Report No. 43353_MN Mongolia Consolidating the Gains, Managing Booms and Busts, and Moving to Better Service Delivery A Public Expenditure and Financial Management Review (In Two Volumes) Volume I: Core Report January 2, 2009 Poverty Reduction and Economic Management Unit East Asia and Pacific Region Document of the World Bank ABBREVIATIONS AND ACRONYMS ALMP Active Labor MarketPrograms MTBF Medium-termBudgetFramework ASEAN Associationof Southeast AsianNations MTFF Medium-termFiscalFramework CAB Cyclically Adjusted Balance MTCSRS Medium-termCivil Service Reform CBC CongressionalBudget Committees MTEF Medium-termExpenditureFramework CCT ccu ConditionalCash Transfer MUB Municipality ofUlaanbaatar Client ConstructionUnit NASA NationalAssessment of Students' Achievements CEE Central and EasternEurope NCD Cardiovascular mental disorder CIS Commonwealth of IndependentStates NDC Notional DefinedContribution CMP Child Money Program NHA NationalHealthAccounts CPI csc Consumer PriceIndex NSO National Statistics Office Civil Service Council ODA Official DevelopmentAssistance DAC DevelopmentAssistance Committee OECD Organizationfor EconomicCooperation and DF Development Fund Development DMD Debt ManagementDivision PEFMR Public Expenditureand FinancialManagement DOTS ObservedTreatment Short Course Review EDCM EducationDonors ConsultativeMechanism PETS Public ExpenditureTrackingSurvey EITI Extractive IndustryTransparency Initiative PFI ProjectFinancialIntermediary EPF EmploymentPromotionFund PFM PublicFinancialManagement FDI ForeignDirect Investment PIP Public InvestmentProgram FTI Fast Track Initiative PIU ProjectImplementingUnits FE&T Finance, Economy and Treasury PMU ProjectManagement Unit FGP Family GroupPractice PPCD Procurement Policy and Coordination Department FMD Foodand Mouth Disease PPLM Public ProcurementLaw of Mongolia GDP Gross Domestic Product PRR Poverty Risk Ratio GFMIS Government FinancialManagement PSMFL Public Sector Managementand FinanceLaw InformationSystem PTR Pupil to Teacher Ratio GoM Government o f Mongolia R&D Research and Development HEIs HigherEducationInstitutions SEG Socio-EconomicGuidelines HIES Household Incomeand Expenditures Survey SES Sanitary EpidemiologicalServices HIF HealthInsuranceFund S I 0 Social InsuranceOffice HIV Humanimmunodeficiencyvirus SHI Social Health Insurance HR HumanResources SNI National Systemof Investment HSMP HealthSector Master Plan SPES Special ProtectedAreas ICOR IncrementalCapital-outputRatio SOE State OwnedEnterprises IDF InstitutionalDevelopmentFund SSIA State Specialized InspectionAgency IE ImpactEvaluation SSIGO State Social InsuranceGeneral Office IMF InternationalMonetaryFund SSMP Social Security Sector Master Plan IMR Infantmortality rates STI Sexually TransmittedInfection INTOSAI InternationalOrganizationof Supreme Audit SWF Social Welfare Fund Institutions TB Tuberculosis LSMS Living StandardMeasurement Survey USMR Under-FiveMortality Rate MECS Ministry of Education, Cultureand Science UNDP UnitedNationsDevelopment Programme MDG Millennium Development Goals UNFPA UnitedNationsFund for PopulationActivities M&E Monitoring and evaluation UNESCO UnitedNationsEducational, Scientific and Cultural MICs Middle Income Countries Organization MLS Minimum Living Subsistence UNICEF UnitedNations InternationalChildren's Emergency MNAO MongoliaNationalAudit Office Fund MNT MongolianTugriks UIF UnemploymentInsuranceFund MOF Ministry ofFinance VAT Value Added Tax MOFALI Ministry of Food and Agriculture and Light WDI World DevelopmentIndicators Industry WIF Work Injury Fund MOH Ministry o f Health WHO World HealthOrganization MOP Ministry o f Planning MOSWL Ministry of Social Welfare and Labor MSL Minimum Subsistence Level MTB Medium-termBudgeting ACKNOWLEDGEMENTS This report has beenpreparedby a core team led by Ms.Genevieve Boyreau (Macroeconomic and Fiscal). The team comprised GiovanaDore (Environment), Andrew Goodland (Agriculture), TserendagvaGerelgua (Procurement), Byambatsogt Jugder (Education, Social Protection), Ulrich Koester (Consultant, Agriculture), Soonman Kwon (Consultant, Health), David I (Financial Management), Xiaoping Li (Procurement), Magnus Lindelow (Health), Ochir Lkhagvasuren (Financial Management), Shabih Mohib, (Public Sector), Katherine Nesmith (Education), Dorjdari Namkhaijantsant (Consultant, Open Society Forum, local governance), Ian Nightingale (Procurement), John Richardson (Consultant, Procurement), Cristobal Ridao-Can0 (Education, Social Protection), Mark Dorfman (Social Insurance), Jinan Shi (Procurement), Altantsetseg Shiilegmaa (overall assistance). Valuable comments on earlier drafts of the chapters, backgroundinputs, and guidance duringthe preparation on this report and PEFMR missions, were received from Messrs.Christopher Finch, BarbaraNunberg, Arshad Sayed, Bert Hofinan. Mr. Chuluunzagd Batbayar and Otgonbayar Yadmaa (EACMF) provided able assistance with the administrative and logistical arrangements for the task team during missions. Ms. Gloria Elmore was responsible for document preparation and formatting, a task she undertook with usual ease and attention to detail. This report would not have been possible without the guidance from H.E. Dr. Bayartsaikhan (Minister of Finance), and State Secretary Khurelbaatar (Ministry of Finance) during the PEFMR missions. The team appreciates the excellent collaborationfrom Mr.Batjargal (Director, FiscalPolicy Department, Ministry of Finance, Head of PEFMR Governmental Working Group and official counterpart, Order of the Minister of Finance #110,see annex) andhis colleagues for makingthe missionsproductive, andproviding critical data for the PEFMR team to do its research and analysis in the subsequent months, the results of which are presentedinthis report. We are grateful for the comments received from the State Secretaries and other officials of the Ministries of: Finance (MoF); Food and Agriculture and Light Industry (MOFALI); Social Welfare and Labor (MSWL); Education, Science and Culture (MESC), Nature and Environment(MNE); Health (MoH); Fuel and Energy (MFE); Roads, Transportation and Tourism (MRTT); NationalStatisticalOffice (NSO); Bank of Mongolia (BoM), at the State Khural the Budget Standing Committee, Economic Standing Committee and Social Policy Standing Committee; the Mongolia National Audit Office (MNAO); Civil Service Council (CSC), Open Society Forum (OSF) of the during the PEFMR working group meetings and workshops in Ulaanbaatar in January and June 2007 that were held to discuss the preliminary findings o f the team as they became available. Peer Reviewers for this Economic Report are Messrs. Mandar Jayawant (Deputy Country Director, Asian Development Bank, Ulaanbaatar, Mongolia), DaehaengKim (Fiscal Affairs Division, IMF) and Francisco Galrao Carneiro (World Bank). Overall guidance for the report was providedby Mr. David Dollar (World Bank Country Director for Mongolia), Vikram Nehru (Acting Regional Chief Economist and Sector Director, PREM, East-Asia and Pacific Region), and Sudarshan Gooptu (Lead Economist, PREM, East- Asia and Pacific Region). TABLE OF CONTENTS VOLUME I EXECUTIVESUMMARY ........................................................................................................... vii CHAPTER 1FISCALTRENDS ..................................................................................................... 1 Progressand Challenges........................................................................................................... 1 KeyFuturePriorities Introduction............................................................................................................................... ................................................................................................................. 1 3 OverallFiscalTrends ................................................................................................................ 3 Economic CompositionofExpenditures ................................................................................ 13 FunctionalCompositionof Expenditures ................................................................................ 16 Local Expenditures ................................................................................................................. 18 CHAPTER2 PUBLIC SECTORFINANCIAL MANAGEMENT.............................................. 23 Progressand Challenges......................................................................................................... Key Future Priorities .............................................................................................................. 23 ............................................................................................................................. 24 The LegislativeFramework.................................................................................................... Introduction 26 26 A reviewofthe PSFMLimplementationprogress........................................................... 27 PublicFinancePlanning .......................................................................................................... 29 Fiscalplanning-Medium-termto annual........................................................................ Fiscalplanning-the PublicInvestmentPlan................................................................... 30 35 ExecutionofPublicFinances .................................................................................................. Procurement............................................................................................................................ 37 41 Internal Audit.......................................................................................................................... 43 DisclosureandTransparency of Public Finances ................................................................... ExternalAudit......................................................................................................................... 43 44 CHAPTER3 EDUCATION .......................................................................................................... 49 Progressand Challenges ......................................................................................................... 49 KeyFuturePriorities ............................................................................................................... Outcomes, Achievements andRemainingChallenges ............................................................ 51 49 .......................................................................................................................... 