INTERNATIONAL DEVELOPMENT ASSOCIATION INTERNATIONAL MONETARY FUND JOINT BANK-FUND NOTE ON PUBLIC SECTOR DEBT DEFINITIONS AND REPORTING IN LOW-INCOME DEVELOPING COUNTRIES January 31, 2020 Prepared by the staffs of the International Monetary Fund and the World Bank Group Approved by Marcello Estevão (World Bank) Kristina Kostial and R.J. Rosales (IMF) Prepared by a World Bank team led by Diego Rivetti (EMFMD) under the guidance of Doerte Doemeland (EMFMD), with inputs from Evis Rucaj and Rubena Sukaj (DEC); and by an IMF team led by Dalia Hakura (SPR) and Andrew Kitili (STA) under the guidance of Mark Flanagan (SPR), with inputs from Charlotte Lundgren and Keiichi Nakatani (SPR) and Noriaki Kinoshita (STA). EXECUTIVE SUMMARY Increasing public debt vulnerabilities in low-income developing countries (LIDCs) have heightened the need for fuller and more transparent accounting of public sector debt (PSD). Public debt transparency is a key prerequisite for effective risk assessment in support of sustainable borrowing and lending practices. However, due to several factors the accounting of PSD in many LIDCs is suboptimal, insufficiently transparent, and may lead to “debt surprises.” In addition, changes in the financing landscape of LIDCs require more clarity about public sector coverage and the statistical treatment of complex and increasingly diverse debt instruments. The framework for reporting on public sector debt is sound. The international statistical standard as described in the Public Sector Debt Statistics: Guide for Compilers and Users (PSDS Guide) provides clear definitions and statistical treatment of all debt-related arrangements including of complex debt instruments. Debt data requirements for the Low- Income Countries Debt Sustainability Framework (LIC-DSF) are closely aligned with the international statistical standard and have been adjusted to better capture debt sustainability and vulnerabilities. Notwithstanding the inadequacies in public sector reporting by the authorities, policy engagements around DSAs and tools and stress tests embedded in the LIC-DSF help to compensate for potential gaps in debt reporting. But there is room for LIDCs to further improve their compilation, reporting, and dissemination of public sector debt data in international databases and more broadly the public domain. Debt data reporting to the Quarterly PSDS database is limited to less than one third of LIDCs (17 countries) and there are considerable differences across countries in national debt definitions. Weaknesses in compilation and reporting of data stem from capacity constraints, weak legal and institutional frameworks, and unclear definition of public debt under national laws. Debt reporting can also suffer from inconsistencies. Several policy priorities arise. Capacity development and institutional reforms in the area of debt data recording and reporting are critical for producing better debt reports. In addition, countries should be encouraged to make more comprehensive and timely debt data available publicly in their national summary data page and through IFIs. In this context, the focus should be on the Quarterly PSDS database and encouraging reporting of additional granular information (e.g., collateralization features of loans and domestic debt) through the Debtor Reporting System (DRS) loan-by-loan debt reporting to the World Bank. To that end, LIDCs would need to overcome impediments for statistical reporting, and the World Bank and the IMF would need to provide required support for capacity development (e.g., the newly-introduced IMF online course on PSDS and scaled-up technical assistance on debt reporting). These steps would complement the ongoing efforts to expand debt coverage in DSAs. 2 TABLE OF CONTENTS Executive Summary .................................................................................................................... 2 I. Introduction ............................................................................................................................. 4 II. Public Sector Debt Statistics .................................................................................................. 5 International Statistical Definition ..................................................................................... 5 Public Debt Data in LIC-DSF ............................................................................................ 9 III. Reporting and Dissemination of Public Sector Debt Data of LIDCs ................................. 12 Statistical Reporting Systems........................................................................................... 12 Data Dissemination Systems ............................................................................................ 14 National Reporting Systems ............................................................................................. 15 Reporting to International Databases and for LIC Debt Sustainability Analysis ............ 15 IV. Statistical treatment of complex debt-creating arrangements ............................................. 17 V. Factors Limiting Reporting of Debt Data by LIDCs ........................................................... 19 VI. Priorities to improve public debt data availability .............................................................. 21 Figures: Figure 1. Public Sector and Its Sub-sectors ................................................................................ 7 Figure 2. Coverage by Instrument and Sector ............................................................................ 8 Figure 3. Improvement in Debt Coverage in LIC-DSAs Since July 2018 ............................... 16 Figure 4. Instrument Identification in Legal Frameworks in LIDCs ........................................ 20 Tables: Table 1. Differences in the Treatment of Public Debt Data Between the Public Sector Debt Statistics (PSDS) and LIC-DSF ................................................................................................ 10 Annexes: Annex Table 1. List of Low-Income Developing Countries .................................................... 23 Annex Table 2. Public Sector Debt Data in the International Databases ................................. 24 Annex Table 3. Reporting Status of Public Sector Debt by LIDCs1 ........................................ 26 3 I. INTRODUCTION 1. Public debt vulnerabilities in low-income developing countries (LIDCs) have increased in recent years reflecting higher public debt and debt service levels and changes in the composition of public debt. Twenty-five out of fifty-seven LIDCs (i.e. forty-four percent) are assessed at high risk of debt distress or already in debt distress as at end-2018,1 almost double the level in 2013. Efforts to prevent and solve debt crises have often been complicated because of inadequate reporting, sometimes with the discovery of “hidden” debts2, and/or the increased use of more complex debt instruments. 2. Public debt transparency is a key prerequisite for effective risk assessment in support of sustainable borrowing and lending practices. It permits borrower countries to accurately track the evolution of their debt situation, and to monitor and manage debt-related risks and vulnerabilities. It is also needed for lenders to accurately assess a borrower’s debt position, borrowing capacity, and creditworthiness before extending fresh credit. Public debt transparency can help lenders and borrowers from not entering into agreements that could cause them financial difficulties in the future. It is also critical for citizen accountability and reducing corruption. For these reasons, debt transparency has been included as one of the key pillars of the joint Bank-Fund multi-pronged approach to reducing emerging debt vulnerabilities. The new IMF/World Bank Low-Income Countries Debt Sustainability Framework (LIC-DSF) operationalized since July 2018 also emphasizes fuller debt coverage and disclosure. 