Report No: KAZAKHSTAN PUBLIC FINANCE REVIEW June 2023 . © 2023 The World Bank 1818 H Street NW, Washington DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved This work is a product of the staff of The World Bank. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages the dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Attribution — Please cite the work as follows: “World Bank. 2023. Kazakhstan: strengthening public finance for inclusive and resilient growth. Public finance review. © World Bank.” Painting credit: Tolepbay, Yerbolat. Cover page: Pathway. 2020. Back cover: Road to the Temple. 2020. Courtesy of the Pygmalion Gallery, Astana. https://www.tolepbay.com/ All 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. KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH PUBLIC FINANCE REVIEW Abbreviations ALMPs Active labor market programs ASPR Agency for Strategic Planning and Reform ASPR Agency for Strategic Planning and Reform BEPS Base Erosion and Profit Shifting BoD Board of Directors BNS Bureau of National Statistics CEQ Commitment to Equity CIPA Committee for Internal Public Audit CPI Consumer Price Index CFCs Controlled foreign corporations CIT Corporate income tax DBK Development Bank of Kazakhstan DSGE Dynamic stochastic general equilibrium EPU Economic Planning Unit ETS Emission Trading System ECA Europe and Central Asia EPT Excess profit tax EBFs Extrabudgetary funds FSC Financial Stability Committee GFSM Government Fiscal Statistics Manual GHG Greenhouse gas GNI Gross national income HIES Household Income and Economy Survey IFIs Independent fiscal institutions ICILS International Computer and Information Literacy Study IMF International Monetary Fund IPSAS International Public Sector Accounting Standards ISSA International Social Security Association IDR Issuer Default Ratings JERP Joint Economic Research Program JSC Joint Stock Company KSF Kazakhstan Sustainability Fund KTG KazTransGas SAI Law on External Audit MTBF Medium-term budget framework MTDS Medium-term debt management strategy MTEF Medium-term expenditure framework MET Mineral extraction tax MENR Ministry of Ecology and Natural Resources MoES Ministry of Education and Science MoE Ministry of Enlightenment/Education MoF Ministry of Finance MLSP Ministry of Labor and Social Protection of Population 4 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH MoSHE Ministry of Science and Higher Education MNE Ministry of the National Economy DPD Debt Policy Department of the Ministry of National Economy SBD State Borrowing Department of the Ministry of Finance M&E Monitoring and evaluation NBK National Bank of Kazakhstan NEDB National Education Database NEP National Education Project NFRK National Fund of the Republic of Kazakhstan NDCs Nationally Determined Contributions NIS Nazarbayev Intellectual Schools NGOs Non-governmental organizations NPLs Non-performing loans OBS Open Budget Survey OECD Organization for Economic Co-operation and Development OBB Outcome-Based Budgeting PCF Per-capita financing PM&E Performance, monitoring, and evaluation PBB Performance-based budgeting PIT Personal income tax PLF Problem Loan Fund PISA Programme for International Student Assessment PEFA Public Expenditure and Financial Accountability PER Public Expenditure Review PFR Public Finance Review PFM Public financial management PPP Purchasing power parity QFAs Quasi-fiscal activities QFEs Quasi-fiscal entities RBC Republican Budget Commission SK Samruk-Kazyna SMEs Small and medium enterprises SWF Sovereign Wealth Fund SEZs Special Economic Zones SHIF State Health Insurance Fund SRC State Revenue Committee SSIF State Social Insurance Fund SOEs State-owned enterprises SNGs Subnational governments SCD Systematic Country Diagnostic TDTs Targeted development transfers TEA Tax expenditure analysis TVET Technical vocational education and training UAPF Unified Accumulative Pension Fund UNDP United Nations Development Programme VAT Value-added tax MFMod World Bank’s macro-fiscal model WHO World Health Organization YoY Year-on-year KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 5 Contents Acknowledgments 9 Introduction 10 1. Strengthening the Fiscal Framework for Inclusive and Resilient Growth 12 1.1. Fiscal Response amid Growing Macro Challenges 13 1.2 Assessment of the Fiscal Posture 16 1.2.1 Size of fiscal operations 16 1.2.2 Rigidity in government spending: modest but rising 19 1.2.3 The use of quasi-fiscal activities 20 1.2.4 Government spending and tax multipliers 21 1.2.5 Cyclicality of the fiscal position 23 1.2.6 Distributive effect of fiscal policies 25 1.3 The Rules Underpinning the Fiscal Framework 28 1.3.1 Improved but with too many targets 28 1.3.2 The new fiscal rules: intention versus implementation 30 1.3.3 Government debt management 33 1.4 Emerging challenges 36 1.4.1 Quasi-fiscal deficits 36 1.4.2 The green transition and climate change 39 1.5 Recommendations 42 Simplify the fiscal rules and strengthen their monitoring framework 42 Broaden the institutional coverage of the fiscal framework by including extra-budgetary units at the central and local levels and public corporations. 42 Strengthen risks monitoring and government debt management 43 2. Managing the Fiscal Footprint and Risk from Quasi-Fiscal Activities 44 2.1 The Prominence of Quasi-Fiscal Activities in Kazakhstan 45 2.2 Constellation of extra-budgetary funds and quasi-fiscal entities in Kazakhstan 46 2.2.1 Social sectors 47 2.2.2 Quasi-Fiscal Activities by development finance institutions and state-owned enterprises 50 2.2.3 Quasi-fiscal activities in the financial sector 51 2.3.2 Fiscal exposure from bank bailouts 52 2.4 Recommendations 54 6 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 3. Improving Domestic Revenue Mobilization 56 3.1 Government Revenue Pressure from Declining Oil Revenue 57 3.2 Tax Collection Compared to Other Countries 60 3.3 Most Tax Collections Are Showing Deficient Performances 62 3.4 Tax Policy Analysis 62 3.4.1 Corporate income tax 62 3.4.2 Special tax regimes for small business entities or presumptive regimes: 66 3.4.3 Tax incentives 67 3.4.4 Personal income tax (PIT) 69 3.4.5 Value-added tax (VAT) 71 3.4.6 Excise tax 74 3.5. Fiscal progressivity of the tax policy 77 3.6 Policy Recommendations 78 3.6.1 Policy recommendations for corporate income tax 78 3.6.2 Policy recommendations for PIT 80 3.6.3 Policy recommendations for VAT 82 3.6.4 Policy recommendations for excise 83 4. Public Spending on Education 84 4.1 Motivation 85 4.2 Overview of the Education System in Kazakhstan 86 4.2.1 Governance and financing of education service delivery 86 4.2.2 Education outcomes 89 4.3. Education Expenditure, Trends over Time, and International Comparisons 91 4.4. Equity in Education Expenditures 94 4.5 The efficiency of Education Financing 96 4.5.1 The efficiency of expenditure across regions 97 4.5.2 The efficiency of expenditure on teachers’ salary 98 4.5.3 The efficiency of expenditure on infrastructure 100 4.6 Reform Options 102 5. Public Spending on Social Protection 106 5.1 Motivation 107 5.2 Overview of Kazakhstan’s Social Protection System 108 5.3 Components of the SP Programs 109 5.3.1 Social assistance 109 5.3.2 Social insurance 112 5.3.3 Active labor market programs 114 5.4 Adequacy, Efficiency, and Effectiveness of the SP System 118 5.4.2 Social insurance 122 5.4.3 Active labor market programs 125 5.5 Areas for Improvement in the Social Protection System 126 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 7 6. Budgeting, Planning, Monitoring: Institutional Foundations for the Public Sector 128 6.1 Strengthening Performance across the Policy Cycle 129 6.2 Overview of Development Efforts, Progress, and Challenges 133 6.2.1 Budget composition and execution rates 134 6.2.2 Monitoring budget performance 135 6.2.3 Strengthening the focus on performance and accountability in public budgeting 138 6.3 Strengthening Linkages between Strategic Planning and Budgeting 139 6.4 Transparency of Public Finances 142 6.4.1 Independence in audit institutions and public sector accounting standards 142 6.4.2 Open-Budget and citizen engagement 143 6.5 Mainstreaming Climate Change in Public Sector Performance 146 6.6 Mainstreaming Gender Agenda in the Public Budgeting 149 6.7 Recommendations and Reform Options 150 7. Improving Intergovernmental Fiscal Relations 154 7.1 Structure of Subnational Governments in Kazakhstan 155 7.1.1 The legal foundation for the division of functions across subnational governments 157 7.1.2 Functional division and spending by subnational governments 159 7.2 Revenue Assignment for Subnational Governments 161 7.3 Transfer Formula and Equalization 162 7.3.1 Subvention and withdrawal 162 7.3.2 Targeted transfers 164 7.3.3 Shortcomings of the existing system of transfers equalization 166 7.4 Possible Reform Directions and Examples from Other Countries 168 Implications for Kazakhstan’s system of subventions and withdrawals 172 Increasing revenue stability at the second and third tiers of subnational governments 173 Annex 1. Analyzing the Distribution Impact of Fiscal Policies 175 Annex 2. Estimating Fiscal Multipliers for Kazakhstan 177 Annex 3. PEFA 2018 Performance Indicators 178 Annex 3. Selected VAT Exemptions and Policy Recommendations 179 Annex 4. Selected Non-export Zero Rating and Policy Recommendations 183 Annex 5. Assignment of Functions in Kazakhstan According to the Budget Code (Summary) 184 Bibliography 186 8 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Acknowledgments This report is led by Sjamsu Rahardja (Senior Economist, ECCM1), with Kathrin Plangemann (Manager, Operations, LCC1C) as a co-task leader. The team consisted of Azamat Agaidarov (Country Economist), Azamat Aldiyarov (Consultant), Sholpan Aitenova (Consultant), Kateryna Bornukova (Consultant), William Dillinger (Consultant), Syedah Aroob Iqbal (Research Analyst, HECED), Magali Junowicz (Consultant), Matthieu Kalbfuss (Consultant, HECED), Salamat Kussainova (Governance Specialist), Tuan Minh Le (Lead Economist), Metin Nebiler (Economist), Alexandra Molchanovskaya (Consultant), Martin Melecky (Lead Economist), Hanna Muria (Consultant), Franz Ulrich Ruch (Senior Economist), Ivan Shulga (Senior Social Protection Specialist), Waewnet Sukkasem (Consultant), Maria Ustinova (Consultant), Murat Temirkhanov (Consultant), Ayesha V. Vawda (Lead Education Specialist), Aristomene Varoudakis (Consultant), Mitchell Wiener (Senior Social Protection Specialist), Usama Zafar (Research Analyst, Social Protection), Almira Zakiyeva (Education Specialist). The team is also grateful for the assistance Gulmira Akshatyrova, Azat Alkeyev, and Symbat Ismat provided. The team is grateful for the help of the Ministry of National Economy and Regional Akimats, especially planning departments in Astana, Mangystau, and Pavlodar, in sharing information. The team is especially grateful for the fruitful exchange of information with the Ministry of Education and Science, Regional Akimats, and especially the Education Departments from Almaty, Atyrau, Mangystau, and Turkestan regions. The team is also grateful for the information exchange with the Ministry of Labor and Social Protection, the Financial Center JSC, the Information Analytical Center, and the National Academy of Education. The team is also grateful for feedback from the National Bank of Kazakhstan, Ministry of National Economy, Ministry of Finance, Ministry of Education, and Samruk Kazyna. The team worked under the overall guidance of Antonio Nucifora (Practice Manager, ECCM1) and previously Sandeep Mahajan and Daniel Boyce (Practice Manager, EECG1). The team benefited from advice from David Stephen Knight (Lead Economist, Program Leader), together with inputs from Henry Aviomoh (Young Professional), Andrei Busuioc (Senior Financial Management Specialist, EECG1), Kai Kaiser (Senior Public Sector Specialist), William Hutchins Seitz (Senior Economist, EECPV), Lorena Vinuela (Senior Public Sector Specialist, EAWG1), Arman Vatyan (Lead FM Specialist), Ian Hawkesworth (Senior Governance Specialist), and Aliya Kim (Senior Financial Management Specialist). The team also thanks the following peer reviewers for their valuable inputs: Richard Record (Lead Economist, EECDR), Tazeen Fazih (Lead Economist, Program Leader, HECDR), Lea Hakim (Senior Economist, EECM1), and Andzs Ubelis (Senior Governance Specialist, EECG1). The team is also grateful for the assistance to Shynar Jetpissova (External Affairs Officer) for guidance on publication and outreach. KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 9 Introduction Kazakhstan’s fiscal policies have undergone several changes to strengthen fiscal discipline and preserve buffers to stabilize growth against short-term shocks. Since the last World Bank 2017 Public Finance Review (PFR), the authorities have implemented new policies on fiscal management, such as discontinuing direct transfers from the National Fund of the Republic of Kazakhstan (NFRK, or the “Oil Fund”) to extrabudgetary funds and state-owned enterprises (SOEs) outside the budgetary process, introducing the non-oil deficit as a fiscal anchor, and improving transparency in budget reporting. The Government has also introduced several changes in the fiscal rules, starting from 2023, to avoid fiscal expansion during good times (i.e., a procyclical fiscal stance), set a conservative rule for regular withdrawals from the NFRK to the budget, and established controls to maintain a sustainable level of government debt. The Government has also introduced changes in budgeting and planning policies to improve results. Kazakhstan started introducing performance-based public administration tools in 2007 and is gradually moving to implement performance-based budgeting (PBB). The Agency for Strategic Planning and Reform (ASPR)—an apex institution at the center of government—was created in 2020 to facilitate progress on policy coordination through a greater focus on strategic planning, budgeting, and monitoring. These changes, however, are inadequate to sustain macro-fiscal resilience and inclusive growth. The substantial increase in quasi-fiscal activities (QFAs) by extra-budgetary funds and SOEs makes it difficult to pin down the overall fiscal stance and complicates macro-fiscal management. Quasi-fiscal activities also generate risks to the government budget from quasi-fiscal deficits that have not been well monitored. The tax revenue-to-GDP ratio is declining, exposing the government budget to more volatile oil revenue. Generous fiscal incentives to attract investment have pressured tax revenues, and their contribution to business expansion has not been systematically analyzed. Productivity in collecting personal income tax and value-added tax is low, and tax policies can do more to improve equity and support Kazakhstan in achieving its green agenda. The reforms also have yet to improve the effectiveness and efficiency of the fiscal program and better progressivity in fiscal policies. Improvement in the budgeting process has focused largely on improving compliance, while weaknesses remain in linking strategy, implementation, and performance monitoring. Large and continuous programs, such as active labor market programs and tax incentives, have yet to be accompanied by a mechanism for impact evaluation. While there has been an improvement in promoting transparency in reporting budget execution, the classification of expenditures is not fully compatible with best practices, making it hard to compare Kazakhstan’s spending with that of other countries. Despite adopting a strategy to achieve carbon neutrality by 2060, the budgeting process has not internalized a mechanism to support the strategy’s implementation. The budgeting process and revenue policy can also support subnational governments (SNGs) to deliver on their apportioned tasks and respond to local development challenges. This public finance review (PFR) focuses on Kazakhstan’s budgeting, revenue, and spending policies critical for sustaining inclusive growth. It aims to strengthen knowledge and capacity for fiscal policy- making and generate public discourse on the challenges and possible ways to improve inclusivity and resilience in public finances. The first three chapters of the PFR review the core fiscal policy and revenue mobilization issues. Chapter 1 discusses the fiscal landscape, fiscal framework, and progressivity of fiscal policy. Chapter 2 looks at the footprint of quasi-fiscal activities, which affects the overall fiscal stance and exposes certain fiscal risks. 10 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Chapter 3 discusses the stagnation in non-oil revenue and collection across taxes and outlines reform options to improve the tax regime. This PFR also covers education and social protection spending, constituting about 42 percent of general government budget spending, and is critical for Kazakhstan’s social agenda and long-term development goals. Chapter 4 analyzes the efficiency of public spending on education, discusses challenges in delivering equitable access to quality education, and offers options for enhancing spending effectiveness through institutional and policy changes. Chapter 5 discusses the efficiency and effectiveness of spending on the social protection system, particularly the coverage and targeting of social assistance programs, issues in implementing active labor market programs, and challenges in delivering social insurance. Because of data constraints, this PFR excludes analysis on social benefits, pensions, and the State Social Insurance Fund. The last two chapters cover the core system of public-finance management issues on budgeting and inter-governmental fiscal relations. Chapter 6 considers options for further improving budgeting, planning, and monitoring to deliver better fiscal outcomes for inclusive and resilient growth. While Chapter 7 examines emerging subnational fiscal issues and options to simplify and improve certainty in the transfer mechanism from central to SNGs and within the SNG hierarchy. KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 11 1. Strengthening the Fiscal Framework for Inclusive and Resilient Growth KEY POINTS • Kazakhstan has successfully used fiscal policy to sustain macroeconomic stability. Sizable oil revenue and relatively low government debt provided the Government with the fiscal buffers needed to implement considerable fiscal measures to weather the negative impact of the COVID-19 crisis. The Government also changed the fiscal rules to reign in the non-oil deficit, avoid a pro-cyclical fiscal stance, and manage government debt. • Despite Kazakhstan’s sizable fiscal reserves, the country faces development challenges that require its fiscal framework to adapt. Kazakhstan faces higher external volatility associated with higher global interest rates, more volatile oil prices, supply chain disruptions, and higher food commodity prices. Non-oil revenues as a percentage of GDP have stagnated, exposing the budget to overreliance on oil revenue, which can limit the budget’s ability to deliver more public spending needed for education, social protection, and the green transition. On the other hand, the size of quasi-fiscal activities required to deliver government programs has increased, complicating the management of macro-fiscal policy. • The Government should further strengthen the fiscal framework and the credibility of fiscal rules. Broadening the institutional coverage of the fiscal framework to quasi-fiscal activities and monitoring the contingent exposure of quasi-fiscal activities are needed. The fiscal framework should also support the long-term growth agenda, mainly through public spending on education and supporting the green transition. Meeting emerging expenditure pressures will also require improved efficiency in delivering public services. In parallel, Kazakhstan needs to implement reforms that increase non-oil revenues. 1.1. Fiscal Response amid Growing Macro Challenges As a resource-rich country, the challenge for Kazakhstan is to use fiscal policy to tame macroeconomic fluctuations, sustain the provision of public services, develop the potential for the non-resource economy, and create fiscal space for a rainy day. Between 2010 and 2021, oil contributed about 65 percent of Kazakhstan’s merchandise exports, and oil revenue represents about 36 percent of the government budget revenue.1 When commodity export prices are high, investors and private companies are confident to spend more, and greater liquidity in the financial sector often responds by making credit abundant. But governments in resource-rich economies often increase spending growth in good times, thus contributing to the economy’s overheating. Such a pro-cyclical fiscal stance tends to exacerbate macroeconomic fluctuations, fueling growth during booms while cutting back on spending during slumps because of lower revenue. Kazakhstan has decisively deployed strong fiscal responses during crises, especially during the COVID-19 pandemic, and has continued to provide fiscal support to sustain economic recovery. These 1 Oil revenue to the Republican Budget, defined as customs duty from oil exports and transfers from the NFRK. 1. STRENGTHENING THE FISCAL FRAMEWORK FOR INCLUSIVE AND RESILIENT GROWTH 13 fiscal responses have led to increasing government debt as a percentage of GDP, with a corresponding decline in the share of NFRK assets. In 2014–2017, responding to the precipitous drop in the oil price, the Government swiftly introduced anti-crisis measures that included a fiscal package of US$20 billion, equivalent to 12 percent of GDP (World Bank 2017). In 2020, the authorities suspended the fiscal rules following the President’s declaration of a National Emergency and the delivery of a substantive fiscal package to weather the COVID-19 crisis. By October 2021, the Government had implemented fiscal support measures in response to the crisis, equivalent to 8.4 percent of 2020 GDP, and stood at the average of comparator countries. Support provided through budgetary measures amounted to 5.5 percent of GDP, of which 0.7 percent was directed to the health sector and 4.8 percent to the rest of the economy. An additional amount, equivalent to 2.9 percent of GDP, was provided “below the line” as liquidity support (see Section 1.4.2 and Chapter 4, or broader analysis of quasi-fiscal activities). The Government has continued to provide fiscal support to help economic recovery. The risk of shocks to the economy underlines the importance of preserving sufficient fiscal space. After the domestic banking crisis in 2009, year-on-year (yoy) output in 2016 sharply declined after a steep drop in global oil prices in 2015. In 2020, the COVID-19 pandemic caused a 2.5 percent decline in Kazakhstan’s GDP. Less than a year after the economy recovered from the pandemic-induced crisis, Kazakhstan is now experiencing other external pressures from the fallout of the Russia-Ukraine war, a global economic slowdown, and tightening global financing conditions. The high inflationary pressure and the upswing in commodity export prices also require calibration in the fiscal policy. The headline Consumer Price Index (CPI) inflation rate reached 8.4 percent yoy in December 2021, the highest rate since 2016 and well above the National Bank of Kazakhstan’s (NBK) target range of 4–6 percent. Supply disruptions and higher global commodity prices due to the war in Ukraine increased the headline inflation rate to 15 percent yoy in July 2022. Such a high inflation rate has fueled concerns because of the disproportionate impact on the poor, as they rely more on wage income and have fewer assets to smooth consumption. The high inflation since 2020 also occurred as real GDP fell below its potential, which suggests the drag from negative supply shocks (higher input prices and supply chain disruptions) contributed to higher prices (Figure 1‑1). The negative impact of inflation on consumers calls for an appropriate fiscal policy response coordinated with monetary policy efforts to preserve price stability. It is also important that quasi-fiscal spending, often sizable (see Chapter 4), does not jeopardize inflation targeting by the monetary authorities. Strengthening the social safety net to protect the poor against negative shocks in real income and improving the quality of public service delivery should be a focus area of the fiscal response. An overly expansionary fiscal stance, particularly through subsidies and quasi-fiscal activities (QFAs), can exacerbate inflationary pressures. The green transition gives rise to another macro-fiscal challenge, as the economy relies heavily on fossil-fuel extraction. The growth model continues to be confronted by high dependency on hydrocarbons, a lack of vibrancy in the private sector, and new risks related to decarbonization. As a highly fossil-fuel- dependent nation, Kazakhstan is vulnerable to climate change and exposed to global efforts to mitigate it. Climate shocks are expected to worsen labor market outcomes for people in Kazakhstan. These estimated macroeconomic impacts are expected to feed through to worsening economic labor market outcomes. Recent analysis suggests that real wages are expected to decrease by 2.1 percent by 2060 and 3.7 percent in 2090 in a combined climate shock scenario (agriculture, flooding, heat shocks) in the RCP8.5 scenario.2 As a result, poverty will be 3 percent higher by 2060. On average, about 30 percent of the budget revenue (6 percent of GDP)3 depends on fossil fuels. Acting early to shift this revenue to other bases will help reduce revenue risk as Kazakhstan, and the rest of the world reduce dependence on fossil fuels. 2 Representative Concentration Pathway (RCP) 8.5 refers to a high-emission global warming, or business as usual. 3 Between 2017 and 2021. 14 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Kazakhstan must manage a transition away from carbon-intensive assets to support longer-term growth while mitigating the costs of that transition and the physical impacts of climate change. While it is important to have market-driven private investments supporting the transition, there will be numerous additional demands for government expenditure arising from the low-carbon transition. But global decarbonization will likely affect Kazakhstan’s oil exports and the revenues of the National Fund of the Republic of Kazakhstan (NFRK, or the “Oil Fund”). A recent study suggests that the energy transition can have a modest impact on the fiscal deficit under a high-oil-price scenario but a high impact under accelerated global decarbonization when the oil price is low (World Bank 2022). In this context, the challenge is ensuring fiscal policy can support growth, economic transition, and inclusiveness. Figure 1‑1. Headline CPI inflation and output Figure 1‑2. Estimated impacts of climate gap4 (%) shocks on poverty (%) and real wages (KZT) Source: World Bank staff calculations. Source: World Bank staff calculations. Figure 1‑3 Consolidated revenues and general Figure 1‑4. The budget deficit, government debt, government spending (% of GDP) and the NFRK’s assets (% of GDP) Source: World Bank staff calculations. Source: World Bank staff calculations. 4 Output gap defined as percentage difference between actual real GDP and potential real GDP (based on Hodrick-Prescott filter on real GDP series). *- 1. STRENGTHENING THE FISCAL FRAMEWORK FOR INCLUSIVE AND RESILIENT GROWTH 15 Although the fiscal resources needed to stabilize growth are currently sizable, Kazakhstan must sustain higher non-oil revenue to support spending for inclusive and resilient growth. Total consolidated revenue as a percentage of GDP declined from an average of 26.4 percent in 2005–2010 to 19 percent in 2016– 2021, owing to weaker non-oil revenues. On the other hand, general government spending as a percentage of GDP fluctuated at around 24.3 and 21.8 percent, respectively, during those two periods. To make up for the revenue shortfall and finance the growing budget deficit, the authorities have used government debt issuance and assets from the NFRK. The Government is also increasingly using extrabudgetary funds (EBFs) and state-owned enterprises (SOEs) to perform QFAs (off-budget spending) for development. At the same time, the risks of these QFAs are not part of the fiscal framework (see Section 1.4 and Chapter 4 on QFAs). Although the NFRK’s assets are now sizable and government debt is relatively low, Kazakhstan could face painful fiscal adjustments if the oil price were to deteriorate. To sustain higher spending needs for human capital, social protection, and a green transition, Kazakhstan needs to increase non-oil revenues by reforming its tax policy and strengthening its tax administration. This chapter reviews the consistency of the fiscal framework with Kazakhstan’s macroeconomic and long-term development challenges. The chapter attempts to answer the following questions: Does the fiscal framework support a counter-cyclical fiscal stance? How does public spending evolve and perform in international comparison? What are the emerging pressures on public expenditure and the fiscal balance? 1.2. Assessment of the Fiscal Posture 1.2.1.  Size of fiscal operations Education, social security, and welfare are the main items of Kazakhstan’s public expenditures that sustained an increase in nominal terms. Since 2005, social security expenditures accounted for the largest share (21 percent) of public expenditures, followed by education expenditures (18 percent) and healthcare expenditures (11 percent). These spending items have increased as a proportion of total spending (Figure 1‑5), reflecting Government’s commitment to education and social agenda. To capture the magnitude of the change in spending in a shorter time grouping, we calculated the fiscal force for two sub-periods, 2014–2017 and 2018–2021 (Figure 1‑6). While the results for the 2014–2017 period indicate that public expenditures were directed at the implementation of anti-crisis measures and infrastructure (captured in the “others” category), 2018–2021 was marked by increased public expenditures in key areas of social development, such as social welfare, education, and healthcare. The fiscal force of education and social security expenditures was most significant during 2018–2021, amounting to 4.8 and 4.5 percent, respectively. Most government spending across economic classifications is on capital, goods and services, and social transfers. These items constituted about 69 percent of general government spending between 2005-2021. The composition somewhat changes, with the share of capital spending sliding from 26.4 percent in 2005- 2010 to 14.5 percent in 2017-2021, while spending allocation on social transfers increased from 18.5 percent to 22.3 percent during the same periods (Figure 1‑7.). Similar to spending based on the functional category, the allocation of spending on “others” increased substantially from 3.8 percent to 11.2 percent, driven by anti-crisis measures in 2017. To examine the magnitude, fiscal force is calculated using shorter time groups (Figure 1‑8. ). Similar to Figure 1‑6, In 2014-2017, spending on others drove almost all fiscal expansion.5 In 2018-2021, spending on capital, social transfer, and the wage bill drove fiscal expansion. The increase in the wage bill in this period was due to higher salaries for health and social workers during COVID-19 and a substantial rise in teachers’ salaries. 5 The measures included financing the infrastructure of Astana Expo. 16 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Figure 1‑5. Сomposition of government spending Figure 1‑6. The fiscal force of government by functional classification (2005–2021) spending by functional classification Source: World Bank staff calculations. Source: World Bank staff calculations. Note: Other expenditures include interest, culture, sport, Note: Fiscal Force = share of expenditure X change in growth rate tourism, fuel, energy, transfers, industry, construction, etc. Figure 1‑7. Composition of government spending Figure 1‑8. The fiscal force of government by economic classification (2005-2021) spending by economic classification Source: World Bank staff calculations Source: World Bank staff calculations. Note: Fiscal Force = share of expenditure X change in growth rate While Kazakhstan’s on-budget public expenditure has fluctuated over the years, it has remained lower compared to oil-exporting countries, upper-middle-income countries, and the global average (Figure 1‑9). Total public expenditure at the general government level represented 23 percent of GDP in 2021, little changed from 21% in 2014. In 2021, the total public expenditure of oil-exporting and upper- middle-income countries stood at 34 and 32 percent of GDP, respectively, while the global average was 39 percent. However, compared to other country groups, these figures should be interpreted cautiously due to the effective use of quasi-fiscal activities for which gross spending in 2017-2021 is estimated at 3.65 percent 1. STRENGTHENING THE FISCAL FRAMEWORK FOR INCLUSIVE AND RESILIENT GROWTH 17 of GDP (see Chapter 4). The civil service wage bill was equivalent to 3.8 percent of GDP in 2021, up from 2.9 percent in 2014 (Figure 1‑10). This was significantly lower than in any other comparator group, although the high level of employment in SOEs partly accounts for the low wage bill at the general government level.6 Public investment financed from the state budget was slightly lower than in oil-exporting countries and the global average and substantially lower than in upper-middle income countries, representing 2.6 percent of GDP in 2021, down from 4.9 percent in 2014 (Figure 1‑11). Large investments by SOEs, particularly in energy and communication infrastructure, also mask the amount of public investment at the general government level. High-quality infrastructure projects can play a key role in boosting growth in jobs and output if selected transparently and strategically using rigorous criteria that assess their economic viability and potential to attract private investment. Figure 1‑9. Total public expenditure in Figure 1‑10. Civil service wage bill in international comparison (% of GDP) international comparison (% of GDP) Source World Bank staff calculations based on data from the Source: World Bank staff calculations based on data from the IMF GFS. IMF GFS. Figure 1‑11. Public investment in international comparison (% of GDP) Source World Bank staff calculations based on data from the IMF GFS. 6 In 2020, about 1.1 million people in Kazakhstan (about 12 percent of employment) were employed by the “quasi-public sector” (formal terminology for SOEs). Meanwhile in the same year about 490 thousand people were employed in public administration and defense. 18 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 1.2.2.  Rigidity in government spending: modest but rising Fiscal rigidities are constraints that limit the ability of the Government to change the level or structure of public spending. Fiscal rigidity reflects the proportion of spending that is naturally inflexible because of certain commitments or contractual obligations, such as public sector wages, pensions, debt service, and transfers to subnational governments (SNGs). Some level of fiscal rigidity can be useful to achieve certain long-term goals. For example, clear commitments to public spending on health and education can support the credibility of the Government’s plans in these areas. However, high fiscal rigidity can also limit the Government’s ability to adjust the budget in response to changing needs, such as cutting spending during downturns or repurposing spending for other public investments offering higher returns. We use a similar approach to Herrera and Olaberria (2022) to determine spending rigidity. We have updated the data to 2020 and used the econometric approach to calculate the structural level of spending on civil service wages and social benefits driven by structural factors beyond immediate government control. The total structural rigid spending comprises estimates of structural spending for civil service wages and social benefits, and actual interest payments. This cross-country analysis excludes variations in specific spending commitments through which the central government transfers to SNGs (e.g., pre-tertiary education in Kazakhstan). While structural rigidity in Kazakhstan’s budget spending is relatively modest, the gap between the structural and actual spending levels suggests some room for spending efficiency. The estimated structural rigidity of spending provides some idea of the spending level driven by long-run economic fundamentals and beyond immediate government control. The proportion of rigid structural spending as a percentage of total spending ranged from 25 percent in 2015 to 23 percent in 2020 (Figure 1‑12), but this was lower than the ECA regional average of about 70 percent in 2017 (Herrera and Olaberria 2022). Part of this is, first, because of Kazakhstan’s relatively low government debt, which affects the size of interest payments. Second is that Kazakhstan has a relatively low structural component of wage bills, social benefits, and pensions compared with the regional average, given the population size, population density, and the share of the urban population.7 But the most interesting finding from this exercise is the sizable gap between the rigid structural component of the spending and the actual spending, suggesting room for policy choices to improve spending efficiency in the wage bill, social benefits, and pensions. Figure 1‑12. Structural rigidity and actual spending in total government spending (%) Source: World Bank staff calculations based on method and updated data from Hererra and Olaberria (2022). Note: Rigid spending is defined as interest payment and the structural components of civil service wages, social benefits, and pensions. 7 A lower population number is associated with lower government size and a lower government wage bill. The data also suggests that higher population density and a higher share of the urban population are associated with a lower government wage bill per capita, which could reflect more people working on non-government activities. A higher population number and more revenue from social contributions are associated with a lower payout for social and pension benefits per capita. Meanwhile higher income per capita is associated with a higher government wage bill per capita and a higher payout of social and pension benefits per capita. 1. STRENGTHENING THE FISCAL FRAMEWORK FOR INCLUSIVE AND RESILIENT GROWTH 19 1.2.3.  The use of quasi-fiscal activities Although Kazakhstan’s government spending ratio to GDP is lower than other countries, the use of quasi-fiscal activities (off-budget spending) to support the economy has increased substantially. This spending is being carried out by quasi-fiscal entities (QFEs) (See chapter 4). While during 2011–2015, on average, spending by QFEs amounted to 1.13 percent of GDP, such spending rose in 2016–2021, on average, to 3.65 percent of GDP (Table 1‑1). Since the 2016 banking crisis, the Government has continued using QFEs to deliver various programs supporting financial stability and the economy. During the COVID-19 pandemic, about one-third of support provided to the economy consisted of liquidity measures channeled below the line. Most of the quasi-fiscal activities to support various programs were carried out by the NBK, which, in 2016–2021, channeled funds of about 1.4 percent of GDP in gross terms compared with 0.6 percent of GDP in 2011–2015, mostly for capital contributions and the purchase of bonds for the PLF. Gross spending by the Pension Fund also increased substantially, mainly due to a spike in payouts in 2021, to allow pensioners to use pension savings for mortgage down payments. The Government also tasked Samruk Kazyna to finance social and development projects beyond its core businesses. Baiterek Holding and its two subsidiaries—the DBK and Otbasy Bank—provide concessional financing on housing ownership and other programs with subsidies that have implications for the fiscal stance, which, in 2016–2021, was estimated at 0.4 percent of GDP. Table 1‑1. Gross QFAs by important extrabudgetary funds and SOEs (% of GDP)   Average Institutions and activities 2011–2015 2016–2021 National Bank of Kazakhstan   Purchase of bonds and shares of SOEs 0.47 - Bond purchase and capital contribution to the Problem Loan Fund 0.07 0.44 Housing and mortgage loan program - 0.19 State Program for Employment Roadmap - 0.21 State program Nurly-Zhol - 0.10 Program to improve stability of the banking sector - 0.38 Program to support priority sectors of the economy   0.04 Samruk-Kazyna Holding   Financing of social and other projects directed by the government 0.16 0.14 Baiterek Holding   Discount for loans below market rate* 0.23 0.38 State Social Insurance Fund   Social payment 0.07 0.55 State Health Insurance Fund   Benefit payout - 0.33 Unified Accumulative Pension Fund   Benefit payout 0.13 1.01 Total gross quasi-fiscal activities 1.13 3.78 Memorandum: General government spending 18.1 21.5 Source: World Bank staff calculations based on published financial statements. * Estimates. 20 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH The intensive use of quasi-fiscal activities (QFAs) can complicate the design of fiscal policy. The fiscal discipline the Government intends to achieve is likely undermined by QFAs directed by the Government. Moreover, deficits from QFAs generate implicit contingent liabilities for the Government to the extent that loss- making SOEs, or else those that do not invest in equipment, maintain efficiency, and develop their networks, will eventually need financial support from the budget (see Chapter 4). Successive bank bailouts have also been a significant source of quasi-fiscal exposure and entail risks for the fiscal position, as further detailed below. 1.2.4.  Government spending and tax multipliers Government spending and tax affect the economy through various channels. One is through components of aggregate demand: consumption, investment, and net government spending. The second is through the supply of factors, mainly labor (vs. leisure) and capital. For analytical purposes, this chapter uses the variation in government spending and tax revenue to approximate their impact on Kazakhstan’s non-oil output. Studies suggest a positive effect on growth from fiscal spending, albeit to a scale that varies depending on institutional aspects and the timing of the intervention. Empirical evidence suggests that the fiscal spending multiplier positively affects output (for example, Blanchard and Perotti (2002)).8 Under a more open economy, increasing government spending can less impact domestic demand. Under a flexible exchange rate, higher government spending can raise domestic interest rates, attracting foreign capital inflows and putting upward pressure on the currency, causing currency appreciation to affect net exports negatively. The fiscal spending multiplier on output can also differ depending on the macroeconomic circumstances—government spending during an economic recession has a higher multiplier than during a period when the economy is expanding (Auerbach and Gorodnichenko 2012). But, overall, the effects of fiscal policy on output in developing economies appear to be smaller than those observed in advanced economies (Hory (2016), Itzetzki et al. (2013)). This difference is often attributed to poorer institutional capability to deliver public services. The analysis finds that Kazakhstan’s fiscal spending multiplier on non-oil GDP is modest. Fiscal multiplier estimates for Kazakhstan suggest that fiscal policy impacts non-oil output, with fiscal consolidation (expansion) reducing (increasing) output. Using quarterly data from 2000 to 2021 (see Annex 2 for the approach), our analysis finds that the government spending multiplier in Kazakhstan is estimated at 0.4 on impact. A KZT 100 increase in spending increases non-oil output by KZT 40 (Figure 1‑13). After two years, the spending multiplier is estimated at KZT 80.9 The magnitude, however, is subject to considerable uncertainty, as implied by the large confidence intervals. Because this analysis looks at fiscal multipliers on non-oil output, the results are also not easily comparable with other studies focusing mostly on total output fiscal multipliers. Nevertheless, the size of Kazakhstan’s spending multiplier appears to be lower than the finding in a recent study on total output fiscal multipliers in selected Asian countries (Dime, Ginting and Zhuang 2021).10 The Government should consider the size of subsequent spending measures, given that the impact of the previous spending on output comes with a time lag. The size of the spending multiplier on non- oil GDP doubles after one year. This finding suggests that the Government should consider this before launching another round of spending measures to support the economy. This consideration is critical when inflationary pressure is high. Improving the quality of public spending can help Kazakhstan increase the size of its fiscal spending multiplier. Membership in the Eurasian Customs Union and a relatively flexible exchange rate regime can limit the size of the government spending multiplier. But these factors can be offset by making public service 8 Studies also suggest that the effects of the fiscal spending multiplier are smaller for economies with a flexible exchange rate or more open trade environment (Ilzetzki, Mendoza and Vegh 2013). 9 Previous estimates of Kazakhstan’s fiscal spending multiplier were between 0.2 and 0.4 in the short run—that is, output increased between KZT 20 and KZT 40 for every KZT 100 increase in spending (World Bank 2015). Other studies, such as (Sheremirov and Spirovska 2022), include Kazakhstan in a panel estimation and find one-year fiscal spending multipliers between 0.75 and 0.85. 10 Their estimate of spending multiplier on total output is about 0.73 after one year and the tax multiplier on total output is about -0.4 over the same period. 1. STRENGTHENING THE FISCAL FRAMEWORK FOR INCLUSIVE AND RESILIENT GROWTH 21 delivery more efficient and effective by improving budgeting and planning processes, giving more autonomy to SNGs to respond to local development priorities, and improving monitoring and evaluation (M&E) of government-funded programs. The size of fiscal multipliers could also be reduced by the prevalence of informality, which implies that public spending efficiency in high-informality countries could be enhanced with structural reforms that encourage the formalization of firms and employment (Furceri, et al. 2022). The estimated tax multiplier is negative—an increase in tax revenue is associated with output losses and vice-versa. The tax multiplier is estimated at -0.6, suggesting that every KZT 100 increase in tax revenues decreases non-oil output by KZT 60. The effect on the economy is largest on impact and then slowly fades, suggesting that households and businesses anticipate the impact and respond pre-emptively to tax increases. Assuming changes in non-oil tax revenues have a symmetrical impact on non-oil output, the negative estimate suggests that the impact of declining tax revenue on non-oil GDP diminishes after the impact. Businesses and households could anticipate the Government’s plan to raise non-oil tax revenue in the future, either through higher taxes or by raising debt, which can raise domestic interest rates. This result provides a reminder of the effectiveness of using tax incentives to support growth. A recent study on selected Asian countries estimates that the tax multiplier on total output is about -0.4.11 Figure 1‑13. Cumulative fiscal multipliers a. Government spending b. Tax revenue Multiplier Multiplier 2.5 0.0 2.0 -0.2 1.5 -0.4 1.0 -0.6 0.5 0.0 -0.8 -0.5 -1.0 On impact 1-year 2-years On impact 1-year 2-years Government spending Tax revenue Source: World Bank staff calculations. Figure 1‑14 Cumulative government spending multiplier after one year Source: World Bank staff calculations. 11 See Dime, Ginting, and Zhuang (2021). 22 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Estimates of spending multipliers for education, transport, and social protection seem to be positive, although their confidence intervals are large. The estimated multiplier of education spending on non-oil GDP after one year is about 5.7, and for social protection is about 4. However, the preciseness of these multipliers is difficult to pin down because of their large confidence intervals (Figure 1‑14). For example, the spending multiplier from transport to non-oil GDP is about 3.6 but with a large lower bound, which implies that such spending could sometimes have a small negative impact on non-oil output. Similarly, the multiplier spending on other items, with carries about half of the budget, is negative with a large confidence interval which suggests it is not statistically significant from zero. As a result, the estimated average spending multiplier after one year is around 0.7. 1.2.5.  Cyclicality of the fiscal position Kazakhstan’s fiscal policy was generally pro-cyclical, except during the COVID-19 pandemic. The fiscal policy stance became pro-cyclical as public spending kept growing in upswings, especially in 2017 and 2019. The pro-cyclicality of the fiscal stance is shown by the positive correlation between real spending growth and changes in the oil price, with a correlation coefficient of 0.36. Half of the data exhibited increases in real spending when the oil price increased (Figure 1‑15). Panel b also shows the deficit-biased stance outside crisis times as the Government injected net spending when the economy was above its potential (i.e., positive output gap). Empirical analysis of the fiscal stance before the COVID-19 pandemic further confirms the pro-cyclicality of fiscal policy (see Box 1‑1. An empirical assessment of the fiscal stance (2000–2019) ). In parallel, the guaranteed transfers, at around 4 percent of GDP, provided a steady revenue to the budget regardless of the economic cycle. To provide a reliable source of expenditure financing throughout the cycle, transfers should rise in “bad times,” when tax revenues shrink, and be curtailed in “good times,” when tax revenues expand due to growing economic activity. The steady guaranteed transfers undermine the usefulness of the non-oil deficit as a fiscal anchor. As a result, the fiscal framework in place has been unable to tame pressure for a pro-cyclical fiscal policy. Figure 1‑15. Growth in real government spending, oil price, and change in the output gap a. Percentage changes in primary spending in goodb. Output gap and primary balance (cyclically adjusted) times and bad times across years (2007 – 2021) Source: World Bank staff calculations Source: World Bank staff calculations. 1. STRENGTHENING THE FISCAL FRAMEWORK FOR INCLUSIVE AND RESILIENT GROWTH 23 Box 1‑1. An empirical assessment of the fiscal stance (2000–2019) A regression model uses quarterly data on budget and macroeconomic variables to confirm the fiscal stance's cyclicality further. Quarterly data provides more ample variation and can pick up supplementary revisions in the revenue and spending, typically done in the third or fourth quarters of the calendar year. The approach is similar to analyzing the counter-cyclicality of budgets in OECD countries (Egert 2014). The fiscal outcome variables used are the overall budget and primary balance. The gap between actual and potential GDP (output gap) measures the economic cycle. The oil price and the previous period’s NFRK assets as a percentage of GDP are used as additional explanatory variables. A one-period lagged dependent variable is included as a regressor to capture persistence. To assess the contributor of the fiscal stance, we use the ordinary least square (columns 1, 2, 3 and 5, 6, 7). We use a two-stage least square (4 and 8) to allow for endogeneity in the output gap (due to variations in oil prices and other factors). Table 1‑2. Regression results / effects on the contributors to the fiscal stance presents results suggesting that a higher real budget surplus (lower real deficit) is associated with the economy performing better than expected (i.e., positive output gap), which supports the countercyclical pattern of the fiscal stance. However, a higher oil price and larger NFRK assets are associated with a more expansionary fiscal stance—higher oil revenue and oil reserves have tended to boost spending and squeeze the budget balance. Table 1‑2. Regression results / effects on the contributors to the fiscal stance     Real overall balance Real primary balance     (1) (2) (3) (4) 2SLS (5) (6) (7) (8) 2SLS Dependent var t-1 -0.083 -0.144 -0.314 -0.314 -0.143 -0.231 -0.297 -0.296   (-0.66) (-1.61) (-2.60)* (-3.01)** (-1.08) (-2.50)* (-2.46)* (-2.81)** Output gap t 2.83 3.69 3.15 3.12 2.61 3.67 3.20 3.09   -1.01 (1.67)+ (1.99)+ (1.73)+ -1.02 (1.98)+ (2.08)* (1.75)+ Oil price t   -0.63 -0.47 -0.47 -0.75 -0.65 -0.65     (-2.89)** (-2.47)* (-2.26)* (-3.84)** (-3.43)** (-3.12)** NFRK/GDP t-1   -0.58 -0.58 -0.33 -0.33     (-2.93)** (-2.84)** (-1.80)+ (-1.70)+ Constant -50.2 -12.0 23.6 23.6 -33.2 13.6 24.2 24.1   (-3.92)** (-0.57) (1.68)+ (1.96)+ (-3.17)** -0.78 -1.65 (1.84)+ Observations 74 74 74 74 74 74 74 74 R-squared 0.02 0.09 0.23 0.23 0.03 0.13 0.17 0.17 Durbin P-value>F a 0.10 0.46 0.78   0.10 0.74 0.93 P-value > F 0.58 0.00 0.00 0.00 b 0.46 0.00 0.00 0.00 b t-statistics in parentheses (z-statistics for 2SLS) + p<0.10; * p<0.05; ** p<0.01. Newey-West standard errors are used. a H0: no serial correlation, b Chi-sq. 24 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 1.2.6.  Distributive effect of fiscal policies Aside from their impact on GDP growth, fiscal policies play a key role in reshaping income distribution based on the final income received by individuals. Governments use fiscal policy to generate revenue to finance public spending, which impacts income distribution and the standard of living. The policies include decisions on direct and indirect taxes, subsidies, pensions, and other direct transfers, as well as public spending on education and health. These policies will affect Market Income (wages, private transfers, capital income, other income) and result in Final Income received by individuals after taxes and government transfers. In Kazakhstan, the role of fiscal policies in income distribution can be important because of the country’s more challenging picture of “inequality.” The national poverty estimates and internationally comparable poverty rates—defined as the share of the country’s population living on less than US$6.85 per day at 2017 purchasing power parity (PPP)—declined sharply from 51 to 14 percent between 2006 and 2018, driven primarily by rising incomes from wage employment. But inequality, as measured by the Gini index, also fell from 30.2 to 29.4 in 2021, ranking Kazakhstan among the most equal countries in the world. However, labor income accounts for only 41 percent of gross national income (GNI) compared with the ECA average of 55 percent, mainly due to underreporting among high earners. According to the World Inequality Database, which adjusts for data often missing in surveys, the top 1 percent of adult income earners in Kazakhstan accounted for 15 percent of total income in 2021,12 nearly as much as the bottom 50 percent of the population combined (accounting for 16.4 percent). The contrast in wealth inequality is even starker, with the top 1 percent accounting for nearly 30 percent13 of total wealth compared with just 5 percent for the bottom 50 percent of the population combined. To study the distributional impact of fiscal policy in Kazakhstan, we use a Commitment to Equity (CEQ) methodology (Lustig 2018). The CEQ is a comprehensive incidence analysis that uses household surveys and national accounts data to assess the impact of taxes and public transfers on household poverty and inequality. The approach has been applied in over 70 countries, which allows us to benchmark Kazakhstan’s performance with relevant peer countries (see Box 1‑1 and Annex for a description of the methodology). This PFR draws findings from recent work by Bornukova and Nebiler (2023), using Kazakhstan’s 2021 Household Income and Economy Survey (HIES). As in many countries, taxes, and transfers impact income distribution in Kazakhstan. Figure 1‑16 shows that income inequality falls in all countries after governments collect taxes and deliver transfers. Compared with the other countries analyzed, Kazakhstan’s smallest decline of the Gini coefficient is found (0.04). The differences lie in the magnitudes and are driven by the relatively moderate inequality-decreasing impacts of direct transfers and taxes and the relatively small effect of in-kind public spending on health and education. Also, a large reduction of the Gini coefficient in Kazakhstan is less likely because (based on HIES) it already has a relatively lower Gini coefficient than other countries. 12 The top 1 percent of adults receiving an income accounted for 19 percent of total income in Turkey, 24 percent in the Russian Federation, 23 percent in Chile, 19 percent in the United States, and 22 percent in Brazil in 2021. 13 The top 1 percent of adults’ wealth accounted for 37 percent of total wealth in Turkey, 48 percent in Russia, 49 percent in Chile, 35 percent in the United States, and 49 percent in Brazil in 2021. 1. STRENGTHENING THE FISCAL FRAMEWORK FOR INCLUSIVE AND RESILIENT GROWTH 25 Box 1‑1 Income concepts under the CEQ analysis The CEQ sequentially measures the distributional impact of fiscal policy by defining four income concepts (Figure 6 1). First, Market Income is each household's income before taxes and transfers. It includes wages and salaries, income from capital (e.g., rents, profits, and dividends), private transfers (e.g., remittances), and other income, all before government taxes, social security contributions, and transfers. Second, Disposable Income adds the impact of direct cash transfers and subtracts personal income taxes and employee contributions to social security from market income. Third, Consumable Income subtracts the impact of indirect taxes on consumption (e.g., VAT and excises) and adds indirect subsidies to disposable income. Lastly, Final Income adds social spending on education and health as in-kind public transfers to consumable income. To measure progressivity, the analysis uses the standard Kakwani index (Kakwani 1977). A benefit (tax) is progressive whenever its entitlement (burden) decreases (rises) with income. In the case of transfers, the Kakwani index is defined as the difference between the Gini coefficient of market income plus pensions (when pensions are treated as deferred income) and the concentration coefficient of the transfers. Meanwhile, for each tax, the Kakwani index is calculated as the difference between the concentration coefficient of the tax and the Gini coefficient of Market Income plus pensions. The Kakwani index for taxes will be positive (negative) if a tax is globally progressive (regressive). In contrast, the Kakwani index for transfers is positive if a transfer is progressive in relative terms. 26 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Figure 1‑16. Gini coefficient before and after fiscal policy in Kazakhstan and peer countries Source: Argentina (Rossignolo 2017); Brazil (Higgins and Pereira 2017); Chile (Martinez-Aguilar et al. 2016); Croatia (Inchauste and Rubil 2015); Mexico (Scott et al. 2018); Poland (Goraus and Inchauste 2016); the Russian Federation (Popova et al. 2018); and the United States (Higgins et al. 2016). Kazakhstan: Bornukova and Nebiler (2023) based on Kazakhstan 2021 HIES. Please note that results do not include education spending for Kazakhstan. Despite the smaller magnitude compared to other countries, Kazakhstan’s fiscal system appears to have positively contributed to reducing income inequality. Figure 1‑17 describes how the fiscal system affects inequality (measured by the Gini coefficient on the vertical axis) across income stages. Without fiscal intervention, in this case, as suggested by the Gini coefficient of 0.41, excluding government transfers for old-age pensioners, the distribution of Market Income is more unequal. A scenario where old-age people receive a pension from deferred income affects the starting point, making the Market Income more equal, as shown in the drop in the Gini coefficient to 0.33. After direct taxes and transfers are included, the Gini declines further to 0.31 for Disposable Income. For Consumable Income, the Gini increases slightly to 0.31 in Kazakhstan after including indirect taxes. Finally, in-kind education and health spending have the largest equalizing effect, with the Gini dropping to 0.29. Overall, inequality in Kazakhstan drops from 0.41 to 0.29 after fiscal interventions, marking a reduction of around 0.1 of a Gini point. Figure 1‑17. Inequality before and after fiscal policy in Kazakhstan, 2021 Source: Bornukova and Nebiler (2023) based on Kazakhstan 2021 HIES. Please note that results do not include education spending for Kazakhstan. Note: The figure shows the Gini coefficient for each income concept described in Figure 1‑17. M.I. refers to before and after fiscal interventions. Disposable income refers to after direct taxes and transfers; Consumable Income refers to after indirect taxes, and Final Income refers to after in-kind (education and health) transfers. 1. STRENGTHENING THE FISCAL FRAMEWORK FOR INCLUSIVE AND RESILIENT GROWTH 27 Although subsidies appear to benefit low- Figure 1‑18 Distributive impact of subsidies income households, they are a less efficient and effective mechanism for supporting the poor and vulnerable. The fiscal system includes subsidized electricity, natural gas, hot water, coal, and fuel prices. In the case of Kazakhstan, utility subsidies are more progressive than fuel subsidies and contribute to a higher reduction in inequality. However, the progressivity of subsidies is much lower than the direct transfers – e.g., targeted social assistance has a greater equalizing effect on income with a Kakwani index of about 1.2 (Bornukova and Nebiler 2023). Moreover, subsidies are less efficient because of the leakage (non-poor receiving the benefit) and less effective because they are not well Source: Bornukova and Nebiler (2023) based on Kazakhstan 2021 targeted to support poor and vulnerable HIES households (see Chapter 7). 1.3. The Rules Underpinning the Fiscal Framework Kazakhstan’s fiscal framework is rules-based and has a clear, forward-looking orientation. The framework includes rules on fiscal balances, the minimum stock of assets of the NFRK, and measures of government debt. The NFRK receives fiscal oil revenues—excepting custom duties from oil exports, which are channeled directly to the budget—and makes transfers to the budget. Over time, the NFRK has accumulated substantial assets, equivalent to 29 percent of GDP at the end of 2021, and projected by the International Monetary Fund (IMF) to reach 32.5 percent in 2022. The rules target predictability in spending and fiscal sustainability. Transfers from the NFRK to the state budget play a stabilization role by allowing the state to deliver public services and meet financing commitments with predictability over time. In addition to the fiscal rules discussed below, a medium-term budget framework (MTBF) is intended to guide expenditures beyond the initial fiscal year. Macroeconomic forecasts covering five years are produced at least once a year. 1.3.1.  Improved but with too many targets The fiscal framework underwent substantial changes in 2016. Rules on the state budget’s overall balance and government debt were first set in 2013. The framework was significantly upgraded with a Presidential Decree in 2016 that set additional rules on the Central Government (Republican) Budget balance, debt, and minimum assets of the NFRK, amending the system of NFRK budget transfers. The 2016 changes improved the fiscal framework and strengthened transparency. In particular, the non-oil deficit was included as a target of the fiscal rules. The NFRK was no longer allowed to directly finance extra-budgetary spending, which had been substantial in the past. Moreover, the NFRK cannot invest in domestic financial instruments. Stronger reporting requirements on NFRK operations were also introduced. The fiscal framework in operation after its 2016 overhaul until 2021 included a set of rules on NFRK transfers, the non-oil deficit, and a set of targets on debt and NFRK assets (see Box 1‑2)). The first three rules are consistent with the 2017 Public Expenditure Review (PER) recommendations concerning the fiscal framework.14 14 See Table 3 of the previous Kazakhstan Public Finance Review (World Bank 2017). 28 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Box 1‑2. Fiscal targets and operational rules of the fiscal framework in 2016–2022 From 2016 and abandoned during the COVID-19 crisis: • Guaranteed NFRK transfers to the budget: a declining path for the maximum annual size of NFRK transfers for 2017–2019, followed by an annual limit of KZT 2 trillion from 2020 onward. • Targeted NFRK transfers to the budget: these can only be allocated by Presidential Decree to finance anti-crisis programs during economic downturns, as well as significant national projects where no alternative sources of financing are available. • The non-oil deficit of the Republican Budget as a percentage of GDP was subject to annual limits, on a declining path, set through 2025. • State debt (central and local governments and the NBK), quasi-sovereign entities' debt not exceeding 60 percent of GDP, and government debt not exceeding 25 percent from 2020 onward. Government debt (including government-guaranteed debt) is to be kept below the foreign exchange assets of the NFRK. • Debt service paid by the Republican Budget is to be kept below 15 percent of budget revenue. • NFRK assets are to be kept above 30 percent of GDP. After changes introduced in 2022 • Targeted NFRK transfers to the budget: they can be allocated only by Presidential Decree to finance anti- crisis programs during economic downturns, as well as significant national projects where no alternative sources of financing are available. • The annual size of guaranteed budget transfers must not exceed the inflow of tax receipts to the NFRK estimated at the cut-off oil price. • During period of economic growth, Republican budget spending cannot grow beyond growth in real GDP and inflation. • The Republican Budget’s non-oil deficit as a percentage of GDP was subject to annual limits, on a declining path, set through 2025. • Debt service payments are to be kept below 10 percent of Republican Budget expenditure. • Debt targets: limit on government debt not to exceed 27.5 percent of GDP by 2030 and overall public debt (government and central bank) and quasi-public sector not to exceed 53.2 percent of GDP by 2030. Figure 1‑19. Transfers from the NFRK and non- However, the detailed operational rules did not oil budget balance (% of GDP) add credibility to the previous fiscal framework. The fiscal stance was generally pro-cyclical after the 2016 overhaul and before the COVID-19 crisis. The non-oil deficit narrowed only in 2018 before widening again in 2019 to support increased social payments, supported by a discretionary transfer (targeted transfers) from the NFRK to the budget (Figure 1‑19).15 Previous World Bank (World Bank 2017) and IMF (IMF 2020) reports suggest that there is no clear advantage to having many numerical targets in addition to an explicit target for the non-oil deficit. The limit to the guaranteed transfers from the NFRK to the budget would have preserved NFRK assets and kept a lid on spending. Source: World Bank staff calculations. 