51 Achievements................................................................................................................... Outcomes 53 Challenges ........................................................................................................................ 53 PublicFinancingof Education ................................................................................................ 54 Educationexpenditures..................................................................................................... 54 Public EducationFinancingSystem........................................................................................ Sourcesof funding............................................................................................................ 56 Equityof PublicFinancingof Education................................................................................ 59 61 Demand-sidefinancing..................................................................................................... 63 Efficiencyand Effectiveness of PublicFinancingof Education............................................. Demand-sidefinancing..................................................................................................... 65 68 1 CHAPTER4 HEALTH ................................................................................................................. 71 Progress and Challenges ......................................................................................................... 71 Key FuturePriorities ............................................................................................................... Outcomes Achievements and RemainingChallenges............................................................. 71 73 Financingof HealthCare........................................................................................................ HealthOutcomes.............................................................................................................. 73 Healthfinancing in Mongolia:the big picture.................................................................. 78 Improvingthe allocationof government healthexpenditures.......................................... 78 81 What future for social health insurance?.......................................................................... 84 Optionsfor the future Improvingpurchasing....................................................................................................... ....................................................................................................... 88 The Delivery of Health Services............................................................................................. 89 Efficiency of service delivery........................................................................................... 90 92 Better payment methods for hospitals .............................................................................. 94 Improvingprimary healthcare services ........................................................................... 96 CHAPTER 5 SOCIAL PROTECTION......................................................................................... 99 Progress andChallenges ......................................................................................................... 99 Key FuturePriorities ............................................................................................................... 99 Overview of SocialProtectionSpending .............................................................................. 101 SocialAssistance ................................................................................................................... 102 Key milestones and overalltrends .................................................................................. ................102 104 The targeted CMP era-a good start with great potential................................................. Social assistanceinthe pre-CMP era -the right idea but inadequatesystem The universalCMP-a costly and unsustainablespending of public resources ............ 110 105 The Makingof a Modern Social Assistance System in Mongolia:A Roadmap...................112 A possible way forward.................................................................................................. Pensions ................................................................................................................................ 114 115 Overview and characteristics ofthe Pensionsystem...................................................... Revenues and expendituresof social insurance.............................................................. 115 120 Key pensionpolicy issues ............................................................................................... 122 CHAPTER 6 AGRICULTURE ................................................................................................... 129 Progress and Challenges ....................................................................................................... 129 Key FuturePriorities ............................................................................................................. 129 Background. Achievements and Challenges .............................................................................................................. ......................................................................... 131 The Budget Cycle.................................................................................................................. Public Expenditures 133 136 Donor Support....................................................................................................................... Efficiency.............................................................................................................................. 137 Optionsfor Reforms-Rethinkingthe Role ofthe Government .......................................... 140 141 BOXES 1. Public finances in Mongolia: key facts............................................................................. viii 1.1 1.2 Commodity saving and stabilizationfunds: Asset allocationand spendingrules ..............11 What fiscal rules aroundthe world?..................................................................................... 9 1.3 Civil Service structure andchanges andwage bill Increases ............................................. Windfall tax and development fund.................................................................................... 12 1.4 14 1.5 Mongolia Public Investment Program ................................................................................ 15 1.6 19 Performance-basedmanagement-Theory and practice in Mongolia.,.............................. Intergovernmental fiscal mechanism in the education sector ............................................. 2.1 28 2.2 A summary of the PSMFL key provisions......................................................................... 29 2.3 Why a centralized public budget managementframework is appropriate in Mongolia ..... 31 2.4 Budget preparation at the Municipality of Ulaanbaatar...................................................... 33 2.5 Good practices on budget deliberations.............................................................................. 34 2.6 Wells in Mongolia -To be or not to be .............................................................................. 34 2.7 Use of country financial management systems................................................................... Good practice on investment programs -Chile.................................................................. 36 2.8 40 2.9 44 62 Food purchasedby schools................................................................................................. Internationalexperience with funding formulas ................................................................. OECD best practices on budget transparency..................................................................... 3.1 3.2 66 4.1 Key features of social health insurance inMongolia .......................................................... 84 5.1 The makingof a modern social assistance system inMongolia: a roadmap .................... 6.1 Inefficiencies in Budget formation and execution: the example of new wells .................114 136 6.2 Analytical requirements inthe Ministryof Food and Agriculture .................................... 142 FIGURES 1.1 public deficits........................................................................................................................ Expenditures have increasedrapidly. but revenues even faster. leading to a reduction in 3 1.2 Mining contributes an increasing share of public revenues ................................................. 4 1.3 Public spending is increasing rapidly driven by investment, wages and social transfers .....5 1.4 Fiscal policy is pro-cyclical .................................................................................................. 6 1.5 Forecastingrevenue has become a more challengingtask ................................................. 10 1.6 Official development assistancecontribution is declining ................................................. 15 1.7a Priorities are given to social sectors ................................................................................... 17 1.7b Butare increasingly shiftingtowards evenmore social spending ...................................... 17 1.8 Local government expenditures were reduced inaccordancewith the transfer o f service delivery responsibilities to the Center ................................................................................ 20 2.1 Mongolia total investment and capital maintenance ........................................................... 37 3.1 Mongolia basic education ranks high in East-Asia but not so high compared to Central Asia..................................................................................................................................... 51 3.2 Students' achievement by location ..................................................................................... 52 3.3 Mongolia spends relatively ore in education than others east and central Asia transition countries.............................................................................................................................. 54 3.4 55 56 Donor assistancefocuses on fixed capital investment........................................................ 