3. This note clarifies definitions and coverage of public sector debt (PSD) in the international statistical standard and the LIC-DSF, and discusses data reporting systems, gaps in debt data collection, and reporting by LIDCs. It is one of the deliverables under the debt transparency pillar of the joint Bank-Fund Multipronged Approach to address debt vulnerabilities. In each section, three main debt reporting dimensions will be assessed: (i) coverage by sector; (ii) coverage by instrument; and (iii) valuation methods. The note is structured as follows: • Section II provides an overview of the international statistical standards for PSD compilation and dissemination and discusses the recommended coverage of the public sector and types of debt instruments. It also discusses the recommended debt coverage in the new LIC-DSF and how it differs from the standard statistical treatment. • Section III takes stock of the actual PSD reporting by LIDCs, including progress in the context of LIC DSAs. 1 This excludes Nigeria and Vietnam that use the debt sustainability analysis (DSA) methodology for market access countries (MAC-DSA). 2 For recent case studies, please see G20 Note: Strengthening Public Debt Transparency. 4 • Section IV provides a refresher on the treatment of complex debt instruments that have become a recent focus of attention. • Section V discusses the factors that are limiting reporting of debt data by LIDCs. • Finally, drawing on the paper’s findings, section VI suggests practical steps to enhance transparency and coverage of PSD data by LIDCs, including in DSAs, and to streamline data reporting. II. PUBLIC SECTOR DEBT STATISTICS International Statistical Definition3 4. To facilitate cross-country comparability and comprehensive debt analyses, public sector debt statistics (PSDS) should be compiled and reported based on internationally accepted statistical definitions and concepts. These standards have been established to foster convergence in debt reporting across countries and are intended for compilers and users of public sector debt statistics. According to such standards, total debt consists of all liabilities that are debt instruments. Debt may be incurred to fund assets that will generate income to meet liabilities.4 The focus of this note is on gross debt only5. 5. A debt instrument is a financial claim that requires payment of interest and/or principal by the debtor to the creditor at a future date, or dates. Debt liabilities are typically established through the provision of economic value by the creditor to the debtor in exchange for a flow of future payments (principal and/or interest). These liabilities are normally under a contractual arrangement but can also be created by the force of law (such as liabilities arising from taxes, penalties, and lawsuits) and by events that require future transfer payments, such as claims on nonlife insurance companies. 6. Debt liabilities should be recorded when goods or assets change ownership, services are rendered, or when funds are made available. Commitments to provide funds in the future do not establish debt liabilities; amounts yet to be disbursed under a loan commitment should not be treated as debt. The definition of debt does not necessarily require that the timing of future payments of principal and/or interest is accurately known. For example, obligations of 3 This subsection draws on the Public Sector Debt Statistics: Guide for Compilers and Users, 2013 (PSDS Guide) which provides the international statistical standard for compiling and reporting PSDS. The PSDS Guide is fully harmonized with the standards set out in the System of National Accounts of 2008 (2008SNA) and the IMF’s Government Finance Statistics Manual of 2014 (GFSM 2014). 4Net debt is calculated as gross debt minus financial assets corresponding to the same debt instruments. 5Data on net debt is generally incomplete, especially in lower income countries, due to the complexity of measuring assets. Moreover, it is not possible to present the DSAs on a net debt instead of a gross debt basis since this implicitly imposes the very strong assumption that government assets and liabilities can perfectly offset each other, which may not always be the case due to liquidity or currency mismatches. 5 employment-related pension funds to their participants are considered debt because payments are due at some point, even though the exact timing and amount of the payment is unknown. 7. Contingent liabilities are excluded from debt liabilities because they are obligations that only arise if a particular event occurs in the future. These include explicit contingent liabilities such as the granting of a guarantee by the government to a state-owned enterprise and implicit contingent liabilities (e.g. future obligations of a social security system, government financial interventions to ensure the solvency of the banking sector during financial crisis, and debt of public sector units without government guarantee which would need to be assumed by the government in case of default). 6 In the case of guaranteed debt, the original underlying liability should be attributed to the original debtor, e.g. the state-owned enterprise—not the guarantor—unless and until the guarantee is called. However, international statistical standards prescribe that explicit and implicit contingent liabilities are reported as memorandum items by the guarantor.7 Public Sector Coverage 8. PSDS should cover the entire public sector as defined by international statistical standards. The public sector consists of general government (budgetary central government, state government, local government, extrabudgetary units, and social security funds) public nonfinancial corporations, and public financial corporations including the central bank (Figure 1).8 6 For more details about contingent liabilities see Chapter 4 of the PSDS Guide. 7In the PSDS Guide, presentation tables for contingent liabilities are provided for Summary of Gross Debt (Table 5.1) which includes publicly guaranteed debt as a memorandum item, and Summary of Explicit Contingent Liabilities and Net Obligations for Future Social Security Benefits (Table 5.12). Standards for measuring contingent liabilities are still evolving because these liabilities are complex arrangements and no single measurement approach can fit all situations. The current international statistical standard is to value them at nominal values (e.g., how much debt would be incurred if the guarantee is called). 8Corporations are defined as entities that are capable of generating a profit or other financial gain for their owners and are set up for the purposes of engaging in market production . 6 Figure 1. Public Sector and Its Sub-sectors 9. Public corporations are defined broadly as entities that are controlled—directly or indirectly—by the government. Control is defined as the ability to determine general corporate policy of the corporation. The 2008SNA and GFSM 2014 list eight indicators that should assist in determining whether a corporation is controlled by a government: (1) ownership of the majority of the voting interest; (2) control of the board or other governing body; (3) control of the appointment and removal of key personnel; (4) control of key committees of the entity; (5) ownership of golden shares and options; (6) capacity to change regulation; (7) control by a dominant public sector customer or group of public sector customers; and (8) control attached to borrowing from the government. 9 The standards recognize that in some circumstances a single indicator may not be sufficient to establish control and classify an entity as a public corporation. Debt Instruments 10. PSDS should cover all debt instruments to the extent data are available. For complete coverage of debt instruments, the following six instruments should be included: special drawing rights (SDRs); 10 currency and deposits; debt securities; loans; insurance, pension, and standardized guarantee schemes (IPSGS); and other accounts payable which in some countries are referred to as pending bills or short-term arrears. 11. To enhance cross-country comparability, debt instruments can be classified into four groups, reflecting data availability (Figure 2). This classification starts with a narrow but commonly applied coverage of only two instruments (D1 on the horizontal axis) which can be expanded to cover all six debt instruments (D4 on the horizontal axis). On the vertical axis, 9See GFSM2014, Box 2.2. Government Control of Corporations. 10SDRs here refer to SDR allocations, not WB/IMF loans denominated in SDR which should be classified as loans. 