Myanmar Public Expenditure Review 2017: Fiscal Space for Economic Growth TABLE OF CONTENTS ACKNOWLEDGMENTS ........................................................................................................................ 6 ACRONYMS............................................................................................................................................. 7 EXECUTIVE SUMMARY ....................................................................................................................... 8 FISCAL SPACE FOR SERVICES AND GROWTH .............................................................................. 16 PER MOTIVATION AND FRAMEWORK ............................................................................................................ 16 RECENT MACROECONOMIC AND FISCAL DEVELOPMENTS ............................................................... 20 MEDIUM-TERM FISCAL FRAMEWORK ............................................................................................................. 30 POLICY OPTIONS........................................................................................................................................................ 35 CAPITAL SPENDING EFFICIENCY ................................................................................................. 43 PUBLIC INVESTMENT AND ECONOMIC GROWTH.................................................................................... 43 ALLOCATIVE EFFICIENCY OF PUBLIC INVESTMENT .............................................................................. 47 PRODUCTIVE EFFICIENCY OF PUBLIC INVESTMENT ............................................................................. 56 POLICY OPTIONS........................................................................................................................................................ 63 FISCAL IMPACT OF STATE ECONOMIC ENTERPRISES ............................................................ 69 SEE INSTITUTIONAL FRAMEWORK AND FISCAL RELATIONS ............................................................ 69 SEE FINANCIAL PERFORMANCE AND FISCAL IMPACT .......................................................................... 81 POLICY OPTIONS........................................................................................................................................................ 90 TAX SYSTEM EFFICIENCY................................................................................................................ 95 REVENUE PROFILE AND PRINCIPLES FOR REFORM ............................................................................... 95 TAX REFORM PROGRESS AND ACHIEVEMENTS ........................................................................................ 99 SHORT AND MEDIUM-TERM SOURCES OF REVENUE LEAKAGE ....................................................105 POLICY OPTIONS......................................................................................................................................................112 SOUND PUBLIC DEBT MANAGEMENT ........................................................................................ 117 PUBLIC DEBT MANAGEMENT FRAMEWORK .............................................................................................117 PUBLIC DEBT LEVELS AND SUSTAINABILITY RISKS ..............................................................................121 MEDIUM-TERM BORROWING STRATEGIES ................................................................................................128 DOMESTIC DEBT MARKET DEVELOPMENT ..............................................................................................133 POLICY OPTIONS......................................................................................................................................................136 LIST OF FIGURES Figure 1: Recurrent spending vs. per capita GDP ........................................................................................................... 16 Figure 2: Access to electricity and road density ............................................................................................................... 16 Figure 3: General Government investment/GDP (5-year moving average) .............................................................. 17 Figure 4: Private investment/GDP (5-year moving average) ........................................................................................ 17 Figure 5: Proposed outline of PER 2 ................................................................................................................................ 18 Figure 6: Real GDP (% change and contribution) .......................................................................................................... 20 1 Figure 7: Consumer Price Index (% change and contribution)..................................................................................... 20 Figure 8: Money Stock (% change and contribution) ..................................................................................................... 21 Figure 9: Nominal and Real Effective Exchange Rates (% change)............................................................................. 21 Figure 10: Fiscal balances (% of GDP) ............................................................................................................................. 23 Figure 11: Revenue and exp. aggregates (% of GDP) .................................................................................................... 23 Figure 12: Overall, recurrent and primary balances of the public sector (% of GDP).............................................. 23 Figure 13: Domestic debt composition (% of GDP)...................................................................................................... 23 Figure 14: Deficit financing sources (% of GDP) ........................................................................................................... 23 Figure 15: Real Effective Interest Rates – T bill actions versus CBM T bills ............................................................. 23 Figure 16: Income tax collections ...................................................................................................................................... 24 Figure 17: Commercial tax collections .............................................................................................................................. 24 Figure 18: Oil/gas tax and contributions outturn ........................................................................................................... 25 Figure 19: Tax and contr. exc. oil/gas and one off ......................................................................................................... 25 Figure 20: Recurrent expenditure outturn ........................................................................................................................ 25 Figure 21: Capital expenditure outturn.............................................................................................................................. 25 Figure 22: Spending growth (Index, 2009/10 = 100) ..................................................................................................... 26 Figure 23: Spending share (% of total ministry) .............................................................................................................. 26 Figure 24: ODA vs. Per capita GDP ................................................................................................................................. 29 Figure 25: ODA commitments and disbursements ........................................................................................................ 29 Figure 26: Project initiation and completion .................................................................................................................... 29 Figure 27: Distribution of project size .............................................................................................................................. 29 Figure 28: General Government revenue and expenditure ........................................................................................... 32 Figure 29: Medium-Term deficit target ............................................................................................................................. 33 Figure 30: Medium-Term public sector revenue ............................................................................................................. 33 Figure 31: Medium-Term tax and non-tax rev. ................................................................................................................ 33 Figure 32: Medium-Term public expenditure .................................................................................................................. 33 Figure 33: Medium-Term ministry spending .................................................................................................................... 33 Figure 34: Medium-Term capital spending ....................................................................................................................... 33 Figure 35: Functional spending allocations in Myanmar and LMICs .......................................................................... 34 Figure 36: Military spending adjustments over time ....................................................................................................... 34 Figure 37: Functional spending allocations in Myanmar and LMICs .......................................................................... 35 Figure 38: Defense personnel (% of labor force) ............................................................................................................ 35 Figure 39: Creating fiscal space and resilience ................................................................................................................. 36 Figure 40: Overall balance (Analytical vs. Accounting) .................................................................................................. 39 Figure 41: Capital expenditure (Analytical vs. Accounting ............................................................................................ 39 Figure 42: Contribution to growth (percentage point) ................................................................................................... 43 Figure 43: GDP per capita growth versus GFCF in LIC and LMIC countries, 2009-14 ......................................... 43 Figure 44: Public investment and public capital stock growth 1960-2009 (MICs) .................................................... 44 Figure 45: Public investment and public capital stock growth 1960-2009 (LICs)...................................................... 44 Figure 46: Public capital stock growth and economic growth 2009-2014 ................................................................... 45 Figure 47: Allocative and productive efficiency of capital spending ............................................................................ 46 Figure 48: Electric Power Consumption (KWh per capita) – Selected ASEAN countries ...................................... 47 Figure 49: Access to Electricity – Selected ASEAN countries (1990-2012) ............................................................... 47 Figure 50: Public sector spending growth in current prices 2012/13-2014/15.......................................................... 48 Figure 51: Share of capital spending 2011/12-2014-15 .................................................................................................. 48 Figure 52: Public sector capital spending by Ministries, States/Regions and SEEs .................................................. 48 Figure 53: Capital spending by Union Ministries 2011/12-2016/17............................................................................ 48 Figure 54: Sector allocations as share of ........................................................................................................................... 49 2 Figure 55: Sectoral spending as share of ........................................................................................................................... 49 Figure 56: Capital Budget Execution by Ministry, ........................................................................................................... 50 Figure 57: Sectoral Allocation of Foreign Loan and Grants in Union Capital Budget ............................................ 51 Figure 58: Number and Total Value of Large Projects .................................................................................................. 51 Figure 59: Capital Spending by States/Regions, 2012/13-2014/15 ............................................................................. 54 Figure 60: S/R capital spending/SR GDP vs. population density ............................................................................... 54 Figure 61: Per capita capital spending by States/Regions, 2014/15 ............................................................................. 55 Figure 62: S/R capex spending/capita vs. S/R GDP/capita, 2014/15 ....................................................................... 55 Figure 63: Allocations to Energy and Transport in State/Region Budgets, 2014/15 BE ........................................ 56 Figure 64: Average annual disbursement rate over 5 years for existing large capital projects, vs. project size ..... 61 Figure 65: Stalled projects among large projects ............................................................................................................ 63 Figure 66: Current number and status of SEEs ................................................................................................. 70 Figure 67: Flow of Funds between Union Budget and SEEs ....................................................................... 73 Figure 68: Gross Profit/Loss by sectors ........................................................................................................................... 81 Figure 69: Gross Profit/Loss drivers ................................................................................................................................ 81 Figure 70: Sector contribution to change in operating costs ......................................................................................... 83 Figure 71: Sector contribution to change in sales revenues ........................................................................................... 83 Figure 72: Operating Profit/Loss....................................................................................................................................... 84 Figure 73: Operating Profit/Loss by Sector ..................................................................................................................... 84 Figure 74: Operating Margin ............................................................................................................................................... 84 Figure 75: Net profits by Sector ......................................................................................................................................... 84 Figure 76: SEE Revenue, Expenses and Balance ............................................................................................................ 85 Figure 77: SEE payments to Union Budget and the oil and gas sector ....................................................................... 85 Figure 78: Current expenses financed out of different accounts .................................................................................. 86 Figure 79: SEE expenses covered by the Union Budget ................................................................................................ 86 Figure 80: Expenses covered by the Union Budget for SEEs with sufficient working capital ................................ 86 Figure 81: Expenses covered by the Union Budget for SEEs with insufficient working capital ............................ 86 Figure 82: SEEs Operating Profit/Loss by SEE category (million) ............................................................................ 87 Figure 83: Losses and Union Budget Support as percent of GDP .............................................................................. 87 Figure 84: Tax/GDP vs. GDP/capita (2010-2016 average).......................................................................................... 96 Figure 85: Tax/GDP across Regions (2000-2016) .......................................................................................................... 96 Figure 86: Tax revenue/Total revenue .............................................................................................................................. 96 Figure 87: Consumption tax/Total revenue ..................................................................................................................... 96 Figure 88: Customs receipts vs. Imports (2010-2016 average) ...................................................................................... 97 Figure 89: Evoluation of customs receipts vs. Imports in recent years ....................................................................... 97 Figure 90: Income tax/Total revenue................................................................................................................................ 98 Figure 91: Labor tax/commercial profits.......................................................................................................................... 98 Figure 92: The Dual Tax Triangle ....................................................................................................................................105 Figure 93: Estimating Tax Expenditures.........................................................................................................................109 Figure 94: Net profits by sector (LTO), FY 2015/16 ...................................................................................................110 Figure 95: Net profits as share of total revenues, top 10 sectors (LTO), FY 2015/16 ...........................................110 Figure 96: Estimated tax expenditures under various scenarios, FY 2015/16 .........................................................111 Figure 97: Tax Expenditures in Select Countries ..........................................................................................................111 Figure 98. Organizational Structure of the Treasury Department ..............................................................................118 Figure 99: External Public Debt Outstanding................................................................................................................123 Figure 100: Currency Composition of External Public Debt Outstanding, as end-March 2017 ...........................123 Figure 101. Present Value of Debt-to-GDP ..................................................................................................................124 Figure 102. Public Debt Service-to-Revenue .................................................................................................................124 3 Figure 103: Gross Debt for LICs (in % of GDP) ........................................................................................................ 130 Figure 104: Total Public Debt (in % of GDP) .............................................................................................................. 130 Figure 105: Change in DSA solvency indicators ........................................................................................................... 130 Figure 106: Change in DSA liquidity indicators ............................................................................................................ 130 Figure 107. Domestic Debt and Real GDP Per Capita: Frontier LICs, 2007 - 2014.............................................. 131 Figure 108. External Commercial Borrowing: Frontier LICs (in US$ million)........................................................ 131 Figure 109. Myanmar: Public debt redemption profile as of end-March 2017, MMK million ............................. 133 Figure 110. Risk Indicators ............................................................................................................................................... 136   LIST OF BOXES Box 1: Achievements and implementation challenges in education and health since PER 2015 ........................... 18 Box 2: Economic assumptions and policy priorities in the 2017/18 Union Budget ................................................ 28 Box 3: Gradual rebalancing of Defense spending in the Union Budget ..................................................................... 36 Box 4: Estimating the budget balance – Accounting vs. Analytical Methods ............................................................ 42 Box 5: Public Investment and Economic Growth – Selected Country Experience ................................................. 48 Box 6: Delivery and fiscal risks from Public Private Partnerships ............................................................................... 55 Box 7: Potential gains from transparent and efficient procurement ....................................................... 63 Box 8: International experiences on institutional setting for appraisal review and approval ................................... 69 Box 9: Global experience with SOE governance arrangements................................................................................... 75 Box 10: Performance monitoring of SOEs in Asia ........................................................................................................ 83 Box 11: Heavy Industry Enterprise No.1......................................................................................................................... 92 Box 12: Options for SEE financial performance monitoring in Myanmar ................................................................ 95 Box 13: Revenue reforms and results – Selected country experiences ...................................................................... 102 Box 14: Myanmar arrears clearance................................................................................................................................. 125 Box 15: Global trends in public debt sustainability ...................................................................................................... 128 Box 16: Myanmar’s First Medium Term Debt Strategy (2016) .................................................................................. 134 Box 17: Progress in establishing an efficient primary debt market in Myanmar...................................................... 139 LIST OF TABLES Table 1: MTFF functional allocation of expenditure ..................................................................................................... 33 Table 2: Policy options for strengthening MOPF macro-fiscal function .................................................................... 43 Table 3: Key macroeconomic indicators1......................................................................................................................... 44 Table 4: Fiscal Operations (% of GDP)1, 2 ....................................................................................................................... 45 Table 5: Breakdown of SEE capital spending (percentage points of total public capital spending) ...................... 56 Table 6: Summary of Findings from Myanmar Public Investment Management Assessment, 2016 ..................... 60 Table 7: Implementation Period and Estimated delays of nationally funded large hydropower projects ............. 64 Table 8: Cost estimate and Cost Increases of nationally funded large hydropower projects................................... 65 Table 9: Funding needed to complete ongoing large projects in Electricity and Energy; and Agriculture, Livestock and Irrigation ...................................................................................................................................................... 68 Table 10: Policy options for strengthening allocative and productive efficiency of public investment ................. 71 Table 11: Cross-country comparison of SOEs ............................................................................................................... 73 Table 12: Expenses covered by Own Account and Union Budget across SEEs ...................................................... 79 Table 13: List of Myanmar’s SEEs with key performance data for most recent fiscal year..................................... 80 Table 14: SEE financial performance and profit indicators .......................................................................................... 85 Table 15: Suggested adjustments to SEE performance indicators ............................................................................... 85 Table 16: Suggested SEE financial solvency indicators ................................................................................................. 90 4     Table 17: Balance Sheet Summary ..................................................................................................................................... 91 Table 18: Policy options for strengthening fiscal oversight of SEEs .......................................................................... 97 Table 19: General Government Revenue (share of GDP) ............................................................................................ 98 Table 20: Tax Revenue Collection by States & Regions / Agencies (% of GDP) .................................................. 101 Table 21: Myanmar - Summary of Number of Self-Assessed Taxpayers by Segment ............................................ 105 Table 22: LTO Tax Collections – 2013/14 to 2015/16 ............................................................................................. 106 Table 23: Tax Rates in Myanmar and neighboring countries (%) .............................................................................. 106 Table 24: Overview of tax instruments in Myanmar .................................................................................................... 107 Table 25: Tax Incentives in East Asia and Pacific ........................................................................................................ 112 Table 26: Base Case and Scenarios for Tax Expenditure Analysis............................................................................. 114 Table 27: Policy options for improving tax system efficiency .................................................................................... 119 Table 28: Total Public Outstanding Debt by Creditor Type ...................................................................................... 125 Table 29: Myanmar Arrears Clearance (Millions US$) ................................................................................................. 125 Table 30: Public debt cost and risk indicators, as of end-March 2017 ...................................................................... 133 Table 31. Macro-fiscal framework ................................................................................................................................... 135 Table 32: Policy options for strengthening public debt management ....................................................................... 144 5     ACKNOWLEDGMENTS This public expenditure review would not have been possible without the leadership of His Excellency Minister of Planning and Finance U Kyaw Win, His Excellency Deputy Minister of Planning and Finance U Maung Maung Win, Director-General (Budget) Daw Nwe Nwe Win, Director General (Planning) U Maung Maung Tint, Director General (Project Appraisal and Progress Reporting Department) Daw Thway Thway Chit and Director General (Internal Revenue Department) U Min Htut. Special thanks are also due to the MOPF counterparts for each of the chapters, for their close collaboration with World Bank technical team members in the preparation of this report. Their help is deeply appreciated. They are not, however, responsible for any of the conclusions in this report. Any errors are the responsibility of the authors. The report was prepared by a multi-donor team led by Habib Rab and Arvind Nair working under the guidance of Ulrich Zachau, Mathew Verghis, Abdoulaye Seck and Shabih Ali Mohib. The team is grateful for their ongoing support and advice. Chapter 1 on fiscal space and growth was led by Habib Rab. The team included Shabih Ali Mohib, Arvind Nair, and May Thet Zin. Chapter 2 on capital expenditure efficiency was led by Arvind Nair. The team included Pajnapa Peamsilpakulchorn, Quang Hong Doan, Martin Kessler and Tini Chatterjee (formerly of the Australian Department for Foreign Affairs and Trade). Chapter 3 on management on fiscal impact of State Economic Enterprises was led by Richard Record and Declan Magee (formerly of the UK Department for International Development). The team consisted of Pajnapa Peamsilpakulchorn and Soe Nandar Linn (consultant). The team benefited from consultations with Chunlin Zhang. Chapter 4 on tax system efficiency was led by Richard Stern. The team consisted of Arvind Nair, Thanapat Reungsri, William Chandler (consultant) and Frederic Tremblay (consultant). The team benefited from consultations with Andrew Minto (IMF), Steven Bowen (USOTA), and Colin Hutchins (consultant). Chapter 5 on sound public debt management was led by William Battaile. The team consisted of May Thet Zin, Lucy Pan and Ian Storkey (consultant). Peer reviewers for the PER were Enrique Blanco Armas and Evgenij Najdov. Review comments were also gratefully received from Morten Larsen, Edie Bowles, Yongzheng Yang (IMF), and Yasuhisa Ojima (IMF). The PER benefited from administrative support from Sandi Soe Lwin, and External Relations support from Kyaw Soe Lynn, Nurani Oktavia Robelus, and Shin Thandar. The PER was generously supported by the Australian Department for Foreign Affairs and Trade, the UK Department of International Development, and the Kingdom of Denmark through the Myanmar Multi-Donor Trust Fund. 6 ACRONYMS ADB Asian Development Bank MOPF Ministry of Planning and Finance ASEAN Association of Southeast Asian Nations MPT Myanmar Post and Telecommunication ATM Average Time to Maturity MSEC Myanmar Securities and Exchange Centre BEPS Base Erosion and Profit Shifting MTDS Medium Term Debt Strategy CBM Central Bank of Myanmar MTFF Medium-Term Framework CCTO Companies Circle Tax Office MTO Medium Taxpayer Office CIT Corporate Income Taxes OAG Office of the Auditor General CMD Cash Management Division OAS Official Administered System CS-DRMS Commonwealth Secretariat Debt OECD Organization for Economic Co-operation Recording and Management System and Development DICA Directorate of Investment and Company OTA Office Technical Assistance Administration DMD Debt Management Division PAPRD Project Appraisal and Progress Reporting Department DSA Debt Sustainability Analysis PDL Public Debt Law ECA Europe and Central Asia PDML Public Debt Management Law EIA Environmental Impact Assessment PER Public Expenditure Review FDI Foreign Direct Investment PFM Public Financial Management FY Fiscal Year PIM Public Investment Management GDP Gross Domestic Product PPG Public and Publicly Guaranteed GFCF Gross Fixed Capital Formation PPP Public Private Partnerships IBRD International Bank for Reconstruction PSCs Production Sharing Contracts and Development IIF Institute of International Finance R&D Research and Development IMF International Monetary Fund RMS Revenue Mobilization Strategy IPPs Independent Power Producers SAS Self-Administered System IRD Internal Revenue Department SCA Special Companies’ Act ITL Income Tax Law SECM Securities and Exchange Commission of Myanmar JPY Japanese Yen SEE State Economic Enterprises JVs Joint Ventures SEZ Special Economic Zone LIC DSF Low Income Country Debt Sustainability SGT Specific Goods Tax Framework LICs Low-income Country SGT Specific Goods Tax LMICS Lower Middle-income Country SOEs State-Owned Enterprises LTO Large Taxpayer Office STO Small Taxpayer Office MADB Myanma Agriculture Development Bank TAPL Tax Administration and Procedures Law MCDC Mandalay City Development Committee T-Bill Treasury Bill MDSA Myanmar Debt Sustainability Analysis T-Bond Treasury Bond MEB Myanma Economic Bank TIN Tax Identification Numbers MESC Mandalay Electricity Supply Corporation UNCTAD United Nations Conference on Trade and Development MFTB Myanma Foreign Trade Bank USD United State Dollar MIC Myanmar Investment Commission VAT Value Added Tax MICB Myanmar Investment and Commercial WB The World Bank Bank MICs Middle Income Countries WEO World Economic Outlook MIS Management Information Systems YCDC Yangon City Development Committee MMK Myanmar Kyat YESC Yangon Electricity Supply Corporation MOGE Myanma Oil and Gas Enterprise YOY Year On Year 7 EXECUTIVE SUMMARY INTRODUCTION Sustained growth will Myanmar had a strong economic take off between 2011 and 2015, but sustaining depend on improvements it will depend on improvements to public services and infrastructure. Yet General in services and Government spending at 15 percent of GDP is much lower than what is needed infrastructure. to deliver these improvements, and well below countries at a similar level of development that spend over 20 percent of GDP on public services. Fiscal space and resilience in Myanmar are limited by a small revenue base, spending inefficiencies, and limited financing options. This second Public Expenditure Review (PER) for Myanmar looks at options for expanding the fiscal space needed to improve public services and infrastructure. The first PER for Myanmar (2015) found that since the country opened up in 2011, it moved quickly to allocate considerably more resources to priority public services. This followed decades of some of the lowest Government spending in the world on essential social and economic services. The ramp up in spending was possible thanks to a big jump in revenue (6 to 12 percent of GDP) from an expanded tax base and higher gas receipts. Expenditure on education quadrupled during 2009- 13 and expenditure on health increased nine-fold during the same period, all while maintaining fiscal deficits at below 3 percent of GDP. Macroeconomic challenges But macroeconomic challenges in the past two years have contributed to from structural and policy deteriorating fiscal conditions. Part of these challenges are structural – Myanmar constraints have led to is dependent on commodity receipts, is prone to natural disasters, and has a narrow deterioration of fiscal production base. Recent shocks such as declining gas prices since 2015/16 and conditions. Cyclone Komen in the summer of 2015 increased risks to strong and stable growth. Growth moderated from an average of 7.8 percent between 2011 to 2015 to an estimated 6.5 percent in 2016/17, inflation peaked at 14 percent (year-on-year) in October 2015, and the Kyat depreciated by nearly 30 percent between 2015 and 2017. These challenges are exacerbated by policy and institutional capacity constraints. Fiscal buffers are limited by low revenue (10 to 12 percent of GDP), with considerable economic activity in either hard-to-tax sectors or dominated by small and micro enterprises. Myanmar has limited options for financing its deficit. When recent economic shocks widened the fiscal deficit, the Government turned to the Central Bank for monetary financing. This exacerbated inflationary pressures from the supply shock. Government has only recently initiated market auctions of Treasury bills and bonds, which will take time to reach volumes large enough to provide significant deficit financing and effectively manage liquidity and monetary pressures. 8 FISCAL SPACE FOR GROWTH Myanmar could reallocate In this context, what can Myanmar do to expand fiscal space for priority spending? resources to more It could reallocate spending from less important areas to priority areas, it could important areas, raise increase the overall fiscal envelope through Government revenue and/or public more revenue, or borrow. debt, or it could do both. The first PER (2015) focused heavily on reallocations both across sectors (for example, from general public services to social sectors), and within sectors (for example, from curative services to preventive care in the health sector). This second PER continues with analysis of the reallocation dimension of fiscal space while also devoting more attention to options for expanding the overall envelope. In line with this framework, the PER considers four options for expanding fiscal space. On the potential for reallocating resources, the PER analyzes (i) the allocative efficiency of capital expenditures, to identify options for reprioritizing spending to higher-valued use, and the productive efficiency of capital expenditures, to minimize waste in project implementation; and (ii) the fiscal impact of State Economic Enterprises (SEEs) to propose a strategy for the Government to maximize returns from and minimize subsidies to SEEs. On increasing the fiscal envelope, the PER analyzes (i) tax system efficiency, focusing particularly on short to medium-term priorities for reducing revenue leakage; and (ii) public debt sustainability, particularly looking at options for Government to more proactively manage debt to reduce the costs and risks to the public debt portfolio. There is scope to both The PER finds that over the medium term, Myanmar (i) could afford to finance a expand the budget fiscal deficit averaging 4 percent of GDP, but that slower-than-anticipated growth envelope and reallocate and fiscal slippages could lead to unsustainable debt; (ii) tax reforms should resources to fund priorities gradually help offset falling SEE payments to the Union Budget and downward over the medium term. adjustments from windfall earnings, but that total collections would remain below potential due to administrative capacity, and structural (for example, informal economy) constraints; (iii) declining SEE expenditure should be offset by increased spending by the General Government, but for this to enable higher Government spending on priority areas, SEE reforms should be pursued within a comprehensive framework; and (iv) increased spending on economic and social services would need to be accommodated in part through adjustments in Ministry of Defense allocations if Myanmar is to inch towards allocations that are closer to Lower Middle Income Country levels. But spending adjustments Even with the additional fiscal space above, medium-term spending adjustments will need to be gradual. towards public service and infrastructure priorities will be gradual. Capital investments are still projected at around 5 percent of GDP, well below needs. The size of General Government would remain below countries at comparative levels of income, as would allocations across important Government functions. However, the options presented in the PER provide the basis for an integrated fiscal strategy to both expand the budget and reallocate spending to priorities, and to do so in a way that helps build fiscal buffers and support longer-term growth. 9 The PER proposes to (i) develop a fiscal strategy that helps break a vicious cycle of low revenue, short-term financing, inefficient public spending, and SEE subsidies; (ii) significantly enhance the fiscal policy function of the Ministry of Planning and Finance (MOPF) building on existing achievements with the development and application of a Medium-Term Fiscal Framework (MTFF); (iii) integrate gas revenue forecasts as part of the MTFF to help better manage risk of revenue volatility and build fiscal buffers; and (iv) enhance fiscal transparency through improved internal data flows in MOPF, adopting internationally accepted definitions of fiscal indicators, and regularly communicating and publishing fiscal policy priorities to sustain market confidence. CAPITAL SPENDING EFFICIENCY Public investment is low Myanmar needs significant public infrastructure investment to address critical and falling, well below infrastructure gaps that are holding back private investment. Despite these needs, needs. public capital spending in Myanmar is low and falling. Capital spending is not sufficiently directed to growth priorities. Further, capital projects are executed inefficiently causing long delays in completing projects and cost overruns. The objectives of the PER were to identify policy options to improve: (i) the allocative efficiency of public investment decisions and (ii) the productive efficiency of public investment management processes. Public capital spending is low, by regional and global standards and has not seen a dramatic change since 2012/13. Union capital spending has to a limited extent been offset by States and Regions, whose share of public capital spending has increased from 5.9 percent in 2012/13 to 17 percent in 2014/15. Public investments are Public capital spending is allocated inefficiently (i) across sectors, as the share of union allocated inefficiently capital spending in transport and energy has declined from 31 percent in 2011/12 across sectors, by project to 24 percent in 2014/15. This is driven by inefficient budget execution. (ii) by scale, and in relation to allocations for public goods, as the SEE share of Government capital spending has risen public goods needs. from 12 percent in 2012/13 to 24 percent in 2014/15, with marked increases in heavy industry. This increases the risk that scarce public capital is being directed to projects that can be financed by the private sector. An assessment of capital expenditure allocations across regions is not possible due to data gaps on the geographic distribution of union capital spending. As a principle, efficient allocation of capital expenditure across regions requires adequate public investment in high-growth regions like Yangon and Mandalay to support continued growth. Imbalanced economic growth across regions is a feature in most countries, but should be balanced by inclusive development of other regions facilitated through increased integration, fiscal transfers and investments in service delivery among other policy measures. The PER chapter analyzes capital expenditure allocations from States’ and Regions’ own budgets -- while the allocations are not material when compared to regional GDP, they are directed to the transport sector reflecting expenditure assignments. 10 Weaknesses in public Public Investment Management (PIM) processes are inefficient, with weaknesses investment management throughout the key PIM functions. The critical weaknesses that worsen project lead to productive execution are (i) weak prioritization and appraisal, as project selection is ad hoc and inefficiencies. projects are not systematically appraised before entering the budget. Myanmar has made progress by developing appraisal guidelines, but critical gaps in regulation and appraisal capacity remain; and (ii) lack of multi-year funding mechanisms, with multi- year projects receiving funds annually through the budget process with no provisions made for carryover of unused funds. These process weaknesses are reflected in large time and cost overruns for projects, with a large number of stalled projects. The World Bank and MOPF jointly developed a large projects database to track these overruns, initially covering the Ministry of Agriculture and Ministry of Energy. Over two-thirds of large projects in these ministries are currently stalled, that is, with over eight years of implementation and less than 30 percent disbursement. Further, these stalled projects were absorbing close to 5 percent of the annual capital budget in 2014/15. The PER proposes to improve allocative efficiency by (i) in the short term, systematically compiling and reporting capital spending by Union Ministries in States and Regions; and (ii) in the medium term, gradually re-allocating capital spending by increasing relative spending on energy and transport, reprioritizing SEE capital expenditure towards public goods and adequately supporting investment needs in Myanmar’s growth poles. The PER proposes to improve productive efficiency by (i) in the short term, prioritizing budget resources for projects nearing completion, and strengthening implementation tracking for all large projects; (ii) in the medium term, finalizing and implementing prioritization and appraisal guidelines; and (iii) adopting policies to allow multi-year budget commitments for capital expenditure. FISCAL IMPACT OF STATE ECONOMIC ENTERPRISES SEEs have gone from State economic enterprises (SEEs) remain a major part of Myanmar’s Union being net contributors to Budget. However, SEEs are increasingly becoming a drain and therefore a the Union Budget to being constraint on fiscal space because of declining profits in industrial-, natural net recipients of subsidies resource-, and infrastructure-related SEEs, in particular. Managing fiscal pressures in the last two years. from SEEs is challenging due to complex institutional arrangements and weaknesses in data on SEE performance. The objectives of the PER were to (i) review the SEE legal and regulatory framework, focusing on SEE performance monitoring; (ii) analyze SEE financial performance, including transfers to and from the Union Budget; and (iii) identify policy options to improve fiscal oversight of SEEs. SEEs retain an important role in Myanmar’s budget, but their importance has declined over time. There are 32 SEEs in Myanmar, down from 44 in 2011/12, as some have been absorbed as administrative units in line ministries. SEEs still account for more than a third of Government expenses and receipts. However, SEE sales account for only around 7 percent of GDP, which is a lower share than in many transition and emerging market economies. 11 SEE activities are varied, covering public goods and commercial activities, across sectors ranging from natural resources to heavy industry. Ownership is largely decentralized to line ministries, with sector legislation playing a primary role in SEE governance. This creates conflict-of-interest concerns, as line ministries act as both owners and regulators, as well as fiscal risks since SEE operational decisions are made purely from a sector perspective with limited input from the Ministry of Planning and Finance. Fiscal flows to and from Fiscal flows between SEEs and the Union Budget are well defined. SEEs use two SEEs in relation to the accounts that are audited and reflected in the public sector balance sheet: (i) the Union Budget are clearly SEE account held at the Myanmar Economic Bank (MEB), which is used as the defined. conduit for Union Budget receipts and expenditures; and (ii) an Own Account, also in MEB, which holds the SEEs’ retained earnings that are used to finance a growing share of SEE expenses. SEE flows are highly concentrated -- the Myanmar Oil and Gas Enterprise, which is the largest SEE, accounted for 15 to 20 percent of union revenues between 2013/14 and 2015/16. Recent reforms have Government reforms have focused on greater SEE autonomy in order to reduce focused on increased SEE reliance on Union Budget transfers. Reforms have included (i) allowing SEEs financial autonomy of to retain more net profits; (ii) classifying SEEs as having sufficient or insufficient SEEs. working capital to finance recurrent expenditures, in order to differentiate the level of budget transfers; and (iii) not allowing SEEs to borrow from each other, except for SEEs that are outside of the Union Fund (e.g. SEEs under the Ministry of Transport and Communications can borrow from each other according to their law). These reforms have led to SEEs financing a larger share of their expenses from own accounts since 2013/14. However, SEE fiscal management is still impeded by (i) fragmented oversight with the SEE unit in MOPF focused solely on fiscal oversight at the exclusion of broader SEE management; (ii) high-volume but low-quality reporting by SEEs to MOPF that inhibits performance monitoring; and (iii) audited balance sheets that do not meet international standards, leading to insufficient understanding of fiscal risks. But success is limited SEE fiscal autonomy is now constrained by lower gross profits, which have without wider reform declined from 4 percent of GDP in 2011/12 to 0.7 percent of GDP in 2014/15. efforts and external factors This is driven by increased operating costs for infrastructure SEEs and declining leading to falling sales revenues for natural resource SEEs due to declining international commodity profitability. prices. Declining profits have led to falling transfers to the Union Budget, with SEEs becoming net recipients of union transfers since 2014/15. SEE losses have increased to 0.7 percent of GDP in 2015/16, which is implicitly absorbed by the Union Budget. 12 In the short term, the PER proposes to (i) adopt a new classification framework for Myanmar’s SEEs with four broad typologies (natural resources, commercially viable SEEs without need for public subsidy, commercially viable SEEs with need for public subsidy, and, non-viable SEEs); (ii) refine existing performance monitoring by adopting consistent accounting standards, automating data collection, simplifying reporting requirements; and publishing simple SEE performance monitoring report as part of the citizen’s budget; and (iii) determine customized contributions for individual SEEs based on financial conditions, future prospects, and current and future investment needs. In the medium term, the PER proposes to build a new policy framework for SEE ownership, regulation and management that moves away from decentralized ownership by line ministries. TAX SYSTEM EFFICIENCY Myanmar’s current Myanmar has one of the lowest levels of revenue mobilization in the world. In revenue collection at 10 to response, Myanmar initiated a comprehensive tax reform program in 2012, to 12 percent of GDP is well widen the tax base, redesign tax instruments and strengthen tax administration below its potential of 15 to capacity. Preliminary reform results are positive. However, the tax system is still 20 percent of GDP. characterized by revenue leakage, a narrow tax base and weak tax administration. The objectives of the PER were to (i) profile Myanmar’s revenue collection in relation to international comparators; (ii) review recent tax administration and policy reforms, including preliminary results achieved; (iii) analyze potentially big sources of near-term revenue leakages (for example, tax incentives and transfer pricing); and (iv) highlight policy options to address these leakages. Within low revenues, Myanmar’s General Government receipts at around 10 to 12 percent of GDP are taxes account for 60 well below potential, which is closer to 15 to 20 percent of GDP. Tax revenues percent of Government account for 60 percent of Government receipts. Corporate income taxes account receipts. for a third of revenue, which is significantly higher than the global average of 10 to 20 percent. Consumption taxes, which tend to be regressive, are a third of revenues, as in other developing countries. Customs revenues, as in other lower- income countries and lower-middle-income countries, have been declining due in part to generalized tariff reductions. Focus on tax Since 2012, Myanmar has tried to balance its short-term need for raising revenue, administration reforms to with its long-term objective of building an effective and equitable tax system. The balance the short-term focus has been on modernizing tax administration, which is a binding constraint to need for revenue with the other improvements, by (i) moving to self-assessment, where the initial burden of long-term objective of compliance and reporting is shifted to the taxpayer, as opposed to payments building an effective and negotiated by tax authorities. Large taxpayers are now self-assessed and audited, equitable tax system. while medium taxpayers are scheduled for self-assessment by 2018/19; (ii) and restructuring tax administration by taxpayer size rather than tax type. The Large Taxpayer Office (LTO), Medium Taxpayer Offices (MTOs) and Small Taxpayer Offices (STOs) have now been established. Tax policy reforms have Tax policy reforms have been incremental, due to limited administrative capacity, been incremental due to with no major new tax instruments introduced. The Income Tax Law (ITL) in administration capacity 2012 expanded the tax base, and reduced and rationalized tax rates. A Specifics constraints. Goods Tax (SGT) was introduced in 2015, similar to an excise tax, to cover alcohol, tobacco, teak, fuel, and natural gas. 13 Important tax reform Tax reforms have achieved notable gains, notably: (i) a filing rate for large taxpayers gains in terms of of over 95 percent; and (ii) the LTO has conducted a number of tax audits, expanding the tax base. unearthing information on economic activity that can be used to expand the tax base. These reforms are yet to have significant impact on revenue collections, which are weighed down by structural constraints including the reliance on volatile natural resource revenues. Nevertheless, the reforms have laid the foundation for a strong tax system. Looking ahead, Myanmar faces revenue leakages, especially from (i) tax base erosion on domestic and international transactions, arising from incomplete information on taxable activity, incomplete specification of the income tax law, and lack of provisions to allow tax administration to tax intra-company group activity; and (ii) provision of tax incentives. Myanmar offers a range of incentives, including tax holidays, that are insufficiently targeted to high social, low private return sectors. There is limited analysis of costs in terms of revenue foregone compared to benefits from additional economic activity. The chapter estimates tax expenditure scenarios from corporate income tax for large taxpayers, which highlights potentially significant tax base leakages from these incentives. To address tax base erosion, the PER proposes to (i) incorporate provisions in the income tax law requiring company groups to present consolidated tax returns, introducing arms-length principles, clearly defining tax residency, limiting interest deduction and loss-carry forward; (ii) strengthen information gathering such that the tax authorities can build business profiles; and (iii) on international transactions, to move cautiously in signing new tax treaties and, in the long term, to introduce regulations on transfer pricing. To reform granting of tax incentives, the PER proposes to (i) increase transparency by consolidating information on incentives provided to investors and publishing tax expenditure estimates; (ii) adjust the mix of incentives available and potentially eliminate tax holidays; (iii) ensure that incentives are based on cost-benefit analysis; (iv) consider placing the legal framework for tax incentives in the income tax code rather than in the investment law. SOUND PUBLIC DEBT MANAGEMENT Public debt management in Myanmar is undergoing transformation at a time when the country’s exposure to domestic and international debt is expanding rapidly. Since 2011, the Government has been modernizing public debt management functions and building institutional and human resource capacity to handle an increasingly complex public debt portfolio. The objectives of the PER are to (i) review the public debt management framework, including recent reforms; and (ii) analyze options for medium-term borrowing to help ensure the public debt portfolio is exposed to reasonable levels of costs and risks, including through expansion of the domestic debt market. Good progress in Debt management reform progress in recent years includes consolidation of public reforming the institutional debt management functions in MOPF, adoption of the Public Debt Law, framework for debt preparation of a Medium-Term Debt Strategy, launch of the Government management, but securities’ market, and increased coordination between debt and cash management important implementation functions. But despite progress in establishing basic institutional structures, the capacity constraints. Debt Management Division (DMD) faces important implementation capacity constraints. DMD is trying to implement big reforms while managing an increased workload with limited staff. 