2006 revenue sources by sector .......................................................................................... Sub-sector shares of education expenditures...................................................................... 3.5 3.6 58 3.7 Higher education financial aid programs benefitthe poor but leakages are significant .....63 3.8 Infant and under-five mortality inMongolia and internally............................................... School variable cost and student achievement by locations ............................................... 67 4.1 74 4.2 74 Immunization rates -a success story.................................................................................. Socio-economic and geographic disparities in mortality.................................................... 4.3 . . 75 4.4 Maternal mortality and attended delivery ........................................................................... 75 4.5 Key risk factors inMongolia and internationally............................................................... 76 4.6 Out-of-pocket spending by households inMongolia.......................................................... 77 4.7 Total health expenditure and GDP...................................................................................... 79 4.8 Public spending on health ................................................................................................... 79 4.9 M O H budget by institutional classification........................................................................ 82 iii 4.10 83 4.11 Coverage of SHI................................................................................................................. Government healthexpenditureat aimag level.................................................................. 84 4.12 Hospitalbednumbershighby internationalcomparison ................................................... SHIcontributionsandbenefits inMongolia....................................................................... 87 4.13 4.14 Healthworkforcehigh as well............................................................................................ 91 92 4.15 Humanresourcesat aimaglevel......................................................................................... 92 4.16 Use of inpatientservices in selectedcountries ................................................................... 93 4.17 Use of impatientservices at aimaglevel ............................................................................ 93 5.1 A dramatic increasein socialassistanceparticularlysince 2005...................................... 103 5.2 Whichhastranslated into large shares of the budget and GDP ........................................ 103 5.3 There is roomfor improvingthe CMP formulato includethe poor................................. 106 5.4 A large leakageof CMP benefitsto non-poor households............................................... A large leakage of CMP benefitsto non-eligiblehouseholds........................................... 108 5.5 108 5.6 There is only roomfor improvement in highschool where CMP beneficiariesseem to have done better ................................................................................................................ 110 5.7 CMP hasa large impact on highschool enrollmentamongthe poor ............................... 110 5.8 CMP accountsfor a large share ofthe budget and GDP .................................................. 110 5.9 Pensionshavebeenthe mainsourceofwelfarepayments but increasinglyless so., ....... 111 5.10,,.as CMP is introduced and later becomesthe main source of welfare payments ............111 5.11 Coverageand pensions systemdependency ratios ........................................................... 117 5.12 Households income composition...................................................................................... Pensioncontributors,beneficiaries and economicallyactive population......................... 118 5.13 119 5.14 Compositionoftotal governmentrevenues...................................................................... 120 5.15 Social Insurancecontributionsandbenefits ..................................................................... 121 5.16 Life expectancyandnormalretirement age 123 6.1 AgricultureGDP growth., ................................................................................................. ...................................................................... 131 6.2 Spendingfor agriculture:foreignassistancecontributes a significantamount ................133 6.3 Internationalcomparisonsof governmentexpenditures for agricultureinselected countries ............................................................................................................................ 133 6.4 Expenditureofthe Ministryof Foodand Agriculture from 2000 to 2007 ....................... 134 TABLES 1 xvm...x 5 38 There are importantspatial and incomedisparitiesin completionrates............................. Regionalpublicfinance managementsystem component comparison .............................. Publicminingrevenuesmostlycome from Erdernet............................................................ Key challenges. futureprioritieswith suggestedsteps for Implementation.................... RegionalPublicFinancemanagementSystemComponent Comparison............................. 2 1.1 2.1 3.1 52 3.2 Sources of fundingofhigher education institutions........................................................... 57 3.3 Current donor commitments for Educationsector by level ................................................ 57 3.4 Householdspendingon publicprimaryand secondaryeducation...................................... 59 3.5 Small, ruralschools havehigherlevels of per-studentspending ........................................ 61 3.6 State trainingfund recipients by program area ................................................................... 64 3.7 Proportionof college costs covered by tuitiodfinancial aid .............................................. 64 4.1 Ten lendingcausesof mortality.......................................................................................... 76 4.2 Government expenditureon healthat a glance................................................................... HealthexpendituresandfinancinginMongolia................................................................. 78 4.3 80 4.4 80 Line itemcompositionfor differenthealth sector institutions............................................ Government healthexpenditureby line item...................................................................... Sources andtrends in governmenthealthfinancing........................................................... 4.5 81 4.6 81 4.7 Allocationsto differenttypes ofhealthcare facilities ........................................................ 83 iv 4.8 Financingarrangement before and after reform ................................................................. 85 4.9 Beds and facility infrastructure inMongolia 90 4.10 Utilization patterns for different facilities .......................................................................... ...................................................................... 91 5.1 Social protectionbenefit expenditures.............................................................................. 101 5.2 Social insurance contribution rates ................................................................................... 116 5.3 Social insurance contributors, beneficiaries, economically active population .................117 5.4 Beneficiaries of social assistance ...................................................................................... 118 5.5 Benefits provided by social assistance .............................................................................. 119 5.6 Average wage, minimumwage and minimumliving standard ........................................ 121 5.7 Minimumliving standard ................................................................................................. 126 6.1 Expenditures for agriculture - 2000 to 2007 .................................................................... 135 V EXECUTIVESUMMARY KEY FACTS MONGOLIA AT A CROSSROAD - 1. Mongolia's external economic outlook is dramatically changing as it faces sharp reductions in the copper price, caused by the financial crisis and global downturn. This compels the government now to drastically cut spending to prudently manage the budget. The budget is extremely dependent on mining revenues. Government is taking the right step in proposing a balanced budget for 2009. But further adjustments will be needed given the continuing fall in copper prices. A prudent fiscal stance will also be needed to manage inflation, which accelerated in the pastyear to over 30 percent (August 2008). 2. The current situationhighlights the need to manage mining revenues better than in recent years. Mongolia saved little during the boom years, but instead dramatically increased expenditures on wages and salaries, and poorly-targetedsocial transfers. Adopting a multi-year fiscal framework-which enforces saving during the boom years, sets limits to expenditure growth and debt, and ensures transparency to the public-can help. Since much of the past windfall revenues have been spent (with Parliament in the driver's seat), the country enters the down-turnwith little savings and high inflation, forcing it to cut expenditures with every drop in the copper price. To avoid such situations in the future, the government has the opportunity to adopt a transparent, multi-yearbudget framework for expenditures and investment.This includes adoptinga new Fiscal ResponsibilityLaw. It would ensure that the government saves during the "boom" years, so that it can continue to spend during the "bust" years. It would also set limits to expenditure growth and public debt. Within the limits set by this framework, Parliamentcan then exercise its constitutionalrightsto amendthe budget. 3. Until recently, Mongolia had seen the most rapid increase in public revenues in decades. Public revenues hade more than doubled in real terms in five years reaching $1.5 billion or 40 percent of GDP in 2007. All this happened in the midst of a rapidly growing economy (9 percent over the last 4 years, on average), driven by a boom in mining and high world mineral prices. This mining boom boosted Mongolia's public finances and fundamentally changed the environment in which fiscal decisions were made. Comparedto previous decade of sluggish growth, budget deficits, and payment arrears, Mongolia was now experiencing a dramatic economic expansion, higherrevenues, lower public debt, and fiscal surpluses. 