15 The social payments cover higher wages for low-paid public workers, housing support for vulnerable households, debt-relief for low- income borrowers, and rural infrastructure. 1. STRENGTHENING THE FISCAL FRAMEWORK FOR INCLUSIVE AND RESILIENT GROWTH 29 But the non-oil deficit to GDP reached 12.4 percent in 2017 and 8 percent in 2019 due to various state programs and higher social payments, which were covered by debt financing and NFRK transfers to the budget. 1.3.2.  The new fiscal rules: intention versus implementation The Government introduced a new Budget Code in 2021 and a Public Finance Concept in 2022, with reformulated rules on the fiscal framework. The Budget Code was amended, among other reasons, to reduce the impact of oil price volatility on the budget, thereby: (i) ensuring inter-generational equity through adequate savings; (ii) securing resources to promote diversification and structural transformation; and (ii) improving the institutional arrangements of fiscal governance. In addition to those mentioned in Box 2, the Government intends to cap public expenditure growth below nominal GDP growth (based on 10-year average real GDP growth plus the expected inflation rate [the expenditure rule]). Furthermore, the Government also intends to reduce the non-oil deficit below 5 percent of GDP by 2030. These additional rules add to the existing ones and will be implemented with the 2023 budget. The reformulated transfer rule and the new expenditure rules intend to confer a stronger countercyclical stance to fiscal policy. In particular, the new rule on using NFRK funds with a limitation on the projected revenue from oil aims to de-link the budget from oil price volatility. As indicated by empirical evidence (see Box 1‑1), a higher oil price tends to be associated with an expansionary fiscal stance, as higher revenue compared with the conservative oil assumption in the budget tends to stimulate spending. The rule can thus strengthen the concept of “saving for a rainy day.” If the actual oil price exceeds the projected oil price for a given output, the difference is saved in the NFRK. And where the actual oil price is below the projection, the authorities can tap into NFRK funds. Expenditure rules may serve policy well, given their potential to reduce the fiscal stance’s pro-cyclical bias and ease of communication. While a large part of tax revenue is sensitive to economic fluctuations and would hence react in a pro-cyclical way during shocks, many expenditure components are not. Therefore, an expenditure rule can protect expenditure from the economic cycle and confer either an a-cyclical or a counter-cyclical pattern to the fiscal balance. If properly designed, such a rule may help to tame expenditure pressure during good times while creating needed fiscal buffers. These buffers can be used for stimulus during bad times to stabilize spending and mitigate the impact of revenue shortfalls that would increase the non-oil deficit. Compared with other fiscal rules, expenditure rules are more transparent and can be easily understood and monitored in real-time. Notably, as of 2017, half of the then-EU member countries had a national expenditure rule in place, referring to the general or the central government, along with the EU-wide rules applying to the fiscal deficit and government debt. Several resource-rich countries (e.g., Azerbaijan, Mongolia, Peru, and the Russian Federation) also have expenditure rules in their fiscal frameworks. An expenditure rule would be a welcome addition to the fiscal toolkit if it helps reach the non-oil deficit targets while reducing the pro-cyclicality of fiscal policy. An expenditure rule linking expenditure growth to GDP growth would need to be calibrated to be compatible with reaching the non-oil deficit target within a desired timeframe. To reach the non-oil deficit target over some reference period, expenditures and/ or taxes must be adjusted accordingly. Capping expenditure growth on a path consistent with the non-oil deficit target would ensure that the latter would be achieved while preventing unnecessary changes in tax rates from hindering the target. That would ensure the stability and predictability of tax policy, thus improving the quality of the overall fiscal framework. The calibration of expenditure growth would need to be based on the non-oil deficit target and GDP and non-resource revenue growth projections over the same period. Ensuring expenditure growth aligned with the target would help reach the non-oil deficit targets. The adopted expenditure rule caps expenditure growth based on past GDP growth augmented by expected inflation. It may not serve the calibration of public expenditure growth well regarding the non-oil deficit target. 30 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH The new fiscal rules and targets are unlikely to add improvement in ensuring macroeconomic stability, and several implementation issues require attention. A model simulation suggests that the new rules and targets may not add improvement in preserving macroeconomic stability under exogenous shocks. In response to a shock (in this case, a drop in the oil price), allowing nominal government spending to grow along with growth in the nominal output trend provides a cushion for aggregate demand (see Box 1‑3). However, despite the potential desirable outcome, implementing the new rules is not without issues. First, the diversity of fiscal targets in the Budget Code remains, creating redundancy, and may unnecessarily complicate the implementation of fiscal policy. Second, fixing NFRK transfers to the budget based on a cut- off oil price can help preserve the NFRK but may impede flexibility in managing assets-liabilities to finance the deficit. For example, under a high-interest rate environment, instead of drawing from the NFRK, the Government may borrow at a higher cost to finance the deficit. Box 1‑3. Simulating the potential impact of exogenous shocks under the new fiscal rules The simulation summarizes findings from work by Abilov, Aviomoh, and Rahardja (2022) using a medium-scale dynamic stochastic general equilibrium (DSGE) for Kazakhstan (Abilov 2021) to simulate the potential impact of the new fiscal rules on macroeconomic variables and welfare. Using a DSGE framework for the analysis is appropriate to capture the general equilibrium impact of a shock on key macroeconomic variables and visualize their dynamic adjustment paths. Their work complements recent findings on an analysis of Kazakhstan using a DSGE model, which suggests that commitment to counter-cyclical fiscal rules can reduce economic volatility from a negative productivity shock (Ybrayev, Kubenbayev and Baimagambetov 2022). The model incorporates important elements of macro-fiscal policy in resource-rich developing countries exposed to commodity price volatility. The model has two types of households (one of which has limited access to the financial market), the non-oil sector, the government, and the central bank. The model links oil prices with the oil sector and fiscal revenue to the budget and sovereign wealth fund. But the model also integrates other relevant features of managing income from natural resources as outlined by Agenor (2016): (i) allocation of the windfall revenue between savings for the future and public spending; (ii) spending inefficiency and absorptive capacity constraints; (iii) the impact of public capital on households’ preferences; and (iv) the production decisions of non-oil firms. The impact of the fiscal rule scenarios is evaluated through impulse responses. The model assumes that all rules are binding. The simulation result of a 10-percentage-point drop in the oil price suggests that the new rules potentially have a marginal impact in stabilizing output compared with the benchmark case (historical fiscal policy). The result also suggests that the new rules can potentially avoid a drain on the NFRK’s assets, while on the other, they can raise government debt to cover spending. The model transmits the shock as follows: a decline in the oil price -will lower export revenue, pushing the trade balance towards a deficit and weakening the exchange rate. Import prices will increase as a result and push up domestic prices. As inflation pressure grows, the central bank must raise its policy rate to stabilize prices, causing aggregate private consumption and investments to contract. The expenditure rule allows the Government to continue spending to sustain demand according to historical growth in nominal output. This spending, however, is likely insufficient to prevent aggregate demand from falling. Since the amount of NFRK transfers to the budget is fixed, the Government will have to raise government debt to finance the budget deficit. 1. STRENGTHENING THE FISCAL FRAMEWORK FOR INCLUSIVE AND RESILIENT GROWTH 31 Benchmark New fiscal rules and targets Source: Abilov et.al (2023). Having too many overlapping debt and fiscal targets can complicate fiscal management. The new set of fiscal rules remains complex, as it comprises several operational rules and fiscal targets (see Box 2 and additional rules on annual transfers from the NFRK and spending rule), some potentially inconsistent with each other and with varying legal standing. The rules include a debt service limit to control government debt but may create inconsistency if annual transfers from the NFRK are fixed according to a predetermined cut-off oil price. The Government should consider further streamlining the fiscal rules. The rules should be streamlined and focus on the non-oil deficit as the main anchor. The structural non-oil fiscal deficit as a target could be considered if potential non-oil GDP and the cyclical component of non-resource revenue can be estimated with sufficient accuracy. Some needed flexibility could be gained if the target for the non-oil deficit is specified concerning a two- or three-year period to avoid sharp policy reversals during external shocks. A medium- term macro-fiscal framework should inform the non-oil deficit target, by which government spending is set based on projected non-oil revenues. The framework could be complemented with an expenditure rule calibrated to the non-oil deficit target to facilitate a counter-cyclical fiscal stance. The Government may wish to consider simplifying the fiscal rules with a focus on achieving the non-oil deficit target, possibly complemented by a consistent target for expenditure growth. Eliminating discretionary targeted transfers and replacing them with an escape clause can further enhance the fiscal rule. It is always possible to cancel the approved budget and switch to a crisis budget in the face of exceptional circumstances, as during the COVID-19 pandemic. However, the benefit of an escape clause would be better framing the scope of discretion by introducing more transparent criteria for departures from the fiscal rule. Operationally, an escape clause can be tied to indicators such as the deviation of GDP growth from its long-term trend or a rise in the unemployment rate. The absence of an escape clause deprives the framework of flexibility against shocks, which could lead to frequent amendments that potentially undermine policy credibility. The design of expenditure rules can be adapted to needs. The expenditure target can be expressed as a ratio to GDP, in numerical terms, or as a growth rate. It can refer to nominal or real expenditure and exclude some specific items from the expenditure aggregate. Interest payments are often excluded since they are not under the direct control of the Government. Cyclically sensitive items, such as unemployment benefits, 32 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH are also usually excluded, as they are not under Government control in the short run. Public investment can also be excluded from the spending rule, especially on climate change mitigation and the green transition (see Box 1‑4). In such a case, closely monitoring expenditure reporting will be important to avoid reclassifying other government spending as public investment. Poorly designed or an excessive number of exceptions could undermine the goals of fiscal rules. A fundamental goal of fiscal rules is to build up fiscal buffers to enable greater spending in extraordinary times without compromising debt sustainability. Box 1‑4. Proposals for safeguarding specific expenditure categories in the EU fiscal framework Within the context of reforming the EU fiscal frameworks, there is a debate about the quality of public finances and how the framework can improve it. Higher quality is typically associated with a larger share of expenditures more beneficial to economic growth, development, and future generations. The European Commission identifies future expenditures with growth and value-added, notably those related to infrastructure investment (especially digital infrastructure), public research, research and development (R&D), climate-related investment, regional policy, investment in education and training, and public employment agencies. A study of various proposals suggests that such expenditure categories should be treated differently in a reformed EU fiscal framework (Reuter 2020). They should be exempt from fiscal rules, making it thus possible to be financed through debt issuance. In what is often called the “Golden Rule,” all proposals focus on exceptions of some form of investment expenditures. One of the main arguments favoring a Golden Rule is that investment expenditures also generate assets, which counterbalance debt increases. Therefore, most proposals focus on the net rather than gross investment expenditures (after deducting capital depreciation), so only adding additional assets would allow debt financing. Deutsche Bundesbank (2019) proposes that such a rule be applied symmetrically so that negative net investment would not only prohibit any debt financing but also require budgetary surpluses concerning negative net investment. Most proposals would implement the deduction of net investment regardless of the economic cycle. In contrast, Barbiero and Darvas (2014) argue in favor of an asymmetric Golden Rule, which provides extra scope for investment only in adverse economic times and gradually reduces it again in more benign times. The credibility of Kazakhstan’s fiscal rules could be enhanced if decisions relating to the circumstances under which the NFRK allocates budget transfers can be independently analyzed and their impact assessed. Credibility would be similarly enhanced if projections of the non-oil fiscal deficit and the underlying economic and fiscal trends could be independently analyzed. A credible and transparent fiscal framework could contribute to the resilience and sustainability of fiscal policy by minimizing political-economy biases that may affect the application of fiscal rules. 1.3.3.  Government debt management Kazakhstan’s government debt is modest by international standards. Government debt16 to GDP in 2021 is projected at 23.7 percent (Figure 1‑20), a slight decline from 24.9 percent during the COVID-19 crisis in 2020. About 60 percent of government debt is denominated in domestic currency and mostly in long-term Treasury bonds (80 percent) with an average time to maturity of 6.5 years (as of December 2022).17 As for 16 Defined as central and local government debt and government-guaranteed debt. 17 Time to maturity is calculated from government treasury data published in KASE (https://kase.kz/en/gsecs/), accessed in December 2022. 1. STRENGTHENING THE FISCAL FRAMEWORK FOR INCLUSIVE AND RESILIENT GROWTH 33 external government debt, about 38 percent is with multilateral development institutions,18 16 percent is in Eurobonds, and the rest is bilateral government debt. Recently, payments for government debt servicing increased to about 2.8 percent of GDP in 2021, compared with 1.8 percent in 2018, owing to higher principal payments (amortization) for domestic borrowing and higher effective interest rates. Considering the central bank’s and local governments’ debt, total public debt stood at 28.3 percent of GDP in 2021. Meanwhile, in the same year, SOE debt was about 23.6 percent of GDP, largely reflecting the debt of large holdings such as Baiterek and Samruk-Kazyna. Government debt management is divided between the Ministry of the National Economy (MNE) and the Ministry of Finance (MoF). Parliament ratifies loan agreements for external debt negotiated with bilateral and multilateral creditors. The MNE’s Debt Policy Department (DPD) sets government debt limits and targets for public debt and develops the medium-term economic and fiscal framework, including forecast parameters for public debt. The MoF’s State Borrowing Department (SBD) uses the framework and the sustainability assessment and then translates them into the Government’s borrowing plan. Kazakhstan’s Parliament (all committees of the lower chamber and the senate) ratifies loan agreements and their amendments for external government debt negotiated with bilateral and multilateral creditors. Parliament also oversees government projects financed by those debts and approves amendments to the loan agreement, which resemble a relatively high degree of ex-ante involvement. Table 1‑3. Degree of ex-ante parliamentary involvement in approving government debt Low degree High degree Parliamentary approval for every loan transaction Specific approvals for certain transactions only, such as those above certain threshold or for foreign borrowing Blanket approval for borrowing under standard terms and conditions Annual legislative approval for government borrowing (e.g., Japan) Delegation of all loan approvals to the Government Source: Reproduced from National Democratic Institute and Westminster Foundation for Democracy.19 The Government issues debt while accumulating NFRK assets partly because the sovereign wealth fund (SWF) serves a dual purpose.20 Kazakhstan is among the many oil-exporting countries that set up oil SWFs (24 out of 29 oil-exporting countries).21 The NFRK, established in 2000, is intended to perform: (i) macroeconomic stabilization through fund transfers to the government budget; and (ii) savings for future generations. These objectives imply that the NFRK should not be the primary source to finance the government budget deficit and that the Government issues debt. The NFRK accumulates assets largely from oil revenue (corporate income tax, royalties, contract-sharing, etc.), investment proceeds from assets, and other sources. For savings purposes, these revenues, net of annual transfers to the Government and other spending, are invested abroad in fixed income, equity, and gold. By the end of 2019, NFRK assets stood at US$61.8 billion (34 percent of GDP). By the end of 2021, they had declined to US$55.3 billion (28 percent of GDP) because of significant transfers to the budget for COVID-19 crisis measures and supporting economic recovery (Figure 1‑20). Government debt stood at 19.7 percent of GDP in 2019 and increased by 4 percentage points in 2021, owing to the higher non-oil deficit in 2020 and debt servicing costs. 18 World Bank, ADB, EBRD, Islamic Development Bank, and AIIB. 19 https://www.wfd.org/sites/default/files/2022-04/wfd_ndi_debt_management_legal_frameworks.pdf 20 It may also not be optimal to have zero government debt because it can help short-term liquidity management (e.g., issuance of monthly treasury bonds) and develop domestic capital market. 21 Ossowski and Halland (2017). 34 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH The aim of increasing NFRK assets appears to suggest that government debt will be determined once the authorities set the annual NFRK transfers to the budget. The authorities intend to increase accumulated assets to US$100 billion by 2030. To support this intention, the annual transfer from the NFRK to the budget is capped to not exceed the projected revenue based on the Government’s oil price forecast. One can argue that the cap on NFRK transfers to the budget should encourage lower non-oil budget deficits through lower budget spending or higher revenue mobilization. However, it may also constrain the choice for financing the budget (e.g., raising government borrowing) during negative economic shocks (as Box 1-3 suggests). In a resource-rich country such as Kazakhstan, the challenge is to balance building up SWF assets and debt repayments. Accumulating SWF assets (NFRK) provides the Government with resources for stabilizing short-term fluctuations and also provides future generations with savings. But the costs of servicing debt would usually be higher than the expected returns of the low-risk portfolio of SWF assets, which implies an opportunity cost for the Government. Indeed, the NFRK annual return22 was 3.7 percent between 2016 and 2017, while the implicit interest rate on government debt was about 4.8 percent (Figure 1‑21). NFRK returns will likely be negative in 2022 due to corrections in asset values due to higher global interest rates. One could argue that, for the Government to access financing, reducing the stock of government debt reduces the risk premium of government bonds, thus lowering the interest rate. But studies suggest that net government debt, as opposed to gross debt, is likely to determine government financing costs.23 The Government could consider exploring scenarios for government debt and scenarios for NFRK assets (net transfers to the budget) and other sovereign financial assets (balance of cash, the Government’s deposits in the banking system, FX reserves) that could potentially result in more favorable financing costs.24   Figure 1‑20 Government debt and NFRK assets Figure 1‑21 Average interest rate of government as a percentage of GDP (2015–2022) debt and return of NFRK assets (2015–2022) Source: World Bank staff calculations. Source: World Bank staff calculations and the National Bank of Kazakhstan 2022 Annual Report Note: The estimate for the 2022 NFRK return is forecasted using the S&P 500 annual return, the annual returns of the US 10-year bond, and the US 3-month treasury bond. 22 National Bank of Kazakhstan, 2022. “Report of the National Bank of Kazakhstan for the year of 2021.” The annual return of NFRK since its inception until December 2021 was 3.72 percent. 23 Hadzi-Vaskov and Ricci (2016). 24 A narrow coverage of sovereign assets and liabilities, focusing on liquid assets. The Government can also explore developing a sovereign balance sheet with broader coverage, such as including fixed assets, assets of SOEs, and other liabilities, such as monetary base and pension. Determining the choice of government debt and SWF asset as part of the sovereign balance sheet is also known as Sovereign Assets-Liability Management (SALM). 1. STRENGTHENING THE FISCAL FRAMEWORK FOR INCLUSIVE AND RESILIENT GROWTH 35 The Government should consider publishing a medium-term public debt management strategy to strengthen overall macro-fiscal management. While fiscal policy determines the debt level, a medium- term debt management strategy (MTDS) provides the Government with a framework to articulate debt management objectives, such as sustaining government financing at the lowest possible cost and acceptable risks. The MTDS also provides a framework to outline the preferred cost-risk trade-off (or strategy) selected by the government. Publication of a debt management strategy boosts transparency because there is a clear articulation of the authorities over the target cost-risk indicators that inform borrowing choices. An annual report published by parliament discusses the implementation of the strategy and any deviations from said targets. Kazakhstan does not publish an MTDS, which implies that the planned debt or liability management operations could be characterized as “opportunistic.” Without a published MTDS, it is unclear if a systematic assessment of cost and risks of alternative debt financing scenarios and funding sources (bilateral, independent fiscal institutions [IFIs], the market) is discussed by the relevant authorities (e.g., the MNE, MoF, NBK, Presidential Administration) and presented to parliament. Also, the MTDS can protect debt managers from criticism for well-intended actions because most government borrowing decisions are taken with uncertainty and because it exposes the budget to future changes. 1.4. Emerging challenges 1.4.1.  Quasi-fiscal deficits State-owned enterprises, or the quasi-public sector, have a significant footprint in Kazakhstan. They comprise about 6,000 entities, which amounted to KZT 51 trillion in assets (72 percent of GDP) or KZT 14.4 trillion in revenues (20 percent of GDP). A significant part of the SOEs (88 percent) is represented by entities from the social sector (health care, education, the arts and sports, entertainment). SOEs are also important employers in the labor market: they accounted for 12 percent of the country’s employed population in 2020 (1.1 million people). In 2020, SOEs experienced a positive net income (KZT 456 billion or 0.65 percent of GDP), mostly from the largest management holding companies Samruk-Kazyna and Baiterek. On the other hand, besides these management holding companies, SOEs suffered losses totaling KZT 200 billion. Income losses of SOEs (quasi-fiscal deficits) and their liabilities expose the state budget to risks. Some SOEs purely deliver public services, with revenue coming mostly from the government budget (directly or through ministerial budgets). Other SOEs undertake commercial operations but suffer losses because of poor operational efficiency or having to provide services at below-cost-recovery (or both). But loss- making SOEs are also embedded under holding companies that booked a positive net income, such as Samruk-Kazyna. Kazakhstan’s SOEs also carry liabilities estimated at around 30 percent of GDP, which is substantial.25 Although the Government does not guarantee all SOEs’ debt, the financial difficulties of a systematically large SOE may leave the Government with limited options other than to provide a bailout. Successive bank bailouts have been a significant source of quasi-fiscal exposure and entail risks for the fiscal position. Government intervention in resolving problem banks has been widespread in Kazakhstan since the 2009 banking crisis (see Chapter 2). Out of 42 banks, 24 were bailed out by the Government, liquidated, merged, or reorganized over the 2009–2021 period. Among bailed-out banks, seven were bailed out once, three were bailed out twice, and two thrice. Over 2009–2020, considering the post-recovery income, the net present value of these interventions as of 2020 is KZT 6.3 trillion (around US$14 billion) 25 (Tang, et al. 2020) 36 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH or about 8.5 percent of Kazakhstan’s 2020 GDP.26 The net cost may increase to KZT 7.4 trillion (US$17.5 billion) if the expected recovery is not realized in the future, owing to non-performing bank exposures and losses from subordinated bonds and the quasi-bailout bank deposits of the public sector. Further work should be considered to systematically monitor risks from quasi-public sector losses and liabilities and publish them in the government budget fiscal risks statement. Currently, the responsibility for monitoring the financial performances of SOEs is scattered across the Committee on State Property and Privatization, the Asset Management Policy Department, and various line ministries. The Government’s effort to establish a clear ownership function of SOEs is an important step toward facilitating coordination to present their performance. The assigned agency (a ministerial department or unit) should develop a depository of financial statements that allows the MNE and MoF to develop a fiscal risk assessment from the performances and liabilities of SOEs and extra-budgetary funds. The Government may wish to consider monitoring fiscal risk from changes in profitability, solvency, and liquidity, as Tang et al. (2020) described. Box 1‑5. The nature of public utilities’ quasi-fiscal deficits in Kazakhstan Public utilities generate quasi-fiscal deficits that encompass losses not offset by budgetary subsidies. Many SOEs in utilities (i.e., electricity, heating, water) provide goods and services to consumers at government-regulated prices, typically at below-cost recovery levels. Such losses can pose contingent risks to the budget, especially if they cause companies to underinvest in better and more efficient technologies. Analysis suggests that utility companies in Astana and Almaty have substantial deficits. Because of the lack of accurate information on cost recovery and the QFAs of SOEs, such deficits are not easy to estimate. Therefore, the analysis focuses on these two cities' major utility companies. The estimate covers under-recovery costs, under-collection of revenues, and technical losses above international norms due to under-investment. The estimate suggests that the overall quasi-fiscal deficit for utility companies in those cities in 2020 was about 2.2 percent of GDP which likely required the cities’ governments to cover. Such a large deficit can undermine the sustainability of utility services and cause substantial fiscal risk to the budget. End-user tariffs set below cost-recovery, non-payment of bills, and high technical losses hurt the financial performance of utilities, creating operating deficits and direct or indirect (“hidden”) needs for subsidies from public budgets to sustain their operations. Usually (but not always), budgetary subsidies cover cost-recovery price gaps. Without explicit subsidies, cost-recovery price gaps generate “hidden” (or unrecorded quasi-fiscal) deficits. Due to inefficient operations, hidden deficits also result from bill collection failures or losses. Following a methodology developed by the World Bank, quasi-fiscal deficits are estimated based on the following three components: • Pricing gaps: losses from end-user tariffs set below cost-recovery rates. The cost-recovery tariff must include operations, maintenance, and interest on outstanding debt. • Collection inefficiency: losses from differences between billed and collected revenue. • Technical inefficiency: losses above normal technical losses from inefficient operations. 26 The net cost of the interventions is calculated by subtracting from the KZT 8.2 trillion incurred expenses, KZT 0.8 trillion of realized recovery, and KZT 1.1 trillion in unrealized recovery. 1. STRENGTHENING THE FISCAL FRAMEWORK FOR INCLUSIVE AND RESILIENT GROWTH 37 This methodology's Preliminary estimates suggest that the deficits are significant mainly in the natural gas and gasoline sectors due to low end-user tariffs. In electricity, drinking water, and district heating, the estimated quasi-fiscal deficits are not significant because CHPs usually generate heating, and electricity tariffs cover the deficit of the heating part of the business. However, the drinking water and district heating estimates cover only the cities of Almaty and Astana. In the natural gas sector, producer prices—the prices subsoil users (i.e., upstream producers) receive—are individually negotiated between producers and KazTransGas (KTG), the natural monopoly in Kazakhstan’s single-buyer model, and ultimately approved by the Ministry of Energy. KTG wields significant power in negotiating gas prices. Because low-regulated end-user prices pressure all aspects of the domestic gas value chain, producers often sell gas at a price that barely covers costs or forces them to incur a loss. KTG generates financial losses in its basic business of selling gas to domestic consumers. However, since 2016, the company has generated positive net income in its overall operations. The company's profitability turnaround was not a result of a major improvement in its main domestic gas sales business activity. Still, it was due to additional revenues from expanded gas exports to China and higher gas transit. Expanding the distribution network from the gasification drive means that the unprofitable segment of KTG’s business will continue to grow. In electricity, a quasi-fiscal deficit from the under-recovery of costs mostly occurs in the distribution grid. Tariffs for power plants cover operational costs plus a 12 percent margin, but regional distribution companies face low tariff levels. This is also the result of fragmented institutional regulation of the energy sector. The Ministry of National Economy regulates heating tariffs and transmission, distribution, and retail electricity prices, whereas the Ministry of Energy regulates the electricity generation tariff. In the oil sector, the Government obligates oil-producing companies to supply a part of the produced crude oil to the domestic market at a cost that may be considerably lower than world prices and may even be set at the level of the cost of oil production. Table 1‑4. Estimated quasi-fiscal deficits of utility SOEs in Almaty and Astana (2020, in percent of GDP) Under-recovery of Under-collection of Technical losses Total QFD cost revenue above norms Natural gas 0.71 0.03 0.02 0.76 Electricity 0.05 0.07 0.05 0.17 Drinking water 0.11 0.01 0.0 0.12 District heating 0.19 0.10 0.0 0.29 Gasoline 0.85 0.02 0.0 0.87 Total 2.21 Source: World Bank staff estimates. 38 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 1.4.2.  The green transition and climate change Global decarbonization may gradually affect Kazakhstan’s oil sector revenue, posing challenges for the fiscal framework. Kazakhstan’s oil production is projected to reach 104 million tons by 2030, 14 percent higher than in 2019, with the completion of several large oil projects. While global economic recovery may sustain demand for oil in the short run, continuing global decarbonization exposes Kazakhstan’s oil and gas sector to higher uncertainty in the future. A decline in investment in the sector might materialize sooner as international oil companies face more pressure from shareholders to commit to a clean energy transition. Decarbonization may thus generate a long-term fiscal challenge for Kazakhstan because of the importance of oil revenue in meeting the Government’s financing needs. As a fossil-fuel-dependent country, an accelerated global decarbonization trend could substantially affect Kazakhstan’s fiscal position. In a world where countries are ramping up decarbonization, Kazakhstan could face lower exports and budget revenue from oil. Modeling using the World Bank’s macro- fiscal model (MFMod) shows that a gradual decline in the oil price may reduce real GDP per capita by 2.5–3 percent compared with the baseline projection by 2050 (Figure 1‑22). At the same time, assuming that higher financing needs will be financed by government debt, lower revenue can potentially increase the Government’s debt-to-GDP ratio by 3 percentage points by 2050 compared with the baseline projection.27 The transition toward net-zero emissions requires higher public investment. The transition to meet the 2060 net zero emission targets the Government sets requires a substantial ramp-up in investment in clean electricity, infrastructure, and the end-use sectors. The role of private capital in these projects will be crucial because public resources are scarce. However, many green projects lack sufficient scale, returns are only expected over a long-term horizon, and perceived risk is high or unclear, decreasing private investors’ appetite. Public budgets will thus need to fill the gap, along with efforts to energize green finance. A recent study (World Bank 2022) suggests that the additional public investment needed for climate mitigation to reach net-zero emissions from 2025 to 2060 is about 0.6 percent of GDP per year (Figure 1‑23). Figure 1‑22. Real GDP and government debt Figure 1‑23. Potential financing need for climate projection under global decarbonization adaptation to reach net-zero commitment (deviation from the baseline projections) Source: World Bank staff calculations using MFMod model and Source: World Bank (2022). baseline and scenario used in World Bank (2022). 27 The results from modeling exercise using World Bank MFMod for Kazakhstan. The same model is also used for the Climate Change Development Report for Kazakhstan (World Bank 2022). The projection on fiscal impact is in line with an earlier study focusing on the fiscal impact from green transition (EBRD 2018). 1. STRENGTHENING THE FISCAL FRAMEWORK FOR INCLUSIVE AND RESILIENT GROWTH 39 Kazakhstan faces the risk of economic loss from climate shocks, and public finances should incorporate the need to spend on climate adaptation. Drought and floods, both exacerbated by climate change, generate average annual losses of 1.2 percent of GDP. These losses are expected to increase under a moderate climate change scenario to 2.2 percent of GDP. The Government should consider using public finance to facilitate spending on climate adaptation. A study finds that preventive measures through investment in adaptation will deliver better outcomes on GDP than corrective actions such as disaster relief.28 Estimates suggest that annual costs of adapting to climate-related hazards are estimated at US$665.4 million, or 0.4 percent of GDP, in four priority areas: (i) water resources management; (ii) dryland agriculture crop production; (iii) improving early warning system; and (iv) improving infrastructure resilience.29 As discussed in Chapter 2, government budgeting and planning must develop the capacity to appraise and identify priority public investments for climate adaptation. Because climate shocks such as natural disasters have a fiscal impact, it is also important to gradually integrate climate risk scenarios into macro-fiscal projections in the Government’s annual budget report and take early action to address the challenges (see (Pigato 2019) on the role of fiscal policies in laying the foundation for low-carbon and climate-resilient growth). While the budget provides contingency reserves, the amount reserved to help offset the impact of natural disasters, including climate-related disasters, is uneven and is likely limited to offset the economic losses from climate shocks. The Civil Protection Law of Kazakhstan establishes relevant contingency reserves, and several contingency reserves are available at the local and national levels (World Bank 2019). For instance, a national reserve fund supports people’s livelihoods after a man-made or natural disaster. A contingency reserve is also established by the provisions of the Committee of Emergency Situations (for disaster response and recovery, such as medical assistance and rescue operations). Local government reserve funds are also available, though their allocations should not exceed 2 percent of the relevant budget revenues. Table 1‑5. Reserve funds allocated annually for disaster relief 2016 2017 2018 2019 2020 Reserves for the elimination of natural and 58.5 2.3 8.8 68.4 44.2 man-made emergencies (US$ million) As percentage of GDP 0.004 0.001 0.005 0.04 0.03 Several other options are useful for managing the fiscal impact of climate-related disasters. The affected population must be able to access social protection services. One way is through the social protection system, which the Government is currently working to improve its coverage and targeting. The improvement is expected to: (i) provide better social protection responses and more effectively communicate the availability of those benefits to the population; (ii) promote the uptake and expansion of climate-related disaster risk insurance products; (ii) expand disaster risk insurance products that can help address financing needs and reduce risks to firms and businesses; and (iv) promote more effective financial management of disaster risks, which calls for a mix of approaches. For instance, high-frequency and low-severity events (such as seasonal flooding) can be cost-effectively managed through adaptation measures that reduce risk, while low-frequency and high-severity events are usually more effectively managed through risk transfer, such as insurance.30 28 A modeling work (Catalano, Forni and Pezzolla 2020). 29 Estimates by UN-ESCAP https://rrp.unescap.org/country-profile/kaz#paragraph-id--23873 30 A study found that countries with lower levels of insurance penetration faced larger declines in economic output and more considerable increases in fiscal deficits in response to disasters than countries with higher levels of insurance penetration (Melecky and Raddatz 2015). 40 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Subsidies, green revenue, and expenditure in the state budget are important to inform a productive public debate. Indeed, there are estimates of subsidies on fossil fuels that do not rely on budget information (such as by the International Energy Agency). But a lack of clarity in the amount allocated for energy and fossil fuels subsidies in Kazakhstan’s government budget undermines transparency in allocating public expenditures. The budget also does not provide information about transfers to SOEs in the heating and power sectors to compensate for their potential losses from selling output below cost-recovery prices. Having these figures reported in the state budget can help the Government to communicate the policy choices and trade-offs from continuing with energy and fuel subsidies. Broadening the excise coverage to all fossil fuels can be a potentially good early action to improve budget resilience against green transition. As discussed in Chapter 3, the current revenue base is concentrated in corporate income tax and value-added tax (VAT), mostly collected from a small group of large firms in a limited set of sectors. Kazakhstan can consider applying excise on all fuels to raise fiscal revenue and complement the Emission Trading System in aligning the incentives to reduce carbon emissions. A recent World Bank study (World Bank 2022) suggests that a gradual increase in excise rates on all fuels to 25 percent of the level specified under the EU directive will gradually increase tax revenue by up to 4 percent of GDP by 2030 (Figure 1‑24). The projections also suggest that recycling 40 percent of the excise revenue as cash transfers for the bottom 40 percent of the income distribution could offset the negative impact of a higher fuel price on their consumption (Figure 1‑25). As discussed in Chapter 5, improving the targeting of social assistance is critical for implementing an efficient transfer program. Figure 1‑24. Projected additional fiscal revenue Figure 1‑25. Projected impact of broadening and from excise on all fossil fuels relative to baseline increasing excise on fossil fuels: Relative mean consumption effect (% consumption in 2030) Source:  World Bank (2022). Calculation using World Bank/IMF Carbon Pricing Assessment Tool (CPAT). 1. STRENGTHENING THE FISCAL FRAMEWORK FOR INCLUSIVE AND RESILIENT GROWTH 41 1.5. Recommendations Simplify the fiscal rules and strengthen their monitoring framework Introduce a well-designed escape clause and eliminate targeted transfers from the NFRK to the budget. In the face of exceptional circumstances, the usual practice is to amend the approved budget and switch to a crisis budget. A case in point is the COVID-19 pandemic, which triggered the suspension of the fiscal rule. The benefit of an escape clause would be better framing the scope of discretion by introducing more transparent criteria for departures from the fiscal rule. Operationally, an escape clause can be tied to indicators such as the deviation of GDP growth from its long-term trend or a rise in the unemployment rate. Guidance on returning to the rules after activating the escape clause is also important. The framework should include a correction mechanism to manage deviations from the rules. A well-designed escape clause would be a superior option to discretionary targeted transfers from the NFRK to the budget because it would reinforce fiscal policy’s transparency and, thus, the credibility of fiscal policy. Consider establishing an independent fiscal institution to monitor the application of the fiscal rule. An independent fiscal institution with well-circumscribed responsibilities could contribute to improved fiscal performance. Independent fiscal institutions (IFIs) have become widespread among several EU member states and emerging and developing economies such as Colombia, South Africa, and Peru.31 IFIs are established with a mandate to assess fiscal policies, plans, rules, and performance. IFIs can be tasked with making independent macroeconomic forecasts or providing assumptions or projections of key variables or parameters on which budget projections can be based. They can also be tasked with making independent revenue and expenditure projections, whether based on current or prospective policies. Assessments of long-run fiscal implications of tax and spending initiatives can also contribute to improving the transparency of budgetary decisions. Evidence suggests that the presence of an IFI is associated with more accurate and realistic fiscal forecasts and greater compliance with fiscal rules (Beetsma, et al. 2018). The IFI could be responsible for monitoring the implementation of the fiscal rules and setting the long-term benchmark oil price used as a reference for the formulation of the budget, following international best practices (Kopits 2011). The case of the Autonomous Fiscal Council of Chile could be a useful example (Larrain and Schmidt- Hebbel 2019). Kazakhstan’s authorities may consider several options, such as tasking the Agency of Strategic Planning and Reform to perform such function or setting up an IFI as an independent partner to support the Parliament (similar to the US Congressional Budget Office). Broaden the institutional coverage of the fiscal framework by including extra- budgetary units at the central and local levels and public corporations. Include spending by the State Social Insurance Fund (SSIF), the State Health Insurance Fund (SHIF), and the Pension Fund in the consolidated government budget reporting and developing a comprehensive approach to systematically monitor quasi-fiscal activities and their overall contribution to the fiscal stance. The overall balance rule applies to the state budget, which comprises the central government and local budgets, while the non-oil balance rule applies to the central government budget (Republican Budget). The fiscal rules exclude extra-budgetary units and public corporations at the central and local levels. Much of the support to the banking system in recent years has been granted off-budget. Part of the housing 31 The number of IFIs has more than quadrupled since the global financial crisis, with 49 countries identified by the IMF as operating IFIs in 2021, while in the OECD zone the proportion rises to three out of four member countries https://www.imf.org/en/Data/Fiscal/fiscal- council-dataset and https://oecdecoscope.blog/2022/04/14/how-can-independent-fiscal-institutions-support-the-resilience-of- public-finances/?print=pdf 42 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH support is provided through interest rate subsidies and construction managed off-budget by subsidiaries of a national management holding company. Similar arrangements are in place for agriculture. Public financial and non-financial corporations carry out a wide range of activities for public policy purposes. SOEs and national holdings fund some social and infrastructure projects. Some extra-budgetary activities are partly funded through budget transfers, but others are not and are thus quasi-fiscal. The COVID-19 crisis forced the Government to spread stimulus measures around various off-budget funds and a few SOEs undertaking quasi-fiscal programs. The lack of a comprehensive government fiscal account complicates properly assessing the fiscal stance. There is thus a risk of deficit-biased policies through excessive stimulus, which would damage macro stability. To address these weaknesses, the Government should form a comprehensive fiscal account by including various activities outside the budget. This would involve properly identifying the institutional sector of extra- budgetary units, setting appropriate reporting requirements, and extending the coverage in fiscal reports. Broader coverage and transparency would help in designing an appropriate fiscal stance, thus contributing toward more effective policy implementation. Strengthen risks monitoring and government debt management Develop and publish an explicit debt management strategy, i.e., a medium-term debt strategy (MTDS). An MTDS would inform the public and parliament about the costs and risks of debt financing that are consistent with the Government’s debt management objective (typically, ensuring that the Government’s financing needs can be met at the lowest possible cost and risk). A debt management strategy and disclosure of borrowing plans increase investor certainty and are expected to lower the Government’s borrowing costs in the long run. In addition, rating agencies, in their sovereign ratings, do consider whether a debt management strategy is in place and is being implemented according to its premises. Systematically monitor fiscal risk from contingent liabilities stemming from the performance of QFAs and SOEs. Consider including in the fiscal risk report/statement to the budget an assessment of the contingent liabilities from QFAs. First, the Government should consider monitoring deficits in certain SOEs reflecting the delivery of services at below-market prices, such as in energy (electricity and heating) and railway transport. The Government may also want to consider developing an approach to monitor aggregate fiscal risk from potential financial bailouts. Second, as part of its SOE reform, the Government should consider assigning a ministry, such as the MOF, responsible for monitoring contingent liability risk from SOE operations. Not all SOEs’ operations are profitable or sustainable, and the Government’s plans to strengthen the ownership function and develop aggregate SOE reporting are major positive steps toward transparency. In this context, the Government could assign a ministry, such as the MOF, to monitor risks from SOE operations, starting with SOEs that account for significant assets and employee numbers, such as those managed by holding companies Baiterek and Samruk-Kazyna. Develop transparency in reporting fossil-fuel subsidies and realistic assessments of possible budget pressure to offset the economic loss from physical damage brought about by climate change and adjust contingency funds accordingly. The budget must explicitly report subsidies and incentives granted to consumption and energy investments, including fossil fuels. This transparency helps align policies with Kazakhstan’s commitment to achieve net-zero emissions by 2060 and inform the public on the potential benefits and trade-offs between spending on those subsidies and incentives vs. other uses of funds. Kazakhstan must also develop the capacity to assess fiscal exposure to climate shocks and natural disasters. This assessment will allow the Government to propose an appropriate contingency fund. 1. STRENGTHENING THE FISCAL FRAMEWORK FOR INCLUSIVE AND RESILIENT GROWTH 43 2. Managing the Fiscal Footprint and Risk from Quasi-Fiscal Activities 44 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH KEY POINTS • Kazakhstan has established extrabudgetary funds to manage and deliver specific services the state provides through the public system, such as social insurance, health insurance, and pensions. But Kazakhstan also uses extrabudgetary funds and SOEs for quasi-fiscal activities (QFAs), i.e., implementing government programs whose spending is either partly or not reflected in the government budget. • The increase in QFAs can undermine fiscal discipline and expose the budget to risks. Some QFAs also distort resource allocations and undermine market discipline. While extrabudgetary funds (EBFs) may be quicker in delivering programs than regular government departments, their performance, and management are not subject to public and parliamentary oversight. • The Government should consider monitoring the size of QFAs and risks emanating from extrabudgetary funds operations in the fiscal framework. The authorities should also consider strengthening the public oversight of extrabudgetary funds to ensure their operations align with their core objectives. Quasi-fiscal risks from recurrent bank bailouts merit special attention. 2.1. The Prominence of Quasi-Fiscal Activities in Kazakhstan Governments often use extrabudgetary funds (EBFs) and state-owned enterprises (SOEs) to deliver economic and social programs classified as quasi-fiscal activities (QFAs). EBFs are a category of government-owned organizations and accounts that, by definition, falls beyond the scope and control of traditional budgetary rules and procedures. In addition, governments also often use SOEs to deliver programs directed by the Government at non-market prices and terms (see Box 2‑1. Quasi-fiscal entities and extra-budgetary funds). These entities are performing QFAs because they are implementing programs on behalf of the Government even though their spending is not part of government expenditure. While justifiable reasons exist to use these entities to deliver public programs, their negative effects should be addressed. EBFs and SOEs can deliver activities on behalf of the Government relatively faster, as they have lower administrative costs associated with budgeting and planning. In most cases, the legal form of a private company provides them with more flexibility and economic freedom in mobilizing human resources than regular government departments. However, there are also significant downside risks. First, their off-budget spending can obscure the Government’s fiscal stance and complicate macroeconomic management. When QFAs become significant, the budget balance is no longer a reliable indicator of the Government’s fiscal position. Public financial statistics without integration of QFAs do not accurately reflect the actual size of the Government and do not capture implicit contingent liabilities that may arise in the absence of budgetary control. Second, they are also a common source of fiscal risks. A cross-countries study estimates that, between 1990-2014, on average, government bailouts of troubled quasi-fiscal entities (QFEs) cost 3 percent of GDP and 15 percent of GDP in the most extreme cases (Bova, et al. 2016). Lastly, their implicit subsidies and function as fiscal agents to deliver or purchase services below market prices distort the allocation of resources. EBFs and certain SOEs with a public policy mandate should not have their financial management, governance, and operations subject to less scrutiny and accountability than the government budget. 2. MANAGING THE FISCAL FOOTPRINT AND RISK FROM QUASI-FISCAL ACTIVITIES 45 In Kazakhstan, using QFAs to implement off-budget spending has intensified since the 2016 banking crisis. As described below (Section 4.2.4) and in Chapter 1, gross spending by major entities on QFAs increased from about 1.1 percent of GDP in 2011–2015 to 3.7 percent in 2016–2021. This increase was driven mostly by COVID-19 crisis measures introduced after the 2016 banking and pandemic crises, such as supporting bank bailouts, credit subsidies, health spending, and social payments. But it also reflects a prolonged use of EBFs and SOEs to deliver government programs. Unlike advanced economies, Kazakhstan does not have a well-established institutional framework for governing QFAs and integrating them into the budgetary process. Box 2‑1. Quasi-fiscal entities and extra-budgetary funds Extra-budgetary funds (EBFs) as a part of “extra-budgetary entities are institutions that are engaged in extra-budgetary transactions, may use extra-budgetary accounts, may have their governance structures and, often, a legal status that is independent of government ministries and departments” (Allen and Radey 2010). Quasi-fiscal entities (QFEs) undertake QFAs: “any activities undertaken by state-owned banks and enterprises, and sometimes by private sector companies at the direction of the government, where the prices charged are less than usual or less than the “market rate.”32 The scope of QFEs is broader than the EBF as QFEs include SOEs in the energy, power sectors, etc. Kazakhstan fixed in the legislation a definition of quasi-public sector entities, which includes state enterprises, limited liability partnerships, joint-stock companies, including national management holdings, national holdings, national companies, in which the state acts as founder, participant, or shareholder, as well as subsidiaries, dependents and other affiliated legal entities (Budget Code). 2.2. Constellation of extra-budgetary funds and quasi-fiscal entities in Kazakhstan There are several extra-budgetary funds and quasi-fiscal entities in Kazakhstan, both of a social and economic nature. These include the State Social Insurance Fund (SSIF), the State Health Insurance Fund (SHIF), the Unified Accumulative Pension Fund (UAPF), the Kazakhstan Sustainability Fund (KSF), the National Fund of the Republic of Kazakhstan (NFRK). There are also quasi-fiscal entities under Baiterek that implement government programs. These EBFs and QFEs receive capital from the state to undertake certain activities on behalf of the Government. This section reviews their operation and magnitude in the social, economic, and financial sectors. SOEs also generate quasi-fiscal deficits that encompass losses not offset by budgetary subsidies. Many SOEs provide goods and services to consumers at low prices, regulated by the Government, typically at below-cost recovery levels. Such deficits are not easy to estimate because of the lack of accurate information on cost recovery. Such quasi-fiscal deficits can create a contingent risk to the budget. Preliminary estimates of the quasi-fiscal deficits of the SOEs are summarized in Chapter 1 for a sample of key utility sectors of local governments’ utility firms in Astana and Almaty. The estimated quasi-fiscal deficits cover the under- 32 https://internationalbudget.org/wp-content/uploads/Looking-Beyond-the-Budget-3-Quasi-Fiscal-Activities.pdf 46 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH recovery of costs, under-collection of revenues, and technical losses above international norms due to under-investment. The overall quasi-fiscal deficit is estimated at 2.2 percent of GDP, which was likely covered by the respective cities’ budgets. 2.2.1 Social sectors State Social Insurance Fund (SSIF) The State Social Insurance Fund (SSIF) was established in 2004 by the Government to implement a compulsory social insurance system through the safe and targeted use of social contributions. The National Bank of Kazakhstan (NBK), the country’s central bank, oversees the investment portfolio of SSIF. The activities of SSIF are not included in the Republican Budget, which is approved by the parliament and are not discussed in parliamentary committees. The main source of funds for SSIF is the collection of social contributions—75 percent of SSIF revenue, on average, in 2017–2019. On average, one-half of the SSIF’s spending goes to programs (social payments to participants) and another half to financial instruments, except during the pandemic in 2020, when there were massive payments to citizens. Up until 2020, revenues and spending were about equal. According to the annual audit report in 2020, during the pandemic, SSIF benefit payments (KZT 752 billion) exceeded the volume of contributions by three times (KZT 273 billion). SSIF has been involved in QFAs to support loss-making SOEs through direct support and investments. In 2018 SSIF provided one-time support to the loss-making KazAgro of around KZT 60 billion.33 As for SSIF investments, the Government had introduced amendments to the legislation and the SSIF’s charter that would allow the SSIF to invest in Kazagro and Kazakhstan engineering, both loss-making SOEs.34 Figure 2‑1. Organizational chart of the SSIF Source: World Bank staff analysis. 33 https://tengrinews.kz/kazakhstan_news/60-milliardov-tenge-fonda-sotsstrahovaniya-napravyat-kazagro-335964/. 34 https://adilet.zan.kz/rus/docs/P2000000196. 2. MANAGING THE FISCAL FOOTPRINT AND RISK FROM QUASI-FISCAL ACTIVITIES 47 State Health Insurance Fund (SHIF) The Government established a Social Health Insurance Fund (SHIF) in 2016. The Fund’s mission is to ensure the financial protection of every citizen of the country from the unforeseen high costs of medical care by improving its availability and quality. The NBK also oversees the investment activities of SHIF assets. The activities of the SHIF are not included in the Republican Budget, which is approved by parliament. The SHIF implements two main programs. First is the mandatory state insurance for free medical care, fully funded through the government budget. Second is social health insurance, for which about 47 percent of the spending for 2021 was contributed from the government budget.35 The SHIF had been accumulating contributions until 2020 when it started spending on benefits for participants, spending around 73 percent of its revenues in 2020. In 2020, 15.9 million people, or 85 percent of the population, were enrolled in the mandatory medical insurance system. According to the annual report, in 2020, KZT 1,411.9 billion was allocated to purchase medical services: KZT 1,010.9 billion from state budget expenditure and KZT 401 billion from the compulsory health insurance system.36 According to the Law on Compulsory Health Insurance, the assets of the SHIF can be invested in government securities, deposits in the NBK, bonds of international financial institutions, bonds of SOEs, and REPO. The investment revenues of the SHIF represent, on average, 5.5 percent of total revenues. Figure 2‑2. An organizational chart of the SHIF Source: World Bank staff analysis. Unified Accumulative Pension Fund (UAPF) The Unified Accumulative Pension Fund (UAPF) is the mandatory state-owned defined contribution pension fund established in 2013 during the reform of the pension system in Kazakhstan. The fund was founded based on the State National Pension Fund and combined all the existing pension funds in the country at that time. The main objective is to ensure higher efficiency of operations, transparency, and reliability, as well as certain guarantees from the state for the safety of pension deposits. The UAPF receives a contribution from employers and employees (Figure 4.3).37 The founder and shareholder of the UAPF is the Government, represented by the Committee of State Property and Privatization of the MoF. 35 SHIF Annual Report 2021. 36 SHIF Annual Report 2020. 37 The UAPF accepts three types of contributions: mandatory pension contributions (MPC), mandatory professional pension contributions (MPPC) and voluntary pension contributions. 48 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH The NBK manages the UAPF’s investment portfolio per the Investment Declaration of the UAPF. Investments are limited to highly liquid, low-risk financial instruments. These instruments include Republic of Kazakhstan government securities, the financial instruments of domestic companies and rated second-tier banks, and bonds issued by international financial institutions. The large pension assets of the UAPF position it as an important institutional investor in the capital markets. Pension savings in the UAPF amounted to ZKT 13.7 billion (16.1 percent of GDP) as of the end of 2021. The UAPF plays an important role in the domestic capital market, particularly through its purchases of the Government’s treasury to finance the budget deficit. As of January 2023, about 50 percent of the UAPF investment’s portfolio is in MoF’s securities, a substantial increase from 39 percent in the previous year.38 However, the difference between the costs and fair values of the investments might suggest that the financing could be at below-market interest rates.39 In 2021 the Government instructed the UAPF to deliver a major QFA by granting options to pensioners to use a part of their pension savings to purchase residential accommodation or to pay for medical treatment. As a result, pension payments increased significantly to KZT 2.9 trillion (3.6 percent of GDP) in 2021. Figure 2‑3. Organizational chart of the UAPF Source: World Bank staff analysis. The SSIF, the SHIF, and the UAPF face common challenges: (i) risk of spending beyond core tasks; (ii) limited accountability on the investment decisions; and (iii) limited systematic parliamentary hearings on these funds’ program and performances. On the other hand, the funds’ “Joint Stock Company” status implies they need to comply with the reporting requirements of a public corporation and likely spend less to track and control the efficiency of resources. To ensure effective risk management, the International Social Security Association (ISSA) recommends that social security funds ensure an actuarial review of the fund’s financial viability. The auditor and actuary reports should be made public to increase stakeholders’ confidence in the scheme’s effective management and as a commitment to openness and transparency. 38 Based on UAPF JSC information on pension assets investment management, January 01, 2023. 39 UAPF’s 2021 statement on Pension Plan Assets, note #18 on Related party transaction, suggests that the transaction on the MoF’s securities accounted at fair value was lower than the that accounted at cost. 2. MANAGING THE FISCAL FOOTPRINT AND RISK FROM QUASI-FISCAL ACTIVITIES 49 2.2.2 Quasi-Fiscal Activities by development finance institutions and state-owned enterprises Baiterek Holding This holding was established by separation from Samruk-Kazyna Holding in 2013, with the main objective of supporting national economic development. The Government is the sole shareholder of the Baiterek and directly participates in the management of the Baiterek Holding. The prime minister is the chairman of the board of directors. Baiterek has been following five key directions to channel state preferential financing to business and the population: (i) large companies financing (industrialization and other state programs); (ii) SME financing; (iii) affordable housing; (iv) support for export of non-primary products; and (v) innovation development. With the merger of KazAgro with Baiterek 2021, the holding also provides financing support to the agriculture sector. The Development Bank of Kazakhstan (DBK) and the Housing Construction and Savings Bank (Otbasy Bank) are the two main subsidiaries of Baiterek Holding, whose size is considerably larger than others. The DBK primarily involves the state concessional financing of large and medium-sized companies under industrialization and other state programs. Otbasy Bank is involved mainly in the state concessional financing of mortgage loans and is a key part of state policy for affordable housing. The dual role of Baiterek as a development financial institution and fiscal agent raises two major issues. By competing directly in financial markets on unequal terms, Baiterek distorts the credit market and crowds out private capital from entering or scaling up certain financial services. The existing governance setup for Baiterek does not foster political independence in decision-making, holistic risk management, and accountability for development impacts (Melecky, Fontan, et al. Forthcoming). Samruk-Kazyna Joint Stock Company (JSC) National Welfare Fund Samruk-Kazyna (SK) was established in 2008. SK was created through the merger of JSC Kazyna Sustainable Development Fund (a holding company for national DFIs) and JSC Kazakhstan Holding for the Management of State Assets Samruk (a holding company for SOEs) and the additional transfer by the state to the SK of its ownership rights in some government-owned enterprises. SK performs four different functions under a single national holding: (i) holding company of strategic SOEs; (ii) serving as a sovereign wealth fund; (iii) performing as a DFI; and (iv) financing state expenditures. While all such functions are important for the development of the economy of Kazakhstan, it is highly questionable whether they should be combined into one single state company. The Government also occasionally uses SK to finance state expenditures. Between 2016-2021, SK spent about 0.14 percent of the GDP on social projects set by Government decrees. For example, SK, following instructions from the Government, financed various initiatives of “Nazarbayev University,” the restoration of damaged social facilities in the city of Arys, the construction of large facilities in Turkestan, promoting mass sports, and many other similar operations. In addition to these activities, SK also provides tariff subsidies on the prices of electricity, rail transport, and oil and gas products. Instead of relying on capital transfers from the Government’s budget, SK finances state programs out of its finances, including the proceeds from ownership shares in large oil projects. This financing of state expenditures through SK and its entities bypasses the scrutiny of parliament and is not subject to public accountability. 