14 Myanmar can borrow to Myanmar is currently at low risk of debt distress, but is still vulnerable to risk in finance MTFF expenses, the event of slower than what is anticipated in the PER medium-term projections, but needs to pursue or fiscal slippages. Some of these risks are structural and exogenous, such as natural complementary reforms to disasters. The combined effect of such shocks on growth and public expenditure, expand fiscal space could quickly lead to significant deterioration in public debt sustainability. A weak outlined in the PER, and policy environment can exacerbate these risks. These include inefficient capital avoid risks from low spending that has little impact on growth, SEE inefficiencies that crowd out private growth and fiscal investments, and a narrow revenue base that leaves little buffer for economic slippages. shocks. Borrowing strategies have The PER considers different borrowing strategies to finance the fiscal deficit in the considered alternative MTFF, and the impact these have on the cost and risk of the debt portfolio. These options for deficit borrowing strategies include different splits between external and domestic debt, financing, providing a with the ultimate objective of eliminating Central Bank of Myanmar financing. strong basis to assess Understandably the strategy with high shares off external concessional funding alternative cost and risk outperforms other strategies. Strategies with higher domestic borrowing have a implications on public higher debt servicing burden due to higher interest rates and shorter maturity, but debt. at the same time are an effective way of supporting domestic debt market development. Enhanced transparency Myanmar is at the very early stages of a long-term journey in developing its and investor relations domestic debt market. Market participation is currently still below potential, which could support efforts to is not unusual at this early stage of development. Broader public finance and debt develop the domestic debt management reforms that support macroeconomic stability and growth are critical market in the near term. to expanding the domestic debt market. But there are also near-term steps that can be taken to ensure confidence in and expansion of the Government’s securities market, including allowing interest rates to rise, enhancing transparency, and strengthening investor relations. The PER proposes to (i) prioritize the upgrade to the debt Management Information System (CS-DRMS) to support improved data management and recording, which should also provide the basis for centralizing debt servicing functions and more timely and rigorous analytical work; (ii) expand training of middle office staff in analytical functions, which otherwise get crowded out by routine tasks; (iii) enhance the Medium-Term Debt Strategy (MTDS) on the basis of the joint WB- IMF Toolkit, submit the MTDS to Parliament concurrently with the Union Budget Law; and (iv) adopt regulations to limit central bank financing to specific emergency purposes and allow interest rates in Government securities to adjust with market conditions. 15 FISCAL SPACE FOR SERVICES AND GROWTH 1. Myanmar has embarked on major fiscal reforms to deliver public services and infrastructure, but faces constraints that hamper its investment and growth potential. Revenue reforms and growth have increased General Government receipts from 6 to 12 percent of GDP between 2009 and 2015. This together with budget reallocations have enabled higher spending on public services within a prudent fiscal stance, with the fiscal deficit averaging 2.4 percent of GDP per year between 2011 and 2015, rising to over 4 percent in 2016, and public debt levels at 33 percent of GDP in 2015. Nevertheless, a narrow fiscal base, vulnerability to shocks, and institutional weaknesses prevent public services and infrastructure to keep up with growing needs. 2. This Public Expenditure (PER) looks at options to create fiscal space and build fiscal resilience, which are essential to reaping the benefits of a high potential, frontier economy . This first introductory chapter sets the context and includes: (i) the motivation and framework for the PER; (ii) recent macroeconomic and fiscal developments; and (iii) a Medium-Term Fiscal Framework for balancing fiscal prudence with expanded public services and infrastructure. PER MOTIVATION AND FRAMEWORK 3. General Government in Myanmar is small relative to public service and infrastructure needs. Myanmar has one of the lowest levels of recurrent spending by General Government, including relative to other countries at a similar level of income (Figure 1). As discussed in the 2015 PER for Myanmar, General Government has historically had a small role in public service delivery. Many decades of underspending led to prohibitive household financing of critical education and health services, resulting in low access. Inefficient public investments left major infrastructure gaps, which are similar to Low Income Countries and significantly higher than those of other Lower Middle Income Countries (Figure 2), yet capital spending is low and falling. Figure 1: Recurrent spending vs. per capita Figure 2: Access to electricity and road density GDP Myanmar Sources: WDI, WB Staff estimates 4. A lack of fiscal space and resilience hamper Myanmar’s ability to absorb economic shocks These include the recent decline in commodity prices, exchange rate volatility in 2015/16, and Cyclone Komen in the summer of 2015. Policy and institutional capacity constraints exacerbate the impact of these shocks e.g. lack of financing options leads to deficit monetization and inflationary pressures; initial pressures on the exchange rate were met with restrictions on foreign exchange access, which fueled parallel markets; a policy of containing interest rate hikes hampers development of the domestic debt market. 16 5. These fiscal challenges impact negatively on Myanmar’s ability to continue attracting high quality private investments. Private investment flows have moderated in the past two years following a big spike in 2013 and 2014, when private investments averaged more than 20 percent of GDP, driven largely by gas and telecommunications. Investment commitments for manufacturing and non-commodity sectors have picked up, but actual inflows lag with investors citing public infrastructure bottlenecks and access to finance as major constraints to doing business. General Government investment to GDP for today’s Middle (and High) Income Countries during their period of infrastructure scale up was close to 10 percent of GDP or higher (Figure 3) compared to only 5 percent in Myanmar today. This played a big role in crowding in private investment, which over time make up a larger share of total investments (Figure 4). Figure 3: General Government Figure 4: Private investment/GDP (5-year moving investment/GDP (5-year moving average) average) Source: IMF Investment and Capital Stock Dataset, 2015 6. Enhanced fiscal space could help address some of these challenges. Fiscal space is defined as the availability of budgetary resources (from revenue, expenditure reallocations – including efficiency gains – and public debt) to spend on priority areas without sacrificing fiscal sustainability. For example, increased borrowing or one off capital receipts can temporarily provide budgetary resources, though may not be fiscally sustainable. On the other hand, borrowing based on a strategy that helps manage the costs and risks of borrowing decisions, or revenue collection based on sound tax policy and administration reforms, could expand budget resources in a fiscally sustainable manner.1 7. Myanmar could consider four possible sources of fiscal space: (i) reallocations from low to high priority spending within Government functions; (ii) reallocations from low to high priority spending across functions; (iii) higher revenue; and (iv) increased borrowing. Reallocation of spending across and within sectors could create fiscal space through efficiency gains. These gains should be considered in line with looking at options for fiscal space through higher levels of overall spending (that is, through higher revenue and public debt). 1There could be several ways to assess current levels of fiscal space in Myanmar (i.e. without measures that may expand future budget resources): tax audits can reveal the difference between actual taxes paid and what should have been paid; debt sustainability analysis can provide an idea of borrowing space; assessing discretionary vs. non-discretionary spending can reveal the amount of budget flexibility available, particularly for countercyclical fiscal measures 17 8. In line with this, the PER looks at four options for expanding fiscal space (Figure 5). On reallocations within and across sectors, the PER analyzes: (i) the allocative efficiency of capital expenditures, to identify options for reprioritizing spending to higher valued use, and the productive efficiency of capital expenditures, to minimize waste in project implementation; and (ii) the fiscal impact of SEEs to propose a strategy for the Government to maximize returns from and minimize subsidies to SEEs. On increasing the fiscal envelope, the PER analyzes: (i) tax system efficiency, focusing in particular on short to medium-term priorities for minimizing revenue leakage; and (ii) public debt sustainability, particularly looking at options for the Government to more proactively manage debt to reduce the costs and risks to the public debt portfolio. Figure 5: Proposed outline of PER 2 Capital expenditure State Public debt Fiscal space Economic Enterprises Tax policy Box 1: Achievements and implementation challenges in education and health since PER 2015 Fiscal space and resilience are necessary but not sufficient for expanding public services and infrastructure as there are also important implementation constraints. The volume of resources to be budgeted, spent, procured and accounted for has increased by over 260 percent in nominal terms between 2009 and 2015 – with the ministries of health, education and agriculture receiving the biggest increases. Whilst implementation issues are not the subject of this PER, we look briefly at achievements in health and education together with implementation challenges. Health sector Large increases in budget allocations for the health sector has helped improve service delivery. The Ministry of Health and Sports (MOHS) has: (i) expanded human resources, in particular at the frontlines; (ii) built skills of the primary health care staff to be able to provide basic emergency and obstetric care; (iii) invested in new primary health care infrastructure (e.g. rural health center and sub RHCs for midwives to operate out of); and (iv) increased operation and maintenance resources for existing health facilities. Increased budgets also enabled MOHS to take over the financing of traditional vaccines previously financed by development assistance. MOHS has formulated a strong National Health Plan, which lays out the path to extending access to a basic package of health services for everyone by 2021. It has also improved its information systems and knowledge base of health and nutritional outcomes and coverage. For example, Demographic and Health Survey, the first ever in the country, provided a solid basis to identify gaps and priorities for policy and interventions. Despite this progress, the MOHS continues to face many challenges related to public finance management. It is important to note that out-of-pocket spending (made by households at the time of seeking care) is still the main source of health expenditure, and Myanmar has one of the lowest global per capita (and GDP share) spending on health. 18 PFM challenges can be found all throughout the cycle. There is a disconnect between planning and budgeting functions. Budgets are incremental, with some consultation with line ministries to discuss needs and budget execution, which is considered for next year budget allocations. Budgets are typically not aligned with needs. There are several bottlenecks to budget execution, including: (i) lack of clear procurement guidelines and rules, which can create inefficient and expensive procurement (e.g. drugs); (ii) late timing of the supplementary budget leading to delayed implementation of the revised budget; (iii) dated financial rules, which are not fit for purpose, particularly around advances and travel allowances; (iv) inflexible budget rules which do not allow re-allocation of funds between budget lines during the year to reflect changing needs; and (vi) low capacity and quantity of financial management staff. Whilst there is a good financial reporting framework in place, it is fully paper based, administratively heavy, and slow. Reporting focuses on inputs and audit, rather than output or achievements. There is very little dissemination of the financial reports, or demand from upper/middle management for timely, quality reporting, and whatever reporting is produced is not regularly used for decision-making. As with budget allocation and execution, low quantity and capacity of financial management staff, particularly lower down the MOH structure, needs to be addressed to improve budget monitoring. Lastly, the disconnect between inadequate financial rules and tailored standard operating procedures for particular projects aiming to address the limitation of the rules are creating issues between townships and audit (particularly around travel allowance). Education sector Between 2011 and 2016, the education budget increased fivefold in nominal terms to help expand access to education services. Government introduced free and compulsory education starting with primary in 2011/12, middle school in 2012/13 and high school in 2015/16. It provided free textbooks and uniforms to all students and removed fees for registration, stationery and parent teacher association in all Government schools. It has also been providing school grants to all basic education schools in order to reduce the burden of school operating costs traditionally borne by communities. A stipends program targeting poor students enrolled in Grade 5 to 11 now covers 55 townships and more than 150,000 students. Over 8,000 new schools have also been built in the past 3-4 years. The increased budget has also been channeled to quality improvements. The Government has launched the restructuring of the basic education system from an 11-year system to a 12-year system plus one year of Kindergarten. Accordingly, a National Curriculum Framework (NCF) for Basic Education, which is aligned with the new KG+12 structure, has been developed with a strong focus on new content, improved teaching methodology and assessment. The Early Grade Reading Assessment (EGRA) has been carried out in 5 more regions since the 2014 pilot in Yangon. Furthermore, an Early Grade Math Assessment (EGMA) has been carried out on a nationally representative sample in 2017. This should help assess the impact of MOE initiatives on learning outcomes. Staffing is among the most significant constraints for implementation. There are a large number of vacancies at all levels including townships, districts and even in Nay Pyi Taw where director positions remain unfilled for long periods of time. The number of teachers trained in Education Colleges (about 7000-8000 a year) remains too low relative to demand. Daily wage teachers continue to be appointed to fill in the gaps. Recent reforms around Financial Rules and Regulations (FRR) have been positive. Prior challenges around the use of internal travel allowance within 5 miles of the Townships Education Offices have been resolved. The new FRR have added some welcome flexibility within codes (e.g. electricity can now be used for the electricity bill but also replacement of meter boxes). The main difficulty is that the FRR do not allow capital expenditures at the school level, which restrict the extent responsibilities over budget management and spending for schools. 19 RECENT MACROECONOMIC AND FISCAL DEVELOPMENTS Macroeconomic developments 9. Myanmar’s fiscal policy shifts in the last 5 years supported, and were the result of, a strong economic take off since the country opened up in 2011. Growth averaged 7.8 percent per year between 2011 and 2015. Reforms boosted private investment particularly in communications, transport and trade. Services on average contributed to around half of GDP growth in the past 4 years (Figure 6). Although public investment (5-7 percent of GDP) has been low, public consumption has accelerated rapidly, growing at 20 percent on average per year in real terms over the past 5 years. Industrial production has expanded thanks to a construction boom and an uptick in gas production. The gas sector accounts for around 7 percent of GDP, close to 40 percent of merchandize exports, and between 15-20 percent of General Government revenue. 10. Growth has moderated more recently to an estimated 6.5 percent in 2016/17 with important implications for the Union Budget. The slowdown reflects the effects of natural disaster, a narrow production base, and increased competition. Cyclone Komen in July 2015 negatively impacted agriculture. This affected food processing, which accounts for nearly two thirds of manufacturing output. The industrial sector in 2016/17 has slowed down across the manufacturing, construction and gas sectors. Therefore, despite positive trends on both the level and composition of growth, recent developments highlight ongoing vulnerabilities, including for the public sector as discussed further below. Figure 6: Real GDP (% change and Figure 7: Consumer Price Index (% change and contribution) contribution) Sources: MOPF, CSO, WB Staff Estimates 11. SEE production has also declined, negatively impacting Government revenue. SEEs make up less than 10 percent of GDP, though accounted for close to 30 percent of General Government receipts in 2013, declining to an estimated 20 percent in 2015. SEEs involved in commercial activities struggle to compete in Myanmar’s growing market place. The Myanmar Oil and Gas Enterprise is experiencing declining profits due to lower commodity prices. Others such as SEEs in the power sector have experienced a sharp increase in operating costs, due to exchange rate adjustments that have increased gas purchase costs. SEEs are now net recipients of Union Budget subsidies, having consistently been net contributors to the budget in past years. 20 12. Despite slowing growth, inflationary pressures have persisted, calling for increased attention to managing demand pressures including from the public sector . Inflation fell from its year-on-year (YOY) peak of 14 percent in October 2015, but continued to rise at 10 percent (YOY) in July 2016, and over 7 percent in February 2017 (Figure 7). Food prices, which account for 58 percent of the CPI basket, accelerated due to supply shocks linked to heavy flooding in the summer of 2015. Whilst second round inflation effects were generally limited, the gap between food and non-food inflation has more recently narrowed. This points to a more broad-based rise in prices expected in demand led inflation. Though data on wage inflation is limited, there are indications of upward pressures – not only for limited skilled labor, but also semi-skilled labor in manufacturing, and agricultural labor. This could enhance risks of upward wage-price spiral. 13. Annual growth in overall money stock (M2) has moderated in the past two years. Although it remains high, including due to fiscal responses to external shocks. Public sector financing needs have historically been a big driver of reserve money growth and M2 (Figure 8). Since 2012/13 there has been a slight rebalancing between private and public sector demand. This was in part thanks to growth in revenue that helped reduce Central Bank financing of the budget deficit, but also increased foreign capital inflows. CBM has adopted a monetary policy framework that targets reserve money growth, implemented through Government securities’ auctions, deposit auctions, and foreign exchange auctions. But as discussed below, these were not sufficient to offset pressures from increased monetization of the fiscal deficit in 2015/16. Figure 8: Money Stock (% change and Figure 9: Nominal and Real Effective Exchange contribution) Rates (% change) Sources: CBM, CEIC, WB Staff Estimates 14. Monetary expansion has contributed to external imbalances and pressures on the exchange rate. Myanmar has experienced widening current account deficits averaging 4 percent of GDP between 2012/13 and 2014/15, and rising to an estimated 4.8 percent in 2015/16. Whilst this is in big part due to increased foreign investment demand, in the last 2 years there has also been a decline in exports and slowing foreign investments, which have enhanced external vulnerabilities. The widening current account deficit, compounded by a general strengthening of the US dollar, has put downward pressure on the Kyat since mid- 2015. The Central Bank of Myanmar has maintained exchange rate flexibility and the Kyat depreciated close to 30 percent between May 2015 and February 2017 (Figure 9). This has also to some extent increased price pressures through imported inflation. 21 Fiscal balances and deficit financing 15. Macroeconomic developments have contributed to a deterioration in fiscal conditions. The public sector deficit has tripled from 1.1 percent of GDP in 2014/15 to an estimated 3.2 percent of GDP in 2015/16. It is estimated to have widened further to 4.5 percent of GDP in 2016/17 (Figure 10). SEE net losses were close -0.5 percent of GDP in 2015/16 (Figure 11). Other indicators also point to growing fiscal constraints, including a narrowing recurrent balance, and a widening primary deficit (Figure 12). 16. Fiscal pressures led to a sharp rise in monetization, which is reflective of economic policy constraints in Myanmar (Figure 13). Monetization had been on a declining trend since 2010, though still an important and steady share of financing due to a lack of either domestic or external financing alternatives. The legacy of monetization has caused the stock of CBM debt to average around 85 percent of the total outstanding domestic debt stock since 2005 compared to less than 20 percent on average for a sample of Low Income Countries in the early 2000s.2 The prolonged period of monetization in Myanmar stands in contrast to other countries where spikes in Central Bank financing is most often in response to fiscal crisis.3 CBM financing in Myanmar has averaged 2.7 percent of GDP per year over the last 10 years, peaking at 5 percent in 2015-2016. 17. Reducing inflationary CBM financing of the Union Budget is high on the Government’s policy agenda. It has recently adopted limits on CBM lending (30 percent of gross financing needs in 2017/18) with the objective of gradually eliminating it over the next three years. In addition, starting in 2016-2017, CBM lending are based on average discount rates determined at Treasury-Bill auctions (7-8 percent to date) and no longer at a flat rate of 4 percent. This should be a disincentive for monetization and is also consistent with development of the domestic debt market. 18. Efforts are also underway to develop the domestic debt market, which is important for gradually reducing monetization and supporting monetary policy independence. The Government has expanded Treasury-Bill auctions since early 2016-2017, including through the introduction of longer tenure bills (6 and 12-month bills). These should over time become an instrument for cash management and short- term liquidity. In addition, the Government launched Treasury-Bond auctions in September 2016. These instruments have longer maturity (2, 3, 5 years), which is important for public investments. 19. While domestic debt market development is a long-term process, near-term participation is below potential partly due to interest rate policy. Despite significant banking sector liquidity, successful bids on a purely competitive basis in T-Bill/Bond auctions are on average less than 50 percent of auction targets. Interest rates do not fully reflect market conditions because of concerns over rising debt servicing costs. Auction discount rates are close to (or below) bank deposit rates. Between Q1 2015/16 and Q1 2016/17, discount rates averaged negative 1.1 percent in real terms per quarter (Figure 15). Maintaining interest rates in line with inflation would likely encourage greater participation and support a lower inflation target by helping to mop up liquidity. Moreover, a higher interest rate at this stage is not likely to significantly impact debt servicing costs as market instruments are still a small share of net financing needs. The alternative of Central Bank financing exacerbates inflation, impacting negatively on the poor and private investments. 2 Bua, G; Pradelli, J; Presbitero, A, “Domestic Public Debt in Low -Income Countries,” World Bank Policy Research Working Paper 6777 (February 2014) 3 Easterly, W; Schmidt-Hebbel, K, “The Macroeconomics of Public Sector Deficits: A Synthesis,” World Bank Working Paper 0775 (October 1991) 22 Figure 10: Fiscal balances (% of GDP) Figure 11: Revenue and exp. aggregates (% of GDP) Figure 12: Overall, recurrent and primary Figure 13: Domestic debt composition (% of balances of the public sector (% of GDP) GDP) Figure 14: Deficit financing sources (% of Figure 15: Real Effective Interest Rates – T bill GDP) actions versus CBM T bills Sources: MOPF, WB Staff estimates 23 Government revenue 20. Revenue collections remain below their current potential and therefore have played small role in supporting fiscal adjustment (Figures 16 and 17). SEE commercial and income taxes have declined as a share of GDP in line with their operating conditions. Total commercial and income taxes (i.e. including from the private sector), which account for around two thirds of General Government receipts, have remained relatively flat as a share of GDP since 2013/14. In other words, revenue buoyancy has been relatively low when considering ongoing efforts to expand the tax base. This is partly due to the lag between reforms and higher revenue collection; revenue leakages discussed further in the revenue chapter; and declining receipts from the gas sector. 21. The drop in international commodity prices has negatively impacted gas receipts for the Union Budget. The natural gas sector is a significant contributor to union Government revenues, through royalties and state participation in production through SEEs. They are conservatively estimated at around 2 percent of GDP per year, or between 15 and 20 percent of Union Government revenues, between 2013/14 and 2015/16. Although tax receipts excluding gas and one off-payments are able to cover non-discretionary expenditure (wages, interest, pension), this is still an important share. 22. Natural resource revenues can pose particular challenges for fiscal management on account of their volatility and exhaustibility. Myanmar’s gas revenues have in the past been impacted by global price shocks. The price of Myanmar’s gas exports is indexed to heavy fuel and a variety of production cost indicators. Price is adjusted every quarter by taking the last 12 months’ average of these variables. This can help smooth short-term volatility. This has also meant a lagged impact of the sharp drop in oil prices in the summer of 2015 on gas receipts in Myanmar. Figure 16: Income tax collections Figure 17: Commercial tax collections Sources: MOPF, WB Staff estimates 23. Data from the last three fiscal years suggest a consistent overestimation of gas receipts in the Union Budget, which ordinarily could feed into unrealistic spending plans. Tax payments and contributions from state participation in the gas sector through MOGE was on average overestimated by around 30 percent in the Union Budget (Figure 18). This would have contributed to a 5-6 percent decline between budgeted revenue and actual. 24 24. At the same time, however, regular tax receipts (i.e. net of MOGE’s tax payments and net of one off payments (e.g. license fees)) were underestimated by on average 60 percent per year over the past three fiscal years (Figure 19). This in turn have added 30-40 percent more revenue than what was estimated in the budget. Since over-realization of regular tax revenue offset any under-realization in gas receipts, the latter did not necessitate overall expenditure adjustments in the Union Budget. As it happens, the Government did under-execute on recurrent expenditure (Figure 20), and on capital expenditure in the last two fiscal years (Figure 21), which led to lower fiscal deficits ex-post. 25. Nevertheless, the last two fiscal years have also experienced cuts to capital expenditure. This is not necessarily due to lower gas receipts. But all other things equal, recent volatility between budgeted and actual gas receipts could affect the quality of public investment. As discussed in the capital spending chapter, the average annual disbursement rate for large capital projects in Myanmar is less than 7 percent, which is driven in part by in-year reductions to budget allocations. This could very plausibly be linked to the impact of gas revenue volatility. Gas revenue are derived from a national asset, and is therefore almost like a capital receipt or financing item for creation of another asset. Myanmar has significant short term infrastructure needs. Channeling exhaustible resources to productive assets to fill these gaps is a sound trade off compared to saving them for future generations. Figure 18: Oil/gas tax and contributions Figure 19: Tax and contr. exc. oil/gas and one outturn off Figure 20: Recurrent expenditure outturn Figure 21: Capital expenditure outturn Sources: MOPF, WB Staff estimates 25 Expenditure allocations 26. Adjustments on the spending side have tried to protect allocations to priority social and economic services. In Myanmar, 6 out 22 ministries account for 80 percent of total ministry spending: Defense, Education, Health, Agriculture, Energy and Planning and Finance. Health Ministry spending has seen the biggest level increase since 2009-2010, followed by Education and Agriculture (Figure 22). Between 2009- 2010 and the 2016-2017 approved budget, spending for these ministries grew at a faster pace than spending for ministries as a whole. Their overall share as a result increased (Figure 23). 27. There are other important areas that are under-resourced. Myanmar has one of the lowest spending on social protection, at less than 0.5 percent of GDP, despite potentially large relative impact of social transfers on promoting social stability and poverty reduction.4 On economic services, public investment in the energy and transport sectors are well below needs. For Myanmar to achieve universal access to electricity by 2030, nearly US$2 billion per year of investments is needed in the power sector over the next 15 years.5 Figure 22: Spending growth (Index, 2009/10 = Figure 23: Spending share (% of total ministry) 100) Sources: MOPF, WB Staff estimates 28. There has been a rebalancing towards recurrent spending, in line with Government efforts to increase resources for front line service delivery. As a share of general Government spending, recurrent expenditures have gone from around 57 percent in 2013/14 to an estimated 70 percent in 2016/17. The wage bill (excluding the military) has on average accounted for 20 percent of total recurrent spending in recent years, which is reasonable compared to other countries. In 2014/15 and 2015/16, the wage bill increased by close to 30 percent in real terms per year. This was in part due to accelerated efforts to recruit new teachers and healthcare workers. The Ministry of Education accounts for close to 60 percent of the general Government wage bill, excluding the military. Other recurrent items that saw big increases include general purpose transfers for States/Regions, which went from 13 percent of recurrent spending share in 2013/14 to around 25 percent in subsequent years. 4 Collier, P; Hoeffler, A, “Aid, Policy and Growth in Post-Conflict Countries,” WB Conflict and Prevention Unit (2002) 5 The World Bank, “Energizing Myanmar: Enhancing Access to Sustainable Energy for All,” (2016) 26 29. The consistent decline in capital spending as a share of the budget and of GDP since 2012/13 is also the result of fiscal space constraints. Part of the decline reflects completion of public works associated with the establishment of Nay Pyi Taw. But more recently, cuts have been in response to fiscal constraints. Capital expenditures, and expenditure on goods and services, unlike spending on wages and salaries, could have a very significant positive impact on long-term growth rates.6 30. Of particular concern is the decline in allocations to physical infrastructure in core economic services. From 2011/12 to 2014/15, there has been a gradual decline in capital spending in the energy and transport sectors. The lack of sustained funding is detrimental, especially considering the multi-year nature of infrastructure projects and the network effect where output of one project is closely linked with overall service delivery. In addition, large capital projects are being de-prioritized.7 The average and total value of large projects has declined significantly since 1995, which may reflect de facto policy to halt new large projects after 2011 in order to put more emphasis on completing existing projects. 31. Among the existing large projects, there is growing evidence of implementation challenges and stalled projects. Stalled projects are defined as those that have been under implementation for at least 8 years and have disbursed more than 80% of total original estimated cost. An assessment of high value large projects highlights that 30 out of 39 large projects are currently stalled, which together account for 29% of the total capital budget. Factors driving slow progress on project execution include lack of strategic project selection, ambitious and rigid project design that is not appraised adequately, lack of multi-year capital budgeting allocation and delays in procurement. These issues are discussed in further detail in the chapter on capital expenditure. Box 2: Economic assumptions and policy priorities in the 2017/18 Union Budget The Union Parliament in March 2017 approved the 2017/18 Union Budget proposal. The 2017/18 Budget process took further steps towards gradual improvements in the policy content and transparency of the Union Budget. For the first time in Myanmar, the detailed budget proposals were published at the same time as they were submitted to the Parliament. Also published were the H1 2016/17 budget execution reports, SEEs’ commercial statements, the Financial Commission recommendation on the Union Budget, among other documents. These are not only important for enabling scrutiny of the Union Budget prior to its final approval, but provides a good indication of fiscal policy priorities. Macroeconomic assumptions: The Government projects real GDP to growth at 7 percent in 2017/18, following deceleration to an estimated 5.9 percent in 2016/17.8 Agriculture is expected to pick up, driven by an expansion in crop and livestock production. Industrial output is expected to accelerate the most, linked to a resumption of FDI, including in Special Economic Zones. This should contribute to services sector growth, particularly in transportation, trade, financial and tourism. Policy priorities: (i) exchange rate and price stability; (ii) increased tax collections and revision of tax exemptions policy; (iii) increased allocations for education, health and social welfare, as well as for rural development, electricity and water; (iv) infrastructure services to support investment and the private sector – prioritize implementation of ongoing projects and minimize the number of new projects; (v) strengthen procurement regulations. 6 Gupta, S. et al, “Expenditure Composition, Fiscal Adjustment and Growth in Low -Income Countries,” IMF Working Paper WP/02/77 (April 2002) 7 This analysis is based on data on large projects (above 2 billion Kyat value (about $1.5 million)) from the previous Ministry of Energy and Electric Power and Ministry of Agriculture, Livestock and Irrigation, which, together with Ministry of Construction, account for all large infrastructure projects in Myanmar. 8 Preliminary Government estimate, subject to change 27 Budget aggregates: The Union Budget projects deteriorating fiscal conditions. Aggregate revenue is expected to decline sharply from 17 percent of GDP (2016/17 BE) to 14 percent. Despite expenditure adjustments, the deficit is expected to increase to 5.7 percent of GDP based on WB GDP projections (6.1 percent based on Government GDP estimates).9 Part of this is driven by widening SEE deficits, estimated at 0.8 percent of GDP. Revenue: A large part of the drop in revenue is from a 2 percentage point decline in SEE revenue. This is linked to oil and gas revenue and closure of loss making factories, which also contributes to expenditure adjustments. Tax and non-tax receipts of the Union Government are projected to decline from 9.1 percent of GDP to 7.7 percent, with a particularly sharp drop in commercial tax collections. Expenditure: Deteriorating fiscal conditions have prompted recurrent and capital expenditure adjustments. SEEs have borne the brunt of the adjustments, though Union Government spending is growing due to health, education, and energy – with the latter expected to receive the biggest increase in 2017/18. Conclusion: Whilst the expansion of the deficit poses some concerns, it is also important to note that: (i) part of the increase is likely due to a pick-up in external financing for development projects, particularly in the energy sector; (ii) there is a recurring tendency to significantly underestimate non-commodity revenue as discussed above, which are expected to pick up with ongoing reforms; (iii) reforms to SEEs in the short-term are expected to be costly, particularly if some are shutting down whilst still requiring operational support. Source: Union Budget documents drawn from MOPF website Official Development Assistance 32. Overseas Development Assistance (ODA) commitments in Myanmar have picked up rapidly, whilst ODA disbursements are well below comparator countries and relatively small when compared to the Union Budget. ODA disbursements in Myanmar averaged around 2.4 percent of GDP in recent years, though this was driven initially by a spike in 2013/14 associated with debt relief (see chapter 5). ODA disbursement rates over the same period for countries at similar level of income averaged around 8 percent of GDP (Figure 24). 33. ODA commitments have grown much more quickly than ODA disbursements, which is to be expected given the nature of IDA funded projects. Available data shows ODA commitments to have risen to a cumulative $8.6 billion in the past 5 years, of which around $2.6 billion (30 percent) has been disbursed (Figure 25). ODA for direct project spending (as opposed to budget support) tend to have long lags between commitments and disbursements, particularly for investments in infrastructure, education and health.10 9 Based on analytical method of estimation not the Government’s accounting method. 10See Hudons, J, “Promises kept, promises broken? The relationship between aid commitments and disbursements,” (2013) 28 Figure 24: ODA vs. Per capita GDP Figure 25: ODA commitments and disbursements Myanmar Source: OECD DAC, WDI Source: Mohinga data base 34. Since 2011, there seems to be a gradual shift towards fewer, larger projects, which could further enhance the gap between commitments and disbursements. The number of new projects peaked in 2013/14 (Figure 26). Average project size has declined between 2015 and 2016 (from $13 million to $7 million), whilst median project size has been steadily rising (from $635 thousand in 2011 to $2.2 billion in 2016) (Figure 27). With normalization of relations with the international community, the number of ODA projects increased from 50 in 2009-2010 to more than 300 projects in 2013-14, falling back to 150 in 2015/16. Of the 1000 projects started since 2012, 600 are still under implementation and 400 have been completed. Available data indicates that average project duration is around 30 months. 35. The shift towards fewer, larger projects is consistent with the sector shift in ODA. In 2012, governance-related projects, which tend to be smaller in volume relative to infrastructure projects for example, drove the large increase in the number of new projects. The following years, however, these were overtaken by infrastructure projects, particularly in energy and transportation. As expected, disbursements in the latter have lagged significantly more than disbursements in governance or food aid sectors. Figure 26: Project initiation and completion Figure 27: Distribution of project size Source: Mohinga 29 36. The number of Development Partners (NGOs, multilateral, bilateral) has increased significantly, though a small number make up the bulk of ODA. Japan remains the largest donor (37% of committed funds), followed by the ADB and the World Bank (14% and 11% of projects in terms of committed funds on the 2012-2016 period respectively). This has implications for use of Government systems and the alignment of aid with Union Budget priorities. For simplicity, aid can be broken down into: (i) on-budget project grants; (ii) on-budget project loans; and (iii) off-budget grants (e.g. to CSOs, INGOs). On-budget project grants and loans would very roughly account for around 7 percent of Union Government expenditures, with ODA dependence across sectors varying significantly. 37. Efforts are under way to enhance aid management, starting with data compilation. Like other countries, however, having reliable data on aid commitments and disbursements is a significant challenge – therefore the above information needs to be treated with some caution. Aid Management Information System is being strengthened through two important initiatives: (i) the MIMU 3W (Who does What, Where) database, which has highly detailed project/village level information, though not spending information; (ii) the Mohinga database,11 which records project commitments and disbursements, directly following the standards set by the International Aid Transparency Initiative (IATI), at the sector/donor/project level. 38. Aid data coverage and comprehensiveness are limiting factors in integrating ODA into the Union Budget process. Mohinga data is self-reported by donors, but reporting standards, frequency, and quality vary across development partners. On-budget loan information is directly managed by the Treasury Department in MOPF with information directly provided by development partners and cross-checked against actual disbursement on a consistent basis based on loan contracts. On-budget grant information is sourced from line ministries (who receive information from development partners), but is not regularly cross-checked with the AIMS and other sources. Off-budget flow is lacking in the PFM systems at present 39. The recent establishment of the Development Assistance Coordination Unit (DACU) offers an opportunity to strengthen the strategic management of ODA. DACU is a high level unit chaired by the State Counsellor. It is tasked with reviewing ODA proposals before they are submitted to the Cabinet’s Economic Committee, to ensure that proposals are aligned with national development priorities. This is a critical function, which could benefit from a broader fiscal policy strategy as discussed further below to help ensure both alignment with development priorities but also clear division of responsibilities e.g. in relation to public debt management. MEDIUM-TERM FISCAL FRAMEWORK Baseline assumptions 40. With the above context, the potential for fiscal space is analyzed below within a Medium-Term Fiscal Framework (MTFF). The MTFF was developed as part of ongoing efforts to strengthen budget preparation. It is an analytical tool that combines top down fiscal aggregates (revenue, expenditure, and financing) with bottom up prioritization across ministries. It covers central Government agencies, including general purpose transfers to States/Regions, and State Economic Enterprises. The MTFF process is led by the MOPF Budget Department, though the PER analysis below was conducted by the World Bank. 41. The MTFF baseline scenario assumes some recovery in economic growth (Table 3), projected to average 7.1 percent per year over the medium-term (2017/18-2019/20). Private investment is expected to rebound, particularly in infrastructure. Inflationary pressures are expected to ease relative to 2015-2016, averaging 5.8 percent over the medium-term. The current account deficit is likely to remain large over the medium-term due to a combination of slowing gas exports, and increased investment-related imports. 11 http://roadto2015.org/mohinga-myanmars-home-grown-aid-information-management-system 30 42. The outlook is based on three overarching assumptions. The first is continued maintenance of fiscal discipline, which should help ease pressures on monetary policy. This is expected to help manage inflationary and exchange rate pressures despite growing external imbalances. The second is acceleration of structural reforms (e.g. energy, financial inclusions, business regulations, skills) to enhance productivity. This includes public investment to address basic infrastructure needs. The third is strengthened communication, predictability and stability of the Government’s macroeconomic and structural reform policies. These are critical to help anchor expectations around economic prospects and thereby sustain investor confidence. 43. Despite the relatively favorable outlook, there are important macroeconomic risks. Persistently low commodity prices could increase fiscal and external imbalances, and exacerbate financing pressures. Myanmar’s relatively narrow production base, increased exposure to competition, relative dependence on primary commodities, and vulnerability to natural disasters pose risks to stable growth. Lack of clarity or delays in policy implementation could prolong economic downturns. And a more challenging external environment may make it more difficult for Myanmar to take advantage of export markets. A combination of these factors enhance the variance around the medium-term baseline projections. Baseline projections 44. The MTFF tries to balance fiscal consolidation with increased spending on priority services. The analysis draws on options for: (i) expanding the fiscal envelope highlighted in the PER chapters on public debt and revenue; and (ii) reallocating resources through adjustments across and within sectors/line ministries proposed in the PER chapters on capital spending and SEEs. 45. First on public debt, the MTFF suggests that Myanmar could afford to finance a fiscal deficit averaging around 4 to 4.5 percent of GDP over the medium-term. This is based on debt sustainability and borrowing options analyzed in chapter 5. The public sector deficit is projected to consolidate from an estimated 4.5 percent of GDP in 2016/17 to 3.1 percent by 2019/20 (Figure 29). The medium-term public sector deficit is therefore projected to average 3.4 percent of GDP. This is below the 4 percent target as it allows some contingency and takes account of the pace of reallocation across ministries (see below discussion on reallocations). It assumes an uptick in external financing, as Myanmar implements growing external debt commitments for development projects. However, there are important risks: a one percent reduction in real growth or fiscal slippage equivalent to two percentage points of GDP could lead to unsustainable debt (see debt management chapter) 46. Second on Government revenue, the MTFF assumes that tax reforms should gradually help offset falling SEE payments and downward adjustments from one-off receipts. SEE revenues (net of transfers to the Union Government) are projected to decline from 7.1 percent of GDP to 6.6 percent, whereas general Government receipts are expected to pick up from an estimated 9.6 percent of GDP in 2016/17 to 10.2 percent in 2019/20 (Figure 30). 47. Tax revenue projections are relatively flat despite ongoing reforms for a number of reasons (Figure 31). Tax and non-tax projections take account of significant administrative capacity and structural (e.g. informal economy) constraints. Another reason is that declining SEE receipts also means reduced SEE tax and dividend payments, particularly from the gas sector, which is likely to suffer due to low international prices and declining production. To offset these, the PER revenue chapter proposes options to potentially plug large sources of revenue leakage (i.e. tax base erosion, transfer pricing, tax incentives), which warrant urgent attention, particularly as projected general Government receipts, currently at 10-12 percent of GDP, remain well below Myanmar’s potential of 15-20 percent of GDP. 31 48. Third on expenditure reallocations across the public sector, the MTFF projects declining SEE expenditure to be offset by increased spending by the general Government (Figure 32). For this reallocation to be effective in terms of enabling higher Government spending on priority areas, SEE reforms should be pursued within a comprehensive framework (see SEE chapter). Recent reforms allow SEEs to retain more of their profits so they can have more autonomy in allocating their resources. This could incentivize efficiency gains and generate benefits (or reduce losses) for the Union Budget. However financial autonomy alone, without broader ownership, private participation, corporate governance, and hard budget constraint reforms is unlikely to deliver efficiency gains. 49. Fourth on expenditure reallocations across ministries, the MTFF projects increased spending on economic and social services through adjustments in Ministry of Defense allocations (Figure 33). Any substantial reallocation would need to come from one of the six largest spending ministries, which are all engaged in strategic priorities: Defense, Agriculture, Energy, Education, Health, and Planning and Finance. The burden of adjustment falls on Defense, not because it is not a strategic priority, but because other areas are significantly underfunded relative to other LMICs (Box 3). Defense, and security more generally, is treated differently in PFM for valid confidentiality reasons. There may nevertheless be scope for some integration in PFM discussions to ensure that spending adjustments do not impact on ability to maintain security (Box 3). 