4. Fortunately,financial managementcontrols had beenstrengthened just prior to the boom in public revenues. The promulgationof the Public Sector Management and Finance Law (PSMFL) in 2002 completely changedthe legal and institutionalenvironmentfor public finances, in particular by recentralizing revenue and expenditure management and strengthening budget predictability of budget execution has greatly improved(see Chapter 2 - Public Sector Financial execution. As a result, the potential for leakages has been reduced to a large extent, and Management). 5. The outlook for Mongolia's medium-term public finances is favorable. High mineral prices have stimulated exploration. With the expected opening o f the new Oyu Tolgoi copper mine in 2011, economic growth is expected to remain at 7-9 percent per year until 2010 and to pick up at 12-14 percent in 2011-12. Even in the event of a decline in mineral prices, the vii Mongolia: Public Expenditure and Financial Management Review upcoming developments o f world class mines in Mongolia i s expected to boost the real economy. As a result, Mongolia's economic growth should remain sustained at high levels over the next decade. Box 1: Public finances in Mongolia: Key facts Since 2002, public revenues have increased almost fourfold (2.5 times in real terms), reaching 40 percent of GDP or more than $1.5 billion. Mining revenues now account for 36 percent of total public revenues (compared to 5 percent in 2002). Expenditures also increased at a fast pace, reachingalmost $1.5 billion in 2007, 3.2 times their 2002 level (2.1 times in real terms). Total government debt fell to 41.7 percentby the end of 2007 (27.6 percent of GDP in net present terms). Mongolia i s at low risk of debt distress, and the public debt burden is likely to continue to decline in the medium term, according to a recent Debt Sustainability Analysis conductedjointly by the World Bank and the IMF. Central government collects most public revenues and executes 91 percent of total expenditures. It i s also responsible for most public service delivery. Revenue and expenditure management was centralized in2003, following the implementation ofthe PSMFL. Social transfersnow account for 29 percent of expenditures,while the public wage bill accounts for another 17 percent at end-2007, as a result of successive wage increasesinrecent years. Public investment increased by 6.6 times between 2002 and 2007, contributing to 31 percent of total public spending in2007, while at the same time budgetary allocations for capital maintenance has declinedto 4 percent oftotal capital spendingin2007. Education and health spending shares of total expenditures were respectively 11 and 6 percent at end 2007, showing a decline since 2005 in relative terms. In recent years, the additional revenue earned from the mining boom has mostly been used for increasingspending in infrastructure, social welfare and public wages across-the-board Mongolia's benefits now from an internally consistent legal framework, a medium term and annual budget planning framework, and an effective countrywide government integrated financial management information system(GFMIS) for budget execution, monitoring and reporting. 6. The extent to which mining revenues translate into developmental outcomes will depend on how the medium-term fiscal space is managed and spent. Fiscal policy in resource-rich economies typically needs to respond to a combination o f objectives o f macroeconomic stability, expenditure smoothing, accumulating human and physical capital and long-term international competitiveness in tradables. Unfortunately, Mongolia seems to be falling prey to the same temptations that have seduced other resource-rich countries. Fiscal policy has turned pro-cyclical in recent years, led by a significant expansion o f public outlays (in real terms, an increase o f 88 percent in three years 2005-2007). Incremental public expenditures have been directed to universal social transfers, public wage increases and new public investments with unclear prioritization. In 2007 and 2008, the change in fiscal stance led to a large fiscal impulse that contributed significantly to inflation (15.1 percent by end-December 2007, and up to 34 percent year over year in 2008, see Chapter 1 on fiscal trends). WHAT THIS REPORTDOES 7. This Public Expenditure and Financial Management Review assesses the achievements and remaining challenges for public expendituresand financial management viii Executive Summary in Mongolia as the government responds to the needs of a rapidly changing economy. It provides factual information and analysis on key service delivery sectors. It also reviews the effectiveness of implementation of the 2002 Public Sector Management and Finance Law (PSMFL), draws on analysis conductedjointly with the Government and civil society in 2007, and lays out options for deepening reforms. The Review is selective -- some issues, such as human resource management, tax policy reforms and public infrastructure,are not addressed in detail 8. In developing practical policy recommendations, the report reflects on Mongolia's experiences during the last five years of economic expansion by examining policy responses, identifying institutional shortcomings, and reviewing key challenges for public expenditure policy. A review of progress and achievements between 2002 and 2008 is presentedin Annex 8. It uses the recommendations of the 2002 World Bank PEFMR to track progress in policy and reforms over that period. Table 2 at the end of this executive summary presents an assessment looking forward of policymakers' challenges, future priorities, and proposes concrete ways to addressthem. ACHIEVEMENTS -CORE PUBLIC FINANCIAL MANAGEMENT SYSTEMS IN PLACE AND FISCAL DISCIPLINE 9. Mongolia has successfully implemented the "first generation" of core public financial management reforms. Since the beginning of the decade, the authorities have stabilizedthe economy following the financial turmoil of the 1990s (Chapter 1 - FiscalTrends). Mongolia has successfully introduced an internally consistent legal framework, established medium term and annual budget planning frameworks, and implemented an effective countrywide government integrated financial management information system (GFMIS) for budget execution, monitoringand reporting- See Table 1 (Chapter 2 - Public Sector Financial Management). This new legal, institutionaland systemic framework has helped control leakages in public finances as well as: (i)improved budget execution significantly, leading to improvement in public spending efficiency in service delivery (see Chapter 3 and 4 for an illustrationof these progress in the Educationand Health sectors); (ii)largely eliminated arrears accumulation on goods, services, and salaries; (iii)achieved fiscal discipline; (iv) clarified roles and responsibilitiesof budget entities; and (vi) increased PFM transparency, with better quality information and timely production of annual budget reports and monthly budget execution reports. The PSMFL also mandates an ambitious performance management framework and accrual accounting; but these measures have not yet been fully implemented, and should be seen more as goals than immediatepriorities. ix Mongolia: Public Expenditure and Financial Management Review Table 1:RegionalPublic FinanceManagementSystem Component Comparison OperationalTreasury Single Account System Yes No No No No IntegratedChart of Accounts Consistentwith InternationalAccountingStandards Yes No Yes No No IntegratedBudgetandTreasury SystemLinkedto the CentralBank Yes No No No No UpdatedDebtManagement System ProvidingInput to the IntegratedTreasury andBudget System Yes ... Yes Yes ... Treasury SystemLinkedwiththe RevenueCollection Agency@) Yes No No No No CHALLENGES MANAGINGMININGREVENUESASAN OPPORTUNITYTO BUILD - FOUNDATIONSFORSUSTAINEDAND INCLUSIVE GROWTH 10. An immediate challenge for policymakers is to manage the commodity booms and busts and to build the foundations for sustained growth and poverty reduction The government's ability to transformthe non-renewablerevenue coming from the mining sector into a sustainable stream of economic resources while preservingmacroeconomic stability will be key to determining Mongolia's medium- to long-term development path. On the spending side, previouspolicy slippages such as large increases in socialtransfers and public investment without clear medium-term prioritization have left the country in weaker position to weather the global downturn. The Government commitment(articulatedin the Medium-termBudget Framework)to limit the public deficit to three to four percent of GDP has become obsolete in the presence of budget surpluses resultingfrom the boom in mineralrevenues. 11. From a macroeconomic perspective, fiscal policy should aim at smoothing the impact of large fiscal revenue inflows on the economy. A first step is to define how much revenue can be spent (as opposed to saved) to preserve macro-stability and sustainability of Mongolia's public finances, and avoid excessive real appreciationofthe exchange rate. A second step is to define medium-termpriority expenditures. To this end, developinga clear, transparent, and comprehensive public investment program with clear priorities and selectioncriteria is key (see Chapters 1 and 2). In doing so, official concessional assistance (ODA) will need to be better integrated into domestic priorities, as it accounts in sectors for a significant share of expenditures and oftenfocuses on new investments. The recent enhanced donor coordinationineducation is an example for other portfolio ministries(see Chapter 3 on Educationand Chapter 6 on Agriculture). 12. In the medium-run, the most significant challenge for Mongolia is moving on to the second generation of reforms: making full use of the core PFM institutions and systems to improve the quality of service delivery. Mongolia's share of public spending goingto education and health compares well to other countries at similar income level. Indeed, Mongolia's main challenge is ensuring that existing resources allocated to service delivery are spent more effectively, rather than allocating additional resources to these services. This is especially important given the increase in migration from rural to urban areas. These migration flows contribute to greater regional disparities by reducing affordability in rural areas and over- burdening the delivery of education, healthcare, and access to basic infrastructure services in Ulaanbaatar. To this end, building consensus on societal choices in social protection (and in X Executive Summary particular social transfers, social insurance) in health (health services and health insurance) and education services (in particular equity in access to education) will be essential to enable policymakersto commit credibly to reforms (see Chapters 3, 4 and 5). In smaller portfolios such as Agriculture, this will mean revisiting and clari@ing priorities to focus on areas where the government's action has the strongest comparative advantage (see Chapter 6). 