50 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 2.2.3 Quasi-fiscal activities in the financial sector Kazakhstan Sustainability Fund (KSF) The Kazakhstan Sustainability Fund (KSF) serves as the NBK’s vehicle to support financial sector stability and the Government’s fiscal agent managed by the NBK. The fund has two-tier management: the Board of Directors (BoD) managing the fund and the Management Board supervising the BoD. There are five members in the BoD, two of which are independent directors.40 As the sole shareholders, the NBK appoints BoD and Management Board members. The KSF’s assets in 2021 were valued at KZT 3.7 trillion, or about 4.8 percent of GDP. In 2020, the KSF implemented several (government) state programs of KZT 3.3 trillion, or about 5 percent of GDP; about 37 of those programs supported the housing mortgage program (7-20-25 program).41 Currently, there is no estimate of the implicit subsidy from KSF interventions. Nevertheless, with many such interventions, the KSF’s programs will likely have a sizable impact on the fiscal stance. Figure 2‑4. Organizational chart of the KSF The classification and definition of the nature of the KSF are more complex than the other funds. The KSF positions itself as a commercial company but simultaneously acquires bonds of second-tier banks with a B- and B ratings at a remuneration rate of 4 percent, which is three times lower than the market rate. At the same time, the KSF places bonds on the domestic securities market at a rate of 8.5 percent or more, which are also purchased by commercial banks. Given the difference in those rates, monitoring the KSF’s liability risk would be important. In 2021, Fitch Rating affirmed KFS’s Long-Term Foreign and Local-Currency Issuer Default Ratings (IDR) at ‘BBB’ with a “Stable” outlook. However, the agency also noted that the KSF’s default could significantly impair investor confidence in the credibility of the NBK and other government- related entities. Problem Loan Fund (PLF) The Problem Loan Fund (PLF) is a separate institutional unit implementing government policy to rescue the banking sector from non-performing loans (NPLs). The PLF was created in 2012 to support national 40 https://kfu.kz/deyatelnost/upravlenie/ 41 Kazakhstan Sustainability Fund annual report, 2020, pages 21–22. https://kfu.kz/wp-content/uploads/2021/12/godovoj-otchet- kfu-2020.pdf 2. MANAGING THE FISCAL FOOTPRINT AND RISK FROM QUASI-FISCAL ACTIVITIES 51 economic recovery. From 2012 through 2017, the PLF was managed by the NBK and then transferred to the Government. The Government bears all risks and losses from the PLF’s activities and provides financial support to PLF from the Republican Budget. The PLF also issues securities, non-listed on the stock exchange, acquired by other QFEs and the NBK. The bulk of the financial support from the PLF is implemented by purchasing assets from banks at a price above their market/fair value and by placing deposits in banks on non-market terms. In 2017, capital transfers to problem banks amounted to KZT 2.539,7 billion, equivalent to 5.4 percent of GDP. In 2018 and 2019, capital transfers to banks from the PLF were lower, equivalent to 0.78 and 0.70 percent of GDP, respectively. The lack of clarity on the limit of the PLF’s purchase of problem assets from banks can expose fiscal risk to the budget (see Section 4.3). 2.3.2 Fiscal exposure from bank bailouts Kazakhstan has experienced frequent Figure 2‑5. Frequency of government interventions in the government interventions to address banking sector banking sector issues. A total of 42 banks operated in Kazakhstan be- tween January 2009 and August 2021, ranging from a maximum of 39 banks operating in 2010 to a minimum of 23 banks active as of August 2021. Of all banks observed from 2009–2021, 24 were subject to government interven- tion or some form of private/market solution, such as liquidation, merger, or other resolution events.42 The cost dif- fers depending on the type of govern- ment intervention to rescue distressed banks: equity injections, purchase of non-performing exposures, and the Source: Mare, Melecky, and Muria (2022). provision of funding in the form of debt financing or bank deposits. The types of intervention have varied over time, with equity injections being used to restore bank financial stability in the early years (Figure 2‑5). Since 2013, the acquisition of non-performing exposures (seven interventions) and the provision of debt or deposit funding (nine interventions) have been used to rescue ailing banks. QFEs played a crucial role in the structuring of bank bailouts. The most recent purchases by the PLF of bank NPLs at 100 percent of their origination price—while their market price can often be lower than 30 cents on the dollar—have been the costliest bailout options to the fiscal space (Figure 2‑6 panel A).43 The recoveries from NPL purchases have been minuscule, and the prospects for further recoveries are deteriorating with time. 42 We classify as a government intervention any action undertaken by the Government of Kazakhstan that involves public sector expenditure associated with helping troubled banks. 43 The average fair value of the portfolios (based on independent appraisal) to the purchase price was 9 percent, the last transaction with the ATF in 2020 had the ratio of 15 percent. 52 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH The use of debt bailouts through the NBK’s Kazakhstan Sustainability Fund has been the second-most- recent type of bailout used by the Government. Injecting public money against senior debt issuance by the banks is a more disciplining bailout for bankers but a very soft one—especially if the debt is underpriced. For example, the bailout of the ATF Bank in 2020 was structured based on a debt with a coupon of 0.1 percent—almost no return for taxpayers’ money. A similar bailout approach has been applied through Samruk-Kazyna44 and PLF deposits, which were placed in banks at well below the market rate and, for example, Kazcom Bank’s case, ultimately with an interest rate as low as 0.1 percent. The expenditures for government bailouts of troubled banks first spiked in 2009 to almost 15 percent of total tax revenues in the same year (Figure 53, Panel B). A few years later, in 2014 and 2015, total government support spiked to over 4 percent of total tax revenues. The cost of government interventions was particularly high in 2017 when the rescue of one bank entailed acquiring a large non-performing exposure to the tune of KZT 2.6 trillion.45 In 2017, bailout costs reached around 48 percent of government tax revenues. In 2018 and 2019, the cost of bailouts was high by historical standards, at around 7 percent of tax revenues each year. In 2020, bailout costs through NPL purchases at 100 percent of the loans’ origination price and financing of underpriced bank bonds represented around 3 percent of tax revenues. Figure 2‑6. Cost of bank bailouts A: Cost of government intervention (% of total tax B: Expenditures on government intervention by revenue) bailout agent and type of instrument Source: Mare, Melecky, and Muria (2022). Note: Panel B reports the cumulative government expenditures on bailouts in a given year. Because of prevailing bank bailout practices, the public bill for rescuing troubled banks reached KZT 6.3 trillion over the 2009–2020 period. The expenses in terms of present value as of 2020 were high at KZT 8.2 trillion (equivalent to 11.6 percent of GDP in 2020), mostly thanks to the PLF purchases of 44 The equity injections by Samruk Kazyna (SK) could be considered as a fairer and more equitable bailout option. Such capital injections could reduce/dilute the ownership stakes of bankers that brought the bank into distress and require them to pay back dividends on the injected equity. The budget may thus earn some return to cover part of the bailout costs. During 2009–2010, SK performed several equity bailouts. Depending on the solvency situation and the amount of capital increase required, SK invested and gained varying equity stakes in the troubled banks, while the original shareholders experienced a proportionate decrease in their stakes. Only the Halyk Bank survived after these bailouts and its majority shareholder acquired the stakes from SK in 2011. Through these bailout interventions and SK exits, the then majority stakeholder increased its stake in the Halyk Bank, while the other shareholders were diluted. In the six bailouts through equity injections that took place between 2009–2010, the Government took ownership stakes of between 20 and 82 percent in the troubled banks that lasted on average 3.2 years and produced an average annual dividend of 1.5 percent. However, the internal rate of return—considering capital gains—averaged -3 percent. For more details see Mare, Melecky and Murina (2022). 45 The cost of bailout per individual banks is available in Mare, Melecky and Murina (2022). 2. MANAGING THE FISCAL FOOTPRINT AND RISK FROM QUASI-FISCAL ACTIVITIES 53 bank NPLs at 100 percent of their origination price. The estimated recoveries were low at KZT 1.9 trillion (equivalent to 2.7 percent of GDP in 2020), which corresponds to a recovery rate of about 24 percent, including both realized and expected future recoveries. Figure 2‑7. Total bailout net cost for 2009–2020 in terms of 2020 present value Source: adapted from Mare, Melecky, and Muria (2022). Globally, the general principles for bailing out banks with public money are that the shareholders and large creditors must bear losses before taxpayers. Taxpayers must have recourse to the future profits of the bailed-out banks to realize a meaningful return. Such returns should cover part of the alternative cost of increasing sovereign borrowing and indebtedness or not investing taxpayers’ money into public services, such as education or health care. 2.4 Recommendations Define extrabudgetary funds (EBFs) in the Budget Code with substantial coverage consistent with the Government Fiscal Statistics Manual (GFSM). The Government has taken steps to monitor QFAs undertaken by EBFs, but the coverage remains limited. Action 40 of the Presidential Decree of September 2020, suggests that the Government will publish an “expanded budget” covering the activities of non- budgetary funds (the State Social Insurance Fund, the Social Health Insurance Fund, and the Unified Accumulative Pension Fund). The Government also amended the Budget Code to expand the concept of a consolidated budget to include payments by the Victims Compensation Fund, receipts, and expenditures of the SSIF and the SHIF (excluding mutually repayable transactions). The Government may want to consider defining in the Budget Code EBFs consistent with the GFSM guidelines, for example, “non-market and non-profit institutions that are both controlled and financed mainly by the state, legally non-government entities but carrying out state policies and program.” This definition can help include QFAs carried out by the UAPF, the Problem Loan Fund, and the Kazakhstan Sustainability Fund. 54 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Consolidate monitoring of QFAs and implement a framework for public accountability to monitor the fiscal cost and risks. Kazakhstan should consider developing an accountability framework to monitor QFAs by extra-budgetary funds (EBFs) and SOEs which allows the Parliament to discuss with the Government to verify the impact of the stated QFAs interventions and the appropriateness of such QFAs against the EBFs and SOEs core operational mandate. The Government may also wish to consider requiring EBFs and SOEs to consolidate the implementation of QFAs under one unit, reporting the funding source, size of the intervention, and, for certain activities (e.g., credit subsidy, pricing below cost), estimations of their implicit subsidies. This disclosure of information should describe the type of QFA, the rationale for performing it through an SOE rather than directly through budget financing, and the cost of the activity. Mechanisms designed to provide financial support to SOEs through concessional lending or tax exemptions should be disclosed, and their opportunity cost should be transparent. It would also be good practice to disclose the non-core activities of SOEs for social spending, mandated by the Government, separately from commercial activities. The SOEs should be compensated for such activities transparently from the budget. Refrain the NBK from implementing QFAs and consolidate the function of the KSF as a fiscal agent within the Government. Refraining the NBK from involvement in QFAs would contribute significantly toward better fiscal discipline and strengthen the NBK’s independence to ensure price stability. Therefore, the NBK’s plan to withdraw from the direct provision of support programs, as stated in its Monetary Policy Strategy 2030, is a major step toward strengthening its independence. The NBK also owns the Kazakhstan Sustainability Fund JSC, which can complicate efforts to put the fund’s QFAs under the Government’s consolidated budget reporting. In this regard, the NBK may wish to consider transferring the fiscal agent function of the KSF to the Government (this type of transfer was already done with the PLF). This move should also incentivize the Government to shift the support measures to firms and individuals from off-budget to on-budget programs or through public financial institutions such as the Baiterek Holding. Improving the governance and transparency of QFEs. The following measures may be considered part of the quasi-public sector’s overall reform. The activities of the QFEs, especially special operations for quasi- fiscal actions beyond their core mandate, should be subject to parliamentary and Supreme Audit oversight. Foundations must establish fund management supervisory boards whose members must be nominated by parliament, the Government, the private sector (employers), and the public (insured). To inform analysts and the public, the UAPF and SHIF should also prepare and publish actuarial reports once every three years. Take concerted action to minimize the fiscal risks from recurrent bank bailouts. The priority action that the Government may wish to consider is to intensify proper and timely supervision, including incorporating forward-looking stress tests into the supervisory approach. The authorities may also wish to consider using greater capital surcharges for unconditionally systemically important banks and greater counter- cyclical capital buffers (capital surcharges) to more proactively, and throughout the credit cycle, limit the occurrence of unstable or failing conditionally systemically important banks. This should also be followed up by amending the Bank Resolution Law to limit the use of public funds to strictly systemically important banks and limiting the room for inappropriate use of public funds in bailouts by narrowing the options for rescuing unstable versus failing systemically important banks. Kazakhstan should also consider strengthening the independence of the Financial Stability Committee (FSC)’s independence by reducing the Government’s representation and increasing the share of representatives from independent technocratic agencies. The Government should also consider publishing, either immediately or with an appropriate time lag, the decisions/recommendations of the FSC on the use of public funds in the resolution of each troubled systemically important bank. 2. MANAGING THE FISCAL FOOTPRINT AND RISK FROM QUASI-FISCAL ACTIVITIES 55 3. Improving Domestic Revenue Mobilization 56 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH KEY POINTS • Kazakhstan continues to rely heavily on oil revenues, while the non-oil tax collection has exhibited a declining trend since 2005. Since 2008/2009, the Government has reduced the major tax rates (corporate income tax and value-added tax) and introduced more generous fiscal incentives to attract investments, which further pressures budget revenue. • Currently, value-added tax (VAT) and corporate income tax (CIT) contribute the lion’s share in total tax collection. In contrast, the revenue contribution from other tax instruments, such as personal income tax (PIT) and excise, remains low. Combined, tax policy-setting and differentiated levels of tax-type compliance attribute to the observed collection composition. To enhance tax collection and equity, reform measures, in part, aim to raise PIT productivity and revenue intakes. Currently, the tax is not buoyant, and the collection declined to 1.3 percent of GDP in 2019, accounting for just about one-fifth of the PIT share of the ECA average and a half of the share in an all-UMICs average. • The Government has taken steps to increase tax revenue, but further reforms are needed to increase revenue and productivity in tax collections. Kazakhstan should consider rationalizing various CIT incentives, simplifying the CIT regime to minimize bunching and loopholes, lowering the threshold for VAT, and rationalizing various exemptions for VAT and PIT. Tax policy can play a different role in promoting the green transition and better health outcomes. Kazakhstan should consider raising the excise on tobacco products and broadening the excise to all fossil fuels. The Government should also consider institutionalizing cost-benefit analysis on tax incentives and regularly publishing tax expenditure reports. 3.1 Government Revenue Pressure from Declining Oil Revenue Kazakhstan’s tax revenues as a percentage of GDP have been trending down during the past 15 years (2005–2021), affecting the ratio of revenue to GDP. After being stable initially, at almost 27 percent of GDP during 2005–08, tax revenues dropped to 21.2 percent in 2009, briefly recovered to 25 percent between 2010 and 2011, and then declined significantly to just 14.2 percent of GDP in 2020 (Figure 3‑1). Non-tax revenues, averaging 10 percent of total revenue during 2005–2021, exhibited a similar oscillating trend over the same 15-year period, at around 1.8 percent of GDP. Therefore, the dynamics of total revenues have been driven largely by the change in tax revenues. The decline in oil revenues following the oil price collapse in 2014–2015 is a major factor behind the contraction of government revenues as a percentage of GDP. Oil revenue accounted for almost half of the general government revenues before the sharp oil price decline in late 2014. They fell to 40 percent in 2015 but remained sizable at 31.1 percent of total revenues over 2016–2019, albeit then declining to about 28.1 percent in 2021 (Figure 3‑2). As a share of GDP, oil tax revenue accounted for 12.5 percent during the 2010–2014 period but dropped significantly to 6.2 percent during 2015–2019 and to 3.4 percent in 2020 before bouncing back to 5.2 percent in 2021. Non-oil tax revenues declined from 11.4 percent of GDP (2010–2014) to 9.3 percent of GDP (2015), then recovering to about 10.7 percent of GDP (2016–2021) (Figure 3‑3). 3. IMPROVING DOMESTIC REVENUE MOBILIZATION 57 The decline in tax revenue as a percentage of GDP has also been accompanied by changes in tax-type- specific contributions. The driving force for down trending in the tax-GDP ratio has been the decline in the CIT and natural resource taxes, both probably due to the performance of the extractive sector. The CIT collection gradually decreased from 11 percent of GDP in 2005 to 3.5 percent in 2020. Notably, the CIT declined sharply during 2008–2009, from 12 to 7.5 percent of GDP, due to a combination of factors: The reduction in the CIT standard rate from 30 to 20 percent under the new Tax Code; and the unfavorable fluctuations in oil prices and decline in oil production. The collection of PIT, property tax, and excise has been relatively constant over the past 15 years. Social tax fell from 2.4 percent of GDP in 2005–2007, remaining stable at around 1.1 percent of GDP since 2010.46 VAT revenue oscillated during 2005–2021: it decreased from 4.6 percent of GDP over 2005–2008 to 2.3 percent in 2015, then recovered to 3.5 percent (2016–2021). Natural resources use revenue averaged 4.1 percent of GDP in 2005–2007 before declining rapidly to 2.9 percent over 2015–2021. Figure 3‑1. Revenue collection and total public expenditures (% of GDP), 2005–2021 Source: World Bank staff calculations Figure 3‑2. Composition of fiscal revenue Figure 3‑3. General government tax revenues (% of total revenue), 2010–2021 (% of GDP), 2010–2021 Source: World Bank staff calculations based on data from IMF Article IV. 46 Social tax in Kazakhstan consists the major schemes, namely, social insurance contributions, medical insurance contributions, and pension contributions. 58 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH VAT, along with CIT, has become a major source of tax revenue, along with CIT, as taxes on natural resource activities have moderated. On average, during 2005–2021, the share of CIT (29.5 percent), taxes on natural resources use (22 percent), and VAT (17.9 percent) were dominant in total tax revenues, inclusive of natural resource use revenues (Figure 3‑5). Figure 3‑6 splits the 2005–2021 period into three sub-periods: 2005–2010, 2011–2016, and 2017–2021 to enhance understanding of the tax composition evolving over time. Only the share of CIT gradually decreased from 38.4 percent in 2005–2010 to 26.9 percent in 2017–2021. In contrast, the share of PIT (from 6.3 to 8.2 percent), property taxes (from 2.7 to 3.4 percent), excise (from 1.5 to 3.3 percent), and trade (from 5.9 to 11.9 percent) rose steadily over the same time interval (all referred to the composition with natural resource use revenues included). When revenues from natural resources use are excluded, CIT and VAT accounted for the two largest shares of tax revenue, at 37.9 and 22.9 percent, respectively; this nearly doubled the shares from trade tax, PIT, and social tax combined (Figure 3‑7). The share of CIT decreased from 49.4 percent in 2005–2010 to 32.8 percent in 2017–2021. In contrast, the share of PIT, VAT, and excise increased, reflecting the various tax- type specific responses, driven largely by changes in the statutory base (Figure 3‑8). Figure 3‑4. Tax revenue, by source and year (% of GDP), 2005–2021 Source: World Bank staff calculations based on data published by the authorities. Figure 3‑5. Composition of tax revenues, Figure 3‑6. Contribution to tax revenues, inclusive of natural resource use revenues inclusive of natural resource use revenues, (average 2005–2021) average 2005–2010, 2011–2016, and 2017– 2021 Source: World Bank staff calculations based on data published by the authorities. 3. IMPROVING DOMESTIC REVENUE MOBILIZATION 59 Figure 3-7. Composition of tax revenues without Figure 3-8. Contribution to tax revenue by a source natural resources use revenue (average 2005– without natural resources use revenue, average 2021) 2005–2010, 2011–2016, and 2017–2021 Source: World Bank staff calculations based on data published by the authorities. 3.2 Tax Collection Compared to Other Countries On aggregate, tax revenue in Kazakhstan as a share of GDP has been lagging behind its peers’ performance. The low revenue efforts narrow the fiscal space, as has the relatively low public spending compared to other countries. Kazakhstan’s tax performance is compared against those countries in five comparable peer groups: (i) all countries in the ECA region; (ii) upper-middle-income countries in ECA; (iii) all upper-middle-income countries across regions; (iv) the world; and (v) the selected resource-dependent countries 47 (Figure 3-9). Figure 3-9. Tax revenue in international comparison (% of GDP) Source: World Bank staff calculations based on data published by the authorities. 47 The sample of ECA groups by income is as follows. (i) The high-income group includes Austria, Belgium, Croatia, Cyprus, the Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Latvia, Lithuania, Luxembourg, the Netherlands, Norway, Poland, Portugal, the Slovak Republic, Slovenia, Spain, Sweden, Switzerland, and the United Kingdom. (ii) The upper middle-income group consists of Albania, Armenia, Azerbaijan, Belarus, Bosnia and Herzegovina, Bulgaria, Georgia, Kazakhstan, Kosovo, Moldova, Montenegro, North Macedonia, Romania, the Russian Federation, Serbia, Turkey, and Turkmenistan. (iii) The lower middle-income group consists of the Kyrgyz Republic, Moldova, Ukraine and Uzbekistan. 60 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH While Kazakhstan’s CIT revenue was highest at 4.6 percent of GDP among the upper-middle-income group in ECA in 2019, revenue from other taxes was the lowest. Kazakhstan’s revenue from PIT, social contribution, VAT, and excise taxes was 1.3, 0.5, 3.9, and 0.5 percent of GDP, respectively. Kazakhstan’s tax performance is also compared against that of structural peers in 2019 (Table 3-1). (These countries, specifically Bolivia, Mongolia, Azerbaijan, and Gabon, share similar socio-economic features: natural resource-rich countries, the size of government (government spending as a percentage of GDP), GDP per capita (constant 2010 US$), and population size). Kazakhstan’s tax collection performance (17.3 percent of GDP) is slightly better than Azerbaijan’s (16.9 percent of GDP) but significantly below the tax intakes in Bolivia and Mongolia. Kazakhstan’s CIT collection was still higher than all structural peers but ranked the second-lowest in VAT (3.9 percent of GDP) and excise (0.5 percent of GDP). Table 3-1. General government revenues (% of GDP) 2019 No. of GDP per Total Tax Pro- Selected countries in capita Excise reve- reve- VAT Trade perty PIT CIT SSC Countries the Income (constant tax nue nue tax group 2010 US$) High Income 64 > 14150 38.73 22.93 6.67 2.13 0.75 1.60 7.28 3.32 8.88 Upper middle 56 4,410 - 14,900 31.08 18.40 6.26 2.02 2.50 0.60 2.94 2.95 4.85 income Lower middle 55 1,070 - 4,700 23.41 16.70 5.70 2.68 2.28 0.36 2.46 3.40 2.43 income Low income 27 < 950 18.65 11.16 4.91 1.60 1.95 0.12 1.91 2.32 0.80 ECA Uzbekistan 1,800 28.06 19.19 6.87 1.61 0.57 0.92 2.61 3.48 4.63 Kyrgyz Lower 1,240 34.05 20.60 8.63 1.69 3.72 0.20 0.92 2.38 5.50 Republic middle Moldova income 4,580 29.96 19.06 9.61 2.96 0.86 0.28 1.89 2.55 8.76 Albania 5,220 27.43 19.53 7.89 2.78 0.39 0.34 1.90 2.88 5.75 Armenia Upper 4,680 23.83 22.01 7.22 1.94 1.45 0.39 6.24 2.76 0.28 Belarus middle 6,370 38.42 24.47 8.36 2.07 3.52 1.19 4.39 3.03 10.81 Kazakhstan income 8,820 19.56 17.34 3.87 0.49 2.25 0.57 1.26 4.63 0.53 Russia 11,250 35.82 26.03 6.48 1.22 2.77 1.09 3.47 4.07 7.41 Hungary 16,530 43.33 24.14 9.69 2.79 0.17 0.95 5.40 1.20 11.30 High income Poland 15,350 41.03 22.09 8.01 3.10 0.19 1.28 5.35 2.23 13.29 Structural Peers (resource-rich countries) Bolivia Lower 3,520 28.84 22.52 7.22 1.84 1.19 0.18 0.18 3.84 6.19 middle Mongolia income 3,790 32.40 23.70 6.70 2.37 2.11 0.44 2.40 4.60 5.40 Azerbaijan Upper 4,490 41.52 16.94 6.33 1.04 2.04 0.34 1.16 3.11 2.35 middle Gabon income 7,170 19.53 12.07 1.34 0.29 2.89 0.07 1.40 4.00 0.24 Source: World Bank staff estimates based on data from the WDI; IMF FAD database; Poland - IMF 2020 Article IV; Mongolia - IMF Request for Purchase Under the Rapid Financing Instrument No. 20/205, June 2020; Hungary - IMF 2021 Article IV; and Azerbaijan - IMF 2021 Article IV. Note: PIT = personal income tax; CIT = corporate income tax; SSC = social security contributions; VAT = value-added tax. 3. IMPROVING DOMESTIC REVENUE MOBILIZATION 61 3.3 Most Tax Collections Table 3-2. R-esponsiveness of revenues to GDP growth, 2005–2020 Are Showing Deficient Tax Type Buoyancy Buoyancy Performances Total Tax Revenue 0.74 The tax buoyancy analysis shows moderate- Personal income tax (PIT) 0.89 ly unsatisfying general tax collection perfor- Corporate income tax (CIT) 0.49 mances, except for excise and trade taxes.48 Social tax 0.60 Table 3-2 shows the tax buoyancy for different tax categories from 2005–2020. The 2005– Property tax 0.91 2020 period’s buoyancy for all combined tax Value-added tax (VAT) 0.84 revenues stood at 0.74, meaning the tax col- Domestic 1.27 lection performance was negative compared Imported 0.66 with economic growth. The highest tax buoy- ancies are shown for tax on international trade Excise 1.09 and excise tax, respectively, at 1.33 and 1.09. Natural resource use revenue 0.77 The buoyancy, however, is at the lowest for CIT Trade tax 1.33 (0.49) and remained non-buoyant for social tax Exported 2.11 (0.6) and tax on natural resource use (0.77). The buoyancy is almost 1 for property tax (0.91), Imported 0.77 PIT (0.89), and VAT (0.84). Source: World Bank staff calculations. 3.4 Tax Policy Analysis The Government aims to increase state budget revenues (net of transfers from NFRK) to 18 percent of GDP in 2030. (Kazakhstan 2022). The Government also intends to strengthen the carbon pricing mechanism to implement the country’s commitments under the Paris Agreement (UNDP n.d.) (Strategy 2050 2021). 49 The draft updated NDC (Republic of Kazakhstan 2021) reinforces the role of the ETS. It proposes additional mitigation measures—such as a new energy tax to cover transport fuels and small-scale coal use (World Bank, 2022). 50 This section provides an overview of the key issues in tax policy and administration directly impacting collection and economic efficiency. 3.4.1 Corporate income tax Statutory rates of major taxes are comparable with peers Kazakhstan’s standard CIT rate is comparable with upper-middle-income peers’ rates, but not its PIT and VAT rates. Figure 3‑10 and Figure 3‑11 summarize the 2020–2021 statutory tax rates for CIT, PIT, and VAT in international comparison. The CIT rate of 20 percent is only lower than that of Turkey (22 percent) but higher than the income group average (15 percent). For CIT, the rate in Kazakhstan is comparable to the ECA average (19 percent). The standard VAT rate, however, is at the lowest (12 percent), significantly lower than the group average (18 percent). The PIT flat rate in Kazakhstan of 10 percent is the second- 48 Tax buoyancy indicates the extent to which tax collection responds to the change in the legally defined base. It is measured as the rate of growth in tax revenues relative to the growth rate of the respective base (in real terms). 49 The November 2022 World Bank Report on Kazakhstan Country Climate and Development Report, however, suggests the country’s 2030 NDC, and particularly its net-zero by 2060 pledge, require strong action: While there are the beginnings of a mitigation policy framework, with an emissions trading system (ETS) and renewable auctions, the framework and its components need to be more ambitious, and additional policies are required to unlock barriers to scaling up, drive emissions reductions across sectors, and support communities through the transition. 50 World Bank 2022. Country Climate and Development Report: KAZAKHSTAN. (Page 20). 62 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH lowest (only higher than the rate of 9 percent in Montenegro) and significantly below the group average of 16 percent. Compared with other income groups among the ECA emerging and developing economies, Kazakhstan’s PIT flat is at the lowest level compared with the average of either the lower middle-income group (13 percent) or the high-income group (28 percent). This signifies the dual detrimental impacts on both PIT efficiency and progressivity. Figure 3‑10. Comparative statutory CIT and PIT rates (%), 2020–2021 Source: World Bank staff calculations based on data published by the authorities, IMF, PwC, and KPMG. Figure 3‑11. VAT rate (%), 2020–2021 Source: World Bank staff calculations based on data published by the authorities, IMF, PwC, and KPMG. 3. IMPROVING DOMESTIC REVENUE MOBILIZATION 63 Corporate income tax performance is relatively high Kazakhstan has a worldwide income tax regime by which resident companies are basically taxable on their worldwide profits. Non-resident companies operating through a permanent establishment in Kazakhstan are subject to Kazakhstan CIT only on the profits attributable to that specific permanent establishment (they are subject to income tax withheld at the source of payment on Kazakhstan-sourced income). Kazakhstan clearly relies more on CIT than most other comparable ECA countries. As in most resource- rich countries, Kazakhstan derives a relatively large part of its revenue from the taxation of corporations, even when oil companies are excluded from the revenue statistics for comparison. On average, CIT in Kazakhstan accounted for 4.5 percent of GDP over 2016–2019, nearly twice the level attained by regional or global peers (Figure 3‑12). During 2005–2021, CIT in Kazakhstan was stronger than regional averages (Figure 3‑13) and contributed, on average, almost 32 percent to total tax revenues (or 7 percent of GDP). The contribution of CIT revenues was about 11.4 percent of GDP between 2005 and 2008, sharply declining to 7.1 percent in 2009 due to the rate reduction and then fluctuating at about 4.3 percent over the period 2015–2021.51 In the aggregate, CIT productivity is measured as the ratio between CIT collection in GDP and the standard CIT rate—which is high in Kazakhstan. The comparative position of Kazakhstan’s revenue and productivity in the ECA region and the structural peers is presented for 2019 (Figure 3‑14).52 In 2019, the Kazakhstan CIT revenue (4.6 percent of GDP) was the highest across all two groups (ECA peers and structural peers). In terms of the CIT productivity ratio, only Uzbekistan (0.248) and the Kyrgyz Republic (0.238), both from the ECA group, did better than Kazakhstan (0.231), reflecting the comparatively broad base in tax policy and the resource sector-driven concentration of large-scaled corporate sector (thereby the ease of administration enforcement). Compared with the ECA countries, Kazakhstan ranked third. Figure 3‑12. CIT revenue in international comparison (% of GDP) Source: World Bank staff calculations based on data published by the authorities. 51 In the past, the Kazakhstan profit tax rate (30 percent) was rather high but has been reduced from 30 to 20 percent since 2009. 52 The CIT productivity is estimated as the ratio between the actual collection as a share of GDP and the standard CIT rate. 64 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Figure 3‑13. CIT collection in regional comparison Figure 3‑14. CIT productivity ratio, 2019 (% GDP) Source: World Bank staff calculations based on data published by the authorities. Corporate income tax structure is comparable with the regional average The CIT rate. The standard CIT rate of 20 percent is similar to the ECA average but is lower than the global comparison. Corporate capital gains or interest income are also taxed at the standard CIT rate, while dividends are exempt. Certain qualified agricultural income is granted a reduced rate of 6 percent. Meanwhile, excess profit tax (EPT) rates are applicable on a graduated schedule and range from 10 to 60 percent.53 The Base. Taxable income is the taxpayer’s aggregate annual income, less allowable deductions. Fiscal depreciation. The rule for fiscal depreciation is clear and follows international best practices. The deduction is based on the declining-balance method, with rates ranging from 10 to 40 percent, applied to four basic categories of assets, specifically: (i) buildings and facilities; (ii) machinery and equipment; (iii) computer and other IT-related equipment; and (iv) others. Loss carry-forward. The granting of loss carried forward is generous, of up to 10 years. International taxation. The Tax Code provides major anti-avoidance provisions, including transfer pricing, thin capitalization, and controlled foreign corporations (CFCs). Significant legislative developments have been made in international taxation and implementation of the Base Erosion and Profit Shifting (BEPS) agenda. The country has ratified the Multilateral Instrument.54 The new law on amendments to tax legisla- tion has been signed with the provision to affect the CFC rules.  53 Starting from 2018, EPT was abolished for subsurface users engaged in the extraction of solid minerals. Instead, the new Tax Code introduced an alternative tax that replaces EPT, mineral extraction tax (MET), and compensation of historical costs and may be applied at the discretion of a taxpayer. (PWC, Worldwide Tax Summaries 2021). 54 The Multilateral Instrument is part of the OECD/G20 Base Erosion and Profit Shifting (BEPS) Agenda. It offers concrete solutions for governments to close loopholes in international tax treaties by transposing results from the BEPS Project into bilateral tax treaties worldwide. The MLI allows governments to implement agreed minimum standards to counter treaty abuse and to improve dispute resolution mechanisms while providing flexibility to accommodate specific tax treaty policies. Please see (OECD n.d.). 3. IMPROVING DOMESTIC REVENUE MOBILIZATION 65 3.4.2 Special tax regimes for small business entities or presumptive regimes: The dual purpose of setting special tax regimes is to simplify compliance and thereby lower collection costs and encourage entrepreneurs to enter the regular tax net progressively. Special tax regimes are categorized into three sub-regimes: patent regime, simplified declaration, and fixed deduction. Their classification and structures are all complex, however (Table 3‑3), and open many loopholes. They essentially add to the long list of tax incentives (to be discussed in 3.4.3), which together are expected to negatively impact neutrality, fairness of the overall tax system, competition, productivity, and private sector job creation. Two observations can be made. The dual eligibility criteria of turnover and the number of employees risk being inherently conflicting. A case in point: a fixed deduction regime is applied to businesses with a certain limit of employees (no more than 50) but is also allowed for those with an annual turnover of 144,184 times Monthly Calculation Indexes (translated to almost US$1 million equivalent). (Note this turnover threshold is significantly higher than the threshold for a regular VAT regime at 20,000 MCI.) Similarly, the eligible turnover for the simplified deduction regime of 24,038 MCI is significantly higher than the current regular VAT threshold. The segregation between those in the patent and simplified declaration category is distortionary and subject to abuse. Still, it could be integrated into the strategy to simplify the special tax regime. It incentivizes businesses to split (so that their turnover stays at the defined level for the patent group) and to hire employees informally (without formal declaration to self-exclude from the simplified declaration group). The same base (gross income without deductions) is applied for both categories, but the rate for the patent regime is just one-third of the simplified declaration. Table 3‑3. A tiered system of special tax regimes Annual turnover eligibility No. of Regime Target group Rates Base MCI a tenge US$ b Employees Gross Patent Individual income 1% 3,528 times <= 10,806,300 <= 23,550 None regime entrepreneurs without deductions Individual Gross Simplified entrepreneurs income 24,038 3% <= 73,628,400 <= 160,460 <= 30 declaration and legal without times entities deductions Gross income Individual less fixed Fixed entrepreneurs 10%/ 144,184 deduction <= 441,635,600 <= 962,500 <= 50 deduction and legal 20% times (30% of entities gross income) Source: Kazakhstan Tax Code. Note: a/ MCI on 2022 = KZT 3,063; b/ average US$/KZT rate as of February 25, 2022, is KZT 458.86 = US$1.00 66 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 3.4.3 Tax incentives Tax incentives are common worldwide (see Box 3-1). The effectiveness of tax incentives depends on various factors: the types of incentives (specifically the profit-based or cost-reducing ones), the overall quality of the legislation, and governance. Box 3‑1. Tax incentives around the world Despite the difference in opinion on their effectiveness, the fact remains that tax incentives, in one form or another, are used by nearly all countries worldwide. A large-scale survey of tax incentives among 140 countries shows that tax holidays are prevalent in all regions except for OECD countries. This reflects the gradual move away from using tax holidays among developed countries due to their ineffectiveness in aligning increased investment incentives with tax benefits. There is now much greater use of tax incentives to encourage research and development, with the OECD countries and those in East Asia and the Pacific using this tax incentive most frequently. Super-deductions, where deductions are allowed for more than the actual cost of certain expenses, are most prevalent in South Asia, mainly to subsidize investment costs when starting a business. The use of tax and duty exemptions in Special Economic Zones is popular across all regions. There is an overall trend in moving away from tax holidays, which is most marked among OECD countries. Source: (Sebastian 2009). In Kazakhstan, overall fiscal incentives are granted through various tax and expenditure instruments and are subject to executive discretion. The incentives are granted through CIT, customs duties, VAT, and investment subsidies. They spread across the Tax Code and other legislation, primarily the Entrepreneurial Code. The CIT-specific incentives vary, including outright tax holidays, reduced rates, accelerating forms of fiscal depreciation, and partial exemption of incomes or sales. Tax incentives are also subject to discretion by executive decision (for example, the Entrepreneurial Code allows for tax incentives for sectors deemed a priority by the Government). This sounds reasonable but tends to create loopholes for such a priority list to be expanded over time. About a dozen Special Economic Zones (SEZs) have been established with different expiry dates. In addition to certain incentives provided (as noted above), the SEZ-specific incentives are even more generous, including CIT tax holidays, zero-rated VAT on non-exports, and exemption from land or property tax payment. Preliminary estimates suggest that fiscal incentives caused a considerable revenue gap. The draft World Bank Report on Tax Expenditures in Kazakhstan (September 2022) provides preliminary estimates of revenue loss due to prevalent VAT and CIT tax expenditures.55 The analysis identifies 197 tax expenditures, of which 145 correspond to VAT and 52 to CIT. In 2021, Kazakhstan’s combined tax gap for CIT and VAT, driven by the revenue forgone triggered by the noted tax expenditure provisions, was estimated at 7.8 percent of GDP, and the estimated total tax gap for VAT and CIT approached 57 percent. The broken-down tax gap projections are as follows: • Tax gap in domestic VAT: 70.6 percent. • Tax gap in customs VAT: 22 percent. • Tax gap with total VAT: 52.6 percent. • Tax gap in CIT: 60.9 percent. The report’s conceptual approach to and methodology for estimating the fiscal costs of tax expenditure provisions are summarized in Box 3‑2 55 Developed. 3. IMPROVING DOMESTIC REVENUE MOBILIZATION 67 Box 3‑2 Tax Expenditures in Kazakhstan: evaluation of this policy instrument There are two main approaches for estimating tax gaps: a top-down approach and a bottom-up approach. The top-down approach aims to provide a comprehensive assessment of all tax revenue foregone by measuring the gap as the difference between the actual base and revenue collection and the estimated potential of that base and its associated revenue using macroeconomic data. In contrast, bottom-up approaches usually rely on microdata from (random) audits, tax returns, and other interventions by the tax authority. These can provide valuable insights into taxpayers’ behavior and related risks. However, since they are based on the results of random audits, they are often costlier for tax administrations than top-down approaches. However, the main drawback of the top-down approach is that it usually does not account for behavioral responses and secondary effects. The World Bank preliminary report (2021) on Tax Expenditures in Kazakhstan: evaluation of this policy instrument has two main objectives: (1) to provide an overview of the universe of tax expenditures in Kazakhstan for the income and value-added tax, and 2) provide preliminary estimates of VAT and CIT tax gaps for the country. The analysis relies on a top-down sectoral approach. Macroeconomic data, for example, from input-output tables, is used to estimate the revenue potential per sector per tax type, in this case, either VAT or CIT. These sectoral estimates are then aggregated to provide an economy-wide measure of the total revenue potential per tax type. Next, the aggregate revenue potential is compared to the actual revenue collection. As opposed to the absolute gap, the relative gap gives a sense of the performance of a certain tax. It provides a more detailed assessment of how close current collection efforts are to their potential. If data on sectoral revenue collection is available, the tax gap could be decomposed into sectoral tax gaps. Calculating the revenue potential per sector involves several steps. As a first step, the tax base needs to be established. Depending on the tax type, the base is either (net) income, for the corporate income tax or value-added, for the value-added tax. Step two is the exclusion of the non-taxable portion. The non-taxable portion refers to the sectoral part of the tax base that cannot be taxed because of the policy choices made by the government. These policy choices could include exemptions, deductions, or thresholds, among other things. In step three, the tax rate is applied to the taxable portion, defined as the tax base excluding the non-taxable portion, to calculate the potential revenue for a particular tax type in a particular sector. The result is the potential revenue or the benchmark for that sector for that tax. It represents the total revenue the government would be able to collect if all taxpayers were to fully comply with their obligations, allowing only for informality and legal reductions in their obligations. Finally, after summing the revenue potential across the different sectors, the total tax gap is calculated by subtracting the actual or observed revenue collection from the estimated total revenue potential. The difference is the tax gap. Given the size of the incentives provided and their unknown impact, Kazakhstan should consider undertaking a systematic analysis of the impact of tax incentives on firms’ growth, investments, and jobs. Evidence suggests that tax expenditures through incentives and special regimes do not always deliver the intended results. A study of over 40 countries in Latin America suggests that while tax holidays can attract FDI, they fail to show any impact in boosting gross capital spending (Klemm and van Parys 2009). On the other hand, a study on a location-based tax incentive in India suggests that such a policy positively attracts larger and more productive firms (Chaurey 2017). A study on special tax regimes in selected European countries (Benedek, et al. 2017), such as reduced CIT for small firms, can also negatively impact their growth and productivity by creating a perverse incentive for them to remain small (the small business trap). Analyses of the impacts of tax incentives and special regimes typically use micro-level data (firm-level data). Therefore, it would be important for the Government to set a solid framework for monitoring and evaluation that can compare the performance of firms/entities receiving the incentives versus those not. 68 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 3.4.4 Personal income tax (PIT) Personal income tax performance is lagging behind the ECA regional average PIT collection is low, has steadily declined over the past four years, and is below the regional average. PIT collection declined from 1.5 percent of GDP in 2016 to 1.3 percent in 2019 (Figure 3‑15). Throughout the 2016–2019 period, the country’s PIT collection in terms of GDP accounts for just one-fifth of the PIT share of the ECA average and one-half of the share in all upper middle-income countries’ average. PIT collection in Kazakhstan was below all three ECA group averages (Figure 3‑16). It is also noteworthy that the PIT revenue trends of Kazakhstan and the ECA emerging markets group are divergent: the trend in Kazakhstan is declining, whereas it is increasing in the emerging markets group. Figure 3‑15. PIT revenue in international comparison (% of GDP) Source: World Bank staff calculations based on data published by the authorities. NOTE: IN TABLE ABOVE CORRECT: Upper middle income (all regions) Figure 3‑16. PIT collection in a regional Figure 3‑17. PIT productivity ratio, 2019 comparison (% GDP) Source: World Bank staff calculations based on data published by the authorities. 3. IMPROVING DOMESTIC REVENUE MOBILIZATION 69 The aggregate PIT productivity of Kazakhstan is also below the regional average. The indicator is the ratio between the PIT share of GDP and PIT’s highest marginal statutory rate. The comparison shows that PIT productivity in 2019 and the comparators are separated into three regional income groups: lower middle- income, upper middle-income, and high-income, and the selected global structural peers. The estimated PIT productivity of Kazakhstan ranked ninth among the 11 selected ECA comparison peers, regardless of income groups. Kazakhstan’s PIT productivity ratio was 0.126 and below the average of any ECA group but higher than most non-ECA structural peers (Bolivia, Gabon, and Mongolia).56 It is worth further noting that the PIT productivity in Kazakhstan (0.126) is significantly lower than the average in UMIC (almost 0.2) and even lower than that of the HIC (0.256). Personal income tax structure Kazakhstan’s PIT rate is flat and low. The PIT rate is set at 10 percent. On the other hand, the personal allowance or basic deduction is set at KZT 510,000 per year (equivalent to US$1,041 per year at the rate of US$1/KZT 490).57 This basic deduction accounts for only a fraction (11 percent) of the 2020 per capita income of US$9,122 (World Bank n.d.). Thus, The PIT regime indicates the opposite impacts and embedded detrimental vertical equity issues: low-wage earners are subject to the same rate as better-off individuals who pay a notably low rate. The PIT base covers all employees’ compensation in cash or in-kind (fringe benefits are also subject to PIT). The legally-defined base consists of employment, individual business income, and income from investments (capital gains, dividends, and interest earned from foreign-affiliated banks). However, a closer look at the individual income tax base reveals multiple channels for revenue leakage. A long list of tax expenditure provisions exists, including exemptions of incomes from alimony, select dividends and interests, incomes of military service members and law enforcement forces, and winnings from lotteries. Equity compensation income exemption. Equity compensation, or a form of compensation representing some type of ‘ownership’ of a company granted to an employee, is income tax exempt. Neither the granting of equity nor the exercise of equity compensation is assessed for income tax purposes. This form of exemption not only contributes to a narrowing of the tax base but also a reduction in equity. Capital gains taxation. Income tax on assessed capital gains is levied at 10 percent. However, multiple ex- emptions acting as tax expenditure provisions exist and are granted for: • Itemized securities, including shares owned for more than three years. • Shares that meet other conditions, namely, those from the company whose shares are sold not regard- ed as sub-surface users.58 (This is a significant tax expenditure provision, provided that more than 50 percent of the company’s shares are sold and are not owned by sub-surface users.) Dividend taxation. Rates differ based on the source of dividend income. Those sourced from Kazakhstan are subject to a rate of 5 percent, while foreign-sourced dividend incomes are taxed at a higher rate of 10 per- cent. There is a long list of exemptions regarding the base in dividends tax. For example, dividends accrued to individuals holding shares for over three years are exempt. This outright exemption is a tax expenditure item. Interest income. Different treatments of interest income from different types of banks: Interest is exempt, except for interest received from foreign banks. 56 The case of Mongolia collection and productivity trending deserves further study. The collection of income taxes (including the PIT) is facilitated with the high concentration of taxable income among a few large taxpayers. 57 The reference exchange rate is as of February 28, 2022. This exchange rate is used throughout the chapter. 58 Oil, gas and mining companies in Kazakhstan are referred to as subsurface users and enter into subsurface use contracts to acquire the rights to extract the mineral resources. 70 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Rental income. Efforts are being made to tax rental income on an assessed net income basis (with deductions allowed for maintenance and repair costs or any costs relating to leasing services). Wealth tax. Kazakhstan neither levies taxes on net wealth nor inheritance or gifts. Combined with the low, flat-rate PIT regime, the omission of the wealth tax exerts a further negative impact on the fairness and progressivity of the income tax instrument. The government intends to correct this – with the plan to revise the Tax Code in 2023 with the possibility of introducing a proxy form of “wealth tax” levied on purchasing high-value real estate and vehicles.59 3.4.5 Value-added tax (VAT) Value-added tax performance VAT is the second-largest source of government tax revenue (17.2 percent)—just behind the CIT, but collection continues to lag behind its peers. As shown in Figure 3‑18, VAT collection as a percentage of GDP in Kazakhstan is lagging behind global, ECA, and the world’s upper middle-income country average. The country’s VAT collection compared with GDP (4 percent) is far lower than the regional upper middle- income average (roughly 8 percent) or the world average (almost 7 percent). Relative to other countries in ECA (Figure 3‑21), Kazakhstan’s VAT performance is declining. As a percentage of GDP, VAT averaged 4.7 percent during 2005–2007, then steadily dropped and oscillated at around a 3.5 percent annual average (2016–2021).60 Figure 3‑18. VAT revenue in international comparison (% of GDP) Source: World Bank staff calculations based on data published by the authorities. VAT collection heavily relies on imports, although such reliance has declined over time (Figure 3‑19). VAT on imports accounted for 78.1 percent of total VAT revenues during 2005–2015 but fell to 57.3 percent during 2016–2021. 59 PWC 2022. World Wide Tax Summaries. Accessed February 12, 2023. 60 To mitigate the COVID-19 impact on food affordability, the Ministry of Finance of Kazakhstan has ordered that the VAT rate on food products be reduced from 12 to 8 percent, and that selected customs duties be eliminated through October 2020. Please see (United States Department of Agriculture 2020). 3. IMPROVING DOMESTIC REVENUE MOBILIZATION 71 Figure 3‑19. VAT collection in regional Figure 3‑20. VAT revenue, by source and year comparison (% of GDP) (% of GDP) Source: World Bank staff calculations based on data published by the authorities. VAT efficiency and productivity fluctuate and are lower than those of peers. VAT productivity and C-efficiency fluctuated over the period 2005–2019, improving in recent years.61 VAT productivity and C-efficiency averaged at 0.29 and 0.61, respectively, dipping to a low in 2015 (Figure 3‑21Figure 3‑21). However, compared with ECA peers, Kazakhstan had the lowest VAT productivity in 2019 (0.32) and below the 2016 average VAT productivity of emerging and developing countries in Europe (0.47) (figure 3‑22Figure 3‑22). In terms of C-efficiency, Kazakhstan ranked seven (0.617) out of 11 ECA countries and slightly higher than the C-efficiency average of emerging and developing countries in Europe (0.56). Figure 3‑21. Kazakhstan’s VAT revenue, VAT Figure 3‑22. VAT productivity and C-efficiency productivity, and C-efficiency, 2005–2019 comparison, 2019 Source: World Bank staff calculations based on data published by the authorities. 61 VAT productivity or VAT C-efficiency are the two indicators that gauge, in aggregate, the VAT efficiency. C-Efficiency is measured as the ratio between the VAT collection in GDP and the product of the standard statutory VAT rate and the share of household consumption in GDP. VAT productivity, on the other hand, represents the ratio between VAT collection in GDP and the standard VAT rate. 72 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Value-added tax structure Kazakhstan is the first country in the Commonwealth of Independent States (CIS) to have introduced VAT, albeit at a relatively low rate. The current standard VAT rate of 12 percent is the lowest compared with the average of the ECA region. Kazakhstan uses a tax-credit invoice type of consumption-based VAT. While the standard rate is low, the VAT regime grants multiple zero ratings for various non-export transactions, mostly related to SEZ operations. The zero-rating VAT of non-exports is subject to abuse, including cross- transactions between SEZs and non-SEZ areas. This entails additional costs for enforcement and, more importantly, VAT-refund audits. The VAT base has a relatively high threshold. From 2022, the mandatory VAT registration threshold is 20,000 MCI (PWC, 2021). (With 1 MCI = KZT 2,778, the threshold is almost KZT 56 million, equivalent to more than US$113,000.) Such a threshold is high, noting that it has already been reduced by one-third. The VAT threshold for individual entrepreneurs using special tax regimes is more than seven times higher, at 114,184 MCI (or almost US$1 million equivalent). Voluntary registration is allowed for those businesses with turnover lower than the defined turnover threshold.62 There are two policy questions related to these threshold levels: • Whether further reform would consider lowering the threshold to conform to international best practic- es as part of an effort to broaden the tax net and thereby enhance VAT efficiency. • Why businesses enjoying the special tax regime are granted such a high VAT threshold. The setting in the latter case would defeat the very purpose of creating a special regime intended exclusively for small businesses to reduce their compliance costs but not to shield them from paying a fair share of tax (or at the level of tax liabilities comparable to those in the standard regime). Table 3‑4 summarizes the thresholds established in the VAT regimes in regional countries (across different income groups). It provides some reference for Kazakhstan to emulate, subject to, among others, the country-specific STC capacity and progress in VAT administration (in particular, the mandatory application of e-invoices). The simple average and the median of thresholds in the sample ranged between US$66,000 and US$83,000 (compared with the current VAT threshold in Kazakhstan of US$113,000 equivalent). Table 3‑4. VAT exemptions GDP per Exemption threshold Statutory Exchange Local Countries capita VAT rate (%) rate currency in LCU in US$ Ratio (current US$) Hungary 27 15,899 IMF HUF 12,000,000 39,762 2.50 Poland 23 15,656 IMF PLN 200,000 51,049 3.26 Albania 20 5,215 3/16/22 ALL 10,000,000 89,440 17.15 Armenia 20 4,267 IMF AMD 115,000,000 232,394 54.46 Belarus 20 6,411 - - - - - Kazakhstan 12 9,056 7/20/21 KZT 56,000,000 113,000 12.48 Russia 18 10,127 - - - - - Uzbekistan 20 1,686 IMF UZS 1,000,000,000 99,917 59.27 Kyrgyz Rep. 12 1,174 3/16/22 KGS 30,000,000 285,816 243.54 Moldova 20 4,551 IMF MDL 1,200,000 65,754 14.45 Ukraine 20 3,727 IMF UAH 1,000,000 36,760 9.86 62 As conforming to the international standards in VAT policy, individuals who are not individual entrepreneurs, professionals with a private practice; state bodies; non-residents not operating in the Republic of Kazakhstan through a branch or a representative office, and structural units of resident legal entities are not eligible for voluntary VAT registration. 3. IMPROVING DOMESTIC REVENUE MOBILIZATION 73 GDP per Exemption threshold Statutory Exchange Local Countries capita VAT rate (%) rate currency in LCU in US$ Ratio (current US$) Azerbaijan 18 4,214 IMF AZN 200,000 117,647 27.92 Gabon 18 7,006 3/16/22 XAF 60,000,000 100,623 14.36 Bolivia 13 3,143 - - - - - Mongolia 10 4,007 IMF MNT 50,000,000 17,415 4.35 Median exemption (US$) 65,754 Average exemption (US$) 83,305 Source: World Bank staff calculations based on data published by the authorities; IMF FAD Database; and PwC Worldwide Tax for Albania, Kyrgyz Republic, and Gabon. Note: Belarus, the Russian Federation, and Bolivia do not have a VAT registration threshold. 3.4.6 Excise tax Excise performance and structure The collection of excise tax in Kazakhstan has been consistently low. Excise tax collection lags behind regional comparisons, driven by a limited tax base, a low-rate structure, and poor compliance. From 2005 to 2019, Kazakhstan’s excise stagnated, falling well below the inter-quartile range of excise in ECA countries (Figure 3‑23). In 2019, excise in Kazakhstan accounted for just 0.5 percent of GDP, compared with the average of 2.6 and 2.7 of GDP, respectively, attained by the two comparative groups of emerging market economies and high-income countries in the ECA. The composition indicates the sensitivity of the excise tax structure in Kazakhstan to collection from alcohol and tobacco (Figure 3‑24). At the same time, other excisable items are relatively undertaxed (in 2017, 77 percent of total excise was derived from tobacco and alcohol). Figure 3‑23. Excise collection in regional Figure 3‑24. Composition of excise tax (%) comparison (% of GDP) Source: World Bank staff calculations based on data published Source: (OECD 2020). by the authorities. The excise tax base is limited to tobacco and alcoholic beverages (e.g., spirits and alcoholic beverages, beer, and alcohol-containing medical products). Other products are also subject to excises, such as petroleum products (e.g., crude oil, gas condensate, petrol/gasoline excluding aviation fuel, and diesel fuel) and motor vehicles (Table 3‑5). 74 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Table 3‑5. Summary of the excise tax regime Type of excisable good Excise tax rate Crude oil, gas condensate, petrol/gasoline, diesel KZT 0 to KZT 24,935 per ton Alcoholic beverages and beer, tobacco KZT 0 to KZT 9,900 per item of measure (kilos, liters, or units) Passenger cars KZT 100 per each cm³ of engine capacity Source: (PWC 2021). Environmental and GHG emissions (carbon tax) in the excise tax structure Despite the importance of climate policy for national commitments and strategies, Kazakhstan has made no extensive use of tax to influence GHG emissions or encourage positive environmental outcomes. Such a climate agenda is reflected in the country’s Concept for Green Economy (2013), the National Development Plan 2025, and the NDCs codified in the Environmental Code. However, the existing tax regime, particularly excise tax, provides numerous exemptions. Coal is not part of the tax base, and certain fuel types are, de facto, excluded as well, as their tax rate is listed at KZT 0. The current excise tax on motor vehicles has certain undesirable features. It is based on car engine size and does not conform to an effective environmental tax. More specifically, this type of excise tax is not based on the types of fuel used, the age of the vehicle, or other factors that negatively affect the environment and thus cannot serve as a relevant proxy for GHG emissions or air pollutants. Other types of green taxes are omitted from the current excise tax structure. Examples include a form of a road tax; and taxes on plastic bags, old tires, and waste, in combination with strong regulatory and/or complementary fiscal measures.63 Road tax is another potential for consideration – that would place a levy on motor vehicles with a rate structure that considers the type and age of a vehicle. A comprehensive road tax in the form of an annual registration fee can be charged based on several criteria: power output, the age of the vehicles, and other environmental criteria, such as CO2 g/km output. The recent trend in restructuring the motor vehicle taxation in OECD presents a good practice for countries like Kazakhstan to emulate (Box 3‑3). 63 The complimentary regulator and fiscal measures would consist of those that target illegal dumping and fiscal subsidies to incentivize activities in garbage collection and recycling. 