7 the coverage expands from budgetary central government (GL1) to the entire public sector (GL5). The standards recognize that a wider coverage is desirable but do not set a minimum standard by instrument and sector. Figure 2. Coverage by Instrument and Sector Valuation Methods 12. The method used for valuing debt should be stated explicitly because it affects debt stock indicators. The PSDS Guide recommends that debt instruments should be valued on the reference date at nominal value, which can be defined as the amount that the debtor owes the creditor at any given point in time, including accrued interest. In addition, traded debt securities should be valued at market value as well. When data and information on market value are not available or not applicable for certain debt instruments, nominal or face value (which is the amount of principal to be repaid at maturity) could be used as a proxy. 8 13. The reference time of recording debt transactions is also important because interest accrues continuously. In macroeconomic statistics, flows and stock positions are recorded when economic value is created, transformed, exchanged, transferred, and extinguished. This principle, referred to as the accrual basis, matches the time of recording with the timing of events giving rise to actual resource flows. The accrual basis provides the most comprehensive information because all resource flows are recorded, including other accounts payables and flows related to pension entitlements. 14. The accrual basis can be particularly important when recording interest. Interest on debt instruments should be recorded as it accumulates, with interest payments then reducing the debt stock. When following a nominal valuation, this same approach should be applied to discounts and premia at issuance of debt securities, with the discounts/premia being accrued across the life of the debt instrument.11 Public Debt Data in LIC-DSF 15. Compilation of debt information for debt sustainability analyses (DSAs) under the LIC-DSF focuses on identifying and assessing debt risks.12 Public debt data used for DSA under the LIC DSF broadly follow the PSDS statistical methodology (Table 1). There are, however, some significant differences to facilitate the identification and assessment of risks, related to the coverage and treatment of public sector debt and other information. 11 Financial statements based on the International Public Sector Accounting Standards provide good source data for compiling reliable public sector debt although certain case-by-case adjustments may be needed. 12This is true also for the DSA for countries with market-access (MAC DSA). 9 Table 1. Differences in the Treatment of Public Debt Data Between the Public Sector Debt Statistics (PSDS) and LIC-DSF PSDS LIC-DSF (key differences from the PSDS) Overall purpose Compile and disseminate internationally- Support LICs’ efforts to achieve comparable debt statistics their development goals while minimizing the risk of debt distress Coverage of debt Public debt instruments as defined in SNA/GFS: Same public debt instruments: SDRs; currency and deposits; debt securities; - minus SDR allocations as IMF loans; insurance, pension, and standardized members are generally under no guarantee schemes; and other accounts payable obligation to reconstitute these13 Private sector debt Treated as contingent liabilities, and not part of Included in the public sector debt guaranteed by the public public sector debt. It can be reported as a memo stock sector (including item (PSDSG, Table 5.1) and using a standard provided for borrowing table (PSDSG, Table 5.8a) by state-owned enterprises) Other contingent Contingent liabilities are not part of public - Could include long-term liabilities. sector debt and can be reported using a standard obligations of the general table (PSDSG, Table 5.12) government, such as unfunded liabilities of social security funds; and known and anticipated recognition of contingent liabilities. - Where the recognition of contingent liabilities is less certain, the LIC DSF’s stress tests should assess the potential impact Coverage of the public Institutional coverage and sectorization as Near-complete coverage in line with sector defined in 2008SNA/GFSM 2014 and PSDS 2008SNA/GFSM 2014 and PSDS Guide; external debt of the public sector, defined Guide, but: excluding public as central, state, and local governments, social financial corporations and including security funds, and public financial and non- the central bank (when it borrows on financial corporations the government’s behalf). Gross vs. net debt Both gross and net debt can be reported Gross debt (PSDSG, Table 5.2) Valuation method Nominal value, and for traded debt securities at Face value market value as well (if market value is not available nominal value or face value could be used as a proxy) Definition of external Both by creditor residency and currency of By creditor residency. However, in debt denomination cases of lack of detailed information, debt denominated in foreign currency can be used as a proxy 13SeeAnnex 4 of “Staff Guidance Note on the Application of the Joint Bank-Fund Debt Sustainability Framework for Low Income Countries,” IMF and World Bank (2013). 10 Public Sector Coverage 16. The LIC-DSF is expected to be based on near-complete coverage of public and publicly guaranteed (PPG) debt of the public sector. This is because broad public debt coverage is important to arrive at an assessment of risk of debt distress that is comparable across countries. More importantly, a narrow definition of public debt can contribute to unexpected increases from sources outside the defined perimeter thus underestimating debt risks of the government (e.g., failure to cover parts of public debt incurred by financially weak public sector elements could lead to the eventual migration of such debt onto the government balance sheet). Some of the tools such as debt coverage assessment and stress tests in the LIC DSF are designed to compensate for weaknesses in limited debt coverage in LIDCs. 17. All non-financial state-owned enterprises (SOEs) that create significant fiscal risks should be included in LIC DSFs. SOEs are defined as public enterprises that are majority- owned by the government, but which can be excluded if they meet certain criteria, as explained in Appendix III of the LIC DSF guidance note. These include: ability to publish annual reports including financial statements, undergo regular independent audits, have independent management, borrow without a guarantee from the government, have no involvement in uncompensated quasi-fiscal activities and a track record of positive operating balances. 18. Public financial corporations are excluded in the LIC DSF. They are excluded because the netting out of assets and liabilities between the government and financial corporations would mask the true extent of debt vulnerabilities of a government (e.g., when government debt is largely held by public financial corporations including the central bank). However, the LIC DSF offers options to capture risks emanating from public financial corporations in the contingent liability stress test. 19. Central bank debt is included in the DSF if it is contracted on behalf of the government. 14 In contrast, central bank debt issuances or foreign exchange swaps for the purpose of conducting monetary policy or reserves management are excluded in the DSF. Debt Instruments 20. The LIC DSF includes loans and debt securities (D1), as well as debt arrears and government guarantees in the public debt stock.15 In addition, the LIC DSF can cover other liabilities, including unfunded obligations of social security systems; ongoing restructurings of financial institutions; demand or other guarantees in public private partnerships (PPPs) that have been or are poised to be triggered; verified and recognized obligations arising from a financial 14 Borrowing from the IMF which is a member’s obligation should also be considered the government’s debt. 15A loan contracted by an SOE and guaranteed by an IDA guarantee—where the government gets into an indemnity agreement with IDA—would also be treated as a debt under the LIC DSF. 11 claim (e.g., ICSID) arbitration awards; and arrears owed to suppliers. Further, any omissions from the public debt would be picked up in the contingent liabilities stress test.16 21. Limited cases where debt should be excluded, or the amount adjusted in the DSA are as follows: • When the validity of a claim or the amount of a claim is in dispute, the entire amount in dispute should be treated as a contingent liability in the LIC DSF stress tests; • Claims that are eligible for internationally-agreed debt relief, for example in post-HIPC completion point countries, should be excluded from the DSA; • The amount of external arrears in the LIC DSF would be adjusted in line with restructuring agreements (e.g., external debt that Paris Club member countries have agreed to cancel but not yet reached legally-binding bilateral agreements). Valuation Method 22. Debt is valued at face value in LIC-DSF. Data provided by national debt offices authorities are the primary source of debt input in the DSA. III. REPORTING AND DISSEMINATION OF PUBLIC SECTOR DEBT DATA OF LIDCS Statistical Reporting Systems 23. LIDCs report debt data to four main statistical databases, hosted by the IMF and World Bank, which are closely aligned with international definitions: Quarterly Public Sector Debt Statistics (QPSDS), Quarterly External Debt Statistics (QEDS), Government Finance Statistics (GFS, annual), and the Debtor Reporting System (DRS) from which the aggregate data are published annually in the International Debt Statistics (IDS) database. 