies on a nested CES approach. The transition matrix is largely diagonal in the current version with consumed commodities directly mapped to supplied commodities. Goods are evaluated at basic prices with tax wedges. The model incorporates trade and transport margins that add an additional wedge between basic prices and end-user prices. The trade and transport margins are differentiated across transport nodes, that is farm or factory gate to domestic markets and the border (for exports), and from port to end-user (for imports). Import demand is modeled using the ubiquitous Armington assumption, that is that goods with the same nomenclature are differentiated by region of origin. Market equilibrium for domestically produced goods sold domestically is assumed through market clearing prices. By default, the small country assumption is assumed for export and import prices and thus they are exogenous, that is export levels do not influence the price received by exporters and import demand does not influence import prices. The current version of the model assumes market clearing wages on the labor markets with the possibility of an upward sloping labor supply schedule and sluggish mobility of labor across sectors. M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 55 Closure of the capital markets depends on the nature of the simulation. In dynamic simulations, new capital, which is generated by recent investments, is allocated across sectors so as to equalize the rate of return across sectors. Old capital remains installed in its original sector unless the sector is in decline. A sector in decline is one in which potential supply, as measured by the capital/output ratio, exceeds ex-post demand. If a sector is in decline, it releases its installed capital using an upward sloping supply schedule and its ex-post return on capital is less than the economy-wide average. Old capital in expanding sectors earns the same rate of return as new capital. The dynamics of MANAGE are composed of three elements. Population and labor stock growth are exogenous—the latter is often equated to the growth of the working age population. The aggregate capital stock grows according to the overall level of saving (enterprises, households, public and foreign), but is also influenced by the investment price index and the rate of depreciation. The third component relies on productivity assumptions. 56 ENDNOTES 1 The Armenian Weekly. Armenia Transformation Strategy 2050. Retrieved (March 20, 2021) from https://armenianweekly.com/2020/09/23/ armenia-transformation-strategy-2050-briefly-explained/ 2 World Bank Data Team. New country classifications by income level: 2018- 2019. Retrieved (April 15, 2021) from. https://blogs.worldbank.org/opendata/ new-country-classifications-income-level-2018-2019 3 World Bank. (2020). World Development Indicators. Retrieved (March 20, 2021) from https://data.worldbank.org/indicator 4 World Bank. (2021). Poverty & Equity Data Portal. Retrieved (April 10, 2021) from . https://povertydata.worldbank.org/poverty/country/ARM 5 World Bank. (2020). Open Knowledge Repository. Europe and Central Asia Economic Update. Fall 2020: COVID-19 and Human Capital. Washington, DC: World Bank. © World Bank. Retrieved (February 25, 2021) from. https://open- knowledge.worldbank.org/handle/10986/34518 License: CC BY 3.0 IGO 6 IMF World Economic Outlook (2019) and IMF World Revenue Longitudinal Data (2019); JLN UHC DRM Report. 7 World Bank. 2020. Europe and Central Asia Economic Update, Fall 2020: COVID- 19 and Human Capital. Washington, DC: World Bank. © World Bank. 8 International Monetary Fund. Government Finance Statistics. Retrieved (March 25, 2021) from https://data.imf.org/?sk=89418059-d5c0-4330-8c41-dbc2d8f90f46 9 International Monetary Fund. Government Finance Statistics. 10 Chukwuma, Adanna; Gurazada, Srinivas; Jain, Manoj; Tsaturyan, Saro; Khcheyan, Makich. 2020. FinHealth Armenia: Reforming Public Financial Management to Improve Health Service Delivery. World Bank, Washington, DC. © World Bank. https://openknowledge.worldbank.org/handle/10986/34747 License: CC BY 3.0 IGO. 11 Chukwuma, A., Gurazada, S., Jain, M., Tsaturyan, S., Khcheyan, M. (2020). FinHealth Armenia: Reforming Public Financial Management to Improve Health Service Delivery. 