3. IMPROVING DOMESTIC REVENUE MOBILIZATION 75 Box 3‑3 OECD Practices in Taxation of Motor Vehicles Most OECD countries levy comprehensive taxes on vehicle purchase, ownership, and usage. While these taxes have become an important source of tax revenue for many governments, countries have integrated environmental and climate objectives into these instruments over time. It is recognized that well-designed taxes can effectively reduce pollution and greenhouse gas emissions. Taxes and charges on vehicles mainly include (i) Taxes on the purchase (including VAT and retail sales taxes) and registration of motor vehicles; (ii) Periodic taxes payable in connection with the ownership or use of the vehicles; and (iii) Taxes on road fuels (an implicit form of carbon pricing). In most countries, total taxes represent a mix between ad valorem and ad quantum taxes. The main criteria for assessing these taxes can include: • The price or value of the vehicle; • The direct environmental impact, i.e., CO₂ emissions and other polluting emissions; • The characteristics of the vehicle, such as the type of fuel used, the weight, the cylinder capacity • , and the engine power. These may be indirectly connected with polluting emissions but were • generally not introduced for environmental purposes; • Social considerations incl. preferential treatment of emergency vehicles, ambulances, vehicles for • people with disabilities, vehicles for public transport, etc.; • The private or commercial use of the vehicle; • The specific features of vehicles for transporting goods, such as the number of axles, cargo room, seats, etc. Some countries apply “feebates,” i.e., rebates or fees, depending on whether the vehicle exceeds a certain emission threshold to incentivize the purchase and use of fuel-efficient vehicles. Source: OECD 2022. Consumption Tax Trend 2022. Broadening the excise coverage to all fossil fuels can be a good early action to improve budget resilience against green transition. As discussed in Chapter 1, a recent study (World Bank 2022) suggests that a gradual increase in excise rates on all fuels to 25 percent of the level specified under the EU directive will gradually increase tax revenue to 4 percent of GDP by 2030. The study also suggests that recycling 40 percent of the excise revenue as cash transfers for the bottom 40 percent of the income distribution could more than offset the negative impact of a higher fuel price on their consumption. Tobacco taxation OECD 2020 notes that the collection from tobacco accounts for more than half of Kazakhstan’s total excise tax. Global evidence also suggests that, while tobacco excise receipts vary across countries, tobacco contributes a significant and stable revenue source (IMF 2016). Kazakhstan has significantly improved taxing tobacco, though further reforms are needed. Kazakhstan became a party to the World Health Organization (WHO) Framework Convention on Tobacco Control on April 22, 2007, and issued regulations to control and discourage smoking. Before 2014, Kazakhstan followed a policy of moderate tobacco excise increases (particularly from 2005 to 2013). The policy resulted in neither significant increases in revenues nor a reduction in tobacco consumption. The mean annual increase in nominal tobacco revenues between 2010 and 2013 was about KZT 8 billion. The sharp tax hike (by 94 percent) in 2014 was, in contrast, a success: not only did it reduce tobacco sales by 9 percent, but it also brought an additional KZT 25 billion (about US$150 million) in government revenue (World Bank 76 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 2019). Despite some improvements in recent years, the price of cigarettes remains extremely low by international standards. As a result, the affordability of cigarettes indicators in Kazakhstan are among the most disadvantaged in the European region (World Bank, 2021). The World Bank (2021) assesses the possible impacts of three alternative reform scenarios in 2022–2024 on revenues, smoking prevalence, and the number of people who stop smoking (Table 3‑6). • Scenario 1: (80 percent of the Eurasian Economic Union (EAEU) indicative rate by 2024): KZT 15,300 per 1,000 cigarettes by 2024, or 11.5 average annual excise growth in 2021–2024; • Scenario 2: (120 percent of the EAEU indicative rate by 2024): KZT 17,000 per 1,000 cigarettes by 2024 or 30 percent average annual excise growth in 2021–2024; • Scenario 3: (optimal according to the WHO guidelines): KZT 25,000 per 1,000 cigarettes by 2024 or 50 percent average annual excise growth in 2021–2024. Table 3‑6. Impacts the reform scenarios, 2020–2024 The number Share of Excise Smoking of people who Average retail excise tax duties prevalence (% quit smoking, Kazakhstan price (KZT) in the retail (KZT of the adult ’000 people price (%) billion) population) (compared with 2020) Initial situation, 2020 508 3 9 218 19.9 0 Scenario 1, 2024 570 4 5 275 19.7 31 Scenario 2, 2024 720 54 365 18.6 139 Scenario 3, 2024 930 6 4 440 17.5 334 Source: World Bank. Analysis of the cigarette market and the prospects for excise policy in Kazakhstan. 2021. Note: All financial figures are adjusted for inflation, i.e., in constant prices (2020). 3.5. Fiscal progressivity of the tax policy Compared with its peers, the direct taxes in Kazakhstan are less progressive. Drawing on findings from Bokurnova and Nebiler (2023), the lack of progressivity in Kazakhstan’s tax policy is more pronounced in the PIT regime. Inherently, the progressivity of PIT in Kazakhstan is similar to that of the Russian Federation and Poland, both with a similar general flat rate structure. However, It is less progressive compared to the PIT regimes in other countries with built-in progressive tax brackets and exempt thresholds inducing low-income households to pay less than those in higher-income groups (Figure 3‑25). As for all the peer countries, in Kazakhstan, the indirect taxes (including the excise on tobacco and VAT) are regressive and have increased inequality. The study indicates, however, that if the potential negative impact on health is considered, excise on tobacco should no longer have a regressive impact on income. Moreover, in the cross-country comparison, income inequality falls in all countries after their respective governments collect more taxes and spend better on health and education and for targeted transfers. 3. IMPROVING DOMESTIC REVENUE MOBILIZATION 77 Figure 3‑25 Progressivity of direct taxes in selected countries Source: Argentina (Rossignolo 2017); Brazil (Higgins and Pereira 2017); Chile (Martinez-Aguilar et al. 2016); Croatia (Inchauste and Rubil 2015); Mexico (Scott 2013); Poland (Goraus and Inchauste 2016); the Russian Federation (Popova et al. 2018); the United States (Higgins et al. 2018); and Turkey (Cuevas et al. 2020). Kazakhstan: Bornukova and Nebiler (2023) based on Kazakhstan 2021 HIES. Note: Marginal contribution to equality is the difference between the Gini coefficient without the particular fiscal intervention, and the Gini coefficient of all income components together. There is no marginal contribution calculated for Argentina. The United States results are preliminary. 3.6 Policy Recommendations Unless consistent reforms are undertaken, Kazakhstan may be unable to reach the ambitious tax collection goal of 25 percent of GDP by 2025.64 Analysis from this chapter highlights elements of tax policies that must be aligned with international best practices and adjusted to attain the multiple objectives of a robust tax system: revenue adequacy, efficiency, neutrality, equity, and responsiveness to environmental and climate change challenges. This chapter proposes major time-bound tax policy reforms that would complement the World Bank-financed project on Tax Administration.65 3.6.1 Policy recommendations for corporate income tax Systematically restructuring the incentives The Government’s Draft Concept Note on Public Finance Management until 2030 envisions the possibility of lowering the CIT rate and differentiating rates by industry. Two suggestions are proposed, as follows: First, it is better to rebalance the structure of tax incentives to attain multiple (seemingly trade-off) objectives rather than bringing down the statutory rate. To attain the established goal of collecting the tax revenues in the order of 25 percent of GDT, the first-order imperative is to cut down the current high level of tax expenditures (estimated to be at 7.8 percent of GDP revenue foregone from main direct and indirect 64 The IMF Report (2020) recognizes the positive outcomes from measures to enhance VAT administration (e.g., e-invoice) and customs in 2019, but also recommends further strong policy and administrative reforms. 65 The four objectives of the Kazakhstan Tax Administration (TA) Reform Project are: (i) to improve the level of voluntary compliance; (ii) enhance effectiveness to fight tax evasion, (iii) increase administrative efficiency; and (iv) reduce the potential for corruption. 78 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH tax expenditures).66 Tax incentives should be conducive to extensive and intensive marginal investment without imposing an undue fiscal burden on the fiscal budget. The current CIT rate is comparable with the regional average (albeit low by global comparisons). Any inducements, such as a lowering of the tax rate, are likely to be nullified, as such efforts would only trigger subsequent rounds of rate reductions by other countries across the region, creating a vicious cycle of a ‘race to the bottom.’ Second, taxation should adhere to the key principle of sectoral neutrality. Differentiating the CIT rates by sector or industry would make the regime more complex and less efficient, further diverting the regime from one of the key principles in taxation. CIT’s existing tax incentive structure is overly generous and compounded by transparency challenges. The Government’s draft Concept Note on PFM until 2030 estimates that the effective CIT rate has fallen too low, varying from just 2.9 to 3.8 percent in recent years. This underscores that generous fiscal incentives drive the significant gap between the standard and the effective rates. Drawing on the lessons from global tax reforms, the following would attribute to the key decision-making ‘points’ in reforming tax incentives under consideration. Here we will focus on three sets of priority policy actions, as follows: (i) Arranging for effective cross-agency review and legal development. Over the short to medium term, with tax incentives spread in different legal documents (primarily the Tax Code and the Entrepreneurial Code), the improved governance of fiscal incentives would be instituted based on close collaboration and coordination between agencies responsible for granting and monitoring incentives. Such an effective mechanism will enhance the likelihood that incentives are aligned with the Government’s strategic investment priorities and harmonize the tax incentives and non-tax regulatory provisions to achieve an improved investment environment. While line ministries are keen to seek more fiscal incentives for investments under their purview, reviewing such proposals needs to be centralized, cost-effective, conducted, and aligned with the national budget perspective. A tax expenditure review would clarify the extent of revenue forgone from fiscal incentives. Over the long term, the country could consider consolidating all tax incentive provisions under the dedicated Tax Code. (ii) Rebalancing the types of incentives with support from cost-benefit analyses. The existing tax incentive regime relies heavily on profit-based instruments (especially reduced rates and outright tax holidays). These are regarded as crude incentive instruments that largely favor already profitable investments and are costly with unclear net benefits (in terms of inducing intended marginal investments). Even worse, they (especially tax holidays) tend to incentivize so-called fly-by-night investments (typically applied to those labor-intensive investments). The IMF, the OECD, and the World Bank (2015) recommend that countries shift from profit-based to cost-reducing tax incentives. The latter would relate to the instruments such as accelerated fiscal depreciation (and even investment tax credit), tax deduction or credit for local workforce training, and R&D. These incentive types could be specifically tailored to those investments that best fit with the NDP 2025. In addition, all incentives need to be set in a time-bound manner with clearly defined “sunset” provisions. (iii) Instituting tax expenditure analysis. Kazakhstan should institute a tax expenditure analysis (TEA). The annual TEA report should become mandatory and be regularly incorporated in the Government’s budget report to the public/parliament. 66 The World Bank plans to provide follow up TA support in institutionalizing and building capacity for tax expenditures analysis. 3. IMPROVING DOMESTIC REVENUE MOBILIZATION 79 Reform the special/simplified regime: The co-existence of three special tax regimes with the mixture of annual turnover and the number of employees makes the system overly complicated and creates loopholes for bunching and avoidance. The fix, however, is straightforward, with a key two-step process. In the first step, structure the special tax regime using annual turnover as the single eligibility criteria. And in the second step, apply the same VAT threshold to classify those that migrate into the standard tax regime and those that remain in the special regime. Under the special regime’s detailed structure (measures to be implemented from year 1 to year 3) is worth noting. • Exempt from CIT are micro-sized household businesses with an annual turnover of less than the current PIT personal allowance/basic deduction of KZT 510,000. As recommended further, such a PIT basic deduction should be raised to match at least the per-capita income. If such an adjustment is made within the PIT regime, the exemption threshold for micro-sized household businesses should also be raised accordingly. Such a proposed exemption threshold for micro-sized household businesses, if instituted, would attribute to equitable tax treatment between salaried and non-salaried income earners. • Businesses (individual entrepreneurs or legal entities) with an annual turnover of between the PIT basic allowance of KZT 510,000 and the VAT threshold of 20,000 MCI (applicable from 2022) will be subject to a simplified declaration (currently set at 3 percent of gross income without deductions). These businesses are also allowed to migrate into the standard regime voluntarily. From year 4, consider raising the gross income-based rate from 3 to perhaps 5 percent.67 Such a policy measure would help to balance short and long-term objectives: reducing the compliance and collection costs for truly small household or business entities and incentivizing them to migrate into the regular regime. 3.6.2 Recommendations for PIT Raise the basic deduction and introduce a progressive tax structure to replace the current low flat rate. The current annual basic deduction (KZT 510,000, equivalent to US41,041), at just 11 percent of 2020 per capita income, would be increased to target a dual objective: protect the poor and enhance the compliance of high-income earners. We propose that the Government raise such an income exemption threshold to around US$9,000 (almost equivalent to the 2020 per capita income). On the one hand, such a measure would exclude the low-wage earners from the PIT regime (hence, likely reducing income tax from this group of individuals). On the other hand, it would free up the State Revenue Committee (SRC) resources to focus on enhancing compliance among those earning higher incomes (thereby resulting in higher PIT collections). Over the mid-term, the current flat rate could be transformed into a progressive, graduated structure consisting of perhaps 4–5 brackets, including the exempt bracket. The reform would help raise revenues from this important tax instrument and improve equity. With a uniform tax rate of 10 percent on all sources of capital income, including dividends (recommended below), we would estimate that the highest marginal rate of the new PIT structure could be set at 28 percent, to harmonize it with the CIT standard rate.68 In concurrence with the PIT rate restructuring, it is necessary to institute an annual review and adjustment for inflation of the exempt bracket, deductions, and income credits. 67 Simply assume the average profit margin of 25 percent. Applying to such profitability the standard CIT rate of 20 percent would make the equivalent gross income tax rate 5 percent. 68 The simple approximate determination of the highest marginal PIT rate that harmonizes with the CIT and dividend rates is formulated as equivalent to the sum of the CIT rate and the dividend rate on the corporate profits (net of CIT). 80 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Review and rationalize the income tax exemptions in PIT. The existing set of numerous exemptions would violate the standards of a comprehensive income tax base. They deplete the base and compromise the integrity of the regime. Over the short term, reviewing and rationalizing the multiple exemptions is sensible, laying the ground for the mid-term transformation of the flat PIT rate into a progressive structure. Capital income taxation (i) Interest income treatment Interest income could be uniformly taxed. The introduction of an interest income tax in no way levies excessively on low-income households or individuals: the interest income can be lumped together with other sources of income subject to the basic exemption or deduction. To put the savings income and investment income on an equal footing, it is recommended that interest income be subject to a rate equal to, or at a lower rate than, legislated for the capital gains or dividend rates (initially, within the first year of the PIT review, the rate of 5 percent—the same lower rate on dividends sourced from Kazakhstan—would be a sensible policy choice. The decision to concurrently raise the capital income tax rates from different sources should be considered over the medium term). (ii) Capital gains taxation A balanced approach to capital gains taxation would help mitigate speculative behavior and ensure tax collection due to long-term gains. We would suggest a two-tier capital gains tax structure applica- ble to both financial and immovable asset transactions: • For short-term holdings of capital assets (typically gains from transactions made within one year from the date of acquisition), the full rate (similar to the standard CIT rate of 20 percent) should be levied. • For long-term holding of securities or immovable assets (those obtained and held for at least two years): A lower rate (at the recommended current capital gains tax rate of 10 percent) should be levied. This new provision will abolish the exemption granted to gains from assets owned for over three years. The recommended tiered structure would simplify the capital gains tax regime, more equitable, effi- cient, and less costly to the government budget. (iii) Unify the rate of capital income from different sources A special rate would be applied to various sources of capital income: interest, dividends, and capital gains. The currently low flat PIT rate (10 percent) can be applied to these capital incomes. (This would mean that the rate will be levied on all dividends, regardless of whether they are derived from foreign or domestic sources). Over the mid-term, while the flat PIT rate of 10 percent will be transitioned to a progressive structure with the highest marginal rate (to stand substantially higher, as suggested above, at 28 percent), the unified 10 percent rate on capital incomes should be retained. This system would constitute a dual- income tax regime. Contemplate introducing the wealth or inheritance tax, but only in the medium term. The new net wealth or inheritance tax could be considered in the mid-term. The compliance with the current annual recurrent property tax could be enhanced in combination with the plan to institute a new asset tax (luxury property or vehicles) that is well suited to proximate the wealth tax and make the tax system fairer. (As proposed, reform priorities should continue focusing on introducing the progressive rate structure over the next three years, rationalizing the exemptions and other tax expenditure provisions, and enhancing the clarity and efficiency in taxing capital incomes). 3. IMPROVING DOMESTIC REVENUE MOBILIZATION 81 3.6.3 Policy recommendations for VAT The key policy recommendations for further reforming the VAT regime focus on broadening the tax base and rationalizing the VAT rate structure. The first (broadening the VAT base) would consider a mixture of provisions to lower the threshold, reduce the number of exemptions, and extend the VAT coverage of e-commerce and digital transactions. The second set of policy actions (rationalizing the rate structure) would target eliminating all non-export zero ratings and raising the standard VAT rate. Lower the VAT threshold. Lowering the VAT threshold over the past year is advancing international best practices. Further reductions of the VAT threshold are recommended to expand the VAT net and use it to harmonize with the threshold defining small and medium enterprises (SMEs), the group of taxpayers granted special tax regimes. Setting a significantly higher VAT threshold for individual entrepreneurs using special tax regimes (at 114,184 MCI) is redundant and inefficient. It is recommended that this provision be abolished in favor of a single, uniform VAT threshold. Streamline the VAT exemptions. The current expansive list of the 47 groups of exempt commodities needs to be reviewed and streamlined. Exemptions violate the integrity of VAT, while the benefits for the poor or their revenue impacts are unclear. Consumers would benefit from lower prices, with VAT relief only if the exemption falls into the final retail sales stage. If granted elsewhere in the middle of the production- distribution chain, exemption breaks the VAT chain. It creates unintended, inefficient cascading effects (like a turnover tax), making the final consumers pay more. International best practice would suggest that exemption be granted only to basic education, health services, and selected financial transactions. As for the treatment of SMEs, the VAT threshold would exempt them from the VAT net or allow them to join (currently already practiced in Kazakhstan) voluntarily. In Annex 1, we propose, with justification, a list of currently exempted commodities or groups of commodities to be either transferred outright to standard VAT or retained but with clear legally binding definitions (leaving no room for liberal legal interpretation and avoidance). The streamlining and rationalizing exemption requires careful, staged wise processing over the suggested timeline of five years accompanied by periodic (annual) reviews. Also, as part of the reform action to broaden the VAT net and ensure the equitable treatment of all sorts of transactions, the VAT regime could cover digital and e-commerce (in line with the 2017 OECD International VAT/GST Guidelines). Limit zero rating to exports only. Following the destination principle, VAT would grant all exports (commodities and services) a zero rate. Additional coverage of zero rating would risk creating an extended backlog of refund claims, depressing revenues, and, thereby, VAT productivity. Annex 3 recommends removing specific items from the current zero-rate listing. The review and rationalization of the zero-rate regime would be done in parallel with a review of VAT exemptions. In addition, we recommend that special VAT treatments for SEZs be eliminated. Specifically, all transactions into or within SEZs will be subject to standard VAT (including applying zero-rating exclusively to exports). Raise the standard VAT rate but only over the long term. The current standard VAT rate of 12 percent is significantly below those in regional and global comparisons. It is worth noting that the rate has consistently been reduced over time: from 20 percent (until 2000) to 16 percent (during 2001–2004), 15 percent (2005–2006), 14 percent (2007), 13 percent (2008), and 12 percent (from 2009). One policy option to consider (as part of the broader strategy to raise the non-oil revenue intakes) is to reverse the declining standard rate trend. While the rate increase may be out of policy discourse over the short to medium term, it can be considered part of the comprehensive review of major tax policies over the long term. From an economic and social perspective, such a policy consideration is advisable gradually, step-wise (for example, the initial increase would be to perhaps 14 or 13 percent, as previously applied in or before 2008). We underscore that the rate adjustment be contemplated only after other more imminent measures to streamline the exemptions and eliminate the non-export zero ratings are legislated. 82 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 3.6.4 Policy recommendations for excise Excise tax reform would require a periodic review of the tax base and rate structure. The net should be widened (including a broader range of fuel/energy sources and new items such as sugary drinks, tires, and plastic bags). At the same time, the rates should be rationalized to mitigate the risks to public health, pollution, the environment, and climate and to raise additional revenues. Expand the scope of carbon pricing. The case for a carbon tax is clear. Kazakhstan has made commitments through Nationally Determined Contributions (NDCs) to adhere to the Paris Agreement in mitigating the impact of climate change. The Government’s PFM Concept Note specifies further that Kazakhstan must prepare to adapt to its carbon pricing mechanism in the face of the imminent EU border carbon tax adjustment.69 The NDCs determine that the country will introduce a carbon tax on energy consumption not covered by the Emission Trading System (ETS) and raise the rates over time. The environment and carbon tax could begin by reviewing the existing excise coverage of fuels with the purview of expanding the base to cover all fuel sources and rationalize the rates. The proposed carbon tax complements the current ETS – and the two would help create a minimum and maximum carbon price. Such policies align with the Government’s vision of achieving carbon neutrality by 2060.70 Taxing motor vehicles. The missing criterion for the excise on motor vehicles is their age. Older vehicles exert higher a negative impact on the environment, road congestion, and traffic accidents. As such, the rationalization of the rate structure may internalize vehicle aging and punitive (higher) rates should be applied as a progressive structure (older vehicles are taxed at higher rates). The Government should also consider aligning the tax on motor vehicles with good practices such as those implemented in OECD countries. Enhancing health outcomes and raising revenues through tobacco excise. The outcomes of the modeling in the World Bank 2021 report show that only a significant increase in excise taxes on cigarettes can impact the prevalence of smoking and, subsequently, contribute to a decrease in mortality. The low elasticity of tobacco demand due to its addictive nature requires greater determination and consistency from national excise policies. Thus, the 2021 report recommends a faster increase in excise tax, corresponding to the WHO recommendations (or the maximum value of the indicative excise rate according to the EAEU plans). Further World Bank support for tax policy and tax administration The World Bank and the Government of Kazakhstan should further their collaboration in various analytical and operational activities on the wide-ranged tax reform agenda. The Tax Administration Reform Project can be emulated with the next phase of operation supporting the State Revenue Committee (SRC) to enhance the quality of voluntary compliance, aligned with the ongoing or planned government reforms, such as e-filing, VAT refund management, and extended coverage of digital transactions. Potential targeted areas of the extended support would include support in tax expenditure analysis (TEA), health taxes (taxation on tobacco and sugary drinks), equity and gender equality tax policy issues, and international taxation (covering, in a phased approach, the actions under the BEPS). 69 The European Commission provides the following: ‘On 14 July 2021, the Commission adopted a proposal for a new Carbon Border Adjustment Mechanism which will put a carbon price on imports of a targeted selection of products so that ambitious climate action in Europe does not lead to ‘carbon leakage.’ This will ensure that European emission reductions contribute to a global emissions decline, instead of pushing carbon-intensive production beyond Europe. It also aims to encourage industry beyond the EU and our international partners to take steps in the same direction. Please see (European Commission n.d.). 70 (Strategy 2050 2021). 3. IMPROVING DOMESTIC REVENUE MOBILIZATION 83 4. Public Spending on Education 84 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH KEY POINTS • During most of the first three decades after independence, the country had vastly underinvested in its public education system while pursuing a policy of promoting an elite school system, much to the detriment of developing a solid human capital base for the country. Today, about two-thirds of Kazakhstan’s 15-year old’s are not prepared to succeed in the modern world of work (PISA, 2018), and recent assessments indicate a worsening of outcomes post CoVID-19. The demands on Kazakhstan’s education system are increasing, both for higher quality and relevant education and skills development due to a rapidly changing world of work and for catering to a growing and changing the concentration of student population due to a population boom and in-migration in certain cities and regions. Kazakhstan has only recently increased public spending on education to meet these demands. The increase must be complemented with measures to enhance efficiency, effectiveness, and equity to ensure that limited public resources are impactful. • Efficiency-enhancing measures include optimum allocation of existing and planned school infrastructure and teacher redeployment from schools with low student-teacher ratios to larger schools facing a shortage of time for teaching and learning. • Improvement in teacher effectiveness, expected to improve the returns from educational investment, include reforming the teacher compensation system (Stavka) to encourage “better” rather than “more” teaching time, incentivizing professional development, and measuring the impact of the multiple teacher reforms (salary increases, optimizing teaching hours, and capacity building). • An equity coefficient in the per capita financing, complemented with reforms to encourage teachers to dedicate efforts to attain learning gains amongst the low performers (rather than high performers), can go a long way in bringing up the skills deficit across the country, including students from disadvantaged backgrounds. 4.1 Motivation The Government of Kazakhstan prioritizes enhanced access to high-quality education. After a period of impasse following independence, substantive reforms were introduced by the Government at the start of the 21st century to assess and improve educational outcomes in the country. These included setting up national assessments of learning to monitor the education system’s effectiveness,71 participation in international assessments to benchmark Kazakhstan globally on key learning outcomes, and the modernization of the school curriculum in 2015. More recently, the Government has invested in expanding early childhood education, raising teacher salaries in line with market norms, introducing per-student financing, and expanding school infrastructure to meet the needs of a growing student population and the demands for a higher quality of education and skills nationwide. 71 The National Testing Center (NTC) was set up in 1993 to develop a coordinated admissions policy and to report directly to MOES (as the republican budget funded most of higher education). The first versions of UNT emerged in 1999 and after a few trials, a compulsory national version was adopted in 2004. 4. PUBLIC SPENDING ON EDUCATION 85 However, Kazakhstan’s education system has yet to deliver quality education. Almost three decades of underinvestment in education, complemented with public finance and educational policies that incentivized academic excellence amongst the top-performing students (who are increasingly more likely to come from higher socio-economic quintiles), has created a weak and unequal educational and skills base across Kazakhstan. The Program for International Student Assessment (PISA) 2018 found that two- thirds of 15-year-olds were functionally illiterate (i.e., unable to operate in the modern world). There are also substantial inequalities in access to preschool and higher education across socio-economic quintiles and substantial socio-economic achievement gaps in learning outcomes. Furthermore, COVID-19-induced school closures have led to large learning losses (close to one year of schooling), which, if not remedied, may generate economic losses valued at over US$4 billion per year.72 The system is under immense pressure to deliver the promise of education to a larger student population widely spread across the country. A baby boom at the start of the 21st century, coupled with continued increases in birth rates in regions with the lowest levels of human capital, and increased migration to urban areas, with three-quarters of secondary schools operating in multiple shifts, has created significant demand for additional school spaces across the country. The economy also demands skills supporting diversification, high-tech skills (54 percent of the jobs are expected to be lost to automation), and “green skills” that could enable Kazakhstan to achieve a green transition. While necessary, recent increases in education financing require a critical review regarding their sustainability, effectiveness, equity, and efficiency. While Kazakhstan spent significantly less on education compared with the average spending levels of middle-income countries for the first 28 years post-independence, the real education budget recently increased by 48 percent between 2018 and 2020. Education spending as a share of GDP has risen to 4.4 percent, approaching the upper middle-income country average of 4.9 percent.73 In this context, and with the economic and social needs to develop a skilled labor force for a green transition that can succeed in the technology-rich world of work, this PFR chapter looks at the adequacy, efficiency, and effectiveness of education spending. Key questions include: How can Kazakhstan maximize results from increased education spending? What are some of the inefficiencies that the system needs to address immediately? How can education spending become more effective? How can education spending become more equitable? 4.2 Overview of the Education System in Kazakhstan 4.2.1 Governance and financing of education service delivery Two ministries manage education service delivery. Pre-tertiary education (pre-school, primary, and secondary) is free and compulsory in Kazakhstan.74 The Ministry of Education and Science (MoES) was split into two ministries in June 2022, namely, the Ministry of Enlightenment/Education (MoE), responsible for pre-tertiary education, and the Ministry of Science and Higher Education (MoSHE), responsible for tertiary education, science, and innovation. The MoE is responsible for planning, policy-making, and monitoring outcomes for pre-tertiary education, and SNGs are responsible for implementing education services through funds channeled directly to them. Education service delivery is spread across the public and private sectors. The public sector is the predominant primary and secondary education service provider, accounting for 97.5 percent of general 72 World Bank. Equitable Human Capital Development in Kazakhstan: Skills Formation During the Foundational Years (English). Washington, DC: World Bank Group. http://documents.worldbank.org/curated/en/099083002082316985/P17487902c284204f086310a12db02babf4 73 World Bank [forthcoming]. Education Finance Watch 2022. 74 The Constitution of the Republic of Kazakhstan (1995) and the National Law on Education (2007) ensure that citizens of Kazakhstan receive free preschool (from the age of 5), primary, basic secondary, general secondary, technical, and vocational education. 86 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH secondary schools and educating 98.7 percent of students. However, private preschools educate 43 percent of preschool students, and private colleges train 42 percent of total vocational students in the country, several through publicly financed vouchers and scholarships. Of 116 higher education institutions in the country, 84 are private, while 63.1 percent of all higher education students study self-paid.75 Pre-tertiary education services are delivered primarily through regional governments and financed through various means. SNGs implement 86 percent of the state’s total education budget (2020) (Figure 4‑1). Interestingly enough, education accounts for the highest share of subnational expenditures (35 percent on average in 2020), with several regions spending as much as 40 percent of their total budget on education service provision (Figure 4‑2). Regional spending on education is largely financed from each region’s general revenue and, as discussed in Chapter 7, is derived from two major sources. First is revenue from taxes assigned to the regions, mainly PIT and social tax. The second is from subventions, calculated as the difference between each region’s expected tax revenues and spending needs.76 In addition to financing by the regions, the MoE also provides targeted transfers to regions for specific developmental expenditures. Figure 4‑1. Amount and share of regional Figure 4‑2. Share of regional budget spent on expenditures on education, 2010–2020 education, 2020 Source: World Bank staff calculations based on data published by Source: World Bank staff calculations based on data published the authorities. by the authorities. The estimate of the total regional budget excludes transfers from the regional budget to the Republican budget. Regional budget planning processes are complex, with a high risk of disconnect from national and regional plans. As documented in Figure 4‑3, determining regional budgets is extremely complex. Yet, regional needs are at a high risk of being left out of budgeting. A review of the Regional Education Plans 2021–2025 (which are part of the Regional Development Plans 2021–2025) shows a disconnect between the regional plans77 and the NEP goals and activities for three (Atyrau, Mangystau, and Turkestan) of the four researched regions in this PFR. Interviews with regional education departments suggest that these disconnects imply that requests for activities submitted by rayons and schools remain largely unmet. In addition, subventions are estimated by the MoE three years in advance. They are not subject to amendments, which burdens the MoE considerably to make precise forecasts. Furthermore, estimates of 75 IAC 2021. Table 2.7.9 76 In the cases of Almaty City, Astana, Atyrau and Mangistau, estimated revenues exceed estimated expenditure needs. These four regional governments are therefore subject to withdrawals, rather than subventions. 77 Five-year regional development plans (including regional education plans) are developed over a six-month planning cycle. Inputs are obtained from all rayons and regional akimat departments work with the respective ministries at the central level to finalize activities under the regional plan. 4. PUBLIC SPENDING ON EDUCATION 87 general transfers are based on current costs and student numbers, as opposed to information updated by schools in the National Education Database (NEDB), which contains information on the existing resources and needs at the school level. Figure 4‑3. Budget development process for regional executive budgets for education Source: World Bank staff analysis using the Budget Code and interviews with regional government officials. Kazakhstan is expanding the implementation of per-capita financing, which has substantial potential to streamline financing procedures, improve efficiency, and reduce persistent regional and socio- economic inequities. Introduced in 2013, per-capita financing (PCF) has expanded considerably since 2018, covering 91 percent of urban public schools and almost all private schools in Kazakhstan. In 2021, the PCF system provided KZT 73.5 billion (US$477.5 million in PPP) to public general secondary schools and KZT 39.6 billion (US$257.2 million in PPP) to private general secondary schools. The PCF system has substantially facilitated the involvement of the private sector in increasing access to preschools. Currently, the PCF Law finances the cost of a limited set of education expenses (purchase of manuals, school furniture, sports equipment, and maintenance of educational infrastructure), which are fixed annually at low nominal monetary values per student, equivalent to just US$68 in PPP for preschool and US$22.7 in PPP for general education. There is a lack of clarity on the allocation and utilization of significant amounts of financing. Over one-quarter (27.5 percent in 2020) of total local education expenditures are classified as ‘other’ expenses without clarification, limiting the transparency of spending and increasing governance risks. Acquisition of services and works (utilities, communication, transport, rent, and other services) and purchasing goods (food, medicines, fuel, etc.) accounted for 30 and 2 percent of the total current expenditures on education, respectively, in 2020. However, almost all expenses under ‘acquisition of services and works’ is classified as ‘payment for other services and works’ without clarity on the services and works procured. This lack of transparency in reporting expenditures can substantially limit the ability of schools, subnational education departments, and the MoE to ensure the effectiveness and efficiency of education expenditures and engage in long-term financial planning for making qualitative improvements in education. 88 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 4.2.2 Education outcomes Kazakhstan has almost universal access to primary, lower, and upper-secondary education, while preschool education has increased. Enrolment rates78 are close to 100 percent for Kazakhstan’s primary, lower-secondary, and upper-secondary education.79 The enrolment rate for preschool has significantly increased over time, although at 70.8 percent, it is well below universal.80 The country’s success in ensuring near-universal access to primary and secondary education is reflected in its strong performance on school attainment, as measured by expected years of schooling, which stood at 13.7 years in 2020—above the OECD average of 13.4 years. The education system performs well in providing basic curricular knowledge in primary education, but gaps exist in attaining high and advanced levels of knowledge. The education system in Kazakhstan performs well in delivering core content knowledge. However, weaknesses begin to appear in the performance of complex tasks in primary education and at the secondary level. The gaps emerge in primary education and accumulate over time by grade 8 when a higher percentage of students perform at or below a low- proficiency benchmark in mathematics and science compared with the ECA average. Sixty-four percent of students (aged 15) perform below minimum proficiency in reading compared with the ECA average of 31 percent and the OECD average of 23 percent on the PISA tests that assess the application of acquired knowledge to real-life problems. Similar gaps exist in student performance in mathematics and science modules in PISA. Furthermore, the percentage of students performing below basic proficiency increased from PISA 2012 to PISA 2018 in reading, mathematics, and science (Figure 4‑4). In addition, students in Kazakhstan lack the basic digital skills required to contribute successfully to the future automated and green economy. According to International Computer and Information Literacy Study (ICILS), 54 percent of grade-8 students in Kazakhstan performed below Level 1 (the lowest level), compared with the ICILS 2018 average of 18 percent. There are also substantial differences in learning outcomes across regions in Kazakhstan. Figure 4‑5 below shows considerable differences in learning outcomes, with the Atyrau region performing lowest in the PISA assessment (around 84 PISA points, roughly equivalent to two years of schooling), despite being an oil- rich region.81 On the other hand, Astana and Almaty cities ranked the highest in learning outcomes, which may reflect various factors, such as better education infrastructure, higher local government capacity, the attractiveness of cities for talented teachers, a higher base of human capital, closer proximity to schools and more competition among schools.82 78 Total net enrolment rates [i.e., the total number of students of the official age group for a given level of education who are enrolled in any level of education, expressed as a percentage of the corresponding population] are used to understand access to education. If total net enrolment is not available, then adjusted net enrolment rates are used. If neither is available, then net enrolment rates are used. 79 World Development Indicators. 80 UNESCO Institute of Statistics. 81 Some regions that are rich in natural resources—as is the case for Atyrau and Mangystau—have lower indexes of human capital than less endowed regions. Such a paradox signals that the benefits of extracting natural resources in resource-rich regions are captured by others beyond the region. It also signals that investments in human capital should prioritize these regions because any increases in work skills would add value to the resources where workers live, thus helping regions capture the benefits of their own resources. World Bank 2023: Equitable Human Capital across the Regions of Kazakhstan. 82 There is a vast literature on this subject, including Gibbons and Olmo (2008) which finds that the proximity of primary and secondary schools within urban areas are strongly associated with slightly better learning results among students in the United Kingdom. A recent study in the Netherlands also finds that children who grow up in a more urban environment are more likely to select into schooling tracks that provide a higher level of human capital accumulation (van Maarseveen, 2021). 4. PUBLIC SPENDING ON EDUCATION 89 Figure 4‑4: Proportion of students’ performance Figure 4‑5: Differences in PISA achievement based on the PISA proficiency test across Kazakhstan’s regions Source: PISA 2018. The education system is also experiencing pressure from the substantial increase in the student population, combined with rural-to-urban migration. Over the past decade, the number of students attending preschool and school education has increased consistently (Figure 4‑6). Pre-school enrolment has more than doubled since 2010, and school enrolment has increased by around 25 percent. These enrolment increases are consistent with the steady increase in children aged 0 to 19, which is expected to continue until 2055 (Figure 4‑7). Population dynamics differ across regions and urban-rural locations, placing differential pressures on the existing school infrastructure. Urbanization in the regions and limited school infrastructure have meant that 75 percent of schools in the country operate multiple shifts,83 and the number of schools operating triple shifts has almost doubled from 2015 to 2020. Figure 4‑6. Number of students by level of Figure 4‑7. Student and youth population education, 2000–2020 projections, 1950–2100 Source: World Bank staff calculations based on data published by Source: United Nations, Department of Economic and Social the authorities. Affairs, Population Division (2019). World Population Prospects 2019, Online Edition. Rev. 1. Furthermore, gaps exist in the universal provision of basic inputs of educational materials such as laboratory equipment. Some schools even lack more basic inputs. For example, 4.3 percent (301 schools) have an undefined type of heating arrangement, 7.3 percent have only outdoor toilets that are not classified as warm, and around 10 percent do not have any library facilities. Kazakhstan has improved in these inputs over time. For example, the proportion of schools without access to an indoor toilet has fallen from 40.5 83 IAC Report 2020-2021 2.2.134. 90 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH percent in 2015. However, these remaining gaps in basic inputs must be filled to ensure an effective and comfortable learning environment for students. In addition, all regions visited highlighted the need for educational inputs such as laboratory equipment. 4.3 Education Expenditure, Trends over Time, and International Comparisons Kazakhstan has spent less on education relative to GDP than most countries. Public expenditure on education in Kazakhstan as a proportion of GDP declined by more than half, from 6.8 percent in 1990 at the time of independence to just 2.9 percent of a much smaller GDP (due to overall economic decline) in 1994.84 Kazakhstan has invested an average of 3.5 percent of its GDP in education over the past 30 years (compared with an average of 4.9 percent for upper-middle-income countries).85 Education financing has increased significantly in response to growing demands on the education system. Only recently, since 2018, has the education budget as a percentage of the total government budget and as a percentage of GDP increased. In 2020, the education budget stood at 19.4 percent of the total government budget and 4.5 percent of the country’s GDP (compared with 17.2 and 3.2 percent, respectively, in 2018). The education budget has also increased in real terms, especially since 2018 (Figure 4‑8). The real education budget increased 48 percent between 2018 and 2020, compared with 30 percent over the previous eight years (2010–2018). This increase in budget is positive and needs to be sustained given the increasing number of students that the education system needs to cater to.86 Figure 4‑8. State education expenditure as a percentage of total government budget and as a percentage of GDP Source: World Bank staff calculations based on data published by the authorities. The majority (81.5 percent) of the recent increase in subnational expenditures on education has been due to increases in current expenses. Real current expenses increased by 48 percent from 2018 to 2020 (Figure 4‑9), more than the increase over the previous eight years (38.6 percent between 2010 and 2018). Of the total real increase in current expenditures, 77.6 percent went toward salaries, 22 percent went toward acquiring goods and services, and 2.5 percent to current transfers. Despite receiving a smaller share of the increase, real capital expenses increased by 81 percent from 2018 to 2020. This is a major reversal in spending patterns, as real capital expenses declined by around 16 percent between 2010 and 2018. Of the total increase in capital expenditures, 54 percent went toward acquiring assets, 30 percent toward 84 PERrepairs, capital Kazakhstan (2000). and 15.7 percent toward new construction. 85 World Bank [forthcoming]. Education Finance Watch 2022. 86 Authors’ calculations using UIS Statistics. 4. PUBLIC SPENDING ON EDUCATION 91 Figure 4‑9 Local government budget on education, 2010–2020 Source: World Bank staff calculations based on data published by the authorities. Most of the recent budgetary increase went toward financing teacher salaries to enhance the teaching force’s professionalization. Teacher salaries were mandated to be more than doubled over four years (2020–2023) through annual increases of 25 percent.87 Perceptions of the status of the teaching profession in Kazakhstan have improved, with average teacher salaries now approaching market values across all levels, except for preschool teachers (Figure 4‑10). The average teacher salary in the country reached KZT  246,900 (US$1,603 in PPP) in 2022, compared with the national average salary of KZT  275,600 (US$1,789 in PPP). Actual salaries can also vary significantly from the average national salary. An analysis of teacher tariff tables from a few schools shows that the average salary for teachers in the select tariff table was KZT 303,000 (US$1,967 in PPP), rising to KZT 578,000 (US$3,752.5 in PPP) for the highest category of teacher (researcher) with 23 years of experience teaching 28 hours a week. Figure 4‑10. Salaries in the education sector, especially for secondary education, are converging to the national average (total) salary over time Source: World Bank staff calculations based on data published by the authorities. The National Education Project is expected to expand the education budget by 12 percent over its implementation period (2021–2025), with 78 percent of the funds going toward general education infrastructure. The National Education Project (NEP) will cost KZT  2.2 trillion (US$14.2 billion in PPP) from 2021 to 2025. A large majority of the funds are intended to expand the country’s general education 87 https://primeminister.kz/en/news/interviews/realizaciya-zakona-o-statuse-pedagoga-i-povyshenie-prestizha-kazahstanskih-uchiteley- intervyu-s-b-asylovoy 92 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH infrastructure (Figure 4‑11) to reduce the number of double- and triple-shift schools, which currently account for 75 percent of the schools in the country. In addition, 15, 14, and 4 percent of funds under the NEP are allocated to achieve the goals of vocational education, higher education, and preschool education, respectively. Box 4‑1 The main goals of the NEP, by education level • Preschool education - Attaining 95 percent coverage for children ages 2–6 years and training 92 percent of preschool teachers in improved preschool methodology. • School education - Improving student results in PISA and reducing the achievement gaps between rural and urban areas, transitioning to a 12-year education system, training 90 percent of teachers by 2025, training 30,178 teachers by 2025, improving the remuneration of teachers in ungraded schools and improving the educational environment and resources, and constructing 1,000 new schools and the modernization of 5,000 existing schools by 2025. • Tertiary and vocational education - Improving the coverage and quality of tertiary education by providing an additional 50,000 educational grants for TVET education and increasing the number of students covered under dual education to 101,000 (35 percent of college students) by 2025. • Higher education - Increasing the number and quality of higher education opportunities available in the country by establishing 20 academic centers of excellence and two universities, following the example of Nazarbayev University, developing a research and industrial university, and opening branches of leading international universities in Kazakhstan. Figure 4‑11. Allocation of funds under the National Education Project 2021–2025 Source: World Bank staff calculations based on data from NEP documents. 4.4 Equity in Education Expenditures Although Kazakhstan generally has spent less per student across all levels of education than countries with similar income levels per capita, per-student spending has been skewed in favor of the country’s elite school system. Kazakhstan spends less per student at all levels of education than most countries 4. PUBLIC SPENDING ON EDUCATION 93 at its income level (Figure 4‑12). However, the Government dedicates funding comparable to developed countries for students enrolled in the Nazarbayev Intellectual Schools (NIS). In 2020, the NIS received KZT 26,121 million, or about 2.7 percent of the Republican (central government) budget for education, as currently targeted transfers from the MoE. There are 20 NISs in the country and 14,697 students in these schools as of 2020, or about 0.4 percent of the secondary education student population.88 The per-student cost at the NIS is KZT 1.78 million (US$12,637 in PPP), which is equivalent to 4.3 times the per-student cost in regular primary, basic, and general secondary schools in the country (Figure 4‑13) and similar to the per-student costs in the Netherlands and Germany. While the budget allocated for the NIS has declined over time, from 14 percent of the central government budget on education in 2017 to 0.6 percent in 2021, the government may consider evaluating the costs and benefits of maintaining the NIS as part of the public education system. Figure 4‑12. Public spending on education per student Source: For Kazakhstan, World Bank staff calculations are based on data published by the authorities and enrolment data from the Informatic Analytic Center for enrolment statistics. For all other countries: OECD/UIS/Eurostat (2021) “Education at a Glance,” Table C1.1. and Annex 2 (Chapter C). See the Source section for more information and Annex 3 for notes (https://www.oecd.org/education/education-at-a-glance/ EAG2021_Annex3_ChapterC.pdf). TUR: Türkiye, LTU: Lithuania; RUS: the Russian Federation; CHL: Chile; MEX: Mexico; GRC: Mexico; COL; Colombia; KAZ: Kazakhstan. The differences in per-student spending between the NIS and mainstream public schools are associated with the divergence in learning outcomes between the two school systems. In the PISA 2018, NIS students significantly outperformed the OECD average in all three subjects. Figure 4‑14 shows that NIS students have significantly lower functional illiteracy (6 percent) than the 22.6 percent for the OECD average. In contrast, 64.2 percent of children in mainstream public schools in Kazakhstan are functionally illiterate.89 The NIS also seems to attract students of better socio-economic backgrounds compared with students in mainstream/regular schools. Students of low and middle socio-economic status are under-represented in NIS. About half of NIS students came from the highest quintile of the income distribution, compared with 15 percent for students in mainstream/regular schools (Figure 4‑15). Also, in 2018, about 3 percent of NIS students came from the first quintile of the income distribution, compared with 23 percent of those in mainstream/regular schools. Although the NIS caters only to a small fraction of students in Kazakhstan’s secondary education, this gap in socio-economic backgrounds between NIS students and those of traditional schools raises an equity issue in the use of public funds. 88 IAC (2020). Republic of Kazakhstan. Statistics of the Education System. Table 2.2.1 and Table 2.2.2. 89 Functional illiteracy is defined as the inability to operate successfully in the modern world economy. For PISA, it is calculated as the proportion of students of 15 years of age who perform below level 2 proficiency in math, science, and reading literacy. 94 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Figure 4‑13. Per-student public spending (2018) Figure 4‑14. Performance of 15-year-old in US$ PPP students in PISA (2018) Source: World Bank staff calculations based on data published by the authorities, PISA (2018, IAC, OECD/UIS/Eurostat (2021). Figure 4‑15. Distribution of students’ socio-economic status in the NIS and mainstream schools across quintiles Source: World Bank staff calculations based on PISA (2018) database data. Publicly-funded scholarships can provide students from low-income households with access to quality education. Kazakhstan finances technical vocational education and training (TVET),90 higher education for those who pass the merit threshold in university entrance examinations, and international higher education opportunities under the Bolashak program. Typically, scholarships in Kazakhstan, primarily funded through the Republican budget, cover the full cost of tuition fees and include a monthly allowance depending on the level of education. The Bolashak International Scholarship was introduced in 1993 for the full-time training of citizens of the Republic of Kazakhstan in leading foreign higher education institutions. In 2020, the Government allocated US$ 33 million91 (0.5 percent of the education budget) to scholarship holders studying abroad, which include Bolashak scholarship holders. This marked a 12 percent increase compared to 2017 when the corresponding amount stood at US$29 million. Since its introduction, over 12,000 students have benefited from a Bolashak scholarship, studying mainly in the humanities (55 percent) and engineering (36 percent). The number of beneficiaries gradually increased until 2022, when 1,055 Bolashak scholarships were granted for training in leading foreign institutions of higher and postgraduate education. 90 World Bank. Equitable Human Capital Development in Kazakhstan: Technical and Vocational Skills Development, Adult Learning and Labor Market Programs (English). Washington, DC: World Bank Group. http://documents.worldbank.org/curated/ en/099083002082333944/P17487900b2b230ae09b920ba9aec6d07c9 91 2020 Budget data from the Ministry of Finance. 4. PUBLIC SPENDING ON EDUCATION 95 The number of state scholarship recipients in Kazakhstan has risen considerably. State educational scholarships are paid to the students of higher education institutions and technical and professional educational institutions. In 2020, the Government allocated US$ 239 million92 (3.4 percent of the education budget) to scholarship holders in Kazakhstan. This is a 97 percent increase from 2017, when the corresponding amount stood at US$ 121 million. The increase reflects the rise in the amounts allocated, which are set to continue growing under the “Amanat” roadmap for 2021-2025, and the rise in the number of scholarship holders. The scholarships supported 36 percent of students enrolled in higher education (207,283 students) in 2021, while only 25 percent of students in higher education were scholarship holders in 2012. In parallel, the scholarships supported 59 percent of students enrolled in TVET (288,413 students) in 2021 compared to 43.5 percent in 2014. Yet, the distribution of education scholarship transfers appears to be regressive. Figure 4‑16 shows that in 2021 households in the last four deciles of the income distribution received only 33.9 percent of scholarship transfers, whereas households in the following five deciles of the income distribution received 57.5 percent of the scholarship transfers. Although some types of scholarships are likely to be more progressive, this shows that scholarships in general in Kazakhstan do not support primarily low-income households. Nevertheless, factors other than income, such as learning capability, may affect student scholarship transfer chances. Figure 4‑16. Scholarship incidence in Kazakhstan (for students under 25 years old) Source: World Bank staff calculations based on data from the 2021 HIES. Note: Transfers from all types of scholarship are captured without scholarship-specific information in HIES data. 4.5 The efficiency of Education Financing There is significant scope for exploring greater efficiencies in Kazakhstan’s education financing. Though students, on average, attend 13.7 years of school, they attain learning equivalent to only 9.1 years of schooling, creating significant inefficiencies within the system. People with general upper-secondary education do not enjoy a statistically significant wage advantage over people with only lower-secondary education, implying a low external efficiency of the system.93 The remainder of this section will explore two major, interconnected sources of increasing efficiency in education service delivery: teachers and school infrastructure. Both expenses are substantial and increasing, requiring a serious review to enhance system efficiency. 92 2020 Budget data from the Ministry of Finance. 93 World Bank. Equitable Human Capital Development in Kazakhstan: Technical and Vocational Skills Development, Adult Learning and Labor Market Programs (English). Washington, DC: World Bank Group. http://documents.worldbank.org/curated/ en/099083002082333944/P17487900b2b230ae09b920ba9aec6d07c9 96 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 4.5.1 The efficiency of expenditure across regions Several inefficiencies are emanating from the current planning and budgeting system in Kazakhstan. First, there is a disconnect between central planning and regional needs. The MoE’s National Education Plan (NEP), which sets the foundation for the MoE’s medium-term financing needs and target outcomes, appears to respond to the region’s underlying dynamics weakly. Over the past 10 years, total student enrolment in the education system has increased considerably in Astana, Shymkent, Mangystau, Almaty city, Turkestan, and Almaty region (from 150 to 340 percent). At the same time, total student enrolment numbers have dropped in North Kazakhstan, Kostanay, Pavlodar, and East Kazakhstan regions (from 70 to 90 percent) due to out-migration from these regions. Despite this migration dynamic and one of the highest increases in enrolment numbers, Turkestan and Shymkent have among the lowest capital expenses per student (half that of the Akmola region). Nominal capital expenditures for South Kazakhstan (now comprising Shymkent and Turkestan) increased by only 6 percent from 2010 to 2020. Meanwhile, nominal capital expenditures increased by nearly five times in Akmola over the same period, where student enrolment has remained constant over the past decade. Figure 4‑17 shows almost no correlation between capital spending per student across regions and their respective net migration rate. Similarly, Figure 4‑18 shows that such spending does not seem to correlate with the regions’ population densities. Figure 4‑17. Capital spending per student and Figure 4‑18. Capital spending per student and net migration rate across regions population density across regions Source: World Bank staff calculations based on data published by the authorities. Second, there is a fragmentation of crucial information for decision-making. National and regional planning for effecting qualitative changes in educational outcomes is hampered by fragmentation across information sources essential for evidence-based decision-making, a fundamental element for accountability. The NEDB, first introduced in 2013, contains information on learning conditions and school personnel data, which schools regularly update. However, it lacks information on learning outcomes or expenditures. Although salary payments constitute the majority of local executive budgets, the central government and the SNGs do not have access to this information, as it is available only in hard copy format at the rayon level. 