13. While the new PFM legal and institutionalarrangements are a step forward in the management of public resources, they have also increased fiscal centralization where local realitiesare less well representedin the budget cycle. Fiscaldiscipline has come at the price of tightened centralized budget controls and planning processes, which in turn have limited the flexibility of budget execution at the local level. This review argues that the centralization of budget execution should be preserved. Yet a more decentralized budgetary planning process would benefit service delivery by better reflecting local needs. This is true for education and health services, where disparities have increased as the economy has become market-led and urbanized(See Chapters 3 and 4). This is also true for smaller ministries where centralizationof public expenditures has weakened local policy effectiveness to an even greater extent, as most of the spending is administered from the center (See Chapter 6). Mongolianow needs to incorporate sectoral perspectives, donor programs, budget entities and local government needs, and civil society demand for accountability andtransparency more effectively into budget planning. Inthis agenda, the first priority will be to improve sectoral planning and strengthen local accountability mechanisms. A coherent, integrated "strategy framework" will be central to an effective public investment program, and will allow the central government to articulate policy priorities in a transparent and accountable manner. 14. Specific thematic and sectoral findings and recommendationsare presented below. FISCAL TRENDS 15. With the global financial crisis unfolding, Mongolia finds itself with prospects of much reduced revenues from mining, on which the government budget heavily relies on, flagging the immediate need for a more prudent fiscal stance. Over the last decade, Mongolia has seen a significant reductionin its budget deficit and debt indicatorsthanks to macroeconomic stability, combinedwith a favorable external environmentand record-highworld mineral prices. Even so, the sharp expansion in expenditures during the recent commodity boom ran two important risks: first, resources may have been used in low priority projects and second, this policy did not account for the likelihood of a significant downturn in commodity prices. To prevent excessive spending and in the view of the large expected increases in revenues from mining, it will be important to delink public spending decisions each year from the revenues from the mining sector that year. To this end, the government should create a stabilizationand saving fund for all public revenues from mining with clear fiscal and governance rules. This fund should replacethe existingDevelopment Fund, and it should be accompaniedby governance rules which follow good international practice. The Government should also adopt transparent fiscal rules applyingto the mediumterm budget framework to smooth expenditures over time and controlthe non-miningdeficit. 16. Not only are expenditure excessive, but the nature of recent public expenditures has increased overall budget rigidities, thereby limiting fiscal space and the ability to run a counter-cyclical fiscal policy. Public wages and social transfers account for 17 and 29 percent respectively of the budget in 2007. These expenditures will be difficult to cut in the event of a commodity price correction. There has also been greater emphasis on capital spending across all sectors, especially infrastructure, which benefited from the creation of the development fund; maintenance investment has remained under-fundedand now represents a significant contingent xi Mongolia: Public Expenditure and Financial ManagementReview liability. Other challenges include unclear criteria for selecting investment projects, political rather than technical rationales dominating investment planning decisions, and low absorptive capacity for investments. In the near term, clarifying priorities for investment project appraisal and selection would increase efficiency and make public investment more transparent for stakeholders. There is also significant need to ensure that political considerations do not overrule the economic merits o f investment projects. PUBLIC SECTOR FINANCIAL MANAGEMENT 17. Implementation of the PSMFL has been a major positive milestone for Mongolia's publicfinances. It is a robust and internally consistent legislative framework, and contains all the elements o f a good public finance management system. 18. Budget planning rests on a clearly-definedlegal framework, including the medium- term fiscal framework (MTFF).Yet the process could benefit from fuller integration o f sectoral priorities, ODA, and local preferences. In addition, the investment program does not include foreign-financed projects, nor does it provide a rationale for investment selectivity. This weakens the link between policy priorities and actual expenditures at the sectoral and local levels. Indeed, making the Public Investment Program (PIP) the sole entry point for capital projects would better align high level strategies and strengthen the links among sectoral strategies, ODA, PIP and the MTFF. Integrating ODA into country systems and moving toward more sector-wide approaches would also lead to greater aid effectiveness, in line with the Paris Declaration. 19. Local governments have a limited role in service delivery. The drop in local expenditures in 2003 reflected implementation o f the PSMFL and the associated centralization o f public revenue and expenditure assignments. While the centralized budget execution framework should be maintained, strengthening accountability at the horizontal level-between local government and service delivery units-would ensure that the budget reflects local needs and realities. 20. Mongolia has made significant headway in improvingthe budget execution system. Since the PSMFL was enacted, a functional Treasury Single Account has been established, a unified and consistent chart o f account was implemented, integrated debt management and a government financial management information system (GFMIS) was deployed, a country-wide electronic payment settlement system was put in place at the Bank o f Mongolia, and a performance contracting framework has been operationalized across the government. Key internal controls surrounding the GFMIS are adequate, but there is still room for improvement. Additionally, these key internal controls should be reviewed periodically (at least once per year) to ensure that they are functioning as designed - details are presented in Chapter 2 and Annex 2B. GFMIS controls around the budget, system integrity, payment security, and system management should be improved upon and all key internal controls should be regularly reviewed. There is also a need to review the efficacy o f performance contracts at the department level and the use o f strategic business plans at the budget entity level. 21. The 2006 Procurement Law was a commendable improvement over previous legislation, but recent amendments represent a significant step away from good practice. The amendments, adopted inFebruary 2007, significantly altered the spirit o fthe law by allowing direct contracting and restricting international competition by significantly increasing the financial threshold for international competitive bids. The Government should consider withdrawing the amendments. In addition, better integration o f public procurement activities with the Government financial management system would help foster procurement efficiency and transparency. xii Executive Summaw 22. The capacity of the Mongolia National Audit Office should be strengthened to conduct robust financial audits. To do so, the overall legislative framework provides an adequate basis for the agency in auditingthe Government. 23. The PSMFL remains silent on disclosure of budget information.In 2006 and 2007, the Government made significant progress in disseminating some key budget documents and information, but outside any legal framework. Amendments with timelines for public release of budget informationcould be proposed for consideration in Parliament.Also, for transparency and communication purposes, the Government could institutionalize the presentation of MBTF policies and proposed State Budgets to media and civil society organizations around the same time that the policiesare presentedto Parliament. EDUCATION 24. Mongolia has achieved high enrollmentand completion rates in education, but there are significant disparities in student performance by area and income levels. Students in rural areas and from poorer families are less likely to enroll and complete education, particularly at upper secondary and higher levels. 25. Mongolia allocates 19 percent of public expenditures to education. This level is similar to or higher than that in other developing countries and OECD countries. However, Mongolia's overall cost for service delivery is higher than in most other countries, presumably because of its large territory andthe relativelysmall population. 26. Donor contribution to the education sector is significant has greatly improved. Domestically-fundedcapitalexpenditures are low, buttend to be offset by donor financing, which i s focused on new investments. Donor coordination has improved significantly with the establishment of the Education Donors Consultative Mechanism (EDCM) and through the development and donor endorsement of Mongolia's Education Sector Master Plan (2006-2015). The budget gap identified in the Master Plan is being addressed by the Fast Track Initiative's Catalytic Fund, which will provide grant resources of $29.4 million to Mongolia's education sector over the period 2007-2009. Other portfolio ministries where donor coordination is a challenge can learn from this successful experience of external partnership coordination and planning. 