12 World Health Organization. (2020). Global Health Expenditure Database. Retrieved (March 30, 2021) from https://apps.who.int/nha/database 13 World Bank. (2020). World Development Indicators. Retrieved (March 20, 2021) from https://data.worldbank.org/indicator/SP.POP.65UP.TO.ZS?end=1990&start=1960 14 World Bank. (2020). World Development Indicators. 15 World Health Organization. (2020). Global Health Expenditure Database. 16 Statistical Committee of RA. (2020). The Demographic Handbook of Armenia 2020. Retrieved (April 2, 2021) from https://www.SCRA.am/am/?nid=82&id=2347 17 National Institute of Health. Ministry of Health. (2018). Analysis of state and pri- vate expenditures related to mother and child healthcare services. Retrieved (March 5, 2021) from http://www.nih.am/am/reports/74/am M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 57 18 World Health Organization. (2020). Global Health Observatory. Retrieved (February 5, 2021) from https://www.who.int/data/gho/data/indicators/indica- tor-details/GHO/adult-mortality-rate-(probability-of-dying-between-15-and-60- years-per-1000-population) 19 Statistical Committee of RA. The Demographic Handbook of Armenia 2020. 20 Fraser, N., Chukwuma, A., Koshkakaryan, M., Yengibaryan, L., Hou, X., Wilkinson, T. D. (2021). Reforming the Basic Benefits Package in Armenia: Modeling Insights from the Health Interventions Prioritization Tool. World Bank: Washington. 21 Statistical Committee of RA. The Demographic Handbook of Armenia 2020. 22 Farrington, J., Kontsevaya, A., Fediaev, D. (2019). Prevention and control of non- communicable diseases in Armenia: the case for investment. World Health Organization. Retrieved (March 30, 2021) from. https://iogt. org/wp-content/ uploads/2019/05/WHO-NCD-case-for-investment_Armenia. pdf. Published (2019) 23 World Health Organization. (2021) Health Topics: Coronavirus. Retrieved (April 15, 2021) from https://www.who.int/health-topics/coronavirus#tab=tab_1 24 The Lancet Global Health. (2020). Supplement to: Martinez R., Lloyd-Sherlock P., Soliz P., et al. Trends in premature avertable mortality from non-communi- cable diseases for 195 countries and territories, 1990–2017: a population-based study. Lancet Glob Health. 8: e511–23. Retrieved (January 30, 2021) from https://www.thelancet.com/journals/langlo/article/PIIS2214-109X(20)30035-8/ fulltext#supplementaryMaterial 25 World Health Organization. Global Health Observatory data repository. Index of ser- vice coverage Data by country. (2020). Retrieved (February 20, 2021) from https:// apps.who.int/gho/data/view.main.INDEXOFESSENTIALSERVICECOVERAGEv 26 Statistical Committee of the Republic of Armenia. Social Snapshot and Poverty in Armenia. (2019). Retrieved (March 20, 2021) from https://armstat.am/ en/?nid=82&id=2217 27 National Statistical Service, Ministry of Health. (2017). Armenia Demographic and Health Survey 2015-16. Yerevan, Armenia. 28 Andreasyan, D., Bazarchyan, A., Manukyan, S., Muradyan, G., Torosyan, A., Chamanyan, A., Bidzyan, L., Zelveyan, P. (2016). Armenia Health System Performance Assessment. NIH/MOH 29 Statistical Committee of the Republic of Armenia. Social Snapshot and Poverty in Armenia. 30 World Health Organization. (2017). Tracking universal health coverage: 2017 Global Monitoring Report. Joint WHO/World Bank Group report. Retrieved (February 10, 2021) from https://www.who.int/healthinfo/universal_health_coverage/ report/2017/en/ 31 The World Bank. Universal Health Coverage Study Series (UNICO). (2017-2018). Retrieved (March 5, 2021) from https://www.worldbank.org/en/topic/health/ publication/universal-health-coverage-study-series 32 Chukwuma, A., Meessen, B., Lylozian, H., Gong, E., Ghazaryan, E. (2020). Strategic Purchasing for Better Health in Armenia. World Bank, Washington, DC. © World Bank. Retrieved (February 2, 2021) from https://openknowledge.worldbank.org/ handle/10986/34491 License: CC BY 3.0 IGO 58 33 World Health Organization. (2020) Global Health Expenditure Database. 