50. Even with the additional fiscal space above, medium-term spending adjustments towards public service and infrastructure priorities will be gradual. Capital investments are still projected at around 5 percent of GDP (Figure 34), though the MTFF reallocates capital investments in Defense, Education, Health and Agriculture, which have been high in recent years, towards energy. The size of general Government would remain below countries at comparative levels of income (Figure 28), as would allocations across important Government functions (Table 1, Figure 35). This is expected and sensible as these are long-term fiscal adjustments. Figure 28: General Government revenue and expenditure Table 1: MTFF functional allocation of expenditure 2013/14 2014/15 2015/16 2016/17 B 2017/18 P 2018/19 P 2019/20 P Economic services 1.7% 1.9% 1.9% 1.9% 2.0% 1.9% 2.2% General services 1.8% 1.8% 1.7% 1.8% 1.9% 2.0% 1.9% Social protection 0.5% 1.1% 1.6% 1.8% 1.7% 1.6% 1.5% Defense 3.8% 3.8% 4.3% 3.4% 3.1% 2.9% 2.6% Education 1.7% 1.9% 2.1% 1.9% 1.9% 2.1% 2.5% Health 1.1% 1.1% 1.1% 1.0% 1.0% 1.2% 1.4% Sources: MOPF, WB Staff estimates, and IMF WEO (for Figure 28) 32 Figure 29: Medium-Term deficit target Figure 30: Medium-Term public sector revenue Figure 31: Medium-Term tax and non-tax rev. Figure 32: Medium-Term public expenditure Figure 33: Medium-Term ministry spending Figure 34: Medium-Term capital spending Sources: MOPF, WB Staff estimates Notes: Figures 33 shows the 5 largest spending ministries excluding MOPF Deficit estimates differ from 2017/18 Budget as these were adjusted based WB staff assumptions 33 Box 3: Gradual rebalancing of Defense spending in the Union Budget Compared to countries at similar levels of income, Myanmar spends a larger share of GDP on defense and a smaller share on health, education, social protection and economic services (Figure 35). This reflects a combination of a relatively small general Government and the crowding out of non- defense priorities in the Union Budget. Myanmar could benefit by a gradual rebalancing towards non- defense priorities. An extra Kyat allocated to economic or social functions could have more marginal impact on citizen's welfare, than an extra Kyat added to an already large defense allocation. Figure 35: Functional spending allocations in Figure 36: Military spending adjustments over Myanmar and LMICs time Sources: WDI, IMF GFS, and WB Staff estimates Such rebalancing of the Union Budget started in 2011/12 and is set to continue in the next three years. Myanmar has reduced the share of spending on defense from 40 percent of the Union Budget in 2010/11 to less than 25 percent in 2016/17 (Figure 36). At the same time, allocations to health and educations sectors more than tripled in the same period. The MTFF proposes to reduce defense allocations share further to 21 percent in 2019/20. Of note is that rebalancing is not due to a decline in nominal defense spending, which is expected to grow every year by 6-7 percent until FY 2019/20. But nominal spending in other sectors is set to increase even faster. Gradual rebalancing away from defense spending reflects international experience of countries that started with a relatively large share of defense spending. Countries with similar level of defense spending as a share of GDP when they embarked on peace process and/or re-engaged with international community show a gradual decline over a ten-year period (Figure 37). Myanmar seems to be following a similar trend. 34 Figure 37: Functional spending allocations in Figure 38: Defense personnel (% of labor Myanmar and LMICs force) Sources: WDI and WB Staff estimates Rationalization of defense spending is subject to similar principles as other Government spending.12 As with other Government spending, defense spending plans should be affordable, evaluated against other sectors, and effective in achieving planned objectives. Myanmar uses the MTFF to ensure that spending is affordable and to balance competing spending demands. The budget formulation process provides an opportunity to review the effectiveness of spending. But there are some differences as well to consider while evaluating defense spending. Some confidentiality of defense spending may be desirable on account of the sensitivity of the information. But, other country experiences highlight that countries can exercise robust external oversight (by legislature and through audits) while maintaining confidentiality.13 POLICY OPTIONS Medium-Term Fiscal Strategy 51. In sum, there is scope to both expand the budget envelope and reallocate resources to fund priorities, but it is important to implement these in an integrated macro-fiscal strategy. The PER finds that whilst Myanmar can afford to borrow more, it remains at risk of unsustainable debt if economic growth is marginally slower or the fiscal deficit is slightly larger than what is projected in the MTFF baseline scenario. Growth could be significantly affected by the efficiency of capital spending, SEE operations and the tax system, and by public debt sustainability. Similarly, the risks of fiscal slippages are greater with inefficient capital spending and SEE operations, a narrow revenue base, and unsustainable debt. Therefore, integrating these elements in a comprehensive strategy that lays out what and how the Government will deliver fiscal space for growth enhances chances of implementation and success. 12 “Securing development: public finance and the security sector”, World Bank (2017). 13 Ibid. 35 52. At the core of this strategy would be to break a vicious cycle of low revenues, short-term financing, inefficient public investments, and SEE subsidies (Figure 39). Myanmar has significant potential to expand revenue collection from current levels of 10-12 percent of GDP to 15-20 percent of GDP by further expanding the tax base and reducing revenue leakage. This should help reduce reliance on short- term financing, particularly if complemented with efforts to more proactively manage public debt and diversify the public debt portfolio. This is essential to match debt maturity to long-term public investments, which could see significant efficiency gains with better allocation and implementation. This includes greater control and oversight over SEE investments, particularly those engaged in commercial activities, which are now facing growing losses and increased subsidies from the Union Budget. Figure 39: Creating fiscal space and resilience • Narrow tax • Monetization • Expand tax • Diversify base • Limited long- base financing • Revenue term financing • Transfer • Med. Term leakage pricing Debt Strategy • Tax incentives Unsustainable Increased tax Sustainable Low revenue debt efficiency debt Inefficient SEEs Crowd in SEE public operating private subsidies investment competitively investment • SEEs • Misallocation • Enhance fiscal • Repriortize uncompetitive of investment oversight spending • Declining nat. • Weak project • Reform SEE • Accelerate resource rents implementation ownership implementation MTFF and MOPF Fiscal Policy function 53. The Medium-Term Fiscal Strategy should be anchored in the MTFF and led by a strong MOPF Fiscal Policy Analysis function. A strong fiscal policy function could help analyze, develop and update the Medium-Term Fiscal Strategy. There is therefore merit in strengthening the MOPF Fiscal Policy Analysis Unit so that it can undertake: (i) regular macro-fiscal monitoring reports; (ii) preparation of fiscal forecasts as part of the MTFF; and (iii) coordination with other economic agencies on data sharing. 54. The PER also explored options to further strengthen the MTFF, which provides the ideal basis for developing MOPF’s Fiscal Policy function. Specific suggestions include: (i) Further deepening of analysis across different components of the MTFF (e.g. macroeconomic assumptions, revenue forecasting, expenditure analysis, budget financing) through division of labor in the MTFF team, whilst ensuring close cooperation; (ii) Coordinating with ongoing IMF technical assistance to MOPF and the Central Bank on financial programming to ensure that up to date and consistent macroeconomic data feeds into the MTFF, and up to date and consistent fiscal data is provided to the financial program; (iii) Linking the MTFF to the Debt Sustainability Analysis (DSA), especially the tailored DSA prepared as part of the PER, to assess liquidity and solvency issues related to fiscal scenarios; (iv) Coordinating more closely with Internal Revenue Department on realistic revenue projections, and engaging with the Ministry of Electricity and Energy on gas revenue forecast assumptions and scenarios; 36 (v) Adopting standard fiscal classifications including for definition of the overall balance (see box 1) and netting out SEE payments to the Union Budget when calculating the consolidated public sector accounts. Gas revenue management 55. The MTFF provides a good basis to manage the impact of resource revenue volatility discussed above. The Government could consider integrating gas revenue forecasts in its MTFF. The Ministry of Electricity and Energy has gas revenue forecasts, which as a first step could be reviewed with MOPF during budget preparation. The review could cover the impact of different price, production and exchange rate assumptions on gas revenues, and implications for spending adjustments. Publishing these forecasts and underlying assumptions could be an important contribution to ongoing fiscal transparency efforts. 56. The Government could target fiscal benchmarks to guide the MTFF. These are not legally binding, unlike some more formal fiscal rules. Several resource-rich countries have experimented with fiscal rules to promote fiscal discipline, but have met with mixed success14 -- they are seen to be effective in countries with already strong fiscal discipline but are less effective in constraining spending in others. Furthermore, in countries with limited data and lack of precise forecasting, as in Myanmar, a rigid fiscal rule may result in inappropriate fiscal policy responses 57. One option could be to target the non-oil and gas balance (i.e. expenditures minus non-oil and gas revenues). This would involve targeting a specified non-oil and gas balance level over the MTFF forecast period – with the level set to ensure that the non-oil and gas balance can be financed in a sustainable manner. Any increase in expenditure or decrease in non-oil and gas revenue that increases the balance above the target level, should, therefore, necessitate an expenditure cut. Conceptually, this involves treating the oil and gas revenues as below the line financing of the non-oil and gas primary balance. This does not necessitate management of funds separate from the budget -- but simply a strong fiscal commitment to maintain the non- oil and gas primary balance at or close to the target level. 58. There has also been discussion on the establishment of a stabilization fund, which would be used to mitigate the impact of short term resource-revenue volatility through deposit and withdrawal rules. These rules are linked to certain “triggers”, usually resource price or resource revenue collected, which ensure a smooth profile of resource revenue on budget – a typical example is as follows: if actual gas price is higher than the expected ‘trigger’ gas price, any revenues associated with the excess gas price are deposited into the fund and not the budget; if actual gas price is lower than the expected ‘trigger’ gas price, any associated revenue shortfall in budget is covered by revenues withdrawn from the stabilization fund. 59. Stabilization funds are widely used in resource-rich countries, but, as with fiscal rules, with mixed success – with several countries facing the prospect of these funds being depleted quickly. 15 Some of the issues countries have faced in managing these funds include: Inaccurately forecasting resource prices and setting an appropriate trigger, owing to the volatility in prices, leading to triggers being set too low or too high over a period of time; lack of fiscal discipline which has led, in some countries, to insufficient revenue deposited in the stabilization fund when prices are high; limited capacity and transparency in PFM systems to manage budget, stabilization fund and interface between the two. These are issues that may need to be addressed if Myanmar decided to pursue a stabilization fund over the medium-term. 14 “Fiscal Management in Resource-Rich Countries” (Halland and Ossowski, World Bank, 2016). 15 See Halland and Ossowski (2016) for detailed survey of country experiences. 37 Fiscal transparency and communication 60. Myanmar has made excellent progress in enhancing the amount of fiscal information available in the public domain. Only six years ago, there was no information on the Union Budget in the public domain. Now however, the Government publishes among other things: the Union Budget proposal at the same time as it is submitted to the Parliament, in-year budget execution reports, Citizens’ Budgets, the Union Budget Law, the Union Tax Law, and financial information on SEEs. There is considerable reporting on Union Budget policies and discussions in the media, and Civil Society Organizations have started to engage actively on budget analysis, including at sub-national level. 61. Building on this progress, it would be important to prioritize transparency efforts in areas that are most likely to have a big impact on economic decisions and outcomes. Some examples are provided below. These are not meant to be exhaustive. They are highlighted more to illustrate the importance of prioritizing and sequencing fiscal transparency efforts based on need and demand. Progress on fiscal transparency depend on many factors including technical feasibility, institutional capacity, and political commitment. Therefore focusing efforts on ensuring that transparency reforms are prioritized according to local needs could have significant impact on economic decisions and outcomes: (i) Internal data sharing: Within MOPF there is scope to improve the completeness, timeliness and reliability of information on: financing flows (i.e. debt disbursements and servicing); pension liabilities; natural resource rents; and domestic revenue. These have affected the credibility of fiscal forecasts and the budget planning process. (ii) Data definitions: As illustrated on the different accounting methods for the Union Budget balance (Box 4), adopting a single definition that is consistent with international practice can be essential for avoiding confusion and presenting an accurate picture of fiscal stance. Another example is the need to net out SEE contributions to the Union Budget when consolidating public sector accounts. Not doing so overinflates the size of the public sector, which impacts on the accuracy of GDP by expenditure estimates. (iii) Reporting on macro-fiscal indicators, policies, and budget priorities: Quarterly (or six-monthly) publication of fiscal operations reports, balance sheet information, investment plans and future policy priorities could have a big impact on market and investor sentiments. Studies have shown how fiscal transparency can impact on fiscal discipline and borrowing costs.16 This could be particularly important for Myanmar’s efforts to develop its domestic debt market. Regular reporting can help anchor economic expectations and strengthen investor sentiments. Aid management 62. Building on existing progress with the Aid Management System (AIMS) by improving data compilation quality could support efforts to better integrate aid flows in the budget. One important step in this regard would be to create capability within the AIMS to disaggregate aid information into on-budget grants, on-budget loans and off-budget flows. At the technical level, there should regular cross-checking and verification of data to ensure consistency with what is reported by donors, what is contained in AIMS, and the information available to MOPF. 63. The newly established DACU offers a good opportunity to set the strategic framework for ODA. This could help: (i) clarify roles and responsibilities of different agencies in reviewing and approving ODA proposals; (ii) the role of ODA in efforts to create fiscal space for public services and infrastructure; (iii) principles of aid management, including efforts to increasingly use and strengthen Government systems. 16 Hameed, F, “Fiscal Transparency and Economic Outcomes,” IMF Working Paper, WP05/225 38 Box 4: Estimating the budget balance – Accounting vs. Analytical Methods Accounting method: The definition of the overall budget balance in Myanmar differs from standard definition. The Accounting Method, as it is referred to in Myanmar is the difference between total receipts (Government revenue and borrowing) and total payments (Government expenditure and debt amortization). Any deficit is covered by gross domestic borrowing. Analytical method: Revenue and expenditure items however should exclude financing operations, and the overall balance should correspond to the change in Government net indebtedness (i.e. new borrowing less repayment). This enables the Government to analyze changes in Government assets and liabilities. Capital receipts therefore also need to be subtracted from capital expense to arrive at net capital expenses (or net acquisition of non-financial assets). Myanmar definition International definition A Revenue (B+C+D+E) Revenue (B+C) B Tax Tax C Non-Tax Non-Tax D Capital receipts Expenditure (E+F) E Financial receipt Recurrent F Expenditure (G+H+I) Capital (G-H) G Recurrent Capital receipts H Capital Capital expense I Financial expense Balance (J-K) J Balance (A-F) Financial receipts K Domestic borrowing Financial expense The analytical method provides a more accurate picture of the fiscal stance. It enables the Government to have a more effective borrowing target. To date, the State Budget Law adopts a borrowing cap, which corresponds to gross domestic financing needs derived from the accounting method (i.e. row K). This however excludes domestic borrowing or foreign financing. Figure 40: Overall balance (Analytical vs. Figure 41: Capital expenditure (Analytical vs. Accounting) Accounting Sources: MOPF, WB staff estimates 39 Table 2: Policy options for strengthening MOPF macro-fiscal function Issue Policy options Fiscal space is limited by Short term: Building on the PER findings, develop a comprehensive fiscal a vicious cycle of low strategy that lays out how the Government will deliver fiscal space for growth. revenues, short-term Short to medium term: Further strengthen the MTFF through deeper analysis financing, inefficient of each component, closer coordination with the Central Bank on public investments, and macroeconomic data, linking the DSA to the MTFF and working with IRD on SEE subsidies realistic revenue projections. Medium term: Strengthen mandate and human resource capacity for the Fiscal Policy Analysis Unit to prepare regular macro-fiscal monitoring reports and fiscal forecasts, coordinate with other economic agencies and provide fiscal policy advice. Natural gas revenue Medium term: Integrate gas revenue forecasts into the MTFF. volatility impacts in-year Medium term: Consider fiscal benchmarks targeting non-oil and gas fiscal budget execution and balance in the MTFF to mitigate impact of volatility. creates pressures on the budget deficit. Fiscal transparency in Short term: Improve internal data sharing within MOPF on financing flows, certain areas can have a pension liabilities, natural resource rents and domestic revenue. These have big impact on economic affected the credibility of fiscal forecasts and budget planning. decisions and outcomes. Medium term: Adopt single, consistent and internationally comparable definitions for key fiscal data including the Union Budget balance. Medium term: Publish quarterly (or six-monthly) reports on fiscal operations, balance sheet information, investment plans and future policy priorities. This could have a big impact on market and investor sentiments. Aid management Medium term: Strengthen Aid Management Information Systems to becoming increasingly rationalize data compilation and improve data cross-checking and verification more complex with across development partner inputs and MOPF records. growing number of Short-term: Develop strategic framework for ODA. donors and level of commitments 40 Table 3: Key macroeconomic indicators 1 2013/14 2014/15 2015/16 2016/17e 2017/18p 2018/19p 2019/20p Output and prices (% change) Real GDP (% change) 8.4 8.0 7.3 6.5 6.9 7.2 7.3 Agriculture 3.6 2.8 3.4 4.3 5.4 5.3 5.3 Industry 11.4 12.1 8.7 4.5 6.5 7.2 7.3 Services 10.3 9.1 9.1 9.5 8.2 8.5 8.5 CPI ( % change, Period average) 5.7 5.9 11.4 8.9 6.3 5.7 5.5 Public sector (% of GDP) Revenue 20.2 22.3 19.9 16.8 16.7 16.3 16.8 Union Government 10.2 12.4 12.4 9.6 9.6 9.5 10.2 Tax 7.7 10.0 8.7 7.7 7.9 7.9 8.7 Non-Tax 2.4 2.2 3.4 1.4 1.3 1.2 1.2 Grants 0.1 0.3 0.3 0.4 0.4 0.4 0.4 Expenditure 21.7 23.4 23.2 21.3 20.5 19.6 19.9 Union Government 13.1 14.6 15.1 13.9 13.5 13.1 13.4 Recurrent 7.4 9.3 10.3 9.8 9.7 9.5 9.6 Wages 1.5 1.7 2.2 2.0 1.9 2.0 2.1 Transfers 1.0 2.6 2.7 2.4 2.2 2.0 1.9 Interest 1.1 1.1 1.0 1.1 1.3 1.4 1.4 Other 3.8 3.8 4.4 4.3 4.2 4.1 4.1 Capital 5.7 5.4 4.9 4.0 3.8 3.7 3.8 Balance -1.5 -1.1 -3.2 -4.5 -3.8 -3.3 -3.1 Public debt 34.2 29.5 33.8 33.8 33.8 34.3 34.7 Money and credit (% change) Reserve money 16.0 5.0 20.0 13.4 9.7 9.4 10.5 Net claims on Government 6.6 13.5 31.5 24.5 12.8 10.0 6.0 Broad money 32.7 19.6 23.2 20.8 16.5 18.6 18.0 Private sector credit 52.5 36.5 34.0 21.0 19.0 25.0 24.0 Balance of payments (% of GDP) Current account balance -4.9 -3.3 -4.8 -6.7 -6.8 -6.7 -6.6 Trade balance -5.1 -6.3 -8.6 -10.2 -10.4 -10.4 -10.2 Financial account 7.4 7.1 6.6 6.5 7.8 8.1 8.3 FDI (net) 4.4 7.1 6.6 5.9 6.4 6.7 6.9 Overall balance 2.0 1.8 -0.7 -0.3 1.0 1.3 1.8 Sources: MOPF, CBM, IMF BoP Statistics, WB Staff estimates 1/ e = estimates; p = projections 41 Table 4: Fiscal Operations (% of GDP) 1 , 2 2013/14 2014/15 2015/16 2016/17B 2017/18P 2018/19P 2019/20P Consolidated Public Sector Revenue 20.2 22.3 19.9 16.8 16.7 16.3 16.8 Expenditure 21.7 23.4 23.2 21.3 20.5 19.6 19.9 Recurrent 14.0 16.4 17.0 16.1 15.5 15.0 15.1 o/w Interest 1.3 1.3 1.2 1.4 1.5 1.6 1.5 Capital 7.6 6.9 6.2 5.2 4.9 4.7 4.8 Balance 3 -1.5 -1.1 -3.2 -4.5 -3.8 -3.3 -3.1 Net Financing 1.5 1.1 3.2 4.5 3.8 3.3 3.1 Domestic 1.3 1.2 3.0 3.9 2.3 1.7 1.2 External 0.5 0.3 0.6 1.1 1.5 1.7 1.9 Other -0.3 -0.4 -0.4 -0.5 0.0 0.0 0.0 Recurrent Balance 6.2 5.8 3.0 0.7 1.1 1.3 1.7 Primary Balance -0.1 0.2 -2.1 -3.2 -2.3 -1.8 -1.6 SEE Operations Revenue 13.0 12.6 10.1 9.1 8.9 8.6 8.2 Net of transfers to UG 10.0 9.8 7.6 7.3 7.1 6.8 6.6 Expenditure 11.6 11.5 10.6 9.3 8.8 8.3 8.1 Recurrent 9.6 9.9 9.3 8.1 7.7 7.2 7.1 Net of transfers to UG 6.6 7.2 6.7 6.3 5.9 5.5 5.5 Capital 2.0 1.5 1.3 1.2 1.1 1.0 1.0 SEE Balance 1.4 1.1 -0.5 -0.2 0.1 0.3 0.1 Union Government Revenue 10.2 12.4 12.4 9.6 9.6 9.5 10.2 Tax 6.7 7.0 6.7 6.4 6.6 7.2 8.0 o/w Income 3.1 3.4 3.2 3.2 3.3 3.6 4.1 o/w Commercial 2.9 2.8 2.9 2.6 2.7 3.0 3.4 Non-Tax 3.4 5.1 5.4 2.8 2.6 1.9 1.8 Grants 0.1 0.3 0.3 0.4 0.4 0.4 0.4 Expenditure 13.1 14.6 15.1 13.9 13.5 13.1 13.4 Recurrent 7.4 9.3 10.3 9.8 9.7 9.5 9.6 Wages 4 1.5 1.7 2.2 2.0 1.9 2.0 2.1 Transfers 1.0 2.6 2.7 2.4 2.2 2.0 1.9 Interest 1.1 1.1 1.0 1.1 1.3 1.4 1.4 Other 3.8 3.8 4.4 4.3 4.2 4.1 4.1 Capital 5.7 5.4 4.9 4.0 3.8 3.7 3.8 Union Government Balance -2.9 -2.2 -2.8 -4.3 -3.9 -3.7 -3.2 Recurrent Balance 2.8 3.2 2.1 -0.3 -0.1 0.0 0.7 Primary Balance -1.8 -1.1 -1.8 -3.2 -2.6 -2.3 -1.8 Sources: MOPF, WB Staff estimates 1/ B = Budget estimates; P = Projections 2/ Union Government is the equivalent of General Government. SEE operations are equivalent to public sector financial and non-financial corporations. The consolidated public sector aggregates Union Government and SEE operations (netting out SEE payments to the Union Government). 3/ Fiscal balances are calculated on the basis of net financing (GFS standards) 4/ Excludes military wages, which are captured in “Other” recurrent expenditure 42 CAPITAL SPENDING EFFICIENCY 1. Despite significant public investment needs, capital spending in Myanmar is low and falling . Recent fiscal pressures have prompted cuts in capital spending. Whilst this may help with short-term macroeconomic stability, it may be harmful for longer-term economic growth. This challenge is exacerbated by a lack of systematic prioritization, appraisal of projects, and long-term affordable financing. Some of these upstream challenges have fed into downstream implementation difficulties, including long delays in project completion and large cost overruns for projects. 2. This chapter reviews capital spending and project implementation in Myanmar in recent years, and presents options for improving public investment efficiency. It includes: (i) an overview of public investment and growth linkages including a summary of the framework for assessing the allocative and productive efficiency of capital spending; (ii) an analysis of capital spending allocations in Union and State/Region level budgets to assess allocative efficiency across sectors, geography, project scale and public good needs; (iii) a review of project implementation processes and outcomes to assess productive efficiency, focusing on project prioritization and appraisal, and the extent and drivers of time and cost overruns; (iv) policy options for improving the allocative and productive efficiency of capital spending. PUBLIC INVESTMENT AND ECONOMIC GROWTH 3. Myanmar has in recent years experienced an investment driven boom, similar to other countries in their early stages of economic take off (Figure 42). Much of this was due to private rather than public investment, which as discussed below remains low in Myanmar. Private investments in telecommunications, manufacturing, and other areas have accelerated thanks to a lifting of economic controls. But sustaining high levels of private investment will require an expansion of public investments in basic infrastructure (e.g. electricity, roads). This is critical to improving Myanmar’s productivity and competitiveness and therefore its attractiveness as an investment destination. Public investment can help crowd in private investment in productive sectors. Over time, private sector participation in public infrastructure could increase (e.g. through Public Private Partnerships, or outright privatization). Figure 42: Contribution to growth (percentage Figure 43: GDP per capita growth versus GFCF point) in LIC and LMIC countries, 2009 -14 Source: WB Staff estimates using Solow decomposition Sources: IMF Investment and Capital Stock dataset, Penn World Tables 9.0 43 4. The importance of capital formation for growth is evident in the experience of other Low and Middle Income Countries. Available data points to strong association between Gross Fixed Capital Formation (GFCF) – a measure of private and public capital investment – and economic growth (Figure 43). Today’s Middle Income Countries (MICs) experienced strong association between public investment levels and public capital stock accumulation between 1960 and 1990, which in turn translated into strong growth (Figure 44). Today’s East Asian MICs in particular grew faster than all other regions combined from 1965 to 1990, driven in large part by capital investment rates that were double the average for South Asia, Sub-Saharan Africa and Latin America.17 Today’s LICs on the other hand experienced relatively low levels of public investments and public capital stock accumulation between 1960 and 2000, and concurrently lower growth over this period (Figure 45). Figure 44: Public investment and public capital Figure 45: Public investment and public capital stock growth 1960-2009 (MICs) stock growth 1960-2009 (LICs) Source: Adapted from “Efficiency Adjusted Public Capital and Growth” (IMF, 2011) Note: t=current period; t-1=previous period 5. Public investment can have a positive short to medium-term impact on growth without necessarily affecting longer-term growth potential. This can be the case for fiscal stimulus-related investments. For example, counter-cyclical investments in response to the 2008 Global Financial Crisis resulted in growth of public capital stock, which was associated with a pick-up in GDP growth (Figure 46). The association was stronger in MICs than in LMICs or LICs. This could be due to weaker Public Investment Management systems in LICs, which leads to greater inefficiencies.18 The impact of fiscal stimulus over the long-term however is less clear.19 The need to offset a short-term shock can lead to poor sector or geographic targeting, which contribute to capital stock in the near term without affecting growth in the long-term. Public investment can also “crowd out” private investment by raising the cost of financing on account of unsustainable debt-financed public investment programs (IMF, 2017).20 17 “The East Asian Growth Miracle”, World Bank (1993). 18 World Bank, 2014:”The Power of Public Investment Management: Transforming Resources into Assets for Growth”. 19 WB, “Fiscal Policy Responses to the Current Financial Crisis: Issues for Developing Countries,” (December 2008) 20 IMF, 2017: “Public Investment, Growth, and Debt Sustainability”, available at: https://www.imf.org/external/np/res/dfidimf/topic2.htm 44 Figure 46: Public capital stock growth and economic growth 2009 -2014 Source: IMF Investment and Capital Stock dataset, Penn World Tables 9.0 6. The potential and limitations of public investment, as a means to stimulate and sustain economic growth are further reflected in the experience of individual LICs and LMICs (Box 5). These reiterate the critical role that public investment can play in stimulating growth and crowding in private investment, but also the risks posed by inefficient infrastructure that is financed by unsustainable means. All of these countries saw a major increase in public investments over a short period of time, motivated by the desire to accelerate public service delivery and private sector development. They faced similar constraints in terms of lack of sustainable financing options, capacity weaknesses, and also in some cases, recovery from conflict. Box 5: Public Investment and Economic Growth – Selected Country Experience Ethiopia: With growth averaging 10.9 percent per year in real terms between 2004 and 2014, Ethiopia was one of the fastest growing economies in Africa in the last decade. Growth was driven in large part by high rates of capital accumulation on a low capital base. Ethiopia had a high economic infrastructure deficit, with road access, electricity consumption and access to water being among the lowest in Africa in the early 2000s. Government re-directed spending towards investment, resulting in one of the highest public investment rates in Africa from 2004-14. Infrastructure spending was financed through the domestic banking sector and foreign borrowing, and partly executed through off-budget state economic enterprises. In order to sustain growth, Ethiopia faces a critical choice: should it continue to direct bulk of domestic financing towards infrastructure or should it allocate more to support private investment? Emerging evidence suggests that private firms are now more constrained in capital than in infrastructure. Source: Ethiopia’s Great Run: The Growth Acceleration and How to Pace It (World Bank, 2016) Vietnam: Between 2000 and 2010, public investment doubled as a share of GDP, from 4 to 8 percent. The goal, to increase access to infrastructure – and electricity in particular, was largely met: the country went from 14% connection to the grid to quasi-universal access in 2010. But this expansion came at a cost: public investment expanded so quickly that it soon became “unaffordable, inefficient and unsustainable” with investment financed in part by increasing domestic debt at high interest rates. This was aggravated by a highly decentralized public investment system. Annual financing needs for infrastructure were projected at above 20% of GDP, which, coupled with Government debt at 57% of GDP, raised fiscal sustainability risks. Vietnam has shifted focus from quantity, lowering the overall investment objective, to emphasizing on project selection, quality, and cost-efficiency – signaled by the new Public Investment Law enacted in 2015. Source: Vietnam Development Report 2012: “Market Economy for a Middle-Income Vietnam” 45 Korea: Between 1967 and 1970, Korean firms came up against significant infrastructure bottlenecks, especially in road transport. In response, the Korean Government engaged in a major construction program, culminating with the Seoul-Busan highway. Through a judicious use of cost comparison, competitive bids to discipline costs and high capacity for implementation, the program was implemented with little cost overruns and in advance of schedule. The highway program was financed initially by growing tax revenues but later through tolls on previously constructed sections. This infrastructure was, in turn, critical in stimulating and sustaining rapid economic growth in subsequent decades. Source: Warner (2014) “Public Investment as an Engine of Growth”, IMF Working Paper 14/148 Rwanda: Coming out of a major civil war in 1994, Rwanda faced significant reconstruction needs. Over the following two decades, this led to a major public infrastructure program. This was financed largely by external aid. The public investment helped support high levels of economic growth. From 2004-14, the level of public capital investment, at 14% of GDP, was double the level of private investment, and Rwanda’s GDP per capita grew 7.7% per year. There are now questions regarding the sustainability of investment on account of external aid dependence. To address this, the Government is working on diversifying financing sources, as well as balancing public and private investment. Sources: World Bank (2015), Rwanda Economic Monitor: “Financing Development”; Clark, Will and Birigir Anarson (2014), “Surging Investment and Declining Aid: Evaluating Debt Sustainability in Rwanda”, IMF Working Paper No 14/51 7. The marginal product of capital spending in Myanmar could be high given existing low levels of public infrastructure. Decades of under-investment in Myanmar have seriously impaired production and supply capacity in the economy. There is therefore significant potential for growth driven through capital deepening, as illustrated by the country’s recent growth acceleration. Diminishing returns to capital spending eventually sets in, as the additional output of each unit of capital spending begins to decline and attention needs to turn to increasing the productivity of the existing stock. But diminishing returns can also set in earlier and at lower levels of capital stock if public investments are not allocated or implemented efficiently. 8. Myanmar’s objective of filling a large infrastructure gap within a highly constrained fiscal environment therefore requires greater attention on the allocative and productive efficiency of capital spending (Figure 47). For public investments to support strong and inclusive growth: (i) the sector, geographic, and public goods allocation of infrastructure needs to be aligned with its highest valued use, which requires an understanding of the demand and production capacity in the economy (allocative efficiency); and (ii) waste has to be minimized so that Government can maximize the amount of infrastructure for each kyat of capital spending through strong public investment management systems and processes (productive efficiency). Figure 47: Allocative and productive efficiency of capital spending Input Output Outcome • Capital • Public • Economic spending infrastructure growth Allocative efficiency Productive efficiency 46 9. To help review the allocative efficiency of capital spending in Myanmar, the PER below looks at three dimensions: (i) capital allocations across sectors - to ascertain if the relative share of economic infrastructure, such as roads and energy, is in line with the high economic service needs of Myanmar; (ii) capital allocations across States/Regions – to assess whether economic infrastructure allocations are sufficiently supportive of ‘growth poles’ i.e. areas in Myanmar that account for a substantial share of economic output; (iii) allocations for public goods – in order to assess if resources are being diverted to economic infrastructure that could be financed by the private sector. 10. This is followed by a review of the productive efficiency of capital spending, which focuses on selected implementation issues. This is based on a joint assessment conducted in 2015 by the World Bank and the Ministry of Planning and Finance of Myanmar’s Public Investment Management (PIM) System. The assessment looked at strengths and challenges across the 8 stages of the PIM cycle: guidance, appraisal, independent review, selection, implementation, adjustment, operation and evaluation.21 The following areas identified in the assessment as critical constraints to the effective execution of capital spending are analyzed further in the PER: (i) upstream assessment of project prioritization and appraisal – to study in more depth the efficacy of prioritization and appraisal processes; and (ii) Downstream assessment of project implementation – to examine the extent and drivers of cost and time overruns in project execution. ALLOCATIVE EFFICIENCY OF PUBLIC INVESTMENT 11. Myanmar ranks very low in the provision of key economic and social infrastructure compared to other countries in the region. It has one of the lowest electrification rates in Southeast Asia (52 percent) and the average annual consumption per capita (164 kWh per person in 2013) is twenty times less than the world average (Figure 48 and 49).22 Myanmar’s road length is 150,816 kilometers with only 33,014 kilometers of paved roads (21.9 percent). It is estimated that around 20 million people, or half of the rural population do not have road access.23 Myanmar has low provision of social infrastructure such as health and education. Figure 48: Electric Power Consumption (KWh Figure 49: Access to Electricity – Selected per capita) – Selected ASEAN countries ASEAN countries (1990-2012) Sources: World Development Indicators and WB Staff estimates 21 “Myanmar Public Investment Management Diagnostic”, World Bank and Ministry of Planning and Finance (2016). Appraisal Document: National Electrification Project”, World Bank (2014). 22“Project 23 “Myanmar Transport Sector Policy Note”, ADB (2015). 47 12. Despite such acute infrastructure challenges, there has been a gradual reprioritization away from capital spending in recent years. The period between 2005 and 2011 saw a major ramp up in public investments linked to the establishment of Nay Pyi Taw. Since then, public investments have been growing more slowly than total and recurrent expenditures, and declining in real terms (Figure 50). As a result, public investment as a share of total spending has declined from 36 percent in 2011 to 26 percent in 2015/16 (Figure 51). This compares to around 10 percent for today’s MICs during their high investment periods. Figure 50: Public sector spending growth in Figure 51: Share of capital spending 2011/12 - current prices 2012/13-2014/15 2014-15 Sources: MOPF and WB Staff estimates 13. In addition to declining public investments, there has also been a rebalancing of capital spending across public sector entities. The States/Regions’ share in capital spending has grown from 3 percent of total capital spending in 2011/12 to 21 percent in FY 2014/15 (Figure 52); and SEEs’ share has also increased significantly – from 12 percent to 18 percent over the same period (Figure 52). The share of Union Government (i.e. general Government) however has declined as a share of the budget and of GDP (Figure 53), which has important implications for public infrastructure delivery as discussed further below. Figure 52: Public sector capital spending by Figure 53: Capital spending by Union Ministries Ministries, States/Regions and SEEs 2011/12-2016/17 Sources: MOPF and WB Staff estimates Composition of Union capital spending 48 14. Within the declining share of capital spending by the Union Government there has also been a rebalancing away from general public services and towards social services since 2009/10. Spending by the Education and Health Ministries has increased from 1 percent of capital spending in 2011/12 to over 10 percent in 2014/15 (Figure 54) and has been used to finance the construction of schools and health facilities, as well as equipment needed for service delivery. The Agriculture Ministry has also seen consistent increase, namely for irrigation facilities. Capital spending by the Defense Ministry has been consistently high, which is likely linked to a mix of military equipment and civil works, including rural infrastructure (Figure 54). Figure 54: Sector allocations as share of Figure 55: Sectoral spending as share of Union capital budget (2011/12-2014/15) Union capital spending (2011/12-2014/15) 35% 30% Share of Union capital budget 25% 20% 15% 10% 5% 0% Agri & Defense Education Energy Industry Transport Others Irrigation & Health & Mining 2011/12 BE 2012/13 BE 2013/14 BE 2014/15 BE Sources: MOPF and WB Staff estimates 15. Although budget allocations to economic services such as transportation and energy have also increased, budget execution has been low. The share of the capital budget allocated to energy has increased from 10 percent in 2011/12 to 13 percent in 2014/15 and for transport has increased from 23 percent to 28 percent over the same period (Figure 54). The share of actual spending in transport and energy however declined from 31 percent in 2011/12 to 24 percent in 2014/15 of capital spending -- with sharp decrease in transport from 23 percent in 2011/12 to 11 percent in 2014/15 and a stagnant share in energy of 12 percent (Figure 55). 16. The low level of budget outturn points to implementation challenges particularly for the transport sector. The transport (construction) sector has the lowest budget execution rate of all ministries – with education, health and defense being the best performers (Figure 56). The challenges in capital budget execution reflect weak project execution, resulting in time and cost overruns, which are discussed below. 17. The relative decline in union capital spending on economic services is a concern given their importance for growth in Myanmar. Despite a doubling in the number of vehicles since 2012, roads remain inadequate. This has led to two to three-fold increase in travel time in Yangon.24 Myanmar’s current transport investment level, from both public and private sources, is estimated at 1-1.5% of GDP and it is estimated that a further 2-2.5 percent of GDP may be needed over the next 15 years to build and maintain a transport network that can support sustained economic growth.25 24Ibid. 25 “National Transport Development Master Plan in the Republic of the Union of Myanmar”, JICA and the Ministry of Transport (2014). 49 18. Additionally, the lack of affordable and reliable power is a key constraint to service delivery for education, health and other services and is cited as a major constraint impeding private sector growth.26 Private investments in power generation have picked up modestly in the last 3-4 years, with Independent Power Producers (IPPs) accounting for 10 percent of generation in 2014/15 from 0 percent in 2011/12.27 However, it is estimated that investments in the energy sector, from the public and private sectors, will need to more than double from current levels in order to meet energy demand and support economic growth over the next 15 years.28 Figure 56: Capital Budget Execution by Min istry, (4-year average 2011/12 -2014/15) 120% 100% Actual/Budget 80% 60% 40% 20% 0% % of BE Executed 2014/15 % of RE Executed 2014/15 % of Supplementary Budget 2014/15 Sources: MOPF and WB Staff estimates 19. One of the challenges in implementing transport and energy projects, in addition to those discussed below, is the availability of long-term financing. Given large and lumpy cash flow needs on the one hand, and long gestation periods for investments on the other, Myanmar needs longer-term concessional resources to fund large transport and energy projects. As discussed further in the chapters on macro-fiscal developments and debt, current domestic debt instruments are not sufficient to implement large, long-term investments. 20. Spending out of growing foreign grant and loan commitments for economic infrastructure are expected to pick-up pace. External loans and grants comprise a relatively small but growing share of on- budget Union capital spending, increasing significantly from less than 1 percent in 2011/12 to 13 percent in 2014/15. External loans and grant disbursements have been directed largely to the industry and mining sector, specifically to commercial projects, rather than to energy and transport (Figure 57). Looking ahead, external loan and grant on-budget commitments, especially from large bilateral and multi-lateral donors such as JICA and the World Bank, show an increasing share of project commitments towards energy and transport projects. 26 “Energizing Myanmar: Enhancing Access to Sustainable Energy for All”, World Bank Policy Notes (2016). 27 Ibid. 28 Ibid. 50 21. Comprehensive analysis of on and off budget external financing is made difficult by incomplete data, but it is important to ensure that external financing is integrated into sector infrastructure plans. The Mohinga Aid Information Management System, launched in 2015, aims to centralize and collect information on all donor supported projects in the country. The system covers self-reported data on project commitments, disbursements, sector and location from development partners. But it has limitations of coverage, as it does not currently include project financing from large bilateral donors such as China and India. Looking ahead, it would be important to ensure that project financing from all external sources are captured and considered together with Government financed projects, within a comprehensive sector infrastructure strategy. This could help ensure that external and public domestic financing are complementary. Figure 57: Sectoral Allocation of Foreign Loan Figure 58: Number and Total Value of Large and Grants in Union Capital Budget Projects Sources: MOPF and WB Staff estimates 22. Private sector participation in economic infrastructure in Myanmar is currently limited but could play an increased role in the future. As in other LICs and LMICs, a combination of high risks, a lack of readily available financing for high fixed costs, and low economic returns makes it difficult to attract private financing for infrastructure. In Myanmar, private investment in transport infrastructure remains low, whilst private investment in power generation is slowly beginning to pick up. A key barrier for private investment in power generation is the lack of viability of investments, especially on account of tariffs being set at levels below cost recovery.29 23. In recent years, the role of Public Private Partnerships (PPPs) has become more prominent. Notable projects include the $250 million, 250 MW Myingyan Independent Power Producer Project.30 This is the first internationally tendered Independent Power Producer (IPP) in Myanmar (225 MW Combined Cycle Gas Turbine or CCGT Myingyan IPP), which resulted in price discovery that is likely to save the Government several billion dollars over the life of the concessions. The Government is currently developing a PPP policy that seeks to (i) develop a consistent approach to project identification, development, procurement and management and (ii) to clarify the roles and responsibilities for Government bodies at each stage in the process31. A sound legal and regulatory would be important to maximize benefits and minimize fiscal and delivery risks of PPPs (Box 6). 29 Ibid. 30 http://www.pppmyanmar.gov.mm/policy-and-guidelines 31 Ibid. 51 Box 6: Delivery and fiscal risks from Public Private Partnerships Countries tend to use PPPs for two major reasons – to enhance efficiency of project execution and as alternative forms of financing. The general expectation is that the private sector can execute projects more efficiently than the public sector, and thus PPPs in general are expected to increase capital expenditure efficiency. Secondly, PPP are sometimes used to ease fiscal constraints while ensuring adequate investment in essential infrastructure. In cases where Governments use PPPs as a financing mechanism, it is important to note that this still implies Government borrowing -- these additional resources need to be paid off in the medium and long term and are not “free money.” PPPs can deliver efficiency gains and provide an alternative financing source, but come with risks. This includes delivery risks arising out of poor demand forecasts and appraisal review, as with traditional public investment projects, and from poor risk allocation and contract management issues which are more specific to PPP projects. For example, in Scotland, the Skye Bridge PPP project faced significantly reduced demand due to lack of coordination with other crossings, which resulted in the Government buying back the whole project from the private partner. Fiscal risks also need to be considered. International experiences highlight potentially large fiscal costs due to poor contract design, optimistic assumptions about revenues from user fees, and minimum income guarantees provided by the Governments. For example, during the 2008 Global Financial Crisis, Portugal was forced to renegotiate its road PPPs when the calling of revenue guarantees by private partners threatened its fiscal position. From a fiscal point of view, it is important to establish whether a Government can maintain the same level of fiscal efficiency and sustainability through PPPs as through conventional means of implementation. Unfortunately, countries with limited experience in PPP projects, and even those with PPP experience, find it hard to calculate how much private involvement or Government liability will be efficient and sustainable. Some measures to help manage fiscal risks include incorporating all PPP fiscal commitments and risks into Government’s routine fiscal screening and monitoring process. For instance, Government could include these commitments in the medium- and longer-term fiscal framework, and report the known and potential future fiscal costs of PPPs in the annual budget. With regards to delivery, PPP requires careful oversight and regular audits. It is useful to develop and announce standard implementation guidelines for deciding procurement strategy, managing bid processes, developing model project agreement and standard clauses, issuing guidelines for output specification and for managing contracts. Source: IMF (2015) Making Public Investment More Efficient; and World Bank (2014) The Power of Public Investment Management (2014) 24. Linked in part to financing constraints is the declining size of large projects. The Project Appraisal and Progress Reporting Department (PAPRD) within the Ministry of Planning and Finance regularly monitors the progress of large projects from the Ministry of Electric Power and the Ministry of Agriculture, Livestockand Irrigation. Analysis of large project performance shows that capital allocation is moving away from large-scale projects – with the average value of projects, the number of projects and thus the total value of new large projects declining steadily (Figure 58). Since 2011 to 2014, there was an average of one large project entering the system each year. The project value in this period ranged from 2,663 million Kyat to 18,000 million Kyat. This may be reflective of financing constraints discussed earlier, where larger projects are fragmented into smaller packages that are easier to obtain financing for. This may also be the result of a de facto policy to halt new projects after 2011 in order to put more emphasis on completing existing projects. 52 Composition of SEE capital spending 25. SEEs account for a significant and rising share of public investment as noted above . This is largely driven by five SEEs, which in FY 2014/15, accounted for 79 percent of SEEs capital expenditure -- No.1 Heavy Industries Enterprise (29 percent), Myanmar Electric Power Enterprise (27 percent), Myanmar Posts and Telecommunications (8 percent), Myanmar Railways (8 percent) and Myanmar Oil and Gas Enterprise (7 percent). It should be noted that while a few SEEs account for most of the public investment, the scale of individual projects within these SEEs is not significant with a number of smaller works and projects. Table 5: Breakdown of SEE capital spending (percentage points of total public capital spending) 2011/12 PA 2012/13 PA 2013/14 PA 2014/15 PA Energy and Electric Power 5.0% 7.3% 9.1% 10.2% Industry 1.5% 12.8% 10.0% 6.9% Communication 1.2% 0.4% 3.5% 1.9% Railways 3.8% 2.7% 2.0% 1.7% Sources: Planning Department, WB Staff Estimates. Note: SEEs in this data are those agencies with SEE status as of 2014/15 and some of these classifications have changed since then 26. SEE capital spending is especially important for the energy sector. In FY 2014/15, the Ministry of Electric Power had the highest share of its capital budget spent by SEEs, at over 37 percent. The Ministry of Industry (31 percent), the Ministry of Energy (9 percent), the Ministry of Railways (8 percent) and Ministry of Communication (8 percent) also had a significant share of their capital budget spent by SEEs in their jurisdiction. 27. Despite the focus on economic infrastructure, SEEs spending is not directed sufficiently to projects that are focused on building public goods. As noted above, SEE capital spending in the Ministry of Industry remains significant. This spending is largely directed to support commercial activities and, as noted in the chapter on SEEs, these activities are not generating high economic return and most recently are turning in large operating losses. Further, there is little evidence to suggest that these activities have high social returns that could potentially compensate for the low economic returns. This suggests that a review of capital spending by SEEs may be warranted, with a view to ascertaining if SEE capital projects are directed towards projects that have high social return (with large positive externalities) and are not crowding out private capital investment in commercial activities. Composition of State/Region capital spending 28. Assessing the geographic distribution of capital spending is key to assessing allocative efficiency, but is hindered by the lack of data. There are two major sources of capital spending in States and Regions, which are: (i) capital spending by the union Government ministries in States/Regions and (ii) capital spending by the States/Regions themselves, drawing on own source revenues and fiscal transfers from the Union Government. Data on spending by Union ministries in States/Regions is not regularly and systematically collected, and is a critical data gap for understanding allocative efficiency. It limits the analysis of efficiency in regional allocations by the Union Government, and is also limiting in regard to the evolving discussion on fiscal decentralization, especially on expenditure and revenue assignments. Addressing this data gap requires changes in budget classification and reporting by union line ministries. 