17 These databases were created for different purposes (Annex Table 2). Though similar data are presented, the coverage and definitions may differ for specific reasons. None of the databases separately collects data on contingent liabilities and, except for the DRS, publicly guaranteed debt, as well as the terms and conditions of contracts. Although extensive documentation accompanies each database, users may find it a challenge to understand the differences in the 16The contingent liabilities stress test is composed of shocks emanating from other elements of the general government, SOE debt (guaranteed and not guaranteed), PPPs, and financial market vulnerabilities that are not already captured in the headline debt indicators. 17The Quarterly Public Sector Debt Statistics and Quarterly External Debt Statistics database are jointly administered by the IMF and the World Bank. The latter database brings together external debt data in line with the classifications and definitions of the External Debt Statistics: Guide for Compilers and Users, 2013 (EDS Guide) and the 6th edition of the Balance of Payments and International Investment Position Manual, 2009. The IMF’s GFS database contains annual data for all reporting countries in the framework of the GFSM2014. 12 data content in terms of coverage and other dimensions. With the exception of the DRS, reporting to the databases is voluntary.18 Quarterly Public Sector Debt Statistics (QPSDS) database 24. The QPSDS database is designed to collect the most comprehensive, detailed, and internationally comparable PSD data. The database covers outstanding (external and domestic) debt of main subsectors of the public sector (i.e., budgetary central government, central government, general government, nonfinancial public corporations, and financial public corporations) with breakdowns by: original and remaining maturity and type of instrument; currency of denomination; and residency of creditors.19 Quarterly External Debt Statistics (QEDS) 25. The QEDS database provides external debt statistics for general government and the central bank, as part of a country’s total external debt. The database contains gross external debt (defined by residency, not currency of denomination) by sector and broken down by debt instruments. The sector breakdowns include “general government”, “central bank”, “deposit-taking corporations, other than the central bank”, and “other sectors,” with public corporations being included in the latter two subsectors. The instrument breakdowns are similar but not identical to those in the QPSDS database. Government Finance Statistics (GFS) 26. The balance sheet data in the IMF’s annual GFS database provides information for reporting countries that could be used in the QPSDS database and the public sector balance sheet database.20 The annual GFS database contains full balance sheet data for general government and its subsectors, covering nonfinancial assets, financial assets and liabilities with more detailed instrument breakdowns than the QPSDS and QEDS. Since the questionnaire to collect data is designed to capture comprehensive information of government operations and balance sheets, the annual GFS database allows analyzing government debt in the context of broader fiscal developments and conducting stock-flow consistency checks for each debt instrument where data are available, including contingent liabilities. World Bank’s Debtor Reporting System (DRS) 18World Bank loans and financing for countries failing to meet basic DRS reporting requirements cannot be presented to the Board unless the country provides an acceptable plan of action for reporting on its external debt (Annex Table 2). IMF program conditionality can be used to strengthen debt statistics and coverage (this would be done on a case-by-case basis and subject to the requirement of macro criticality). 19 The presentation tables in the PSDS Guide, chapter 5 provide the standardized format suitable for comprehensive debt reporting and analysis. 20The October 2018 Fiscal Monitor emphasizes the importance of capturing assets in the public sector balance sheet for a more thorough analysis of debt vulnerabilities and discusses the construction of a new public sector balance sheet database. - 13 27. The World Bank’s DRS is the most comprehensive database on LIDCs external debt, collecting loan-by-loan information on PPG debt. It was established in 1951 but reporting to the DRS is only mandatory to active and potential borrowers of the World Bank. Though only aggregate information is published in the public domain in the IDS database for confidentiality reasons, the DRS is based on detailed debt information for public and publicly guaranteed external borrowing on a loan-by-loan basis.21 The database is updated regularly and has granular information on individual debts, including on debt service schedule, concessionality, in addition to the basic lending terms such as maturity, grace period, and interest. Data accuracy and comprehensiveness is ensured by validation with other sources such as market data, creditor data, other external statistics such as BoP/IIP, QEDS —including data used in debt analytical exercises led by the World Bank and IMF, such as the medium-term debt strategy (MTDS) or DSA—and rigorous follow-up with government authorities. Using this database, users can conduct a meaningful analysis on both debt outstanding and future debt services as well as debt decomposition by creditor category and concessionality. 28. While the statistical databases serve different purposes, they pose demanding data reporting requirements on LIDCs. This is particularly important for countries that face capacity constraints in collecting and compiling debt statistics (Section V). For instance, the IMF/World Bank QPSDS and QEDS require countries to report over 560 series of data on a quarterly basis (although minimum requirements focus on a narrow set of data, and therefore should promote participation in these database). Given the “public good” nature of such data, enhanced use of modern tools for data exchange would go a long way in reducing the reporting burden. Data Dissemination Systems 29. Under the IMF’s data dissemination standards, LIDCs are encouraged to compile and publish timely and comprehensive statistics, including public sector debt data. In particular, the enhanced General Data Dissemination System (e-GDDS) established in 2015 includes central government gross debt (to be disseminated quarterly within two quarters) as one of the data categories to be disseminated. The e-GDDS is designed to support transparency, encourage statistical development, and help create synergies between data dissemination and surveillance through the effective use of modern information technology tools for data dissemination—common reporting platform, provision of sufficient metadata, and adherence to disciplined timetable for data dissemination. Debt data transparency could be improved further by subscribing to the higher tiers of the IMF’s data dissemination standards where debt data is a mandated category under the observance procedures.22 21The DRS also includes data on SOEs’ debt in 41 countries. 22 https://dsbb.imf.org/sdds/overview. 14 National Reporting Systems 30. The debt coverage at the national level is often narrower than the internationally agreed statistical definition. This is due to (i) national or regional definition of debt which deviate from international standards, or (ii) institutional frameworks that do not give explicit mandate to the debt office or statistics department to collect the relevant data. As a result, coverage of ‘public sector’ debt statistics is limited to the budgetary central government in many LIDCs. 