27. The 2002 PSMFL has greatly improvedthe predictabilityof schoolfinancingand of fiscal discipline at the school level, yet school staff perceived the budget situation as more rigid than before the law was implemented. Teachers are paid on a timely schedule, whereas before arrears were common, but schools are limited in their use of savings and performance- based contracts fail to evaluate teachers on measurable outcomes. Measures that increase budget execution flexibility at the budget entity level such as carrying over savings are difficult to implementinpractice.To addressthe perceived lack of flexibility in budget execution, improving budget planningat the local levelis recommended. 28. There is a needto refineboth fixed and variablefund allocationschemes.Allocations for fixed costs such as heating are determined on an incremental basis and do not encourage savings, while differences in infrastructure make it difficult to estimate needs. To reduce wastages associated with fixed costs, the government could explore several potentialsolutions:(i) increasing maintenance funds to reduce energy losses; (ii)renovating buildings; (iii)installing more fuel-efficient heating systems or alternative technologies; and (iv) installing meters for utilities. Aside from fixed utility costs, the government allocates resources to schools based on a simple andtransparent per-student funding formula. Micro-coefficients allocate more resourcesto small schools than the per-capita funding formula, but this creates inequities across similar xiii Mongolia: Public Expenditure and Financial ManagementReview schools located in different aimags. Lack of budget planning skills among local authoritiesand school management limits the efficiency gains from the per-capitafunding formula. Providing trainingto local staff, parents and councils for budget planningand financial management could strengthen localaccountability. 29. The delivery of education services is challenged by rural-to-urban migration, which threatens the survival of small, isolated, rural schools and over-burdens schools in Ulaanbaatar. Funding allocations for schools need to reflect these changes. To address high rural-urban disparities and promote equity, the government should support policies which encourage multi- grade and multi-subjectteaching in rural schools, and could use national-levelmicro-coefficients to improvethe current formula. 30. Financial-aid programs in higher education such as the State training Fund loan program benefit the poor and have a positive impact on student performance, but the coverage is limited and there are significant leakagesto the non-poor. To improve demand-side financing programs, the government could better target the poor with mean-testing and build a stronger system of monitoringand impact evaluation. HEALTH 3 1. Infant and child mortality are low in Mongolia compared to other countries at similar levels of economic development, but these rates are higher among poorer families. By contrast, adult mortality is on the rise, due largely to the growing impact of non-communicable diseases such as heart disease, stroke, cancers, hypertension, diabetes, and chronic lungdisease. 32. Total and government health financing appear high by international standards. Budget financing remains predominant, and the HealthFinancingFund accounts for slightly less than a quarter of total financing. Out-of-pocket payments have been increasing since 1998, reachinga third oftotal financingin2004. Expenditureslargely go toward salaries, secondaryand tertiary providers, and soum and inter-soum hospitals receive a large share of the recurrent budget. Geographic expenditure variations are significant. In order to refine resource allocation, there is a needto explore the reasons for these observed geographic disparities in spending and the determinants of unit costs indifferentcontexts. 33. There is considerable scope for improving efficiency while improving health outcomes and equity. Since the beginningof the decade, efforts to rationalize the healthsector have included rationalizingsoum hospitals and creating Family Group Practices. The hospital rationalizationreformmomentumcouldbenefitfrom an inclusiveprocess of consultationto build reformbuy-in.Improvementsin efficiencyrequire a rationalizationof facility infrastructure,a re- structuring of the health workforce, and the creation of systems and incentives for reducing unnecessary inpatientadmissions. 34. I t is not clear whether the Social Health Insurance (SHI) Scheme has succeeded in its goals of generating additional funding. The design of the SHI also raises questions about equity in access to services. The SHI system has not contributed in any significant way to improvements in service delivery. Policymakers therefore might consider unifying government healthexpenditures under a single agency. The government should at least clarify the role of the SHI. Phasing out budget subsidies to the SHI and reducingthe extent of cross-subsidization by members making payroll contributions would be an important step in this direction. If, on the other hand, the SHI scheme is seen as an instrument for achieving universal coverage, finding ways for convertingthe SHI system into the primarypurchaser of health services would then be necessary. 35. The 2006 reforms have introduced significant changes and have the potential to xiv Executive Summary considerably improve primary and hospital care. The move away from line-item budgeting and fee-for-servicepayment towards case-based payment, introduced in the 2006 reforms, is a step in the right direction. This change will help improve hospital efficiency, but also poses significant challenges to current financial management practices. The introductionof case-based payment in hospitals is rarely without its problems. Capitation payments financed out of state budgets have great potential to improve primary care services, but implementation requires carefulattention.For instance, capitationpayment rates should be set sufficiently highto attract a qualified workforce, while strong quality monitoring systems should be put in place to avoid under-servicing. The appropriateness of referrals should be also assessed, and gate-keeping arrangements enforced. Also, combining capitation payments with fee-for-service payments or performancebonuses should be considered in order to incentivizethe delivery of priority public health interventions.The capacity within the Ministry of Health, Ministry of Social Welfare and Labor, and Civil Service Council to implement recent provider payment reforms should be strengthened by providing training. This is essential in order to manage risks associated with case-based payments such as those related to quality or volume of services. Finally, it will therefore be importantto monitor the consequences of the recent reforms on the use of services and the financial burden for different groups, and to promote debate over whether the outcomes are compatible with the prioritiesof the population. SOCIAL ASSISTANCE 36. The 2003 Social Security Sector Master Plan (SSMP) aimed at consolidatingand rationalizing the system and improving cost-efficiency. The Child Money Program (CMP) responded to these objectives and consolidated all child-relatedallowance programs. The CMP has had little impact on poverty, but has been successful in conditioning benefits on positive social behavior.For example, the program had a large impact on high school enrollment among the poor. The move to a proxy-meanstest was an important improvement. However, the impact among the poor came at high cost because of mis-targetingto the non-poor. The CMP was universalizedin July 2006 and the benefit increasedto 136,000 MNT per child as of January 2008. CMP income is estimatedto account for 34 and 20 percent of total household consumption among the poorest and second poorest quintiles. The large size of the program raised concerns about its fiscal sustainabilityand pro-cyclicalnature. 37. In order to maximize social impact with the lowest cost, the CMP should return to cost-effectively targeting the poor and condition benefits on human capital investments. Preparation work should focus on refining the transparency and credibility of the targeting criteria, ex-ante analysis on the benefit rate, and implementation arrangements. An implementation strategy should accompany the SSMP for consolidating and rationalizing the system. The first step is to evaluate the main social risks faced by different groups and identify who is most at risk. The second step is to identify the most cost-effectivepolicy instruments to address those risks. For instance, besides a direct social transfer scheme, policies to enhance productiveemployment opportunitiesmay be neededto addressjoblessness. SOCIAL INSURANCE 38. Examining contributions and benefits together, the overall difference between contributions and benefits has grown from less than one percent of GDP in 2001 to almost three percentof GDP in 2007. The adoption of a NotionalDefinedContribution(NDC) scheme in 1999 improvedthe inter-generationalsustainability, equity and predictability of the Pension Insurance Fund. But it did not accommodatethe special needs of herders, those in the informal sector, and those formal sector workers with insufficient lifetime contributions to support a xv Mongolia: Public Expenditure and Financial Management Review pension or who otherwise fall into poverty during retirement.Problems include:(i)the minimum pension framework is too generous and creates weak incentives; (ii)the retirementage is too low to ensure either financial sustainability or a meaningfulreplacement rate; (iii)actual indexation has often resultedin a depreciationof the real benefit value; (iv) disability benefits are generous enough that they pose a fiscal burden and weak incentives under the defined-benefit scheme and, inthe case ofNDC, are inappropriatelyfinanced. 39. The societal objectives of the combined public and private pension and social security system should be reviewed in order to ensure that the system provides: (i) a minimum living subsistence to the elderly, permanently disabled and survivors/dependents; (ii)mandatory consumption smoothingfor most formal sector workers and (iii) voluntary contractualsavings for informalsector workers. While the basic design features of the current system can be retained, the retirement age, benefit indexation, minimum pension provisions, wage base for determining benefits, formulation and financing of the disability benefit, and select other provisions should be revisedin order to strengthen fiscal sustainability,affordability, predictability and equity. AGRICULTURE 40. The agricultural (livestock) sector is still the largest employer in the economy, despite its declining share of GDP. The sector is prone to exogenous shocks which can have a dramatic impacton the economy and population. 