53 29. It is for example not possible to determine whether sufficient resources are allocated to support the public investment needs of growth poles such as Yangon and Mandalay. Growth poles are areas within a country with high economic potential and strong locational advantages. Globally, infrastructure investment in such growth poles increases growth potential by helping reduce congestion, reduce the cost of doing business, facilitating spill-over effects and thereby increase the productivity of private investment.32 Investing in growth poles also gives rise to concerns of imbalanced economic growth – international experience highlights that trying to correct these imbalances by looking to spread economic growth across regions is not usually successful and succeeds only in reducing the country’s growth prospects.33 A more feasible approach is to ensure inclusive development across regions on the back of imbalanced growth, through policies such as fiscal transfers, investments in service delivery and a focus on economic integration.34. Yangon and Mandalay are potential growth poles within Myanmar, on account of their large contribution to national GDP – in 2014/15, these two areas accounted for 32 percent of national GDP; had relatively high level of private sector development; and had potential for economic spill-over. 30. It is on the other hand possible to analyze States/Regions’ own capital spending, which is a growing, albeit relatively small share of public capital investment. Capital spending by States/Regions has increased four-fold in nominal terms from 280 billion Kyat in 2012/13 to 1.1 trillion Kyat in 2014/15, in contrast to a modest 5 percent increase in Union capital spending in the same period. This has been driven by increased funding available through interGovernmental fiscal transfer and is especially pronounced for poorer States/Regions such as Chin, Kachin and Kayin which have seen an over ten-fold increase in capital spending (Figure 59). Yangon and Mandalay have also experienced a doubling of capital expenditure from a higher base, and still account for 25% of capital spending by States/Regions. Figure 59: Capital Spending by States/Regions, Figure 60: S/R capital spending/SR GDP vs. 2012/13-2014/15 population density Source: MOPF, WB Staff estimates Source: MOPF, WB Staff estimates Note: Data for Nay Pyi Taw was not available. Note: NPT and Yangon excluded 32 “World Development Report 2009: Reshaping Economic Geography”, World Bank (2009) 33 Ibid. 34 Ibid. 54 31. Capital spending per capita highlights significant variance amongst States/Regions (Figure 60). In 2014/15, Chin spent close to 177,400 Kyat per capita in capital spending as contrasted with Bago, which only spent 8,300 Kyat per capita. Densely populated States/Regions like Yangon, Shan, Mandalay and Ayeyarwady have the lowest capital spending per capita (Figure 61). This reflects the nature of union fiscal transfers which are designed to provide more transfers per capita to less densely populated regions. This, in turn, reflects higher need to improve connectivity and higher investment costs in low density areas, and economies of scale and lower unit cost of service delivery in higher density areas. 32. State/Region capital spending in high growth States/Regions has also been relatively low, which may be due to higher levels of direct Union capital spending in those jurisdictions. As an illustration, Mandalay, which accounted for close to 10 percent of national GDP in 2014/15 had the second lowest per capita capital spending from its own budget. Yangon, which accounted for 22 percent of national GDP in 2014/15 had a per capita capital spend that was six times lower than Chin State (Figure 61 and Figure 62). This is positive if capital spending in poorer States/Regions are helping to strengthen connectivity with high growth areas. Figure 61: Per capita capital spending by Figure 62: S/R capex spending/capita vs. S/R States/Regions, 2014/15 GDP/capita, 2014/15 Source: MOPF, WB Staff estimates Source: MOPF, WB Staff estimates Note: Data for Nay Pyi Taw was not available. 33. Unlike the Union Budget, capital spending by States/Regions focuses on the transport sector, reflecting the nature of expenditure assignments. State/Regions have responsibility for addressing local economic services and specifically, the construction and maintenance of secondary roads (non-highways) and urban roads35. On the other hand, States/Regions have limited expenditure assignments when it comes to power generation, which is, in turn, reflected in a low share of regional capital spending dedicated to the energy sector. 34. In 2014/15, close to half of all capital spending by States/Regions was directed to the transport sector but with wide variance: in some like Ayarwaddy and Magwe, close to 80 percent of the States/Regions’ capital budget was directed to transport (Figure 63). In Yangon and Mandalay, the share was slightly lower at around 65-70 percent of total regional capital budget, with the majority of Yangon and Mandalay’s spending executed by municipal bodies such as Yangon City Development Committee (YCDC) and Mandalay City Development Committee (MCDC). 35 “Realigning the Union Budget to Myanmar’s Development Priorities”, World Bank (2015). 55 35. The share of capital spending dedicated to the energy sector is low, at less than 3% with some exceptions like Kayah state (Figure 63). It appears that the major proportion of energy spending in States and Regions are made by the Union Government. For example, investments in distribution and connections are made by YESC, MESC and Electricity Supply Enterprise through Union Budget. Investments in generation are in nature centralized (especially large-scale power plants) and the majority of which was also carried out by electric power SEEs, hence, geographical spending may not necessarily reflect direct the benefits of investment going into that region. 36. Despite the significant increase and positive allocation trends, however, capital spending by State/Region authorities is not material when it comes to overall economic activity. Capital expenditure from states and region budgets accounted for a small share of region GDP – ranging from 0.1% of GDP in Bago to 1.1% of GDP in Chin in FY 2014/15. This highlights that this source of capital spending, while important, is not material in magnitude when it comes to addressing the infrastructure needs at the subnational level. Figure 63: Allocations to Energy and Transport in State/Region Budgets, 2014/15 BE Source: Ministry of Planning, WB Staff Estimates. Data for Nay Pyi Taw is not available Note: Spending by YCDC and MCDC is estimated based on share of construction in regional spending. PRODUCTIVE EFFICIENCY OF PUBLIC INVESTMENT 37. The productive efficiency of capital spending depends in big part on Public Investment Management (PIM) systems and processes that impact on project preparation and implementation. PIM systems and processes impact on project selection, project design, adequacy of project financing, quality and speed of procurement, timeliness of delivery and other factors. In 2016, the Ministry of Planning and Finance with support from the World Bank conducted a review of Myanmar’s PIM systems and processes across 8 standard good practice functions. These functions include strategic guidance, appraisal, independent review, selection based on technical merit and priority, timely and cost-effective implementation, adjustment mechanism to respond to changes, proper operation and maintenance of assets and post evaluation to create feedback loop.36 A summary of the key findings from the joint MOPF and World Bank review is provided below. 36 “The Power of Public Investment Management: transforming resources into assets for growth”, World Bank (2014). 56 Table 6: Summary of Findings from Myanma r Public Investment Management Assessment, 2016 PIM System Function Assessment Findings Strategic Guidance Myanmar has a 20-year and 5-year development plan, and while the plans may be inconsistent in quality, they provide a basis for identifying priority areas for investment. However, currently, no guidance is provided to line ministries on priorities based on plans. Appraisal Appraisal function was reinstated since 2012, but, in practice, the function is curtailed by lack of a formal legal mandate for appraisal, limited scope, lack of adequate safeguards and limited capacity. As a result, projects are not systematically and comprehensively appraised. Independent Review Independent review by the central planning ministry of project proposals is currently limited, as with appraisal, by lack of mandate and capacity. As a result, the optimism bias inherent in line ministry proposals is not addressed. Selection based on technical Project selection relies on ministerial committees and does not rely on a merit and priority specified and consistent set of selection criteria. In practice, project selection has moved away from large projects towards small projects that are very thinly spread across the country. Timely and cost-effective Large projects suffer from long time and cost overruns, with a number implementation of projects suffering from continuously low disbursement. The drivers of this are complex, and include lack of effective multi-year budgeting and funding mechanisms, procurement delays and layers of approvals with limited delegation. Adjustment mechanism to Project progress is tracked for large projects, which provides an respond to changes opportunity to adjust in case of slow progress. However, in practice, adjustment in project design and scope is ad-hoc in nature without a systematic process in this regard. Operation and maintenance Some line ministries maintain fixed asset registers for completed capital of assets projects, but a full assessment of the comprehensiveness of these could not be completed. Further, the process for maintaining assets could also not be assessed comprehensively, but could be the focus of subsequent PIM assessments. Post evaluation for feedback Impact evaluation of capital investments, including a retrospective study loops of selection, design and implementation, is not currently being undertaken – and to do, would first require a significant strengthening of project monitoring systems. 38. Out of the list of issues identified, the following areas were picked as priorities in the near term for addressing implementation bottlenecks: strengthening prioritization, appraisal and improving implementation to reduce time and cost overruns. Addressing other PIM areas, without addressing these are unlikely to yield significant benefits. For instance, guidance would not be impactful unless selection processes are robust; and downstream activities of adjustment, operation and maintenance and evaluation, would yield limited benefits if the projects are not selected appropriately and are not completed on time. 57 Project Selection based on prioritization 39. Project selection is currently conducted in an ad hoc manner by ministerial committees, causing delays and insufficient scrutiny of projects proposed by line ministries. Project selection involves reviews by ad-hoc committees and relies on unclear prioritization criteria that may vary across projects. The lack of clarity on criteria and complexity of process also affects donor supported projects, which often suffer from approval and implementation delays. Almost all projects proposed by line ministries are typically approved with delays, and the only review and challenge function occurs at the late stage in parliament. 40. The net result of this selection process is a lack of prioritization in capital spending, leading to serious allocative inefficiencies. As discussed above, the lack of prioritization in Union capital spending is reflected in (i) inefficiencies in allocation, with inadequate resources for transport and energy projects (ii) increased fragmentation, with a relative increase in the share of capital spending directed towards smaller scale works and purchase of equipment, and further, a decline in the average value of projects and a proliferation in the number of projects. Project Appraisal 41. The Project Appraisal and Progress Reporting Department has stepped up its oversight function in managing public investment with the reintroduction of appraisal and strengthened effort on project monitoring. After over 20 years of the Government not undertaking any project appraisal, procedures for appraising projects were reintroduced in 2012. 42. However, actual practice in regard to appraisal is still severely constrained. Some of the challenges with regards to appraisal include: (i) Lack of formal mandate for appraisal: Currently, there is no formal legal mandate in place for ministries and agencies to submit project appraisals to PAPRD or to require appraisal approval to be a condition for funding. As a result, appraisal is done on a voluntary and ad-hoc basis and since it was reintroduced, only one ministry regularly submits project proposals to MOPF.37 (ii) Limited scope for appraisal: Appraisal scope, when appraisals are undertaken, do not include all the potential costs and benefits – specifically externalities, environmental risks and beneficiary consultation and feedback are missing. (iii) Limited capacity for appraisal: Practical experience of undertaking appraisals is very limited throughout most of the Government (although there is a small number of staff that have had experience working on donor-financed projects prior to 1988 and were exposed to appraisal procedures). (iv) Lack of adequate environmental and social safeguards: An important factor limiting the assessment of risks at appraisal stage is the absence of environmental and social safeguards, which increases vulnerability to costly mistakes. A broad legal framework on Environmental Impact Assessment (EIA) is included in an Environment Ministry Notification,38 which provides for the Ministry to determine categories of projects, business, service or activity which shall conduct EIA. The full adoption of an EIA system is still underway, with further regulations, significant capacity building and implementation required. 37In 2015, Planning received 14 appraisal reports from the Ministry of Agriculture, Livestockand Irrigation. 38Environmental Conversation Rules in the Ministry of Environmental Conservation and Forestry Notification No. 50/2014 dated June 2014, Chapter 11. 58 43. As a result, independent review and appraisal of projects is still lagging in Myanmar. Projects are proposed by line ministries with limited independent review in regard to strategic importance and appraisal of costs and benefits. This, in turn, leads to common problems of optimism bias and strategic misrepresentation that is inherent in any project proposed by a sponsoring ministry, and the lack of considerations of alternative project designs for achieving the project’s objective more effectively. Community views on the project which could help improve design of the project to maximize net benefits to the target population are also not effectively incorporated. Project Implementation 44. Inefficiencies in project execution in Myanmar starts with a lack of upstream project preparation. Ministries and agencies responsible for project execution identified inadequately prepared projects as a major hurdle, especially with unrealistic and inflexible project designs that, in turn, cannot be executed. The agencies also highlighted other issues during project preparation, such as a lack of well-prepared bidding documents to support procurement, a lack of resettlement planning and community consultation, which hurt project implementation down the line. 45. Inadequately prepared projects interact with weak budgeting processes to create challenges for budget execution. As highlighted earlier (Figure 53), budget execution rates are low especially for critical sectors such as transport. Discussions with line ministries highlight challenges in budget preparation that compound problems of inadequately prepared projects, such as (i) the lack of adequate information during budget planning with spending ceilings provided with some delays; (ii) absence of a sound cost estimation methodology; and (iii) variety of procurement procedures, which causes confusion. 46. The problems are further exacerbated by a lack of multi-year funding commitment mechanism for projects. Multi-year capital projects are still budgeted and executed on an annual basis, and, as per current financial rules and regulations, unused funds allocated to projects cannot be rolled over to subsequent years. In addition, slow and inefficient processes for issuance of project tenders result in project contractors being mobilized towards the end of the fiscal year. Contractors cannot make expected progress on a project in a given fiscal year, but, as funds cannot be rolled over across fiscal years, project activity is stalled for a considerable time in the next fiscal year and additional costs are incurred. This combination of lack of rollover and slow process of issuance of project tenders thus results in significant cost increases and time delays. 47. In addition, project execution is hampered by inconsistent and inefficient procurement practices. Procurement practices are not guided by a consistent policy with rules and procedures differing across line ministries. These rules tend to place a focus on integrity (control of corruption) and quality, but are not designed for efficiency and value for money. This contributes to time and cost overruns.39 Myanmar has now commenced significant reforms to improve public procurement, including updating procurement directives to harmonize procurement practices across spending agencies, with an eventual goal of adopting a Procurement Law that could apply to the whole public sector.40 Some immediate gains from efficient and transparent procurement can already be seen in ODA funded projects, as highlighted in Box 7. 39 “First Macroeconomic Stability and Fiscal Resilience Development Policy Operation, Program Document”, World Bank (2017). 40 Ibid. 59 Box 7: Potential gains from transparent and efficient procurement The World Bank financed National Energy Project’s off-grid component targets communities in rural areas. These communities are located far beyond the existing national grid and unlikely to receive grid access in the next 10 or more years. With assistance from the World Bank, the Department of Rural Development (DRD) of Myanmar takes ownership and responsibility for managing and implementing this project. The NEP covers partial costs of solar system and services, with the rest financed by the Myanmar Government and small contribution from beneficiary families. The contribution from families is estimated to be 10-20% of system cost, or USD 23 from each family. The Project also aims to install 19,000 public street lights and electricity connections for 11,400 health clinics, schools and other community buildings which will further positively impact these local communities that have had difficulty having access to higher quality social services due to their remoteness. The first procurement package for the supply and installation of solar home systems (12 lots by regions/locations) was launched through international competition in early 2016. The estimated budget was USD 64 million. After international and national advertisements, 354 bids for a total of 12 lots were submitted by the bid submission deadline. It took two days to open the bids and 3 months for DRD to evaluate them. The Bank issued its No Objection to contract awards within 4 working days. The contracts awarded at the total amount of USD 27 million. The result of the procurement was an estimated saving of USD 37 million. This was partially used for increasing the number of solar home systems for the benefit of more families and reducing the family’s contribution by 30%. The project procurement process saved time, resources and ultimately led to greater benefits for the families. The process also helped build capacity within the DRD to manage such complex procurements going forward, promising more savings in time and cost in the future. Source: World Bank blog (2017) entitled “A Success Story from Myanmar”. 48. Inefficient project execution in Myanmar is reflected in significant time overruns and low annual disbursement. The PAPRD in MOPF tracks disbursement for 39 large projects in the Ministry of Electricity and Energy, and the Ministry of Agriculture, Livestock and Irrigation, and find that the average annual disbursement rate was 6 percent with a range of less than 1 percent annual disbursement to over 20 percent (Figure 64). An average annual disbursement rate of 6 percent means that a large project could take up to 17 years to complete. Nine projects had average disbursement rates of less than 1 percent, which are unlikely to be completed, if they do not receive allocation on a consecutive yearly basis. 49. Low annual disbursements have also contributed to a large number of projects that are classified as ‘ongoing’ under implementation. The number of ‘ongoing’ projects has increased from 236 in FY 2011/12 to 468 in FY 2014/15, with a particularly large addition of new projects FY 14/15. Most of the ongoing projects are in the road sector. These suffer from low annual disbursement and, in some cases, a pause in funding to projects that are no longer considered priorities but are also not yet closed. 60 50. The case of hydropower projects under the Ministry of Electricity and Energy helps illustrate the problem of time overruns for large projects. Data on time and costs for 11 large hydropower projects implemented since 2002 through the national budget highlights significant deviation between originally estimated completion date and actual completion date (Table 7).41 For the three projects that were completed, the average time overrun is 2.6 years and average time taken to complete a project is 9.7 years. For 8 ongoing projects, the average estimated delay is 2 years and estimated time taken to complete a project is 10 years on average. Long implementation delays for hydropower projects result in critical delays in service delivery and also results in cost increases. Figure 64: Average annual disbursement rate over 5 years for existing large capital projects, vs. project size Source: PAPRD, WB Staff Calculation . Table 7: Implementation Period and Estimated delays of nationally funded large hydropower projects Estimated Estimated Installed Actual Original Revised or Actual Project Project Name Capacity Start Date Estimated End Date Delays Period (MW) End Date / Years /Years Upper Paunglaung (Completed) 140 2004-2005 2010-2011 2015-2016 5 12 Middle Paunglaung 100 2014-2015 2019-2020 2020-2021 1 7 Nancho (Completed) 40 2006-2007 2010-2011 2013-2014 3 8 Shweli -3 1,050 2010-2011 2019-2020 2021-2022 2 12 Upper Namtham 3 2014-2015 2016-2017 2017-2018 1 4 Dee Doke 40-80 2015-2016 2019-2020 2020-2021 1 6 Upper Yeywa 280 2008-2009 2018-2019 2019-2020 1 12 Tha Htay 111 2005-2006 2018-2019 2023-2024 5 19 Upper Kengtawng 51 2010-2011 2017-2018 2019-2020 2 10 Upper Baluchaung 30 2010-2011 2015-2016 2019-2021 5 11 Shwe Gyin (Completed) 75 2002-2003 2010-2011 2010-2011 0 9 Source: Department of Hydropower Implementation, Ministry of Energy and Electric Power, WB Staff Calculations 41 Data provided by the Department of Hydropower Implementation within the Ministry of Electricity and Energy. 61 51. Data on project implementation for large hydropower projects also highlights large cost overruns – in certain cases, an over four-fold increase from original budgets. Of the 11 large projects studied, only 3 were completed and actual cost deviation could be calculated (difference between original cost estimate and actual cost). For the other 8 projects, the projects are ongoing and only estimated cost increases (difference between original cost estimate and revised cost estimate) were calculated. The findings (Table 8) suggest that cost increases during implementation are common and range from 20 percent to over 400 percent. Drivers behind cost increases include extended project life, method and accuracy of cost estimation and uncontrollable factors which can vary significantly during long project life such as exchange rate and inflation. Table 8: Cost estimate and Cost Increases of nationally funded large hydropower projects Installed Total Project Total Project Cost (Revised Cost Capacity (MW) Cost (Original) Estimate) Increase (%) Upper Paunglaung (Completed) 140 121,396 312,268 157% Middle Paunglaung 100 193,325 169,157 -13% Nancho (Completed) 40 19,337 60,095 211% Shweli -3 1,050 903,343 1,412,368 56% Upper Namtham 3 19,706 25,666 30% Dee Doke 40-80 81,552 115,323 41% Upper Yeywa 280 192,257 359,689 87% Tha Htay 111 69,494 354,601 410% Upper Kengtawng 51 25,342 92,058 263% Upper Baluchaung 30 50,720 59,844 18% Shwe Gyin (Completed) 75 58,799 76,770 31% Source: Department of Hydropower Implementation, Ministry of Energy and Electric Power, WB Staff Calculations 52. Understanding the nature of stalled projects, and estimating the extent to which these projects are prevalent in the portfolio, is critical to improving overall capital expenditure efficiency in Myanmar. The problem of incomplete projects is widespread but estimating the size of this problem is hindered by the lack of singular identifying codes for projects, frequent changes in classification, and the lack of centralized data management on project monitoring. 53. In this regard, MOPF has recently developed a Large Projects Database, which also categorized projects as ‘on-track’ or stalled. The database currently covers 39 large projects, above Kyat 10 billion, from the Ministry of Electric Power and Agriculture, Livestock and Irrigation, and includes information on location, value, disbursements by year and project progress as reported by the line ministry. Stalled projects are defined as those which have been under implementation for more than 8 years. Preliminary discussions with the two line ministries highlight that the lack availability of timely financing, inappropriate design and lack of implementation capacity, are key drivers for these projects stalling. The discussions are not conclusive but reflect the need for MOPF to engage with the line ministries to understand these challenges and take action. 54. The database further classifies projects into three tiers, in order of priority for completion: (i) Tier 1 projects are those that have been implemented more than 8 years and have disbursed more than 80 percent of total project cost. Tier 1 projects are urgent priority for completion as resources have been heavily invested in these projects and they are considered well-performing. Among Tier 1 projects, those which have been implemented for more than 15 years are of particular concern. (ii) Tier 2 projects are those that have been under implementation for more than 8 years and have disbursed more than 30 percent but less than 80 percent of total project cost. Similarly, a significant 62 amount of resources has been invested in Tier 2 projects and they should be prioritized for completion. (iii) Tier 3 projects are those which have been under implementation for more than 8 years and have disbursed less than 30 percent of total project cost. Tier 3 projects may require a deeper review to understand the issues behind particularly low disbursement, and may require critical reprioritization to ascertain whether these projects are still in line with Government’s priorities. If not, some of these projects could be terminated so that resources could be freed up to fund other ongoing projects or new high priority projects. 55. In the large project portfolio, the majority of projects or 30 out of 39 projects are considered stalled projects while 9 projects are categorized as on-track. Among stalled projects, 12 are in Tier 1, 8 are in Tier 2 and 10 are in Tier 3 (Figure 65). Figure 65: Stalled projects among large projects On-track Stalled Projects (< 8 years of implementation) (> 8 years of implementation) Tier 1: Urgent Top performing priority for completion (12) On track Tier 2: Priority for completion (8) Review to explain below-average Tier 3: Require performance critical reprioritization (10) Source: PAPRD data and WB staff calculation. POLICY OPTIONS Addressing data gaps on union capital spending 56. The total capital expenditure by states and regions have been recorded and analyzed. However, sector wise data on geographic distribution of union capital spending is not available.. This prevents an understanding of geographic allocation of spending, and the extent to which this is aligned with infrastructure gaps, or connectivity needs. Available information on capital spending by States/Regions directly 63 out of their budgets is not sufficient for a comprehensive assessment, given that State/Region spending is only 20 percent of the capital budget. 57. Addressing this data gap could help improve the allocative efficiency of general Government capital spending and could be achieved without significant administrative burden. Selected line ministries already maintain data on spending across States/Regions (e.g. Ministry of Construction). Furthermore, budget execution data compiled by the Treasury Department and the Myanma Economic Bank, could also be used to determine the level of Union capital spending in States/Regions. Compilation of this data, including cross-checking to ensure accuracy, could be led by MOPF, and ultimately be incorporated into the regular budget reporting cycle. Reallocating capital spending to support growth and leveraging financing 58. The geographic, sector and public goods allocation of capital spending warrants review, as current expenditure patterns may run quickly into diminishing economic returns. The annual budget process offers the opportunity to take a strategic look at reallocating public investment resources to better support inclusive economic growth, in particular through: (i) Accelerating implementation of public investments in energy and transport sectors. Despite increased allocations to energy and transport, budget execution has not kept pace. Some of the issues for implementation are gradually being addressed by MOPF, including issuing spending ceilings in advance and standardizing procurement rules. While other issues, specifically adequate project preparation fall squarely under the responsibility of line ministries. (ii) Aligning public investments in support of growth poles. As highlighted, it is important that sufficient financing is directed to address public investment needs in rapidly growing regions of Yangon and Mandalay. In the absence of such investments, there is risk that growth in these regions will be constrained in the future. (iii) Rationalizing SEE capital spending and reallocate to public goods. In light of its increasing share and emerging evidence of insufficient focus on public goods, there is potential to rationalize capital expenditure by SEEs. In order to support such rationalization, MOPF could consider certain principles for capital investment by SEEs that are financed through the budget, including: 42 (a) Restricting capital expenditure to those SEEs that are expected to remain under public ownership for some time – specifically, global experience highlights that capital expenditure in SEEs that are expected to be privatized in the short to medium term is unlikely to generate sufficient returns through higher sale price; (b) Ensuring adequate financial oversight of SEE investments, to ensure that SEE financial reporting is comprehensive, up-to-date and is audited, with a view to ensure cost- effectiveness of capital expenditure; (c) Operations and maintenance expenditure required for capital assets should be covered by SEE own accounts – in general, the Union Budget should not finance operating expenditure associated with capital investments, as also highlighted in the SEE chapter; (d) Evaluating SEE capital investments as a part of the MTFF process and PAPRD appraisal review, against other budget proposals by line ministries, to ensure that these investments are expected to generate economic return and represent value-for-money. 42This section draws from a World Bank Public Expenditure Review conducted in Afghanistan in 2004, entitled “Afghanistan: Managing Public Finances for Development”. 64 59. In addition, the Government could look to increasingly leverage other sources of financing. External concessional financing is increasing and now being directed to priority sectors such as energy and transport. The Government could leverage this financing, along with other sources of financing from non- traditional donors, within a comprehensive sector strategy. The Government is also developing a PPP framework that could potentially crowd in more private investment to priority areas, while managing the risks. Optimizing existing investments and tracking implementation 60. The authorities could consider using the ‘Large Projects Database’ to reallocate funding to complete on-track projects and consider reprioritization of funding to stalled projects. The ‘Large Projects Database’ only covers two large capital spending ministries, but the findings indicate that the problem of stalled projects in Myanmar is prevalent and persistent -- with 30 out of 39 large/high priority classified as stalled. If no action is taken, these projects are likely to remain under implementation, leading to cost overruns and crowding out of scarce fiscal space for new projects. Options include: (i) Allocating funding to complete on-track projects. It is estimated that 11 percent of total capital spending will be needed to complete all on-track projects within 4 years (Table 9). (ii) Assessing the funding allocations to stalled projects. It is estimated that Myanmar will require an additional 8 percent of capital budget to finish all stalled projects in 4 years. However, if the Government considers to terminate a group of stalled projects which have the lowest disbursement rate (e.g. Tier 2 and 3), it could lead to saving of up to 5 percent of total capital spending every year.. Table 9: Funding needed to complete ongoing large projects in Electric ity and Energy; and Agriculture, Livestock and Irrigation Funding % of Annual If complete in 4 years No. of Project Category Need (Million Capital (% of Annual Capital Projects Kyat) Budget Budget) On-track projects (Disbursed > 80%, 9 1,953,382 42% 11% < 8 years of implementation) Stalled Projects Tier 1 (Disbursed > 12 807,298 18% 5% 80%, high priority for completion) Stalled Projects Tier 2 (Disbursed 8 435,665 9% 2% >30%, <80%, priority for completion) Stalled Projects Tier 3 (Disbursed < 10 93,514 2% 1% 30%, require critical reprioritization) Total 39 3,289,859 72% 18% Source: PAPRD. 61. In addition, Myanmar could build on the ‘Large Projects Database’ to develop a system of systematic portfolio monitoring. Currently, the database covers only two ministries but has already aided in identifying savings up to 5 percent of the annual capital budget. Covering more ministries, especially in high capital spending ministries such as Ministry of Construction could be beneficial in unearthing fiscal space. Furthermore, producing portfolio performance reports (such as Figure 63 and Table 9), could prove beneficial to policymakers in proactively identifying stalled projects and working with line ministries to address roadblocks to implementation. 65 Improving project appraisal and prioritization processes 62. Strengthening project appraisal can help avoid making big mistakes in terms of project selection and design. These mistakes can have significant costs during implementation, wasting resources and potential creating adverse environmental and social impacts. To improve project appraisal, the following actions have been identified: (i) Formalizing the mandate for central appraisal review. Myanmar re-introduced project appraisal in 2012, but only one line ministry is currently submitting appraisal reports regularly to MOPF. Myanmar could benefit from strengthening appraisal regulations to mandate line agencies to submit appraisal reports, for all projects including donor-funded and PPP projects, for review by PAPRD. (ii) Accelerating the roll out of new project proposal form and appraisal guidelines system-wide. New project proposal forms which provides more detailed information on the project and guidelines for appraisal has been developed, with technical inputs from the World Bank43. Using these guidelines requires intensive training, developing technical guidelines to conduct economic analysis and other analytics44 and strong collaboration between PAPRD and line ministries. (iii) Focusing scarce administrative resources to review large projects. The guidelines recognize that appraisal has to be proportionate to project scale and complexity, with differing requirements for small versus large, complex projects to avoid creating excessive burden. Appraisal and review should vary by size of the project. (iv) Linking project appraisal with funding. Finally, it is critical that appraisal review is instituted as a condition for eligibility for funding within the budget. The key question is which agency (i.e. MOPF) or institutional body (e.g. Commission, Committee) shall have the authority to grant approval after the technical review. Review of international experiences suggested that there is wide variance in the scope and intensity of involvement of the central finance and planning ministry in project appraisal (Box 8). Detailed consultation within Government may be needed to select the appropriate institutional framework for Myanmar. Box 8: International experiences on institutional setting for appraisal review and approval In the UK, where there is a great degree of decentralization of planning to functional ministries, appraisal review and approval could be done within sponsoring department/ministries. However, higher approval from HM Treasury is required (after sponsoring department approves projects) when expenditure proposals exceed delegated limits or when a project is novel or contentious. The key is that both sponsoring department and Treasury approval must be granted prior to starting project. In Vietnam, for project of national importance, the National Assembly approves investment guidelines and pre-feasibility studies while feasibility studies are approved by PM and State Appraisal Council chaired by Minister of Planning and Investment. For projects with less value, they are approved by Provincial People’s Committee, Minister of relevant line Ministry, line ministries or other designated authorities. 43 These guidelines have been developed with technical assistance under the ongoing Myanmar Integrated Public Financial Management Reform project, which is executed by the MOPF. The guidelines cover reporting templates, methodologies for assessing costs and benefits, determining appropriate discount rates and highlight the role of safeguards. 44 Further technical guidance such as on economic analysis of projects in different sectors (methodologies for assessing costs and benefits, determining appropriate discount rates) could also facilitate the implementation of the guidelines. 66 In Thailand, project above 1 billion THB is reviewed by the National Economic and Social Development Board and must seek Cabinet approval before inclusion into budget. For project less than 1 billion THB, project appraisal review is done as part of annual budget process and appraisal analysis is scrutinized by budget analyst. In some other countries, an independent agency is tasked to review (e.g. Ireland) or even conduct (e.g. Korea) appraisal and decision is made based on the recommendations of independent expert. 63. In addition to a robust appraisal process, prioritizing among high-return and well-prepared projects is also crucial. The existence of a pipeline of feasible, appraised projects following the strengthening of the appraisal system is critical but is insufficient. Establishing a prioritization framework or criteria could help bring greater clarity and rule-based process/approach to select the highest return projects ready for execution. The key principles are to adopt the framework and tool that are (i) accepted by all Ministries, (ii) simple to use in practice: and (iii) implementable with available data. The following options are proposed: (i) At the policy level, in the short term, the PER findings suggests that potential criteria for prioritization in the next 4-5 years include (i) a focus on projects that address transport and energy sector constraints; and (ii) projects that target constraints in the growth poles. These criteria for strategic prioritization could be reviewed and updated to respond to changing circumstances. (ii) At the project level, in the long term, the Government could consider scoring projects against pre- determined criteria. Myanmar could potentially draw on a number of examples from around the world, including a recent prioritization framework developed by the World Bank group in Vietnam,45 which (i) uses information from project appraisal reports and Government selection criteria to create social-environmental and financial-economic indices to score particular projects; (ii) combines the project score with value of projects to determine the amount of funding required to finance high-scoring new projects and (iii) compares this funding with available budget constraint, to determine the list of new projects. This system of scoring relies on strong data and analysis at the project level both on financial/economic analysis (EIRR, FIRR) and accuracy of project cost estimation – thus making this a long term recommendation. Medium term capital expenditure commitments 64. Moving towards medium term funding commitments for capital projects addresses one of the major impediments to effective project implementation. Lack of funding to execute projects on schedule lead to inefficiency and exposes the project to additional cost increases from inflation. Addressing this challenge is potentially a complex reform, and requires close and careful coordination linking the planning process to the Medium Term Fiscal Framework (MTFF)46 preparation steps. In this regard, some potential steps for Government consideration include: (i) Estimating the medium term expenditure need for large capital projects, which could draw on the information provided in the Large Projects Database, once ministry coverage is expanded; (ii) Linking expenditure needs to the MTFF framework, which is already in use for medium term budget preparation, in an iterative manner, taking into account prioritization and budget constraints. (iii) Linking expenditure needs to the MTFF, could then provide the basis for creating a 3-year medium term commitment budget for capital projects. This could be accompanied by changes in the annual budget law or an overarching PFM law, allowing for multi-year commitment budgets to be appropriated for large capital projects. 45 “Prioritizing Infrastructure Investment: A Framework for Government Decision Making”, World Bank (2016). 46 Please see chapter on macro-fiscal developments and outlook. 67 (iv) A concurrent step could be to reflect in Financial Rules and Regulations the provision for fund rollover for the capital projects that have a medium term commitment budget approved. (v) Over time, perhaps over several budget cycles, this process could potentially be refined to create a medium term Public Investment Plan (PIP) that is linked to the MTFF -- with a process in place such that agencies can learn from the implementation of the commitment budget. Table 10: Policy options for strengthening allocative and productive efficiency of public investment Issue Policy options Allocative efficiency Data constraints prevent adequate Short-term: Systematically compile and report on assessment of allocative efficiency of capital spending by Union Ministries across Union Budget and the Government ’s States/Regions. ability to effectively plan public investments. Current allocation of public Medium-term: Gradually re-allocate capital investments across sectors and public spending by (i) increasing relative spending on goods inadequately prioritize areas that energy and transport; (ii) reprioritizing SEE are important for long-term inclusive capital expenditure towards public goods and growth. (iii) adequately supporting investment needs in the growth poles Other sources of financing are not Medium term: Consider external concessional currently being leveraged to support and non-concessional financing, and public growth. investments, within a comprehensive sector strategy. Look to develop a PPP legal framework that effectively manages fiscal and delivery risks. Productive efficiency Large cost and time overruns on Short-term: Prioritize budget resources for tier 1 projects crowding out productive projects, and strengthen implementation trackin g capital investments. systems for all projects. Lack of project prioritization and Medium-term: Finalize prioritization and appraisal causing downstream appraisal guidelines and roll out across implementation problems. Government, including training program for application. Inability to rollover budgets leads to Medium-term: Adopt policies for multi-year weak implementation and fragmentation budget commitments for capital expenditure. of large projects into smaller ones. 68 FISCAL IMPACT OF STATE ECONOMIC ENTERPRISES 1. State Economic Enterprises (SEEs) have historically been an important part of the public sector in Myanmar though are increasingly becoming a drain on the Union Budget. SEEs in the oil and gas sector, which dominate SEE turnover, have seen declining profits due to lower commodity prices. Sales of SEEs in industrial sectors have dropped due to falling demand and lower market prices. A combination of these has contributed to declining net transfers from SEEs to the Union Budget, which have turned negative in the last two fiscal years. Efforts to manage the increased fiscal pressures felt by Myanmar’s SEEs are underway, but an outdated and complex institutional framework pose challenges for a comprehensive reform agenda. 2. This chapter reviews the institutional framework, performance and fiscal impact of SEEs in Myanmar, and presents options for strengthening fiscal oversight and containing fiscal risks. It includes: (i) a review of the SEE institutional framework including the legal and regulatory framework, fiscal flows between SEEs and the Union Budget, recent fiscal reforms in relation to SEEs, and SEE performance monitoring; (ii) an analysis of the financial performance of SEEs including their impact on the Union Budget and recent trends in SEE borrowing; (iii) policy options for strengthening fiscal oversight, including through improved performance monitoring, and containing fiscal risk of SEEs. The approach of this chapter is to assess the performance of Myanmar’s SEEs primarily from the perspective of the fiscal links between SEEs and the Union Budget. SEE INSTITUTIONAL FRAMEWORK AND FISCAL RELATIONS Overview 3. SEEs in Myanmar are the equivalent of State Owned Enterprises (SOEs) in other countries, though they have some distinct characteristics. Most SEEs function as operational departments within line ministries rather than (semi-) autonomous corporate entities. They are an integral part of the Government’s budget, receiving annual allocations for expenses. More recently, SEEs have been presented as a separate group in the consolidated public sector accounts, separate from the general Government (or Union Government). This is important so as to distinguish between service delivery out of tax and non-tax receipts, and production of goods and services through sales receipts. 4. The role of SEEs in the economy has gradually declined over time. SEEs were introduced in Myanmar in 1962 under General Ne Win. Considerable nationalization of firms took place and the total number of SEEs is anecdotally thought to have been around 15,000. Shortly after General Ne Win’s time in power ended in 1988, Government policy on SEEs changed significantly. This culminated in the establishment of a Privatization Commission in 1995, which privatized 772 entities between 1995 and 2011.47 5. SEEs have continued to evolve since 2011 with the total number of SEEs falling from 44 to 32 today. Twelve SEEs were integrated into ministries and departments, ceasing to be SEEs and converting into regular administrative units in Government, although still earning non-tax revenues and owning assets. Of the remaining 32, a total of 7 SEEs have been corporatized and have been able to operate with a greater degree of autonomy. Companies under the Ministry of Defense such as United Myanma Economic Holdings Limited or Myanma Economic Company are not under the supervision of the MOPF and are not included in the PER. 47 WTO Trade Policy Review 2014 69 6. SEE reforms are listed as a high priority on the Government agenda though have been pursued in an ad hoc manner. In addition to the changes listed above, a number of SEEs have been involved in Joint Ventures; some SEEs have been associated with Public Private Partnerships; and SEEs have been exploring different approaches to managing their asset base. This suggests a dynamic reform environment for SEEs which provides important opportunities for further reform. It also presents potential fiscal risks, particularly where corporate governance is not strong or the approach to reform is fragmented. Figure 66: Current number and status of SEEs 12 converted to administrative units 44 SEEs 32 remaining SEEs 7 Corporatized WB Staff based on discussions with MOPF 7. SEEs continue to have an important bearing on fiscal outcomes in Myanmar, though they are a smaller part of the economy compared to other transition countries. SEEs have accounted for between 35 and 45 percent of public sector expenses and receipts in the past 4 years. On the other hand, SEE sales account for around 7 percent of GDP, a lower share than in many transition and emerging market economies (Table 11), and OECD countries where sales account for around 15 percent of GDP. Cross-country comparisons of SOEs need to be undertaken with care due to definitional issues, but for the most part, Myanmar’s SEEs can be considered to be reasonably manageable in terms of size and complexity. Table 11: Cross-country comparison of SOEs SOEs Employment Revenue Net profit (number) (persons) (% of GDP) (% of GDP) China 160,000 several million 12 2 India 290 16 4 Indonesia 141 3 0 Myanmar 32 145,000 7 0 Russia 4,100 16 3 Vietnam 2,970 1,300,000 32 WB Staff estimates 8. SEEs account for a sizable share of public sector employment, though a relatively modest share of the overall labor force. Myanmar’s 32 SEEs have more than 145,000 employees,48 which represents around 15 percent of total public sector employment and less than 1 percent of the 22.7 million labor force. In contrast, China has some 155,000 State-Owned firms with several million employees. Vietnam’s 2,970 State- owned (or state majority owned) enterprises employ approximately 1,300,000 people.49, 50 9. SEEs in Myanmar are a varied group, some involved in commercial activities and others responsible for non-commercial and public service-type activities. SEEs are involved in a broad range of sectors including natural resource extraction; the provision of services such as transport, telecommunications and electricity; and industrial activities including heavy industry and downstream activities in the oil and gas sector (see table 13 below). 48 Data correct as of 1st of December 2016 and shared with the team by MOPF 49The Government of Vietnam has majority ownership in some 2,970 firms, of which 650 are 100 percent state owned. 50 Data on SOEs from Vietnam and China are based on WB Staff Estimates 70 Legal and regulatory framework 10. The legal and regulatory framework for SEEs is relatively thin and fragmented, underpinned firstly by an SEE Law introduced in 1989 by the State Law and Order Restoration Council. This law provides for the Government’s sole right to carry out activities in 12 different sectors.51 It also enables the Government to engage in other production activities whether through joint ventures or sole proprietorship. The procedures linked to this law give the right to the line ministry that creates the SEE to determine the duties, powers, governance and financial arrangements of the SEE. 11. Sector legislation and the Companies’ Act also play important roles in governing SEE operations. The day-to-day operations of most SEEs are governed by sector legislation overseen by the parent line ministry, which also sets sector policy and acts as regulator. The Special Companies’ Act (SCA) of 1950 allows SEEs to be corporatized by being registered as companies with the Government as a shareholder. They would be referred to as SCA Companies and would need to abide by the Companies’ Act. In practice the corporatized SEEs have mostly not been registered in this way (though Myanmar National Airlines is an exception). The SCA is used mostly to facilitate joint private and public ownership. The act says that SCA Companies are established as a separate legal entity and therefore any assets and liabilities of an SCA company are technically held in its own name and not by its shareholders (i.e. the Government). 12. Checks and balances around SEE Corporate Policy do exist but there is a lack of an overarching policy. Joint ventures (JVs) between SEEs and the private sector must be approved by the Myanmar Investment Commission unless these are small in nature, defined as JVs that last less than five years. Major changes, such as moving an SEE or part of an SEE to being an administrative unit in a line ministry requires Cabinet approval. Such changes are generally proposed by the line ministry and are considered by Cabinet on a case-by-case basis. In addition, a Privatization Commission headed by Vice-President 1 has a role in approving privatization plans submitted by line ministries. Despite all these arrangements, it is unclear whether privatization, corporatization, absorption back into the line ministry or other major structural changes are guided by an overarching SEE policy. 