31. Debt statistics in LIDCs refer mostly to the narrowest coverage (loans and securities), and often guarantees. The only additional category partially covered (in 8 percent of LIDCs) are other accounts payable which can be quite large, especially in countries facing tight liquidity constraints. The least reported instrument across all countries is constituted by IPSGS, which do not enter in the debt definitions in most of the national legislations. Similarly, contingent liabilities are rarely monitored or quantified. 32. Most LIDCs are recording debt at face value only. Nominal and face value definitions tend to be used interchangeably. Given that most LIDCs use debt recording systems (e.g. Commonwealth’s COMSEC and UNCTAD’s DMFAS) that define debt in their software at face value and do not allow computation of market value, it is reasonable to infer that debt is evaluated at face value only. This is consistent with DSA requirements but deviates from international standards. 33. With respect to method of accounting, two thirds of LIDCs still use a cash basis. From an analytical perspective, the use of cash accounting can lead to unexpected surprises in the form of accounts payable through accrued obligations (i.e. arrears) that are only discovered when the payment is requested, or inaccurate valuation of securities issued below or above par and recorded at face value.23 Reporting to International Databases and for LIC Debt Sustainability Analysis 34. The DRS has the broadest country and public sector coverage for LIDCs and the most granular information on external public sector debt (Annex Table 3). However, reporting to the DRS is limited only to active and potential borrowers of the World Bank. All but two LIDCs (out of 59 countries) have reported loan-by-loan debt information to DRS of which 53 countries have reported through 2018. The coverage of public sector is broadest under DRS with around 85 percent of countries having reported external debt contracted by their development banks and/or SOEs. 35. Only half of LIDCs have been reporting to other databases as it is voluntary. Coverage of QPSDS—which is the most consistent with international standards—is the most 23Theimportance of recognizing all commitments has also been noted in the recent G-20 Note: Strengthening Public Debt Transparency– the Role of the IMF and the World Bank. 15 limited with less than one third of LIDCs (17 countries) having reported in the past and only 10 countries through 2019Q3. Countries tended to use the central government for public sector coverage whereas debt instrument coverage varies across countries with 7 countries using the narrowest definition (D1). Despite extensive engagement and support provided by the IMF Statistics Department to member countries’ GFS compilers, compliance remains uneven especially among LIDCs.24 36. Most LIDCs have produced a DSA using the new LIC DSF, and its implementation since July 2018 has begun to improve disclosure of public debt statistics. LIDCs are typically required to have a DSA prepared on an annual basis in the context of IMF Article IV consultations and annual IDA credit-grant allocations. 25 For this purpose, all countries are reporting debt data tailored to the LIC DSF requirements. As of end-December, 56 countries have prepared a DSA under the new framework. So far, the strengthened data provision in the revised LIC DSF has led to larger disclosure of public sector debt data in the DSAs of eleven countries, including on non-guaranteed SOEs’ debt, which was previously only occasionally reported.26 Further progress is expected to take time, including related to the implementation of capacity development. Guarantees are now included in over 90 percent of the LIC DSAs, and the number of countries reporting state/local government and SOEs’ debt has increased over the last two years (Figure 3). Figure 3. Improvement in Debt Coverage in LIC-DSAs Since July 2018 (in percent, DSAs to date) Central Gov. A. State and local Gov. B. Other elements in Gen. Gov. (SS or EBF) Guarantees Central Bank Non-guarantees 0 20 40 60 80 100 Sources: Fund staff calculations based on DSAs conducted under the new LIC-DSF (as of mid-October 2019). Yellow-shaded parts represent progress under the implementation of the new LIC DSF 24http://data.imf.org/?sk=A0867067-D23C-4EBC-AD23-D3B015045405. 25 Around 48 countries have conducted a DSA each year during the past 5 years. 26It is worthwhile noting that there are many LIDCs where state/local governments and SOEs cannot borrow debt independently without a government guarantee under their relevant laws. 16 IV. STATISTICAL TREATMENT OF COMPLEX DEBT-CREATING ARRANGEMENTS 37. In recent years, certain debt-creating arrangements have given rise to debt transparency issues. These generally relate to “off-balance sheet” exposures. Care is needed in accounting for them in official statistics and LIC DSFs. This section discusses PPPs, collateral and collateral-like debt, debt obligations relating to pension entitlements, and trade credits. Public-Private Partnerships 38. According to the international standards, PPP contracts could give rise to debt liabilities of the public sector depending on their design. PPPs are long-term contracts between two entities, whereby typically one entity acquires or builds an asset or a set of assets, operates it for a period, and then hands over the asset to a second entity (e.g., a general government or public sector unit). The statistical treatment as to whether PPP contracts create debt liability depends on the economic ownership (not legal ownership) of the assets involved: the economic owner of an asset is the entity which is entitled to claim benefits associated with the use of the asset by virtue of accepting the associated risks. If the government is assessed as being the economic owner of the assets during the contract period but does not make any initial payment for the purchase of the assets at the beginning of the contract period, a transaction must be imputed to cover the acquisition of the assets. In such a case a loan should also be imputed and recorded and subsequent actual government payments to the private partner could be partitioned so that a portion of each payment represents repayment of the loan. If the private partner is assessed as being the economic owner of the asset during the contract period, any debt associated with the acquisition of the asset should be attributed to the private partner. 39. Government guarantees provided for payments under a Power Purchase Agreement (PPA) in the context of a PPP agreement do not constitute debt. While the central government often guarantees payment obligations owed by a SOE as an off-taker to a private independent power producer (IPP), this guarantee would not be included in the public debt stock until the IPP calls on the guarantee. In the LIC DSF, however, this should be evaluated as a potential contingent liability in the stress test.27 Collateral and Collateral-like Debt 40. Collateralized debt obligations or asset-backed securities issued by a public sector unit constitute PSD. Most collateral borrowing is correctly accounted for as debt, but some collateral arrangements create confusion. In the case of an indirect collateralized arrangement, where the collateral is for example assigned to a special purpose vehicle (SPV) which then grants it (or shares in the SPV) to the creditors as collateral, this should be included in the public 27Generally,the risks related to PPPs are illustrated through a default shock where 35 percent of the country’s PPP capital stock (proxying for the present value of direct and potential future fiscal costs from PPP distress and/or cancellations) is triggered under the contingent liabilities stress test in all DSAs when the PPP stock is larger than 3 percent of GDP. 17 debt in DSAs when the government can become liable for the SPV’s obligations even if the SPV is a separate and fully independent entity from the government.28 41. Some commodity-backed arrangements are not collateralized loans in a strict legal sense but are collateral-like transactions and thus need to be reported as debt. For example, a commodity barter transaction or a pre-purchase agreement related to forward sales of commodities. These can be considered as payment for delivery of a good