41. The government's objectives for agriculture are somewhat disconnected from actual policies and programs. Policy objective statements are consistent with a market-oriented economy. Inpractice, however, public expendituresgo toward direct interventionsinthe market, instead of toward correcting market failures and providing public goods. For instance, the Government could play a role in limiting risk and uncertainty for herders by improvingaccess to informationand providingtraining, education, and extension services. 42. Public spending on agricultureas a percentage of GDP is relativelylow compared to other countries, and donor funding contributes a significant share. Yet there is little mechanism for coordinatingdonor funding with government policy priorities.Programsappear to be under-costed and spending is biased toward recurrent expenditures. Central-local accountability is weakened by information asymmetry, lack of incentives at the local level to implementcentrally planned programs in a cost-efficientway, and low monitoringand evaluation capacity at the central level. Budget planning could be strengthened by: (i)including donor financing ingovernment priorities; (ii)partly decentralizingbudget planningto reflect localneeds and preferences. In addition, policy analysis capacity needs to be improved to allow for better monitoringand evaluationof programs. A UNIQUEOPPORTUNITY 43. Mongolia's public finances face today a unique opportunity to address medium- term development challenges by allocating large increases in public revenue to service delivery improvements (the so-called "second generation" of public financial management reforms). To this end, the government benefitsfrom strong legal, institutionaland systemic PFM foundations (the so called "first generation" of reforms). Improving service delivery will require (i) a strong public commitment to macroeconomic stability and sustainability that avoids pro- cyclical spending; (ii)a system to ensure clear and credible medium-term revenue allocation priorities for public spending; (iii)the policymakers' ability to adapt public services to the country's new structural characteristics, which include: (i)a completedeconomic transition to a market economy with a change in the government's role from a planner to a regulator and xvi Executive Summaw facilitator of private sector growth; (ii)a rural- urban transition where migration is a source o f spatial inefficiencies and inequities for access to public services; and (iii)a sectoral transition with an increasing reliance ofthe economy on the miningsector, with associated large amount o f public revenues. 44. Table 2 lays out key challenges for policymakers, future priorities, and suggests concrete ways to address them. xvii s"b 0 .-C0 .-2 c1 w .-Xx a 8 5 e, I 1 . 9 CHAPTER 1. FISCAL TRENDS PROGRESSAND CHALLENGES 1 Mongolia's external economic outlook has recently dramatically changed as it faces sharp reductions inthe copperprice, causedby the global downturn. This will compelthe governmentto drastically cut spendingto prudently managethe budget. 1 Over the last decade, Mongolia has seen a significant reduction in its budget deficit and debt indicators. This has occurred because of improvements in budget execution and a commitment to macroeconomic stability, in combination with a favorable external environment with record-high worldmineralprices. Public revenues have increasedat a fast pace thanks to sustained economic growth and record high mineralprices in recent years, and will most likely continue increasing in the medium term as new minescome onstream. . 1 Public expenditureshave also increasedrapidly, thereby limiting fiscal space. Public wages and un- targetedsocialtransfers account for an increasingshare ofthe budget. There has been an increased emphasis on capital spending across all sectors but maintenance investment is under-funded. This is especially true in the Ulaanbaatar area, where the population is growing rapidly and where existing infkastructure need attention. Other problems include unclear ..criteria for selecting investment projects and the fact that over the past year the budgeted capital expenditureswere under-spentdue to absorptioncapacity constraints. Parliamentarydiscretionover the annualbudgetis significant. The drop inlocalexpenditures in2003 reflectedimplementationofthe Public Sector Managementand FinanceLaw (PSMFL) fkom that year onwardsandthe associatedcentralization ofpublicrevenuesand expenditures assignments. Local governments have had a limitedrole in service delivery, in line with their revenueandexpenditureassignments. KEYFUTURE PRIORITIES i. IntheviewofthelargeboomsandbustsinexpectedrevenuesthatMongoliawillseeinthecoming years from mining, set-upa stabilizationand saving fund for all miningrevenues and adopt transparent fiscal rules as part of the medium term budget fkamework to smooth expenditures over time, mitigate real appreciation, andmanageexpenditures inline with absorptivecapacity. > Suggested"How to": Indoing so, the Government may consider3 priorities: a) Delink public spending decisions each year fkom the revenuesfrom the mining sector that year. This can be done by transferring all public revenues from mining to a stabilization and saving fund. This find should replacethe existing Development Fund, and it should be accompaniedby governance rules which follow good internationalpractice(describedinthis chapter); b) Define how much should be transferred each year from the Stabilization Fund to the General Budget. While there might be more than one rule that fits Mongolia (see Box 1.2), it is recommendedthat the rule is defined so that: (i) smoothes available resources over time for sustainablespendingand intergenerationalequity; isolatespublic spendingand the economy fkom copper price volatility; (ii) line with the economy's absorptive capacity and with a moderate in 1 Mongolia: Public Expenditure and Financial Management Review real exchange appreciation. c) Adopt fiscal rules which apply to the Medium Term Budget Framework and General Budget to keep the non-mineral deficit under control (see chapter for good principles to be followed and fiscal rule options). ii. ReevaluatetheextentofparliamentarydiscretionovertheannualBudget. Suggested"How to": Review good international practice, conduct consultations, and draft legal provisions according to internationalpractice and Mongolia's institutional setting. 2 Chapter I -Fiscal Trends INTRODUCTION 1.1 Mongolia's public finances have changed quite significantly since the beginningof the decade, following two major events. The first is the enactment of the new Public Sector Management and Finance Law (PSMFL), which introduceda profoundchange in the way public monies were managed, includingrigorous budget execution, recentralizationof public resources, and medium-termbudgetingamong others. The second is the boom and bust in mineralprices in world markets, which, from 2005 onwards, led to a rapid increase in public revenues and an expansion of public spending, followed in September 2008 by a significant decline in metal prices which promptsthe needto adopt more prudentfiscal policy. OVERALLFISCAL TRENDS 1.2 Over the past decade, Mongolia has seen a significant reduction in the size of its budget deficit.In 2000, the deficit was 6.1 percent of GDP; by 2006, there was a surplus of 8.1 percent. 2007 preliminary numbers suggest a surplus of 2.2 percent of GDP (Figure 1.1). This trend represents a greater commitment to macroeconomic stability as well as improved macroeconomic conditions,relative to the late 1990s period, which was associated with double- digit leveldeficits,weak budgetperformanceand revenue shortfalls. Figure1.1: Expenditureshave increased rapidly, but revenueseven faster, leadingto a reductionin publicdeficits 50% 120% 40% 100% 30% 80% 20% 60% 10% 40% 0% 20% -10% 0% Source: MOF 1.3 Internal factors, such as improvements in budget execution, have played an important role in reducing the public deficit. Some fiscal consolidation has occurred, as a result of a sudden increase in revenue, which in turn was generated by high economic growth, improvements in budget execution and tax administration,and a general policy commitment to macroeconomic stability. In addition, revenue forecasts have been consistently pessimistic- unnecessarily so. Revenue collection plans have over-performed, and improvements in tax administration have led to improved tax collection efficiency and higher revenues. On the expenditures side, the Government FinancialManagement InformationSystem (GFMIS) greatly contributedto better execution of public spending, better treasury management, and additional public savings. 3 Mongolia: Public Expenditure and Financial ManagementReview 1.4 Following the macroeconomic turmoil of the 1990s, the Government made a commitment to macroeconomic stability, but fiscal consolidation was not a result of a reductionin overall spending (Figure 1.2). Thanks to economic growth and increased revenues, it was possible for the Government to increase expenditures without compromising macroeconomic stability. Over the last decade, GDP has grown faster than the increase in expenditures; expenditures remained stable as a share of GDP, whereas in realterms expenditures have more than tripled since 2000. 2005 is an exception, when spending slowedsignificantly.The adoption in 2003 of a medium-term budget framework which set a binding deficit target for the following fiscal year helped support the commitment to macroeconomic stability. This deficit ceiling was set at approximately 3 percent of GDP and subsequently endorsed in the planned budget. Figure 1.2: Miningcontributes an increasingshare of public revenues thanks to the increase in copper and gold prices I"`" I Non-minina revenues Minina revenues I 40% 35% 30% 25% 20% 15% 10% 5% 0% -5`10 -10% -15% - -20% - -25% - Non-mineral revenue and grants (%GDP) -Mineral revenue (%GDP) i\\ Non-mining GDP growth share (YO) Mining GDP growth share (%) +Overall budget balance (%GDP) *Overall non-mining balance (%GDP) Source: MOF and NSO 1.5 Improvementsin exogenous factors such as terms of trade and weather conditions have also played a critical role. Since 2001, the Mongolian economy has seen sustained economic growth. Economic growth has been supported by (i)improvement in weather conditions, which allowed the agricultural economy to recover from the 2001-2002 consecutive dzuds; (ii)sustained FDI in the construction, mining and services sectors; and (iii)a continued improvement in the terms of trade, with a marked increase in world metal prices from 2005 onwards. 