4.5.2 The efficiency of expenditure on teachers’ salary While the increase in teacher salaries will help raise the competitiveness of the education profession, 4. PUBLIC SPENDING ON EDUCATION 97 there remains considerable scope to enhance the efficiency of the wage bill and teaching effectiveness. International experience suggests that increasing teacher salaries does not guarantee teacher effectiveness (i.e., improved teaching and better learning outcomes), so the salary increase cannot be considered an end target and would need to be complemented with other policies to support teaching effectiveness. Significant inefficiencies in Kazakhstan emanate from low student-to-teacher ratios, especially in rural areas, and low teaching workloads. As the Government plans to build more schools under the NEP, it would be important to consider the reallocation of existing teachers to fulfill the needs. Moreover, while mandated teaching loads are low, the current teacher remuneration system (Stavka) incentivizes teachers to spend more hours teaching for higher remuneration, focusing on bringing up the best performers. This then constrains the time left available to prepare teaching materials, support students’ learning needs and teachers’ professional development needs. This risks their effectiveness as teachers and their capacity to deliver the results of a high-performing education system. Finally, teachers’ professional development and incentives to work with low-performing students will be crucial to render teaching in Kazakhstan effective. Student-to-teacher ratio The student-to-teacher ratio is low on average across all levels of education in Kazakhstan, with wide variations across levels and regions. The national average preschool student-to-teacher ratio of 9:1 is low compared with the average for Europe (where only five countries have student-to-teacher ratios of less than 10). It also varies across the country: from a minimum of 6.8 students per teacher in North Kazakhstan to 11.7 students per teacher in Astana city. Similarly, student-to-teacher ratios for general secondary education and colleges are low. On average, there are 9.5 students per teacher in general secondary education (2020/21) and 13 students per teacher in colleges (2018).94 These overall ratios are low compared with the OECD average of 13 students per teacher for all secondary and vocational programs and 18 students per teacher for post-secondary non-tertiary education. In the 40 percent of small schools across the country, the student-to-teacher ratio is as low as 4.4, with some schools reporting having more teachers than students. Meanwhile, 75 percent of schools are overcrowded. Redeployment of the current stock of teachers will be important to allow more optimal and balanced student-to-teacher ratios. The low average student-to-teacher ratio suggests that Kazakhstan can provide services to the increasing student population without increasing the teacher stock. There is, however, a need to address cases of overcrowded schools, as this will likely lead to lower learning outcomes. Overcrowded schools imply drastically reduced time for individual attention and student-to-teacher interaction, further risking learning gains among children, especially those who need additional support. Improving teaching effectiveness through a better remuneration system While teacher workloads are inefficiently low in Kazakhstan, the remuneration system (Stavka) incentivizes increasing in-class teaching hours, not teaching quality, risking teacher effectiveness. The official teaching load in Kazakhstan is 16 hours a week, which is low by international standards and a source of inefficiency in the education system. For example, full-time teachers in the EU report teaching almost 20 hours per week.95 Teachers are paid based on the hours they spend teaching in classrooms, with one Stavka equivalent to 16 hours. A critical issue with this payment mechanism is that teachers perceive all types of work besides teaching in class as “unpaid work.” This harms the quality of teaching and learning, which includes diverse kinds of work, for example, preparing lesson plans, checking assignments, discussing student performance with parents, holding office hours, supporting students who are lagging, and 94 The highest STR is 16.25 for general secondary education and 18.7 for colleges in Astana. 95 European Commission/EACEA/Eurydice, 2021. Teachers in Europe: Careers, Development, and Well-being. Eurydice report. Luxembourg: Publications Office of the European Union. 98 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH spending time on professional development provides an example of hours allocated for in-service training across countries). In fact, the official time to be dedicated to professional development in Kazakhstan is low. These “non-teaching” activities are generally mandatory and make up the non-teaching time of the statutory workload of teachers in most countries. The second critical issue is that teachers prefer additional lesson time to increase their take-home pay. While there is a limit of 24 hours of classroom work per week, this limit is not implemented in practice. According to teacher tariff tables obtained for this PFR, a Russian language teacher working in a school in Munaily rayon of Mangistau region taught 42 hours a week. Most physical education teachers in large urban schools teach 32–36 hours per week, and quite a few physics teachers in Almaty teach classes for about 30 hours per week.96 Box 4‑2. Time for collegial work and ongoing professional development. Globally, time for collegial work and ongoing professional development is included in teachers’ workloads. For example, in Shanghai, every teacher is required to take 360 hours of professional development over five years compared with the current 72 hours in Kazakhstan. A teacher must complete another 540 hours of training if they aim to apply for a senior grade. Much of this professional development is designed to be school-based, collaborative, and targeted toward instructional improvement. Schools in Shanghai spend 7 percent of their total operating budget on teachers’ professional development, and teacher growth is identified as one of the responsibilities of school principals. Principals are responsible for developing each teacher's professional development profiles (based on teacher needs, for example, scores on National Qualification Tests) and are involved in teacher evaluations. Figure 4‑19. Officially required days of professional training per year, secondary education Source: Liang, Xiaoyan, Huma Kidwai, and Minxuan Zhang. 2016. How Shanghai Does It: Insights and Lessons from the Highest- Ranking Education System in the World. Directions in Development. Washington, DC: World Bank. doi:10.1596/978-1-4648-0790-9. License: Creative Commons Attribution CC BY 3.0 IGO. 96 Source: tariff tables, interviews. 4. PUBLIC SPENDING ON EDUCATION 99 The current teacher incentive system is skewed toward supporting already high-performing students. The only additional payment based on student learning outcomes is for training potential winners of Olympiads of different levels. A one-time reward payment of a threefold salary is made for the teachers of the winners of high-level Olympiads.97 This is an important incentive but can create problems if it is not balanced by comparable incentives that would support other areas of work. For example, suppose a teacher can earn incomparably more in the case of a student’s high achievement. In that case, s/he will be motivated to work with the most gifted and well-performing children as much as possible. This will be financially justified, and given the large payout, there is an incentive to invest significant effort in such work. If this is the only big payment, the educational process overall may suffer. Children of average success will receive less attention, and there will be no incentive to work with the lagging students, engage in extracurricular activities, etc. The incentive system should be balanced, support all important areas of the teacher’s work, and incentivize teacher support toward students falling behind in learning outcomes. Capacity building to improve teaching effectiveness Since most of the increase in Kazakhstan’s education spending has gone toward teacher salaries, ensuring the competency and quality instruction of the teaching force is critical. The minimum qualification requirement of the MoE for teachers of pre-university education (from preschool to post- secondary, non-tertiary education) is post-secondary, non-tertiary (ISCED 4).98 These requirements for teacher qualifications are lower than the EU standards of a minimum of a bachelor’s qualification.99 The MoE has taken steps toward improving teachers’ pedagogical and content knowledge, but the contribution to learning outcomes has not been analyzed. It will be important for the Government to rigorously evaluate the impact of the different initiatives on instruction quality and student learning outcomes. For example, teachers are undertaking training, and a possible vertical trajectory (teacher- master, researcher, expert, moderator) associated with substantial salary increments is available to motivate teachers to improve performance. However, it is unclear if the different policy pieces fit together to ensure that the training and incentive structures lead to improved instruction and student learning outcomes. 4.5.3 The efficiency of expenditure on infrastructure Kazakhstan has successfully provided access to basic and secondary education services across its vast and diverse landscape. Still, weak planning and limited investments have created complex inefficiencies in its school infrastructure. Despite repeated plans to reduce the number of schools with triple shifts to zero, seventy-five percent of schools in the country operate multiple shifts100 , and the number of schools operating triple shifts has almost doubled from 2015 to 2020101 due to the population boom and in-migration to urban centers. Average class sizes in Astana, Almaty city, and Skymkent are much higher than the average for the country (e.g., 26, 22, and 23 students per class, respectively, compared with an average of 16 students per class in lower secondary grades 5 to 9). In contrast to these infrastructural 97 https://strategy2050.kz/ru/news/chto-izmenitsya-dlya-uchiteley-s-vvedeniem-zakona-o-statuse-pedagoga/ 98 Order of the Minister of Education and Science of the Republic of Kazakhstan dated July 13, 2009 No. 338. Registered with the Ministry of Justice of the Republic of Kazakhstan on August 17, 2009 No. 5750. On the approval of the Standard qualification characteristics of the position of teachers. (ECE: Para 4, clause 20; Secondary Para 7, clause 66; TVET: Para 12, clause 286: “higher and (or) postgraduate pedagogical education or vocational education in the relevant profiles without presenting requirements for work experience”). Accessed at: https://adilet.zan.kz/rus/docs/V090005750_ 99 European Commission/EACEA/Eurydice, 2021. Teachers in Europe: Careers, Development and Well-being. Eurydice report. Luxembourg: Publications Office of the European Union. 100 IAC Report 2020-2021 2.2.134. 101 IAC Report 2020-2021. 2.2.132. 100 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH shortages, 40 percent of schools102 in the country are small schools catering to only 5.8 percent of the total student population, with an occupancy rate of seven students per class103 and an average student- to-teacher ratio of 3.9.104 Most of these small schools are primary or general secondary education schools that require higher resources than primary schools. Larger class sizes are associated with reduced learning, impacting external efficiency, i.e., returns on education. Given that the school shift lasts only 4 hours, the actual instruction time is much less than the compulsory instruction time in comparator countries. As a result, a student in Kazakhstan, on average, receives around 20 percent less instruction time than an average student in the OECD, impacting student learning outcomes (Figure 4‑20). ). Figure 4‑20. On average, a student in Kazakhstan receives around 20 percent less instruction time than an average student in the OECD Source: World Bank staff addition for Kazakhstan to OECD (2021), Table D1.1. Optimization of the existing school infrastructure and improved planning for new infrastructure development will help Kazakhstan make more efficient use of limited public resources. Given the input-based financing system with limited accountability to deliver outcomes, regional governments are more interested in maintaining the physical structures of schools and teacher numbers, even with very small student numbers. A policy that incentivizes and empowers local governments to decide on the most efficient use of resources using their greater contextual knowledge and data will ensure that the education infrastructure is efficiently structured and can respond to changing student demographics. In addition, school-level data is available on school locations (geographical coordinates) and the number of classes, students, and teachers in the NEDB and can be leveraged to improve the efficiency of operational expenses and infrastructure investments. The information in the NEDB can be combined with financial information to develop a localized strategy for school network optimization. Box 4‑3 presents an example of a successful school network optimization in Bulgaria. All new infrastructure plans need to consider student population projections and the existence of facilities that could be retrofitted rather than reconstructed. Secondary schools generally require more resources, including more subject specialist teachers and laboratories. Therefore these schools could be separated from primary schools and preschools and be in areas where they could serve larger catchment areas. 102 IAC (2021). Table 2.2.148. 103 IAC (2021) Tables 2.2.151 & 2.2.157. 104 IAC (2021). Tables 2.2.151 & 2.2.158. 4. PUBLIC SPENDING ON EDUCATION 101 Box 4‑3. Developing a successful localized strategy for school network optimization – Lessons from Bulgaria Starting on January 1, 2007, the Bulgarian Government shifted the system for financing primary education from one based on inputs to one based on enrolment. The bulk of financing for education to municipalities shifted to large earmarked grants based on the number of students in the municipality. Specific features of the policy allowed school network optimization in Bulgaria to be successful: Bulgaria conceded considerable autonomy over the determinants of education costs to their local governments. While salary scales in Bulgaria are set nationally, local governments have the authority to consolidate classrooms, dismiss teachers, and, subject to certain limitations, close schools. Municipalities decide which schools should be closed to ensure the effectiveness and efficiency of the education budget. The Ministry, however, had the final say in school closures and monitored compliance with elaborate procedures that municipalities had to undertake before closing a school. For instance, a school closure proposal had to include a discussion of how the transportation needs of affected students would be addressed and demonstrate that nearby schools had sufficient capacity to accommodate the additional students. The Government also introduced a program to monitor dropout rates and established additional national programs to support municipalities. For instance, municipalities with school closures can apply for school buses, additional payments to laid-off teachers, and resources to refurbish “central” or new “merged” schools. The Ministry also put together a list of “protected schools,” i.e., schools that cannot be closed because no nearby schools exist, to ensure access to education for all, while additional resources to finance such small schools were also provided. 4.6 Reform Options As documented in this chapter, Kazakhstan’s educational development needs are significant. These include the need to provide access to preschool and higher education students, keep up with population increases for all levels of education while also addressing the severe infrastructure deficits across schools, and significantly raise the quality and relevance of the system and the skills imparted for a just green transition and economic diversification, etc. The recent increases in education spending were long overdue, but critical inefficiencies and inequities limit the country’s ability to meet its diverse needs adequately and sustainably. The following section proposes recommendations for improving the effectiveness, efficiency, and equity of education financing. These recommendations complement those developed in the 2023 policy notes for building equitable human capital in Kazakhstan.105 They are also linked with the discussion on the effectiveness of the budgeting and planning system in delivering results (Chapter 6) and the need to simplify and clarify fiscal transfers from central to subnational governments (Chapter 7). 105 World Bank 2023. Equitable Human Capital Development in Kazakhstan: Skills Formation During the Foundational Years (https://documents.worldbank.org/en/publication/documents-reports/documentdetail/099083002082316985/ p17487902c284204f086310a12db02babf4); Equitable Human Capital Development in Kazakhstan: The State of Human Capital Across the Regions of Kazakhstan (https://documents.worldbank.org/en/publication/documents-reports/ documentdetail/099083002082337789/p1748790b92e770f50b5490c8bbe615fa2b); Equitable Human Capital Development in Kazakhstan: A Policy Overview (https://documents.worldbank.org/en/publication/documents-reports/ documentdetail/099083002082350371/p17487901152a409b0915702a100c0800ec); Equitable Human Capital Development in Kazakhstan: Technical and Vocational Skills Development, Adult Learning and Labor Market Programs (https://documents.worldbank.org/ en/publication/documents-reports/documentdetail/099083002082333944/p17487900b2b230ae09b920ba9aec6d07c9 102 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Improve the efficiency and effectiveness of key educational inputs such as educational infrastructure and teachers by optimizing the school infrastructure network and improving teacher effectiveness. Specifically, • Develop a long-term localized plan of improving education inputs and infrastructure in collaboration with regional- and district-level governments while ensuring the efficiency of the existing infrastructure and additional infrastructural investments through a localized strategy of school network optimization. School-level data is available on school location (geographical coordinates), the number of classes, and the students and teachers in the NEDB, and such data can be leveraged to improve the efficiency of operational expenses and infrastructure investments. The information in the NEDB can be combined with financial information to develop a localized strategy for school network optimization. This school network optimization might require the closure of certain schools and the conversion of others from public education schools providing services to all grades from 1 to 11, to primary schools feeding students to a nearby, larger, well-equipped secondary school. • Expand PCF to all districts and schools, and actively engage the private sector in mobilizing finances to expand school infrastructure in the country. The PCF formula in Kazakhstan includes a component for the depreciation of expenses for new construction, capital repairs, and the acquisition of fixed assets. This component allows the private sector to actively participate in expanding preschool services in the country and can allow the Government to expand infrastructure for general education. • Rationalize teacher recruitment and deploy existing teachers to better address teacher needs across existing and new schools. As student-to-teacher ratios are low, Kazakhstan can limit teacher recruitment to the replacement rate and, with improved salaries, be extremely selective in the selection, preparation, and recruitment process. This is important as one-quarter of the current teaching force in Kazakhstan failed subject content assessments. After selective recruitment, teachers must receive continuous professional development to improve their competence. • Modify the Stavka (pay-per-lesson) system to a comprehensive weekly workload system. It is important to revise the Stavka system to a comprehensive weekly workload system for teacher salaries, whereby average expected hours are recommended for different responsibilities. There is a need to lay out the prescribed distribution of statutory official workload across teaching and non-teaching duties, such as lesson preparation, correction, collegial work, staff meetings, parent-teacher meetings, professional development, etc. • Assess the efficacy of the teacher reforms. There is no evidence that salary increases lead to improved student learning outcomes. A similar reform in Indonesia led to an almost twofold increase in wages.30, And while teachers’ satisfaction with working conditions increased, the expected improvements in the quality of young teaching professionals and improved educational outcomes of students failed to materialize. Kazakhstan needs to assess critically its recent reforms for teachers to better understand the impact of composite reforms and to adjust as required to enhance teaching effectiveness for improved learning. Target public financing to meet the needs of socio-economically disadvantaged students and those lagging in their human capital accumulation through revisiting the efficacy of the NIS financing model, introducing the equity coefficient in PCF, and incentivizing teachers to focus on improving the educational outcomes of the lowest performers. Kazakhstan could: • Revise the PCF Law to include a needs-based component. PCF can be employed as an equity- enhancing tool by introducing an equity coefficient. Such a coefficient could provide additional financial support in schools catering to larger shares of students with low performance or those from socio- economically disadvantaged backgrounds. Currently, the PCF Law provides an extremely low fixed monetary amount of funds for educational expenses. The expenditure on educational processes 4. PUBLIC SPENDING ON EDUCATION 103 needs to increase, especially to provide additional resources for schools catering to students from disadvantaged backgrounds. Students from socio-economically disadvantaged backgrounds likely have limited educational resources and tools at home and need additional support at school to overcome the achievement gap. • Revise the Law on Teacher Remuneration to allocate time for, and incentivize additional support to, low-performing students. The Law on Teacher Remuneration that incentivizes support to high- performing students and allocates high financial rewards for participation and success in national and international Olympiads needs to be amended to provide instructional support to students lagging in outcomes. • Evaluate the economic benefits of sustaining the NIS as part of the Kazakhstan public education system. NIS schools can theoretically serve as centers of excellence for high-quality teaching and attracting high-quality students. Their presence can also create positive spillovers from their alumnus and inspiration for other schools. Nevertheless, given that NIS schools cost more in public resources than other public schools, the Government may wish to evaluate the impact of having the NIS from a benefit incidence across students’ income background, returns from attending the NIS, and other metrics. Enhance system accountability through improved data collection and utilization and empowerment of SNGs to deliver education results. It is important to: • Improve data (adequacy, transparency, credibility, and integration), and utilize it for decision-making and reporting on outcomes across regions. Data should be complete, transparent, integrated, and used for decision-making. A single entity at the Republican or regional level, such as the Information Analytics Center, could be mandated to collect comprehensive and detailed information on education financing, inputs, and outputs to guide the effectiveness and efficiency of education expenditures. The MoE’s NEDB could be developed further to fulfill this purpose. Financial information needs to be combined with information on educational inputs and outcomes in the NEDB. SNGs should be obligated to share financial reports with the MoE, and the MoE should work closely with them to develop national plans. Issuing district report cards indicating performance along several outcome indicators can go a long way in collating the information required to ascertain how the regions and the country are progressing toward the intended outcomes. Expanding PCF to all schools in the country and channeling PCF through the MoE can simplify and streamline the financing process. Empower SNGs to actively plan and monitor educational outcomes. While administering most education expenses, SNGs face capacity constraints such as severe staffing shortages, not being accountable for education quality, not being involved in monitoring and assessment, and not being actively engaged in educational planning for their regions and the country. As such, according to national guidelines, they act primarily as conduits of funds to schools. The MoE and education departments in SNGs must actively collaborate to plan and monitor education outcomes based on national and international student assessments. Regions require training in using national assessment data and the NEDB to plan how education expenditures can help contribute to improved outcomes. SNGs need to be involved in the monitoring of educational inputs and outcomes and held accountable for education quality in their regions. The MoE also needs to engage more actively with subnational education departments to ensure that national plans effectively address the needs and realities on the ground across different regions. 104 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 4. PUBLIC SPENDING ON EDUCATION 105 5. Public Spending on Social Protection 106 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH KEY POINTS • Kazakhstan has a comprehensive social protection (SP) system that includes programs to address a wide range of vulnerabilities and social risks and perform a critical function in reducing poverty. • The social assistance (SA) system is diverse and generous but is focused mainly on supporting selected categories of the population rather than directly targeting the most impoverished. Improving social assistance design and implementation practices can minimize exclusion and inclusion errors. As for targeted social assistance (TSA), the national poverty line used to define eligibility for and the size of TSA payment is set at a relatively low level, resulting in limited population coverage. An increase in TSA scope based on the revised poverty line could result in improved coverage of the poor. • The social insurance (SI) system protects citizens from different social risks and supports different categories of people (pensioners, the unemployed, families with children, and the disabled). However, informal workers are excluded from the social insurance system. The challenge is to bring them into formal employment so that they can participate in the system and receive benefits. • While Kazakhstan implements a comprehensive set of active labor market programs (ALMPs), their contribution to the likelihood of recipients getting quality jobs has not been systematically analyzed. Existing and future AMLPs need additional quality assessment and regular monitoring of how the activities could benefit the population’s poor and other vulnerable groups in the longer term. 5.1 Motivation The main aim of social protection systems is to help individuals and families, especially the poor and vulnerable, to cope with crises and shocks, find jobs, improve productivity, invest in the health and education of their children, and protect the aging population. Typically, the social protection (SP) system includes various policies and programs that can be grouped into three main categories: (i) social assistance (SA); (ii) social insurance (SI); and (iii) employment support programs, including active labor market programs (ALMPs). The SA programs can include: (i) different unconditional cash transfers; (ii) conditional cash transfers; (iii) social pensions; (iv) food and in-kind support; (v) school feeding programs; (vi) public works; (vii) fee waivers; and (viii) targeted subsidies and other interventions.106 Kazakhstan’s SP system must respond to the growing challenges in the population’s livelihood and well- being. First, the relatively low official poverty rate (5.2 percent in 2021) is unlikely to reflect the country’s real poverty situation. Using an international poverty line of US$6.85 in 2017, PPP for upper middle-income countries suggests that the poverty rate in Kazakhstan was about 13 percent in 2021. Second, while the official unemployment rate in Kazakhstan remains relatively low (4.9 percent in 2021), the high level of 106 Bowen, Thomas, Carlo del Ninno, Colin Andrews, Sarah Coll-Black, Ugo Gentilini, Kelly Johnson, Yasuhiro Kawasoe, Adea Kryeziu, Barry Maher, and Asha Williams. 2020. Adaptive Social Protection: Building Resilience to Shocks. International Development in Focus. Washington, DC: World Bank. doi:10.1596/978-1-4648-1575-1. 5. PUBLIC SPENDING ON SOCIAL PROTECTION 107 labor market informality places many workers at a higher risk of vulnerability. A significant share of the economically active population is employed in the informal sector and does not contribute to the SI system. The official level of informal employment in Kazakhstan is 14 percent in 2021. Still, it could be higher based on the coverage data of employed people provided by the pension system. Only 77.9 percent of employees and 58.7 percent of self-employed contributed to the Unified Accumulative Pension Fund (Q3 2022); contributions could be at a minimum level. About 2.1 million people (or around one-quarter of all workers) are self-employed, with limited access to SI.107 Third, pressure on the labor market will significantly increase during the next five years due to the growing number of young people. According to the Ministry of Labor and Social Protection of the Population (MLSP) estimates, youth aged 15–20 will increase by 36 percent, accounting for 1.8 million in 2025 (compared with 1.3 million in 2018). The number of young people of working age entering the labor market will be 1.5 times higher in 2030 compared with 220,000 young people in 2020. Therefore, developing skills among the youth entering the labor market is one of the government’s major concerns.108 Assessing the SP performance can inform policymakers’ options to improve the efficiency and effectiveness of the SP system. Kazakhstan’s SP covers programs addressing a wide range of vulnerabilities and risks. There are categorical and means-tested social assistance (SA) programs (including a Guaranteed Minimum Income [GMI]-type targeted social assistance [TSA] program), contributory SI programs, and a comprehensive set of ALMPs. The Government has substantially increased budget allocation for SP programs, from an average of 20 percent of total general government spending in 2014–2017 to 23.6 percent in 2018–2021. This increase reflects not only the Government’s responses to the impact of the COVID-19 pandemic but also its political commitment to improving social welfare, as mentioned by President Tokayev in his annual speeches in 2022 and 2023. In this context, this assessment seeks to inform policymakers about the issues that may undermine the efficiency and effectiveness of the SP system in meeting its objective and highlight the potential reform areas for consideration. 5.2 Overview of Kazakhstan’s Social Protection System Kazakhstan’s SP system is financed from three main channels: the republican (central) government budget (and transfers to lower-level budgets), the State Social Insurance Fund, and the Unified Accumulative Pension Fund. Several categorical SA (cash and in-kind) programs and TSA are financed from the own-source revenues of local (district) budgets and targeted transfers from higher-level budgets (see Annex 1 for more details). The housing allowance is paid from local budgets, including local revenue and equalization transfers from the higher-level budget. The SI system is financed through mandatory contributions collected by the State Social Insurance Fund (SSIF). The SSIF assigns and pays social benefits to selected members of the compulsory SI system who have encountered specific social risks, for example, unemployment, disability, the birth of a child, etc. The pension payments for the funded pillar are financed through the mandatory pension contributions collected by the Unified Accumulative Pension Fund (UAPF).109 In parallel, the base pension (pillar zero) is financed through the republican (central) government budget. In recent years, spending on SP in Kazakhstan has been rising as a share of the government budget and GDP. The share of total government budget expenditures allocated to SP has grown steadily in recent years to an average of 23.6 percent in 2018–2021 compared with 19 percent in 2006–2009, confirming SP as 107 Self-employed workers can pay Social Aggregated Payment (includes contributions to the pension and social security funds) and benefit from social and medical insurance systems support. However, to obtain the status of an insured person and receive assistance, it was necessary to have paid contributions for the past 12 months and continue to make monthly contributions, so the worker did not lose the status. 108 Workforce Development Center (2021). 109 Collected from 10 percent of monthly income eligible for mandatory pension contribution. 108 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH one of the highest budget priorities. Relative to GDP, total SP spending has increased continuously from 4.9 percent of GDP in 2018 to 5.5 percent in 2021, driven largely by higher spending on social assistance. In response to the COVID-19 crisis and a sharp decline in GDP, total spending for SP reached about 7 percent of GDP in 2020 (see Box 5-1). ). Figure 5‑1. Spending on SP programs as Figure 5‑2. Share of SP programs spending, percentage of GDP, 2016–2021 2016–2021 Source: World Bank staff calculations based on data published Source: World Bank staff calculations based on data published by by the authorities. the authorities. Box 5‑1. Kazakhstan’s social protection measures during the COVID-19 pandemic The Government implemented comprehensive measures to support citizens against hardship during the COVID-19 pandemic. These measures included: (i) lump-sum payments from the SSIF to people who temporarily lost jobs/income during the state of emergency (April and May 2022) and the lock- down (July to August 2022); (ii) a large-scale public works program (Employment Roadmap 2020); (iii) compensation of utility bills for socially vulnerable population; (iv) provision of a “food-and-kits” package to beneficiaries of TSA; and (v) one-time social payment to health-care workers who contracted the COVID-19 virus at work. Source: MLSP. 5.3 Components of the SP Programs 5.3.1 Social assistance Kazakhstan has two types of SA programs: the means-tested SA programs, which focus on alleviating poverty, and categorical SA programs, which support selected categories of people. It is important to note that in Kazakhstan, the SA system provides selected allowances for childbirth/adoption, child support, and survivors. These programs aim to cover the risks of vulnerable populations not participating in contributory SI. The detailed list of existing SA programs is provided in Box 5-2 below. 5. PUBLIC SPENDING ON SOCIAL PROTECTION 109 Box 5‑2. Social assistance programs in Kazakhstan Poverty-targeted (means-tested) social assistance programs: 1. Targeted social assistance (TSA) is a cash transfer provided by the state to individuals with monthly average per-capita income below an established threshold of 70 percent of the minimum subsistence level. TSA is divided into two sub-types, namely: • Unconditional TSA, provided to: (i) single low-income people with limited opportunities to participate in employment in connection with retirement age; (ii) those with disabilities; (iii) those incapacitated by the presence of a disease, which contributes to temporary work incapacity for more than two months; (iv) low-income families in which there are no able-bodied persons; or (v) the only able- bodied member cares for: a child under the age of 3; a child with a disability; a person with a disability; the elderly; and those in need of outside care and help. • Conditional TSA is provided to single low-income working-age individuals or families and low- income families with an able-bodied member/members, including individuals who are payers of a single aggregate payment.110 The conditional TSA benefit is available to those who conclude a social contract with the state and actively participate in finding new employment, completing retraining, public works, or are involved in entrepreneurial activities. 2. Housing allowance is a cash transfer provided to low-income households from local budgets to partially reimburse utility costs, maintain a dwelling, and rent accommodation in public housing.111 Non-poverty targeted (categorical) social assistance programs: 3. Family and child allowances, including: • State allowance to large families with four or more minor children living together; • State allowance to mothers of large families, awarded pendants "Altyn Alka," "Kumis Alka" or previously received the title "Heroine Mother" and awarded the Order of the "Parent Glory;" • State allowance to disabled children 0–16 years old and 16–18 years old; • State allowance to a caregiver of a disabled child; and • State allowance for families (persons) caring for a disabled person of the first group since childhood. 4. Disability social allowance/benefits, including: • State social benefits for disability; • Compensation for harm caused to life and health imposed by the court on the state in the event of termination of the activity of a legal entity; and • Reimbursement of expenses for home education of disabled children. 5. Pensions, including: • State Basic pension (Pillar Zero); and • Special state benefits for 17 categories of citizens. 6. Supplementary payments for those not covered by the SI system: • State basic social benefits for the loss of a breadwinner; • Allowance for the care of a child under the age of one and a half years; and • Allowance for childbirth. 7. Public works • “Employment Roadmap 2020–2021.” 110 Law of the Republic of Kazakhstan "On State Social Targeted Assistance". 111 https://adilet.zan.kz/rus/docs/V09PH000078 110 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 8. Other categorical, including: • Special allowances for Lists 1 and 2; • Special state benefits (including people with disabilities); • Lump sum payment to citizens who became victims of nuclear tests at the Semipalatinsk nuclear test site; and • Compensation paid to victims of mass political repression. Kazakhstan’s spending on SA as a share of GDP has grown steadily for categorical programs. Spending on the overall SA programs accounts for 2.8 percent of GDP compared with 1.8 percent of GDP spent on average by comparator UMICs. In the composition of SA programs, the base pension remains the largest during these years. During 2016–2021, spending on state basic pensions112 rose from 0.6 percent of GDP in 2016 to 1 percent in 2021. The family and child allowances also steadily increased from 0.5 percent of GDP in 2016 to 0.7 percent in 2021. The expenditures for disability allowances remained relatively stable throughout the period, accounting for 0.4 percent of GDP in 2021. Three groups of programs have remained the largest in the overall composition of SA over the past five years: the base pension, family and child allowances, and disability allowances. Almost 50 percent of total SA funding went to the state basic pension until pre-pandemic 2019, following a decrease to 30 percent in 2020 and a rapid increase to 40 percent in 2021 (Figure 5‑3). Disability allowances steadily declined from 26 to 12 percent of total SA funding but remained a key policy priority following the increase to 17 percent in 2021. The support provided through the family and child allowances as a share of total SA funding declined from 29 percent in 2016 to 16 percent in 2019 but started to pick up the pace again in 2020 by reaching 19 percent (which is equivalent to the 2018 level and explained by the move of the large family allowance from the means-tested to a categorical program). In 2021, the share increased and accounted for 27 percent, explained again by another change in the design of TSA and categorical programs. Figure 5‑3. Key SA programs as % of GDP, Figure 5‑4. Key SA programs as a share of total 2016–2021 SA funding, 2016–2021 Source: World Bank staff calculations based on data published by the authorities. 112 The base pension is a basic state pension for all citizens (pillar zero level) who reach retirement age. Since 2018, it has been granted depending on the length of service and is equal to 54 percent of the subsistence minimum when a person worked 0–10 years. For each year worked over 10 years, the base pension increases by 2 percent and can reach a maximum of 100 percent of the subsistence minimum. 5. PUBLIC SPENDING ON SOCIAL PROTECTION 111 While the number of people who received the base pension and disability allowance remained stable during 2016–2021, the number of family and child allowance beneficiaries varied considerably over the same period. The number of beneficiaries of poverty alleviation programs fluctuated at around 2 million recipients in 2019 (Figure 5‑5), showing the same TSA design change pattern observed in the expenditure data. Social pressure following a tragedy that cost the lives of a low-income family in February 2019 prompted the Government to temporarily provide additional family and child allowances.113 Subsequently, the number of beneficiaries of family and child allowances decreased in 2020 by almost 1 million recipients and remained stable throughout 2021. The number of public work beneficiaries increased to 239,000 recipients in 2020, reflecting the socio-economic situation during the COVID-19 pandemic. Figure 5‑5. Number of beneficiaries of the SA program, 2016–2021, ’000 people Source: World Bank staff calculations based on data published by the authorities; ADePT. 5.3.2 Social insurance Currently, Kazakhstan operates two SI programs: pensions and programs to overcome certain social risks. The SI pensions represent solidarity and funded pensions. The solidarity pension is only for those who enrolled in a defined benefit pension before the 1998 pension reform, while the funded pension is based on contributions and managed by UAPF. The first three SI programs related to social risks were introduced by the Government in 2005 to address: (i) disability; (ii) loss of a breadwinner; and (iii) unemployment. Since 2008, the SI system has supported two additional types of benefits: (iv) loss of income due to pregnancy and childbirth or adoption of a child; and (v) childcare for infants until age 1. SI expenditures have been steadily increasing in nominal terms and spending as a share of GDP over five years. The size of unemployment benefits and survivors and disability benefits have been slowly increasing, but their shares remain low, accounting for 0.02, 0.02, and 0.03 percent of GDP in 2021, respectively. The benefit covering pregnancy or the adoption of a child increased slightly from 0.12 percent in 2017 to 0.15 percent of GDP in 2021 after experiencing a drop to 0.11 percent in 2018–2019. The expenditures for the monthly benefit in support of an infant aged up to 1 year increased in nominal terms from 15 percent in 2016 to 18 percent in 2021, with a decline to 13 percent in 2019. 113 In February 2019, fire destroyed a house in Astana and killed five children whose parents were working on night shifts, triggering a nationwide protest. The benefits provided to families in 2019 were different from those provided in 2018 and 2020. They were also calculated in a different way: in 2019 each child was provided KZT 20,000 for several months. While before and after that, the size of the payment depended on the number of children in the family. 112 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Figure 5‑6. Social insurance 2016–2021, % of GDP Figure 5‑7. Social insurance 2016–2021, % of total SP program 2016–2021 Source: World Bank staff calculations based on data published by the authorities. For citizens not covered by SI, the Government provides additional support to reduce the risk of them falling into poverty. The Government provides additional support to those that do not participate in the SI system through: (i) support for the loss of a breadwinner; (ii) one-time payment for childbirth; and (iii) benefit for childcare for infants under one and a half years of age. These additional supports mimic benefits provided to those enrolled in the SI system and are funded through the Republican (central government) budget (Table 8-3). However, the additional support for those not participating in the SI system is generally less valuable than the same SI benefits, except for the allowance for the loss of a breadwinner (survivors’ benefit). The eligibility criteria for receiving additional support from the budget appear not to be well-targeted toward the poor and vulnerable. The eligibility criteria for additional support are only based on whether the individual regularly pays salaried contributions to the SI. Meanwhile, the payment for survivorship, motherhood, and childhood provided through SA-type categorical programs is not means-tested and could burden the Republican budget. Table 5‑1. SP benefits for similar types of social risk under SI, and additional SA supports for citizens not covered by SI Additional support (SA from the SI benefits from SSIF Republican budget Expenditures, 2021, Average Expenditures, 2021, Average share of SP (%) benefit, KZT the share of SP (%) benefit, KZT State social allowance for the 1.8 29, 499 114 0.4 25,274 115 loss of a breadwinner One-time social allowance for 1.3 110,846116 2.7 498,732117 childbirth Social allowance for caring 0.3 16,802118 3.4 32,054119 for a child under the age of 1 114 Monthly, for a household with one dependent. 115 Monthly. 116 Lump-sum payment. 117 Lump-sum payment. 118 Monthly. 119 Monthly. 5. PUBLIC SPENDING ON SOCIAL PROTECTION 113 Source: World Bank staff calculations based on data published by the authorities. 5.3.3 Active labor market programs Kazakhstan has implemented various state programs to support employment over the past 20 years. In 2000–2002, the Government adopted a program to fight poverty and unemployment, which aimed at creating new jobs and provisions of employment. Then the poverty measures were separated from employment support programs, and in 2005, the Government adopted the Program of Employment for 2005–2007 and the Action Plan for Improvement of the Employment System of the Republic of Kazakhstan for 2008–2010. There was also a special Employment Roadmap to support the population during the banking crisis in 2008. In 2011, the Employment Program 2020 was developed to increase the population’s income by promoting sustainable and productive employment. In 2013, a comprehensive Employment Roadmap was established to support the self-employed and the unemployed in productive employment, improve the TSA system, and develop human capital (WDC 2021). In 2017, the Government implemented the State Program for Productive Employment and Mass Entrepreneurship Development (Enbek) as a major employment support program. This program focused on promoting productive employment and entrepreneurship. It consisted of sub-programs, such as skills training, various entrepreneurial microloans and grants, and labor mobility and employment support (see Box 7-3). Between 2017 and 2021, government spending for the Enbek program increased by an average of 19 percent annually in nominal terms. As a proportion to GDP, spending for Enbek increased substantially in 2019–2020 before moderating to 0.19 percent in 2021. Box 5‑3. State Program for Productive Employment and Mass Entrepreneurship Development (Enbek) Enbek was organized around three key components, or “directions,” and each had its objectives and indicators: i. Skills training. This involves the provision of market-relevant TVET and short-term vocational training to the eligible participants of the state program: young people under the age of 29, the unemployed, and certain categories of the employed. ii. Entrepreneurship. This involves: training on basic entrepreneurship skills, micro-crediting, and providing government grants for implementing new business ideas. This component also includes the Zhas Kasipker sub-program, which helps develop entrepreneurial skills among young people. It also provides financial support to young entrepreneurs. iii. Supporting employment promotion and labor force mobility. This involves: (i) the provision of active measures to promote employment (recruitment and retainment incentives through wage subsidies, youth internships, and jobs, social jobs, and public works); (ii) career guidance and job-placement services; and (iii) labor market information to the unemployed and some groups of the employed, creating a unified digital employment platform. Source: World Bank, 2020. 114 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Table 5‑2. Size of spending for employment support programs compared, as a % of GDP and KZT million 2017 2018 2019 2020 2021 Total Enbek (KZT 79,132 106,831 169,365 180,388 159,344 million) Total Enbek GDP share 0.15 0.17 0.24 0.26 0.19 (%) Training (KZT, million) 14,091 22,479 23,966 22,478 25,220 Training GDP share (%) 0.03 0.04 0.03 0.03 0.03 Entrepreneurship 45,156 62,607 90,490 66,755 48,543 support (KZT, million) Entrepreneurship 0.08 0.10 0.13 0.09 0.06 support GDP share (%) Wage subsidies (KZT, 7,397 9,173 12,459 12,812 13,441 million) Wage subsidies GDP 0.01 0.01 0.02 0.02 0.02 share (%) Public works (KZT, 1,827 2,947 6,454 38,875 48,922 million) Public works GDP 0.003 0.005 0.01 0.06 0.06 share (%) Mobility (KZT, million) 552 1,423 4,973 5,968 4,202 Mobility GDP share (%) 0.001 0.002 0.007 0.008 0.005 Other sub-programs 9,076 5,873 3,014 2,069 1,633 (KZT, million) Other sub-programs 0.017 0.009 0.004 0.003 0.002 GDP share (%) Source: MLSP, 2022. Enbek provided a comprehensive set of active labor market programs (ALMP), but its efficiency remains unclear. In 2021, the ALMPs covered 7 percent of the working-age population, almost at the level of Italy and Denmark. Among all the Enbek beneficiaries, 9 percent received skills training (VET, short courses, and online). Almost 52 percent received direct job placements, 24 percent were provided with internship/ job placements through wage subsidies (see Box 5-4), 2 percent were provided with career development services, and 1 percent received mobility support (Figure 5‑8). It is difficult to say whether the programs facilitated stable employment due to the absence of regular impact assessment. The programs must be evaluated regarding achieving their objectives and balancing costs and benefits. Existing and future AMLPs need additional quality assessment and regular monitoring of how the activities could benefit the poor and vulnerable populations. 5. PUBLIC SPENDING ON SOCIAL PROTECTION 115 Figure 5‑8. Distribution of Enbek participants by program, 2017–2022 Source: Workforce Development Center, 2022. 116 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Box 5‑4. Typology of wage subsidy programs in Kazakhstan Social Jobs This is a temporary workplace created by the employer and the employment center for the unemployed with subsidized salaries from the state. Priority is given to citizens from target groups: persons with disabilities, orphans, persons released from prison, kandas (ethnic Kazakhs, who reside outside the Republic of Kazakhstan and are eligible for the repatriation program),120 and single parents with many children. Duration: no more than 12 months Salary: it is set by the employer (35 percent of the salary is subsidized by the state (but no more than 20 MCI), 65 percent by the employer). Youth Apprenticeship This is a temporary job for obtaining initial experience for college and university graduates no older than 29 years121 who graduated no more than three years ago and have no work experience in their specialty. Duration: no more than 12 months (since 2022).122 Salary: 30 MCI (since 2022),123 but can be set higher by the decision of the local executive authority. First Workplace This job is for obtaining professional knowledge and skills at the first workplace. For young people, including those of the NEET category, without professional education and work experience. Upon completing the first job, an unemployed person is provided a permanent job for at least 18 months. Duration: no more than 18 months (since 2022)124 Salary: 30 MCI (since 2022),125 but can be set higher by the decision of the local executive body. Contract “Generation” This workplace was created to transfer experience and skills from existing employees of pre-retirement age to graduates of educational institutions who graduated no more than two years ago. Upon completion of the Generational Contract, an unemployed person is provided with a permanent job or replaces an employee who has reached retirement age. The employer sets work requirements. Duration: no more than six months Salary: 30 MCI, but can be set higher by decision of the local executive authority Silver Age The employer created this workplace and the employment center to employ persons of pre-retirement age (over 50 years). Upon completion of the Silver Age, an unemployed person is provided with a permanent job until retirement. The employer sets work requirements. Duration: no more than 36 months Salary: set by the employer126 Source: WDC, 2022. 120 https://egov.kz/cms/en/articles/kandas_rk 121 In 2023 the youth threshold was extended to 35 years. 122 Before it was 6 months. 123 It was 25 MCI in 2021. 124 It was 12 months in 2021. 125 It was 20 MCI in 2021. 126 Wages are set and paid by the employer, while the state covers the first 12 months - 70% of wages; from 13 to 24 months - 65% of wages; from 25 to 36 months - 60% of wages, but not more than 30 monthly calculation indices (MCI). 5. PUBLIC SPENDING ON SOCIAL PROTECTION 117 5.4 Adequacy, Efficiency, and Effectiveness of the SP System The SP system in Kazakhstan covers 57.6 percent of the total population, varying across different income groups. About 87 percent of the poorest decile are covered by SP, compared with 54 percent of the richest decile. Also, 80 percent of the poorest decile are covered by at least one SA program (with the prevalence of child and family benefits at 45.1 percent and disability at 32.6 percent). The coverage of the poorest decile by the old-age pension is 30.1 percent, while for the richest decile, it is 48.7 percent. Figure 5‑9. Distribution of SP program coverage across a broad category of beneficiaries (%) by deciles127 Source: World Bank staff calculations based on data from the ADePT database. Overall, the benefit adequacy of Kazakhstan’s SP programs is higher than the regional average. The adequacy of the benefit indicates what share a mean transfer amount represents in the total income of the beneficiaries (in total or for the income distribution group). For the beneficiaries of SP programs, SP transfers represented 47.9 percent of total household income in 2021, above the average in ECA countries (46 percent). The adequacy of SA (20.6 percent in 2021) is well above the ECA average (15 percent). The benefit adequacy of the SP system in the bottom quintile of population income varies across programs. For individuals in Q1 of the income distribution, the adequacy of Kazakhstan’s SP transfers represents 48.7 percent in 2021 (Figure 5-10), slightly lower than the ECA average (59.2 percent). Specifically for SA, the adequacy of the transfers for a Q1 beneficiary is the highest (29.3 percent in 2021) compared with other income groups. Also, within Q1, old age-pensioner have the highest adequacy among individuals in the lowest quintile of distribution (52.4 percent). Overall, old-age pension transfers represent a significant share of the total income of each quintile (around 50–60 percent). However, due to the limitation of the HBS questionnaire, it is impossible to distinguish which type of pension (base, solidarity, or funded) is assigned to the beneficiaries of each income group. Compared with other SA programs in Kazakhstan, TSA performs better in reaching the poor. Beneficiaries of Kazakhstan’s SA system are mostly the poor. Nevertheless, the distribution of beneficiaries displays that TSA has the highest share among all direct and indirect beneficiaries (82.5 percent) in the lowest quintile compared with overall SA programs (Figure 5‑11). 127 Calculated by ADePT using the net of all SA transfers scenario: removes all SA benefit amounts from the aggregate welfare before ranking. 118 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Figure 5‑10. Adequacy of SP benefits by program in 2021 across income quintiles128 Source: World Bank staff calculations based on data from ADePT based on HBS 2021 data. The analysis of the distribution of benefits also suggests that benefits provided under TSA are reaching the poorest more effectively than in other SA programs. The benefit incidence129 for SA for Q1 (49 percent in 2021) shows that the SA targeting is in line with the average for the ECA region (49.1 percent for Q1). While SA is progressive (49 percent of all transfers go to Q1, compared with 10.3 percent accruing to Q5), at the level of individual programs, TSA appears to be more progressive than others, where Q1 receives 85.3 percent of benefits, and Q5 receives 2.2 percent (Figure 5-12). Figure 5‑11. Distribution by beneficiaries, 2021, Figure 5‑12. Distribution by benefits, 2021, percent percent Source: World Bank staff calculations using ADePT based on HIES 2021 data. Certain beneficiaries of SP programs appear to belong to the non-poor group. The leakage represents the percentage of individuals who receive SP transfers and are not poor (do not belong to the lowest 10 percent of the population). This incidence is especially higher for old-age pension (75.3 percent of the population); housing allowance (50.5 percent); disability benefits (44.3 percent); and child and family benefits (35 percent). Some people are not poor but receive TSA, designed specifically to support the poor population on a means-tested basis. Additional assessment is needed to understand the reasons for such leakage. 128 For adequacy households are analyzed by their per-capita welfare, including SA transfers. 129 Benefit incidence is the transfer amount received by the group as a percentage of total transfers received by the population. 5. PUBLIC SPENDING ON SOCIAL PROTECTION 119 Figure 5‑13. Estimated leakage of SA programs (percent) Source: World Bank staff calculations based on data from the ADePT database. At the same time, there are overlaps between selected SP programs, in which citizens can benefit from multiple programs simultaneously. For example, the recipients of TSA can also receive the old-age pension (15.2 percent), child and family benefits (41.9 percent), and housing allowance (1.9 percent) (Table 5-4). The old-age pension, child/family benefits, and disability significantly overlap with other programs. This could be explained by the functional purpose of each benefit and the fact that a person can receive a combination of benefits due to his/her particular life situation. The means-tested approach could be introduced to the selected child and family benefits to avoid possible inclusion errors in SA programs. At the same time, the eligibility assessment for housing allowance could be improved. More analysis of SP program performance is needed to understand the fragmentation, overlap, and duplication of SP programs and determine how to streamline them. Table 5‑3. SA programs: benefits overlap (%), 2021 Other benefits received Old Age Targeted Social Child and family Housing and Disability Recipient of Pension Assistance benefits utilities benefit Old Age Pension130 0.1 11.1 0.1 8.1 Targeted Social 15.2 41.9 1.9 Assistance 8.6 Child and family 33.6 1.1 0.1 benefits 11.8 Housing and 50.0 14.3 20.9 utilities 17.6 Disability benefit 29.8 0.3 14.4 0.1 Children are less likely to be covered by SP programs. Despite the positive demographic trends and increasing social benefits for large families with children, SP programs cover children aged 0–17 less than the average in the ECA region, accounting for 44.9 and 55 percent, respectively. UNICEF identifies the following barriers for families with children to receive benefits: limited awareness, lack of information, the onerous application process, poor user experience, distance-related factors, and social stigma. In addition, children from small families (one or two children) can remain excluded from SP programs as they do not qualify for TSA.131 130 The ADePT software does the analysis based on HBS data. The HBS questionnaire captures whether the individual/household received a pension, without distinguishing it by pillar (base, solidarity or funded). 131 UNICEF. 2017. Barriers to access social assistance and special social services in Kazakhstan. URL: https://www.unicef.org/ kazakhstan/en/reports/barriers-access-social-assistance-and-special-social-services-kazakhstan 120 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH The coverage of overall SA is higher in rural areas (31.4 percent) than in urban areas (25.8 percent). This pattern repeats in the coverage of family and child benefits and disability allowance. It is also higher in the rural areas (15.5 and 12.2 percent, respectively). The coverage of TSA is equal in urban and rural areas (0.3 percent). The analysis of poverty gap reduction Figure 5‑14. The benefit-cost ratio of SP programs, 2021 concerning the cost of each social program shows that TSA achieves a higher ratio between the poverty gap reduction and the total amount of costs spent on the program among all SA programs (0.76 versus 0.4). At the individual SA program level, TSA shows the better value for money than child and family benefits (0.53) and disability benefits (0.34). It is also the only income- tested program with such results, whereas the second means-tested program, the Source: World Bank staff calculations based on data from the ADePT housing allowance, accounts for 0.49. database, (2021). Increasing the scope of poverty-targeted programs can help improve the overall targeting of SA and the benefit-cost ratio for SA programs, which is currently in line with the average compared with other ECA countries (Figure 5-15). Figure 5‑15 Cross-country benefit-cost ratio comparison for SA programs Source: World Bank staff calculations based on data from the SPEED database(2021). Aligning the national poverty line with that of countries with a similar income level is key to increasing the coverage of TSA programs for people experiencing poverty. Kazakhstan’s national poverty line (equal to the minimum subsistence level) is below typical for countries with similar income levels. This has an impact on the eligibility threshold that is used to assign the TSA benefit (see Box 7-5). 5. PUBLIC SPENDING ON SOCIAL PROTECTION 121 Box 5‑5. Kazakhstan’s poverty line Kazakhstan is at the upper end of UMICs with a 2020 per capita GDP of US$25,363 (constant 2017 PPP) compared with the average of US$17,103 for UMICs. But Kazakhstan’s poverty line is well below the average of UMICs. The current poverty line (KZT 37,389)132 is around 10 percent of GDP per capita, lower than the average for UMICs (15 percent of GDP per capita) and the global average (22 percent of GDP per capita for countries at all income levels) as shown by the green line below. The national poverty line defines eligibility for assistance and the size of means-tested SA payments. Therefore such a low level impacts the coverage of poverty-targeted programs. Figure 5‑16. National poverty line comparison Source: Seitz W. (2022). Benchmarking Poverty Lines: Considerations for Kazakhstan. 5.4.2 Social insurance Despite the low entry threshold, the SI system does not provide sufficient incentives for voluntary enrolment. Concerned by the economic impact of the COVID-19 pandemic and the relatively low coverage of SI among citizens, in 2020, the Government lowered the threshold for informal workers who did not enroll in SI to receive social insurance payments. Individuals who had paid the Single Aggregate Payment (SAP) for the self-employed even once became eligible for social support payments of KZT 42,500 during the COVID- 19-related state of emergency.133 The authorities envisaged that this new enrolment would sustain higher citizens’ participation in the SI system. However, most people left the system immediately after receiving the emergency benefit (Figure 5‑17). The number of SAP contributors reached 3 million in April 2020, when 132 Закон Республики Казахстан «О внесении изменений и дополнений в Закон Республики Казахстан «О республиканском бюджете на 2022-2024 годы» от 12 мая 2022 года № 121-VII. URL: https://online.zakon.kz/document/?doc_id=1026672&pos=2;-71#pos=2;-71 133 Rules on social payments Order of the Minister of Labor and Social Protection of the Population of the Republic of Kazakhstan dated March 26, 2020, No. 110 “On approval of the Rules for the implementation of social payments to participants in the compulsory social insurance system and individuals receiving income under civil law contracts, the subject of which is the performance of work (provision of services) for which tax agents have paid mandatory pension contributions for the period of the state of emergency.” URL: https://online. zakon.kz/Document/?doc_id=37609220 122 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH the Government imposed the first strict quarantine Figure 5‑17 Number of SAP contributors, 2020 and mobility restrictions,134 before dropping to 99,000 once mobility restrictions were lifted. SAP contributors increased again in July 2020 during the second quarantine restrictions and before stabilizing. While the monthly SAP contributors during the second half of 2020 were slightly higher than in the first quarter, lowering the threshold for informal workers to receive SI benefits did not sustain higher SAP contributors. This is because the benefits of the SI system and the required contribution are unattractive for informal workers. But the complaints of not receiving SI benefits Source: World Bank staff calculations based on data published by despite enrolment suggest other problems in the authorities. the system. During the COVID-19 crisis, many self- employed did not receive support from the system despite enrolling. According to a UNDP assessment, 72 percent of the surveyed self-employed applied for emergency benefits during the first wave of the pandemic, but less than half of the self-employed (45 percent) received them. In almost 25 percent of cases, the self-employed obtained no social benefits. In 20 percent of all cases, respondents cited the complexity of the application procedure. A possible reason was that the requirements for application changed several times during the pandemic and that, to qualify for state support, the self-employed had to have worked while contributing to the SSIF for the previous six months (UN 2020).135 Monetary support provided by the Government to the unemployed through unemployment benefits (UB) is low. The average amount of UB relative to the average nominal salary (the replacement income) has been around 13 to 14 percent, except during the COVID-19 pandemic in 2020 (Table 5‑4.). The provided replacement income appears to be lower than international benchmarks. For example, a household with two children, where one partner has no job, and the other is unemployed and receives a job support benefit (based on average salary), can achieve a 60 percent replacement rate, on average, in OECD countries (varying from 15 percent in Hungary up to 87 percent in Denmark). Table 5‑4. Share of average unemployment benefit compared with the average nominal wage 2016 2017 2018 2019 2020 2021 Average nominal wage (KZT) 142,898 150,827 162,673 186,815 213,003 250,311 Average unemployment benefit (KZT) 20,354 21,698 24,816 23,223 43,576 32,174 Number of recipients 32,895 35,748 47,648 63,869 99,123 104,160 Share unemployment benefit/wage (%) 14.2 14.4 15.3 12.4 20.5 12.8 Source: World Bank staff calculations based on data published by the authorities. 134 March to May 2020. 