13. This situation arises in part due to the highly decentralized ownership of SEEs. Decentralized ownership appears to have its origins in a set of commercial guidelines for SEEs developed by the Ministry of Planning and Finance in 1975. These set out a number of principles for developing SEEs on a commercial basis with the intention of gradually providing more autonomy to SEEs. The responsibility for implementing these guidelines is left to line ministries through sector legislation, which is consistent with a decentralized approach to SEE management. SEEs in Myanmar do not have professional boards and anecdotal evidence suggests that SEE staff, including management, are often appointed to SEEs from within the parent line ministry and may move back into the line ministry after working in the SEE. 14. This ownership model presents some risks and challenges. A decentralized model can result in inconsistencies in SEE governance, including a lack of standardized practices across core corporate governance functions. The model also highlights some potential weaknesses in the overall governance framework for SEEs. In particular, line ministries play multiple roles which can result in conflicts of interest, for example between operator and regulator. This potential conflict can be exacerbated by recruitment policies if staff are regularly moving between working for the operator, the policymaker and the regulator. 51These sectors are Extraction and sale of teak; Cultivation and conservation of forest plantation; Exploration, extraction and sale of petroleum and natural gas and production of products of the same; Exploration, extraction and export of pearl, jade and precious stones; Breeding and production of fish and prawn in fisheries which have been reserved for research by the Government; Postal and Telecommunications Service; Air Transport Service and Railway Transport Service; Banking Service and Insurance Service; Broadcasting Service and Television Service; Exploration, extraction and export of metals; Electricity Generating Services; and Manufacture of products relating to security and defense. 71 15. The decentralized model could also exacerbate fiscal risks. In the case of a poorly performing SEEs, taking the difficult decision for the state to divest might be better made by a ministry with overall responsibility for public expenditure quality rather than a line ministry focused only on sector policy. If the latter drives policy, it could result in the state shouldering losses or cross-subsidizing underperforming SEEs for longer than necessary. On the other hand, in the case of a strong performing SEE that is accumulating profits or an SEE which owns valuable assets, decisions over how to use excess profits or returns might be better made by an institution representing the interests of the country as a whole rather than the specific sector interests of one ministry. Box 9: Global experie nce with SOE governance arrangements A strong legal and regulatory framework is important for providing clarity on the relationship between the Government, the SOE board and SOE management as well as creating a level playing field between SOEs and the private sector. There are different approaches such as SOE specific legislation, using sectoral legislation or embedding SOEs under company laws. State ownership must separate state ownership functions from the state’s policymaking and regulatory functions. This helps to minimize conflicts of interests between these functions and ensures that the SOE is free to operate as a company without excessive state interference. The Centralized Model (where ownership is centralized in a single entity) is increasingly seen as the model of good practice and is widely used in Asia by China, Indonesia, Korea, Bhutan and Singapore. In practice, the Decentralized Model (where ministries exercise the ownership function) is still used by a number of countries but it does increase the scope for political interference in SOEs and conflicts of interests as well as resulting in a fragmented approach. Another approach is the Dual Model where a ministry of finance might retain strong oversight of the SOEs whilst the ownership function is still played by line ministries. This model is essentially a compromise but can be particularly useful where the SOE sector is extremely large and the centralized model might not be practical. Both India and the Philippines use a Coordinating Agency approach where a particular Government unit provides advice to the ministries playing the shareholding function.52 Professional SOE boards that act on behalf of the state to look after its shareholding are seen as an important part of SOE corporate governance. Increasingly it is seen as good practice to reduce Government representation on these boards, including prohibiting ministers or regulators to be on boards to avoid conflicts of interest. Performance monitoring of SOEs is important in ensuring transparency around SOE activities and accountability in the use of public funds. For this to function well it is important that SOE mandates are well-defined, in particular where there are non-commercial objectives. Beyond that it is important that a structure for monitoring is agreed between the SOE owner and their boards and that performance indicators are agreed upon. Strong financial and fiscal discipline are important to reduce Government liabilities and to improve incentives for better SOE governance and performance. It is also important in ensuring that any subsidy received by the SOE is sensible given the Public Service Obligation of the SOE which is also key in levelling the playing field with the private sector. High transparency standards are important for SOEs to ensure that they are subject to the same standards as the private sector and so that the ownership entity can ensure public funds are being well used. In addition, strong auditing, oversight and accounting standards are also important. Source: The World Bank, “Corporate governance of state owned enterprises: A Toolkit,” 52 Some of the information from this section was drawn from State-Owned Enterprises in Asia, OECD 2016 72 SEEs and fiscal flows 16. The flow of funds between SEEs and the Union Budget are clearly defined and an important element of the overall SEE institutional framework. SEEs use two accounts to manage their funds: (i) the SEE account held at the Myanmar Economic Bank (MEB), which is used as the conduit for Union Budget receipts and expenditures; and (ii) an Own Account, also in MEB, which holds the SEEs’ retained earnings that are used to finance a growing share of SEE expenses. 17. To illustrate, SEEs receive an allocation from the Union Budget into their SEE Account to finance their expenses (Step 1, Figure 67). The SEE Account also collects non-production revenues of SEEs (Step 2, Figure 67), which are not derived from their core business, for example earnings from the rental of buildings owned by the SEEs, administrative fees/charges imposed by the SEE (e.g. fines), or other. The SEE Account is in turn used to finance SEEs’ recurrent and capital expenses (Step 3, Figure 67) for the production of goods and services.53 The SEE account is also used for SEEs’ debt service payments (i.e. interest and principal repayment). Figure 67: Flow of Funds between Union Budget and SEEs Union Budget (1) Union Budget allocation for: o Recurrent expenses (Y/N) o Loan interest (Y/N) (6) SEE payments including: o Capital investment (Y/N) o Contribution/dividend (Y/N) o Income tax (Y/N) o Commercial tax (Y/N) State Economic (2) SEE non-production revenue from: Enterprise o Other income (e.g. rent) o Administrative fees/charges (3) State funded o Other expenses SEE Account Annual Annual expenses Own Account revenues (5) SEE funded expenses (4) SEE production revenue from: o Sales of goods and services (7) SEE Net profits: o Service fees/charges o Sales revenue minus production expenses minus tax and contributions o Joint venture profit share o Net profits are rolled over in the Own Account for next fiscal year o Other Source: WB Staff based on discussion with MOPF 53 Union Budget allocations for profitable SEEs that do not face liquidity constraints are declining as they are required to finance a growing share of expenses out of their own profits, as discussed further below. 73 18. SEE production revenues (e.g. from sales, service fees, join venture profits) are transferred to SEEs’ Own Accounts (Step 4, Figure 67). The SEEs’ Own Account is used to finance selected expenses (e.g. raw materials) (Step 5, Figure 67). Tax liabilities and contributions (i.e. dividends) are also transferred from the Own Account to the Union Budget (Step 6, Figure 67). Commercial taxes are applied on sales. Income taxes and contributions are charged at the rate of 25 percent and 20 percent respectively on operating profit (i.e. sales revenue minus recurrent expenses). The SEEs net profits (i.e. the remaining 55 percent of operating profits) remain in the SEE Own Account and rolled over to the next fiscal year (Step 7, Figure 67).54 19. The structure of the accounts is provided for in the Ministry of Finance instruction 547/201255 whilst the rules on the flow of funds are updated every year in the annual Union Budget Law. Instruction 547/2012 provides for the establishment of SEE and Own Accounts. It enables the rollover of positive cash balances in Own Accounts from one year to the next as noted above. The Union Budget Law, on the other hand, specifies each year which expenses are eligible for funding through the SEE and Own Accounts depending on the financial condition of the individual SEE (see below). 20. Instruction 547/2012 and concurrent updates in the Union Budget Law only apply to the 25 non-corporatized SEEs. It does not cover the 7 corporatized SEEs nor the 12 that have been converted into administrative units. There is more autonomy in determining the flow of funds for corporatized SEEs. For example, line ministries are able to decide whether or not corporatized SEEs should cover capital and debt service liabilities out of their Own Accounts. Currently all corporatized SEEs service their own debt. They also have more flexibility in the use of Own Account balances, including for investment in Treasury Bills and Bonds. 21. SEE income taxes are paid to the Internal Revenue Department, and contributions are transferred to the Treasury. The treatment of commercial tax is more complex. This is calculated by and paid to the Internal Revenue Department. The tax can vary depending on the sector and the stage of the value chain at which it is applied (e.g. production, sales, distributions). SEEs do not receive specific tax exemptions. Though they may be eligible for some exemptions if for example they participate in a Joint Venture or they import goods funded out of a grant. 22. SEE tax payments are made ahead of actual liability, which is designed to manage cash flow pressures but creates inefficiencies in financial management. For example, SEEs estimate their commercial tax liability based on production and sale forecasts. This in turn forms the basis of monthly commercial tax payments by SEEs, who produce a quarterly report summarizing taxes paid. At the end of the year the SEE submits an annual commercial tax report once this has been audited (within 3 months FY end). The annual report is used to reconcile projected and actual tax payments, and is used to make any adjustments i.e. either to reimburse SEEs for over payment or receive transfers from SEEs for under payment. This adjustment is made within 12 months, which causes delays in finalizing end of year accounts. The process for Income Tax is similar with the exception that the payments are made quarterly rather than monthly. 23. The Myanmar Oil and Gas Enterprise is the largest SEE contributor to the Union Budget. Revenues from natural gas are conservatively estimated at an average of 2 percent of GDP per year, or 15-20 percent of Union Government revenues, between 2013/14 and 2015/16. Revenues from natural gas come through royalties and state participation in production through SEEs. The fiscal regime for natural gas is governed through Production Sharing Contracts (PSCs). Myanmar’s PSC structure includes: (i) Royalties, which are payable on production; (ii) Profit share of overall gas production net of cost recovery; (iii) Corporate taxes, which are payable on profits; and (iv) Dividends accruing from participation of SEEs such as MOGE as a partner for private firms. 54 Profitable SEEs that do not face liquidity constraints now need to draw down balances from their Own Accounts to finance expenses, which helps to reduce their allocations from the Union Budget. This is discussed further below. 55 Based on documentation provided to the PER team by MOPF staff 74 24. State Owned Banks have different provisions for tax payments compared to SEEs. In particular, State Owned Banks have a more diversified approach in determining contributions to the State, though they are still set in percentage terms, suggesting some flexibility at the individual bank level to adjust dividends based on the economic performance or operational needs of the bank.56 SEEs and fiscal reforms 25. Starting in 2012 reforms were initiated to grant increased financial autonomy to SEEs by allowing them to retain more of their profits. Instruction 547/201257 reduced the SEE contribution to Government from 70 to 20 percent of operating profits; and the Income Tax Law reduced the Corporate Income Tax from 30 to 25 percent. SEEs could therefore retain more of their net profits in their Own Account, which in turn had to cover an increasing share of their expenses. Holding all other things constant this should mean lower net transfers to SEEs from the Union Budget.58 When the policy was first applied in 2012/13, all non-corporatized SEEs were treated the same: 78 percent of the costs of raw materials were to be funded through Own Accounts; whereas the remaining 22 percent of raw materials, together with other recurrent expenditures, as well as capital expenditures and debt servicing costs were funded by the Union Budget. Over the next two years, MOPF required profitable SEEs to not only cover all raw material expenses but also tax and contribution liabilities out of their Own Account. 26. By 2016/17 and 2017/18, additional measures were introduced to further extend the fiscal autonomy of SEEs. In particular, MOPF grouped SEEs according to: (i) those that had sufficient working capital in Own Accounts to cover recurrent expenses; and (ii) those that had insufficient working capital in Own Accounts to cover recurrent expenses. The treatment of each category is summarized below (Table 12). SEEs that were profitable and had sufficient working capital had to cover all recurrent expenses, with the exception interest on external loans. From 2017/18, SEEs in this same category would also have to cover capital expenses. For these SEEs, the only expense covered by the Union Budget would be debt repayment costs. 56 In particular: (i) MEB and the Myanma Foreign Trade Bank (MFTB) pay a 25 percent Income Tax. After this payment, they must make a payment of 2 percent of profits against bad debts. 75 percent of the remaining amounts are paid as a contribution to the state and the remaining 25 percent go into a fund reserve; (ii) The Myanmar Investment and Commercial Bank (MICB) pays 25 percent Income Tax, 10 percent as an emergency reserve and 2 percent go towards bad debts. Of the remaining profits, 75 percent goes as a contribution to the state and the remaining 25 percent goes into a reserve fund; (iii) The Central Bank of Myanmar (CBM) pays 40 percent of profit into General Reserve Funds and 60 percent as a contribution to the State; (iv) Myanma Insurance pays 25 percent of profits as Income Tax. Of the remaining 75 percent, 90 percent is paid as a contribution to the State and 10 percent goes into a reserve fund; (v) The Security Printing Works pays 25 percent of profits as Income Tax whilst 75 percent goes to the State as a contribution; (vi) The Myanma Agriculture Development Bank (MADB) does not pay Income Tax. Instead it allocates 2 percent of its profits to bad debt and, of the remaining profit, 75 percent goes as a contribution to the State and 25 percent of it goes to a Reserve Fund 57 Based on documentation provided to the PER team by MOPF staff. 58 The focus of these instructions and calendar notifications are those organizations running businesses (business enterprises) – they do not include the Central Bank, other banks, the Myanmar Insurance Business, Financial Regulatory Agencies or Social Welfare Organizations formed under the social security law. They include the following documents: • Letter No. bakha – 3/20 (547/2012) • Letter No. Bakha 3/20 (915/2012) dated the 12th of July 2012 • Letter No. Bakha 3/20 (1474/2013) dated the 16th of September 2013 • Letter No. Bakha – 3/20 (5875/2014) dated the 14th of November 2014 • Letter No. Finance – 4/1/2 (1531/2015) dated 16th of November 2015 • Letter No. SaBa/Finance – 4/1/1 (3995/2016) dated the 22nd of November 2016 75 Table 12: Expenses covered by Own Ac count and Union Budget across SEEs Sufficient Insufficient o Own Account: All current expenditures o Own Account: raw materials, production and (with the exception of interest payments on service expenses, income tax, state external loans), income tax, state contribution contribution and commercial tax Profitable and commercial tax. o Union Budget: remainder of current o Union Budget: Interest payments on expenditure, capital and debt repayment external loans, capital expenditure and debt repayment o Own account: current expenditures (with the o Own Account: cover the cost of raw exception of interest payments on loans) as material, the cost of production and services, well as commercial tax using their own fund and commercial tax Not account profitable o Union Budget: The remainder of current o Union Budget: All other current expenditure (less working capital), capital expenditures, capital expenditure and debt expenditure and debt repayment repayment 27. These reforms effectively grant SEEs autonomy over management of their finances, which in theory should incentivize them to pursue efficiency gains. Prior to 2012, SEEs would transfer all their profits to the Union Budget, which then allocated funds for SEE expenses. SEEs can now retain their profits and allocate these directly for their own use. This is a positive move in terms of separating the activities of de facto public corporations from the public service functions of General Government. 28. The difficulty however is that the external environment has also become a lot more challenging, which has dented SEE profits and increased losses. Exogenous shocks from falling commodity prices have negatively impacted profits of SEEs in Natural Resources and Extractives in particular (Table 13). Similarly, intensified competition has led to losses for SEEs in Industry and Services. Whilst competition could also incentivize efficiency gains, investments needed to upgrade production capacity far outweigh what SEEs can afford out of their own resources. Moreover, even if such SEEs did have enough resources there is a larger question as to whether it would be in the best interests of Government as a whole to invest public funds in Industry and Services. As a result, a number of SEEs have gone the other way by converting into administrative units in line ministries (“Transformed” SEEs in Table 13). This removes their income tax and contribution liabilities and makes them entirely dependent on the Union Budget. 29. Increased autonomy for SEEs is unlikely to enable self-sufficiency and reduce the fiscal burden over the medium to long-term unless it is complemented by a broader reform strategy. For example, financial autonomy should be underpinned by strong SEE business planning and hard budget constraints. This could be an integral part of the Union Budget process, as is the case in many other countries, to help prevent unnecessary investments, restrict subsidies, and manage fiscal risks. Some SEEs may require restructuring or privatization, which should be pursued in an open and transparent manner. It may also require revisiting some aspects of the SEE ownership model to prevent ad hoc integration of SEEs as administrative units in line ministries, which are contrary to the above SEE reform objectives. 76 Table 13: List of Myanmar’s SEEs with key performance data for most recent fiscal year SEE MOPF Category Sales Profit/loss (Kyat millions) (Kyat millions) (NET) Natural Resources and Extractives [Total 6] Myanma Gems Enterprise Sufficient 150,251 116,564 Myanma Oil and Gas Enterprise Sufficient 228,584 412,156 Myanma Pearl Enterprise Sufficient 14,005 6,676 Myanma Timber Enterprise Sufficient 300,289 42,612 No. 1 Mining Enterprise Sufficient - 3,902 No. 2 Mining Enterprise Sufficient 34,523 15,983 Infrastructure [Total 10] Electric Power Generation Enterprise Insufficient 1,040,142 (377,751) Electricity Supply Enterprise Insufficient 378,727 3,542 Inland Water Transport Corporatized 7,649 (5,882) Mandalay Electricity Supply Corporation Corporatized 243,102 18,361 Myanma Post and Telecommunications Sufficient 223,800 150,918 Myanma Railways Insufficient 57,405 (81,716) Myanma Port Authority Corporatized 64,002 24,900 Public Works (2015/16) Transformed 519,637 9,357 Road Transport Insufficient 7,322 (3,677) Yangon Electricity Supply Corporation Corporatized 552,460 8,737 Industry and Services [Total 20] Co-operative Export Import Enterprise (14/15) Transformed 315 21 No. 1 Heavy Industries Enterprise Insufficient 111,390 (11,705) No. 2 Heavy Industries Enterprise Insufficient 49,853 (25,100) No. 3 Heavy Industries Enterprise Insufficient 16,022 (32,362) Livestock, Feedstuff and Milk Products Enterprise Transformed 8756 (762) (14/15) Transformed 36 (818) Myanma Agricultural Product Trading (13/14) Corporatized 133,899 303 Myanma Airways Transformed 3400 945 Myanmar Hotels and Tourism Services (14/15) Insufficient 501,773 49,750 Myanma Petrochemical Enterprise Insufficient 670,545 51,600 Myanma Petroleum Products Enterprise Insufficient 32,447 162 Myanma Pharmaceuticals Industries Insufficient 7,558 (12,425) Myanma Posts Transformed - - Myanmar Salt and Marine Chemical Enterprise (16/17) Transformed 3,096 (16,728) Myanma Agriculture Service (12/13) Transformed 22,810 (15,310) Myanma Industrial Crops Development Enterprise Transformed 67 (330) (12/13) Corporatized 2,320 28 Myanma Motion Picture Enterprise (14/15) Sufficient 5,760 855 Myanma Shipyards Transformed 11,293 (126) News and Periodicals Enterprise Transformed 7,976 3,463 Printing and Publishing Enterprise (16/17) Social Security Board (13/14) Banking and Other Financial Services [Total 8] Central Bank of Myanmar Corporatized 434,945 310,939 Myanma Agricultural Development Bank Financial Sector 48,900 11,767 Myanma Economic Bank Financial Sector 303,244 1 Myanma Foreign Trade Bank Financial Sector 23,960 21,461 Myanma Insurance Financial Sector 31,902 8,391 Myanma Investment and Commercial Bank Financial Sector 15,191 8,867 Myanma Microfinance Supervisory Enterprise (14/15) Transformed 3,194 216 Security Printing Works Financial Sector 89,950 44,836 77 Source: World Bank staff calculations based on MOPF data Data points are for most recent fiscal year (FY16/17) unless otherwise noted Category is based on MOPF’s definition. Transformed is transformed into administrative units. Financial data is presented in millions of kyats (nominal) 30. An important element of this broader reform strategy are reforms associated with public debt management. Previously SEEs were able to borrow directly with little central oversight, which created contingent liabilities and associated risks for the public sector, exacerbated by weak corporate governance of SEEs. As a result, SEEs are currently saddled with significant legacy debt, which are discussed further below. Since the adoption of the Public Debt Law, all public sector borrowing decisions have to be reviewed and approved by the MOPF, and subsequently by the Parliament. SEEs are no longer permitted to lend to one another, except for those SEEs operating outside of the Union Budget (e.g. SEEs under the Ministry of Transport and Communications, who can borrow from each other based on their respective guiding legislation. They are also no longer allowed to borrow directly from State Banks, or external creditors, including through on-lending or guarantees without the requisite MOPF and Parliamentary approvals. This is a good step forward in terms of managing potential fiscal risks, particularly as SEEs are granted increased autonomy. 31. Closely linked to this however is the lack of clarity over prioritization or appraisal of SEE investments. As discussed in the capital expenditure chapter, SEEs account for a sizeable share of public investments. This includes not only utilities but also commercial enterprises.59 Additionally, starting in 2017/18, 8 SEEs will be able to fund capital investments out of their Own Accounts. But this poses risks if not underpinned by strengthened public investment management capacity and clear prioritization criteria. As discussed in the capital expenditure chapter, Myanmar could benefit from adopting clear principles for capital investments by SEEs. SEE fiscal oversight and performance monitoring 32. SEE oversight and performance monitoring is spread across many agencies, which disperses responsibility and risks diluting accountability. The SEE Division under the Budget Department of MOPF is responsible for overseeing fiscal links of SEEs. The Planning Department provides views on SEE restructuring including through its engagement in the Privatization Commission. SEEs that have converted to administrative units are overseen by the Budget Department. SEE Debt is overseen by the Debt Department within the State Treasury. State Banks are overseen by the Central Bank of Myanmar. 33. SEEs submit a substantial amount of data to the SEE Division as part of the budget process. Before 1988, in addition to a Budget Calendar, the MOPF issued an SEE Budget Calendar setting out the reporting requirements for SEEs. The two processes were integrated after 1988. Since 2012, Instruction 547/2012 discussed above and the Annual Budget Law and Budget Calendar govern SEE reporting. SEEs must submit 18 forms in total. They submit the first 17 forms as part of the Budget preparation process but form 18, the Balance Sheet, is only submitted after it has been audited by the Office of the Auditor General. 34. The quality of reporting is mixed, which means that consolidating data is a major task for the MOPF. Information is not submitted electronically. Much of staff time in the SEE Division is spent compiling and verifying data, and revising submissions to ensure compliance with reporting templates. This greatly reduces the ability of the SEE Division to focus on analysis and policy-setting. Consequently, the SEE Division focuses on Forms 2 (Profit and Loss) and 14 (Cash Accounts) in terms of its analysis. 59The top five SEEs are No.1 Heavy Industries Enterprise (29 percent), Myanmar Electric Power Enterprise (27 percent), Myanmar Posts and Telecommunications (8 percent), Myanmar Railways (8 percent) and Myanmar Oil and Gas Enterprise (7 percent). 78 35. Financial audits are conducted for each SEE by the Office of the Auditor General (OAG). SEEs were subjected to two audits a year up to 2015/16 and, starting from 2016/17 this was sensibly reduced to once a year. The OAG conducts risk-based audits on both the SEE Account and the Own Account. The audits are submitted to SEE management. The OAG then submits a combined audit to parliament. This includes a Budget Provision report and an Audit Objection report. Some of the information from the SEE audits is presented in this report but the SEE audits are not submitted to parliament as standalone documents. In the case of a Joint Venture or a Corporatized SEE, the JV or SEE can choose between a public and private auditor. In the case that a private auditor is chosen, the OAG must approve the selection of the auditor. 36. The audited financial statements of these SEEs consist of a balance sheet and income statement, which fall well below international good practice standards. The audited financial statements are prepared in accordance with the current accounting standards for SEEs which are based on ‘gene rally accepted accounting principles’.60 However, it is not possible to determine from the financial statements any detail on the accounting standards applied as they do not include statements of accounting policies, notes to the financial statements or audit reports.61 It is also not possible to determine whether the financial statements provide a fair view of the SEEs’ financial position or financial performance or how the manage their risk exposures. This hampers decisions on allocating resources to SEEs, whether in the form of Union Budget allocations, bank loans, or private investments. 37. Despite these challenges, Myanmar has managed to put in place the basic building blocks for strengthening fiscal oversight and performance monitoring of SEEs. There is already a dedicated division within MOPF with committed staff that have in depth knowledge of SEEs and their relationship to the Union Budget. There is a wealth of existing data, which can be mined and reviewed, as is already happening and informing policies to progressively grant greater independence to SEEs. Data quality could be strengthened by simplifying reporting requirements, automating data collection and submission; and by working with SEEs to promote more consistent and higher quality data reporting. Ensuring consistency in definitions and form submissions could help improve reporting. 38. Improving the quality of data and automating the system would free up staff time in the SEE Division to review and analyze SEE data in more depth. This could be formalized into a short report on SEE performance submitted to the MOPF periodically. The MOPF’s role vis-à-vis the shareholding of SEEs could be strengthened to formalize performance discussions between the SEE and the MOPF and eventually lay the groundwork for a stronger performance monitoring framework overseen by a professional board on behalf of the shareholding entity. 60 In the SEE context generally accepted accounting principles means the existing double entry accounting procedures with some accrual entries incorporated in the financial statements. However, there are no accounting standards as such. 61 The fact that audit reports were not included was likely due to a lack of understanding at the SEEs of what audited financial statements normally comprise. 79 Box 10: Performance monitoring of SOEs in Asia Strengthening the definition of the state’s objectives through clear performance indicators and monitoring can enhance SOE performance and transparency. Some use Performance Contracts which outline yearly performance targets signed between owners and manager. Others use Performance Indicators where performance is evaluated against a number of financial and non-financial targets. In some countries, performance evaluations are completely undertaken by SOE boards, whereas elsewhere evaluations are either led or overseen by the state ownership entity. In most countries, performance evaluations are conducted at least yearly, with some conducting evaluation as much as quarterly. Most produce an annual aggregate performance report as well. In many countries the remuneration of SEE management is often linked to performance evaluations. Asian countries use a variety of indicators to evaluate performance: Quantitative Qualitative Financial • Net Interest Margin (Bhutan) • Policy Directed Activities (Bhutan) • Economic Value Added (China) • Quality of risk management (Indonesia) • Net Profit (India) • Transparency of budget process (Korea) • Financial Ratios (Indonesia) • Labor Productivity (Korea) • Return on Investment (Philippines) Non- • Customer Satisfaction Index (Bhutan) • Quality of corporate governance (Bhutan) Financial • Number of new products (China) • Commitment to corporate social • Project cost overrun (India) responsibility (India) • Number of corporate events (India) • Timely submission of reports to regulators • Achievement of core business targets (Indonesia) (Korea) • Development of gender equality policy • Percentage of beneficiaries saved (Korea) (Philippines) • Certifications indicating compliance with international standards (Philippines) Summary of performance management approach used in Vietnam and Indonesia: Indonesia Vietnam Financial Performance: Financial Performance: • Financial Ratios • Revenue • Net Profit • Profit and Return on Equity • Growth • Overdue liabilities and ability to repay • Risk Management • Share performance Non-Financial Performance: Non-Financial Performance: • Corporate Events • Supplying public goods and services • Corporate Social Responsibility Programs Legal compliance vis-à-vis taxes, budget • Corporate Soundness Level remittances, credit, insurance, environmental • Timely submission of reports to regulators protection, employment, wages, accounting and auditing, submitting financial and other • Public Service Obligations reports • Implementation of Corp. Gov. Practices • Awards Source: Adapted from State-Owned Enterprises in Asia, OECD 2016 80 SEE FINANCIAL PERFORMANCE AND FISCAL IMPACT SEE financial performance and profits 39. Building on the review of institutions and policies linking SEEs to the Union Budget, the PER analyzes the financial performance of SEEs and how this is impacting the Union Budget. This draws on data from SEE profit and loss statements. There are two caveats: (i) indicators of financial performance are not entirely consistent with international norms (Table 14). Whilst these do not materially affect the final fiscal outcomes, they do matter in terms of interpreting SEE viability, profitability, and efficiency. The analysis is based mostly on Myanmar norms due to data constraints, whilst suggestions are made for future adjustments (Table 15); (ii) Data gaps across SEEs affect the comprehensiveness of the analysis. This explains some of the difference between the results below and those in the Medium-Term Fiscal Framework. 40. SEEs’ Gross Profits, which measure commercial viability, have been on a declining trend over the past four years.62 Gross profits have declined to an estimated 0.7 percent of GDP in the 2016/17 Union Budget from 4.0 percent of GDP in 2012/13. The sharp drop began in 2014/15. Infrastructure and natural resources SEEs have experienced particularly large drops (Figure 68).63 Falling Gross Profits would ordinarily imply that Sales Revenue are covering a declining share of Direct Production Expenses. This could be due to different factors e.g. declining market demand impacting sales; increased cost of inputs impacting production costs. This should prompt a cut back to, or in some cases even a ceasing of, production. 41. The decline in Gross Profits has been due to increased Operating Expenses relative to slowing or declining Sales Revenue (Figure 69). Sales did contract in 2014/15, but rising Operating Expenses have had a relatively bigger impact in dragging down Gross Profits. The infrastructure sector contributed most to SEE Operating Expenses (Figure 70). This is due to higher gas purchase costs for the power sector resulting from exchange rate adjustment in 2013, and further depreciation between 2015 and 2017. It is also linked to the expansion of power and telecommunications, which has required increased Operating Expenses. 64 Figure 68: Gross Profit/Loss by sectors Figure 69: Gross Profit/Loss drivers Sources: MOPF, WB Staff estimates Note: The above exclude Other Income from Gross Profits, as these are not derived from own production activities. This is why the share of Gross Profits to GDP is so low. 62 If only Cost of Goods are deducted, and not all Operating Costs as seems to be done in Myanmar, then the result on Gross Margin (i.e. (Gross Profit/Loss)/Sales Revenue) would be higher. 63 Note: This excludes Other Income, which accounts for the largest source of revenue for MOGE. 64 Prices are no longer administered in Myanmar and so, for the most part, SEEs generally compete on market terms with respect to pricing. Natural resource prices are set according to international market trends. However, the exception is in the domestic utility market where retail prices are subject to regulatory control. 81 Table 13: SEE financial performance and profit indicators Indicator Inputs in Myanmar SEE Definition and comments profit/loss statements Receipts Sales Revenue Price x Quantity sold Gross receipts from sale of goods and services from SEE core business. An indicator of business growth, particularly for non -commodity SEEs. Other Income Administrative fees (e.g. Receipts that are unrelated to SEE core fines), rental receipts, production. These are not necessarily recurring. investment income, tax In Myanmar a large portion of Other Income is refunds, other . MOGE income from Joint Venture Projects Expenses Direct Direct labor expenses, raw Direct costs of producing goods and services. Production material inputs, machinery. Also referred to as Cost of Goods Sold. Not Expenses presented separately in P&L statement. Operating General administration, Expenses associated with keeping t he business Expenses salaries, product going, and not linked to Direct Production distribution, advertising, Expenses. Includes overheads. R&D Interest Interest on SEE debt. Part of debt servicing cost. Principal repayment Payment part of Union Budget financing. Profits Gross = Sales Revenue Indicator of commercial viability of SEE. Profit/Loss – Operating Expenses Inability to cover operating expenses from sales indicates need to scale back operations or potentially shut down. Operating = Gross Profits Indicator of business profitability and efficiency. Profit/Loss + Other Income Ordinarily derived: without adding Other – [Overheads Income; by deducting capital depreciation; and + Interest payment before interest and tax payments. Used in + Commercial tax] Myanmar to derive taxable income. Operating = Operating profits Indicator of business profitability that can be Margin / Sales revenue compared across similar businesses. Net Profits = Operating profits Profits retained by SEEs in their Own Account – [Income Tax to be spent in accordance with Instruction 547 + Contributions] and Union Budget Law. Payments to the Union Budget Commercial tax 5% of Sales revenue Indirect consumption tax Income tax 25 % of Operating profits Direct income tax Contributions 20% of Operating profits Equivalent to dividend payment Source: Derived from Form 2 of SEE reports Table 14: Suggested adjustments to SEE performance indicators Current practice in Myanmar Suggested adjustments Sales Revenue Sales Revenue – Operating Expenses – Cost of Goods and Services = Gross Profit/Loss = Gross Profit/Loss + Other Income – [Operating Expenses + Overheads] – [Overheads + Interest + Com. Tax] = Operating Profit/Loss = Operating Profit/Loss + Other Income – Income Tax – Interest – Commercial Tax – [Com. Tax + Inc. Tax + Contribution] = Net Profits = Net Profits Source: WB Staff 82 42. Sales have declined in the natural resource sector, though picked up in the infrastructure sector (Figure 71). The gas sector through MOGE has the highest turnover among all SEEs, though 85 percent of MOGE’s income is recorded as Other Income (Table 14) rather than Sales Revenue. The Other Income represents MOGE’s profit share from Joint Venture projects. The decline in MOGE’s Other Income started in 2015/16 with the impact of declining international commodity prices. The actual drop in natural resource sector sales is coming from the Myanmar Timber Enterprise and the Myanmar Gem Enterprise. Sales in the Infrastructure sector have propped up aggregate SEE Sales Revenue. Infrastructure Sales Revenue comes from the power utilities reflecting rapidly increasing demand for electricity. But this masks poor performance across other SEEs within the Industry and Services sector, reflective of low competitiveness and market demand. Figure 70: Sector contribution to change in Figure 71: Sector contribution to change in operating costs sales revenues Sources: MOPF, WB Staff estimates 43. In addition to Gross Profits, Operating Profits, as assessed using international norms, are also declining, which means lower taxable earnings. Assessed by Myanmar standards, Operating Profits declined from 4 percent of GDP to an estimated 0.75 percent of GDP between 2013 and 2016 (Figure 72). But this assessment of Operating Profits: (i) includes Other Income, which should be part of taxable income, but not part of Operating Profit/Loss; and (ii) subtracts interest payments and commercial tax, which should not be deducted at this stage. Other Income is not derived from SEE production and sales. This is the case for MOGE, for example, which derives income from Joint Venture projects in which production and sale are the responsibility of foreign operators. By not including Other Income (and not subtracting tax and interest), Operating Profit/Loss declines from 3.6 percent of GDP to an estimated 0.5 percent between 2013 and 2016. 65 44. Whilst the gas sector accounts for the biggest drag on Operating Profit/Loss, the infrastructure sector has also started to experience Operating Losses in the past 2 years. This is as noted above due to Operating Costs in the power sector with rising demand for electricity services. Power generation is the key driver for rising Operating Costs (Figure 73), followed by the distribution business in Yangon and Mandalay. This reiterates the importance of a comprehensive policy for electricity pricing, which promotes the commercial viability of the power sector whilst protecting poor and vulnerable households. 65The MTFF in the macro-fiscal developments and outlook chapter shows small operating losses in 2015/16 and 2016/17 for SEEs as a whole. There are 2 reasons for this. The first is that data gaps at individual SEE level in the information collected for this chapter creates difficulties for the comprehensiveness of this exercise. The second is that the MTFF approximates Operating Balance by subtracting total recurrent expenditure from total revenue, which is reported separately in Form 14. 83 Figure 72: Operating Profit/Loss Figure 73: Operating Profit/Loss by Sector Figure 74: Operating Margin Figure 75: Net profits by Sector Sources: MOPF, WB Staff estimates. Note: Other Income excluded from Operating Profits. 45. Declining profits and performance reaffirms the challenge of increasing the financial autonomy of SEEs in the absence of a broader reform program. Faced with declining sales due to lack of productivity and competitiveness, some SEEs under the Ministry of Industries have sensibly scaled back production. This helps bring down Direct Production Expenses, but Operating Expenses still need to be covered until a decision is made on the future of these SEEs. In other cases, a rapid expansion in services, for example in the electricity sector, has required increased operating (and capital) expenses. Whilst sales have also increased, a lack of pricing adjustments have contributed to widening Gross and Operating Losses. This impacts negatively on the commercial viability of the power sector and its ability to attract investments. At the same time, any adjustment in prices has to be in line with protecting poor and vulnerable consumers. 46. Moreover, Net Profits and therefore SEEs’ ability to retain earnings to exercise autonomy are either declining or turning negative (Figure 75). With the exception of SEEs in the Natural Resource sector, the other three sectors all face Net Losses. This means lower earnings in SEEs’ Own Accounts and thereby less own working capital, which will need to be offset by Union Budget allocations. Monitoring these developments systematically is important for effective policy response. An important element of this is the need to update accounting definitions as discussed above to help monitor SEE performance and fiscal impact, report to policy makers, and take timely actions. The proposed adjustments would not require additional data, and could significantly enhance the fiscal oversight of SEEs. 84 Fiscal impact of SEEs 47. With declining financial performance, SEEs are estimated to have transitioned in the past two years from being net contributors to the Union Budget to being net recipients. In the Union Budget and MTFF, SEEs in the period 2015 – 2018 are expected to make net losses of between 0.5-1 percent of GDP (see chapter on fiscal space). The data from the Profit and Loss Statement show that SEEs are still making a net positive contribution, but as noted above this is largely due to the bottom up data gaps in the Profit and Loss Statements, which show SEEs as a group experiencing a decline in overall surplus from 3 percent of GDP in 2013/14 to 1 percent of GDP in 2015/16.66 48. Fiscal outcomes are driven by the performance of a small number of SEEs. Five SEEs, all in either the oil and gas sector or the electric power sector, represent some 60 percent of SEE revenue and 80 percent of SEE income (Myanmar Oil and Gas Enterprise, Electricity Generation Enterprise, Myanma Petroleum Products Enterprise, Yangon City Electricity Supply Corporation and Myanma Petrochemical Enterprise). These same five SEEs represent about 55 percent of total current expenditure. Unsurprisingly MOGE stands out as the single most important SEE, accounting for around 81 percent of commercial tax payments, 48 percent of Income tax payments, and 24 percent of contributions (Figure 77). Figure 76: SEE Revenue, Expenses and Balance Figure 77: SEE payments to Union Budget and the oil and gas sector Sources: MOPF, WB Staff estimates 49. In the short-term, reforms granting more financial autonomy to SEEs have reduced transfers from the Union Budget, as SEEs have dipped into Other Account balances to cover expenses (Figure 78). From 2016/17, corporatized SEEs are responsible for covering 100 percent of their recurrent expenses from earnings in their Current State Owned Enterprise Account (Own Account equivalent). SEE expenses funded by the Union Budget are estimated to have decrease from Kyat 1.9 trillion in 2013/14 to Kyat 0.8 trillion in 2016/17 (Figure 79). 50. Much of the decline in Union Budget allocations therefore is due to SEEs with sufficient working capital in their Own Accounts. Union Budget allocations for these SEEs fell from Kyat 1.1 trillion in 2013/14 to an estimated Kyat 0.1 trillion in 2016/17 (Figure 80). Support from the Union Budget for SEEs with insufficient working capital has increased since 2013/14 although the amount started to decline in 2016/17 (Figure 81). In 2016/17, wages and salaries, administrative costs, interest payments (and some taxes and contributions), are still (pre)funded by the Union Budget. 66 Based on revised estimates in 2015/16 85 Figure 78: Current expenses financed out of Figure 79: SEE expenses covered by the Union different accounts Budget Figure 80: Expenses covered by the Union Figure 81: Expenses covered by the Union Budget for SEEs with sufficient working Budget for SEEs with insufficient working capital capital Sources: MOPF, WB Staff estimates 51. Despite the short-term relief on Union Budget allocations, the sustainability of this policy will naturally depend on the current and future profitability of SEEs which, as discussed above, is declining rapidly. SEEs in the insufficient category started experiencing Net Losses in 2016/17. In response, two enterprises (Electric Power Generation Enterprise and Inland Waterways) were allowed to cover only 50 percent of raw materials from their Own Account, which was a regressive step. Additionally, Net Profits for SEEs in the sufficient category have also started to decline, which means slower accumulation of earnings in Own Accounts. Corporatized SEEs, especially CBM, have seen steady increases in their Operating Profits. Among corporatized transport SEEs, only Myanmar Port Authority saw some improvement in financial performance while others have not shown significant improvement. 86 Figure 82: SEEs Operating Profit/Loss by SEE Figure 83: Losses and Union Budget Support as category (million) percent of GDP Sources: MOPF, WB Staff estimates 52. In sum, even with the reduction in Union Budget allocations to SEEs, recent policy reforms have had limited impact in reducing implicit subsidies due to growing Operating Losses. SEE losses increased from 0.2 percent of GDP to 0.7 percent of GDP (Figure 83). These were in effect absorbed by the Union Budget and is also treated as a subsidy for SEEs. Unless the financial performance of SEEs improves, cash flow will be insufficient to cover operating expenses, and progressively capital investments. Close monitoring of both financial and fiscal performance is critical in this regard. SEE balance sheet and fiscal risks 53. Data availability and quality constrain a full understanding of SEEs’ net financial positions , net worth and solvency, and therefore any fiscal risks arising. Main data sources for this analysis are the SEEs’ balance sheets and debt profiles. Balance sheet data is not regularly reviewed or analyzed by the Budget Department as part of fiscal oversight. Given the lack of quality data, it is not possible to draw any general conclusions on the net worth of solvency of SEEs. This is a major gap in Myanmar’s SEEs management framework. The limited data made available highlights the importance of further analysis. Key indicators used in preliminary analysis are summarized below. Table 15: Suggested SEE financial solvency indicators Indicators Calculation Description Return on = Net Profit [Plus Contribution] Profits relative to assets; indicates how efficient Asset /Total Assets management is at using its assets to generate earnings. Return on = Net Profit [Plus Contribution] Profits per unit of shareholders' equity (Government in Equity /Total Equity or Capital this case). Used to measure financial leverage; indicates how much Debt to = Total Liabilities debt a company is using to finance its assets relative to Equity Ratio /Shareholders’ Equity the amount of value represented in shareholders' equity. A ratio of below 1 is deemed financially sound. Source: WB Staff 87 54. The 10 SEEs that reported as part of the PER have sizable assets and liabilities (Table 17). Their total assets were valued at 25 percent of GDP whilst total liabilities were valued at 10 percent of GDP. MPT’s assets alone were valued at 7.5 percent of GDP followed by MPE and MOGE at around 6 percent. Myanma railways’ assets were valued at 0.5 percent of GDP despite its vast holdings of land, which raised the question of valuation method and whether assets value of SEEs is underestimated in general. 55. Obtaining a comprehensive and timely picture of SEE assets and liabilities is critical to good fiscal management. Productive use of these assets can generate important fiscal returns. Assets need to be well managed; the role of state as shareholder should be to optimize returns on assets. It is also of critical importance to undertake proper asset valuation and exercise due caution against the risk of leasing out or selling assets managed by SEEs at below market price. The average return on assets and on equity of the 10 SEEs were 3 percent and 5 percent respectively. The average debt to equity ratio for 10 SEEs is less than one which implies that these SEEs are moderately leveraged and do not pose concern in terms of fiscal risks. However, further understanding of how balance sheet are prepared would help determine fiscal risk from SEEs financial positions more accurately. Table 16: Balance Sheet Summary SEEs Total Assets Total Total Capital Liabilities (Equity) Electricity Supply Enterprise 586,781 91,531 495,249 Myanma Oil and Gas Enterprise 4,347,505 2,201,685 2,145,820 Myanma Petrochemical Enterprise 4,862,775 1,817,465 3,045,311 Myanma Petroleum Products Ent. 177,295 31,037 146,258 Myanma Post and Telecoms 5,449,853 3,157,618 2,292,234 Myanma Railways 1,072,608 86,198 986,411 News and Periodical Enterprise 43,079 723 42,356 Road Transport 22,450 1,669 20,781 Yangon City Electricity Supply Corp. 370,774 45,289 325,485 Total 10 Selected SEEs 17,842,202 7,906,660 9,985,544 Source: MOPF 56. The Own Account balances for some SEEs, in particular Natural Resources and Telecommunications, are very large, which has implications for the opportunity cost of these resources. Own Account balances are captured in the above balance sheets. MPT (due to licensing payments) and MOGE’s Own Account balances were close to 9 percent of GDP at the beginning of the 2015/16 financial year. MOGE alone held own account balances of some 6 percent of GDP. As discussed above, there are now clear policies on the use of Own Account balances, which should be for the exclusive purposes of SEE expenses and not the Union Budget. This is important as it is likely to affect SEE incentives to pursue efficiency gains and generate higher Net Profits. 57. If on the other hand, the State felt that SEE earnings are best channeled to the Union Budget, then this should be done through adjustments to dividend/contribution policies. Currently, all SEEs are subject to the same dividend policy, which is that they transfer 20 percent of their earnings to the Union Budget. Contribution or dividend policy is something that should be determined at the individual SEE level. Contributions should be a function of factors such as existing financial conditions (Table 17), future prospects, and current and future investment needs. This might imply different contribution rates across different (types) of SEEs. This would be more difficult to administer, but is also likely to be a more efficient use of resources. Another issue to take into account is investment policy for Own Account balances. In the absence of State Enterprise Boards of Directors, which would ordinarily take decisions on such matters, the MOPF could invest Own Account resources to ensure a sound return on these important assets. 88 58. Debts are concentrated in a small number of SEEs. Outstanding debt of SEE as of June 2016 was around 8.35 percent of GDP. MOGE has the highest outstanding debt at 2.5 percent of GDP followed by Heavy Industry No.1 (1.8 percent), Electricity Generation Enterprise (1.1 percent), Myanma Railways (1.1 percent) and Myanma Petrochemical Enterprise (1.1 percent). Together these SEEs account for 90 percent of total SEE debt. Among the five SEEs with the largest amount of outstanding loans, three are loss-making– Heavy Industry No.1, Electricity Generation Enterprise, and Myanma Railways – which raises concerns on their ability to repay loans and more broadly on potential fiscal risks. 