or service but can create an obligation for repayment over an extended period of time. Debt Obligations Relating to Pension Entitlements 42. Pension entitlements of public sector employees with employment-related pension systems constitute debt of the public sector. Pension entitlements are financial claims that existing and future pensioners hold against the government as an employer or a fund designated by the government to pay the pension earned. The statistical method to calculate the debt liability arising from employment-related pension schemes depends on whether the scheme is a defined-benefit scheme or a defined-contribution scheme. In the former, the level of pension benefits is pre-determined and guaranteed, and the present value of any unfunded obligations (i.e., future obligations that would exceed the amount of assets held by the pension fund) is considered debt liability. In the latter, benefits that will be paid eventually depend on the financial performance of the pension fund, and the market value of financial assets held by the pension fund (which could change depending on market conditions) is not considered debt liability. Unfunded liabilities of social security funds, when they are not explicitly recognized as part of general government debt, can be included in the LIC DSF. Trade Credits 43. Trade credits used to meet long-term investment needs should be recorded as debt. Unlike in the statistical treatment, “self-liquidating” trade credits where importers play only an intermediation role by purchasing goods from exporters for immediate onward sale can be excluded from the LIC DSF. Trade credit with a maturity longer than one year should be included in the DSA, because (i) proceeds of sales might be used for different purposes than to repay the trade credit; and (ii) currency mismatches might become an issue. The SOE’s financial soundness is also an important consideration in determining whether trade credit is risky. There have been some cases where state enterprises have built up significant short-term facilities for importing capital goods. Judgement is needed about when short-term facilities may be substituting for longer-term facilities and thus should be included in analytical measures of debt. 28SPV arrangements should be assessed on a case-by-case basis. It is possible that underlying legal documentation may grant investors claims on government resources in the event of default, notwithstanding the assignment of collateral to the SPV. Moreover, a determination should be made, based on the GFSM2014, about whether the SPV is truly an independent entity or if it should be classified as part of the general government. 18 V. FACTORS LIMITING REPORTING OF DEBT DATA BY LIDCS29 44. A number of factors impede LIDCs from compiling and reporting comprehensive public debt data. These include capacity constraints, the treatment of debt in their legal frameworks and unclear definitions of public debt under national laws, and weak governance.30 Capacity Constraints to Collect, Compile and Disseminate Debt Statistics 45. In LIDCs, government resources—both human resources and IT infrastructure — are scarce and constrain the capacity to collect, compile and disseminate debt statistics. Results from the World Bank’s Debt Management Performance Assessment (DeMPA) conducted since 2015, suggest that less than 50 percent of the LIDCs meet the minimum requirements in staff capacity and HR management. Investing in statistical capacity is difficult to achieve in LIDCs where other developmental objectives may have higher priority than providing high quality statistics, which is a long-term commitment with results that are less tangible to the general public. These capacity constraints are especially important when considering the debt of public corporations, social security funds, extrabudgetary funds and subnational governments who may themselves have weak governance structures or limited incentives to co-operate with government officials. Where debt management offices have limited capacity to collect information on debt, data collection will be limited to central government. Also, legal capacity for governments to appropriately evaluate loan contracts is sometimes limited. Legal Framework 46. LIDCs often lack a clearly-defined legal framework for debt management requiring the compilation and reporting of debt statistics. The responsibility to compile, record, and report debt statistics should be clearly established in a country’s legal framework and delegated to a specific agency with credible enforcement mechanisms in the case of noncompliance. The most comprehensive evaluation of debt management in LIDCs, the DeMPA framework, found that only half of a sample of seventeen LICs and LMICs between 2015 and 2017 “have legal frameworks that clearly define the delegation of authority to borrow and undertake debt management activities including the issuance of guarantees, all on behalf of the central government.” Ambiguously defined authority and responsibilities would limit the ability of central governments to manage and monitor public debt in a comprehensive manner. In order to give an idea of instrument coverage in legal frameworks (Debt Management/Public 29Formore detailed discussion, please see G-20 Note: Improving Public Debt Recording, Monitoring, and Reporting Capacity in Low and Lower Middle-Income Countries: Proposed Reforms. 30It should be noted that enhancing debt transparency also depends on increased efforts made by creditors. For example, there have been cases where a non-disclosure clause embedded in a loan agreement prevented debtor countries from disclosing the important nature of debt contracts. The G20 operational guidelines for sustainable financing encourage creditors to share information on their lending and contractual terms. 19 Finance Management Acts), Figure 4 below provides a breakdown of the language used (if any) relating to debt instruments covered. Figure 4. Instrument Identification in Legal Frameworks in LIDCs 5.5% 15.8% 14.4% 24.7% 20.5% 6.8% 4.1% 8.2% n/a not identified securities securities and 'debt instruments' securities and guarantees securities and loans securities, loans, guarantees total debt liabilities Source: Bank staff calculation, based on Debt Management and/or Public Finance Management Laws (n. countries =59) 47. Where coverage of public sector debt is narrowly defined in legislations, debt compilers do not have the legal backing to collect debt statistics from broader public agencies. Debt definitions in national laws often do not cover debt of public sector entities outside of the central or general government such as public corporations making it difficult for PSDS compilers to seek cooperation from broader public sector entities, especially when faced with the capacity constraints discussed above.31 Governance 48. The lack of strong incentives for senior administrative and political management is one of the impediments underpinning weaknesses in debt recording, monitoring and reporting. The weak incentives relate to lack of demand for reliable, timely and comprehensive data, limited public scrutiny, and limited integration with other PFM systems and in some cases, poor alignment between statistical reporting entities and the accountability structure of government. Also, audits of debt management operations in LIDCs are rare. This might have 31 Forinstance, in both CEMAC and EAMU convergence criteria, the debt coverage in the debt rules is for the central government only. In Mongolia, the 2015 Debt Management Law has narrowed the debt definition from public to general government. 