135 UN (2020) UN Kazakhstan Covid-19 Socio-Economic Response & Recovery Plan. URL: https://unsdg.un.org/sites/default/ files/2020-08/KAZ_Socioeconomic-Response-Plan_2020_0.pdf 5. PUBLIC SPENDING ON SOCIAL PROTECTION 123 Figure 5‑18. Net replacement rate in unemployment (couple with two children, partner without work, six months of unemployment) Source: OECD, 2022. The Government has acknowledged the problem with informal workers and has tried to expand SI coverage by introducing a special enrolment option to the SI system. The Single Aggregate Payment (SAP), introduced in 2019, provides a simplified procedure to register informal workers with the tax authorities. Box 5‑6. Procedures to register informal workers with the tax authorities Single Aggregate Payment (SAP) payers are individuals engaged in entrepreneurial activities without registration as an individual entrepreneur who simultaneously meets the following conditions: • paid a Single Aggregate Payment; • do not have employees; and • provide services exclusively to individuals who are not tax agents and (or) sell agricultural products from their personal subsidiary farms (except for excisable products) exclusively to individuals who are not tax agents. The monthly amount of the SAP is calculated as follows: • 1 minimal indices (MI) – for individuals living in cities of Republican level, the capital and regional importance;136 and • 0.5 MI – for individuals living in other settlements.137 The threshold is rather low, representing only around 5.3 percent of the minimum wage.138 This amount includes four payments: (i) individual tax contributions at 10 percent; (ii) SI contributions at 20 percent; (iii) mandatory pensions contributions at 30 percent; and (iv) medical insurance contributions at 40 percent. The payment of SAP by self-employed individuals allows them to receive the following benefits: • be a member of the compulsory social health insurance system; • make pension savings and, depending on the length of participation in the pension system, receive the right to a basic pension; and • receive benefits in the case of disability, loss of employment, loss of a breadwinner, pregnancy, and childbirth, or the adoption of a child or caring for a child under the age of 1. 136 As of April 1, 2022, the minimal indices are equal to KZT 3,180 (US$6.65). 137 https://egov.kz/cms/ru/articles/tax_report/edinyiplatezh 138 As of April 1, 2022, the minimum wage is equal to KZT 60,000 (US$125.56). 124 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 5.4.3 Active labor market programs Despite the national demand for skills development, the employment support program prioritizes entrepreneurship programs. Funding for skills training represents 13 percent of the Enbek program’s spending, entrepreneurship programs 49 percent, and employment measures 33 percent (this sub- component increased in 2020–2021 due to the pandemic). The entrepreneurship component provides support through micro-loans, grants for entrepreneurial activities, and specific training on entrepreneurship (under the Bastau Business program). The higher share of the budget allocated for entrepreneurship in Kazakhstan’s Enbek program is the opposite of OECD countries, which spend more on skills development and employment subsidies. Compared with other countries’ ALMPs, training/skills development in Kazakhstan receives a lower budget allocation even though this is a key policy task for Kazakhstan (the productivity of the young population is low, and there is a need to develop a new set of skills) (WDC 2021).139 This can be explained by the absence of stimuli for enterprises to provide continuous vocational training. At the same time, the private sector shows little interest in attracting young people and employing graduates. The entrepreneurship programs differ from typical ALMPs. Entrepreneurship programs might target people who do not have a predisposition/skillset for entrepreneurial activities or do not have opportunities for subsequent entrepreneurship once the program support ends. For example, the current priority group is socially vulnerable people, such as people with numerous children, people with disabilities, poor people, kandas,140 and recipients of survivors’ benefits. Microloans mostly support the rural population (89 percent), specifically women and the youth. Half of the state grants are divided between women (25.6 percent) and young people (24.3 percent). The majority of state grant recipients live in rural areas (66 percent) and represent the unemployed population (69 percent) (WDC 2020). While providing grants or microloans to entrepreneurs may help alleviate some of their financing constraints, it is critical for the Government to systematically evaluate them against the objectives before deciding to extend or expand. Also, the support provided covers a specific area of the economy where it could be difficult to create high-productivity jobs. The provision of microloans mostly supports agricultural activities (79 percent), often associated with low-productivity jobs (WDC 2020). Finally, the size of the grants provided was small (100 or 200 MCI, equivalent to US$667 and US$1,334, respectively) to develop and maintain the entrepreneurial activity. For future programs, evaluating the cost efficiency of loans and grants supporting entrepreneurial activities will be important. The weak profiling of participants contributes to the possibility that the same people will reapply again for a similar type of support. For example, 17.7 percent of the total grant recipients reapplied for the grants and received other ALMP measures. This can be explained by the insufficient profiling of the participants and a lack of social case management to analyze the individual situation of beneficiaries, their needs, and the barriers to sustainable, productive employment. The effectiveness against officially identified program objectives is difficult to assess because the existing AMLPs lack regular and systematic impact evaluations. It is difficult to judge whether the current distribution of Enbek expenditures achieved the best value for money because there is little evidence on how these programs work in practice. The target indicators used in the Enbek program did not provide an opportunity to evaluate effectiveness, especially regarding the sustainability of participants’ employment in the future and entrepreneurship projects launched under the program. ALMPs also have other 139 https://www.akorda.kz/ru/addresses/addresses_of_president/poslanie-glavy-gosudarstva-kasym-zhomarta-tokaeva-narodu- kazahstana 140 Kandas are ethnic Kazakhs and (or) members of their families, who have not previously held citizenship of the Republic of Kazakhstan, and who returned to Kazakhstan and received the appropriate status. More details - https://egov.kz/cms/en/articles/for_foreigners/ kandas_rights_conditions 5. PUBLIC SPENDING ON SOCIAL PROTECTION 125 implementation issues that require attention (see Box 5‑7). In addition, AMLPs need further assessment on how the activities benefited the poor population in the longer term. Such assessment should inform on the budgeting and planning of ALMPs and the choice for monitoring and evaluation indicators (see Chapter 6 for budgeting and planning). Box 5‑7. Discussion among experts in the Workforce Development Center also highlighted other gaps in the existing ALMPs:141 • Lack of a case management system in working with program beneficiaries. Specifically, experts point out the insufficient profiling of program participants, which should identify the needs of potential beneficiaries and better target them toward specific sub-programs. • The incentives for employers to attract participants are low. There are clear criteria for selecting employers to participate in the program. However, large regional employers do not want to participate in ALMPs to avoid labor inspections, likely also connected with public funds usage. • Capacity differs at various public employment centers/employment departments, and there is considerable potential for improvement. There is evidence of weak coordination of certain policy measures at the local level. • Design of sub-programs can be improved. Contractual arrangements between job seekers and employers for the early termination of work (resignation) and graduation mechanisms could be improved. 5.5 Areas for Improvement in the Social Protection System The system would benefit from a regular and systematic performance analysis of SP programs, especially SA and ALMP programs. The impact assessment of specific SA programs would help to assess the effectiveness of current funding allocations among different activities and their long-term effects on recipients. This is especially relevant in the case of poverty-targeted programs. The Government should evaluate the current monitoring and evaluation (M&E) capacity of the governmental agencies responsible for SP policy implementation, identify gaps, and plan ways to design and strengthen their M&E systems (also discussed in Chapter 6). The focus on means-tested poverty-alleviation programs should be increased in the composition of SA programs. Specifically, TSA and housing allowance should be given priority. The increase of TSA should be followed by serious SA performance analysis, especially focusing on program overlaps and consolidating selected programs. The Government can consider applying the means-tested approach to child benefit programs and consolidate it with the TSA program. Such an effort was previously made in 2018 when the benefits for families with many children were integrated into the TSA program. However, the design was reversed with the ad-hoc introduction of a “KZT 21,000 child payment for TSA recipients.”142 It may be a good time to return to the option of converting child benefit programs into means-tested poverty-alleviation programs. 141 Based on comments from the Workforce Development Center. 142 https://informburo.kz/novosti/nazarbaev-minimalnaya-vyplata-mnogodetnym-semyam-na-odnogo-rebyonka-sostavit-21-000- tenge-86016.html 126 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH The eligibility criteria for poverty-targeted SA programs could be improved to minimize exclusion and inclusion errors. In the case of TSA, the program could benefit from expanding the eligibility threshold to include other poor households not covered by the current program. In contrast, the housing allowance program should benefit those with the lowest income distribution by reviewing the eligibility criteria and assessment practices. More proactive usage of poverty-targeted SA programs during crises will provide more targeted support for poor and vulnerable populations during the crisis. The Government should facilitate efforts to involve different worker types in the formal sector. There is a need to continue developing various measures to lower labor market informality and stimulate different types of workers to enter formal employment (self-employed, informal workers). This would allow workers to participate in the SI system, paying a contribution and receiving medical insurance, survivorship, employment, and maternity benefits. Nevertheless, further analytical work is needed to understand the barriers to bringing informal workers into formal employment status. In addition, the Government should consider reviewing the performance analysis of the SSIF as the main operator of SI payments, which could enable it to improve the performance of the SI system. The existing set of ALMPs should be reviewed based on a comprehensive analysis, and its new scope should reflect productivity needs. ALMPs should address major challenges in the labor market (productivity, mobility, etc.). At the same time, ALMPs should match national policy priorities in economics, labor policy, and social protection policy. The scope of ALMPs should be redefined, streamlined, and given a new sense of priority to face the challenges of the upcoming labor productivity needs. The mechanisms for monitoring and evaluating the effectiveness of employment-active promotion measures should be developed based on international best practices. At the local level, responsible practitioners should use evidence-based methods to assess and prioritize potential beneficiaries, build partnerships with different types of employees, and improve awareness regarding the programs among potential recipients. The Government can capitalize on the lessons learned during the COVID-19 pandemic to improve shock responsiveness and the adaptability of the SP system. The existing support channels, such as TSA and emergency payments from the SI system, did not fully cover all the potential recipients. As a result, the implementation of the measures should be evaluated to identify the advantages and disadvantages of these methods. Successful ad-hoc solutions can be part of a systematic response to future crises. In addition, future shock responses should use the advantages of digital services, the Digital Family Card, and envisage contingency funds in the state budget that could expand support to vulnerable groups during a crisis. 5. PUBLIC SPENDING ON SOCIAL PROTECTION 127 6. Budgeting, Planning, Monitoring: Institutional Foundations for the Public Sector 128 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH KEY POINTS • Improving the public sector’s capacity is one of Kazakhstan’s development pillars to achieve high-income status by 2050. The country has implemented steps to improve the accountability and transparency of the public sector in delivering state programs. The commitment was further stressed in the state-leadership announcement to establish a “Listening State” where the public sector is accountable and able to deliver results. • Despite the progress, Kazakhstan needs a stronger and more integrated approach to public sector management that can translate strategic planning into performance through better budgeting, monitoring and evaluation, and increased transparency. The possible entry points are to: (i) strengthen the performance orientation of state budget by improving public spending allocative and operational efficiency and effectiveness; (ii) facilitate the integration of strategies and plans with corresponding budgets, especially for key national priorities; and (iii) enhance the quality and transparency of public finance reporting for improved responsiveness and evidential policy-making. • Despite the climate and green transition challenges, Kazakhstan has not integrated these aspects into the budgeting and planning of state programs and a risk assessment of its public finances. The Government should consider mainstreaming climate assessment and risk and encourage green public sector performance throughout the budget cycle. 6.1 Strengthening Performance across the Policy Cycle Kazakhstan has made important progress in public sector management and toward greater alignment of national systems with international best practices, including OECD countries. Kazakhstan started introducing performance-based public administration tools in 2007 and is gradually moving to implement performance-based budgeting (PBB) to improve the quality of budgets, including their alignment with strategic plans and performance focus (Beazley, Downes and Nicol 2019). The system has been strengthened and reformed for over a decade, including through the development of M&E functions. In September 2020, President Tokayev announced short-term reforms to address the COVID-19 pandemic and longer-term reforms to revamp public sector performance to improve development outcomes under government strategic priorities. The Government’s public sector management agenda focuses on strengthening strategic planning to drive the design of core national programs and projects and ensure their implementation through strengthened domestic resource mobilization and public financial management (PFM). The Agency for Strategic Planning and Reform (ASPR)—an apex institution at the center of government—was created in 2020 to facilitate progress on policy coordination through a greater focus on strategic planning, budgeting, and monitoring. The President also introduced the concept of a “Listening State,” called for building an accountable government, and raised the need to better plan, manage, and monitor public spending to achieve better development outcomes expressed in the National Development Plan (NDP) 2025. Despite these achievements, the public sector still faces challenges in delivering inclusive services to citizens and the private sector. For example, performance gaps in education with OECD countries remain substantial according to the PISA comparisons (Avvisati, et al. 2019). Persistent regional and rural-urban disparities lead to significant inequalities, with rural areas and specific regions experiencing lower educational attainment, lower average incomes, poorer access to clean water and sanitation, substandard heating, poor local roads and infrastructure, and higher poverty rates. This is further aggravated by limited accountability and capacity by SNGs, as reflected in Chapter 5. Disparities in health access and outcomes are also wide, with national averages in health outcomes hiding troubling regional differences in healthcare access and utilization, health outcomes, and out-of-pocket expenditures. Despite Kazakhstan’s progress on gender equality, there 6. BUDGETING, PLANNING, MONITORING: INSTITUTIONAL FOUNDATIONS FOR THE PUBLIC SECTOR 129 are still significant gender gaps, including labor force participation and wages (Beazley, Downes and Nicol 2019),143 as reflected by the deterioration of the country´s position in the Global Gender Index over the past 15 years (Global Gender Gap Report 2021).144 These outcomes are, in part, a result of poor planning, weak monitoring and evaluation of outcomes, limited execution capacity in certain sectors and regions, and the inefficient use of resources.145 Improving public sector performance will reinforce service delivery and spread economic opportunities, improving inclusion for disadvantaged and vulnerable households and citizens. The discussion in this chapter is linked to other chapters in this report. The budgeting and planning system is part of the public finance management (PFM) system that supports the effectiveness and efficiency of fiscal policies. This chapter discusses utilizing the medium-expenditure framework (MTEF), a tool for fiscal projection, as an instrument for effective budgeting and planning. The issues discussed in this chapter also reflect challenges with budgeting, planning, monitoring, and evaluating the results of key public spending programs, such as education, social assistance, and the active labor market (Chapters 4 and 5). The issues are also linked with the role of subnational governments in delivering public services (chapter 7). Legal and Institutional framework for PFM Over the past 15 years, Kazakhstan has developed its PFM structures and management systems inspired by OECD countries’ experiences. Most PFM aspects are regulated by the 2008 Budget Code, most recently revised in 2022, which covers both the central government and SNGs (see Box 6‑1). The Code sets out how the budgets of both the central government and SNGs will be planned, prepared, and executed, including the timing of different stages and the responsibilities of different institutions and the two Houses of Parliament. The Republican Budget (RB) is revised twice a year and can be amended twice a year during the fiscal year by law and government decree.146 Box 6‑1 . Different budget types and levels in Kazakhstan The Budget Code (BC) provides for distinct budget types and levels: 1) The Republican Budget. This is the centralized monetary fund formed from the proceeds, defined by the BC, and designed to financially support the tasks and functions of the central state bodies, their subordinate state institutions, and the implementation of national state policy. 2) Local budgets: The country is divided into 17 regions (oblasts), plus Astana (capital city), Almaty, and Shymkent, which have the status of cities of Republican significance and are also divided into raions and villages. Local budgets include oblast budgets, budgets of cities of Republican significance and the capital city, raions’ budgets, and the budgets of villages or rural districts. The State Budget is the centralized monetary fund of the state, combining both the Republican and local budgets. Most of the data in this chapter are based on the State Budget. The consolidated budget is the centralized monetary fund of the state consisting of the Republican Budget, local budgets, and revenues and expenditures of the National Fund of the Republic of Kazakhstan (NFRK) and the Victims Compensation Fund. Source: World Bank staff analysis based on information from the Budget Code of the Republic of Kazakhstan (Article 6). 143 “Achievements include women's integration in the labor market and entrepreneurial activity; significant drop of maternal mortality rates; close to gender parity in access to primary and secondary education, ranking among the top countries in this area. (…) Yet, significant gaps remain in terms of delivering gender-responsive policies and services. Women still make up only 8.4% of political-level civil servants; wide gender gaps persist in earnings; and violence against women remains a major problem across the country.” 144 Kazakhstan ranked 32nd in the Gap Gender Ranking in 2006, and 65nd in 2022. 145 Accounts Committee for Control over Execution of the Republican Budget was transformed into the Supreme Audit Chamber by Law of June 8, 2022, “On Amendments and Additions to the Constitution of the Republic of Kazakhstan.” 146 The Budget Code of the Republic of Kazakhstan (chapters 19, 20) provides for amendments to republican and local budgets in situations that threaten the political, economic, ecological, and social stability of the country, in cases of a change in revenues (increase or decrease) by more than 10 percent of their annual approved amount and when there is expected failure to absorb expenditures by more than 10 percent of their annual approved amount. 130 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Three main institutions have key coordinating and leading roles in strategic and budget planning. These are the MNE, MoF, and the recently created ASPR. The division of role illustrates the institutional architecture established by existing norms and regulations. Table 6‑1. Main institutional architecture for planning and budgeting   MNE MoF ASPR Kazakhstan 2050147 National • Reviews/clears • Develops the national • Reviews/clears draft priorities148 draft priorities • Reviews/clears draft • Develops the NDP • Monitors implementation • Prepares assessment/ National • Prepares the annual report on NDP • Reviews/clears opinion on NDP´s Development implementation and submits it to draft implementation and Plan (NDP) the Government and the Presidential submits to Presidential Administration Administration • Develops the NSS National • Develops National Security Risk Security Management Action Plan (NSRMAP) Strategy (NSS) • Monitors and reports to the Security Council NSRMAP implementation Territorial Dev. • Develops and monitors TDP and • Reviews/clears • Reviews/clears draft TDP Plan (TDP) generates reports draft TDP • Reviews/clears draft • Submits opinion on the draft • Monitors and prepares the annual Concepts to the Presidential Concepts for report and assessment on the Administration Development implementation • Reviews/clears • Prepares assessment/ of industry/ • Submits it to Government and then to draft opinion on implementation sphere the Presidential Administration and also submits it to the • Publishes summary on Concepts Presidential Administration implementation on its website • Clears the list of NPs and • Develops the list of NPs approves—with the MNE— • Approves, jointly with ASPR, the the methodology for the methodology for the development, • Reviews the development, monitoring, National monitoring, and reporting of NPs availability and reporting of NPs Projects • Reviews economic feasibility of financial • Reviews compliance with (NPs) • Prepares annual report and opinion on resources for NDP, NSS, and UN SDGs implementation and submits it to the NPs • Prepares its own opinion Government on implementation for submission to the Presidential Administration 147 Developed by the Prime Minister´s Office, Kazakhstan 2050 is being implemented through the subordinate/lower-level documents of the State Planning System. The State Planning System (Government decree of 26 February 2021 #99) is silent on the functions of the agencies in terms of development, monitoring, and evaluation of this document. 148 The implementation of the National Priorities is ensured through the implementation of the subordinate/lower-level documents of the State Planning System. 6. BUDGETING, PLANNING, MONITORING: INSTITUTIONAL FOUNDATIONS FOR THE PUBLIC SECTOR 131   MNE MoF ASPR • Defines format and methodology of Government • Reviews DP DPs agencies’ as part of • Reviews drafts of DPs development ministerial • Prepares annual report on the plans (DPs) budget requests implementation • Sets spending limits • Develops five-year Social and • Prepares Budget Economic Development Forecast conclusions to development • Prepares conclusions to the drafts of budget requests process strategic plans and budget programs • Develops the draft law on the RB. Source: World Bank staff analysis. At the central government level, socio-economic and medium-term fiscal planning is the responsibility of the MNE, which coordinates intergovernmental fiscal relations and regional development and undertakes an economic appraisal of investment projects. The MNE has also been designated the executive body leading and coordinating project management. The newly-created Project Management Office (PMO)149 is co-housed by the Project Management Center of the Prime Minister’s Office and MNE’s Project Management Department. It is mandated to provide operational support to different ministries, departments, and agencies (MDAs) and monitor the implementation of national projects. The Ministry of Finance (MoF) prepares the annual budget and monitors budget execution. It prepares the draft RB and monitors and provides consolidated reports on budget execution.150 It also provides sectoral coordination and methodological guidance on public procurement, budget execution, accounting, auditing, internal public audit, and financial control and reporting.151 The overall budget process and the definition of the expenditure ceilings for each line ministry are supervised by the Republican Budget Commission (RBC) appointed by a Government Resolution.152 The RBC, which includes several ministers and representatives of the National Bank of Kazakhstan (central bank) and parliament, also reviews the draft RB before its submission to parliament. Conceived as an apex institution for high-level coordination, ASPR has a broad mandate ranging from strategic planning and the public administration system to promoting competitiveness, people’s well- being, and state statistics. The ASPR was established in 2020 under the Presidency of the Republic of Kazakhstan as part of institutional reforms to strengthen public administration and strategic planning and improve government program coordination.153 Its institutional objectives include the “development of proposals for the creation and implementation of state policy in strategic planning; participation in the improvement and modernization of the public administration system; development of approaches to the 149 Established by Prime-Ministerial resolution on February 23, 2021, and the Law of the Republic of Kazakhstan dated January 2, 2021, #399-VI ZRK (ЗРК). 150 As reported by the (World Bank 2018) D16.2 -Medium-term expenditure ceilings. Score A. 151 Regulation on the MoF approved by Government decree dated April 24, 2008, # 387. 152 “About the Republican Budget Commission” approved by the Government decree dated 15 January 2018 # 10. Membership reflects the position held. 153 Address of the Head of State Kassym-Jomart Tokayev to the people of Kazakhstan. September 1, 2020: Kazakhstan in the new reality: time for action. 132 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH implementation of reforms aimed at increasing national competitiveness and the well-being of the people; and the development of state statistical activity.”154 Complementing the MNE, MoF, and ASPR, the Administration of the President of the Republic of Kazakhstan (Presidential Administration) is a key player in the existing monitoring and evaluation system. The Presidential Administration (PA) is a key actor providing informational, analytical, legal, protocol, organizational, and other support for the President’s activities. The President approves key strategic documents, including the National Priorities, National Development Plans, Concepts, and the list of National Projects, among others. This requires close involvement and engagement of his office (PA) in the planning process, from development to monitoring and evaluating strategic documents. As part of its activities, the PA established the institutional framework for effectiveness assessments of MDAs and local governments based on the annual evaluation system.155 The Accounts Committee provides Independent oversight of financial management for Control over the Execution of the Republican Budget, which was transformed into the Supreme Audit Chamber through amendments to the Constitution of Kazakhstan in June 2022.156 Its mandate is set out in the Law on State Audit and Financial Control and its Regulations, which establishes that the work of the Accounts Committee is independent, understood as the interdiction of any interference in the performance of state audit and financial control. The Accounts Committee also leads the Annual Effectiveness Assessment of the MDA’s development plans at the central and local levels. There is an opportunity to leverage the existing institutional framework working together across administrative boundaries to achieve an integrated performance-oriented PFM, reducing the duplication of effort, and enhancing coordination between institutions. Several agencies, including the MNE, MoF, and ASPR, play a role throughout the planning and monitoring processes, suggesting there might be some overlaps in their mandates and opportunities to strengthen integration and coordination across these institutions and with other relevant stakeholders. On the one hand, the MNE leads the monitoring and development of consolidated reports for the Government and the PA for Concepts and National Projects. Meanwhile, the ASPR also prepares and submits to the PA its assessment on implementing key strategic documents, including the NDP, the Concept for developing specific industries, and National Projects. As the ASPR has only recently been created, the final design of the mandate, roles, responsibilities, priorities, institutional set-up, and its role vis-à-vis other institutions, particularly the MNE, requires further refinement to promote coordination of M&E efforts. 6.2 Overview of Development Efforts, Progress, and Challenges For over a decade, Kazakhstan has moved toward public sector modernization, including the gradual introduction of PBB, the development of strategic planning, and the establishment of M&E functions and instruments to enhance government effectiveness, efficiency, and transparency. Kazakhstan first introduced some performance-oriented public administration functions and tools in 2007, and since then, it has gradually moved to develop the PBB. Since the 2013 amendment of the Budget Code, the Kazakhstani budget has undergone a series of reforms to improve the quality of budgets, including their alignment with strategic plans and performance focus. Reforms have included the introduction of spending caps for administrators, reducing the number of programs, establishing Public Councils, an updated audit law, and streamlined budget procedures (World Bank 2019). 154 Regulations on the ASPR approved by the decree of the President of the RoK (#427 of October 5, 2020). 155 Annual Effectiveness Assessment System, approved by Presidential Decree #954 of March 19, 2010. 156 Law of the Republic of Kazakhstan of June 8, 2022, “On Amendments and Additions to the Constitution of the Republic of Kazakhstan.” 6. BUDGETING, PLANNING, MONITORING: INSTITUTIONAL FOUNDATIONS FOR THE PUBLIC SECTOR 133 6.2.1 Budget composition and execution rates The PFM system is one of the key tools governments use to turn policy statements and intentions into delivering goods and services.  The PFM system in Kazakhstan includes the Government’s budgets and supporting systems. This chapter focuses on key Public Expenditure and Financial Accountability (PEFA) indicators on budget reliability, transparency of public finances, and policy-based fiscal strategy and budgeting. It is not a full assessment, but rather it identifies some highlights that might be of interest to this PFR. The last PEFA assessment performed in 2018 provides a baseline to assess progress in the most recent years in these areas.157 Budget execution was affected by the impact of the COVID-19 pandemic. The PEFA framework assesses the reliability of government budgets by comparing actual expenditure and revenue outturns with the originally approved budgets. If the budget is reliable, actual expenditure and revenue will be close to what was originally intended, planned, and approved. PEFA 2018 reflected that actual expenditure deviated from the original budget by 5.8 percent in 2015, 6.7 percent in 2016, and 29.0 percent in 2017.158 As shown in Table 6‑2, the difference between actual aggregate expenditure against the originally approved budget for these three years was 6.5 percent in 2018, 11.5 percent in 2019, and 12.7 percent in 2020. The difference was less than 10 percent in only one of the three years, as opposed to the 2015–2017 period when the difference was less than 10 percent in two of the three years. The 2020 figure needs to be analyzed, considering the COVID-19 outbreak at the start of the year. Table 6‑2. State budget execution rate for total expenditures (KZT billion) 2018 2019 2020 Budget total expenditure* 10,655.1 12,143.4 14,833.2 11,346.1 13,535.6 16,725.1 Actual expenditure159 Difference between actual and originally approved 691.0 1,392.2 1,891.9 budgeted expenditure Actual aggregate expenditure as % of originally approved 106.5 111.5 112.8 budget expenditure Source: World Bank staff calculations based on data published by the authorities. Assessing actual spending and the budget execution rate poses a challenge, as the Republican and State Budgets do not represent the full picture of government spending. As indicated in the 2018 PEFA report, expenditure data included in the Republican and State Budgets does not represent the full picture of government spending, as it does not account for spending on services on behalf of the Government provided by SOEs and off-budget funds, which are an integral part of Kazakhstan’s PFM system.160 Currently, spending by the State Pension Fund, the State Social Insurance Fund (SSIF), and the Social Health Insurance Fund (SHIF) is not integrated into the consolidated budget. According to the Bureau of National Statistics, the gross value added of the quasi-public sector in GDP decreased from 18.3 percent in 2018 to 14.9 percent in 2020 (while still representing an increase in absolute terms of 21 percent, from KZT 8,547.2 157 Annex 2 presents a summary of PEFA indicators and scores for Kazakhstan. As a reference, indicators are scored on a scale from A (highest) to D (lowest). 158 This extraordinarily large deviation in 2017 reflects support for national banking sector recovery. 159 Pursuant to Art. 12 of the BC, budget expenditures include expenses, budget credits, purchase of financial assets, and repayment of loans. 160 Subjects of the quasi-public sector: State enterprises, Limited Liability Partnerships, Joint-Stock Companies. Budget Code, article 3, item 1, sub-item 31. 134 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH billion to KZT 10,355.5 billion).161 Expenditures by the economic classification are inconsistent with the Government Finance Statistics Manual (Table 6‑3). For example, Treasury funds allocated to fund SOEs are recorded as acquisitions of goods and services, resulting in disproportionally inflated expenditures on this category, while other expenditure categories, for instance, wages or subsidies, are not fully reflected. This makes it hard to compare the Government Finance Statistics of Kazakhstan with those of other countries, as highlighted by the IMF (IMF 2019).162 Table 6‑3. State budget expenditures by economic classification (% of GDP)   2018 2019 2020 Functional group KZT billion % GDP KZT billion % GDP KZT billion % GDP Total budget expenditures 11,346.05 18.35 13,535.58 19.47 16,725.10 23.67 Current expenses 9,536.67 15.43 11,587.94 16.67 13,732.59 19.44 Wage bill 1,625.74 2.63 1,986.66 2.86 2,583.37 3.66 Goods and services 3,193.80 5.17 3,909.16 5.62 4,393.50 6.22 Interest payments 587.02 0.95 679.08 0.98 767.46 1.09 Subsidies and current 4,130.11 6.68 5,013.04 7.21 5,988.27 8.48 transfers Capital expenses 1,657.22 2.68 1,843.45 2.65 2,822.17 3.99 Acquisition of fixed assets 423.91 0.69 389.05 0.56 683.85 0.97 Fixed capital creation and 1,233.31 2.00 1,454.40 2.09 2,138.32 3.03 repairs Acquisition of financial assets 152.2 0.25 104.2 0.15 170.3 0.24 Source: World Bank staff calculations based on data published by the authorities. 6.2.2 Monitoring budget performance Government effectiveness indicators have improved over the past 20 years, though they are still well below OECD levels and ECA regional averages. Government effectiveness encompasses the quality of public service, civil service, and policy formulation and implementation. Kazakhstan has shown a marked improvement in government effectiveness indicators, measuring government capacity to implement its political vision and deliver desired changes. As measured by the government effectiveness indicator in the World Bank’s Worldwide Governance Indicators, the country ranked close to the 25th percentile in 2000. It improved to the 60th percentile in 2020, as shown in Figure 6‑1. Over the past 20 years, this improvement seems to reflect a path in the right direction, though it still places Kazakhstan far from OECD levels and behind the ECA regional average ranking in the 69th percentile (Figure 6‑2). 161 Conclusion to the (Accounts Committee of the Republic of Kazakhstan 2020). 162 This refers to SEOs which provide public services and relates to their operating budgets. 6. BUDGETING, PLANNING, MONITORING: INSTITUTIONAL FOUNDATIONS FOR THE PUBLIC SECTOR 135 Figure 6‑1. Kazakhstan government Figure 6‑2. Kazakhstan government effectiveness percentile rank (0–100) effectiveness benchmarking, 2021 Source: World Bank staff analysis based on data from World Governance Indicators. A key challenge for Kazakhstan is strengthening the linkages between budget expenditures and government strategic priorities, using performance indicators to inform budget allocation and prioritization. One of the key challenges identified in the final report of the Joint Economic Research Program (JERP) between the Government of Kazakhstan and the World Bank Group (WBG) is the “implementation gap” between the stated plans and policies and the results on the ground for citizens and the private sector.163 There is a wide variation in the quantity and quality of indicators and targets for budget indicators across MDAs. While the MNE provides some central guidance and oversight of developing strategic plans and budget programs, each MDA defines programs and key performance indicators. Performance information is presented in a separate “Explanatory Note” accompanying the main budget document. The annual report on budget execution includes detailed reports on the achievement of targets prepared by the MNE based on reports from line ministries (Beazley, Downes and Nicol 2019). An annual national evaluation system was introduced in 2010 to assess the performance of MDAs and local governments.164 This complex system covers three main areas (Table 6‑4). The first area focuses on achieving goals, including assessing mid-term plans and budget programs. The second area focuses on citizens and other stakeholder engagement. It assesses public services provision, grievance redressal, and open government, including open government portals, such as open budgets, legal acts, data, and open dialogue. The third area focuses on organizational development and assesses human resources, personnel management, and the use of IT. These evaluations are implemented throughout the calendar year by different state agencies. An SOE under the Accounts Committee manages the whole evaluation system.165 Evaluation results are presented to the President and the heads of MDAs at the end of the calendar year, along with specific tasks and recommendations. Based on the evaluation results across the three areas, each MDA and local government is ranked, and the ranking and aggregate results are published online on a dedicated web portal: www.bagalau.kz. The specific recommendations for each MDA and local government are not published. Complementing the annual evaluation system, an M&E framework is in place for the documents of the state planning system (e.g., national and sectorial strategies, National Projects, MDAs, and SNGs’ mid- term development plans) as described in Table 6‑4. Budget evaluation is also part of the budget planning process, focusing on evaluating mid-term development plans and budget program execution. 163 World Bank. 2020. “Kazakhstan: Strategy 2025 From Global Developments to National Policies.” Prepared as part of the Joint Economic Research Program between the Government of Kazakhstan and the World Bank Group. 164 Decree of the President of the Republic of Kazakhstan “On the System for the annual evaluation of the effectiveness of the activities of central state and local executive bodies of regions, cities of republican significance, and the capital city” (#954 of March 19, 2010). 165 LLP “Center for Research, Analysis and Performance Evaluation” of the Accounts Committee for Control over Implementation of the Republican Budget 136 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Table 6‑4. Annual evaluation system of MDAs and local governments Evaluation area Component The agency responsible for the evaluation 166 Strategic component Accounts Committee Goals achievement Budget component Ministry of Finance Agency For Civil Service Affairs Quality of public services Ministry of Digital Development, Innovation, and Aerospace Citizens’ engagement Committee on Legal Statistics and Special Grievance redressal Accounts of the General Prosecutor’s Office Open Government Ministry of Information and Social Development Personnel/HR management Agency for Civil Service Affairs Organizational development Ministry of Digital Development, Innovation, and IT application Aerospace Source: World Bank staff analysis based on the Presidential Decree dated 19 March 2010 # 954 “On the system for the annual assessment of the effectiveness of the activities of central state and local executive bodies of regions, cities of republican significance, and the capital.” While there are several M&E systems and frameworks, these are not functioning as integrated tools to manage performance. First, there are many non-coordinated initiatives, such as the new state planning system, which emphasizes the M&E of strategic documents with a leading role of the MNE and ASPR, and the annual evaluation system, led by the Accounts Committee and MoF. Second, monitoring tools collect routine information on progress against key indicators and budget disbursement. Aggregate disclosed results do not reflect outcomes, but statistical information such as the number of budget programs and indicators, the rate of disbursement, the percentage of budget programs that have achieved or not achieved results, the percentage of achieved/not achieved indicators, etc. There is no clear evidence on the impact and use of performance, monitoring, and evaluation (PM&E) findings on-budget programs, partly due to a mismatch between the budget preparation process and availability of evaluation results, which are distributed to MDAs at the end of the calendar year, eight months after the initiation of the budget process in April of the same year. The existing institutional framework, tools, guidelines, processes, and incentives can be further refined in design, integrated, and rolled out to ensure their use has the desired impact to help deliver the agreed outcomes. This could be done by including implementation status (on-track or off-track) and corrective actions agreed upon. Indicators will need to be linked to work and budget programs. This will also help identify budgeting gaps or duplications, strengthening budget efficiency and effectiveness. Once done, budget and planning indicators (National Development Plan, Concepts, National Projects, development plans of MDAs) must be aligned at different levels so that PM&E can track the implementation of the strategic plans and budget execution (World Bank 2019). The second is to have a periodic spending review of the PM&E system. Spending reviews are widely used as a strategic budgeting tool in OECD countries to systematically analyze expenditures and identify scope for efficiency gains strengthening government control over the level of aggregate expenditure and to improve expenditure prioritization (OECD 2019). Spending reviews also play a crucial role in helping governments manage the enormous budgetary challenges of the post-pandemic era, particularly in identifying reductions to fund new priorities (reallocation) and control total expenditures (Robinson 2022). Requiring completion reviews for key projects, including actual and planned cost comparisons, could facilitate implementation. Independent ex-post evaluation of large and strategic government projects and programs, with lessons learned widely disseminated, is also being considered. 166 Measures are taken to eliminate conflicts of interest. For example, if the MDA is responsible for evaluation, its evaluation is performed by a high level institution. For example, evaluation of the Ministry of Finance on Goals achievement is performed by the Chancellery of the Prime-Minister; MDAs subordinated and accountable to the President are evaluated by the Presidential Administration; and so on. 6. BUDGETING, PLANNING, MONITORING: INSTITUTIONAL FOUNDATIONS FOR THE PUBLIC SECTOR 137 6.2.3 Strengthening the focus on performance and accountability in public budgeting Over the next few years, considerable effort will be required to build capacity in the central government and line ministries to prepare and appraise programs, establish performance accountability processes, and develop program statements to accompany the Annual Budget Law (World Bank 2018). Currently, activities are tracked in a control and punish mode rather than embedded in a culture of continuous performance improvements based on testing, innovating, identifying lessons learned, and ensuring continuous adjustment.167 The quality and relevance of performance measures could be further improved by incorporating both process and outcomes indicators on service delivery, including service delivery quality and efficiency and measures of citizen satisfaction focused on citizen experience. A cultural shift toward performance, supported by a change in management strategy, could be promoted. France, for example, developed a comprehensive training and communication program for public sector officials as part of the country’s PFM reforms, initiated by the Organic Law relating to Finance Laws, the Loi Organique relative aux Lois de Finances (see Box 6‑2). Box 6‑2. France - Building capacity for performance measurement Performance budgeting was the centerpiece of overhauling the French PFM system, the Loi Organique relative aux Lois de Finances, the law that initiated the reforms. Implementation required analytical and process skills beyond a traditional budget department's typical financial and economic skills. It included understanding sectoral policies, program intervention logic, and the science of performance measurement. Starting in 2001, the reforms were implemented over five years. They included the development of a programmatic structure for the budget, the relaxation of line-item budget controls, and the introduction of a new financial management information system (CHORUS). Key to successful implementation was the extensive training of personnel and an active program to communicate the changes. The new system was implemented only in 2006 after five years of intensive preparation and piloting and has remained broadly stable since that date. The Financial Committee of the National Assembly wrote in a 2011 report, “No doubt one of the most important successes of the Loi Organique relative aux Lois de Finances after five years is the dissemination of a performance culture and management in the French civil service, thanks to the chain of accountability.” Source: (OECD 2019). 6.3 Strengthening Linkages between Strategic Planning and Budgeting To help concretize and implement the vision, the Government approved National Priorities that cascade down from the Kazakhstan Strategy 2050. National Priorities are endorsed by the High Council on Reforms under the Presidency of Kazakhstan and approved by the President. The National Development Plan (NDP) 2025 establishes the country’s strategic direction in the medium term to achieve National Priorities. In 2021, Kazakhstan changed the traditional system of national planning. Acknowledging the abundance and, at times, duplication of national and sectoral programs, the Government has reduced the number of 167 Legislators introduced responsibility in the Budget Code (for example, art. 104-1 of the Budget Code). 138 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH national programs that guide the NDP. The new four-layer State Planning System168 and strategic framework is based on a long-term strategic vision, Kazakhstan Strategy 2050, as the high-level expression of strategic aspirations, as shown in Figure 6‑3. National Projects—which define inter-regional or inter-sectoral policies— aim to contribute to achieving National Priorities and objectives and require priority budget funding.169 The President has recently approved the list of national projects, “Concepts” covering areas such as health care, education, strong regions, the agro-industrial complex, and others.170 The Concept is a document that defines the development vision for 5–10 years of a specific sphere/industry, as well as the basic principles and approaches to implementing the relevant policy.171 The expected results of the Concept require alignment with the NDP objectives. The Concept is implemented through National Projects, plans for developing state bodies, regions, and cities, plans for developing national companies, laws, and other regulatory legal acts. Figure 6‑3. Kazakhstan National Planning System Source: World Bank staff analysis. The Budget Code (BC) requires budget program administrators to link expenditures with goals set in their strategic plans. Kazakhstan has a system of three-year rolling budgets, with the outer two years representing indicative rather than binding commitments. As highlighted by the PEFA 2018, ministries’ medium-term strategic plans are prepared within the framework set by the MNE, specifying the ministries’ strategic objectives and actions to be taken and aligned to medium-term budget estimates.172 Expenditures are linked to the MDA’s functions and mandates without a strategic plan. Budget programs covering current and capital expenditures for the three years ahead must be elaborated based on strategic plans. Strategic plans are part of the budget documentation, as the Budget Code stipulates, and should be reviewed and 168 The Renewed State Planning system approved by the Government Decree of February 26, 2021, No. 99. 169 On February 26, 2021, the authorities approved National Priorities, the National Development Plan 2025, the Public Administration Development Concept until 2030, and the State Planning System. 170 The Decree of the President of the Republic of Kazakhstan, # 670 2021. List of national projects include: “High-quality and affordable health care for every citizen,” “A healthy nation, ” "Quality education," and “Educated nation,” "Ulttyk rukhani zhagyru," "Technological breakthrough through digitalization, science and innovation," “Development of Entrepreneurship,” "Strong regions - the driver of the country's development," "Sustainable economic growth aimed at improving the welfare of Kazakhstanis," "Green Kazakhstan,” “Development of the agro-industrial complex,”, and a "Safe Country". 171 Examples of recently approved Concepts include the Public Administration Development Concept until 2030; Local Self-Government Development Concept in the RoK until 2025; and the Concept of the Legal Policy of the RoK until 2030. 172 (World Bank 2018). Indicator D-16.3 scores A (the highest score). Budget programs covering both current and capital expenditure for the three years ahead should be elaborated based on strategic plans. The costs of the strategic plans thus align with budget programs. 6. BUDGETING, PLANNING, MONITORING: INSTITUTIONAL FOUNDATIONS FOR THE PUBLIC SECTOR 139 agreed on by the MNE and MoF before submission to parliament. The MNE analyses the consistency of strategic plans with the Government’s NDP and other national program documents. MoF reviews the alignment of the costs and financial resources for implementing strategic plans with medium-term budget estimates, considering budget constraints and ceilings. Budget programs are required to have both ex-ante and ex-post monitoring.173 Ex-ante monitoring is performed during the budget preparation and clearance process. All budget programs go through preliminary clearance with MoF, MNE, and the RBC and are finally submitted to parliament. Ex-post monitoring reports are also mandated by the Budget Code (Article 112), which requires administrators to submit monthly progress reports to the MoF on the execution of the Republican Budget. These reports showcase outputs and direct results, indicate actual and planned expenditures, and provide reasons for undisbursed amounts. As such, most indicators cover the MDA’s current functional activities, with weak links to strategic actions, a lack of performance or outcome orientation, and poor information regarding investment projects’ economic, budget, and social effects. Despite these instruments, strategic plans and budget linkages reflect formal compliance with BC requirements rather than a powerful performance-oriented tool. According to the Accounts Committee’s report on the 2016 Budget Execution, about 60 percent of the expenditure of budget programs is covered by the strategic plans with performance targets.174 However, a more recent report by the Accounts Committee finds that budget programs are mapped to many strategic objectives. Indicators for budget programs are mainly process or output-oriented—reporting, presentations, polls, research, etc.—not linked to actual program results and much less to strategic goals (Accounts Committee of the Republic of Kazakhstan 2020). Hence, PM&E and budget reorientation based on public spending efficiency and effectiveness become challenging. In its Commentary on the Government’s Report on the RB 2020 execution, the Accounts Committee has also pointed to insufficient links between investment project cycles, budgeting, and PM&E, including economic and social effects (Accounts Committee of the Republic of Kazakhstan 2020). Table 6‑5 summarizes key findings of the Accounts Committee Report on 2020 Republican Budget execution. Table 6‑5 Examples of findings of the Accounts Committee report on the 2020 Republican Budget Execution SP for development SP for development State Program (SP) Industrial – Innovation of employment and of the agro-industrial Nurly Zher 2020–2025 Program 2020–2025 entrepreneurship 2017–2021 sector 2017–2021 • Inadequate quality of • Weak links across • Lack of strategic planning and execution • Insufficient planning indicators at approach for of state program (program was corrected 10 different levels. housing provision. indicators. times after its adoption). • Absence of unified • Risks in the • Weak coordination • Improper costing of the approach to Government’s and inefficient expenditures. agricultural sector targeting approach. implementation • Discrepancies between statistical data. monitoring. • Weak control of the Government reports and • Weak monitoring distribution of social • Failures to comply State program financing of indicators and housing. with budget and other amount and sources.175 projects. legislation. Source: World Bank staff calculations based on data published by the authorities. 173 OECD, 2019. 174 Given that the coverage of the strategic plans and performance targets was less than 75 percent of the budget expenditure (World Bank 2018) score for the indicator “performance plans for service delivery” was a C. 175 For example, the use of funds from local budgets (KZT 55 billion), was not accounted for in the State Program for 2020–21, resulting in a 33 percent excess of allocated funds over the originally approved amount. 140 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH The recently approved State Planning System can potentially strengthen linkages between planning instruments and objectives and the budget process, including performance targets and results. The Government is committed to prioritizing budget allocations and enhancing the efficiency of public spending, as reflected in the President’s latest address.176 Building on the recommendations made by the JERP 2020–2021, ministries could propose increased operational performance indicators linked to their budgeted programs and subject to a budget envelope. The ministries could propose budget increments (or decrements) for important individual programs, along with the corresponding increase of the operational performance indicators. This would inform the budget negotiation between the Treasury, MNE, and the respective ministry, to agree on the budget and performance indicators. In this sense, budgeting would be bottom-up and top-down and more informed by performance decisions. Over time, PM&E could be further aligned so the monitoring to include the design of performance indicators, respective reporting templates that link the indicators with information on financial performance, the tracking of their achievement, and an adjustment to the different budget phases to enhance the alignment. The medium-term expenditure framework (MTEF) can be further used as a tool for strategic resource allocation by providing a more detailed rationale for decisions to shift resources between and within sectors based on alignment with policy priorities. MTEFs can improve the effectiveness of public spending by aligning public expenditure with national priorities and giving government agencies greater certainty of resource availability over a multi-year period, promoting more effective planning and resourcing of policies that require an extended time horizon for implementation, such as large capital projects, new programs, and organizational restructuring. As many international best practices demonstrate, strategic planning has its greatest impact when designed not as a separate, stand-alone, aspirational function but as an integrated one. Box 6‑3 summarizes the Malaysian experience. Box 6‑3. The impact of strong linkages between national development planning and budgeting: The case of Malaysia National development planning in Malaysia has been central in guiding economic policy-making for over 60 years. Many dividends have been due to good planning and implementation, with political commitment from the highest levels. This was backed further by budget resources and a strong mandate for the Economic Planning Unit (EPU), which is also responsible for the capital budget. Malaysia practices a dual budgeting system, whereby the development expenditure (DE) is prepared by the EPU, and the operating expenditure comes under the Ministry of Finance (MoF). For every five-year plan, the EPU determines the ceiling for sectoral development expenditure based on a general DE allocation and ensures that the projects proposed by the ministries, departments, and other government entities are consistent with the development plan and are within the project ceilings. The DE amount is estimated in consultation with and based on feedback from the MoF. The MTFF and adoption of the new performance budgeting system—Outcome-Based Budgeting (OBB)— are reforms that have occurred in parallel with the development planning process. Using outcome-based budgeting, Malaysia has successfully linked high-level national strategies to specific budget programs and activities using a common results framework to define accountability. The outcome-based budgeting reform was instrumental in enabling the Government to re-orient budget preparation toward achieving clearly-defined policy outcomes. Change management, staff training, and stakeholder outreach were central to the implementation strategy to promote buy-in from key actors, mitigate resistance to change, and strengthen sustainability. Source: (Brownt and Govindasamy 2019). 6.4 Transparency of Public Finances Enhancing the transparency of public finances has been a priority over the past ten years, though implementation 176 remains Address from the Head of Statea key challenge. Kassym-Jomart Transparency Tokayev and accountability are the greatest challenges to the people of Kazakhstan. 6. BUDGETING, PLANNING, MONITORING: INSTITUTIONAL FOUNDATIONS FOR THE PUBLIC SECTOR 141 to inclusive growth, investment, and competitiveness, hindering reform in most areas, as highlighted by the Systematic Country Diagnostic (SCD) (World Bank 2018). Kazakhstan’s NDP 2025 includes the concept of the “Listening State,” a pledge to further transparency and engagement with civil society and a mechanism for ensuring a continuous dialogue between the Government and society on important reform areas. The Public Administration Development Concept 2030177 includes an “accountable State” as one of its principles. However, current efforts seem driven to comply with legislative requirements, with little evidence of citizen engagement/participation. Transparency International’s Corruption Perceptions Index ranked the country 94 out of 180 countries in 2020. With a score of 38 out of 100, Kazakhstan has made important improvements, jumping 11 points since 2011. Still, it remains close to the average of the ECA region—the world’s second-worst performing region—and far from the average score of 66 for Western and central European countries.178 The January 2022 events, marred by violence and attempts at destabilization, clearly pointed to the need for faster progress on reforms to achieve sustainable growth and shared national prosperity, including transparency. 6.4.1 Independence in audit institutions and public sector accounting standards Legislative scrutiny and independent oversight occur throughout the budget cycle but with limited independence of audit institutions. Parliament participates in the initial formulation of the Government’s budget through its representatives in the RBC, which approves the MoF’s budget ceilings. After that, the budget committee requires that government proposals are submitted to parliament by September 1 each year.179 Parliament also approves the Government’s annual report on the Republican Budget execution and may set up special hearings inviting the Government and MDAs. Hearings are broadcasted through the parliament´s websites and national TV channels, and the parliament’s recommendations to the Government are also published on the parliament’s website. A similar process takes place at the local level, where local budgets are developed by local executive bodies (akimats) and approved by local representative bodies “maslikhats.” However, the fairness of inter-budgetary fiscal relations is being questioned, along with the fund allocation process, which seems to be associated with the personal influence of the heads of MDAs and regions.180 Independent oversight is provided by the Accounts Committee. However, as a Supreme Audit Institutions Independence Index reported, Kazakhstan meets few independence indicators and has been assessed as having a Supreme Audit Institution lacking independence (World Bank 2021).181 Some shortcomings identified in reviewing compliance with International Public Sector Accounting Standards (IPSAS) undermine the reliability and verifiability of financial reports. The Consolidated Public Financial Statement for 2019 was reviewed for compliance with IPSAS under JERP 2020–2021.182 The key findings of the report and respective recommendations on further steps for full IPSAS implementation include 183 (i) the absence of certain important disclosures required by IPSAS; (ii) quasi-state entities are not included in consolidated financial statements; and (iii) segment reporting is missing. The accounting policy 177 Adopted by Presidential Decree of February 26, 2021, #522. 178 Transparency International Corruption Perception Index 2020. 179 Verified by (World Bank 2018) Indicator D-17.3 Budget Submission to the Legislature. Score 2018: A. Drafts laws with the proposed republican budget for 2018–2020 were also submitted on time. 180 State of the Nation Address by the President of the Republic of Kazakhstan Kassym-Jomart Tokayev: September 1, 2021. Speech by the Head of State K.K. Tokayev at a meeting of the Mazhilis of the Parliament of the Republic of Kazakhstan; “Lessons of the tragic January: the unity of society is a guarantee of independence: January 11, 2022. 181 Supreme Audit Institutions Independence Index 2021 Global Synthesis Report. Indicators measure the constitutional and legal framework, transparency in appointing the SAI head, financial autonomy, types of audits, operational autonomy, staffing autonomy, audit mandate, audit scope autonomy, access to records and information, and right and obligation on audit reporting. 182 World Bank (2019) Joint Economic Research Program (JERP). 183 International Public Sector Accounting Standards are a set of accounting standards issued by the IPSAS Board for use by public sector entities around the world in the preparation of financial statements. 142 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH lacks qualitative elements concerning financial information that IPSAS requires, such as those relevant to reliable representation, timeliness, and verifiability. The audit report issued by the Accounts Committee for 2020 confirmed the above findings. 