59. SEEs have not borrowed on particularly favorable terms, raising some concern over debt sustainability. Almost 80 percent of SEE debt is non-concessional. SEEs in the insufficient category are the most indebted, which also raises concern over debt sustainability and risks. The principal repayment profile up to 2054, shows that repayment will peak in 2021 with a steep rise in the repayment profile from 2016 and 2021. This enhances fiscal risks in light of declining profitability, slow growth in revenue, and rising operating expenses, which may further constrain ability to repay over the next five years if there is no turnaround in performance. SEE debt dominates the overall debt profile of the Government accounting for over 70 percent of interest payments and 59 percent of principal repayments. Box 11: Heavy Industry Enterprise No.1 Heavy Industry No.1 has six factories in operation employing more than 3,000 people. These include factories that produce diesel engines, excavation and lifting equipment, agricultural machinery and tractors. The remaining two are steel factories which were transferred to Heavy Industry No. 1 from the Myanmar Economic Corporation in 2012. Factory 16 and No. 1 Steel Mill are the most important employers each with more than 800 staff. The factories use technology and machinery from China, Germany, Japan, India and the Czech Republic to assemble their products. They produce for the private market and for the public sector. They sell to other line ministries, other SEEs as well as factories under Heavy Industry No. 1 – for example Factory 14 which produces diesel engines used in the excavation equipment produced by Factory 15. Other SEEs and Development Affairs Organizations (municipalities) are the most important market for this enterprise. Set up under the socialist era, Heavy Industry No. 1 has a mandate for regional development. Consequently, it has factories set up all over the country including Sinde (Bago), Malun (Magwe), Myingyan (Magwe), PangPet (Shan) and Thagaya (Bago). Unsurprisingly location has in certain cases been an important impediment to private investment, with high transportation costs deterring would be investors. As Myanmar opened up the economy, competition with the private sector has also been a major challenge for this SEE. In particular management cited that the machine refurbishment and renting of machines has been a major factor, significantly undercutting the price of the newly made machines produced by this SEE. The enterprise falls into the insufficient balance category and has run an operational loss in each of the last five financial years. This means that it is partly responsible for covering current expenditures. Anecdotal evidence suggests this may have incentivized some efficiency improvements – management described producing only for order where possible thereby reducing excess stock and wastage. Despite this, challenges remain and management cited the difficult of obtaining sufficient capital budget and increasing loan servicing costs on external loans as a result of dollar appreciation. 89 Myanmar’s Industrial Strategy sets out a policy to reduce state operational involvement in industry to focus more on policy and regulation. Heavy Industry No. 1 has already undergone significant restructuring in recent years. Factories 11, 12 and 13 which produce jeeps, light trucks and larger trucks were transferred to the Ministry of Defense last year. Privatization has been explored for Factories 14, 15, 16 and 17, as well as No. 1 Steel Mill. These privatizations have moved at different speeds and some have stalled completely prompting the exploration of different approaches to securing private investment such as Joint Ventures and contracting. Whatever the approach, these changes suggest much reduced operations for the SEE in the future. This seems sensible given operational losses and rising exposure to competition from the private sector however the Government will need to keep a close eye on risks such as the mispricing or valuation of assets and possible staff redundancies. Source: WB Staff based on interviews POLICY OPTIONS 60. Whilst SEEs have been important contributors to the Union Budget, recent developments and a challenging financial outlook suggests the need to further strengthen SEE reforms and fiscal oversight. The financial performance of the SEEs sector as a whole is likely to remain under pressure in the coming years. Firstly, gas prices coupled with declining production from existing fields will impact gas sectors receipts, which dominate SEE payments to the Union Budget. Secondly, commercial SEEs will continue to find it increasingly difficult to compete against cheaper and better quality products from the private sector, bolstered by foreign investment and trade. Thirdly, the rapid expansion of public sector utility provision without cost recovery, particularly in the power sector, has resulted in increasing Operating Losses in the Infrastructure sub-sector. 61. Experience from other countries suggests that reform of Myanmar’s SEEs will take time, and given the inherent risks, should be undertaken via a phased approach. While the number of distinct SEEs and their total employment makes reform of Myanmar’s SEEs somewhat more manageable compared to other transitional economies such as Vietnam and China, the fact that SEE-driven revenues and expenditures account for a substantial share of the Union Budget suggests a cautious and incremental approach to reform is warranted. Suggestions are made below, focusing on the fiscal management aspects of Myanmar’s SEEs. Prioritizing reform and fiscal oversight by SEE typologies 62. Myanmar’s SEEs are diverse in terms of sectors, levels of employment, financial performance, asset base, and impact on fiscal outcomes. This calls for a differentiated approach, particularly as far as reform priorities and fiscal policies are concerned. Therefore, the first policy option is to put in place a new classification framework for Myanmar’s SEEs with four broad typologies. The MOPF should then develop a specific policy for each group for the state as a shareholder. The suggested groups and policy direction is below: (i) Natural resource SEEs: where the objective is to maximize the return on publicly owned endowments, within a broader framework to ensure environmental and social sustainability. Natural resource SEEs warrant separate treatment given the uniquely high returns from national assets. 90 Elements of such a treatment could include: (i) a clear separation of operational and regulatory functions, in this case removing the regulatory functions from the Ministry of Energy and Electric Power and the Ministry of Mines and from sectoral associations, which among other things is critical to avoid conflicts of interest in negotiating fiscal terms and conditions with private companies; (ii) a sound and transparent fiscal regime for natural resources with adequate administrative and audit capacity to ensure efficient collection of taxes, royalties, and other revenue; (iii) policies for managing revenue from natural resource SEEs including adequate forecasting for medium-term expenditure management, managing revenue volatility, and developing a clear savings policy. Putting in place a customized fiscal regime for natural resource SEEs would also allow for a greater net transferal of resources from SEE operating accounts to the Union Budget, recognizing that Myanmar’s abundant natural resources endowments provide the country with the opportunity to transform a portion of this natural capital into investments in human and physical capital, thereby laying the foundations for sustainable growth and poverty reduction. The rate at which this transformation takes place, the necessary safeguards that need to be in place to ensure the sustainable extraction of natural resources, and the decision making process for investing the proceeds, are all important issues of public policy. Such decisions should be taken at a level of Government that can weigh these broad tradeoffs, rather than at the sectoral or SEE level. (ii) Commercially viable SEE, without need for Union Budget subsidy: where activities present a net positive return on investment over the medium-term with potential for full cost recovery. These SEEs could progressively be granted increased fiscal autonomy through greater retention of profits for potential investments, which would need to be appraised and overseen by the MOPF just like any other public investment (see capital expenditure chapter) or a professional board acting in the interest of the MOPF. These SEEs should be corporatized and made to adopt modern corporate governance standards, reporting requirements and disclosure standards. This could form the basis for private sector participation at a following stage. (iii) Commercially viable SEE, with need for Union Budget subsidy: where a net positive return on investment is possible over the medium-term, but not necessarily under conditions of full cost recovery. Public utilities, which provide positive societal benefits would fall under this category and could justify additional subsidy. Such public service obligations would need to be carefully defined and costed for the purposes of transparency. These could be underpinned by a broader restructuring plan, which would cover aspects such as human resources, corporate finances, and others, which would help attract much needed external investment. This should lead to a consideration of the best way to deliver these public service requirements. Options could include provision through SEEs through public service agreements or under the supervision of an independent regulator but also directly contracting the private sector. (iv) Non-viable SEE: where the objective of the state is to cease activities, either by reabsorption of publically desirable functions into the activities of Government, or the sale/divestiture/closure of SEEs. It is critical that this be conducted in a highly transparent and open manner. The Government could even consider granting the responsibility for managing sale or divestiture processes to an independent agency to ensure a process that is beyond reproach. 91 Performance monitoring 63. The PER highlights opportunities to build on existing monitoring practices to significantly strengthen fiscal oversight of SEEs and implement the types of policies highlighted above. One first important step is to adopt consistent accounting standards and definitions in reporting SEE performance information collected in the annual forms 1-18. As highlighted above, some adjustment is warranted to move towards more accepted standards in terms of performance indicators. For little cost, this would significantly improve the interpretation of performance information for more timely policy responses. Secondly, there is scope to automate data collection. The current manual process is very time consuming and leads to challenges in terms of completeness and accuracy of reporting, with considerable variation in quality and coverage across SEEs. This would help free up staff time in the MOPF Division for analysis and reporting rather than data management. Thirdly, there is also scope to simplify reporting requirements and reduce redundancy and duplication, which would in turn also help with the quality, and potentially the frequency, of information submitted by SEEs.67 Box 12: Options for SEE financial performance monitoring in Myanmar Based on the experience of other transition economies, and existing data availability in Myanmar, the PER proposes some options for performance monitoring of SEEs in Myanmar, to complement information covered above. Economic Value Added is calculated by deducting the cost of capital from operating profit with adjustments made for taxes. The idea behind this indicator is that it gives an indication of the true economic profit of a company as it deducts the opportunity cost of making a specific investment (the cost of capital) from the investment. It would be possible to assess this with existing data in Myanmar. Net Profit measures profitability of a company after all costs have been accounted for. In Myanmar the net profit line does not include all income lines which is why the PER focuses on Operating Profit/Loss in the section above. Myanmar would be able to calculate this with existing data but it may want to make some adjustment to the calculation to ensure all relevant incomes are included. Financial Ratios refer to set of indicators that can be used to assess company performance, activity, financing and liquidity. Measuring company performance includes many of the measures covered in this box such as net profit and return on investment. It includes the debt to equity ratio as defined above – it would be possible to calculate these subject to improved coverage of SEE balance sheet information. Liquidity measures include interest coverage and working capital, which provides an indication of whether an SEE can meet its recurring expenses. Activity indicators include asset and inventory turnover and provide an indication of how efficient the company is at turning assets into sales – better data on assets and inventories would be needed to estimate this. Labor Productivity measures the amount of goods and services produced by a given unit of labor. It should already be possible to estimate this in Myanmar. 67 In terms of specific improvements to the forms, these could include the following: i) clearly distinguished forms according to purpose (for example, for the purpose of budget submission, performance monitoring, financial monitoring); ii) streamlined to reduce administrative burden in preparing and compiling the information (eliminate forms which are no longer relevant and for which data is regularly used by MOPF) and instead ensure that critical information is submitted in a timely manner (e.g. Balance Sheet); and iii) review the calculation methodology and how items are classified for profit and loss and balance sheet forms to further align with international accounting standards and standardize among SEEs (redesigned forms accompanied with guidance on how to fill them in), which will be key to accurate interpretation of data and analysis of fiscal risks. 92 Return on Investment measures net profit as a percentage of the investment. It is also common to see return on capital, return on assets and return on equity calculated. This provides a measure of how efficient the investment, the equity, the assets or the capital is at generating profit. This would require better balance sheet information. Benchmarking with private sector data: It is often useful to compare SOE indicators with private sector comparators. This enables a Government to better assess whether a SOE is performing well in comparison to the private sector. To do this well it is important to have good performance data on private sector performance in comparable sectors which often involves high quality survey work. Identifying comparators needs to be carefully done so that a Government is setting the correct benchmark for its SOEs. Some Governments use such data to set performance targets for their SOE sector. Medium-Term policy framework for SEEs 64. Recent policy shifts offer an opportunity to reform Myanmar’s approach to SEE ownership, regulation and management. A major challenge with the current system is that too many functions are fulfilled with sector ministries, which causes conflicts of interest, exacerbated by staff moving within the Ministry from function to function. The regulatory function should be separated from the SEE. It should not be fulfilled by sector associations and should it stay in the ministry, there should be appropriate firewalls between the regulatory function and the SEE including rules about staff movement between the two. As SEEs move towards corporatization, professional managers should be recruited, ideally reporting to a professional board acting in the interests of the ultimate shareholder – MOPF on behalf of the Government. 65. Further, it would be helpful to build over the medium term a policy framework for the overall management of SEEs. This would set out the overarching objectives and rational for state investment via an SEE, list the range of corporate structures that can be applied to Myanmar’s SEEs, based on a revised typology of activities, and establish clear governance arrangements with specific roles and responsibilities for Myanmar’s key fiscal institutions. The development of an updated SEEs policy framework could also bring clarity to the current fragmentation of governing legislation in the sector, and establish clear principles of ownership. 66. Given the relatively limited number of SEEs, a centralized ownership policy would be feasible. A distinction would need to be made between commercially-oriented activities versus public-service obligations, with the establishment of a clear framework for costing such public service obligations. This policy framework would allow for progressive increases to the operational independence of some SEEs, with stronger corporate governance and oversight mechanisms including the application of private sector norms, accounting standards and board supervision to maintain fiscal oversight. However, for many activities, remaining close to the state budget oversight and control framework would be appropriate in order to manage fiscal risks. 93 Table 17: Policy options for strengthening fiscal oversight of SEEs Issue Policy options Myanmar’s SEEs are diverse in terms of Short-term: Put in place a new classification framework sectors, levels of employment, financial for Myanmar’s SEEs with four broad typologies (natural performance, asset base, and impact on fiscal resources; commercially viable SEEs without need for outcomes—differences are particularly public subsidy; commercially viable SEEs with need for apparent between natural resources, public subsidy; and, non-viable SEEs) infrastructure and other SEEs. Incremental Short-term: Determine customized contributions at reforms have served Myanmar well thus far, individual SEE levels (instead of uniform rates) to take but a more strategic classification framework is into account financial conditions, future prospects, necessary for future growth and management. current and future investment needs The current manual fiscal reporting process is Short-term: Build on existing monitoring practices to time consuming and leads to challenges in significantly strengthen fiscal oversight, and in particular: terms of completeness and accuracy of data, with considerable variation in quality and • Adopt consistent accounting standards and coverage across SEEs. This makes it hard to definitions in reporting SEE performance policymakers to understand emerging fiscal information collected in the annual forms 1-18. trends and risks associated with Myanmar’s • Automate data collection SEEs. • Simplify reporting requirements Short-term: Prepare annual simple SEEs performance monitoring report as part of the Citizen’s Budget Weaker commodity prices and greater Medium-term: Build a policy framework for the overall competitive pressures suggests the need for a management of the SEEs sector with clear and specific medium-term policy framework that would roles for Myanmar’s fiscal institutions. chart the next phase of reforms for Myanmar’s SEEs. 94 TAX SYSTEM EFFICIENCY 1. Myanmar’s comprehensive tax reform program is gradually improving revenue collections, which nevertheless remain well below their current potential. Tax reforms initiated in 2012 include efforts to widen and redefine the tax base, redesign tax instruments, and strengthen tax administration capacity. If successful, these reforms could set Myanmar on a strong revenue growth path. Preliminary reform results, in regard to revenue collection and administrative performance, have on the whole been positive. Though there is a long way to go. Despite important structural constraints to revenue collection, there could be potential to further stem near-term revenue leakage, expand the tax base, and focus limited administrative capacity where they are most needed. 2. This chapter analyzes the qualitative and quantitative results of tax reforms to date, and highlights potential short-term priorities for enhancing revenue collection and improving tax system efficiency. It includes: (i) an overview of revenue profile, including international comparators and trends; (ii) a review of tax administration and policy reforms, including preliminary results achieved; (iii) an analysis of the potentially big sources of near-term revenue leakages (e.g. tax incentives and transfer pricing); and (iv) policy options to help address some of these challenges. REVENUE PROFILE AND PRINCIPLES FOR REFORM 3. Myanmar has one of the lowest levels of Government revenue mobilization at 10-12 percent of GDP whereas its current potential could be in the range of 15-20 percent of GDP.68 A lot of economic activity is in ‘hard to tax’ sectors like agriculture, and/or dominated by many small and micro enterprises that are outside of the tax net. At the same time, the legacy tax administration system, discussed below, together with decades of underinvestment, have resulted in a major erosion of Myanmar’s tax base. Despite good progress in recent years, tax receipts are at around 6-7 percent of GDP (Table 19) compared to between 10-20 percent of GDP for countries at similar levels of income (Figure 84). In other fragile states, tax collections have recently ranged between 14-16 percent of GDP (Figure 85), driven by relatively high natural resource rents. Table 18: General Government Revenue (share of GDP) 2013/14 2014/15 2015/16 2016/17 Revenue 10.1% 12.2% 12.1% 9.2% Tax 6.7% 7.0% 6.7% 6.4% Income Tax 3.1% 3.4% 3.2% 3.2% o/w SEEs 1.0% 1.2% 0.9% 0.6% Commercial Tax 2.9% 2.8% 2.9% 2.6% Customs duties 0.6% 0.8% 0.6% 0.6% Excise Tax 0.0% 0.0% 0.0% 0.0% Non Tax 3.4% 5.1% 5.4% 2.8% Receipts on use of national properties 0.4% 2.3% 1.3% 1.0% o/w oil and gas 0.4% 0.5% 0.5% 0.8% o/w telecommunications 0.0% 1.7% 0.7% 0.2% SEE contributions 1.3% 0.8% 1.1% 0.7% License and fees 0.5% 0.5% 0.4% 0.2% Other 1.2% 1.5% 2.5% 0.9% Sources: MOPF, WB Staff estimates 68Excluding tax and non-tax receipts from large mining activities. The potential revenue range is derived from revenue collection in countries at similar levels of income. 95 Figure 84: Tax/GDP vs. GDP/capita (2010 - Figure 85: Tax/GDP across Regions (2000 - 2016 average) 2016) Myanmar Sources: WDI, MOPF and WB Staff estimates. Note: Regional averages exclude High Income Countries 4. Tax revenues account for around 60 percent of Union Government receipts, which in Myanmar are more stable and correlated with economic activity than non-tax revenues. Thus anchoring spending decisions against expected tax receipts is important for fiscal sustainability. Non-tax receipts have recently included one off telecom licensing fees, which contributed to windfall earnings; and gas sector royalties, which depend on gross earnings from gas sales. Other important non-tax receipts are dividend payments (or contributions) from State Economic Enterprises, which are on a declining path (see chapter on SEEs). 5. Myanmar also has high non-tax potential, particularly from natural resource rents. Though not a focus topic of the PER, it is important to note the significant revenue leakage from mining operations in border areas. For countries in the Middle East and North Africa Region over half of General Government revenue are from non-tax receipts, with much coming from natural resources (Figure 86). Outside of this, there are fees and charges on public services, which in Myanmar could become more important in the context of potential electricity tariff adjustments. At higher income categories, social contributions can play a big role in non-tax receipts, but are less relevant for Myanmar now. Figure 86: Tax revenue/Total revenue Figure 87: Consumption tax/Total rev enue Sources: WDI, MOPF and WB Staff estimates. Note: Regional averages exclude High Income Countries 96 6. Indirect taxes in Myanmar account for around one third of revenue collections, similar to other developing countries, whose dependence on these regressive taxes have on average increased (Figure 87). Indirect taxes, the bulk of which come from commercial taxes in Myanmar, are levied on consumption and tend to be easier to administer compared to direct taxes, which are levied on earnings of individuals and firms. But indirect taxes with uniform rates have higher relative incidence on poorer households’ incomes, unless multiple rates are introduced, which in turn could reduce the efficiency of the tax system. The regressive impact of higher consumption taxes could be offset by targeted social protection spending. In Myanmar, however, Government expenditure on social protection is very low, though targeted assistance programs (e.g. in response to electricity tariff adjustments noted above) are likely to become more important. 7. Within indirect taxes, customs receipts do not play a significant role, as is the case for a growing number of developing countries following implementation of trade agreements (Figure 88). Customs receipts as a share of total revenue in Myanmar is similar to the average for 40 Low and Lower Middle Income Countries, despite import levels being significantly less. Nevertheless, recent data shows that customs receipts have been growing at a slower pace than imports despite ongoing customs reforms (Figure 89). This is likely due to declining tariffs or exemptions on customs duties, particularly for new investments. Until recently, all new foreign investments approved by the Myanmar Investment Commission were exempt from customs duties in the project establishment phase. Figure 88: Customs receipts vs. Imports (2010 - Figure 89: Evoluation of customs receipts vs. 2016 average) Imports in recent years Sources: WDI, MOPF and WB Staff estimates. Note: Normal Trade excludes Border Trade. 8. The share of direct taxes in total revenue in Myanmar is somewhat higher than in other LICs and LMICs, where direct taxes to GDP nevertheless remains higher. Corporate income tax receipts in Myanmar account for over a third of total revenue collections, compared to around 10-20 percent for countries in other regions (Figure 90). In many other countries, corporate income tax rates have been declining with the aim of boosting production and supply side response. Concurrent efforts to expand the tax base have helped to sustain corporate tax receipts, which as a share of GDP remain higher than in Myanmar. Nevertheless, from the 1990s to the early 2000s, corporate tax receipts as a whole fell by roughly one-fifth with the most marked drop taking place in the ECA Region.69 Personal income tax (PIT) on the other hand, accounts for a very small share of overall receipts, and is also a small share of commercial profits in Myanmar. It is also far below the average for other LICs and LMICs (Figure 91). PIT falls on formal sector employees subject to withholding. 69 Keen, Michael and Simone, Alejandro, “Tax Policy in Developing Countries: Some Lessons from the 1990s and Some Challenges Ahead.” In Clements, Benedict, Gupta, Sanjeev and Inchauste, Gabriela, eds. Helping Countries Develop, the Role of Fiscal Policy.Washington: IMF, 2004 97 Figure 90: Income tax/Total revenue Figure 91: Labor tax/commercial profits Sources: WDI, MOPF and WB Staff estimates 9. Revenue collection, disaggregated by States/Regions as well as agencies, highlights the important role played by Yangon in overall revenue collection. Yangon and union territories (which draw revenue largely from economic activity in Yangon) accounted for around 40 percent of total revenue (Table 20). The Companies’ Circle Tax Office, which covered all large and medium enterprises, unsurprisingly accounted for the bulk of revenue collection until FY 2014/15. It was thereafter carved out into the Large Taxpayers’ Office and the Medium Taxpayers’ Offices (1, 2 and 3), which explains the sharp drop in CCTO collections in 2015/16 being offset by the LTO and MTO. Table 19: Tax Revenue Collection by States & Regions / Agencies (% of GDP) Source 2013/14 2014/15 2015/16 Mon 0.01 0.01 0.02 Ayeyarwaddy 0.03 0.03 0.04 CCTO 2.76 2.86 - Chin 0.00 0.01 0.01 Kachin 0.01 0.03 0.03 Kayah 0.00 0.00 0.01 Kayin 0.03 0.04 0.06 Lottery Sales Central 0.01 0.01 0.01 Lottery Sales Mandalay - - 0.01 LTO - 0.21 1.49 Magwe 0.05 0.04 0.04 Mandalay 0.09 0.12 0.21 MTO (1) - - 0.60 MTO (2) - - 0.90 MTO (3) - - - Pago 0.04 0.04 0.04 Rakhine 0.01 0.01 0.02 Sagaing 0.03 0.04 0.05 Shan 0.23 0.27 0.28 Thanintharyi 0.03 0.04 0.04 Union Territory 1.51 1.43 1.03 Yangon 1.20 1.47 1.35 Total 6.05 6.67 6.24 Source: Myanmar Internal Revenue Department (IRD) 98 TAX REFORM PROGRESS AND ACHIEVEMENTS 10. Tax reforms in developing countries typically have three main components: expanding the tax base, rationalizing rates and improving tax administration.70 Expanding the tax base can improve both efficiency and equity of the system. It is typically achieved by (i) bringing untaxed goods and services into the tax system; (ii) reducing exemptions and (iii) improving taxpayer identification and registration functions. Rationalizing tax rates involves following certain principles including (i) simplicity, through elimination of excessive slabs and exemptions; (ii) for indirect taxes such as VAT, a broad base, single rate and high thresholds, with possible exemptions for certain goods; (iii) for CIT and PIT, coordinating top rates to ensure international competitiveness and minimizing exemptions and incentives. Tax policy needs to be adequately supported by a strong tax administration that (i) ensures compliance; (ii) reduces administrative burden on taxpayers, and (iii) reduces discretion of tax authorities so as minimize the scope of corruption and leakage. 11. The sequencing of reforms along these three components varies by country (Box 13), but in countries rebuilding fiscal institutions, such as Myanmar, the binding constraint in the short run is administrative capacity.71 In such contexts, trying to expand the narrow tax base by introducing new taxes is likely to be ineffective, as revenue administration tend to lack basic infrastructure, trained staff and capabilities for carrying out basic taxpayer functions such as identification, registration, filing, payment and audits72. Therefore, strengthening tax administration, while ensuring that tax policy is simple to administer with few tax handles and limited exemptions, is the most feasible short-run reform approach. Box 13: Revenue reforms and results – Selected country experiences Cambodia has seen significant gains in revenue mobilization since 2011, with revenue to GDP increasing from 15.7 % of GDP in 2011 to 18.8% of GDP in 2016 (projected). This has been underpinned by an increase in tax collection, from 10.2% of GDP in 2011 to 14.9% of GDP in 2015, with a rising share of direct taxes in overall tax collection. The gains in revenue performance have been driven by an expansion of formal taxable sectors such as construction and their increasing share in the economy as well as the Government’s Revenue Mobilization Strategy (RMS) reforms. The RMS has focused on strengthening tax administration and improving tax compliance, especially through improved audits, enhanced enforcement, arrears collection and implementation of import tariffs. During this period, Cambodia has refrained from major tax policy reforms such as introducing new taxes or significant reform of tax rates or structures. Source: Cambodia Economic Updates, April and October 2016, World Bank. Vietnam transitioned from an Officially Administered System (OAS) to a Self-Administered System (SAS) in the early 2000s – first, by piloting the SAS in two provinces in 2004 and then building on the pilot to introduce SAS to all provinces in 2006. The reform was accompanied by significant building of technical capacity and retraining of tax inspectors and auditors, as well as institutional redesign. The results of the change have been impressive, contributing to lower compliance costs and increased tax yields. Vietnam’s Government revenues increased from 18 percent of GDP in the late 1990s to 27 percent in 2011, partly on the back of these tax administration reforms. Source: Myanmar Policy Notes, January 2016, World Bank. 70 “Creating Fiscal Space Through Revenue Mobilization: South Asia Economic Focus”, World Bank (2012) 71 “Rebuilding Fiscal Institutions in Post-Conflict Countries”, IMF (2004) 72 Ibid. 99 Georgia’s public administration at the time of the ‘Rose Revolution’ in 2003, including tax administration, was in crisis – confidence in the tax system was low and revenue collections of 12 percent of GDP were inadequate to finance Government expenditures creating arrears. The Government undertook major legal, regulatory and administrative reforms to improve the tax system, including: (i) reducing the number of taxes from 21 to 6; (ii) lowering tax rates across the board, and unifying tax rates for all businesses getting rid of ‘special’ tax regimes for different types of businesses; (iii) streamlining and automating most processes, including introduction of risk based audits and e-filing and (iv) strengthening electronic information sharing between the revenue service, banks, etc. and (v) improving taxpayer services and introducing greater transparency in interactions between citizens and tax authorities. These reforms have been very successful (i) in doubling revenues from 12 percent of GDP in 2003 to 25% in 2012; (ii) in reducing administrative costs for tax authorities and compliance costs for taxpayers; and (iii) significantly reducing the number of bribery cases filed against tax authority officials. Source: Examples of Successful Domestic Revenue Mobilization Reforms, OECD, 2015. Paraguay underwent an economic crisis from 2000 to 2004, with domestic revenues falling to 8.8 percent of GDP in 2003, among the lowest in the world. In response, the Government of Paraguay undertook wide ranging reforms including (i) tax policy reforms including reduction of corporate tax rates on profits, introduction of a special tax on the agriculture sector and reduction of VAT exemptions and (ii) tax administration reforms, including creation of a Large Taxpayer Unit and a focus on increasing the VAT effective tax base. The reforms have led to an increase in revenues, from 8.8 percent in 2003 to 13.8 percent in 2013, with tax administration reforms especially bearing fruit – with reduced compliance and administrative costs, and a 50 percent increase in VAT gross compliance ratio. Source: Examples of Successful Domestic Revenue Mobilization Reforms, OECD, 2015. 12. Myanmar commenced tax reforms in 2012, starting from an outmoded tax administration and low tax collection. Prior to reform in 2010/11, the tax to GDP ratio (at less than 4 percent) and revenue to GDP ratio (at 6.3 percent) was considerably lower than other countries at similar levels of income. 73 The tax administration was characterized by manual processes and major gaps in basic taxpayer functions, contributing to low levels of compliance.74 In addition, tax liabilities were administratively assessed by tax authorities following negotiations with taxpayers, lowering the transparency of the system and raising risks of collusion.75 Myanmar authorities identified the key objective of reform as “arriving at a system that is efficient, equitable and simple, delivering necessary resources to Government while minimizing the administrative burden on taxpayers”.76 13. Tax reforms are phased according to capacity constraints and the complexity of changes. The first phase of reforms from 2012 to 2017 has focused on modernizing tax administration with some gradual changes on expanding the tax base. Reforms have focused on two fundamental transformations: from a system of administrative assessment to a system of self-assessment, and moving from a tax-type organization to a functional organization. Some reforms in the first phase have also focused on expanding the tax base – such as the revision of the Income Tax Law to cover public servants and military owned companies and the adoption of Specific Goods Tax Law.77 In 2016, the authorities rationalized tax rates for Commercial taxes. Income and commercial tax rates remain relatively uniform. 73 Myanmar Public Expenditure Review 2015, World Bank. 74 “Modernization of Public Financial Management Project: Project Appraisal Document”, World Bank (2014). 75 Ibid. 76 Ibid. 77 The tax base includes activities, revenue streams and conditions that are subject to tax. 100 14. Reform progress on modernizing tax administration has been positive, as highlighted in the section below, supported by assistance from development partners such as the World Bank. Myanmar has successfully established the Large Taxpayer Office, and piloted improved taxpayer functions within the LTO. The most significant has been the successful introduction of self-assessment, coupled with tax audits, within the LTO. The changes have been supported under the World Bank Modernization of Public Finance Management Project (MPFM project), and with technical assistance from other development partners including the International Monetary Fund (IMF) and United States Treasury Office of Technical Assistance (USOTA). 15. Translating reforms into improved revenue collections has proved challenging in the short term. There are both policy and structural constraints to the efficiency and equity of the tax system, which impact on the sustainability of reforms and how quickly they can generate higher revenues. On policy constraints for example, the widespread use of tax incentives (discussed below) means that effective tax rates (i.e. actual tax liability over tax earnings) can vary across and within sectors, even though tax rates are relatively uniform. This is not only inefficient, but could also imply revenue foregone without incremental investment. On more structural, long-term constraints, natural gas related revenues still comprise a major part of overall revenue collection. As discussed in detail in Chapter 1, gas related revenues are volatile and subject to changes in global gas prices which have trended downwards in the last two years. Tax administration reform progress 16. A major hallmark of Myanmar’s tax administration reform in Phase 1 is the transformation from official assessment to self-assessment. An official assessment system implies that the taxpayer presents his/her books to the tax authority and, subsequent to negotiations, the taxpayer and the authority agree upon a tax liability. Self-assessment shifts the initial burden of compliance and reporting to the taxpayer78, and, since 2012, the InternalRevenue Department (IRD) has been implementing a plan to sequentially move tax administration to self-assessment over time. 17. The move to self-assessment has also been accompanied by a move to restructure tax administration to focus on taxpayer segments by size. The early phase of reform, since 2012, has focused on the establishment of: (i) A Large Taxpayer Office (LTO) to handle not just large businesses by businesses in complex sectors; (ii) 3 Medium-sized taxpayer offices (MTO1, MTO2, MTO3); (iii) Small Taxpayer Offices in the regions. Segmentation by taxpayer size is considered global good practice, as it allows both audit and taxpayer service to focus on the needs of the taxpayer segment. For example, large taxpayers usually like to have a relationship with the tax administration to ask for guidance (and rulings eventually) ahead of making transactions to prevent a problem later, while small taxpayers usually need taxpayer service to guide them on how to fill out tax forms.79 78 The tax payer is required to build a case that supports the level of economic activity (profit, revenue, and cost) which underpins the final tax liability. Thus a level of “trust” must be accorded to the tax payer, which over time with a compliance track record is confirmed or not. The tax payer is required to be the basic source of data for the LTO, but in order to corroborate data, the LTO must collect other information from the third party sources. Finally the burden is on the tax payer to prove any challenges to the data or “story”. 79 Administration of SAS requires building a taxpayer profile, which involves: (a) Compulsory documentation (filing forms) which allow the LTO to understand business operations including profits and losses and their components, financing operations and tax planning strategies; and (b) Third party information to supplement taxpayer data which includes taxpayer data, establishing market conditions and competitiveness and establishing the international environment for comparison and to understand how a multinational firm works 101 18. Self-assessment rollout started with the LTO, since FY 2014/15 and will continue with the MTO in FY 2017/18. The LTO was established in 2013, with the first two years focused on building institutional capacity, especially on conducting audits, to facilitate self-assessment. FY 2014/15 was the first self-assessment filing season for the LTO, with all 544 taxpayers in this segment being self-assessed, and the third filing season is currently (FY 2016/17) underway. MTO1, MTO2, and MTO3 have been created out of the Companies Circles Tax Office (CCTO), which oversees medium (and large-medium) enterprises. This is the first year (FY 2016/17 filing season) where MTO1, which covers the largest of the medium sized enterprises, will submit self-assessment forms, while MTO2 and MTO3 will transition to self-assessment in the next two fiscal years (Table 21). 19. The establishment of the LTO has been supported by a multi-donor capacity building effort focused on core operations, especially around audit capacity. The core of LTO work rests on knowing your client, which necessitates both understanding of your clients’ business structure, the competitive environment in which it operates, and also the psychology of the client’s willingness to comply with the law. Each of these areas is unique to a self-assessment system. LTO has benefited greatly from a multi-donor capacity building effort to train all aspects of LTO operations, and is a key component of the World Bank MPFM project. 20. The LTO has already achieved some notable gains, including a filing rate of 95 percent in their first year and in conducting tax audits. The LTO has also successfully issued unique Tax Identification Numbers (TIN) to all large taxpayers. In the first two years of self-assessment, filing rate of 95 percent was achieved – a notable accomplishment when compared to tax administrations in other countries. With multi- donor support, LTO has also launched an audit process and carried out 37 planned audits in FY 15/16 and plans on increasing the number of audits to 60 during the current fiscal year. Table 20: Myanmar - Summary of Number of Self -Assessed Taxpayers by Segment Taxpayer segment No. of taxpaying firms Start of self-assessment Firms self-assessing as of FY 2016/17 Large Taxpayer Office 544 FY 2014/15 544 Medium Taxpayer Office 1 1,200 FY 2016/17 1,200 Medium Taxpayer Office 2 1,199 FY 2017/18 N/A Medium Taxpayer Office 3 16,905 FY 2018/19 N/A Small taxpayer offices (TBD) TBD TBD TBD 19,848 1,744 Source: Myanmar Internal Revenue Department (IRD) 21. An important result has been the availability of greater information on tax returns of large taxpayers, which allows analysis of sources of revenue leakages including insufficient coverage of the tax base and tax exemptions. For instance, a US Office of Technical Assistance (OTA) led analysis, based on the LTO audit data, identified that large multi-national enterprises seemed to be absent from the list of the first 100 large taxpayers to be audited, suggesting major gaps in the coverage of the tax base. Furthermore, the LTO tax return data highlights that income tax exemptions for 13 large taxpayers resulted in a revenue loss of K41 billion, or 4% of total LTO collections, in FY 2015/16, hinting at the scale of potential revenue leakage. 22. LTO revenue collection (preliminary estimates) highlights a slow but continuous growth in nominal income and commercial tax collection from large taxpayers (Table 22). Although commercial tax is not subject to self-assessment, the growth in commercial tax is likely a result of strengthened administration resulting in a spillover effect, in line with taxpayer expectations of greater scrutiny. 102 Table 21: LTO Tax Collections – 2013/14 to 2015/16 2013/14 2014/15 2015/16 Billions of Kyat Income Tax 619,491 629,211 637,883 Commercial Tax 374,981 394,689 420,378 Total 994,472 1,023,900 1,058,261 % Growth from previous year Income Tax 1.57 1.36 Commercial Tax 5.26 6.11 Total 2.96 3.25 Source: WB staff estimates, IRD/LTO preliminary estimates 23. However, despite the progress made, especially in the increase in filing and compliance of large taxpayers, revenue increase has been modest and weighed down by changes in depreciation and carryover losses. Estimates from the LTO suggest that losses from depreciation and carryover amounted to K 136 billion, or approximately 13 percent of total tax take from LTO in FY 2015/16. With regard to carryover losses, the key concern derives from the official assessment of medium taxpayers by the former CCTO, wherein taxpayer report of losses was insufficiently verified and reviewed. These large quantum of losses have subsequently been carried over by firms as they migrate to LTO, which results in claim for tax deductions – for instance, two new telecom companies, beverage companies and others that initiated business in 2013/2014 reported no income, with unverified losses with entire amount was carried forward by taxpayers to 2014/2015 and beyond. Tax policy reform progress 24. Tax policy reforms in Phase 1 have focused on streamlining Myanmar’s major direct tax instruments, through the enactment of an Income Tax Law (ITL) in 2012, and, on the indirect tax side, the enactment of the Specific Goods Tax (SGT) in 2015. Myanmar’s tax instruments involve a mix of direct and indirect taxes, as in other countries in the region. The tax structure currently comprises 15 different types of direct and indirect taxes -- details on Myanmar’s tax handles are summarized in Table 24 below. The ITL allowed foreign transactions to be covered; military owned companies also now pay income tax. The ITL rationalized rates by lowering them on Corporate Income Tax, Personal Income Tax, and other withholding taxes. This effort has resulted in Myanmar’s tax rates being comparable to other countries in the region, with a caveat that Myanmar has a commercial tax rather than a VAT (Table 23). The SGT, functioning similar to an excise tax, covers tobacco and alcohol products, teak and hardwood logs, precious stones, jewelry, luxury cars, fuel and natural gas. Table 22: Tax Rates in Myanmar and neighboring countries (%) Country VAT CIT PIT Rate Rate Rate Cambodia 10 20 20 Lao PDR 10 24 24 Malaysia 6 25 25 Myanmar 5 25 20 Thailand 7 20 35 Sources: PWC and KPMG 2016 Note: Myanmar uses Commercial Tax instead of VAT. 103 25. Reform and modernization of tax policy is still in process, and is proceeding as per the Government’s reform roadmap, prepared with assistance from the International Monetary Fund in 2014. The reform vision envisages a deepening of direct tax policy (both personal and corporate) to include provisions to tax passive income (for example, on dividends, capital gains, and other investment income); international tax provisions (including on arm’s length principles and transfer pricing); and clarifying rules such as on permanent establishment. On the indirect tax side, the reform vision includes a plan to replace the commercial tax with a value-added tax and reform excise taxes, over the long term. Both direct and indirect tax reform are envisaged through a comprehensive redraft of primary legislation (income, VAT, and indirect tax bills), along with secondary legislation and guidance. Table 23: Overview of tax instruments in Myanmar Tax Type Legal Basis Key Provisions Personal Income Tax PIT applies to all resident individuals in Myanmar, foreign non- residents who Income Tax Law (2012) have income sourced in Myanmar, and citizens of Myanmar who reside outside the country (non-residents).80 Corporate Income Tax Tax Rate: All active income is taxed at a 25 percent, but dividends and interest Income Tax Law (2012) income are not taxed. Sector fiscal regimes legally override any CIT provision in the income tax law (provision 5d of the Income Tax Law) Loss carry forward: Losses can only be carried forward for 3 years, but some companies that receive investment incentives may be allowed to carry forward losses indefinitely. The SEZ law allows carry over for 5 years from the year of the loss. Pooling of losses from all companies within a conglomerate is not allowed, but as returns are not consolidated by company groups, pooling is difficult to detect. Residency definitions: Individuals who reside in Myanmar for no less than 183 days in any one fiscal year are considered resident for tax purposes. A person that resides in Myanmar for 183 days (or even 360 days) divided equally over two fiscal years is not considered a resident of Myanmar.81 Permanent Establishment: No provision exists in the tax law and thus, in each tax treaty, the permanent establishment clause effectively defines PE for investment from the treaty partner country. Commercial Commercial Introduced as a sales or turnover tax. Commercial tax is levied at various rates Tax Tax Law (1990) for different products, varying from 5 percent for primary products and escalating to a higher rate as sales move up the value chain. More than half of the tax is collected at the border, and only a limited set of services (trading— both wholesale and retail), domestic transportation, hotels, and restaurants constitute the main sources of inland commercial tax receipts82. Excise Duties Various annual Excise Duties are levied ad valorem, with rates ranging from 8-100 percent. MOPF Latest rates for select goods include: 100 percent for cigarettes, 50 percent for notifications alcoholic beverages and 8 percent for natural gas. Sources: MOPF 26. In the next 5 years, major legislative priorities that are being debated include: (i) A new Tax Administration and Procedures Law (TAPL) to standardize administrative processes to support self- assessment; (ii) A new Income Tax Law, that looks to clarify income tax rules, and in particular, includes provisions for related party groups, transfer pricing (arms-length principle) and dealing with incentives through the tax law instead of the investment law. In addition, the law is expected to be updated to align with a self- assessment system, and not an official assessment system; and (iii) Policy for enactment for Value Added Tax (VAT) – the enactment of VAT, however, is likely to be a long term effort subordinate to building the required administrative capacity. 104 SHORT AND MEDIUM-TERM SOURCES OF REVENUE LEAKAGE 27. As Myanmar transitions further towards a modern tax system, there are two short to medium- term sources of revenue leakage that warrant close attention: (i) tax base erosion on domestic and international transactions; and (ii) provision of tax incentives. These two issues are part of a much broader set of tax reform challenges, which are covered elsewhere and not repeated in the PER. The PER instead focuses on these major sources of revenue leakage that arise largely from the largest domestic and foreign taxpayers. This, in turn, provides options to raise revenues in an efficient manner by focusing limited administrative resources. 28. An overarching and related challenge before considering these two topics is the definition of the tax base. The roll-out of self-assessment to large taxpayers has helped to partially uncover the tax base, which was not known under the official assessment system. Economic activity was either not covered by tax provisions or the lack of supporting documentation filed with tax returns was missing (e.g. balance sheet information). The rollout of self-assessment to large taxpayers has generated greater information on taxpayer returns, facilitated tax audits and, in turn, created incentives to report on economic activity for taxation purposes, which, taken together, has helped expand the mapping of the tax base. The focus on large taxpayers for initial rollout of self-assessment is justified, as it: (i) has allowed for the testing of the system on taxpayers with higher administrative capacity and (ii) has focused limited resources on the segment of taxpayers that likely account for a large share of tax revenue (Figure 92).83 Figure 92: The Dual Tax Triangle Number of taxpayers Amount of tax paid 29. The mapping of the tax base is not complete, especially as self-assessment rollout is yet to be rolled out for medium and small taxpayers. Medium and small enterprises are still operating under an officially administered regime, with limited information on economic activity and, thus, on the tax base. The presence of related party groups, which are companies that share ownership or partial ownership structures, across medium and large tax payer segments creates further information gaps and contributes to revenue leakages (see discussion on base erosion below). The rollout of self-assessment at the MTO1 and MTO2 in the next two fiscal years thus constitutes a critical step in the reform process to define the tax base. 80 It covers seven categories of income: (1) salary, including gratuities and pensions; (2) professional (self-employed); (3) business income, including interest income; (4) property (rental income); (5) capital gains from all sources, within and outside Myanmar; (6) undisclosed sources; and (7) other. Tax treatment differs across types of income, residence status of the recipient, and in some cases citizenship. 