20 discouraged LIDCs to pursue more comprehensive coverage of debt statistics, adopt modern IT infrastructure, and strengthen legal backing with credible enforcement mechanisms for noncompliance. Moreover, accounting for grey areas can create incentives to minimize disclosed debt. This is particularly true in the case of PPP contracts where the ability to treat these risks as off-balance sheet items (contingent liabilities) may lead to a ‘PPP bias’ where governments engage in these type of contracts as they will have no effect on public sector debt statistics even if they increase the risk of future debt surprises. VI. PRIORITIES TO IMPROVE PUBLIC DEBT DATA AVAILABILITY 49. Against this backdrop, there are a number of priorities to improve debt data availability. There are already initiatives underway in some of these areas: • Strengthen the legal framework and institutional capacity to enhance debt reporting and debt transparency. This would require capacity development in these areas to be prioritized in the context of a country’s capacity development strategy. • Promote the use of the standard definitions and concepts of PSD to enhance sector and debt instrument coverage. The PSDS Guide developed jointly with several international institutions provides such a definition. This definition is also informing the LIC-DSF. Further promoting the use of the definition can be achieved by encouraging debt managers in LIDCs to take the newly-launched IMF online course on PSDS and other capacity development activities. • Enhance the QPSDS database, which can serve—together with DRS—as a global source of timely and comprehensive public debt data.32 The QPSDS database, which is intended to cover all countries, is already being used for monitoring progress under the G-20 Data Gaps Initiative. To adequately serve this global purpose QPSDS coverage (country, sector, and instrument), countries’ compliance, and data validation all need to be improved through intensive technical assistance. Concerted and sustained efforts are needed from both the IMF/World Bank and the reporting countries to enhance awareness (through outreach), strengthen motivation given voluntary reporting (by highlighting that improved transparency would facilitate creditors’ lending decisions by reducing uncertainty), and extend capacity development support. • Enhance the World Bank’s DRS to capture more granular details on the terms and conditions of loans, including collateralization features and domestic debt. This would provide more detail for use in the DSAs and address some of the debt transparency issues highlighted by the G20, such as the risks arising from collateralized debt obligations. This would require systematic collection of additional instrument-level 32 Nearly all LIDCs are currently reporting to the DRS. This therefore begs the question as to why the DRS should not just be used to collect both external and domestic public debt. Nonetheless, it is important to have a database to which all countries can report debt data in a comparable and consistent format. 21 information. Collecting additional granular information associated with specific transactions would require both capacities to assess appropriate statistical treatment and implementation of a process that encourages and supports provision of data. On this front, the World Bank’s DRS is already in the process of piloting an initiative to expand debt coverage to domestic debt. • Reduce the reporting burden. The harmonization of the debt definition and reporting templates currently used by IFIs would simplify the data compilation. At a country level, there should be a single reporting channel that would source multiple databases. This can be achieved by encouraging LIDCs to use data structure definitions to unify their databases and to utilize modern IT tools for data dissemination (a common reporting platform, provision of sufficient metadata, and adherence to disciplined timetable for data dissemination). The IMF and the World Bank will continue to collaborate with the two main debt software providers (COMSEC and UNCTAD) to encourage LIDCs to harmonize debt definitions and compile debt based on international standards. • Continue to implement the new LIC DSF requirements. Write-ups should include a full description of the data used for the analysis and this can be posted on the respective IMF-World Bank DSF websites to give it greater visibility. Disclosure of the coverage of public sector and debt instruments also needs to continue being strengthened under the LIC DSF. A continued review and support of debt data reporting in DSAs is warranted and can be discussed in the context of updates on the multi-pronged approach. Further guidance may need to be issued by the IMF/World Bank on how to treat complex debt arrangements. 22 Annex Table 1. List of Low-Income Developing Countries Afghanistan Honduras Rwanda Bangladesh Kenya Sao Tome and Principe Benin Kiribati Senegal Bhutan Kyrgyz Republic Siera Leone Burkina Faso Lao P.D.R. Solomon Islands Burundi Lesotho Somalia Cambodia Liberia South Sudan Cameroon Madagascar Sudan Central African Republic Malawi Tajikistan Chad Mali Tanzania Comoros Mauritania The Gambia Cote d'Ivore Moldova Timor-Leste Democratic Republic of the Congo Mozambique Togo Djibouti Myanmar Uganda Eritrea Nepal Uzbekistan Ethiopia Nicaragua Vietnam Ghana Niger Yemen Guinea Nigeria Zambia Guniea-Bissau Papua New Guinea Zimbabwe Haiti Republic of Congo Note: The country grouping is according to the IMF World Economic Outlook, April 2019. 23 Annex Table 2. Public Sector Debt Data in the International Databases Database Institution Coverage1 Main purposes/collection mechanism Frequency Latest available data 1. Statistical database Quarterly IMF/WB BCG, CG, Collect and disseminate public sector debt Quarterly 2019Q2 Public Sector GG, NFC, statistics based on the PSDS Guide. Debt FC, PS Statistics Data are reported on voluntary basis using a questionnaire form. Quarterly IMF/WB CG, GG, Collect and disseminate comparable cross- Quarterly 2019Q2 External NFC, FC, country external debt statistics based on the Debt PS EDS Guide. Statistics Data are reported on voluntary basis using a questionnaire form. Government IMF BCG, Collect and disseminate government finance Annual 2017 Financial (Statistics EBF, SSF, statistics (including balance sheet data) based Statistics Department) CG, SG, on the GFS Manual. LG, GG Data are reported on voluntary basis using a questionnaire form. 2. Database for IMF surveillance World IMF GG Contains selected macroeconomic data series Semi- 2019H2 Economic (Research from the statistical appendix of the World annual Outlook Department) Economic Outlook report, which presents the Database IMF staff's projections of economic developments at the global level, in major country groups and in many individual countries. Data provided for IMF surveillance are sent from IMF country teams to the database. Global Debt IMF CG, GG, Comprises total gross debt of the (private and Annual 2017 Database (Fiscal NFC, PS public) nonfinancial sector for an unbalanced Affairs panel of 190 advanced economies, emerging Department) market economies and low-income countries, dating back to 1950. It adopts a multidimensional approach by offering multiple debt series with different coverages. The integrity of the data has been checked through bilateral consultations with officials and IMF country desks of all countries in the sample, setting a higher data quality standard. Public Sector IMF CG, GG, Shows comprehensive estimates of public Annual 2016 Balance (Fiscal NFC, FC, sector assets and liabilities that formed the Sheet Affairs PS basis for the analysis presented in the October Department) 2018 edition of the Fiscal Monitor. The 24 database originally covered public sector balance sheets for a broad sample of 31 countries, covering 61 percent of the global economy. Since October 2018, the database has been updated with PSBS data for another 7 countries and now covers 63 percent of the global economy. The PSBS database is compiled on a best efforts basis, using the conceptual framework of the GFS Manual 2014. Data for the central and general government generally replicate data reported by country authorities in the IMF’s Government Finance Statistics database. Data for the central bank generally replicate data reported by country authorities in the IMF’s Monetary and Finance Statistics database. Where these data fail to cover all categories of assets and liabilities, they are complemented by other data reported by statistical authorities at the national level, other international organizations, or staff estimates. Data sources for public corporations are country specific and are captured in the country specific metadata documents. 