6.4.2 Open-Budget and citizen engagement Kazakhstan has made budget documentation widely available. An Access to Information Law, approved in 2015, further reinforces the constitutional right of citizens to obtain information held by public bodies (with limited exceptions). It encompasses a right to request and receive information and an obligation for governments to publish information proactively. Publication of budget documentation includes pre-budget statements, the Executive’s budget proposals, the enacted budget, citizens’ budget, in-year reports, mid- year reviews, year-end reports, and audit reports (International Budget Partnership 2019). The Government has developed different Open Government initiatives, including (i) monitoring and communication channels with citizens (e.g., a single window for communication); (ii) commitments for in-time responses to the requests from non-governmental organizations (NGOs), media, and citizens; and (iii) information disclosure, etc. In addition, audit reports prepared by the AC, the Committee for Internal Public Audit (CIPA), and audit reports prepared by the Internal Audit units of most ministries are published on their respective official websites. Table 6‑6 summarizes Open Government tools and portals for enhanced transparency.184 Table 6‑6 Open Government for enhanced transparency Instrument Open data A portal providing access to several datasets generated by open.egov.kz/ government agencies Open budgets A website created to facilitate citizens’ understanding of the budget.egov.kHz/ budget, promoting engagement through public monitoring of budget spending MNE’s website. www.gov.kz/ MNE publishes development plans of state agencies on its portal memleket/entities/ economy MoF’s website www.gov.kz/ MoF publishes information on Republican, state, consolidated, and memleket/entities/minfin citizens’ budgets. Open legal Acts A portal for public consultations on draft legal acts. Republican legalacts.egov.kz Budget draft laws are published on the portal and open for comment for two weeks Open dialogue An open platform to facilitate dialogue between the citizens and dialog.egov.kz the state. Users can directly submit an appeal and send proposals, report on the quality of services, and participate in socially significant surveys Interactive map of open budgets A portal with access to the regional budget for schools, publicbudget.kz kindergartens, hospitals, polyclinics, and roads Source: World Bank staff analysis. These initiatives have positively impacted Kazakhstan’s Open Budget Survey (OBS) transparency indicators rankings. The OBS measures public access to information on how the central government raises and spends public resources and other variables. It assesses the online availability, timeliness, and comprehensiveness of key budget documents and scores each country on a scale of 0 to 100. A transparency score of 61 or above indicates a country is likely publishing enough material to support informed public debate on the budget. Kazakhstan ranks 28 out of 120 countries, and its overall score has improved from 58/100 in 2019 to 63/100 in 2021, slightly above the threshold marking a sufficing score (61) and ahead of the ECA region and global 184 averages (Figure 6‑4) (International Budget Partnership 2021). The OBS public participation indicator, Approved budget programs, monthly and annual budget execution reports are disclosed on the official websites of all ministries. 6. BUDGETING, PLANNING, MONITORING: INSTITUTIONAL FOUNDATIONS FOR THE PUBLIC SECTOR 143 which measures the formal opportunities offered to the public for meaningful participation in the different stages of the budget process, assigns Kazakhstan a score of 9 out of 100 in public participation in 2021, a deterioration from its 2019 score of 17/100 and below regional comparators (such as Georgia at 44/100) and the global average (14/100). This is confirmed by the voice and accountability indicator of the WGI 2020, where Kazakhstan ranks in the 16th percentile, significantly below the ECA region average of 66th percentile and high-income OECD countries (87th percentile) (World Bank 2020). Figure 6‑4. Kazakhstan Open Budget Survey 2021 transparency indicator Source: World Bank staff calculations based on data from Open Budget Survey. While the publication of budget information and documentation has been an important step toward increased transparency, there is still little evidence that this has facilitated and promoted citizen engagement. The draft and approved versions of the Republican Budget and budget programs are disclosed to the public through government websites. Parliament receives more comprehensive and detailed information than the public, including estimations and calculations, and can request further information and justification of estimations. There appears to be little consistency in the disclosure of audit reports by the Accounts Committee. Only some audit reports are available on the official website, while most other audit results are missing. Information about public procurement and other government activities is available on government websites. But the availability of information has not translated into participation. While a deeper understanding of what drives the citizen-state relationship is required to progress in this area, it seems that the country has put considerable emphasis on formal compliance with regulatory requirements but not on actual use. While the information on public procurement and other government activities is available on government websites, it is not presented in a consolidated form that would aid public understanding. Civil budgets are presented in a general format with information not meant to generate citizens’ feedback.185 The OBS highlights the need to expand mechanisms to engage civil society during budget formulation and implementation. Civil society organizations are relatively underdeveloped in Kazakhstan. There is limited access to venues for expressing citizens’ needs, and responsiveness to those needs is weak. Social accountability mechanisms are fragmented, and citizens have little voice. As an example of low participation and engagement, the draft law with amendments to the 2021–2023 Republican Budget has not received any comments (either because the public is not informed and trained to comment, or else its senses that its comments are 185 Analysis and recommendations of the expert group on the format of citizen’s budget. INTEGRITY ASTANA. 2020. 144 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH not used for policy-making).186 Possible initiatives to further facilitate civil society participation in the legislative process could include capacity-building activities, public awareness campaigns, modernized and more user-friendly public consultation processes, and a greater disposition to consider and incorporate citizens’ feedback into policy-making decisions (World Bank; OECD; European Bank for Reconstruction and Development; Islamic Development Bank; Asian Development Bank 2019) (U.S Department of State 2020). Box 6‑4 presents Mexico’s experience in developing a Budget Transparency Portal to allow public access to budget information, focusing not only on making information available but also doing it in a way that is useful and accessible for users. Box 6‑4. Mexico - Budget Transparency Portal Since 2011, the Federal Government of Mexico has had a Budget Transparency Portal that presents performance information that users can interpret without deep knowledge of budgetary processes, using infographics and geo-referenced information. It also provides several open datasets that analysts and researchers can use. Mexico was the first country to formally adopt the international open fiscal data package’s specification promoted by the Global Initiative for Fiscal Transparency, Open Knowledge International, and the World Bank. Mexico’s Federal Government also formally implements the Open Contracting Data Standard and has formally released an open contracting portal as part of its Extractive Industries Transparency Initiative membership. This strategy includes information and data accessible through the internet, dissemination through social networks, training of public officials and other stakeholders in performance budgeting, as well as the use and interpretation of performance information, and exercises involving civil society organizations to promote the use of this information. The Public Councils initiative aims to broaden opportunities for citizen engagement in the decision- making process and can be further strengthened for improved impact. The Law on Public Councils of 2015 established the legal and regulatory framework for Public Councils in Kazakhstan and named the Ministry of Information and Social Development as the implementing institution. The Public Councils are advisory bodies established across all MDAs, local governments, and the largest SOEs to provide a platform for civil society to voice its opinion on important public issues. Most members come from the non- governmental sector (Ministry of Information and Public Development of the Republic of Kazakhstan 2021). Box 6‑5 summarizes the main activities of the Public Councils. Over five years after their implementation, an OECD evaluation revealed that Public Councils had been established to comply with the 2015 Law and need further improvements in terms of performance, greater transparency, and diversification of their composition (OECD 2020) (Ministry of Information and Social Development of the Republic of Kazakhstan 2020). The latest sociological survey illustrated that 70 percent of respondents were unaware of the existence of Public Councils or their activities (Ministry of Information and Social Development of the Republic of Kazakhstan 2020). 186 “On Amendments and Additions to the Resolution of the Government of the Republic of Kazakhstan” dated December 10, 2020 No. 840 and "On the implementation of the Law of the Republic of Kazakhstan" was placed on the legal acts e-portal from May 28 to June 11, 2021. 6. BUDGETING, PLANNING, MONITORING: INSTITUTIONAL FOUNDATIONS FOR THE PUBLIC SECTOR 145 Box 6‑5. Kazakhstan’s Public Council activities • Discuss draft budget programs, draft strategic plans or regional development programs, and draft state and governmental programs. • Discuss budget program performance, implementation, and results of strategic plans or regional development programs, state and governmental programs. • Discuss executive bodies’ reports on progress against target indicators. • Discuss reports of the budget program administrator regarding the implementation of budget programs, the implementation of revenue and expenditure plans in connection with the selling of goods (works, services), and the revenue and expenditure in connection with charitable initiatives. • Participate in developing and discussing draft legal acts regarding citizens' rights, liberties, and duties. • Consider appeals lodged by natural and legal persons regarding improving public administration and ensuring transparency of state operations, including the observance of the regulations of service ethics. • Develop and discuss with authorities proposals for improving the legislation of the Republic of Kazakhstan. • Carry out public control in other forms as prescribed by this Law. Source: (KAZKENES n.d.). Building on the pilot program experiences, the Government should expand the participatory budgeting practice. Pilot initiatives were undertaken in Almaty and Astana in 2019,187 which promoted citizens’ involvement in public investment decision-making related to local infrastructure. The information about certain projects is disclosed on local executive bodies’ (akimats) websites, with details and deadlines for submitting proposals and votes. Citizens participated in: decisions about landscaping; construction and repair of sidewalks, ramps, and irrigation ditches; creation, repair, and lighting in common areas; elimination of landfills; children’s playgrounds in courtyards and public areas; and others. While this is an important first step, pilot initiatives revealed that the participatory budgeting program could be further refined, strengthening the linkages with PBB and budget transparency activities and clarifying the roles and responsibilities of key stakeholders (the MoF is responsible for participatory budgeting as the state agency in charge of budget execution, while the MNE leads on budget policy) and promoting the development of budget literacy among the population, including in settlements with local self-governance. 6.5 Mainstreaming Climate Change in Public Sector Performance Kazakhstan has made some progress in addressing climate change and green growth challenges, but stronger action is needed to integrate climate-responsive and green growth strategies into the budget cycle. The Government is committed to developing a Concept of Low Carbon Development until 2050 to achieve the Paris Agreement commitments. Under the Paris Agreement, Kazakhstan has committed to an unconditional 15 percent reduction of greenhouse gas (GHG) emissions from 1990 levels or a 25 percent reduction conditional on international support. Kazakhstan Vision 2050 emphasizes the country’s transition to a “green” development path. In 2013, the President approved the “Concept for Transition of the Republic of Kazakhstan to a green economy.”188 The country also includes climate change-related 187 MoF Decree of March 12, 2020, # 254. 188 May 30, 2013, # 577. 146 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH actions in its National Development Plan 2025. However, much must be done to translate these priorities and climate responsive short-, medium-, and longer-term goals into action. The 2021 Environmental Code has institutionalized climate change governance, but gaps in implementation and coordination remain. According to the Environmental Code, the Ministry of Ecology and Natural Resources (MENR) is the leading agency responsible for the climate change agenda, including implementing regulations on emissions and the removal of GHG, meteorological and hydrological monitoring, developing and approving carbon budgets and serving as a focal point on implementation of international treaties.189 At the same time, almost all MDAs and SNGs have a role in climate change adaptation, mitigation, or reporting. They are required to assist MENR in implementing its functions and take actions to ensure the implementation of the Nationally Determined Contributions (NDCs). However, the existing institutional architecture does not specify the coordination mechanisms for strategy setting, budget planning and monitoring, and evaluation of climate change policy. MENR also lacks some of the political powers, resources, and capacity to set and coordinate the climate change agenda. On the budget, for example, while the Environmental Code provides for the implementation of green finance, the Budget Code (the country´s main PFM regulation) does not include such a provision. The emergence of different “green initiatives” spread throughout MDAs leads to the dispersion of effort and resources, both human and financial, risking impact and sustainability. The National Strategy 2050, the National Development Plan (NDP) 2025, and many National Projects include “green” initiatives, such as building alternative energy sources or using green financing. However, these are not streamlined into an integrated green growth strategy; neither are they coordinated across the public sector, compromising the achievement of national goals. For example, the NDP 2025 aims to facilitate investments in waste and water management, energy-efficient and smart technologies, and renewable energy sources and to streamline green financial instruments for financing “green projects.” However, the NDP 2025 and the Mid- term National Project “Green Kazakhstan” have not established specific goals or joint initiatives to support Kazakhstan’s NDC implementation, nor have they provided projections of GHG emissions against the target, risks assessments, estimated costs, and financing sources.190 The lack of an integrated approach to address climate change priorities was also reflected in the 10 National Projects approved in October 2021, with only one-third having scattered “green” actions. There are no clear estimations of “green” revenues and expenditures, and financing sources to smooth the climate change transition are also poorly identified. The lack of tools to identify a green budget throughout the budget planning process (such as climate budget tagging, green financing, green subsidies, green budget credits, sovereign green bonds, etc.) limits the possibility of measuring both green budget allocation and impact. Economic stimulus tools for environmental purposes are currently scattered under provisions of the Environmental Code across different regulations and MDAs.191 The disbursement of revenues collected from environmental fees is complicated, as there is no consolidated statement of total collected amounts and, much less, any assessment of their contribution to environmental objectives. Climate change impact assessments were not incorporated into the existing state M&E system or the state’s audit and financial control framework. By integrating those into the budget cycle, Kazakhstan could strengthen climate-responsive and green growth strategies, policies, plans, and actions. The budget cycle is the single-most-important action-forcing mechanism in government, expressing its priorities, strategic investments, and ultimately, the Government’s 189 Order of the President of the Republic of Kazakhstan dated 17 June 2019 #17 “On measures to further improve the public administration system of the Republic of Kazakhstan.” Environmental Code. 190 National Project “Green Kazakhstan” approved by Government decree dated 12 October, 2021. 191 EcoCode, art 126 6. BUDGETING, PLANNING, MONITORING: INSTITUTIONAL FOUNDATIONS FOR THE PUBLIC SECTOR 147 short, medium, and longer-term goals. Green budgeting integrates climate and environmental perspectives into existing budgetary frameworks, tools, and processes, which include fiscal risk management, medium and long-term fiscal statements, and performance budgeting. OECD countries have been introducing green budgeting initiatives to manage resources of climate and environmental impact (OECD 2022), and below are some examples that could be considered as entry points: • A clear, integrated legal and policy framework for climate change could be established at the planning and policy-making level, building on the Paris Agreement commitments and the National Development Plan 2025 and upgrading relevant strategic and policy documents as needed. • At the PM&E level, performance indicators relevant to climate and environmental policy objectives could be included. • At the medium-term budgeting level, integrate climate appraisal to assess government programs. • From the budget’s revenue side, options to be considered include green taxes and bonds. • Different instruments can be developed from the expenditure side, including climate change expenditure reviews, budget tagging, and green public investment management and procurement. One of the most often used tools to strengthen climate budgeting is budget tagging, identifying, measuring, and monitoring climate-relevant activities and expenditures. It builds on prior experience in tagging for other whole-of-government policy objectives, such as poverty, gender, and the international development goals outlined by the United Nations. Several countries already apply climate budget tagging alongside tags for these policy objectives. There are three essential design elements to climate budget tagging methodologies: (i) definition of climate-relevant expenditure; (ii) definition of appropriate coverage; and (iii) estimation of climate-relevant spending (Box 6‑6 presents the implementation of budget tagging in Indonesia). Box 6‑6. Budget tagging in Indonesia The Ministry of Finance (MoF) of the Government of Indonesia implements climate budget tagging at the national and regional levels to determine the State Budget's and Subnational Budget's contributions to climate change. It involves specific interventions in the revenue, expenditure, and financing policy. State revenue policy: To support the development of renewable energy, as well as environmentally friendly business areas, MoF has implemented tax facilities in the form of tax holidays, tax allowances, import duty exemptions, VAT reductions, government-borne income tax, and reductions of property tax to support geothermal and other renewable energy. State expenditure policy: To support low-carbon and climate-resilient government spending, MoF has implemented climate budget tagging at the national and regional levels to determine the State and Subnational budgets' contributions to climate change. Financing policy: To support comprehensive fiscal policy through the development of innovative financing instruments, such as the Sovereign Green Sukuk (Green Islamic Bond), including both global green sukuk and green sukuk retail, to finance climate mitigation and adaptation projects. Source: (World Bank 2021). Disaster-resilient and responsive PFM could also be implemented to determine the capacity of the PFM systems to respond to the impacts of disasters. It does so by measuring the extent to which disaster resilience and recovery considerations are integrated into key PFM functions and processes. Specific entry points can also be explored on SOE governance (e.g., climate change-related financial disclosure, green corporate governance, procurement, transparency, and reporting) and sub-national governance (intergovernmental transfers and fiscal decentralization). 148 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 6.6 Mainstreaming Gender Agenda in the Public Budgeting In Kazakhstan, steps have been made to promote the gender agenda, though initiatives are still dispersed among strategic documents and budget programs. The Concept on Family and Gender Policy 2030, approved in 2016, provides the conceptual framework for gender policy. It reaffirms the country’s focus on strengthening gender-based equality, expanding economic and political opportunities for women, its commitment to combating violence against women, and promoting gender-related knowledge.192 However, cascading from the Concept to sectoral strategic and policy documents has been weak and fragmented. For example, while the 2016 Family and Gender Concept calls for the introduction of gender-based budgeting, with initial pilots in 2022, such an initiative has not been incorporated into the Public Administration Development Concept 2030, nor does it appear in the existing Budget Policy Concept or the draft of PFM Concept circulated for public consultations.193 The NDP 2025 has not included the gender agenda, apart from a few actions linked to women’s health care and supporting women who have experienced domestic violence. Some MDAs, such as the Ministry of Labor and Social Protection, track gender-based actions and have dedicated budget programs to support women with many children, pregnant women and women on maternity leave, and those who have suffered domestic violence.194 While these constitute important steps, a systematic approach to gender budgeting could be implemented to track and support achieving gender equality objectives in Kazakhstan. Gender budgeting is the application of gender mainstreaming in the budgetary process. It incorporates a gender lens into the budget process to ensure that governments are acutely aware of the impact of their choices on gender outcomes (Curristine, et al. 2021). Kazakhstan could strengthen its application of gender budgeting practices across the budget cycle by mainstreaming the national gender equality strategy within MDAs and focusing on analyzing the differential impact of the Government’s budget on women and men, thereby translating government commitments on gender equality into budgetary commitments, developing a Gender Budget Statement illustrating how the annual budget supports national gender equality goals, and applying a gender perspective to program evaluations and impact assessments. Box 6‑7 summarizes certain applications in OECD countries that range from the legal and institutional architecture for gender budgeting to budget preparation and execution and performance monitoring. Lessons learned from gender budgeting implementation point to political support, binding legislative requirements, and the active support of finance ministries are key drivers of gender-budgeting success. The Government also needs to make gender-disaggregated data available for gender aspects to be properly considered in policy design and for gender outcomes to be properly evaluated (Curristine, et al. 2021) (Downes, Trapp and Nicol 2017). 192 Approved by Decree of the President of the Republic of Kazakhstan on December 6, 2016, #384. 193 Public Administration Development Concept approved by the decree of the President of the RoK dated 26 February 2021 #522. New Budget Policy Concept approved by the Decree of the President on June 26, 2013, #590. 194 According to the report of the Ministry of Labor and Social Protection on the implementation of its Development Plan for 2020-2024 approved by the decree of its Minister on 30 December 2019 #709. 6. BUDGETING, PLANNING, MONITORING: INSTITUTIONAL FOUNDATIONS FOR THE PUBLIC SECTOR 149 Box 6‑7. Gender budgeting tools and practices – OECD countries’ experience Emerging best practice points to three areas to strengthen gender budgeting implementation: (i) ensuring that gender-sensitive programs can be identified and spending can be tracked, both if the gender impact is complete or partial; (ii) improving and developing the use of performance indicators to track and monitor program performance against specific outcomes; and (iii) developing granular ex-post analysis, which requires strong analytical capacity and resources. Below is a summary of experiences from OECD countries: • Austria’s gender budgeting was introduced as part of the PBB framework codified in the budget law. Each chapter within the Annual Budget Statement should have outcome objectives, with at least one objective directly addressing gender equality (for example, higher female participation in the labor market, improvement of the state of health of males above 50, and reduction of the gender pay gap). • In Canada, an ex-ante gender impact assessment is applied to assess the potential impact of policies based on gender and other identity factors to develop more inclusive budget measures. Programs are tagged to the pillars of the Gender Results Framework when relevant to gender equality objectives. • In France, the Gender Budget Statement assesses fiscal policies and the budget from a gender perspective. It presents the programs that contribute to gender equality with their budgets and performance indicators. • Italy has coded public expenditure using three digits, 0, 1, and 2, where 0 denotes gender- neutrality, 1 implies gender-sensitivity (i.e., expenditures which have a different impact on men and women), and 2 is used for those expenditures whose goal is reducing gender inequality. • In Iceland, ministries are required to undertake gender analysis of budget proposals, as well as legislative proposals. The budget bill submitted to parliament should also outline the impact of changes in revenue and expenditure policies on gender equality targets. • In Sweden, an annual appendix to the budget bill, entitled Economic Equality between Women and Men, is published, showing the distribution of economic resources between the sexes. Source: (Curristine, et al. 2021) and (Downes, Trapp and Nicol 2017). 6.7 Recommendations and Reform Options Strengthening the performance orientation of the PFM will require improving the quality of spending and PM&E. A multi-year strategy to improve the spending quality could be developed, focusing on building capacity in both the central government and line ministries to prepare and appraise programs and establish performance accountability processes. A single, integrated new high-level PM&E function and system could be designed to track performance on priority outcomes and used (both the tool and its findings) to inform policy-making at the highest levels. This includes both strengthening monitoring and the development of a strong M&E system, including both spending reviews and completion reviews for key projects funded from central or local budgets with comparisons of actual and planned costs and the measurement of outputs and independent ex-post impact evaluations of large and strategic government projects and programs, with lessons learned widely disseminated. Ex-ante gender impact assessments, as well as ex-post program evaluations applying a gender lens, could also be implemented to support the achievement of gender equity goals. Enhancing institutional, organizational, managerial, and technical skills in budgeting and PM&E capacities will ensure conceptualization, design, and use of the tools and findings. This might benefit from 150 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH a broader change-management strategy to facilitate a cultural change from compliance- to performance- oriented public sector management in the MoF, MNE, ASPR, and across MDAs. Reinforcing the linkages between policy instruments and strategic objectives, strategic planning, PBB, and PM&E will be crucial for a greater results-focus considering the NDP 2025. This requires harmonization of the planning and budget architecture, the PM&E framework, and organizational structure by improving business processes and using common program coding and a consistent planning and budget classification structure to fully implement PBB. This also includes strengthening the capacity of strategic planning and budget units in line ministries and SNGs to set objectives and select appropriate indicators to facilitate the assessment of public spending effectiveness and build on past efforts to introduce PBB to create stronger links between resource allocation, outputs, and outcomes. Developing budget-tagged expenditures and outputs for measuring results achieved under National Priorities and thematic programs could also enhance results focus. Finally, strengthening the medium-term perspective in planning and budgeting through the MTEF would foster better medium-term planning, providing MGAs with greater certainty of resource availability over a multi-year period, which is crucial for policies that require an extended time horizon for implementation. Improving the transparency of the budget process and, more broadly, public sector transparency will require further work on implementing the Law on Public Audit and Financial Control, as well as greater transparency and citizens’ participation in PFM decision-making. On the one hand, a clear separation of responsibilities is needed between the Accounts Committee and the CIPA. Specifically, the Law on External Audit (SAI) could be a separate one, as that covers the external audit of the public sector and is regulated by a completely different set of standards and methodologies (INTOSAI/ISSAI). Internal audit can be potentially combined with public internal (financial) control in a single Law on Public Audit and Financial Control, in terms of the system and processes, but not in terms of “financial control,” as their respective functions for financial audit and control are completely different in their objectives, roles, key performance indicators, external assessments, standards, methodologies and tools applied. On the other hand, there is a need to facilitate greater transparency and citizens’ participation in fiscal issues and decision-making. This can be done by ensuring the availability of comprehensive budget information in accessible and user-friendly formats, such as online portals and citizens’ budgets, and by helping citizens, civil society, and the media to monitor performance. It can also include expanding mechanisms during budget formulation and implementation to engage civil society organizations and citizens, including actively engaging vulnerable and underrepresented communities. This also includes capacity-strengthening activities for Public Councils, related NGOs, and the public, including budget literacy; public awareness campaigns; modernized and more user-friendly public consultation processes, deepening the participatory budget initiatives; and a greater disposition to consider and incorporate citizens’ feedback into policy-making decisions. Participatory budgeting can be further strengthened, leveraging pilot initiatives, through the design of a special program aligned with PBB and budget transparency initiatives, with specific resource allocation to ensure implementation and capacity-building activities for responsible MDAs at central and local levels, NGOs, the expert community, and the public in general. Promoting green public sector performance will require a strengthened and coordinated institutional framework and specific actions that can be implemented throughout the budget cycle. High-level leadership is needed to drive reforms, such as from the Presidential Administration, coupled with the necessary policy and funding support from MNE and MoF. MENR could play a special role as the authorized government entity for developing and implementing environmental policy, including protection, control, and oversight over the rational use of natural resources. ASPR can facilitate an integrated approach and help mainstream green, resilient, and inclusive development planning across sectors in and beyond government. The Accounts Committee, as an organ exercising public oversight, could be incorporated into the reform agenda and help build its internal program and capacity in auditing activities, strategic plans, and budget programs linked to climate change reforms. 6. BUDGETING, PLANNING, MONITORING: INSTITUTIONAL FOUNDATIONS FOR THE PUBLIC SECTOR 151 Different entry points can be used to include green public sector performance actions in the budget cycle. In addition, different entry points can be leveraged to incorporate green public sector performance actions throughout the budget cycle, including: (i) climate-informed fiscal risk assessment; (ii) climate-informed long/medium-term macro-fiscal framework; (iii) budget guidance on the integration of climate change policy considerations in the preparation of agency expenditure plans and budgets; (iv) the Government’s short and medium-term financing strategy, which includes disaster risks and financing, to improve preparedness for extreme climate events; (v) climate-related expenditure disclosed in budget documents and execution reports; and (vi) green bonds and/or dedicated climate change financing to support the Government’s climate change strategy. Their implementation will require developing guidelines for line ministries to support the development and implementation of climate change actions into their budget  submissions, including regulations and technical instructions. Finally, the Government can facilitate broad public access to comprehensive, timely, and useful information on all climate-relevant revenues and expenditures, along with new opportunities for participation in decision-making and oversight. Table 6‑7. Summary of recommendations Challenges Recommendations Timeframe 1. Performance orientation of PFM • Strengthen PFM focus on • Multi-year strategy to improve the quality of • Short term outcomes and results spending with a focus on capacity building • Strengthen PM&E function • Single new high-level PM&E, with a focus • Improve institutional on evaluations and their use in the budget • Medium-term capacity preparation process • Continuous • Change management strategy 2. Linkages between strategic planning and budgeting • Implementation of PBB and • Harmonize and streamline the planning, • Medium-term its linkages with SP budgeting, and PM&E architecture and • Medium-term • MTEF and strategic resource business processes, also clarifying roles and allocation responsibilities and the organizational structure • Strengthen the MTEF process to allow line ministries to undertake better medium-term planning 3. Transparency of Public Finance • Separate internal and • Further work on the implementation of the Law • Short-term external audit functions on Public Audit and Financial Control • Facilitate and promote • Ensure availability of timely and user- participation friendly budget information and strengthen • Medium-term • Improve the capacity of Public mechanisms for participation • Continuous Councils • Improve transparency in the allocation of budget funds • Capacity-building activities for Public Councils, related NGOs, and the general public • Continuous 4. Green public sector performance • Integrate institutional and • Strengthening institutional framework for • Medium to policy framework climate change long term • Leverage entry points for • Incorporate green public sector performance implementing climate change actions throughout the budget cycle across the budget cycle • Medium to long term 152 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 6. BUDGETING, PLANNING, MONITORING: INSTITUTIONAL FOUNDATIONS FOR THE PUBLIC SECTOR 153 7. Improving Intergovernmental Fiscal Relations 154 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH KEY POINTS • Despite the centralized political structure, subnational governments (SNGs) are already important providers of public services in Kazakhstan, accounting for about 40 percent of total government spending. Although their political autonomy is limited—particularly at the higher tiers—they have significant discretion over allocating their budgets. Moreover, this is desirable given the country’s large territory and dispersed population. In principle, subnational autonomy can help local governments improve the public sector and respond faster to local development priorities. • To support the functioning of SNGs, the intergovernmental fiscal system must allow SNGs to have the incentives and resources to deliver key services and develop their regions. The current system for financing SNGs relies heavily on transfers. This is sensible, given the wide variations in revenue-generating capacity among individual SNGs. But the current system for allocating transfers is extremely complicated, attempting (in vain) to equalize the level and quality of services nationwide. Moreover, the system for financing lower tiers of SNGs is not clearly spelled out, leaving raions, okrugs, and villages subject to the whims of their respective oblast governments. • The Government should simplify the system for determining the level of central- government transfers to first-tier SNGs and establish clear rules for allocating transfers from oblasts to subordinate levels of SNGs. At the same time, it should improve the clarity in the system for the financing of lower tiers of SNGs (rayons, okrugs) to reduce negotiation over resources with their respective oblasts and improve the revenue base for SNGs from existing revenue sources. 7.1 Structure of Subnational Governments in Kazakhstan Strengthening the capacity of subnational governments (SNGs) to address local development priorities is a key policy agenda. Like most countries, SNGs are at the frontline in Kazakhstan to engage with the population and deliver key public services. The central government can develop and control the implemen- tation of public programs with a high degree of homogeneity and a strong spillover effect across locations (e.g., national highway transport, communication, pandemic control, etc.). However, the central government may be unable to deliver public programs to respond to specific local needs so effectively.195 Improving the capacity of SNGs to deliver public programs that can effectively respond to local development needs there- fore becomes a critical item on the policy agenda. Kazakhstan is a unitary state that has three tiers of subnational administration. The top tier consists of cities of republican significance, namely, the capital city Astana, the cities of Almaty and Shymkent, and 14 regions (oblasts),196 whose territory is divided into cities of oblast subordination and districts (raions) (Figure 7‑1Figure 7‑1. Territorial structure of subnational governments in Kazakhstan 195 Bardhan (2002) explained that the agency and transaction costs for central government to design and deliver public services suitable for local conditions are likely to be high. SNGs have better local information and ability to customize public programs to local conditions. But more importantly, accountability to local electorates is a crucial factor to ensure an effective delivery of local public services. 196 As of 2022, the number of oblasts has been increased by three. 7. IMPROVING INTERGOVERNMENTAL FISCAL RELATIONS 155 ). The territory of each raion is, in turn, divided into cities of raion significance and rural districts (okrugs).197 In addition, some larger villages (auls) have the de-facto status of okrugs because they have their own executives. (Of the roughly 6,500 recognized villages, only 2,300 have this characteristic.) These ‘independent villages’ are subordinate to their respective raions rather than the okrugs surrounding them. The rest of the ‘non-independent’ villages are directly governed by their respective rural district governments (okrugs) and have no separate administrative structure. Figure 7‑1. Territorial structure of subnational governments in Kazakhstan Despite their important role in delivering local services, the political structure of Kazakhstan’s SNGs remains highly centralized. SNGs account for roughly 40 percent of total government spending. They deliv- er many public services, including primary and secondary education, regional and local transportation, and public utilities. But the political structure of SNGs remains highly centralized. • The executives (aims) of the highest tier of SNGs are appointed and dismissed by the President with the consent of parliament. In turn, the akims of the second tier are appointed by those Presidential appointees. Only akims of the third tier of local government are directly elected—and this remains an ongoing develop- ment since 2021 that should be complete by 2025. • Candidates for executive (akim) positions in the second tier (cities of oblast significance and raions) are proposed by their respective oblast akims and elected (usually without opposition) by the raion maslikhat. The direct election of akims is limited only to the third tier of SNGs. Before 2021, candidates for akim positions in the third tier of SNGs were proposed by their respective raion akims and elected (again, usually without opposition) by their respective assemblies. Beginning in 2021, those akims began to be directly elected. Akims appointed under the former system can complete their terms but will be replaced by directly elected akims once their terms expire. This process is expected to be completed by 2025. Table 7‑1. The political structure of SNGs according to the Local Government Law 197 The Russian terms ‘rayon’ and ‘okrug’ are both translated as ‘district’ in English. For purposes of clarity, the term okrug will be used to refer only to rural districts in this report. 156 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Representative Type of Jurisdiction Executive Body Astana, Almaty City, Shym- Indirectly elected: candidate proposed by Presi- 1st tier Maslikhat elected kent, oblasts dent, subject to approval by parliament Cities of oblast significance Indirectly elected: candidate proposed by oblast 2nd tier Maslikhat elected and raions akim, subject to approval by raion maslikhat Cities of raion significance, 3rd tier Directly elected (to be completed by 2025) okrugs, large villages 7.1.1 The legal foundation for the division of functions across subnational governments From Kazakhstan’s independence at the end of 1991 until 1998, the division of functions between the tiers of government largely reflected the practices of the previous Soviet era. Functional responsibilities were not formalized and were often subject to negotiation. However, since the beginning of 1998, attempts have been made to distinguish the functional responsibilities of the various tiers of government through legislation. But overlaps and ambiguities remain, both de jure and de facto. Two main laws govern the assignment of functions between the three tiers of government. The Law on Local Government and Self-Government in the Republic of Kazakhstan (Adilet legal information system n.d.) and the Budget Code. The Law on Local Government sets out the responsibilities of the first, second, and third tiers of SNGs in three articles (namely, Articles 27, 31, and 35, respectively). Each of these articles sets out a long list of the functions assigned to its respective tier: • Article 27 refers to the functions of first-tier SNGs (oblasts and cities of Republic significance), listing 37 such functions (not counting various subfunctions). In the social sectors, these include ‘ensuring the realization of the right to free education,’ ‘ensuring the right to guaranteed free medical care (in the context of the system of compulsory social health insurance),’ and ‘coordinating social assistance to vulnerable so- cial groups.’ In the infrastructure sectors, the functions listed include constructing and maintaining roads, organizing public transport, and ‘organizing the construction of water pipes and heating and electrical networks.’ They also include repairing facades, roofs, and the common areas of apartment buildings and collecting and disposing of solid waste. • Article 31 lists the functions assigned to the second tier of SNGs: raions and cities of oblast signifi- cance. It requires second-tier SNGs to ‘ensure the realization of the right to free education at the primary, basic secondary, and general secondary levels’ and to ‘render assistance to vulnerable social groups.’ (Health care is not mentioned.) In the infrastructure sectors, SNGs at this tier are assigned responsibility for constructing and maintaining roads of raion (and cities of oblast) significance, organizing public trans- port, and organizing the construction of water pipes and heating and electrical networks. In addition, their responsibilities include repairing the roofs, facades, and common areas of apartment buildings and solid waste management. • Article 35 sets out the functional responsibilities of third-tier SNGs—okrugs, villages, and cities of raion significance. In the social sectors, these include providing preschool education, organizing free transporta- tion of pupils to school, identifying low-income households to determine eligibility for social assistance pro- vided by higher levels of government, and coordinating charitable assistance to vulnerable social groups. In the health sector, third-tier SNGs are responsible only for rendering emergency care and delivering patients to the nearest public hospital. In the infrastructure sectors, they are responsible for constructing and main- 7. IMPROVING INTERGOVERNMENTAL FISCAL RELATIONS 157 taining roads of okrug, village, and cities of raion significance and (once again) organizing the construction of water pipes and heating and electrical networks. In addition, they are assigned responsibility for lighting, landscaping, and ‘sanitary cleaning’ (which may mean solid waste management or street cleaning). The Budget Code sets out a slightly different and more detailed division of functions between levels of government (see Box 7‑1). Box 7‑1. The division of functions according to the Budget Code A separate division of functions is set out in the Budget Code (as amended). As summarized in Annex 3, the Budget Code sets out the respective responsibilities of the central government, the three cities of Republic significance, the oblasts, and the two lower levels of SNGs. In the case of education, for example, it assigns oblasts only responsibility for technical and vocational education. Raions (and cities of oblast significance) are responsible for preschool and primary and secondary education. Okrugs and villages are responsible for the transportation of students to the nearest rural school. (Primary and secondary education responsibility has since been transferred to the oblast/Republic city level.) In the case of health care, the central government is assigned the principal responsibility in the context of the national healthcare system. Oblasts can provide ‘additional’ (unspecified) free medical care. Raions and okrugs have no role in health care whatsoever other than (in the case of okrugs and villages) transporting critically ill patients to the nearest healthcare facility. In the case of social assistance, the central government is again assigned the primary responsibility through the national social assistance system. Oblasts are responsible for providing social assistance to particularly vulnerable populations: orphans, children left without parental care, the elderly, and the disabled. Raions are responsible only for material support for disabled persons under individual treatment programs. Third-tier SNGs are not assigned any social assistance responsibilities at all. In the case of transportation, the central government is assigned responsibility for constructing and maintaining roads of national and international importance. Oblasts are assigned responsibility for roads of oblast significance, raions for roads of raion significance, and okrugs for roads of oblast significance. The central government is assigned exclusive responsibility for the rail system, while oblasts and raions organize other forms of passenger transport. In the case of urban infrastructure (water supply, sewerage, drainage, district heating, solid waste management), the raions (along with the three Republic-level cities and the cities of oblast significance) bear primary responsibility. Okrugs and villages are responsible only for solid waste management and public lighting. Overall, the picture of the division of functional responsibilities that emerges from the current legisla- tion is one in which: 1. The central (Republican) government is the primary provider of national defense, health care, and social assistance; 2. Oblasts and cities of Republican significance are the primary providers of primary and secondary edu- cation; 3. Raions (along with the three Republic-level cities and the cities of oblast significance) are responsible for local public services (water supply, sewerage, drainage, district heating, solid waste management); and 4. Each tier of government is responsible for the roads under its jurisdiction. 158 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 7.1.2 Functional division and spending by subnational governments Data on spending by each tier of government suggest that the division of functions between the central government and SNGs (as a group) is fairly clear-cut.198 As shown in Figure 56, health care, social assistance, and education are by far the largest functional categories of public expenditure. Among those three, the central government accounts for most of the spending on social assistance and health care. On the other hand, spending on education is largely the domain of SNGs. Figure 7‑2. (Central government spending on education is largely limited to the tertiary level.) Spending on housing and communal services (including urban water supply and district heating) is entirely the domain of SNGs. Spending on transport is evenly split, as one might expect. Central government spending on transport includes spending on the national highway network, as well as rail and air transport. SNG spending on transport largely consists of the oblast, city, raion, and village-level roads. Figure 7‑2. Division of spending responsibilities between the central government and SNGs (level of spending by level of govt, KZT trillion, 2021) Source: World Bank staff calculations based on data published by the authorities. The functional division of spending between levels of SNGs is more difficult to document. Aggregate data suggest that oblast governments account for about half of SNG spending within their territories. Recent data from the Treasury indicates the scale of total spending by each tier of SNG in each oblast. As shown in Figure 7‑3, first-tier SNGs (i.e., oblast governments) account for about half of the total SNG expenditures within their territories, ranging from 30 percent in Pavlodar to 56 percent in Mangistau.199 Second-tier SNGs (raions and cities of oblast significance) account for nearly all the remainder. The spending of third-tier SNGs—okrugs, villages, and cities of raion significance—is extremely small. As shown in Figure 57, third-tier SNGs account for an average of 3 percent of total SNG expenditures in Kazakhstan’s 14 oblasts, ranging from 1.5 percent in Turkestan to 4.3 percent in Kzyl-Orda. 198 Face-to-face interviews with representatives of SNG officials were conducted in Astana, Aktau, and Pavlodar, while an online meeting was organized with officials from Almaty oblast. 199 In the cities of Astana, Almaty, and Shymkent, all three tiers of SNG are combined. 7. IMPROVING INTERGOVERNMENTAL FISCAL RELATIONS 159 Figure 7‑3. Spending by each tier of sub-national government (% of spending by each tier, by oblast, 2021) Source: World Bank staff calculations based on data published by the authorities. A case study from Mangistau oblast provides a more precise picture of the division of spending between the oblast government and lower levels of SNGs. As shown, education accounts for nearly 60 percent of the oblast government’s total expenditures (Figure 7‑4). The remaining 40 percent is divided between a wide range of functions, none of which accounts for more than 10 percent of the total. (Transport accounts for 9 percent of the total, agriculture and environmental protection accounts for 8 percent, and culture for 6 percent). As for the second and third-tier SNGs within Mangistau oblast,200 housing and communal services (utilities) account for half of the total spending in each tier (Figure 7‑5). Social assistance accounts for another 17 percent. No other sector accounts for more than 10 percent. SNGs at the second and third tiers of SNG spend nothing at all on education or health. These charts suggest that the oblast government is exclusively responsible for education (and health, to the extent that it is not financed directly by the central government). Lower levels of SNG are largely responsible for housing and communal services (and social assistance to the extent that it is not financed directly by the central government).201 Further analysis would be required to determine whether this conclusion applies to other oblasts. While the Budget Code assigns a range of taxes to subnational governments, they have virtually no control over their yields. The central government sets tax rates and exemption policies. In addition, SNGs have little control over the quality of tax administration. All SNG taxes are administered by the State Revenue Committee (SRC), an agency of the central government, and then transferred to SNGs. The SRC has a department in each oblast and city of Republican significance. With its subordinate customs units and tax divisions, the SRC is responsible for administering every tax, fee, mandatory payment, and customs duty collected in Kazakhstan, even if it is a small fee for a harvesting license. 200 These consist of the cities of Aktau and Zhana-Ozen and five raions. Separate data on the spending of third-tier SNGs within each raion is not available. 201 Both charts exclude subnational spending on transfers. This category of functional expenditures accounts for a significant level of spending in some of Mangistau’s raions, with the proportion ranging from 26 percent in Tupkaraganskii raion to 67 percent in Karakiyanskii raion. Unfortunately, it is not clear what ‘transfers’ means in this context. Under Kazakhstan’s budget classification system, ‘transfers to other levels of government’ (functional classification 330) include both transfers to higher levels of government (i.e., withdrawals) and transfers to lower levels of government, i.e., to raions, okrugs, and villages. The published data do not provide further details. 160 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Figure 7‑4. Functional distribution of Figure 7‑5. Functional distribution of expenditures by Mangistau oblast government expenditures by second and third-tier SNGs in Mangistau oblast Source: World Bank staff calculations based on data published by the authorities. 7.2 Revenue Assignment for Subnational Governments Oblast (first-tier) SNGs have the authority to Figure 7‑6. Sources of subnational determine the proportion of these taxes to be governments’ revenues, 2021 shared with subordinate SNGs. Article 52 of the Budget Code authorizes raions and cities of oblast significance (second-tier SNGs) to retain a share of PIT, social tax, and CIT raised within their jurisdictions. But the proportion is to be determined by the oblast maslikhat except in the case of PIT on individuals ‘who have declared in the territory of a city of raion significance, village, or okrug.’ Those revenues are assigned exclusively to these third- tier SNGs. Second-tier SNGs are also assigned the proceeds of the property tax and the vehicle tax, except for taxes on property and vehicles located and registered in third-tier jurisdictions. Second-tier SNGs are also assigned the proceeds from excise taxes on alcohol, tobacco, and gasoline. Source: World Bank staff calculations based on data published by the authorities. The assignment of revenues to third-tier SNGs follows from the exceptions noted above. Cities of raion significance, okrugs, and independent villages are assigned the proceeds of PIT, property tax, and vehicle taxes collected within their jurisdictions. This arrangement is summarized in Table 7‑1. 7. IMPROVING INTERGOVERNMENTAL FISCAL RELATIONS 161 Table 7‑2. Summary of subnational tax assignments 1st tier 2nd tier 3rd tier Personal income tax except Share of oblast PIT, as determined by ‘Local’ PIT ‘local*.’ oblast Social tax Share of oblast social tax as determined by oblast Corporate income tax on small/ Share of oblast CIT as determined by medium non-oil firms oblast Property and vehicle taxes tax, except ‘Local’ property and ‘local’ vehicle taxes Excise taxes on alcohol, tobacco, and gasoline *The term ‘local’ refers to persons and property registered in a city of raion significance, village, or okrug. 7.3 Transfer Formula and Equalization 7.3.1 Subvention and withdrawal This system of tax assignments results in widely varying levels of tax revenues among the various juris- dictions, in both absolute and per capita terms. This is due to wide variations in the regional levels of economic activity. To offset the resulting disparities, Kazakhstan employs a system of subventions and withdrawals. The amount of the subventions or withdrawals is based on calculating the projected revenues and ‘expenditure needs’ of each first-tier SNG (including the lower tiers of SNGs within it). First-tier SNGs whose projected expenditures exceed their projected revenues are allocated additional funds (subventions) to make up the difference. First-tier SNGs whose revenues are projected to exceed their expenditure needs are subject to withdrawals. This system was introduced in 2004 and has remained in place, with several modifications, ever since. The methodology for calculating projected revenues and expenditure needs is complicated.202 As noted earlier, the most important subnational taxes are PIT and social tax. The actual revenues of those taxes in the previous year are taken as a starting point. Those figures are then adjusted to reflect revenue chang- es in the coming year. These include projected GDP growth and inflation and changes in the number of taxpayers, the minimum wage, and the tax rate. They also include an adjustment for improvements in tax administration. 202 The Ministry of National Economy (MNE) Order No. 139 (2014) set out the methodology to be used over the period 2015–2022. The specific methodology to be used in forecasting revenues is set out in MNE Order No. 34 (2015). 162 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Box 7‑2. Evolution of general transfer formula The system of transfers was first codified in conjunction with Kazakhstan’s first Budget Code in 2004. This established the gap-filling approach: the transfers to first-tier SNGs were based on the difference between each jurisdiction’s projected own-source revenues and estimated expenditure needs. In theory, the Government was committed to filling the gap. If projected expenditure needs exceeded projected own-source revenues, the Government was obligated to make up the entire difference through a transfer. By the same token, if own-source revenues were projected to exceed expenditure needs, the Government was entitled to extract the surplus. Initially, the estimates of each jurisdiction’s own-source revenues and expenditure needs were based on the previous year’s results. This approach resulted in wide disparities in the level of per capita resources among first-tier SNGs. By basing projected revenues and expenditures on the previous year’s results, it froze in place the disparities that existed before the system went into effect. In response, the Government introduced a new methodology for estimating expenditures in 2007. Under the new methodology, the expenditure needs of each jurisdiction were calculated based on the number of clients or beneficiaries of each function assigned to the sub-national level. This was then adjusted to reflect jurisdiction variations in population density, age structure, environmental conditions, and other factors. However, this approach proved unworkable. As a result, as of 2023, Kazakhstan will adopt a new variant combining aspects of the first and second approaches. Actual own-source revenues and expenditures in 2022 (budget) will be the basis for calculating subventions and withdrawals in 2023–2025. Projections for increases in expenditure needs will, however, be based on projections of the number of clients for each service, along with other factors. The methodology for estimating expenditure needs is more complicated.203 This exercise calculates the expenditure needs of each first-tier SNG in each function assigned to the sub-national level under the Budget Code.204 Under the methodology in effect before 2023, the calculation of expenditure needs began by taking the total amount of spending on that function by all SNGs in the previous year and adjusting it for projected inflation and any central government decisions that would affect the cost of delivering that service. Those total costs were then distributed to each jurisdiction according to the number of clients for that function in each jurisdiction, as adjusted for factors that would affect the unit costs of that function in that particular jurisdiction. These included: projected increases in population, variations in the unit cost of providing a function in different regions (such as road maintenance), population density, the proportion of students enrolled in ‘small’ schools, the age structure of the jurisdiction’s population, poverty levels, and the duration of the heating season. The exercise also uses different factors for different functions.205 Given its complexity and the radical redistribution of funds it would represent when first introduced, it is unclear whether this methodology was ever precisely applied. The methodology adopted for the 2023–2025 period makes two major changes. First, the Ministry of Ed- ucation (MoE) will calculate the expenditure needs of each oblast (along with the cities of Astana, Almaty, and Shymkent) in the education sector rather than relying on the formula set out in MoE Order No. 139. The 203 Ministry of the National Economy. On Approval of the Methodology for Calculating Transfers of a General Nature. Minister of National Economy Order No. 139, dated December 11, 2014, as amended through 2022. 204 Articles 54, 55, 56, and 56-1. 205 In the case of primary and secondary education, for example, expenditure needs were based on the number of children aged 6–18 in each jurisdiction, but adjusted to take into account the proportion of students enrolled in small schools, the costs of bonuses paid to teachers in rural areas, and the length of the heating season. In the case of social assistance, expenditure needs were based on the number of children aged 6–18 and the number of people over retirement age, but adjusted to reflect variations in population density, costs of bonuses paid to social workers in rural areas, and the length of the heating season. 7. IMPROVING INTERGOVERNMENTAL FISCAL RELATIONS 163 MoE will calculate education costs in most urban schools based on enrolment and standard costs per stu- dent. The education costs in most rural schools will be based on historical data. Second, the basis for pro- jections of expenditures in other sectors will change. Rather than starting with an estimate of the number of clients for each service and the unit costs of providing that service to those clients, the projections will now be based on actual spending on each function in each oblast. Thus, the projected expenditure needs for road maintenance in Mangistau oblast will be based on the actual level of spending—or, more precisely, the budgeted expenditure on that function in the preceding year in that oblast.206 The spending on road maintenance in Almaty City will be based on that city’s expenditures on that function in the preceding year. 7.3.2 Targeted transfers In addition to general subventions, SNGs are eligible to receive targeted transfers for recurrent or cap- ital (development) spending. Unlike general subventions, targeted transfers are earmarked: they can only be spent for the particular purpose designated in their respective enabling legislation. Targeted recurrent transfers aim to compensate SNGs for additional expenditures imposed by central government decisions. In 2021, one of the largest targeted recurrent transfers compensated SNGs for the costs imposed by a government-decreed increase in rural teachers’ salaries. • Targeted recurrent transfers are intended to be temporary. Suppose a central government decision implies a permanent increase in the cost of a subnational service. In that case, this is to be reflected in the expenditure calculation when subventions and withdrawals are recalculated. Thus, the targeted recurrent transfer that compensated SNGs for the increase in rural teachers’ salaries will expire at the end of 2022. Those costs will instead be reflected in the MoE’s calculation of each oblast’s expenditure needs for the period 2023–2025. • Targeted development transfers (TDTs) are, as noted above, intended to finance capital works. They are allocated on a project-by-project basis. The procedures for the identification, preparation, appraisal, and funding of projects financed from TDTs are specified in legislation, including Article 46 of the Budget Code and MoF Order No. 126207 (2015), as amended. Overall, the system of transfers and withdrawals in effect in 2021 appears to have reduced disparities in per capita resources among first-tier SNGs. Figure 7‑7 illustrates the level variations in per capita own-source revenues among first-tier SNGs in 2021. As shown, these revenues are far higher in three jurisdictions—Almaty City, Astana, and oil-rich Atyrau—than in any other first-tier jurisdiction, largely on the strength of their revenues from PIT, CIT, and social tax and, in the case of Atyrau, other taxes. Per capita, own-source revenues in the Turkestan oblast were only KZT 54,000 (US$127) in 2021. In Atyrau oblast, they were more than 10 times higher (KZT 657,000, or US$1,542). These disparities are offset to a great degree by the system of subventions and withdrawals. Figure 7‑8 illustrates the variation in per capita recurrent revenues among first-tier jurisdictions once transfers and withdrawals are considered. As shown, net revenues ranged from KZT 375,000 (US$880) in Mangistau oblast to KZT 633,000 (US$1,485) in North Kazakhstan. Most first-tier SNGs fall within a much narrower range: with the exceptions of Akmola, Astana, and North Kazakhstan, they fall within a range of KZT 375,000 and KZT 466,000. 206 See ‘On Amendments to the Order of the Minister of National Economy of the Republic of Kazakhstan dated December 11, 2014 No. 139 “On Approval of the Methodology for Calculating General Transfers.’ 207 ‘About Approval of Rules of Consideration and Selection of Target Transfers on Development.’ 164 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Figure 7‑7. Variations in SNG own-source revenues per capita (KZT ’000, 2021) Source: World Bank staff calculations based on data published by the authorities. Figure 7‑8. Variations in SNG total net revenues per capita (KZT ’000, 2021) Source: World Bank staff calculations based on data published by the authorities. 