81 For this reason, international rules usually provide that residency is based upon presence in the country for any 183 days in any consecutive twelve-month period. This is a crucial provision for determining who is taxable (especially foreign nationals doing business on behalf of a multinational) and links with Permanent Establishment definition. 82 Data from IMF 2013 83 The “dual tax triangle” is a phenomenon found in all countries, whereby the largest 10 -20 percent of taxpayers pay 80-90 percent of the tax revenue (and the bottom 10-20 percent of taxpayers pay less than 10 percent of tax revenue). Evidence highlights that this is likely the case in Myanmar as well, thus necessitating an administrative and taxpayer service focus on large taxpayers. 105 Tax base erosion 30. Tax base erosion is defined as legal tax avoidance and is different from illegal tax evasion and/or corruption. Tax avoidance implies that a company is either following the law or acting where there is no law or procedure to stop the activity. Tax evasion is intentionally breaking the law. In Myanmar, tax avoidance on domestic transactions is facilitated by a number of domestic and international factors. Some of the primary domestic factors facilitating tax avoidance, through domestic sources, are discussed below. 31. The first domestic factor facilitating tax avoidance is incomplete information on taxable economic activity. The short run information challenge is driven by the stepwise rollout of self-assessment, with limited information available on income generated by medium and small enterprises.84 With the medium to long term rollout schedule for self-assessment, combined with the time taken to build capacity both in the tax administration to effectively audit returns from small and small/medium sized enterprises, this challenge will remain for years to come. However, it is expected that the rollout of self-assessment at the MTO 1 in FY 2017/18, and the concurrent collection of income tax return information, will add to the knowledge of the tax base. 32. The second domestic factor enabling tax avoidance is incomplete specification of the income tax law. A number of innovations in income tax instruments that have been seen to be effective globally are absent in Myanmar (e.g. firm disclosure of related companies and/or individuals). These innovations allow the tax administration to better understand firm behavior—especially within groups of companies, domestic and foreign, the taxability of firms (permanent establishment/residency data), and the use and abuse of deductions. 33. To put this into effect, the Income Tax Code could be amended to require firms to disclose related party companies or individuals. This change could allow tax administrators to assess how costs are booked, and to judge whether the costs booked match the economic activity in each individual company. This, in turn, helps to ascertain if the company group is profit shifting to lower tax burden, especially if one group member is receiving a tax incentive or is in a special economic zone (see below for discussion on related party transactions). In addition, the tax law could introduce ring-fencing provisions. This could be provisions around deductions, specifically if certain deductions should be ring-fenced around individual investments or whether they can be spread throughout the company group. An important cost to be considered in this regard is intra- company finance (interest generated and the deductibility), as well as the use of loss-carry forward. 34. The third domestic factor enabling tax avoidance is the lack of income tax provisions that allow the tax administration to monitor intra company group activity. The IRD is now beginning to focus on tax base erosion arising out of tax planning from related party (domestic and foreign) transactions. These transactions are relevant to related party domestic and foreign companies, which use differences in the tax system (incentives, zones, etc.) to tax plan and shift profits to the lowest tax regime. A widely used way to shift profits is transfer mispricing, which refers to the manipulation of price of goods and services provided by and procured by related companies within a group, with an aim to shift profits to the company(ies) subject to a lower tax regime 84Outside of commercial and excise tax data, but even here there is no benchmarks to verify whether the levels reported make sense without income tax to corroborate and verify that the reported levels make sense. 106 35. Addressing the tax base erosion from related party transactions requires a legal review, build- up of administrative capacity, and enhanced scope of LTO audits to cover related party companies. Drawing on global experience, addressing challenges of tax base erosion from domestic related party transactions requires: (i) review of laws and procedures to clearly define what tax planning is permissible within the law and the punitive actions for non-compliance; (ii) administrative capacity building to identify risk, monitor related party transactions and audit transactions be able to separate between transfer pricing and mispricing; and (iii) expanding LTO mandate to audit related party companies as a group – currently, only the parent company (say a large Multi-National Enterprises) is subject to audit and self-assessment in the LTO while its domestic subsidiaries are not subject to audit or self-assessment regime, on account of being under the jurisdiction of the MTO. 36. In addition to tax base erosion in domestic transactions, Myanmar is increasingly challenged by tax avoidance on international transactions. There has been increased attention globally on rapidly accelerating multinational operations in LICs and LMICs, who are either unaware or unable to identify taxable events due to multinational activities within their borders. Though no precise measurement of tax base erosion from international sources exist, estimates consistently suggest that revenue losses due to unidentified international tax evasion/avoidance is sizable, which is consistent with estimates, such as that of UNCTAD (2015) of multi-national enterprises accounting for close to 10% of revenue in developing countries. 37. As a result, international tax issues that have received attention globally include: (i) Transfer mispricing, which refers to the manipulation of the price of goods and services provided by and procured by related parties within a Multi-National Enterprise; (ii) Misuse of ‘within group’ financing mechanisms, especially the use of debt finance to claim interest deductions; (iii) Manipulating the pricing of intangibles such as intellectual property; (iv) Use of tax treaties and withholding taxes to minimize taxation, referred to as ‘treaty shopping’; (v) Avoiding a taxable event in a country by creating a shell structure that prevents a country from the right to tax – which is referred to as avoiding the creation of a “permanent establishment” for tax purposes; (vi) Digitization of many transactions, making issues of jurisdiction identification open to interpretation and manipulation. Use of tax incentives 38. The second area that warrants close attention in terms of addressing revenue leakage and tax system inefficiency is the use of tax incentives. Tax incentives could be good for long-term growth if they generate incremental investment (i.e. new investment that would not have taken place in the absence of incentives) that could, in turn, lead to positive externalities in the form of backward linkages to local industries and knowledge spillovers. On the other hand, foregone revenue from incentives could exacerbate fiscal imbalances, or shift the tax burden to other areas, to the detriment of overall investment, growth and poverty reduction. 39. In Myanmar, prior to the new Investment Law (2016), foreign firm applications by the Myanmar Investment Commission received automatic tax relief (first on imports during the development phase, and then on the other taxes for the next 5-7 years). The system was not well targeted, unnecessarily generous, and did not involve the IRD. In addition, under the “officially assessed” tax compliance system, which is gradually being phased out, many taxpayers were able to negotiate exemptions with the tax authorities, which were not reflected in tax submissions. 107 40. As Myanmar reforms tax incentives, it may be important to consider 4 issues.85 The first is the types of tax incentives. Myanmar offers a range of tax incentives, including tax holidays, reduced rates, tax credit, specific incentives for SEZs, and other discretionary incentives. The least effective are tax holidays, which offer blanket exemptions over a period of time – despite the time limits, these are often renewed or companies can close businesses at the end of the period and then reopen as a new investment. The length of the tax holiday period in Myanmar is on the generous side.86 41. The second issue is the targeting of incentives. The Investment Law does away with blanket exemptions and is looking to develop policy for better targeted incentives. Global surveys highlight that tax incentives seem to contribute to FDI attraction in advanced economies, but the response is limited in developing economies87. Further, tax incentives have also been found to be mostly effective to attracting efficiency-seeking FDI, such as, for example, garment producers that are looking to minimizing costs and exporting to developed countries. On the other hand, in the case of resource-seeking FDI (e.g. mining) or market-seeking FDI, tax incentives are less effective, since investors are driven in those cases by existing natural resources, agglomeration economies, or the size and characteristics of the domestic market, where they plan to commercialize their products88. Tax incentives, thus, should be adapted to the type of investments. 42. An overarching principle is that incentives should be targeted to sectors that generate significant economic (social) returns, but limited financial (private returns). Such targeting should help ensure that the benefits from incentives, from increased economic returns, outweigh the costs in regard to revenue foregone. The revenue foregone from tax incentives is equivalent to reduced expenditure on public goods, which is why tax incentives are also referred to as tax expenditures. By focusing on sectors with high social return, the tax expenditures could offset lower Government spending on public goods. 43. The third issue is around the process of administering tax incentives. The eligibility criteria for incentives should be simple and transparent. The policy could be consolidated into a single legal instrument and should have the engagement of the IRD in regard to assessing the costs in regard to revenue foregone. This is discussed further below. 44. The fourth related issue is transparency, especially to assess and publish estimates of revenue foregone from tax incentives. A tax expenditure statement, which provides an estimate of the revenue foregone from tax incentives, and makes this information part of the annual budget is important to allow for debate on costs and benefits of tax incentives. 45. To illustrate potential impact of incentives, the PER uses LTO corporate income tax data to assess possible range of revenue foregone under different scenarios. On account of the incompleteness of the data on revenue foregone (‘cost’) and also on additional FDI generated by incentives (‘benefit’), a complete cost-benefit analysis of existing incentives in Myanmar could not be conducted. The chapter uses the “revenue foregone” approach to measure tax expenditures. It is a simple accounting exercise that amounts to summing all of the additional tax revenue that hypothetically would have been collected if the tax incentive were not in place. In other words, it compares the current/prospective treatment to the benchmark treatment, assuming taxpayer behavior is unchanged. 85 Adapted from “Rethinking the Use of Tax Incentives in East Asia and Pacific – East Asia and Pacific Economic Update,” The World Bank (October 2015). 86 However, comparing tax holiday durations across countries is complicated by the fact that the start of the tax holiday is triggered by different factors: first year of sales in Myanmar; commencement of operations in Lao P.D.R. and Thailand; and realization of taxable income in Vietnam. 87 “Empirical Evidence on the Effects of Tax Incentives”, IMF (2009). 88 “Empirical Asymmetries in Foreign Direct Investment and Taxation”, Grubert and Mutti, Journal of International Economics (2004). 108 Table 24: Tax Incentives in East Asia and Pacific Max tax Tax Reduced Investment R&D Super SEZs Discretionary holiday, Holiday/ Tax rate allowance/ incentives deductions or Incentives years Exemption Tax credit EPZs Cambodia 9 x x x … … x x China 2 x x x x x x x Hong Kong … … … x x … … … Japan … … x x x … … … S. Korea 5 x x x … … x x Indonesia 20 x … x … … x x Lao PDR 10 x x … x … x x Malaysia 10 x x x x x x x Myanmar 5 x x x x … x x Philippines 6 x x … x … x … Singapore Negotiated x x x x x x x Thailand 11 x x x x x x x Vietnam 4 x x … x … x x Note: “x” denotes incentive granted; “…” denotes that the relevant incentive is not applicable in the country. Source: World Bank, East Asia and Pacific Economic Update October 2015 46. In practice, the “revenue foregone” approach uses the difference between two revenue estimates (Figure 93). The first is an aggregation of the actual tax revenues or a simulation along the lines of the actual tax code, which is called the ‘base case’. The second is a simulation of tax revenues under a scenario where the TE is removed from the code, along the lines of a specified ‘benchmark’ tax system. This revenue is called the ‘hypothetical case’. The difference between the two measures is the revenue foregone from the tax expenditure. Figure 93: Estimating Tax Expenditures Source: IMF (2015) 47. Alternative approaches were not used, as the revenue foregone approach requires no behavioral assumptions and is also consistent with expenditure statements published in OECD countries.89 Alternative approaches include the “revenue gain” approach, which measures how much revenue could increase if a particular tax incentive were removed, or the “outlay equivalence” approach, which estimates how much direct expenditure would be required to provide a benefit equivalent to the tax expenditure. Both the alternatives require significant additional data on benefits from tax expenditure (outlay approach) and on behavioral responses to tax rate changes (revenue gain approach), which is not available in Myanmar. Australia publishes some estimates using the “revenue gain” approach, but those are supplementary to “revenue forgone” estima tes. 89 Only Sweden and the United States have tried the “outlay equivalence” approach (IMF Fiscal Monitor, April 2011). 109 48. The analysis is based on data from large taxpayer CIT returns received from the LTO for 2014/2015 and 2015/2016. Ideally, the analysis should be extended to all taxpayers and to other tax types, especially consumption taxes and customs duties – however, data on returns was not available. In the future, with the rollout of self-assessment to other taxpayer segments, this data could be collected and analyzed for tax expenditures using the same methodology and analytical tools. 49. The LTO CIT data provides insights on the most profitable sectors in Myanmar, which may be relevant to targeting incentives. The data from FY 2014/15 and FY 2015/16 highlights that profits are concentrated, especially in the oil and gas industry (extraction, transportation and sale of refined production) which together account for 59 % of total profits in FY 2015/16 and, to a lesser extent, in the telecommunications industry which accounted for 11 % of total profits in FY 2015/16 (Figure 94). Other sectors are significantly smaller in scale when it comes to overall corporate profits in Myanmar. When looking at profitability within sector in Myanmar, some smaller sectors tend to have higher margins (measured as profits as a share of revenue) but the oil and gas industry still has high margins (Figure 95). Figure 94: Net profits by sector (LTO), FY Figure 95: Net profits as share of total revenues, 2015/16 top 10 sectors (LTO), FY 2015/16 Source: WB Staff estimates on LTO data 50. The tax expenditure estimates draw on four scenarios, corresponding to de facto plausible policies in Myanmar (Table 26), and compares it to the base case. The base case is an estimate of CIT that assumes an effective tax rate of close to 25%, which is the statutory rate applicable under Myanmar Income Tax laws. Scenario 1 applies a tax holiday to 54 sectors that are considered as promoted sectors eligible for tax incentives under the Myanmar Investment Law regulations90. Scenario 2 applies a reduced tax rate to 5% across all sectors – this scenario is plausible on account of widespread non-compliance discussed below that reduces the average effective tax rate. Scenario 3 applies a tax holiday to sectors that have low profitability, but potentially high social returns such as transport infrastructure, water supply and electricity generation. Finally, Scenario 4 applies a reduced tax rate of 15% for the oil and gas sector, which may be plausible as the effective tax rate in the sector may be reduced due to exemptions. 90 The list of promoted sectors for 2017 is available at http://www.dica.gov.mm/sites/dica.gov.mm/files/document- files/promotedsector_notification032017eng_1.pdf 110 51. The tax expenditure estimates by scenario (Figure 96) are significant, as a share of the base and as a share of GDP. The estimates range from 0.04 percent of GDP or 2.4 percent of the tax base in Scenario 3 to as high as 1.2 percent of GDP or 80 percent of the tax base in Scenario 2. Some important observations are that (i) the estimate for an across the board non targeted corporate tax reduction is significantly greater than other scenarios (ii) the estimated tax expenditure from a tax holiday for 54 promoted sectors is the same as tax expenditures from reduced taxes for the oil and gas sector, which indicates the level of concentration and profitability of the sector as compared to other sectors in Myanmar; and finally (iii) the estimated tax expenditures from focusing on public goods is the lowest. These observations underscore the gains from targeting incentives to sectors that have high social, but low private returns as well as the losses incurred if tax compliance in large profitable sectors, like oil and gas, is not improved. Table 25: Base Case and Scenarios for Tax Expenditure Analysis Scenario Key Specification Base Case Effective corporate tax rate of close to 25% is applied to all sectors in the large taxpayer list, as per the provisions of the Myanmar Income Tax Law. Scenario 1: Tax holiday Effective corporate tax rate of 0% (tax holiday) is applied to all 54 promoted for promoted sector sectors under the Investment Law regulations. All other sectors in the large taxpayer list have effective tax rate of close to 25%. Scenario 2: Across the Reduced effective corporate tax rate of close to 5% is applied to all sectors in the board corporate tax rate large tax payer database. reduction Scenario 3: Tax holiday Effective corporate tax rate of 0% for sectors that are assumed to produce public for sectors with low goods. All other sectors in the large taxpayer list have effective tax rate of close private, but high social to 25%. (economic) return Scenario 4: Reduced tax Effective corporate tax rate of close to 15% for the oil and gas sector. All other rate for the oil and gas sectors in the large taxpayer list have effective tax rate of close to 25%. sector Figure 96: Estimated tax expenditures under Figure 97: Tax Expenditures in Select Countries various scenarios, FY 2015/16 Sources: WB Staff estimates, LTO Sources: WB, OECD 111 52. When compared to other countries, Myanmar’s estimates may seem on the lower end but this is because other tax types and non-LTO taxpayers, for which tax expenditures are likely to be greater than for LTO CIT. Reported tax expenditures from other countries range from over 8 percent of GDP in Guatemala to less than 1 percent of GDP in Germany (Figure 97). However, these countries have significantly higher tax revenue collections as a share of GDP, thus these expenditures as a share of the tax base are significantly less than in Myanmar. Furthermore, once analysis is completed on tax returns data from other tax payers and tax types, Myanmar’s estimated tax expenditures could increase significantly. Discussions with accounting firms confirmed the difficulty in calculating what tax is foregone, given low compliance rates, ambiguities in the income tax code allowing for legal tax avoidance and the challenges arising from firms allocating profits to smaller businesses for tax avoidance (discussed above). 53. The design of tax incentives in Myanmar is not sufficiently tailored to achieve targeted policy objective of maximizing the benefits from FDI. In this regard, global experience with tax incentives highlights the need to design incentives with a view to support innovation, link local private sector to global value chains of production and focus on higher value-added sectors. For Myanmar, similar to other fast growing regional economies (such as Thailand and Malaysia), this could imply a shift away from blanket tax exemptions for companies to a focus on (i) “super-deductions” for R&D and staff training spending, which are deductions from taxable profits of more than 100 percent of the cost of the investment; and (ii) accelerated depreciation schemes on the acquisition of machinery and equipment incentivizing the upgrading of production facilities and the attraction of higher value added industries. A further advantage of the introduction of these kind of deductions, rather than plain exemptions, is that firms would need to fill in their tax returns, which could result in better taxpayer information and reduced scope for tax evasion. 54. Tax incentives are provided, largely by MIC and DICA, without adequate provisions for cost- benefit analysis. A cost-benefit analysis of tax incentives at the sectoral or industry level is essential in order to determine both the cost in regard to revenue foregone and the perceived benefits arising from increased investment that is stimulated by the incentives. In addition, in the absence of such analysis, it is also not possible to (i) isolate the “marginal investors” for whom incentives tip the balance to invest and (ii) to effectively forecast tax revenues. 55. In addition, the analysis of revenue foregone is not carried out by the Internal Revenue Department prior to granting of the tax incentive. The IRD has the mandate for revenue collection and the requisite capacity, within Government, to carry out a revenue foregone assessment, similar to the exercise highlighted above based on LTO income tax returns. Despite this, such an assessment is not currently carried out, in part due to the absence of information sharing between the incentive granting authority and IRD on the nature of proposed incentives, sector for which the incentive is applicable and other relevant information. POLICY OPTIONS Reforms to improve understanding of the tax base 56. In the short-term, increased efforts at better defining and identifying the tax base could have a major positive impact on revenue collections and tax system efficiency. It would also improve the chances that the audit process will identify underreporting, tax avoidance, and even tax evasion. The tax policy and administration reforms already in motion are likely to improve understanding of the tax base especially as (i) self-assessment is rolled out with more of the taxpaying population, with MTO1 beginning in 2017, and the other taxpayer segment starting from 2018; and (ii) on the tax policy side, revisions in the income tax law, and eventual adoption of a VAT to replace the commercial tax will help ensure that the tax base is fully specified and that the IRD has the legal basis to administer all aspects of the tax base. 112 57. Notwithstanding these significant reforms, there are potential tax policy and tax administration that could further enhance the understanding of the tax base. Starting with tax policy, the authorities have several options to in terms of provisions that could be incorporated in the forthcoming income tax law – the options are not focused on indirect taxes as VAT reform is not likely to be feasible from an administration point of view in the next 3 years: (i) Related party provisions: This is a legal requirement for company groups and individuals to be identified (via the tax reporting forms) and present consolidated tax returns, in addition to individual subsidiary tax returns. This is essential to understand whether costs and income is reported properly and whether profit shifting through costs manipulation is happening. (ii) Arms-length principles introduced into the tax code: In order to have a comparator to base related party transactions, the concept of “arms-length principle” should be added to the income tax law, stipulating that related party transactions (domestic or foreign) should be the basis (comparing related party reported prices to how a similar transaction between un-related parties would be priced). (iii) Clear definition of residency principle and a domestic law definition (wide) of permanent establishments. These provisions, which are companion pieces to related-party provisions are targeted at defining a “taxable event” by a foreign entity. This is particularly important for service providers, distributors, and eventually, digital transactions. (iv) On the cost side, limitation of abuse of deductions such as interest deduction ceilings (thin capitalization, but interest deduction ceilings are considered more effective—BEPS action 4). (v) Rules to limit loss-carry forward (perhaps by investment, then by group), with the elimination of incentives which extend the loss carry forward period. 58. On tax administration, strengthening the system of information gathering could help further identify and expand the tax base, building on successful efforts to reorganize the tax authority by taxpayer segment, and concurrent rollout of self-assessment. The ability to do audit, taxpayer service, and collection all depend on understanding the taxpaying population. This is especially true for the LTO, which in particular needs to understand the business, sector, market conditions (domestically and internationally), and risks. The easiest way build this information comes first from tax returns, but also from outside research (including internet searches, reading corporate annual reports, reading the newspaper, attending chamber of commerce or trade group meetings, etc.) This approach of building business profiles so the tax administration knows profitability parameters, cyclical events (for example when an agricultural industry revolves around the harvest), and specific events that could affect a sector (a shortage of cement, bad weather, etc.). Addressing transfer pricing and international tax issues 59. The OECD is currently leading a coordinated global response (the “BEPS” program91) to create a single set of consensus-based international tax rules to realign taxation with economic activities and value creation, which will then be implemented by individual countries, designed to protect tax bases while offering increased certainty and predictability to taxpayers. 91The “BEPS” program was adopted by the G20 and most international organizations, including the World Bank, in 2015. The program aims at 14 actions, policies, and procedures aimed at curbing profit shifting between related parties mainly resulting in international transactions. For a detailed description, see http://www.oecd.org/tax/beps/beps- actions.htm. 113 60. Implementing the complete actions under the BEPS agenda is not currently feasible in Myanmar, but certain aspects may be considered. The challenges posed by tax evasion, tax avoidance and transfer mis-pricing on international transactions will become increasingly relevant for Myanmar as FDI flows increase in coming years. While implementing the complete BEPS actions is not feasible, Myanmar could consider bringing in international support to implement the following actions: (i) In the short term, moving cautiously on tax treaties. In comparison to other countries in the region, and other developing countries, Myanmar has a relatively small tax treaty network. This is optimal from an international tax policy and administration perspective, as it minimizes the chance of treaty provisions being excessively generous, in conflict with domestic tax law and also the chance of “treaty shopping”. In the absence of treaty with a country, Myanmar levies a 25 percent withholding tax on all foreign companies. (ii) In the longer term, introducing some regulations or procedures on transfer pricing, specifically on audit. Currently, Myanmar does not have any transfer pricing related legislation or procedures and furthermore, no clarity is provided to auditors on transfer pricing audits on appropriate price adjustments to be used to reflect “market prices”. (iii) In the longer term, introducing some related party reporting requirements. This could include requiring firms to present a consolidated tax return that shows the distribution of costs among subsidiaries. This will, in turn, allow the LTO or MTO to determine if costs are correctly being booked against the proposed economic activity. For example, if one of the firms in the group is a distribution company, it is unlikely that it would be requiring high levels of finance and have intellectual property booked to this operation. (iv) In the longer term, introducing “permanent establishment” provisions in income tax law. Such provisions don’t currently exist in law, and, in practice, vary by treaty. With increased foreign direct investment, the lack of such provisions could be used as a tax avoidance mechanism by companies. Options for tax incentives reform 61. Despite rationalization of tax incentives under the new Investment Law, there is merit in doing more systematic assessment of costs and benefits, and an analysis of the appropriate types of incentives. Addressing this challenge requires reform of the legal framework for granting incentives, addressing the mix of incentives currently in place and also improvements in the process and underlying analysis before granting of incentives. 62. In the short term, there is scope to enhance the transparency of incentives available to investors and the amount of tax expenditures resulting from those incentives. This would entail the consolidation and streamlining of information on incentives that is available to investors, and also publishing the specific criteria used to grant tax incentives, and a detailed explanation of the approval and appeal process. It would also be desirable to produce annual tax expenditure estimates. Tax expenditures are a use of public funds, and thus information on them should be made available to the public. The tax expenditure estimates based on LTO data could be annually updated and complemented with further analysis. The consolidated assessment could then be published as a yearly tax expenditure report e.g. as an annex to the Union Budget. 63. There is merit in adjusting the specific mix of tax incentives currently in place, gradually replacing tax holidays with more cost-effective incentives. As discussed above, incentives such as super- deductions and accelerated depreciation could help increase the productivity of the physical and human capital stock and foster the diversification of the economy by encouraging investment in higher value-added sectors. 114 64. On the other hand, tax holidays could be deemed to some extent redundant, in the sense that three of each four investors surveyed in the IMS mention they were not considering alternative locations to Myanmar. Ideally, tax holidays could be targeted to the marginal investor, that whom would not have invested in Myanmar without incentives. However, lack of sufficient information on potential investors (not yet established in Myanmar) makes their targeting unfeasible. 65. If policymakers deem that tax holidays are still necessary, their fiscal impact could be lessened by introducing partial tax holidays. This could take the form of a 50 percent tax exemption rather than a total tax holiday, or gradually decreasing exemption rates over time. Alternatively, the introduction of deferred tax payment, with liabilities that can be carried forward for a few years and ascribed when taxes are paid could be another way to promoting long-term investors. In addition, Myanmar could use sector profitability in the decision to grant incentives. Sectors with high profit margins should not require tax incentives. 66. In terms of the process for granting tax incentives, Myanmar could carry out a cost-benefit analysis at the industry or sector level. Inter-institutional coordination, especially between DICA and IRD, is essential to balance the objectives of investment attraction and revenue generation. Such collaboration is important to assess the costs and benefits of tax incentives, with IRD best placed to assess the potential costs (in terms of revenue foregone) and DICA and other incentive granting authority best placed to assess the benefits. 67. It would be desirable for Myanmar to engage in a regular review of its incentives regime. Policy objectives should also be clearly defined and agreed across public and private actors, including the type of FDI Myanmar would like to attract. The cost-effectiveness of incentives in achieving the targeted policy objectives could then be assessed. This would include an evaluation of the short- and long-term benefits and costs of the incentives provided. Incentive programs should also include indicators to measures their results, as well as monitoring mechanisms to make sure public money is being spent effectively. 68. In the longer term, the legal framework for tax incentives could be consolidated and placed in the Tax Code, rather than the Investment Law or sector legislation. Following international best practices, moving the tax incentives to the Tax Code would allow taxpayers to access all the needed information, and prevent the proliferation of incentives that could result in tax expenditure. Tax authorities are also better placed to monitor companies’ compliance with the incentive regime than other institutions, given the complementarities of these checks with the primary functions of the tax administration. 115 Table 26: Policy options for improving tax system efficiency Issue Policy options The extent and nature of Short term (tax policy): Introduce provisions in the income tax law taxable economic activity (‘tax related to: related party provisions requiring company groups to present base’) is not well understood. consolidated tax returns, introducing arms-length principles, defining residency, and limiting interest deductions and loss-carry forward. Short term (tax administration): Strengthen information gathering such that tax administrations can build business profiles. Tax avoidance, tax evasion and Short term: Move cautiously on signing new global tax treaties. transfer mis-pricing on Long term: Introduce regulations related to transfer pricing, especially on international transactions has a audit. severe impact on tax base erosion in developing Long term: Introduce related party reporting requirements and countries like Myanmar. provisions for permanent establishments in the income tax law. Tax incentives are not well- Short term: Increase transparency by providing consolidated information targeted nor not systematically on incentives to investors and also by publishing information on tax evaluated for costs versus expenditures incurred. benefits. They result in Medium term: Adjust the mix of incentives in place. Consider moving significant erosion of the tax away from tax holidays. base. Medium term: Coordinate between IRD and DICA such that granting of incentives is preceded by a cost-benefit analysis for the sector or industry. Long term: Consider placing tax incentives under the income tax code rather than the Investment Law 116 SOUND PUBLIC DEBT MANAGEMENT 1. Public debt management in Myanmar is undergoing transformation at a time when the country’s exposure to domestic and international financing is expanding rapidly. Since 2011, the Government has been modernizing public debt management functions and building institutional and human resource capacity to handle an increasingly complex public debt portfolio. Starting from a low base, there is significant scope to further modernize debt management functions. The PER Medium-Term Fiscal Framework (chapter 1) proposes the level of financing needed to afford an expansion of public services and infrastructure. This needs to be underpinned by analysis of whether those borrowing levels are affordable for Myanmar over the long-term, how the composition of borrowing affects the costs and risks of the public debt portfolio, and the institutional needs to manage this debt in an effective way. 2. This chapter reviews public debt management reforms and sustainability in Myanmar, and presents options to borrow affordably for essential public services and infrastructure . It includes: (i) a review of the public debt management framework, including recent reforms; (ii) an analysis of current public debt levels and sustainability; (iii) an analysis of options available for medium-term borrowing to help ensure that Myanmar’s public debt portfolio is exposed to reasonable levels of costs and risks; (iv) a review of options to help further develop the domestic debt market; and (v) a summary of policy options for strengthening public debt management. PUBLIC DEBT MANAGEMENT FRAMEWORK 3. There has been much progress in the past four years in gradually modernizing the institutional framework for public debt management in Myanmar, even though implementation capacity remains a constraint. A Debt Management Reform Plan developed in 2013 provided the basis for priority actions across five pillars of reform: (i) institutional arrangements and capacity building; (ii) legal framework; (iii) debt strategy and reporting; (iv) domestic debt market development; and (v) cash and broader public financial management. The PER reviewed progress across these pillars, and suggests priorities going forward. Pillar 1: Institutional Arrangements and Capacity Building 4. A big achievement in recent years has been the consolidation of public debt management functions in MOPF, moving away from a system where public sector entities could borrow with little central control or oversight. Prior to the establishment of the MOPF Treasury Department in September 2014, public debt management responsibilities were very fragmented. The erstwhile Ministry of Finance’s External Debt Division was only responsible for back office operations, including recording and processing of all external loan contracts, disbursements, and debt service payments. Although access to external loans was limited prior to 2011, line ministries and SEEs borrowed directly from external creditors, which ultimately became the Ministry of Finance’s liabilities. 117 5. The institutional structure in MOPF now follows a more traditional Treasury setup.92 A Debt Management Division (DMD) has been established to integrate external and domestic debt management, linking closely with cash management and other Treasury functions. The DMD has: (i) a Back Office responsible for handling settlement of transactions and maintenance of financial records; (ii) a Middle Office tasked with risk analysis and monitors, reporting portfolio-related risks, and assessing the performance of debt managers against any strategic targets/benchmarks; and (iii) a Front Office responsible for executing transactions in financial markets, including the management of auctions and other forms of borrowing, and all other funding operations. Figure 98. Organizational Structure of the Treasury Department Cash Management Divison Front Office Debt Management Deputy Director General Middle Office Division Director General Information Technology Back office Division Administration and Accounts Division Accounting and Reporting Deputy Director General Division Treasury Policy and Quality Promotion Division Source: MOPF 6. Despite progress in establishing the institutional structure, the DMD faces important implementation capacity constraints. The Division has 30 staff, which is small relative to its increasing and more complex responsibilities. Officials end up working across all functions of DMD, which limits the ability of staff to specialize within their Back, Middle or Front Office responsibilities. Officials in the Front Office for example, get drawn into commenting on debt management policies or manage IT-related issues. Officials in the Middle Office get drawn into assisting colleagues in the Front and Back Office, which detracts from their analytical and reporting responsibilities. With 12 new staff joining the DMD in May 2017, reallocation of staff to boost the Middle Office will be possible. But a big part of the challenge is that a small number of DMD staff are carrying a rapidly increasing workload, whilst in parallel having to implement major, technically complex reforms 7. A major reform underway is the effort to modernize the DMD’s Management Information Systems (MIS) for public debt data. Until recently, the DMD was working from manual ledgers with only five computers. In September 2016, the DMD started to transfer its data to an electronic database, the Commonwealth Secretariat Debt Recording and Management System (CS-DRMS). CS-DRMS has been installed and all debt data has been input. Further training is being provided covering debt data validation, analysis and generation of reports. It is expected that the implementation of the system will be completed by end of 2017, but further training over the 4th quarter of 2017 will be needed to ensure that the system is fully maintained and MOPF is able to gain full benefit from using the system. 92“Roles and Responsibilities of divisions within the Treasury Department” downloaded from http://www.mof.gov.mm/my/?q=node/196. 118 8. The transition to the CS-DRMS could significantly improve debt data management and quality, though it involves a major change management process. CS-DRMS should become the official public debt database with secure, accurate, up-to-date, and regular data validation. All debt reporting and analysis should originate from the CS-DRMS. But as experience from other Lower Middle Income Countries (LMICs) shows, MIS upgrades, especially of this magnitude switching from manual ledgers to a computerized system, take a long time to become fully established and relied on. In Myanmar, the shift to CS-DRMS from a manual system entails a cultural change, which will require significant handholding and training before DMD is able to benefit fully from the system’s functionality. Pillar 2: Legal and regulatory framework 9. The adoption of the Public Debt Management Law (PDML) by Parliament in January 2016 was a major milestone in public debt management reforms. The PDML for the first time assigns borrowing responsibilities to the Minister of Planning and Finance with the Parliament’s approval. It prohibits any borrowing or loan discussions without permission from the Planning and Finance Minister, effectively ceasing further independent debt contracting by line ministries and SEEs. The Planning and Finance Minister also has authority to issue loan guarantees and other forms of guarantees provided the guarantees conform to terms and conditions set by the Government and the Parliament. 10. Regulations for the PDML, currently under development, could usefully centralize debt servicing projections and obligations, which still rest with line ministries. Line ministries, SEEs, development committees and States/Regions are required to submit their estimated loan disbursements and debt servicing costs during the budget preparation process. The Union Budget Law then appropriates funds to each entity for debt service payments to creditors. Whilst it makes sense for line ministries to provide loan disbursement projections, MOPF should be centrally responsible for oversight of debt service obligations and transactions with creditors. 11. The lack of complete and timely information on debt service obligations, and the lack of centralized control over transactions with creditors poses substantial operational risks. The DMD needs to have an accurate and comprehensive picture of payment obligations at all times, which should be possible through the CS-DRMS. This is an essential element of sound public debt management to ensure that debt service obligations are met in a timely manner, budget estimates are complete and reliable, and financial accounts are comprehensive and transparent. It also provides creditors with a single point of contact, and can positively impact creditors’ perceptions of risk and their pricing of debt. 12. The PDML regulations could also require that the Medium-Term Debt Strategy be submitted to Parliament at the same time as the Union Budget Law. This would enable Parliament to consider the level of financing required for the Union Budget, the composition of that financing, and the resulting implications for the cost and risks of the public debt portfolio. The PDML states that the debt strategy needs to be published before the beginning of the fiscal year. The strategy for FY2016-17 was approved in August 2016―four months after the beginning of the fiscal year and also after the publication of the Union Budget Law. For FY2017-18, Parliament had approved the Union Budget Law before the debt strategy was debated in the Parliament. 119 Pillar 3: Debt strategy and reporting 13. Myanmar prepared its first debt strategy in 2016. This was approved by the Parliament and made publicly available as per best practice in August 2016. The strategy outlined the overall debt management objectives and discussed the existing debt portfolio, though it did not include an analysis of costs and risks. To be aligned with other well-designed debt management strategies, the next strategy could include a plan that operationalizes the Government’s overall debt management objectives with more forward-looking descriptions of future fiscal and debt flow projections, costs and risks of financing options and assumptions on financial instruments. It may also be desirable for the debt strategy to be made publically available in English, to inform potential international investors and creditors. 14. MOPF plans to publish a semi-annual public debt bulletin on the levels and breakdown of domestic and external debt, which would be an important contribution to fiscal transparency. In the meantime, the first Annual Debt Report was submitted to the Minister of Planning and Finance in January 2017 and considered by the Parliament in May 2017. The Middle Office leads on debt reporting issues, and can take example from other countries in the region. In Cambodia, the General Department of International Cooperation and Debt Management of the Ministry of Economy and Finance have built up the size and capacity of their Middle Office over the past five years. They now have a capable team undertaking comprehensive debt portfolio analysis, publishing a debt statistical bulletin, and currently finalizing the first medium term debt management strategy. Indonesia is another country that has developed a well-functioning Middle Office that has prepared a comprehensive debt management strategy and is currently developing a credit model for guarantees and an asset-liability management framework. Pillar 4: Domestic debt market development 15. There has been good progress in initiating a primary market for domestic public debt with the launch of Government securities’ auctions. Treasury bill and Treasury bond auctions are now in place. MOPF is responsible for decisions on the instruments to be issued, including the coupon rates, the offering amount in each auction and the auction calendar for both bills and bonds. The first move from the administered interest rate regime to market-based borrowing started with the introduction of 3-month Treasury bills in January 2015 using single price auctions as the mechanism for issuance, which was changed to multiple price auctions from August 2015. The Treasury bill auctions saw two expansions in 2016 with 6-month bills in May 2016 and 12-months bills in July 2016 and all three maturities are now offered six times each quarter. Most recently, the Government launched multiple price Treasury bond auctions in September 2016. 16. CBM no longer issues bills and bonds on behalf of the Government and now acts as fiscal and paying agent of MOPF and undertakes the auctions, maintains the registry, and processes debt related transactions. The Central Bank of Myanmar Law, approved in July 2013, defines CBM’s role as a fiscal agent of the Government and has provisions for CBM to provide loans to the Government with Parliamentary approval as well as issue its own debt. When the Treasury Department was established in September 2014, it took over responsibility for management of domestic debt from CBM’s Financial Markets Department. Under a fiscal agency agreement reached in January 2015, CBM undertakes the auctions, maintains the registry, and processes debt related transactions as fiscal and paying agent for MOPF. 120 17. Continued improvements in fiscal transparency and allowing interest rates to rise in Government securities’ auctions could potentially improve participation in the domestic debt market. As noted in chapter 1, market participation remains below auction targets. Communication and public awareness over the fiscal policy stance and priorities could help anchor expectations and build investor confidence. In addition to some of the measures noted above and in chapter 1, MOPF has initiated surveys of investors to better understand market sentiments, and accordingly adjust policy. In this regard, as discussed in chapter 1, allowing interest rates to better reflect market conditions is essential for increasing market participation, but also supporting monetary policy objectives of maintaining price stability. Pillar 5: Cash and public finance management 18. The Government has made good progress in developing cash forecasts and a consolidated view of Government cash balances. Effective cash management ensures that the Government is able to fund its expenditures in a timely manner and meet its obligations as they fall due. In doing so, reliable cash forecasts are essential. With the introduction of weekly settlement between CBM and the Myanma Economic Bank (MEB) in November 2014, the cash position of the Union Fund Account is now available within three weeks of the close of the month – a significant improvement from a lag of more than two months. Cash Management Division (CMD) has started working with line ministries since April 2015 to obtain their cash plans and spending reports. The actuals help determine whether the cash flow forecasts from a working cash flow forecasting model are consistent. As cash flow forecasting is still a relatively new function in Myanmar, it may be some time before MOPF can work with accurate and timely cash flow forecasts to manage cash balances effectively. 19. While issuance and buyback of Treasury bills are currently not used to manage aggregate cash balances, it will be important to do so once a target balance has been developed. Currently, any deficit in the Union Fund Account is covered by CBM buying Treasury bills, however, as of FY2016-17 onwards their interest rate is set at market auction rate instead of the previous fixed 4 percent. At the time of settlement, auctioned Treasury bills help offset the amount financed by CBM. In addition, CBM financing is limited to a maximum of thirty percent of the financial requirement for the current fiscal year of 2017-2018. PUBLIC DEBT LEVELS AND SUSTAINABILITY RISKS 20. Unlike most other LMICs and due to its historical circumstances, Myanmar’s pub lic debt portfolio is dominated by short-term domestic debt. As of end March 2017, total Government debt amounted to US$21.4 billion, or 33.2 percent of GDP (Table 1). External debt accounted for 13.4 percent of GDP and domestic debt 19.8 percent. Approximately three-fourths of domestic debt (14 percent of GDP) is debt with CBM. Most of these are 3-month interest-bearing Treasury bills that the CBM purchased from the Government. Once these short-term bills reach maturity, most are resold and de facto treated as long term Treasury bills with interest paid at four percent every quarter.93 93 Myanmar Medium Term Debt Strategy based on the 2016-17 Budget Law. 121 Table 27: Total Public Outstanding Debt by Creditor Type 94 Creditor Type Mar-14 Mar-15 Mar-16 Mar-17 External 16.3 13.4 16.0 13.4 Multilateral 2.5 1.9 2.3 2.1 Bilateral 13.8 11.5 13.7 11.3 Paris Club 6.2 4.7 5.7 5.1 Non-Paris Club 7.7 6.8 8.0 6.2 Commercial 0.0 0.0 0.0 0.0 Domestic 19.9 17.8 21.6 19.8 Total 36.2 31.2 37.6 33.2 Source: MOPF, WB staff estimates. 21. Myanmar has reengaged with several official creditors following clearance of longstanding debt arrears in 2012/13 (Box 14), which has led to a pick-up in external loan commitments. Around half of public debt outstanding is owed to Non-Paris Club bilateral creditors, and around a quarter to multilaterals (Figure 2). In terms of currency composition, around a third of Myanmar’s external debt is denominated in Euros, and a quarter in Japanese Yen. The recent depreciation of the Euro against the US Dollar contributed to a drop in debt outstanding as a share of GDP. Box 14: Myanmar arrears clearance Myanmar settled arrears with external creditors in 2012 and 2013, opening the door to international re-engagement. Prior to international reengagement, Myanmar was not able to service its external debt obligations and accumulated large amounts of principal and interest arrears. By end-2012, total arrears with multilateral and Paris Club creditors were estimated at US$11 billion (72 percent of total external debt outstanding). In April 2012, Myanmar agreed with Japan on a debt restructuring plan to resolve its arrears, and with the World Bank and the Asian Development Bank in January 2013. Myanmar also reached an agreement with Paris Club Creditors in January 2013 to a concessional debt treatment that wrote off 50 percent of all arrears and rescheduled the remainder over 15 years with a 7-year grace period. Subsequent to these arrears clearance operations, the outstanding public external debt was lowered to sustainable levels, and the country has since maintained its debt servicing obligations. Table 28: Myanmar Arrears Clearance (Millions US$) December 31, 2012 March 31, 2013 Total arrears Outstanding Debt Total Debt Total Debt External 10,984 4,341 15,345 9,176 Multilateral 933 717 1,670 1,511 Bilateral 10,051 3,524 13,575 7,665 Paris Club 10,051 364 10,415 3,806 Non-Paris Club - 3,160 3,160 3,859 Commercial - 100 100 - Source: MOPF, WB Staff estimates. 