3. Loan-by-loan data Debtor World Bank CG, GG, Since 1951, World Bank Debtor Reporting Quarterly 2019Q3 Reporting NFC, FC, System requirements were instituted (as per OP reporting System PS 14.10), any country (Government Authority) of the new that borrows from IBRD or IDA is required to commitme provide comprehensive information on its nt external debt obligations until all obligations to IBRD and IDA are expunged. The underlying Annual 2018 rationale for the collection of these data was to reporting enable the World Bank to assure itself of the for debt servicing capacity of the countries to individual which it lent. In existence for over sixty years, transactio the rationale for the DRS was an institutional ns of the need to be able to assess the creditworthiness debt and debt servicing capacity of Bank borrowers. instrument s Reporting requirements demand for quarterly reporting of new borrowing commitments of public and publicly guaranteed debt, an annual loan-by-loan statement of stocks and flows, and an aggregate reporting of stocks and flows on private non-guaranteed debt. 1BCG: Budgetary central government; CG: Central Government; EBF: Extrabudgetary Funds; LG: Local Government; SG: State Government; SSF: Social Security Funds; GG: General Government; NFC: Nonfinancial Public Corporations; FC: Financial Public Corporations; and PS: Public Sector. Central bank data are included in the data for financial public corporations (FC). 25 Annex Table 3. Reporting Status of Public Sector Debt by LIDCs1 QPSDS2/ QEDS3/ GFS2/ DRS4/ e-GDDS2/ First Last First Last First Last First Last First Last 1 Afghanistan Data reported 2017Q1 2019Q2 2004 2006 2006 2018 Coverage N/A PSE, PrSE/SRD, C&D, Ln BCG/D2 GG, CB & SOEs N/A 2 Bangladesh Data reported 2009Q3 2019Q2 2013Q3 2019Q2 1972 2018 2011Q1 2019Q2 Coverage PS/D4 CG, PrSE/ADI N/A GG, CB & SOEs CG/Domestic 3 Benin Data reported 1970 2018 2015Q1 2016Q3 Coverage N/A N/A N/A GG, CB & SOEs CG/D1 4 Bhutan Data reported 2003 2018 1981 2018 2012 2018 Coverage N/A N/A BCG/D2 GG, CB & SOEs CG/D1 5 Burkina Faso Data reported 2008Q1 2019Q2 1970 2018 2005 2017 Coverage N/A CG/SDR,Ln N/A GG, CB, Dev. Banks & SOEs CG/External 6 Burundi Data reported 1970 2018 Coverage N/A N/A N/A GG, CB, Dev. Banks & SOEs N/A 7 Cambodia Data reported 2008Q4 2019Q2 1981 2018 1995 2016 Coverage N/A PSE, PrSE/SRD, C&D, Ln, N/A GG, CB & SOEs CG/D1 Other 8 Cameroon Data reported 2007Q4 2019Q2 1970 2018 2017Q1 2019Q2 Coverage N/A PSE, PrSE/DS, Ln N/A GG, CB, Dev. Banks & SOEs CG/D1 9 Central African Republic Data reported 2009Q2 2010Q3 1970 2018 Coverage N/A PSE/ Ln N/A GG, CB, Dev. Banks & SOEs N/A # Chad Data reported 1970 2015 Coverage N/A N/A N/A GG, CB, Dev. Banks & SOEs N/A # Comoros Data reported 1970 2018 Coverage N/A N/A N/A GG, CB, Dev. Banks & SOEs N/A # Congo, DR Data reported 2013Q3 2019Q2 1972 1989 1970 2018 Coverage GG/D3 N/A CG/D1 GG, CB, Dev. Banks & SOEs N/A # Congo, Republic of Data reported 2009 2010 1970 2018 Coverage N/A N/A CG/D3 GG, CB, Dev. Banks & SOEs N/A # Cote d'Ivoire Data reported 2017Q1 2019Q2 2010Q2 2019Q2 1970 2018 2006 2015 Coverage PS/D4 PSE/DS, Ln N/A GG, CB, Dev. Banks & SOEs CG/D1 # Djibouti Data reported 2017Q4 2019Q2 1970 2018 Coverage N/A PSE, PrSE/SRD, C&D, Ln, N/A GG, CB & SOEs N/A Other # Eritrea Data reported 1994 2009 Coverage N/A N/A N/A CG & CB N/A # Ethiopia Data reported 2007Q1 2019Q2 2014 2018 1970 2018 Coverage N/A GG/SDR, Ln, DS BCG/D4 GG, CB, Dev. Banks & SOEs N/A # Gambia, The Data reported 2005 2009 1970 2018 Coverage N/A N/A CG/D1 GG, CB & SOEs N/A # Ghana Data reported 2006Q3 2011Q3 1970 2018 2018M1 2019M3 Coverage N/A CG/SDR, DS, Ln N/A CG, CB, Dev. Banks & SOEs CG/D1 # Guinea Data reported 1970 2018 Coverage N/A N/A N/A GG, CB, Dev. Banks & SOEs N/A 26 Annex Table 3. Reporting Status of Public Sector Debt by LIDCs1 (continued) QPSDS2/ QEDS3/ GFS2/ DRS4/ e-GDDS2/ First Last First Last First Last First Last First Last # Guinea-Bissau Data reported 1975 2018 Coverage N/A N/A N/A CG & CB N/A # Haiti Data reported 1970 2018 Coverage N/A N/A N/A GG, CB & SOEs N/A # Honduras Data reported 2000Q1 2019Q2 2002Q4 2019Q2 1970 2018 Coverage PS/D3 PSE, PrSE/SRD,DS, Ln N/A GG, CB & SOEs N/A # Kenya Data reported 2009Q2 2017Q4 2008Q2 2019Q2 2009 2011 1970 2018 2017Q3 2018Q3 Coverage PS/D1 PSE/ Ln BCG/D1 GG, CB, SOEs CG/D1 # Kiribati Data reported 2013Q1 2014Q4 Coverage N/A PSE, PrSE/SRD, Ln N/A N/A N/A # Kyrgyz Republic Data reported 2014Q1 2019Q1 2003Q3 2019Q2 2014 2018 1970 2018 Coverage GG/D1 PSE, PrSE/SRD, C&D, DS, Ln, GG/D4 GG & CB N/A Other # Lao PDR Data reported 1970 2018 Coverage N/A N/A N/A GG, CB & SOEs N/A # Lesotho Data reported 1970 2018 2004 2015 Coverage N/A N/A N/A GG, CB, SOEs CG/D1 # Liberia Data reported 2011Q4 2016Q4 2011 2012 1970 2018 Coverage N/A PSE/Ln CG/D2 GG, CB & SOEs N/A # Madagascar Data reported 2011Q1 2019Q2 2007Q1 2019Q1 1972 1974 1970 2018 Coverage CG/D1 PSE, PrSE/SRD, C&D, TC&A, CG/D1 GG, CB, Dev. Banks & SOEs N/A Ln, Other # Malawi Data reported 2011Q1 2014Q3 2009 2018 1970 2018 2014Q3 2016Q2 Coverage PS/D2 N/A BCG/D2 GG, CB & SOEs CG/D2 # Mali Data reported 1980 1986 1970 2018 Coverage N/A N/A CG/D1 GG, CB, Dev. Banks, & SOEs N/A # Mauritania Data reported 1970 2018 Coverage N/A N/A N/A GG, CB, Dev. Banks & SOEs N/A # Moldova Data reported 2009Q3 2019Q2 2004Q1 2019Q2 2011 2018 1992 2018 Coverage PS/D2 PSE, PrSE/ADI GG/D4 GG, CB, Dev. Banks & SOEs N/A # Mozambique Data reported 2016 2018 1984 2018 Coverage N/A N/A BCG/D4 GG, CB & SOEs N/A # Myanmar Data reported 1970 2018 Coverage N/A N/A N/A GG, CB, Dev. Banks & SOEs N/A # Nepal Data reported 2009Q1-2016Q4 2018Q3-2019Q2 2009Q2 2019Q2 1974 1989 1970 2018 2013Q3 2016Q4 Coverage BCG/D1 GG, PrSE/SDR, C&D, Ln CG/D1 GG, CB & SOEs CG/D1 # Nicaragua Data reported 2010Q1 2019Q1 2007Q3 2019Q2 1970 2018 Coverage PS/D3 PSE, PrSE/SRD, C&D, TC&A, N/A GG, CB, Dev. Banks & SOEs N/A Ln # Niger Data reported 1970 2018 Coverage N/A N/A N/A GG, CB, Dev. Banks & SOEs N/A 27 Annex Table 3. Reporting Status of Public Sector Debt by LIDCs1(concluded) QPSDS2/ QEDS3/ GFS2/ DRS4/ e-GDDS2/ First Last First Last First Last First Last First Last # Nigeria Data reported 2009Q4 2019Q1 2007Q4 2015Q4 1970 2018 2013Q4 2015Q4 Coverage PS/D1 PSE/SDR, LN N/A GG, CB, Dev. Banks & SOEs CG/D1 # Papua New Guinea Data reported 2011Q4 2019Q2 2014 2018 1970 2018 Coverage N/A PSE, PrSE/SRD, C&D, DS, BCG/D2 CG, CB, Dev. Banks & SOEs N/A TC&A, Ln # Rwanda Data reported 2017Q1 2019Q2 2006Q3 2019Q2 1977 1989 1970 2018 2015Q4 2019Q2 Coverage CG/D1 CG, PrSE/SRD, C&D, DS, Ln CG/D1 GG, CB, Dev. Banks & SOEs CG/D1 # Sao Tome & Principe Data reported 1970 2018 Coverage N/A N/A N/A GG & CB N/A # Senegal Data reported 2014Q1 2019Q2 1970 2018 2017Q2 2019Q3 Coverage CG/D2 N/A N/A GG, CB, Dev. Banks & SOEs CG/D1 # Sierra Leone Data reported 2007Q4 2018Q3 1970 2018 Coverage N/A CG/Ln N/A CG, CB & SOEs N/A # Solomon Islands Data reported 2011Q1 2019Q2 2012 2018 1978 2018 Coverage N/A PSE, PrSE/SRD, C&D, DS, BCG/D4 GG & CB N/A TC&A, Ln, other # Somalia Data reported 1970 1992 Coverage N/A N/A N/A GG, CB, Dev. Banks & SOEs N/A # South Sudan Data reported Coverage N/A N/A N/A N/A N/A # Sudan Data reported 1970 2018 Coverage N/A N/A N/A GG, CB & SOEs N/A # Tajikistan Data reported 2008Q1 2019Q2 1992 2018 Coverage N/A PSE, PrSE/ADI N/A GG, CB & SOEs N/A # Tanzania Data reported 2010Q2 2014Q2 2010Q1 2013Q2 1970 2015 2014Q1 2015Q4 Coverage GG/D1 GG/SDR, Ln, Other N/A GG, CB & SOEs CG/D1 # Timor-Leste Data reported 2012 2018 Coverage N/A N/A N/A CG & CB N/A # Togo Data reported 2011Q1 2011Q4 1983 1986 1970 2018 2008 2016 Coverage CG / D2 N/A CG/D1 GG, CB, Dev. Banks & SOEs CG/D1 # Uganda Data reported 2009Q3 2019Q2 2006Q3 2019Q2 2018 2018 1970 2018 2015Q3 2019Q1 Coverage PS/D3 CG, PrSE/SRD, C&D, DS, Ln, GG/D4 GG, CB, Dev. Banks & SOEs CG/D1 TC&A, Other # Uzbekistan Data reported 1991 2018 2017Q1 2019Q2 Coverage N/A N/A N/A GG, CB, Dev. Banks & SOEs CG/External # Yemen, Republic of Data reported 2006Q3 2018Q4 1970 2018 Coverage N/A CG/C&D, Ln N/A GG, CB, Dev. Banks & SOEs N/A # Zambia Data reported 2011Q1 2019Q1 2010 2018 1970 2018 2009 2016 Coverage N/A PrSe/Ln GG, CB, Dev. Banks & SOEs CG/D1 # Zimbabwe Data reported 1970 2018 Coverage N/A N/A GG, CB, Dev. Banks & SOEs N/A # Vietnam Data reported 1981 2018 Coverage N/A N/A GG, CB, Dev. Banks & SOEs N/A 1/ BCG: budgetary central government, CG: central government, GG: general government, PS: public sector, CB: central bank, SOEs: state-owned enterprises, and Dev. Banks: official development banks. 2/ D1: debt securities and loans, D2: D1 plus SDRs and currency and deposits, D3: D2 plus accounts payable, and D5: D4 plus insurance, pension, and standardized guarantee schemes. 3/PSE: public sector external debt, PrSE: private sector external debt, SDR: special drawing rights, C&D: currency and deposits, DS: debt securities, Ln: loans, TC&A: trade credits and advances, Other: other debt liabilites, and ADI : all debt instruments. 4/ For DRS, debt data is reported on a loan-by-loan basis for all countries. 28