7. IMPROVING INTERGOVERNMENTAL FISCAL RELATIONS 165 Box 7‑3. Calculating per capita SNG revenues Calculating the level of SNG revenues and expenditure needs in Kazakhstan is complicated. The Statistical Bulletin publishes summary data on subnational revenues and expenditures by oblast. The revenue data, however, are reported gross of withdrawals, i.e., before withdrawals from the richer jurisdictions are subtracted. As a result, the published figures overstate the revenues available to these jurisdictions. Withdrawals are instead reported as expenditures under the classification ‘transfers.’ But the figures for transfers (budget functional classification 330) include both transfers from lower levels of government to higher ones (e.g., withdrawals) and transfers from higher levels of government to lower ones (e.g., transfers from an oblast to raions and cities of oblast significance). Thus, subtracting ‘transfers’ from total revenues overstates the extent of withdrawals. For purposes of Figure 7‑8, net revenues are calculated as total recurrent revenues minus withdrawals as reported in a separate column of the Bulletin. The remaining disparities may be due to several factors. It is possible that the various factors used to calculate expenditure needs in each jurisdiction really do yield unusually high figures for those at the upper end of the range (Akmola, Astana, and North Kazakhstan) and lower figures for oblasts such as Mangistau and Almaty at the lower end. If the cost of living or salaries were a major factor in the calculations, it could lead to this result. Second, the level of net revenues may be affected by variations in targeted transfers— both recurrent and capital. This may explain the high net per capita revenues in North Kazakhstan, for example. Due to data constraints, the impact of each type of transfer on net per capita revenues cannot be determined.208 Another possible explanation is that the official formula acts only as a guideline—a starting place for negotiations between the MNE and each of the oblast akims—and that the ultimate result reflects the outcome of these negotiations rather than the formula itself. 7.3.3 Shortcomings of the existing system of transfers equalization Despite its success in reducing disparities in net per capita revenues among first-tier SNGs, the current system of tax assignments and withdrawals/subventions has several shortcomings. Excessive complexity with little value added The methodology for calculating subventions and withdrawals is extremely complex, undermining con- fidence—at least on the part of akims, if not the general public—that it is being applied fairly. The methodology for calculating expenditure needs, as set out in Order No. 139, is one of the most complicated in the world. As described above, in the case of social assistance, expenditure needs are based on the number of children aged 6–18 and the number of people over retirement age, but then adjusted to reflect variations in population density, costs of bonuses paid to social workers in rural areas, and the length of the heating season. The Order does not specify how these factors will be incorporated into the calculation—e.g., how much weight is to be assigned to population density as opposed to the length of the heating season. This is presumably set out in yet more orders. It is not clear that this complexity accomplishes much. Social assistance, for example, accounts for only 5 percent of SNG expenditures. (As noted earlier, the central government accounts for 90 percent of spending 208 In 2021, some oblasts also received large loans from the central budget. These are not reflected in the figures shown in Figure 6. 166 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH on this function.) It is not obvious that a detailed methodology is required to calculate expenditure needs for so small an expenditure category. This is particularly true now that the MoE will calculate expenditure needs in the education sector. As noted earlier, education accounts for 40 percent of subnational expenditure, leaving the formula to calculate expenditure needs in various functions, none of which constitutes more than 15 percent of SNG expenditures. The objective of the system of subnational finances should be more modest. Rather than attempting to equalize outcomes, it should ensure that all jurisdictions have the minimum resources to perform the assigned functions. It is not obvious that equalization of service outcomes is a feasible or desirable objective. Given the diversity of circumstances in Kazakhstan, an equal level of service outcomes would be very difficult to achieve. Raising the quality of local services in every village in the country to the levels prevailing in Astana or Almaty City would require massive capital investment in water supply, sewerage, road paving, and sanitary landfills. As Kazakhstan urbanizes, many of these villages are losing population. It could be argued that such funds would be better spent on expanding infrastructure in places that people are migrating to, not where they are migrating from. One can certainly argue that every child in Kazakhstan deserves a basic level of education. Since ed- ucation is financed through first-tier SNG budgets, the sub-national finance system should ensure this is achieved. But even in this case, ensuring equal service levels would be difficult to accomplish. Education is subject to economies of scale: costs per student fall as classroom size increases. The number of students in urban areas is large enough to permit separate classes for each grade. In large cities, there are even enough eligible students to fill specialized classes at the high school level (e.g., calculus). But the number of students in villages is too small to fill such classes. As a result, education must be provided through multi- grade classes (one-room schoolhouses) by transporting students over long distances or relying on online remote learning. There is yet another justification for inequality. It is not unreasonable to expect that those jurisdictions in which residents pay higher levels of taxes should have higher-quality services. People who pay high PIT should expect to have better services than those who do not. This is recognized throughout the developed world. No OECD country attempts to equalize levels of per capita spending locally—let alone outcomes. Disincentives for local economic development The system discourages subnational efforts to develop their economies. It is true that a growing econ- omy would lead to increased revenues from PIT and social tax and, up to a point, CIT.209 But, under the present methodology, this would automatically lead to reduced subventions or increased withdrawals. As a result, SNGs would derive no financial benefit from attracting new investment. Revenue precariousness at the second and third tiers of subnational governments The arrangement for financing second and third-tier SNGs leaves them vulnerable to the whims of oblast governments. While SNGs at this level are assigned several own-source revenues (as noted earlier), the yields of these revenues appear to be very small. Instead, SNGs at the second tier are dependent upon shares of PIT, social tax, and CIT that is collected within their territories and on subventions from their re- spective oblasts. The shares of these taxes and the level of subventions are both set at the discretion of the oblasts. Third-tier SNGs, for their part, are largely dependent on shares of the property tax and vehicle taxes 209 Since SNGs receive only the CIT on SMEs, successful firms eventually graduate out of the local tax base; once they pass the size threshold, their CIT goes straight to the Republic budget. 7. IMPROVING INTERGOVERNMENTAL FISCAL RELATIONS 167 that are collected within their territories and subventions from their respective raions. Again, the levels of both sources are set at the discretion of their respective raions. 7.4 Possible Reform Directions and Examples from Other Countries To address these shortcomings, the Government could consider adopting a simpler subnational finance system while guaranteeing a stable source of revenues to the two lower tiers of SNG. One good option is to focus on reducing, but not eliminating—disparities in per capita resources—rather than attempting to close the gap between each jurisdiction’s expenditure needs and its own-source revenues. By focusing on a single, readily measurable indicator, such a system would be simpler and easier to understand. It would still ensure that all jurisdictions have at least the minimum resources required for their functional obligations. As long as the equalization formula allowed some variation in after-transfer per capita revenues, it would also allow SNGs that succeed in expanding their tax bases to benefit from that effort. Germany’s example The principal regional (Lander) revenue sources in Germany are PIT and VAT. Revenues from PIT are divided three ways: 42.5 percent to the federal government, 42.5 percent to the Lander, and 15 percent to the municipalities (gemeinden). The shares assigned to the Lander and gemeinden are distributed based on origin (i.e., according to where they were collected). Revenues from VAT are also divided among the three levels: 49.8 percent to the federal government, 47.2 percent to the Lander, and 3.2 percent to the gemeinden.210 But the Landers’ share of VAT is not distributed based on origin. Instead, 75 percent of the Landers’ share is distributed based on population. Box 7‑4. Determining the total amount of funds to be transferred All transfer systems confront two basic design questions: (i) how to determine the total amount to be transferred; and (ii) how to distribute that amount among individual jurisdictions. In Kazakhstan, the current system of transfers and subventions addresses these two issues simultaneously, at least in theory. Once the revenue gap—the difference between projected revenues and projected expenditure needs—is calculated for each jurisdiction, the Republic budget is, in theory, obligated to fill it—regardless of the cost. If this were implemented in practice, it would represent an open-ended commitment on the part of the Republic—and one that could become increasingly expensive. An alternative would be to set the amount of the equalization transfer as a separate exercise. The amount could be fixed in the central government’s annual budget exercise. Many countries, however, prefer to reduce the potential for unpredictable year-to-year changes by tying the transfer level to an indicator of overall resource availability, such as a fixed percentage of GDP or central government tax revenues. Some countries tie the equalization transfer amount to a specific central government tax yield, such as VAT. However, this last approach has disadvantages as it leaves SNGs vulnerable to fluctuations in a single tax base. The remainder of the Landers’ share of VAT is targeted exclusively to bring up per-capita revenues of the poorer Landers, specifically, Landers whose per capita revenues from PIT (as well as minor business CIT is also divided between the federal government and the Landers, but is not a major source of revenue for the latter. 210 168 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH taxes) fall below the national average.211 A statistical program is used to determine how much the revenues of poorer Landers will be topped up. In no cases are the per capita revenues of poorer jurisdictions brought up to 100 percent of the national average. For example, in 2020, a Lander whose per capita revenues were equal to 70 percent of the national average before equalization would be brought up to 91 percent of the national average after equalization. The German system also embodies a small ‘fraternal’ equalization element where Landers with above-av- erage per capita revenues must contribute part of their revenues to poorer jurisdictions. A statistical program determines how much the richer Landers must contribute. Under the formula in effect in 2020, a Lander whose per capita revenues were 130 percent of the national average would be brought down to 109 percent of it. The Russian Federation’s example The principal taxes assigned to regional governments (subjects of the Federation) are PIT and CIT. All PIT revenues are transferred to SNGs, with regional governments (and cities of Republic significance) retain- ing 85 percent of PIT generated in their jurisdictions. Cities of oblast subordination retain the remaining 15 percent of the tax collected in their jurisdictions. PIT revenues are divided between the raion government and second-tier municipalities in raions. Ninety percent of CIT is assigned to regional governments, with the federal government retaining the remainder. In principle, CIT is distributed based on origin, although this involves a complicated arraignment to determine where that origin is (see Box 19). As in Germany, the Russian Federation’s equalization transfer is designed to raise poorer regions’ per capita budget revenues (those with per capita revenues below the national average) to a target na- tional average percentage. Calculating the equalization target excludes the 10 richest and the 10 poorest regions. Adjustments are also made to reflect variations in the strength of tax bases among different re- gions, as well as differences in factors that affect the costs of providing services (for example, labor costs, living costs, and population density). As in Kazakhstan, the total amount of the transfer is determined en- dogenously (i.e., the federal government is required to contribute whatever sum is needed to achieve the equalization target). Box 7‑5. Determining the ‘origin’ of CIT in the Russian Federation Until the mid-2010s, the proceeds of CIT were retained in the jurisdiction where they were collected. Consequently, the City of Moscow retained a disproportionate share of CIT, where corporate headquarters tended to be. Currently, CIT paid by vertically integrated companies with operations in more than one region is distributed among the regions where the company does business according to the value of the company’s assets and employees’ salaries in each region. This has resulted in a reduction in Moscow’s share of CIT and a corresponding increase in the shares of other regional governments. Even so, CIT revenues are still concentrated in Moscow. 211 See: The Federal Financial Equalization System in Germany (Bundesministerium der Finanzen 2018). 7. IMPROVING INTERGOVERNMENTAL FISCAL RELATIONS 169 Poland’s example The principal sources of subnational own-source revenues in Poland are PIT and the property tax.212 PIT is shared between the central government and the three tiers of SNGs (see Box 7‑6), with the central government taking roughly 50 percent, the lowest tier of SNGs (gmina) 37 percent, the powiat 10 percent, and the voivodships about 1.5 percent. The property tax is administered and retained by the gmina. As in Germany and the Russian Federation, Poland’s equalization transfer aims to reduce disparities in per capita revenues. The equalization transfer is allocated to all SNGs whose per capita revenues from shared taxes and local taxes (in the case of gminas and miastos) are lower than the national average.213 Suppose a gmina’s per capita revenues are between 40 and 75 percent of the national average. In that case, the transfer (per capita) equals only 75 percent of the difference between the gmina’s per capita revenues and 83 percent of the national average. If the gap is larger, the proportion of it that the grant fills is proportionately larger. The formula for higher tiers of SNGs is simpler: all powiats with per capita revenues less than the national average receive a transfer equal to 90 percent of the difference. All voivodships with per capita revenues less than the national average receive a transfer equal to 72 percent of the difference. In Poland, central government funding for equalization transfers is supplemented by contributions from richer jurisdictions. The formula for calculating each jurisdiction’s contribution to the equalization mechanism is complicated. For each inhabitant, gminas with per capita tax revenues of between 150 percent and 200 percent of the national average must contribute 20 percent of the difference between their own per capita revenues and 150 percent of the average per capita revenues of all gminas. Gminas with per capita tax revenues of between 200 and 300 percent of the national average must contribute, for each inhabitant, 25 percent of the difference between their own per capita revenues and 200 percent of the average per capita revenues of all gminas, plus 10 percent of the average per capita revenues of all gminas. Box 7‑6. Poland decentralizes to the lowest tier of sub-national government Poland has three tiers of SNG. At the top are 16 voivodships. These are divided into powiats, which are, in turn, divided into gminas. (Some jurisdictions have the combined status of gminas and powiats, termed miastos.) Unlike in Germany, the Russian Federation, or Kazakhstan, most subnational spending in Poland occurs at the lowest tier of SNG: the gmina/miasto level. This is largely because funding for preschool, primary, and secondary education passes through the gmina/miasto budget. According to the most recent available data, 78 percent of subnational spending occurs at this level. Only 12 percent occurs at the powiat level and only 9 percent at the voivodship level. As in the Russian Federation, the total amount of the equalization formula is determined endogenously. After taking the contributions from richer jurisdictions, the contributions from the central budget must be sufficient to achieve the equalization target. In addition to the equalization transfer, SNGs in Poland receive earmarked transfers for education. This is intended to compensate SNGs for the costs of providing the level of education assigned to their respec- tive tiers. Until 2003, the total amount for the education subvention was defined as a percentage of the national budget. Since 2004, the amount has been determined through the annual budget process, with the proviso that it must be equal to the previous year’s amount. The annual budget law, in turn, does not lay 212 SNGs are also entitled to shares of the centrally administered CIT, but those shares are small: roughly 15 percent of CIT is distributed to down a specific voivodships process for allocating the subvention among individual jurisdictions but instead delegates ; 1 percent to the powiats, and 7 percent to the gminas—all on the basis of origin. The remaining 77 percent is retained by this role to an ordinance the central government. issued by the Ministry of Education. Although the actual distribution algorithm varies 213 from year In calculating to year, revenues the taxes—where from own bulk of the education gminas subvention have some is control over allocated the based on rate—the subvention enrolment, formula assumes thatwith high- the maximum rate has been applied. 170 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH er weights for pupils in rural areas and small towns, pupils with disabilities, schools in which classes are taught in minority languages and vocational schools. Since gminas are responsible for most of the public education below the university level, they are the principal recipients of the educational subvention. Implications for Kazakhstan’s system of subventions and withdrawals These examples of systems differ from Kazakhstan’s current system of subventions and withdrawals in two important respects. First, they employ a simple, readily measured indicator of the equalization target. Rather than attempting to calculate the gap between own-source revenues and expenditure needs, they measure inequality regarding variations in per capita revenues. Second, they do not attempt to eliminate disparities among individual jurisdictions; they simply aim to reduce them. In Kazakhstan, setting the target based on variations in own-source revenues would not be immediately practical, as the level of per capita own-source revenues—except in the cities of Atyrau, Astana, and Almaty—is extremely low. In Kazakhstan, as a whole, average per capita SNG own-source revenues are only KZT 215,000 (US$500.) A more feasible approach in the first year of implementation would be to set the target as a percentage of total per capita revenues in the preceding (pre-reform) year.214 Figure 7‑9 shows the total per capita revenues resulting from setting the initial target at 80 percent of average per capita revenues in 2021. It also shows how much of the total would consist of own-source revenues and how much of transfers. As shown, all but three jurisdictions would receive transfers, but the proportions would vary considerably. In Turkestan, transfers would account for 85 percent of total revenues; in Mangistau, they would account for only 30 percent, while the cities of Almaty, Astana, and Atyrau would receive no transfers at all. Figure 7‑9. Per capita revenues after equalization at 80 percent of average total revenues (KZT ’000) Source: World Bank staff calculations based on data published by the authorities. Many variants of this approach are, of course, possible. For example, the Government might consider setting up a separate, earmarked transfer for education, as is done in Poland. This would ensure that the education spending needs in each oblast, as determined by the MoE, are fully financed and are spent on that function. If so, the amount of equalization transfer would be reduced proportionately. The Government might also con- sider adding a ‘fraternal’ equalization element to the transfer methodology along the lines of Germany and Poland. This would resemble the current system of withdrawals in the sense that it would require contributions from richer jurisdictions. But, in this case, ‘richer’ would be defined based on per capita revenue: contributions would be required from jurisdictions whose per capita revenues were above the threshold amount. However, 214 this approach Total per capita would revenues would not consist immediately of own-source solve revenues plusthe second subventions inquestion of transfer the year preceding design: how to de- the reform. 7. IMPROVING INTERGOVERNMENTAL FISCAL RELATIONS 171 termine the total amount of the transfer. As noted above, the transfer systems in the Russian Federation, Poland, and, to a lesser degree, Germany are open-ended: the central government is obligated to transfer as much as required to meet the equalization target. Kazakhstan could address this in two ways. First, it could adjust the target each year depending on the medium-term fiscal framework: the total transfers enve- lope should align with total government spending consistent with the non-oil deficit target. In revenue-flushed years, the target could be increased. It could be kept constant or even reduced in more austere times. Alterna- tively, the Government could tie the total amount of the transfer to indicators of resource availability and then allocate that amount to individual jurisdictions proportionately. In other words, if the total amount required to bring all poorer jurisdiction up to 80 percent of the national average were KZT 100 billion and the total amount available for transfers were only KZT 75 billion, each jurisdiction would receive 75 percent of the amount it would otherwise receive. Increasing revenue stability at the second and third tiers of subnational governments As noted above, the levels of SNGs’ revenues at the second and third tiers are subject to the whims of their respective oblast governments. Oblasts determine the percentage of PIT, social tax, and CIT that subordinate levels of government receive each year. They also determine the amounts of subvention that these levels of government receive. Reportedly, oblasts generally have no fixed criteria for making these decisions. Instead, akims at the second and third tiers are forced to negotiate for such resources. Although second and third-tier SNGs do have some tax bases that are exclusively assigned to them, they are unlikely to provide substantial revenues.215 Second-tier SNGs can retain 100 percent of the property and vehicle taxes collected within their territories, except for taxes on property and vehicles located and registered in third-tier jurisdictions. They are also assigned proceeds from excise taxes on alcohol, tobacco, and gasoline. Third-tier SNGs are permitted to retain 100 percent of PIT on individuals ‘who have declared in the territory of the city of raion significance, village, township the registration of the state revenue agency,’ as well as the taxes on property and vehicles registered within their territories. However, these latter taxes do not apparently raise much revenue. Assigning more revenue to second and third tiers of SNGs needs to consider the functions that lower tiers of SNGs are required to perform. Finance follows function. Because public education is directly financed out of oblast budgets, the responsibilities of lower tiers of government seem to be small. Judging from the available data, the largest category of functional expenditure at the second and third tiers of SNGs is housing and communal services. According to the budget classification system (and confirmed in interviews during the mission), much of this consists of constructing, maintaining, and repairing the public housing stock and capital investments in urban public utilities (water, heating, and sanitation).216 It reportedly also includes the costs of parks and street cleaning. If, in fact, most spending at this level is devoted to road repair, solid waste management, and other services that benefit residents at large, then it is appropriate to finance these func- tions from local budgets. However, suppose much local spending goes on public housing construction and maintenance and related investment in extending utility networks. In that case, it raises questions regarding why local governments should be in the business of constructing housing and why the costs of maintaining public housing should be borne by local taxpayers rather than by the occupants of the individual housing units. Different countries use different methods to finance such local services. Figure 7‑10 illustrates the sources of revenues of the lowest tiers of local government in Germany, Poland, and the Russian Federation.217As shown, transfers are the largest source of local revenues in all three countries. But these countries are not 215 The yield of these taxes is not revealed in the published version of the statistical bulletin. 216 The operating costs of urban utilities are reportedly covered by tariffs, which are retained by the SOEs that provide them. 217 Source: IMF Government Finance Statistics https://data.imf.org/?sk=A0867067-D23C-4EBC-AD23-D3B015045405 172 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH strictly comparable to Kazakhstan in one important respect. In all three cases, local governments spend a significant proportion of their budgets on education. (In Germany, the proportion is 17 percent; in Poland, 25 percent; and in the Russian Federation, 41 percent). Transfer systems in all three countries are intended, in part, to finance these costs. In Kazakhstan, of course, education is not a responsibility of the lower tiers of government. Figure 7‑10. Sources of local revenue: Germany, Poland, and the Russian Federation (% of total, 2020) Source: World Bank staff calculations. Kazakhstan might consider addressing constraints in the revenue sources of the second and third tiers of SNGs. As described earlier, Kazakhstan already assigns certain taxes to the lower tiers of SNGs, but they do not generate much revenue. If the Government wishes to increase the yield of local taxes, it could address the constraints on these revenue sources. It could, for example, authorize second and third- tier SNGs to retain a share of the entire PIT generated in their territories, not just the revenue from locally registered taxpayers. By the same token, it could address possible constraints on the yield of the property tax (see Box 21). All three countries (Germany, Poland, and the Russian Federation) provide SNGs with independent tax bases that they are free to use for any function that falls under their responsibility. In all three countries, the principal source of tax revenues is PIT, as shown in Figure 7‑10. There, SNGs are also permitted to impose local taxes on property, goods, and services, although neither of these latter sources constitutes a major source of local tax revenue. Overall, the focus of reform in the system of intergovernmental finance should be on the transfer system. The system of transfers from first-tier subnational governments to second and third-tier SNGs should be based on stable and objective formulas. And the system of transfers from the central government to first-tier SNGs should be simplified, focusing on reducing disparities in per capita revenues rather than attempting to equalize the quality of public services in every jurisdiction in the country. 7. IMPROVING INTERGOVERNMENTAL FISCAL RELATIONS 173 Box 7-7 Why are the property tax yields so low? In Kazakhstan, cities (at all three levels of SNG) and rural districts can impose two forms of property tax: land and buildings. The land tax is imposed on agricultural land and ‘land in populated areas,’ including residential and non-residential land. The building tax is imposed on residential, commercial, and industrial buildings. However, international experience suggests that problems at all four administration steps may limit property tax yield. Discovery: Properties may be missing from the tax rolls, particularly if the tax authority relies only on owner declarations to discover taxable properties. Valuation. Properties may be undervalued. In Kazakhstan, properties owned by physical persons are valued using a mass-appraisal technique, in which the physical characteristics of a land parcel (e.g., its size, location, and use) and a building (its square footage, location, construction materials, age, use, etc.) are converted into an estimate of value based on formulas. Mass appraisal is a widely used objective basis for determining property values but is subject to two vulnerabilities. First, data on the physical characteristics of properties may be incomplete or inaccurate. Second, the formula used to convert physical characteristics to value may be inaccurate or outdated. Ideally, the value assigned to a given characteristic would be based on recent market data. But market data may be inaccurate or out of date.218 The value of a property owned by a legal person (e.g., a corporation) is based on its book value. This raises problems of its own, as book values are often based on original purchase prices, adjusted for depreciation, and may not reflect current market values. Rates and exemptions: Tax rates may be too low. This appears to be the case in Kazakhstan. The rates on substantial buildings are low. For example, the land tax on a 600 m2 residential property in Almaty would be only KZT 575 (US$1.25). If the residence resting on it were valued at KZT 50 million (US$110,000), the building tax would be just KZT 141,555 (US$310). In many counties, exemptions also tend to be too generous, although this does not appear to be true in Kazakhstan. The Tax Code exempts pensioners living alone, WWII veterans, and certain classes of farmworkers but does not extend to large groups of urban property owners. Collection: Collection is the often the weakest point in property tax administration. Poor collection performance often reflects the obstacles facing taxpayers who would be willing to pay: bills are not delivered to the correct taxpayers, and the process of paying taxes involves waiting in long lines or presenting obscure documents. But much of the problem in collection lies in the taxing authority’s failure to enforce penalties. Like many countries, the Tax Code in Kazakhstan provides draconian penalties for failure to pay property taxes in full and on time. But such penalties only affect collection if enforced. In many countries, enforcement is patchy. It is not known whether this is the case in Kazakhstan. 218 A report prepared in the mid-2000s found substantial problems in Kazakhstan’s valuation system, although these may have been addressed since then. https://www.registrucentras.lt/bylos/dokumentai/conferency/KAZAKHSTAN_Vilnius%20version.pdf 174 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Annex 1. Analyzing the Distribution Impact of Fiscal Policies To study the distributional impact of fiscal policy in Kazakhstan, we use the CEQ methodology (Lustig 2018). The CEQ is a comprehensive incidence analysis that uses data from household surveys and national accounts to assess the impact of taxes and public transfers on household poverty and inequality. The approach has been applied in over 70 countries, which allows us to benchmark Kazakhstan’s performance with relevant peer countries. The method is based on an accounting approach; it adds and subtracts taxes and transfers to household per-capita income to measure income before and after each fiscal intervention. The per-capita household income after transfers and taxes Yh for household h is given by (1) where Ih is the income before taxes and transfers, Ti is the taxes paid by households (i is the range of taxes analyzed), Bj are the transfers received by households (j is the range of transfers studied), and Sih and Sjh are the shares of tax i and transfer j paid and received by households, respectively. Data The main source of information is the 2021 Household Income and Expenditure Survey (HIES) collected by the Bureau of National Statistics (BNS). The HIES is a quarterly representative household survey that collects detailed information on consumption, labor income, social assistance, pensions, remittances, financial income, assets, housing characteristics, accessibility to the labor and financial market, health services, education opportunities, and individual characteristics such as education, health, and labor market status and experience. The survey has been collected on an annual basis since 2001. We use the 2021 round, the latest available HIES round at the start of the study. Relative to standard incidence analysis, the CEQ methodology’s main strength is to provide a framework to analyze the personal impact of different taxes and transfers and their overall combined impact on poverty and inequality. As such, the CEQ is a tool that enables the generation of evidence on both the ‘small picture’ (the impact of a specific fiscal intervention) and the ‘big picture’ of overall taxes and spending for policy research. Moreover, once the overall system has been estimated, it offers a platform to simulate policy changes, assess their distributional implications, and contribute an equity lens to policy discussions about fiscal reform. At the same time, the CEQ shares certain limitations with standard incidence analysis. Among these caveats are: (i) it is a partial equilibrium analysis: it does not model behavioral responses, lifecycle, and spillover effects; (ii) it does not consider externalities, for instance, long-term increases in national productivity that arise from higher investment in education; (iii) the approach assumes that indirect taxes and contributions are borne entirely by the income earner, and indirect taxes are borne entirely by the consumer; (iv) the methodology cannot analyze all taxes and spending, and interventions such as corporate profit taxes, corporate subsidies, infrastructure investment (e.g., water projects) are left out; and (v) the approach does not consider the quality of public services provided. Overall, the CEQ is not only a first-order approximation to the impact of taxes and social spending on poverty and inequality at a given time but the most comprehensive methodology to do so to date. This analysis follows standard practice to measure progressivity and uses the Kakwani index (Kakwani 1977). A tax (benefit) is progressive whenever its burden (entitlement) rises (decreases) with income. In the case of transfers, the Kakwani index is defined as the difference between the Gini coefficient of Market Income plus pensions (when pensions are treated as deferred income) and the concentration coefficient of the transfers. Meanwhile, for each tax, the Kakwani index is calculated as the difference between the 1. STRENGTHENING THE FISCAL FRAMEWORK FOR INCLUSIVE AND RESILIENT GROWTH 175 concentration coefficient of the tax and the Gini coefficient of Market Income plus pensions. The Kakwani index for taxes will be positive (negative) if a tax is globally progressive (regressive). In contrast, the Kakwani index for transfers is positive if a transfer is progressive in relative terms. To analyze whether a tax or transfer is equalizing or not, the analysis uses the marginal contribution of taxes and transfers to income inequality measured by the Gini coefficient. The marginal contribution measures the marginal reduction in inequality due to a tax or a transfer. It is the difference between the Gini coefficient without fiscal intervention and the Gini coefficient of all income components together. The intervention is equalizing whenever the marginal contribution is positive. By comparing the marginal contribution and the Kakwani index, we can determine whether a fiscal intervention is equalizing (un-equalizing) despite being regressive (progressive). Selected countries are used for benchmarking. These upper-middle-income and high-income countries have implemented the same analysis; Argentina, Brazil, Chile, Mexico, Croatia, Poland, Russia, the United States, and Turkey. 176 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Annex 2. Estimating Fiscal Multipliers for Kazakhstan Methodological Approach Fiscal multipliers are based on a three-variable structural Bayesian vector autoregressive (SBVAR) model, including central government spending, real GDP, and tax revenue, as in Blanchard and Perotti (2002). The model is specified as follows: Where is an vector of endogenous variables, is a matrix of lagged dependent variables including an intercept, B is a matrix of parameters, and is a matrix of independently and identically distributed errors. The model is estimated on quarterly data from 2000 to 2019 with four lags. It is estimated using Bayesian techniques. A dummy variable is included for 2017Q3. The Minnesota prior is used with an overall tightness parameter set at 0.2, cross-variable weighting at 0.9, and lag decay set to 1.5. The model is estimated on 25,000 draws, with the first 5000 discarded and the 5th draw kept. To identify the structural shocks of the model, the following sign restrictions are imposed: where a structural government spending shock (ɛ) is defined as that which increases government spending and output, a business cycle shock raises output and tax, and a tax shock increases tax revenue but decreases output. These restrictions are adapted from (Mountford and Uhlig 2009), (Caldara and Kamps, What are the effects of fiscal policy shocks? A VAR-based comparative analysis 2008) and (Caldara and Kamps 2017). Restrictions are imposed for two quarters. The variables used are real government expenditure, excluding operations with financial assets and interest payments on government debt. Tax revenue is general government tax receipts. These are both deflated using the GDP deflator. As a proxy for non-oil output, the difference between total output and output in mining and quarrying is used. Since mining output data only starts in 2012 every quarter, the data is extended backward based on a simple linear regression of mining output on Kazakhstan’s oil-production data. All data is seasonally adjusted. The variables are scaled as in (Ramey and Zubairy 2018) by trend GDP estimated using the Hodrick-Prescott filter. This rescaling transforms all variables into the same units and allows direct estimation of fiscal multipliers. Alternatively, regressions that estimate fiscal multipliers using logarithms of the variable (real output and government spending) produce estimates of elasticities that must be converted to multiplier equivalents using an ex-post conversion factor of the sample average for the ratio of GDP to government spending. Such conversions may be particularly sensitive to the choice of the sample period as the ratio of GDP to government spending tends to be volatile. KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 177 Annex 3. PEFA 2018 Performance Indicators Table 0‑1. Summary of Performance Indicators and Dimensions Scores Source: (World Bank 2018) Table A1.1 shows the scores for each of the Performance Indicators and Dimensions. Indicators are scored on a scale from A (highest) to D (lowest). Indicators marked M1 base the overall score on the lowest score of any dimension (the Weakest link method); a + indicates that other dimension(s) received higher scores. For indicators marked M2, the scores are averaged (the Averaging method) according to a table in the PEFA Handbook. 178 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Annex 3. Selected VAT Exemptions and Policy Recommendations Current VAT Exemptions List Recommended Actions Justification Turnover from the sale of financial Define specifically what The definition of specific transactions exempt from VAT. financial transactions are financial transactions to be to be kept exempt and exempt or subject to the what to be transferred to standard regime is based on the standard regime. whether their value-added can be determined. Buildings and structures sold by the Recommended to This item should conform to state Islamic special financing company be transferred to the the standard VAT treatment to the standard regime. on housing/property. If this authorized body for state property exemption is to remain, it is a management, which were earlier tax expenditure provision; as acquired under contracts concluded per such, its revenue loss is to be the terms of issuance of state Islamic estimated and incorporated in securities, and land plots occupied by the report on tax expenditure such property. analysis for transparency and accountability. Property in the form of winnings given by Recommended for This should conform to the a lottery operator to a lottery participant. transfer to the standard standard VAT treatment of VAT regime. property/asset transfers. Services for the processing and (or) To be defined clearly Destination principle in VAT is to repair of goods imported into the and unambiguously. If be adhered to. customs territory imported and consumed of the Eurasian Economic Union under within Kazakhstan, a the customs procedure for processing in standard VAT rate is the customs territory. applied. If consumed outside the country, zero- rated. Services within the framework of the Specific consideration on The standard VAT is to apply activities of a cooperative of owners of whether to retain in the if a business (regardless of premises (apartments), associations exemption list or transfer its corporate forms) is a VAT- of property owners of an apartment to the standard VAT registered entity. building for the management of a regime would be related condominium facility, carried out per the to the status of the VAT housing legislation of the Republic of registration (e.g., the size Kazakhstan. of the annual turnover above the VAT registration threshold). KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 179 Current VAT Exemptions List Recommended Actions Justification For goods works, services provided in Tax benefits can be The definition of the VAT the taxable period of a sale, and also in considered in the form exemption threshold can be the four preceding taxable periods, one of income tax exemption used to screen businesses as of the following conditions is observed: or tax credits, not within being in or outside of the VAT the average number of disabled people the paradigm of VAT net. is at least 51 percent of the total exemption. Adding further criteria for number of employees; expenses for defining the standard VAT the remuneration of labor of disabled regime would make the VAT people make up at least 51 percent (in policy and administration more specialized organizations employing complicated. people with the loss of hearing, speech, vision - at least 35 percent) of total labor expenses. The provisions of this subparagraph do not apply to turnovers from the sale of excisable goods. Works, services for free repair and To be transferred to the The listed works and services (or) maintenance of goods during the standard VAT regime. are not part of the standard warranty period set by a deal, including exemption criteria and risk being the value of spare parts and their subject to abuse. components, if the deal terms provide for the taxpayer’s warranty of goods sold, works performed, and services rendered. Scrap and waste of non-ferrous and Transferred to the These items are not part of the ferrous metals. standard VAT rate regime. standard VAT exemption list. Religious items by religious associations Transferred to the These are not standard registered with registering authority. standard VAT regime. exemptions. The list of specified goods and criteria for its formation shall be approved by the Government of the Republic of Kazakhstan. Special social services provided by Transferred to the These are not standard non-commercial organizations per the standard VAT regime. exemptions and are subject to legislation of the Republic of Kazakhstan If non-commercial abuse. on special social services. organizations’ annual turnover is below the VAT threshold, they are not obliged to register the VAT. Services for conducting socially Transferred to the These are not standard significant events in the field of culture, standard VAT regime. exemptions and are subject to spectacular cultural events held as part abuse. of a state tasks per the legislation of the Republic of Kazakhstan on culture. 180 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Current VAT Exemptions List Recommended Actions Justification Services for exercising cultural, Transferred to the These are not standard educational, scientific, and research standard VAT regime. exemptions and are subject to functions by abuse. museums and ensuring the popularization of the historical and cultural heritage of the Republic of Kazakhstan. Services for exercising information, Transferred to the These are not standard cultural, and educational functions of standard VAT regime. exemptions and are subject to libraries. abuse. Services and works in the field of culture Transferred to the These are not standard and education carried out by theaters, standard VAT regime. exemptions and are subject to philharmonic societies, cultural and abuse. recreational organizations. Scientific and restoration works at Transferred to the These are not standard historical and cultural sites conducted on standard VAT regime. exemptions and are subject to the basis of a license for such activity. abuse. Additional education services provided These services are to be Exempting basic/primary and by an educational organization licensed defined and classified secondary/post-secondary for educational activity. if they are part of basic education should be retained. education. Blanket exemption of all types of services labeled under education would risk compromising both equity and revenue-enhancing objectives. Vehicles and (or) agricultural machinery Agriculture should be Exemption of agricultural inputs provided all of the following requirements treated as another would risk cascading along the are met: a realizable vehicle and (or) economic sector: Those production-distribution chain. agricultural machinery include earlier small (with turnover imported raw materials and (or) under the VAT registration materials that are exempt from VAT threshold) are exempt per subparagraph 15) of paragraph 1 – whereas others are to of Article 399 or subparagraph 4) of be taxed in the standard paragraph 2 of Article 451 of this Code; regime. importation of raw materials and (or) materials as part of a realizable vehicle and (or) agricultural machinery is carried out by a legal entity selling these vehicles and (or) agricultural machinery; vehicles and (or) agricultural machinery are included on the list of vehicles and agricultural machinery, the sale of which is exempt from VAT, approved by the authorized body for state support to industrial and innovation activity in coordination with the central authorized body for state planning and the authorized body. KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 181 Current VAT Exemptions List Recommended Actions Justification Goods, works, and services sold in the Transferred to the VAT incentives for SEZs should territory of the special economic zone, standard VAT regime. be abolished. the limits of which fully or partially coincide with the sections of the customs border of the Eurasian Economic Union. Services rendered by sports Transferred to the These services are not part of organizations based on state-task standard VAT regime. standard VAT exemptions. contracts. Works and services performed and Transferred to the These services are not part of provided by the cinematographic standard VAT regime. standard VAT exemptions. organization for the investor in the production of films. Goods produced and sold by the Standard VAT regime to Similar to the case of SEZs, VAT participants of the Astana-Hub be applied. incentives should be abolished. international technological park that meet the conditions of Paragraph 4-3 of Article 293 of this Code. Works, services sold by the participants Standard VAT regime to Similar to the case of SEZs, of the Astana-Hub international be applied. incentives through VAT should technological be abolished. park that meet the conditions of Paragraph 4-3 of Article 293 of this Code. 182 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Annex 4. Selected Non-export Zero Rating and Policy Recommendations VAT Zero rating List Recommended Actions Justification Taxation of goods realizable to the territory of Transferred to standard According to the a special economic zone. VAT. destination principle, only exports are zero- rated. Non-export zero rating would raise the compliance and enforcement costs (increased backlog of refund claims) and be subject to abuse. VAT incentives for SEZs should be abolished. Features of taxation of goods realizable to the Transferred to standard The same justification as territory of the special economic zone “Astana VAT. above. - a new city.” Features of taxation of goods sold on the Transferred to standard The same justification as territory of the special economic zone, the VAT. above. limits of which fully or partially coincide with the sections of the customs border of the Eurasian Economic Union. KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 183 Annex 5. Assignment of Functions in Kazakhstan According to the Budget Code (Summary) Raions, cities Capital, Republic Central Oblast of oblast Okrugs Cities significance Operations of Operations of Operations of city Operations of Operations of parliament, oblast maslikhat maslikhat and raion maslikhat okrug, village presidency, and executive executive bodies and executive executive bodies General and central bodies bodies Administra- govt tion executive bodies, fiscal transfers to SNGs National Protection of Protection of      defense public order, public order, Defense, material support material support public order to the central to the central military military Training of Technical and Preschool and Preschool, Preschool specialists vocational, secondary primary, basic transportation of post-secondary education, secondary, students to the Education education technical and and general nearest school in vocational, secondary rural areas post-secondary education education219 Provision of  ‘Additional’ ‘Additional’ Health care Transport of free medical provision of free provision of free for school the seriously ill care medical care medical care employees to the nearest Healthcare hospital, medical assistance in emergency cases Pensions, Social Social assistance Material support social assistance to orphans, for disabled assistance, to orphans, children left persons per maternity children left without parental individual Social benefits without parental care, the elderly, treatment assistance care, the elderly, and disabled programs, and disabled domestic social assistance to citizens in need 219 In the version of the Budget Code enacted in 2019 (and in effect as of January 2021) there is no general provision for primary education in Republic cities. Republic cities are, however, required to provide text books and other teaching materials to ‘public education organizations that implement general educational programs for primary, basic secondary and general secondary education’. 184 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Support of Organization Support of Support for theaters, of sports and theaters, cultural and protection cultural events protection of leisure work at Culture of historical historical artifacts, the local level artifacts sports, and cultural events Protection Construction of and water pipelines Water regulation of regional resources, of water significance; forestry resources, reforestation forests Construction, Provision/ Construction, Construction, Construction, repair, and regulation of repair, and repair, and repair, and maintenance passenger maintenance maintenance of maintenance of of roads of transportation of streets of roads of raion roads of okrug, international on intercity, cities, traffic importance, village, township and national interregional, management, of provision/ significance Transporta- importance and commuter traffic, provision/ regulation of tion lines regulation of passenger control passenger transportation transportation on urban, on urban and suburban, and suburban roads inter-raion lines, except for railroads Development Organization of Provision of Provision of Provision of of district architectural, water supply, water supply, sanitation of heating town planning, sewage, drainage, sewage, settlements; supply and construction sewerage, heating, drainage, street lighting in electricity, activity and electricity heating, and populated areas gas, networks that electricity water, and are in community networks in Urban sanitation, property;220 ‘settlements,’ infrastruc- within provision of provision ture settlements. sanitation in of housing settlements; for certain organization of categories of architectural, citizens town planning, and construction activity 220 In the version of the Budget Code enacted in 2019 (and in effect as of January 2021) there is no general provision for primary education in Republic cities. Republic cities are, however, required to provide text books and other teaching materials to ‘public education organizations that implement general educational programs for primary, basic secondary and general secondary education’. KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 185 Bibliography Abilov, Nurdaulet. 2021. “A medium-scale Bayesian DSGE model for Kazakhstan with incomplete exchange rate pass-through.” International Economic Journal 35 (4). Abilov, Nurdaulet, Henry Aviomoh, and Sjamsu Rahardja. 2022. In search for an optimal fiscal rule: the case of Kazakhstan. Unpublished paper. Accounts Committee of the Republic of Kazakhstan. 2020. “Conclusion to the report of the Government of the Republic of Kazakhstan on implementation of Republican Budget for 2020.” n.d. Address of the President of the Republic of Kazakhstan on Strategy Kazakhstan 2050: New polit- ical course of the established state. Accessed April 7, 2022. https://policy.asiapacificenergy. org/sites/default/files/Presidential%20Address%20%27Strategy%20Kazakhstan-2050%27%20 %28EN%29.pdf. Adilet legal information system. n.d. On Local Government and Self-government in the Republic of Kazakh- stan. Agenor, Pierre-Richard. 2016. “Optimal fiscal management of commodity price shocks.” Journal of Devel- opment Economics 122: 183-196. Auerbach, Alan J, and Yuriy Gorodnichenko. 2012. “Measuring the Output Responses to Fiscal Policy.” American Economic Journal: Economic Policy 4 (2): 1-27. Avvisati, F., A. Echazarra, P. Givord, and M. Schwabe. 2019. “PISA 2018 Country Note for Kazakhstan.” OECD. https://www.oecd.org/pisa/publications/PISA2018_CN_KAZ.pdf. Barbiero, Francesca, and Zsolt Darvas. 2014. “In Sickness and in Health: Protecting and Supporting Public Investment in Europe.” Bruegel Policy Contribution, February. Bardhan, Pranab. 2002. “Decentralization of Governance and Development.” Journal of Economic Per- spectives 16 (Fall): 185-205. Beazley, Ivor, Ronnie Downes, and Scherie Nicol. 2019. Budgeting in Kazakhstan: A roadmap for continued budgetary governance reform. OECD Journal on Budgeting. Blanchard, Olivier, and Roberto Perotti. 2002. “An Empirical Characterization of the Dynamic Effects of Changes in Government Spending and Taxes on Output.” The Quarterly Journal of Economics 117 (4): 1329–1368. Brownt, Deryck, and Jeevakumar Govindasamy. 2019. “National development planning in Malaysia: Learn- ing from the past to chart the future.” World Bank Blogs. https://blogs.worldbank.org/eastasiapa- cific/national-development-planning-malaysia-learning-past-chart-future. Bundesministerium der Finanzen. 2018. “Bundesfinanzministerium.de.” https://www.bundesfinanzminis- terium.de/Web/DE/Home/home.html. Accessed December 2022. https://www.bundesfinanzmin- isterium.de/Content/DE/Downloads/Oeffentliche-Finanzen/Foederale-Finanzbeziehungen/ Laenderfinanzausgleich/bundesstaatlicher-finanzausgleich-englisch.pdf?__blob=publication- File&v=5. Caldara, Dario, and Christophe Kamps. 2017. “The Analytics of SVARs: A Unified Framework to Measure Fiscal Multipliers.” The Review of Economic Studies 1015–1040. 186 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Caldara, Dario, and Christophe Kamps. 2017. “The Analytics of SVARs: A Unified Framework to Measure Fiscal Multipliers.” The Review of Economic Studies, Volume 84, Issue 3 1015–1040. Caldara, Dario, and Christophe Kamps. 2008. What are the effects of fiscal policy shocks? A VAR-based comparative analysis. Frankfurt am Main: European Central Bank. Catalano, Michele, Lorenzo Forni, and Emilia Pezzolla. 2020. “Climate-change adaptation: The role of fiscal policy.” Resource and Energy Economics 59 (C). Curristine, Teresa, Nino Tchelishvili, Virginia Alonso-Albarran, Alberto Soler, Gemma Preston, and Sureni Weerathunga. 2021. Gender Budgeting in G20 Countries. IMF Working Paper. Deutsche Bundesbank. 2019. European Stability and Growth Pact: individual reform options. Monthly Re- port, Deustche Bundesbank. Dime, Roselle, Edimon Ginting, and Juzhong Zhuang. 2021. Estimating Fiscal Multipliers in Selected Asian Economies. ADB Economics Working Paper Series. Downes, Ronnie, Lisa von Trapp, and Scherie Nicol. 2017. “Gender budgeting in OECD countries.” OECD Journal on Budgeting. EBRD. 2018. The fiscal implication for Kazakhstan of the worldwide transition to a greener global economy. London: European Bank for Reconstruction and Development. Egert, Bala’zs. 2014. “Fiscal Policy Reaction to the Cycle in the OECD: Pro or Counter-Cyclical?” Mondes en Development 42–2014: 35–52. European Commission. n.d. Carbon Border Adjustment Mechanism. Accessed April 19, 2022. Furceri, Davide, Prakash Loungani, Jonathan D. Ostry, and Pietro Pizzuto. 2022. “Will COVID-19 Have Long-Lasting Effects on Inequality? Evidence from Past Pandemics.” The Journal of Economic In- equality. 2021. Global Gender Gap Report 2021. World Economic Forum. Hererra, Santiago, and Eduardo Olaberria. 2022. Budget Ridigity in Latin America and the Caribbean: Caus- es, Consequences, and Policy Implications. Washington, DC: World Bank. Hory, Marie-Pierre. 2016. “Fiscal Multipliers in Emerging Market Economies: Can We Learn Something From Advanced Economies?” International Economics 146: 59–84. Iltzetzki, Ethan, and Carlos Vegh. 2008. “Procyclical Fiscal Policy in Developing Countries: Truth or Fiction?” Working Paper. National Bureau of Economic Research. Ilzetzki, Ethan, G Enrique Mendoza, and Carlos A Vegh. 2013. “How Big (Small?) Are Fiscal Multipliers?” Journal of Monetary Economics 114: 109–125. IMF. 2019. ARTICLE IV CONSULTATION. IMF. IMF. 2016. How to Design and Enforce Tobacco Excises? IMF. IMF. 2020. Kazakhstan IMF Country Report No 20/38, Selected Issues. Washington, DC: International Monetary Fund. IMF. 2019. Report on technical assistance mission on government finance statistics. International Mone- tary Fund. KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 187 IMF. 2021. Republic of Kazakhstan: 2021 Article IV. IMF. International Budget Partnership. 2019. Open Budget Survey 2019. International Budget Partnership. International Budget Partnership. 2021. Open Budget Survey 2021. International Budget Partnership. n.d. Kazakhstan 2050 Strategy: A New Course for the Republic of Kazakhstan. https://kazakhstan2050. com. KAZKENES. n.d. https://kazkenes.kz/ru/. Madani, Dorsati, and Ilyas Sarsenov. 2015. Kazakhstan: Adjusting to Low Oil Prices, Challenging Times Ahead. Washington, DC: World Bank. Mare, Davide S, Martin Melecky, and Hanna Murina. 2022. Bank Bailouts and Fiscal Contingent Liabilities. World Bank Policy Research (forthcoming). Ministry of Information and Public Development of the Republic of Kazakhstan. 2021. “National report on public councils in the Republic of Kazakhstan.” Ministry of Information and Social Development of the Republic of Kazakhstan. 2020. “Annual report on public councils activities in Kazakhstan.” Mountford, Andrew, and Harald Uhlig. 2009. “What are the effects of fiscal policy shocks?” Journal of Ap- plied Econometrics 960-992. OECD. n.d. Action 15 Multilateral Instrument. https://www.oecd.org/tax/beps/beps-actions/action15/. OECD. 2019. Budgeting and Public Expenditures in OECD Countries. Paris: OECD. OECD. 2016. Financing Climate Action in Kazakhstan. Paris: OECD. OECD. 2022. Green Budgeting: The Way Forward. Paris: OECD. OECD. 2019. OECD Good Practices for Performance Budgeting. Paris: OECD Publishing. OECD. 2020. OECD Tax Policy Reviews: Kazakhstan. Paris: OECD. OECD. 2020. Transparent and Inclusive Stakeholder Participation through Public Councils in Kazakhstan. Paris: OECD. PWC. 2021. Worldwide Tax Summaries. PWC. Ramey, Valerie, and Sarah Zubairy. 2018. “Government Spending Multipliers in Good Times and in Bad: Evidence from US Historical Data.” Journal of Political Economy 850 - 901. Reuter, Wolf Heinrich. 2020. Benefits and drawbacks of an “expenditure rule,” as well as of a “golden rule,” in the EU fiscal framework. Study, Economic Governance Support Unit of the European Parliament. https://www.europarl.europa.eu/RegData/etudes/STUD/2020/645732/IPOL_ STU(2020)645732_EN.pdf. Robinson, Marc. 2022. “Public finances after the COVID-19 pandemic.” OECD Journal on Budgeting. Sebastian, James. 2009. Incentives and Investments: Evidence and Policy Implications. Washington, DC: World Bank. Sheremirov, Viacheslav, and Sandra Spirovska. 2022. “Fiscal multipliers in advanced and developing coun- tries: Evidence from military spending.” Journal of Public Economics. 188 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH Stone, Vincent, Alberto Soler, John Zohrab, Wei Shi, Avril Halstead, and Paul Gardiner. 2020. Kazakhstan: Assessing Macro Fiscal Risk and State-Owned Enterprises. IMF. Strategy 2050. 2021. Carbon neutrality: scenario announced to achieve zero emissions in Kazakh- stan. https://strategy2050-kz.translate.goog/ru/news/uglerodnaya-neytralnost-ozvuchen-st- senariy-po-dostizheniyu-nulevykh-vybrosov-v-kazakhstane/?_x_tr_sl=ru&_x_tr_tl=en&_x_tr_ hl=ru&_x_tr_pto=sc. Tang, Vincent, Alberto Soler, John Zohrab, Wei Shi, Avril Halstead, and Paul Gardiner. 2020. Kazakhstan: Assessing Macro Fiscal Risks and State-Owned Enterprises. Technical Report, IMF. The Astana Times. 2021. New Law on Elections Aims to Engage Kazakh Citizens in Political Institutions, Experts Say. https://astanatimes.com/2021/05/new-law-on-elections-aims-to-engage-kazakh-cit- izens-in-political-institutions-experts-say/. U.S. Department of State. 2020. Investment Climate Statements: Kazakhstan. https://www.state.gov/re- ports/2020-investment-climate-statements/kazakhstan/. UNDP. n.d. Kazakhstan. Accessed April 16, 2022. https://www.ndcs.undp.org/content/ndc-support-pro- gramme/en/home/our-work/geographic/europe-and-cis/kazakhstan. UNESCAP. 2021. Accessed May 2022. https://rrp.unescap.org/country-profile/KAZ. —. 2021. Risk and Resilience Portal. Accessed April 2022. https://rrp.unescap.org/country-profile/ KAZ. UNESCO. n.d. The UNESCO Institute for Statistics. http://data.uis.unesco.org/. United States Department of Agriculture. 2020. Kazakhstan COVID-19 Updates and Impacts. United States Department of Agriculture. World Bank. n.d. World Bank Open Data. https://data.worldbank.org/indicator/NY.GDP.PCAP.CD?loca- tions=KZ. World Bank. 2021. Climate Change Budget Tagging: A Review of International Experience. Washington, DC: World Bank. World Bank. 2021. Climate Change Budget Tagging: A Review of International Experience. Washington, DC: World Bank. World Bank. 2022. Climate Change Development Report for Kazakhstan. Washington, DC: World Bank. World Bank. 2021. Green, Resilient, and Inclusive Development. Washington, DC: World Bank. World Bank. 2019. Joint Economic Research Program (JERP) between the Government of Kazakhstan and the World Bank Group to support to Kazakhstan’s 2025 Strategic Development Plan Implementa- tion. Washington, DC: World Bank. World Bank. 2019. Kazakhstan - Country Partnership Framework for the Period FY20-25. Washington, DC: World Bank. World Bank. 2018. Kazakhstan - Systematic Country Diagnostic: A New Growth Model for Building a Secure Middle Class: Kazakhstan - Systematic country diagnostic: a new growth model for building a se- cure middle class. Washington, DC: World Bank. World Bank. 2019. Kazakhstan Monitoring and Evaluation Assessment. Washington, DC: World Bank. KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH 189 World Bank. 2018. Kazakhstan Public Expenditure and Financial Accountability Assessment. Washington, DC: World Bank. World Bank. 2017. Kazakhstan: Enhancing the Fiscal Framework to Support Economic Transformation. Washington, DC.: World Bank. World Bank. 2019. Overview of Tobacco Use, Tobacco Control Legislation, and Taxation. Washington, DC: World Bank. World Bank. n.d. Spending for inclusive and quality education. Washington, DC: World Bank. World Bank. 2021. Supreme Audit Institutions Independence Index: 2021 Global Synthesis Report. Wash- ington, DC: World Bank. —. 2020. Worldwide Governance Indicators. https://info.worldbank.org/governance/wgi/. World Bank; OECD; European Bank for Reconstruction and Development; Islamic Development Bank; Asian Development Bank. 2019. Assessment of the Public Procurement Systems of the Republic of Ka- zakhstan and the Sovereign Wealth Fund Samruk-Kazyna. Washington, DC: World Bank. World Health Organization. n.d. Global Health Expenditure database. https://apps.who.int/nha/database/ Error/en?aspxerrorpath=/nha/database. Ybrayev, Zhandos, Olzhas Kubenbayev, and Akylzhan Baimagambetov. 2022. “Macroeconomic effects of fiscal rules for a commodity-exporting economy: avoiding procyclical bias in Kazakhstan.” Macro- economics and Finance in Emerging Market Economies 1–24. 190 KAZAKHSTAN: STRENGTHENING PUBLIC FINANCE FOR INCLUSIVE AND RESILIENT GROWTH