2020. Retrieved (February 2, 2021) from https://apps.who.int/nha/database/ Regional_Averages/Index/en 34 World Health Organization. (2020). Global health expenditure database. Retrieved (April 3, 2021) from https://www.who.int/data/gho/data/themes/ topics/health-financing 35 World Bank. (2020). World Development Indicators. Retrieved (March 20, 2021) from https://data.worldbank.org/indicator/SH.UHC.OOPC.10.ZS 36 Wagstaff, A., Flores G., Hsu J., Smitz M.F., Chepynoga K., Buisman L.R., van Wilgenburg K., Eozenou P. (2018). Progress on catastrophic health spending in 133 countries: a retrospective observational study. The Lancet Global Health 6.2: e169-e179 37 World Health Organization. (2020). Global Health Expenditure Database. 38 Chukwuma, A. et al. Strategic Purchasing for Better Health in Armenia. 39 Fraser, N. et al. Reforming the Basic Benefits Package in Armenia: Modeling Insights from the Health Interventions Prioritization Tool. 40 Angel-Urdinola, F., Jain, Sh. (2006). Do Subsidized Health Programs in Armenia Increase Utilization among the Poor? Policy Research Working Paper; No. 4017. World Bank, Washington, DC. © World Bank. Retrieved (February 2, 2021) from https://openknowledge.worldbank.org/handle/10986/9277 License: CC BY 3.0 IGO 41 Fraser, N. et al. Reforming the Basic Benefits Package in Armenia: Modeling Insights from the Health Interventions Prioritization Tool. 42 National Statistical Service of the Republic of Armenia. World Bank. (2018). Armenia Integrated Living Conditions Survey (ILCS) 43 Ministry of Finance of the Republic of Armenia. (2019). Strategy for the Implementation of Tax Reforms in the Republic Armenia. Draft Report. 44 Coady, David. Creating fiscal space. International Monetary Fund. Finance and Development, December 2018, Vol. 55, No. 4. 45 Chukwuma, A. et al. Strategic Purchasing for Better Health in Armenia. 46 World Bank. 2020. Macroeconomic Impacts of Climate Change and Mitigation Policies in Vietnam. World Bank: Washington, DC. 47 Jensen, J., Tarr, D., and Shepotylo, O. (2007). GTAP Data Bases: I-O Table Submission. Retrieved (January 20, 2021) from. https://www.gtap.agecon.purdue. edu/databases/IO/table_display.asp?IO_ID=242 48 Aguiar, A., Chepeliev, M., Corong, E., McDougall, R., van der Mensbrugghe, D. (2019). The GTAP Data Base: Version 10. Journal of Global Economic Analysis, 4(1). 1-27. Retrieved (March 5, 2021) from https://www.jgea.org/ojs/index.php/ jgea/article/view/77 49 Sherman Robinson, Andrea Cattaneo & Moataz El-Said (2001) Updating and Estimating a Social Accounting Matrix Using Cross Entropy Methods, Economic Systems Research, 13:1, 47-64, DOI: 10.1080/09535310120026247 M ACRO ECO NOMI C EFFECTS OF FI N A N CI N G U N I VER SA L HEA LTH COVER AG E I N A RMEN I A 59 50 Heckelei, T., Mittelhammer, R. C., & Jansson, T. (2008). A Bayesian alternative to generalized cross entropy solutions for underdetermined econometric models. University of Bonn Institute for Food and Resource Economics Discussion Papers No. 1548-2016-132440. 51 Statistical Committee of the Republic of Armenia. (2020). Main Indicators of National Accounts. Retrieved (November 11, 2021) from https://www.armstat.am/ en/?nid=202 52 Fraser, N. et al. Reforming the Basic Benefits Package in Armenia: Modeling Insights from the Health Interventions Prioritization Tool. 53 Robinson, L. A., Hammitt, J. K., Chang, A. Y., Resch, S. (2017). Understanding and improving the one and three times GDP per capita cost-effectiveness thresh- olds. Health Policy and Planning. Volume 32, Issue 1. Pages 141–145, https://doi. org/10.1093/heapol/czw096 54 Maduko, F., Chukwuma, A., Minasyan, G., Manookian, A., Noel Miguel AS., Tandon, A. (2021). More Money for Health: Resource Mobilization for Universal Health Coverage in Armenia. World Bank: Washington. 55 Coady, D. Creating fiscal space. 56 Cotlear, D., Nagpal, S., Smith, O., Tandon, A., Cortez, R. (2015). Going universal: how 24 developing countries are implementing universal health coverage from the bottom up. World Bank Publications. 57 van der Mensbrugghe, D., 2017. MANAGE Model Version 2.0f documentation, Purdue University Global Trade Analysis Project Center (unpublished model doc- umentation available upon request). E N D N OT E S 60