94 The outstanding debt numbers differ slightly from the records maintained by the Debt Management Division on account of differing GDP and exchange rate assumptions. The differences are not material in magnitude. 122 Figure 99: External Public Debt Outstanding Figure 100: Currency Composition of External Public Debt Outstanding, as end -March 2017 Source: Government of Myanmar, WB staff estimates. 22. To analyze the sustainability of Myanmar’s public debt, a customized Debt Sustainability Analysis tool was developed as part of the PER. The Myanmar DSA (MDSA) is premised on a similar analytical framework to the World Bank-IMF Low Income Country Debt Sustainability Framework (LIC DSF)95 with a focus on public debt (external and domestic). The evolution of public debt is captured by firstly identifying the obligations associated with the aggregate stock of debt from the previous period. Projections of the primary fiscal balance and other factors are drawn from the PER Medium-Term Fiscal Framework (see chapter 1). 23. The MDSA is a simplified tool designed to enable the authorities to carry out their own debt sustainability analyses. There are differences compared to the World Bank-IMF LIC DSF; for example the MDSA does not take into account residual factors that lead to changing public debt levels such as asset disposition. This reduces data intensity and focuses on key aspects of the evolution of public sector debt, which is the largest component of the country’s total debt and the largest source of risk. The MDSA requires collaboration with the Planning Department and other economic agencies such as the Central Bank of Myanmar on macroeconomic assumptions, and with the Budget Department on MTFF assumptions. DSA would traditionally be undertaken by a Fiscal Policy Analysis Unit, but in the absence of this is led by the DMD. 24. In contrast to many LICs, Myanmar’s current risk of debt distress is low under the WB -IMF DSA and the MDSA. Under the baseline of both assessments, public and publicly guaranteed (PPG) debt burden indicators are projected to remain below the benchmark for the projection period. The present value of PPG debt is expected to rise from 31.2 percent of GDP in FY2016-17 to 34.0 percent of GDP in FY2030- 31, towards the indicative benchmark of 38 percent of GDP. More of a concern, present value of PPG debt service moves in a sharp upward trajectory from 16.0 percent of revenues in FY2016-17 to 23.2 percent of revenues in FY2030-31. The trajectory of both indicators suggest increased risk of debt distress over the long run. Deficit levels and interest-growth differential seem to dominate debt dynamics. While strong growth was able to partly counter balances the contribution of fiscal deficits in the past, its ability to do so diminishes over the long run as rising deficits outpace economic growth. 95 Please see: http://www.worldbank.org/en/topic/debt/brief/dsf. 123 Figure 101. Present Value of Debt-to-GDP Figure 102. Public Debt Service -to-Revenue Source: WB Staff Estimates based on the MTFF and data from authorities 25. External vulnerabilities such as exchange rate depreciation pose limited risks to Myanmar’s debt sustainability, due to the highly concessional nature of much of the external obligations. Growing current account deficits and tightening US monetary policy could put further downward pressure on the value of the Kyat relative to the US dollar. However, a scenario whereby the domestic currency depreciates by 26.5 percent against the US dollar between FY2016-17 and FY2019-20, leading to a pickup in domestic inflation and cost of borrowing is expected to have a limited impact on debt sustainability. Under this scenario, the present value of PPG external debt is expected to moderate slightly from 34.6 percent of GDP in FY2016-17 to 33.6 percent of GDP in FY2030-31 due to exchange rate movements. However, the present value of PPG debt service moves in a sharp upward trajectory from 16.6 percent of revenues in FY2016-17 to 23.0 percent of revenues in FY2030-31. This is partly explained by the low stock of existing external debt and the highly concessional nature of much of the external debt. The adverse effects of a one-off depreciation could also partly be offset by its support of the external competitiveness of non-resource exports. 26. Myanmar’s debt sustainability is more sensitive to a lower growth scenario. As discussed in chapter 1, the baseline projections reflect an expected pick-up in investments but there are several downside risks, which could lead to lower real growth in the long run. A one percent reduction in real growth leads to a significant breach of the indicative public debt and debt service indicators in the long term. Under the lower growth scenario, present value of PPG debt is expected to rise from 31.5 percent of GDP in FY2016-17 to 48.7 percent of GDP in FY2030-31, breaching the indicative benchmark of 38 percent of GDP from FY2026- 27 onwards. Present value of PPG debt service doubles from 16.0 percent of revenues in FY2016-17 to 32.0 percent of revenues in FY2030-31. This scenario would suggest a significantly heightened risk of debt distress for Myanmar. 27. When combining low growth with countercyclical Government spending, debt sustainability indicators deteriorate even more significantly. As noted in chapter 1, structural weaknesses (e.g. narrow production base, proneness to natural disasters, commodity dependence) heighten risks to strong and stable growth. These risks can be exacerbated by weak policy responses, a number of which have been highlighted in other parts of the PER (i.e. inefficient capital spending, SEE operations, and tax system), which can prolong recovery from economic shocks. In light of these risks, the MDSA tested a scenario whereby in addition to a 1 percent lower GDP, there is also a one off increase in expenditure equivalent to 5 percent of GDP. This scenario leads to the most severe deterioration of debt sustainability indicators, with present value of PPG debt expected to rise from 31.5 percent of GDP in FY2016-17 to 50.2 percent of GDP in FY2030-31, breaching the indicative benchmark of 38 percent of GDP from FY2023-24 onwards. Present value of PPG debt service more than doubles from 16.0 percent of revenues in FY2016-17 to 33.4 percent of revenues in FY2030-31. 124 28. Fiscal slippages pose a significant additional risk to Myanmar’s risk of debt distress. The baseline projects a moderately expansionary fiscal deficit. However, given Myanmar’s dependence on natural resource revenues, sustained low commodity prices, lower growth or a fall in revenue buoyancy (where revenue growth does not keep pace with economic growth) could also result in higher primary deficits. Similarly, should expenditure restraint loosen or SEEs performance weaken further, this could also result in higher primary deficits or the realization of contingent liabilities. Recent economic developments led to a widening of the fiscal deficit from 1.1 percent of GDP in FY2014-15 to an estimated 3.2 percent of GDP in FY2015-16. A scenario whereby fiscal slippages equivalent to 2 percent of GDP over the forecasting horizon leads to a large deterioration of debt sustainability indicators. The present value of PPG debt is expected to rise from 33.2 percent of GDP in FY2016-17 to 46.6 percent of GDP in FY2030-31, breaching the indicative benchmark of 38 percent of GDP from FY2023-24 onwards. Present value of PPG debt service more than doubles from 16.0 percent of revenues in FY2016-17 to 32.0 percent of revenues in FY2030-31. 29. In sum, the MDSA finds that Myanmar can afford to finance the MTFF if it pursues the complementary efforts to enhance fiscal space outlined in the PER but that there are important risks. Some of these risks are structural and exogenous, such as natural disasters. As highlighted above, the combined effect of such shocks on growth and public expenditure, could quickly lead to significant deterioration in public debt sustainability. A weak policy environment can exacerbate these risks. These include inefficient capital spending that has little impact on growth, SEE inefficiencies that crowd out private investments, or a narrow revenue base leaves little buffer for economic shocks. Box 15: Global trends in public debt sustainability Myanmar’s reengagement with international creditors draws parallel to increased integration of LICs to global markets, and in turn greater exposure to market risks. Myanmar has seen a significant reduction in debt stock over the past several years. Like other LICs, this is partly explained by the debt restructuring efforts of 2012 and 2013, as well as higher growth and commodity prices. Similarly, Myanmar is embarking on a path of greater integration to global markets and domestic financial market deepening, increasingly exposing the country to global market factors. The share of low income countries (LICs) at high risk of debt distress has almost halved between 2007 and 2014, but liquidity buffers have been reduced. Since the Global Financial Crisis (2008), public debt dynamics have moved favorably for most LICs due to general economic growth, higher commodity prices and debt relief in some HIPC countries. In aggregate terms, this translated to a reduction of debt-to-GDP ratio from 66 percent in 2006 to around 48 percent at the end of 2014. However, the general improvement in solvency risk indicators across LICs is set against worsening liquidity indicators for all groups except commodity exporters. While risk ratings may not have deteriorated, countries’ liquidity buffers have generally decreased, especially for frontier LICs. 125 Figure 103: Gross Debt for LICs (in % of Figure 104: Total Public Debt (in % of GDP) GDP) Source: IMF Fiscal Monitor (October 2016) Source: LIC DSF database Figure 105: Change in DSA solvency Figure 106: Change in DSA liquidity indicators indicators Source: IMF Fiscal Monitor (October 2016) The deterioration in liquidity buffers is partly explained by expanding financing sources enabled by good macroeconomic performance. In terms of external sources, the share of non-concessional to total external debt broadly doubled between 2007 and 2014. LICs have also increasingly tapped the global capital markets to issue Eurobonds, driven by greater economic development and supported by ample global liquidity. While there is no significant evidence of linkage between these issuances with short term fiscal policy loosening, access to international capital markets may be providing LICs with more scope to conduct discretionary fiscal policy. Borrowing from domestic markets for frontier markets rose from 14 to 19 percent of GDP between 2007 and 2014, but remained stable at 13 percent for the average LIC. The rise in domestic borrowing for many countries appear to be associated with financial deepening and economic development, but tend to carry a higher interest rate and when in excess, can crowd out private credit and investment. The closer integration of LICs’ to the global market also carries greater exposure to market factors includin g US interest rates and global volatility. 126 Figure 107. Domestic Debt and Real GDP Per Figure 108. External Commercial Borrowing: Capita: Frontier LICs, 2007 - 2014 Frontier LICs (in US$ million) Source: WEO for countries with available data. Sources: Dealogic, Bloomberg. In light of these trends and recent developments in the global economic environment, LICs can be especially vulnerable to shocks to growth, exports and exchange rates. The weaker growth outlook in key advanced and emerging market economies are likely to impact LICs through lower demand for traded goods, with commodity producers particularly affected.96 Contingent liability shocks, which often materialize in low growth scenarios can also increase indebtedness and spike in debt service. At the same time, the prospect of tighter global liquidity conditions could lead to higher global interest rates. This may have implications on the ability of LICs to access the international capital markets to rollover existing bonds and raise new financing at affordable rates. This may also lead to higher cost of domestic borrowing for countries that rely on foreign participation in their domestic debt markets. In addition, exchange rate pressures in LICs, for example those resulting from the recent appreciation of the US dollar, could have significant implications for both the level of debt and the repayment capacity of debtor countries. Safeguarding countries’ debt sustainability in the context of greater financial integration and a changing global environment will likely require improved institutional capacity, including in the area of debt management. Strong domestic policies – maintaining prudent macroeconomic policies with an eye to downside risks, and being ready to quickly and forcefully implement policy corrections where warranted by adverse shifts in the external environment, including to terms of trade, export markets, and financing terms – will also be critical for managing public debt vulnerabilities in this context 96For many low-income commodity exporters, lower commodity prices have already necessitated fiscal adjustment. Additional adjustment may be warranted to stem debt pressures as low commodity prices are likely to be sustained, and exchange rate depreciation are likely to exacerbate debt service burden. Adjustments will have repercussions for growth especially if capital spending bears the brunt of the adjustment burden. 127 MEDIUM-TERM BORROWING STRATEGIES 30. After considering the sustainability of overall financing needs, the PER now turns to the composition of that financing. The key job of the DMD is to decide what borrowing instruments to use, and in what combination, to meet the financing needs decided in the MTFF. These decisions directly impact evolving costs and risks of the debt portfolio. Given the prominence of long tenor instruments in the portfolio, the structure of existing debt implies significant inertia going forward in terms of costs and risks. 31. Borrowing needs are determined by the primary fiscal deficit and debt servicing requirements. Financing required to meet any excess of primary expenditures (i.e., non-interest payment expenditures) over revenues plus grants is determined by fiscal policymakers, as part of the annual budget process. Borrowing needs in a particular fiscal year include financing for the primary deficit as well as any additional financing needed to service debt obligations falling due in the year. Incorporating projections for debt service as well as the primary deficit over a multi-year horizon is thus important for fiscal policy makers to plan properly. 32. The PER analyzes below the characteristics of the existing debt portfolio and the effects of alternative borrowing strategies going forward. The aim is to help debt managers decide the best course forward for the objective of meeting financing needs at a reasonable level of risk. Characteristics of the existing portfolio are as of end FY2016-17, and include coverage of standard cost and risk indicators. A large share of the existing portfolio (55.5 percent) reflects debt obligations to CBM.97 Several future borrowing strategies are considered to explore the impact on the debt portfolio costs and risks. This is intended to be illustrative, with scope for refinement by the DMD in subsequent analytical work. Costs and risks of existing debt portfolio 33. Interest payments on existing public debt total 1.4 percent of GDP, and the weighted average interest rate is 4.1 percent (Table 30). The majority of interest payments are made on domestic obligations, largely reflecting the continuous rollover of legacy CBM debt (0.6 percent of GDP). The relatively low weighted average interest rate is driven by the characteristics of external debt (2.5 percent average weighted interest rate) and the concessional 4.0 percent interest rate on legacy CBM debt. The average for non-legacy CBM domestic debt is 7.2 percent. As Government moves toward more market-based financing for domestic borrowing, the weighted average interest rate is expected to gradually rise. This of course would significantly increase if legacy CBM debt were converted to market-based Treasury bills over time. 97CBM outstanding debt stood at MMK 12,460 billion (14.7 percent of GDP) at end-March 2017. The baseline analysis distinguishes legacy CBM financing (95.3 percent of the total) from FY2016/17 CBM financing. The former has a concessionary interest rate of 4 percent with associated interest payments assumed in perpetuity. The latter is treated on similarly, with interest rates consistent with the auctioned 3-month Treasury bills. This assumption is made for analytical purposes, and thus the baseline analysis reflects the current situation that Treasury is not making principal payments on CBM financing (legacy or more recent). 128 Table 29: Public debt cost and risk indicators, as of end -March 2017 Risk Indicators External debt Domestic debt Total debt Cost of debt Interest payments (% of GDP) 0.4 1.0 1.4 Weighted Av. IR (%) 2.5 5.3 4.1 Refinancing risk ATM (years) 11.4 1.6 8.8 Debt maturing in 1yr (% of total) 4.4 37.4 13.0 Debt maturing in 1yr (% of GDP) 0.6 1.9 2.6 Interest rate risk ATR (years) 11.4 1.6 8.8 Debt refixing in 1yr (% of total) 4.4 37.4 13.0 Fixed rate debt (% of total) 100.0 100.0 100.0 FX risk FX debt (% of total debt) 42.2 ST FX debt (% of reserves) 8.9 Sources: MOPF, WB Staff estimates. 34. Given all debt is contracted at fixed interest rates, refinancing and interest rate risk indicators are equal and show refinancing/rollover risk is low for external debt but significant for domestic debt. The average time to maturity (ATM) for external debt is 11.4 years, while for domestic debt this figure is 1.6 years. External debt is made up of loans with very long maturities, some with close to 40 years of amortization remaining, while auctioned domestic debt on the other hand is comprised of Treasury bills and bonds of much shorter maturity. Over one-third of domestic debt matures in one year. As a result, the redemption profile (Figure 110) shows a peak of principal payments in FY2018-19 driven by the portfolio of outstanding Treasury bonds. The redemption profile for external debt is initially high as non-concessional external loans contracted before arrears clearance start to mature, but declines afterwards. Importantly, the redemption profile does not reflect the large amount of outstanding CBM obligations, given there is currently no repayment. Figure 109. Myanmar: Public debt redemption profile as of end -March 2017, MMK million Source: GoM, WB staff estimates. Note: The profile reflects current practice of no CBM repayments. 129 35. Exchange rate depreciation can have substantial valuation effects on the public debt stock due to the high share of external debt. As external debt is mostly concessional, i.e., very low interest rates and significant grace periods, the recent depreciation of the Kyat did not significantly increase the interest cost in the budget. However, the debt stock, as a percentage of GDP, did increase substantially. A breakdown of the SDR into the component currencies shows that the main exposure of exchange risk to foreign currencies was the Euro (36.8 percent of total debt), followed by the Japanese yen (25.1 percent) and US dollar (21.6 percent). Future borrowing strategies and impact on debt portfolio costs and risks 36. As financing options evolve for Myanmar, more rigorous analysis is needed to explore the debt cost and risk implications. The country’s potential sources of finance have changed dramatically over the last five years from both external and domestic sources. This changing and increasingly diverse menu implies that debt managers need to consider more than one cost indicator, calculated on a regular basis and tracked over time. It also implies that market risks translate into different impacts for different debt compositions, and risk exposure indicators (e.g., share of external debt in the portfolio, share of domestic debt maturing in a year, ratios of Treasury bills versus bonds, etc.) can provide useful targets to drive desired adjustments to the debt portfolio to match risk tolerance. 37. Myanmar’s current Medium-Term Debt Strategy provides general guidance for borrowing decisions over a three-year period. The MTDS focuses on three main goals: maintaining debt at sustainable levels; monitoring the composition of the public debt portfolio so as to identify financial risks and what is needed to ensure that these risks are kept to prudent levels; and continuing to develop the domestic market for Treasury securities. While the strategy provides some guidance for borrowing activities and management of the risk exposure of the debt, it is expressed as broad guidelines. Overall, external funding is encouraged at concessional terms from multilateral and bilateral lending agencies, and is set by existing and planned infrastructure projects and disbursement schedules. Domestic debt market development is a primary objective of the debt management strategy. There is no coverage of potential Public Private Partnerships. Box 16 : Myanmar’s First Medium Term Debt Strategy (2016) The PDML stipulates that MOPF is responsible for submitting a rolling three-year Medium Term Debt Strategy (MTDS) annually to the Parliament for approval after it has been reviewed by Cabinet. While there was a gap of six months between the Parliamentary approval of the Union Budget Law 2016/17 in January 2016 and the first MTDS in August 2016, future debt strategies are expected to be prepared together with the Union Budget (and MTFF) and submitted to the Parliament at the same time. This is important to help strengthen links between Treasury and Budget Departments as discussions take place during MTFF and budget preparation on various fiscal scenarios and financing options affecting debt sustainability and servicing costs. Submitting the MTDS and the Annual Budget Law together to the Parliament could help ensure consistency between strategy, planning, and budgeting. The first MTDS in 2016 outlined plans to reduce reliance on monetization of the budget deficit by CBM through auctions of Treasury bills and bonds. In addition to presenting the debt portfolio composition as of end-2015 through breakdowns by creditors and currencies, the MTDS identifies that 78 percent of total domestic debt (42 percent of total debt) is Treasury bills sold to CBM at 4 percent interest rate and that nearly all of them have not been redeemed though maturity dates have passed. In order to carry out the strategies outlined, the 2016 Myanmar MTDS concludes with a plan to develop an Annual Borrowing Plan. The Plan will take into account outstanding debt, debt sustainability levels, risks and development needs financing. 130 38. A medium term analysis was conducted to explore the impact of various borrowing scenarios on the cost/risk tradeoffs of the debt portfolio. The analysis combines alternative borrowing strategies under baseline macroeconomic assumptions taken from the MTFF (Table 4) with shocks to market risk variables to assess impacts on the debt portfolio’s cost and risk. The timeframe covers three years. Importantly, the analysis incorporates the current policy to phase out reliance on CBM purchase of Treasury bills for deficit financing. What this indicates is that the proportion of Treasury bonds will increase each year to effectively replace issuance of CBM Treasury bills, with the proportion of auctioned Treasury bills issued remaining constant at around 30 percent. Under the baseline scenario, the issuance of CBM Treasury bills will be phased out by FY2020-21, indicative from the maximum of 30 percent in FY2017-18 to zero by FY2020-21. Table 30. Macro-fiscal framework FY16/17 FY17/18 FY18/19 FY19/20 Revenue (% of GDP) 18.7 18.5 18.1 18.5 Primary expenditure (% of GDP) 21.8 20.7 19.8 20.0 Primary balance (% of GDP) 2.9 2.2 1.7 1.5 Nominal GDP growth (%) 16.4 15.6 15.3 14.8 Source: MTFF 39. Baseline financial market conditions are expected to remain stable. After a steady lowering of the domestic yield curve in the last few years, reducing the 25-year yield from 10 percent to 6 percent, and the T-bill rate close to zero, the MTDS projects a stabilization of the yield curve. International market rates of the USD and JPY are expected to gradually increase, which will affect the rates on ADB and IBRD debt and global bond pricing. 40. Four alternative borrowing strategies were considered. The analysis focuses on the impact of the choice between domestic and external financing, and on the choice between short- and long-term domestic debt from a market development and cost perspective. The specific borrowing strategies considered are: • Strategy 1 ‘40/60 short’: 40/60 split between external/domestic borrowing to cover gross financing needs; domestic borrowing moves gradually away from reliance on T-bills (both auction and CBM) and gradually increasing 2-3 year bonds (bonds share of domestic borrowing increases 40%→50%→60% in the 3-year horizon); the distribution within T-bonds is 40% 2-yr; 40% 3-yr; 20% 5-yr for all years; external borrowing is balanced equally across multilateral concessional and bilateral semi-concessional instruments • Strategy 2 ‘40/60 long’: similar external/domestic split to S1; domestic assumes market development toward longer tenor, split between 3-yr, 5-yr and 10-yr issuance (40%, 40%, 20% in FY2017-18 transitioning to 20%, 40%, 40% by FY2019-20); external borrowing remains balanced across instruments • Strategy 3 ‘60/40 max concessional’: 60/40 split between external and domestic; external 60% concessional; domestic similar to baseline in S1 • Strategy 4 ‘60/40 Eurobond’: 60/40 split between external and domestic; external balanced between concessional, semi-concessional and Eurobond; domestic similar to baseline in S 131 41. The costs associated with each strategy were evaluated over a 3-year horizon (FY2017/18 through FY2019/20), using two cost measures: the ratio of interest payments to budget revenues; and the ratio of total Government debt to GDP.98 Risk is defined as the difference between the cost measures under the baseline and under a shock scenario. For the purposes of the analysis, the focus was on the maximum risk observed across all shock scenarios in the last year of the projection period. 42. The cost-risk analysis clearly demonstrates the importance of not relying on one single cost measure, as the ranking of the strategies differed for different indicators. Furthermore, the analysis demonstrated that the choice of strategy––even when applying detailed scenario analysis––is far from trivial. Not surprisingly, S3 outperforms all other strategies given the high share of external concessional funding, i.e., funding at very low interest rates and at very long maturities, resulting in a significantly lower present value of public debt to GDP at the end of the 3-year horizon. The strategies with higher domestic borrowing (S1 and S2) have higher debt servicing due to relatively shorter maturity, however are an effective way of meeting the objective of supporting domestic debt market development. S1 and S2 also minimize forex risk, as shown by the external debt service indicator. S4 in particular has more exposure to foreign exchange risk. Under the Eurobond scenario the external debt service is higher than with concessional external terms and thus susceptible to exchange rate shocks. Figure 110. Risk Indicators Source: WB staff estimates. 98The ratio of total Government debt to GDP is used as a short-hand measure intended to capture capital gains/losses in the nominal stock of debt arising from variations in the exchange rate. 132 Directions forward 43. Implementation of the MTDS would provide a good basis for balancing increasing borrowing needs and maintaining debt sustainability. The MTDS helps to operationalize a country’s debt management objectives by outlining cost-risk trade-offs and debt service profiles associated with alternative borrowing strategies for meeting the Government’s financing needs and payment obligations. For example, it will analyze the cost and risk of moving towards more concessional financing (with a significant grant element) which could help finance development needs, but limit the rise in debt burden in present value terms, which would however increase the vulnerability of the debt portfolio to exchange rate movements. A MTDS could also help address the vulnerabilities uncovered in the DSA, such as those related to increased interest rates. 44. The DMD can take important initial steps toward improving the analytical basis of its medium term debt strategy. As demonstrated above, applying the MTDS Toolkit can help by further assessing the performance of other strategies not examined by the PER and fine tune the analysis. This will add substantial value in the process of developing a formal medium-term debt management strategy that should be updated annually as part of the budget process. 45. The progress in developing the domestic debt market will be a key determinant in the evolution of the debt portfolio over the medium term. This would phase out the reliance on short term borrowing for deficit financing, and the eventual ability to extend the maturity of domestic issuances toward 5+ years. These issues are discussed further in the next section. DOMESTIC DEBT MARKET DEVELOPMENT 46. Myanmar is at the very early stages of a long-term journey in developing its domestic debt market. In the near-term, Government securities’ auctions should help to gradually reduce dependence on short-term Central Bank financing. In the long-term, reliance on external financing should also diminish. Treasury bonds with longer maturities (5, 7, 10-year) could be introduced subject to market conditions and demand. A stronger domestic debt market could help reduce exposure to interest rate risks from external debt, and enable the Government to raise financing urgently. In particular, short-term liquidity issues can be managed initially through the issuance and buy back of Treasury bills, and eventually through repurchase agreements (repo) when the repo market is developed. 47. The development of the domestic debt market will go hand in hand with the gradual strengthening of wider policy and institutional capacity. As noted in chapter 1, market participation at the moment is still below potential, which is not unusual at this early stage of development. Other Lower Middle Income countries with more mature public debt management systems have been able to develop deeper and more reliable domestic debt markets over a 10 to 15-year period. This required broader public finance and debt management reforms that supported macroeconomic stability and growth. In the near or medium-term therefore financing for public investment will not be driven by domestic debt. 48. There are nevertheless important near term steps that can be taken to ensure confidence in and expansion of the Government securities’ market. This includes, as discussed in chapter 1, allowing the interest rate in Treasury bill and bond auctions to rise. Currently, there is significant unmet demand (unsuccessful bids), which suggests that bill and bond purchases could be capped at a target effective interest rate. Or it could also imply that liquidity needs are met through existing purchases. Either way, the lack of transparency, coupled with negative real interest rates, could hamper the development of the debt market. 133 49. International practice shows that auctions should rarely be cancelled because of market conditions. Award amounts should also not be below the pre-announced auction amount in order to achieve short-term debt service obligations. While MOPF has not cancelled an auction, it has cut off the amounts awarded below the offered amount when the yields bid have been deemed to be too high. Such practices can affect credibility and damage the integrity of the auction process, causing the risk premium to rise, hampering market development, and causing long-run debt service cost to increase. As this has been a practice of MOPF, it is recommended that a policy change is needed to avoid continuing this practice. The section below suggests other near-term priorities, that could potentially help expand market participation. 50. The Government has made a commitment to develop the market through the use of auctions, which have so far been reasonably successful. While there have been auctions which have not been fully bid, evidence suggests that this is diminishing as more participants are attracted to the market. In particular, the insurance companies and securities companies on behalf of retail investors. It is also expected that once pension funds are established, this will further increase the demand for Treasury securities. Syndication has been option that was considered for issuing Treasury securities, but this was rejected due to the commissions that would be paid to the syndicate of banks selected and the lack of understanding within the Government to undertake syndication at this point. Enhanced transparency 51. More transparent dealings with the market not only reduce the long-term cost of borrowing by promoting a more efficient and low-risk market, and are often an end in itself. Transparency can enhance good governance through greater accountability of MOPF. This includes publicly disclosing debt management objectives, pre-announcing borrowing or financing programs, as well as maintaining professional relationships with market participants. MOPF can follow other countries by establishing and maintaining a website specifically to cover debt management, and particularly for the domestic market. Some countries have a publication calendar which commits the debt managers to publish important information throughout the year. 52. This includes high-quality information about the debt structure, funding needs, and debt management strategies and operations. This is best achieved through the publication of a medium-term debt management strategy (MTDS) and regular (e.g., quarterly and/or monthly) publications on domestic market activities. Accepting the market price also demonstrates commitment to market development; even though this may increase debt-servicing costs in the short-term, there are offsetting factors such as reduced dependence on inflationary financing and improved capacity to manage monetary pressures. 53. Increased transparency can have important implications for the operation of the primary market. Regardless of the mechanism used to raise funds, international experience suggests that borrowing costs are typically minimized and the market functions most efficiently when MOPF’s operations are transparent–for example, by publishing borrowing plans well in advance and acting consistently when issuing new Treasury bonds–and when MOPF creates a level playing field for investors. The terms and conditions of new Treasury bond issues should be publicly disclosed and clearly understood by investors. 54. Communication of market relevant information by Governments is now increasingly through an Investor Relations Website. MOPF could consider establishing a website dedicated to supporting the information needs of institutional investors active in the Myanmar bond market. The website will further improve transparency and access to information describing MOPF’s debt activities and pertinent fiscal, macroeconomic and other statistics. Information that is highly market relevant will include: composition of Government debt (domestic v. external, currency, structure by maturity and interest rate); non-resident holdings of Government Treasury bonds (when applicable); breakdown of public, publicly guaranteed, and SEE debt; risk profile of Government debt; Government contingent liabilities and other fiscal risks; budget forecasts and out-turns; and institutional (legal and regulatory) information relating to the Government, central Government and regulatory authorities (i.e., relevant laws and regulations applicable to Treasury bonds). 134 Box 17: Progress in establishing an efficient primary debt market in Myanmar • Establishing clear objectives for Treasury bond issuance and debt management: this has been achieved by MOPF through the enactment of the Public Debt Management Law and publication of the annual MTDS. • Developing basic projections of MOPF’s cash/liquidity needs: this is waiting on the development of accurate projections of revenue and expenditures along with cash flow forecasts under the responsibility of the Cash Management Division in the Treasury Department. • Creating safe and efficient channels for the distribution of Treasury bonds targeted to investor needs: this has been achieved through the introduction of CBM Net in 2016. • Progressively extending the maturity of Treasury bonds (when market conditions are suitable): this has been achieved as MOPF has extended the maturities of Treasury bonds in the auctions from May 2018 to May 2020, and extension to a 5-year maturity is planned for FY2017-18. • Consolidating the number of Treasury bonds issued and creating standardized Treasury bonds with conventional maturities to provide market “benchmark” bonds: this was achieved on 31 December 2015 with the consolidation of Treasury bonds into 22 bonds across 11 maturities. • Moving to a predictable and transparent Treasury bond issuance program, with pre-announced issuance calendars and disclosure of funding needs and issuance outcomes: this has been partially achieved with MOPF publishing an auction calendar at the beginning of each quarter and with auction results published on the CBM website on the same day, but funding needs are not yet fully disclosed in the MTDS. • Establishing a website specifically for announcements and dissemination of auction results and other information on Treasury bonds: This is to be achieved when DMD implements a public website that has been designed and is ready to launch. Treasury Department has decided to create a website treasury.gov.mm where DMD will have a subdomain dmd.treasury.gov.mm that can be updated directly by DMD staff to keep the information current and timely. Strengthening investor relations and diversifying the investor base 55. Enhanced transparency could be complemented with efforts to proactively engage with investors. The MOPF now meets with financial institutions and investors at least twice a year in order to understand investor preferences and to develop relationships with “end” investors. In this way, MOPF can factor in investor preferences in making decisions on new debt issuance. MOPF can also use these relationships to promote new debt instruments and initiatives. As DMD staffing increases, there could dedicated officials to carry out the functions of an investor relations program. This is unlikely to be a full-time position in the initial stages of the development of the Treasury bond program, but in the longer term an investor relations team may evolve when the Treasury bond issuance program becomes more active. 135 56. The Institute of International Finance (IIF) has published reports on investor relations and data dissemination, which provide useful guidance in these matters for Myanmar.99 The report, which looks at 20 criteria to assess country practices, offers the investor community a comprehensive comparative evaluation of the efforts emerging market issuers have made to strengthen their dialogue and communication practices as well as a guide of their adherence to best practices in data dissemination relevant to sovereign risk analysis. Other debt management offices have used the IIF criteria to formulate their investor relations and data dissemination policies. 57. These efforts could help broaden the investor base for Government securities. While the local commercial banks have been the main participants in the auction of Treasury bonds, recent auctions have seen the Myanmar Securities and Exchange Centre (MSEC) submit bids on behalf of other securities companies, insurance companies, and individuals. The Securities and Exchange Commission of Myanmar (SECM) requires a securities regulation to be prepared and introduced to enable other securities companies to participate in the auctions as underwriters and to trade Government securities in the secondary market. 58. While foreign banks have been granted a commercial banking license to operate in Myanmar and therefore eligible to participate in the auctions of Treasury securities, there has been very little participation to-date. Foreign banks are often not willing to acquire Treasury bonds if there is no secondary market as they require the ability to quickly liquidate their holdings or to enter into repurchase agreements (repos) should they need to increase their cash position. As there is no secondary market for Treasury securities in Myanmar, foreign banks are unlikely to be a major purchaser of Treasury securities in the short-term. This supports the case to establish a series of benchmark bonds across the yield curve to facilitate the development of a secondary market. POLICY OPTIONS Pillar 1: Institutional Arrangements and Capacity Building 59. A priority for strengthening the institutional capacity for DMD is ensuring the roll out and implementation of the CS-DRMS. This could have a significant impact not only on the quality and timeliness of debt data, but also in terms of relieving staff pressures from maintaining and accessing data from manual ledgers, which invariably create inaccuracies. This could also enable the consolidation of debt servicing functions within DMD, and provide complete and immediate information on debt service obligations. The implementation of the CS-DRMS is supported by a large capacity building effort by the Asian Development Bank and the Commonwealth Secretariat. Experience of MIS reforms in other countries have highlighted the importance of moving gradually in terms of rolling out functionalities and providing constant support to dedicated counterparts responsible for managing the MIS. 60. In parallel, it is important to scale up capacity building efforts for Middle Office staff. Despite ongoing efforts including the addition of more Middle Office staff, this tends to get crowded out by day to day operational tasks. Yet building analytical capacity in debt analysis and reporting takes time and is critical to informing policy. Building analytical capacity is also more complex than training officials in more routine tasks. This may therefore warrant a different approach, including more formal class room based instruction for staff that are dedicated to only Middle Office functions. This could be applied practically by requiring regular analytical outputs. Without a dedicated effort, there is a big risk that this gets relegated. “Sovereign Investor Relations: 2014 Evaluation on Investor Relations and Dissemination Practices by Key Emerging 99 Market Borrowing Countries” https://www.iif.com/system/files/investor_relations_report_2014_0.pdf 136 Pillar 2: Legal and regulatory framework 61. It is proposed that regulations are developed that would enable centralization of debt servicing functions. Whilst this may have to be sequenced with the implementation of the CS-DRMS, gradually moving towards this model is essential to improve operational efficiency, fiscal oversight, and creditor relations. When DMD communicates directly with creditors and is more in charge of status of payments, rather than the line ministries themselves, payment obligations will be at minimum cost and risk as set out in the PDML. 62. It is also proposed to develop regulations requiring concurrent submission of the Union Budget Law and the annual MTDS at the same time to Parliament. This would help better understand how the Government proposes to finance the Union Budget deficit, including sources, implications on costs and risks to the debt portfolio, and broader macroeconomic implications. Otherwise it is not possible to understand how financing choices enable Government manage the trade-offs between more borrowing and more investment. Pillar 3: Debt strategy and reporting 63. The Government has made good progress in developing a debt strategy, which offers a strong basis for further deepening the analytical content. This could be supported through application of the WB- IMF MTDS Toolkit. In particular, it would enable MOPF to carry out basic cost and risk analysis of feasible borrowing strategies, similar to the indicative exercise done for the PER. This can also provide the basis for closer integration of debt strategy development with medium term fiscal planning. The MTDS could also include a plan that helps operationalize the Government’s overall debt management objectives and be more forward looking with descriptions of future fiscal and debt flow projections, costs and risks of financing options and assumptions on financial instruments. These would help move the strategy closer to international good practice. The World Bank stands ready to provide further support in this area. 64. The PER has highlighted the importance of transparency in building market confidence and promoting market participation. Disclosing the funding needs and annual borrowing plan in the MTDS, and publishing this in both English and Myanmar could make an important contribution to this effort. This could be complemented with preparation and publication of a semi-annual public debt bulletin on the levels and breakdown of domestic and external debt and more regular publications on the market and Government’s annual borrowing plan and issuance program when CS-DRMS is fully operational. The DMD could use the IIF Investor Relations and Data Transparency Guidelines to inform transparency reforms. Pillar 4: Domestic debt market development 65. Government policy to limit CBM financing is an important element of balancing fiscal prudence with longer-term investment needs, as well as broader macroeconomic stability. A quantitative limit on CBM financing is a sensible approach in the short-term, as it can support CBM efforts with its Reserve Money target. This could help anchor inflation expectations. Over the medium-term, a prudent policy goal would be to only resort to monetization for specific emergency purposes (e.g. natural disasters, severe cash shortages). This could further strengthen the credibility of Government financing reforms and commitment to more independent monetary policy. 137 66. Ongoing efforts to expand the domestic debt market, and by implication lower reliance on monetization, could be supported by allowing interest rates to rise. This is likely to improve market participation, support monetary independence, and enable counter-cyclical fiscal policies. Whilst this implies higher borrowing costs, this could be offset by greater revenue contributions to the Union Budget and lower net costs from reduced inflationary financing. The domestic debt market could be an important driver of fiscal and monetary policy transparency and credibility. In the short-term, they are unlikely to crowd out financing for the private sector given the small size of the Government securities market and other more binding constraints on private sector access to credit. 67. Efforts to proactively engage investors could usefully complement ongoing reforms to improve transparency. MOPF is already engaging with investors including through regular meetings with the banks and investors on at least a six-monthly basis. As DMD staffing improves, it is also recommended to have a dedicated staff assigned only to investor relations. 68. To develop the secondary market for Government securities, MOPF could prepare and introduce regulations through the Securities and Exchange Commission of Myanmar (SECM) to enable securities companies to trade Government securities in the secondary market. The Securities and Exchange Law 2013 makes provision for securities companies to undertake Securities Dealing by buying and selling Government Treasury bills and bonds for its own account and Securities Brokerage acting as an intermediary to carry out Government Treasury bills and bonds buying or selling orders on behalf of any institution or individual in consideration of a commission, fee or other remuneration. The Law stipulates that securities companies can establish an Over-the-Counter Market (OTC) for trading of Government Treasury bills and bonds. This would formalize the process to allow securities companies to trade Government securities in the secondary market. This will address a major shortcoming whereby securities companies do not have the authority from SECM to provide securities trading and brokerage services. Pillar 5: Cash and public finance management 69. Government should continue the policy to gradually phase out Central Bank financing. There are two key issues to address with the transition from CBM financing. First, the conversion of the existing stock of legacy CBM 4 percent Treasury bills to a combination of equity and market based instruments. Second, the process that MOPF needs to follow in order to ensure that the market fully meets the financing requirement without the need for CBM financing except in the case of an emergency, which should be limited to only short- term of less than three months. Reduced/eliminated reliance on monetization will provide the Government with additional leverage to mitigate against external shocks such as inflationary pressures arising from exchange rate shock pass-through, as the CBM will have more space to adjust liquidity conditions. 70. Government should consider converting some of the large stock of legacy CBM Treasury bills to equity, with the remainder converted to interest-bearing negotiable instruments of Government securities. A debt-equity swap, for example, can be considered. This would involve the Union Government to enter into a debt-equity swap whereby the face value of the Treasury bills would be swapped to the equivalent value in equity, thereby increasing the Union Government’s equity position with CBM. This has the advantage that it reduces the stock of debt and the associated debt service interest cost, and increases the strength of the balance sheet of CBM with an increase in the Government’s equity holdings. The balance of the legacy debt would be progressively converted to market-based Treasury bills and bonds, which would have the benefit of providing the instruments to assist with the development of the secondary market for Treasury bonds and for the development of the repo market. This will ensure compliance with Clause 93 of the CBM Law, which requires conversion to interest-bearing negotiable instruments of Government securities, and would have the benefit of providing the instruments to assist with the development of the secondary market for Treasury bonds and for the development of the repo market. 138 Table 31: Policy options for strengthening public debt management Issue Policy options Institutional arrangements and capacity building Current debt data management based on Short-term: Prioritize transition to CS -DRMS manual ledgers and Excel entries creates with adequate capacity building to ensure staff pressure, prevents access to complete proper change management. Focus on priority and timely information. functions, without overloading stretched capacity. Capacity building in analytical functions of Short to Medium-term: Expand formal class Middle Office crowded out by more room based instruction for staff that are routine debt management tasks. dedicated to only Middle Office functions. Legal and regulatory framework Decentralized debt servicing enhances Short-term: Develop regulations to assign MOPF sole operational risks. responsibility for debt servicing function. Medium-term: Implement regulation in line with roll out of CS-DRMS. Debt strategy and reporting Separate submission of Debt Strategy and Union Short-term: Develop regulations requiring Budget Law impedes analysis of financing concurrent submission of Debt Strategy and options. Union Budget Law to Parliament. Current MTDS provides incomplete picture of Short to medium-term: Use IMF-WB MTDS borrowing strategy and cost-risk implications Toolkit to enhance analytical content (cost - risk analysis, forward projections). Lack of transparency affects market confidence Short to medium-term: Disclose MTDS; and participation in Government securities’ prepare and publish semi-annual debt reports. auctions Domestic debt market development Insufficient supply of market-based domestic Short-term: allow interest rates to rise to credit, and need for greater market participation ensure at least positive real rates. Short-term: increase market outreach and engagement with potential investors Medium-term: prepare/introduce regulations to enable securities companies to participate in the auctions as underwriters Cash and public finance management CBM financing of budget deficit is highly Short-term: quantitative target over CBM inflationary and not aligned with investment financing. needs Medium-term: adopt regulations with policies to limit monetization for specific emergency purposes (e.g. natural disasters, severe cash shortages) 139 The World Bank Myanmar No 57, Pyay Road 61/2 Mile, Hlaing Township, Yangon www.worldbank.org/myanmar www.facebook.com/WorldBankMyanmar myanmar@worldbank.org