1.6 From 2005 on, record-high world mineral prices increased Mongolia's public revenues even further. The budget surpluses of the most recent years (2005 onwards) mostly reflect the significant improvement in terms of trade, resulting from the boom in world mineral prices.' Significant changes in fiscal policy were implemented in 2006, includingthe introduction of a large "windfall" tax on the copper and gold sectors and an increase in mining royalties from 2.5 percentto 5 percent. A development fund was also created to hold revenuesfrom the windfall ' Mongolia i s a resource-richeconomy which is dependent on copper and gold, which account as a whole for 69 percent o f exports, 27 percent of nominal GDP and 36 percent of government revenues in 2007. While the share of the budget coming from mining has significantly increased, the share of GDP in constant prices has remainedthe same, illustrating the limitedreal impact of the mining boom in the short term. 4 Chapter I -Fiscal Trends Billions of Tugriks 2001 2002 2003 2004 2005 2006 2007* 2008** CIT from Erdenet 11.3 9.5 29.8 52.9 61.3 130.4 75.1 75.6 Windfall Tax (95% from Erdemet) 0.0 0.0 0.0 0.0 0.0 177.5 353.7 363.3 Royalty 3.7 10.8 11.3 25.0 31.7 44.3 106.2 132.1 Dividends from Erdenet 1.6 2.5 0.1 1.9 18.1 37.0 136.6 105.5 Total 16.6 22.8 41.2 79.8 111.1 389.2 671.6 676.5 Total (%GDP) 1% 2% 2% 4% 4% 10% 15% 12% Total (YORevenues & grants) 4% 5% 7% 11% 13% 29% 36% 28% Source: MOF and IMF 1.7 With rapidly increasing Figure 1.3: Public spending is increasing rapidly in real terms driven by investment and wages, and social transfers revenue and a stable fiscal deficit target of 3 percent in the medium-term budget framework, significant fiscal W C a p I t a I expenditure and net +GDP Qrowth space remained for ox expansion of public outlays. With rapid 4m increase in minerals prices and revenues, 2002 2003 2004 the objective of I 1% maintaining a deficit at an unchanged level has resulted in a fast expansion o f expenditures. Expenditures between 2005 and 2007 increased by 88 percent at constant prices, with the expansion of social welfare, increased civil service wages, and large investments in infrastructure (Figure 1.3). 1.8 An analysis of the fiscal stance in Mongolia before and after the mining boom suggests that the fiscal position began improvingprior to the increase in mineralprices and that fiscal policy was significantlypro-cyclicalin 2007 and so is the planned fiscal policy in 2008. A Cyclical decomposition o f fiscal balance is used to measure the true fiscal stance. The actual budgetary position o f the government can be a misleading indicator o f the thrust o f fiscal policy, because o f the presence o f automatic stabilizers (e.g., unemployment and welfare benefits) which move in a countercyclical fashion even in the absence o f any discretionary action by the government. Determining the true fiscal stance requires purging the fiscal balance from the Starting in2006, official records of gold production and exports started declining, as more and more gold was allegedly smuggledout of the country. 5 Mongolia: Public Expenditure and Financial ManagementReview effects of these automatic stabilizers. This can be achieved by decomposing the primary balance into the cyclically adjusted balance (CAB) that would prevail if the economy was operating at its full potential, and the cyclical component that reflects the impact of the cycle on the fiscal balance. The CAB thus permits measurement of the contribution made by cycle and in particular the improvement o fterms and trade to the improvemento fthe fiscal position in Mongolia. 1.9 The cyclically adjusted balance (CAB) is defined as the difference between cyclically adjusted revenues and expenditures. Cyclically adjusted revenues and expenditures are estimated using the OECD approach. Cyclical factors are eliminated from revenues by estimating revenue levels when output is at its potential, assuming average elasticity for the period 1990-2007.3 1-10 The fiscal stance is then defined as the first difference in the CAB. The fiscal stance can be a usefulmeasure of the impact of fiscal policy on aggregatedemand and can be employed to determine whether the fiscal policy is procyclical or counter-cyclical. A positive correlation between the fiscal stance and the output gap in a particular period indicates that fiscal policy was procyclical duringthat period. 1.11 The analysis of the CAB indicates that the improvementof the underlying fiscal position predates the improvementin the terms of trade. Developments in cyclically-adjusted revenues and expenditures suggest that fiscal policy was tightened from 1999 (Figure 1.4). Cyclically-adjusted revenues increased, while spending remained contained. Since 2006-7, fiscal policy has turned significantly I procyclical, with a fiscal impulse as Figure 1.4: Fiscal policy is pro-cyclical large as 7.2 percent of GDP in 2007, 4000 1 - Output and PotentialOutput and more then 5 percent o f GDP - 1 planned in 2008. 1.12 Procyclicality of fiscal policy may have important 2500 implications for macroeconomic m stability and fiscal sustainability. A recent IMF study concludes that 15004 2000 2001 2002 2003 2004 2005 2006 2007 2W8 procyclical fiscal policy exacerbates -+-GDP economic fluctuations, with adverse -8- PotentialGDP consequences for savings, investment, Fiscal stance and Fiscal inpulse economic growth, and welfare. It also ---_____ suggests that a stronger procyclical 8.0 bias in the upturn may result in a 6.0 4.0 deteriorating fiscal position over time, B 2.0 since deficits and debt built up during 0 0.0 bad times are in general not offset -2.0 duringgood times. -4.0 -6.0 1.13 Mongolia's external II t.OverallBalance-m- Fiscallnpulse CycliiaityAdjusted Balance economic outlook has recently 3The formula usedto calculate the CAB in year t is: CABt=Rt(Ypt/Yt)a -Et, where R is actual revenue; Yp is potential output; Y i s actual output; a i s the average elasticity of revenue with respect to YpN for the period 1990-2004; and E is the actual expenditure. Potentialoutput is calculatedby de-trendingtime-series data of real GDP, using the Hodrick- Prescott (HP) filter. It is also assumed that cyclically adjusted expenditures are equal to actual expenditures 6 Chauter 1-Fiscal Trends sharp reductionsin the copper price from levels above US $8,000 per ton over the summer 2008 to below U S $4,000 per ton in early November 2008, caused by the global downturn. This will compel the government to drasticallycut spendingto prudentlymanagethe budget. 1.14 The global financial crisis has heightened the vulnerability of many developing countries to exports slowdown, investment flight, credit default, banking system failures, budgetary imbalances, balance of payments shortfalls, and has put the poor at greater risk.These vulnerabilities apply to Mongolia, too. Although Mongolia appears to have limited direct exposure to foreign financial markets, the global financial turmoil and weaker global economy will affect Mongolia through a sharp fall in copper prices, decline in foreign investments, withdrawal of foreign lines of credit, and potential for further delays in planned mining developments. 1.15 Given the budget dependency on mining revenues, the Government is taking the right step in proposinga balanced budget for 20094.But further adjustments will be needed given the continuingfall in copper prices.A prudent fiscal stance will also be needed to manage inflation, which accelerated in the past year to over 30 percent.The sources of this inflation have shifted gradually from importedto domestic. Wages and salaries have more than doubled in two years. The Central Bank has tightened monetary policy somewhat, but interest rates are well below inflation. As oil and food prices are now coming down, the focus needs to shift towards containing domestic consumption (public and private). To do this, the most effective policy instrumentin the currentpolicy settingis fiscal tightening, first, and monetarytightening, second. A more flexible exchange ratecouldhelp softenthe adjustment. 1.16 The current situation highlightsthe need to manage mining revenues better than in recent years. Mongolia saved little during the boom years, but instead dramaticallyincreased expenditures on wages and salaries, and poorly-targeted social transfers. Adopting a multi-year fiscal framework-which enforces saving during the boom years, sets limits to expenditure growth and debt, and ensures transparency to the public-can help. Since much of the recent windfall revenues have been spent (with Parliament in the driver's seat), the country enters the down-turnwith little savings and high inflation, forcing it to cut expenditures with every drop in the copper price. To avoid such situations in the future, the government has the opportunity to adopt a transparent, multi-year budget framework for expenditures and investment. This includes adoptinga new Fiscal ResponsibilityLaw. It would ensure that the government saves during the "boom" years, so that it can continueto spend during the "bust" years. It would also set limits to expenditure growth and public debt. Within the limits set by this framework, Parliamentcan then exercise its constitutionalrightsto amend the budget. 1.17 The government is looking for ways to redefine the framework in which fiscal policy is operating given Mongolia's new mineral wealth prospects. Indeed, the practice of setting the overall deficit at a constant level of 3 percent of GDP leaves spending-and capital spending in particular-as a residuals5Large swings in minerals prices and revenues could result in excessive fluctuations in capital spending, and complicate soundmacroeconomic management. The resulting volatility in capital spending could also be an obstacle to using resources efficiently, sustaining high growth momentum, and promoting non-mineral sector development. A variety of mechanisms are being explored, including adopting a fiscal responsibility frameworks with bindingfiscal rules, and also establishinga stabilizatingand saving fund, which couldfollow the Alaska fund modelby distributingdividendsto the population. 4 'InThefiscal 2008, miningrevenues constituted43 percent o fthe budget. deficit target was set at 3, 4 and 4 percent respectively in the 2008-2010, 2007-2009, 2006- 2008 Medium-TermBudgetFrameworks. 7 Mongolia: Public Expenditure and Financial ManagementReview 1.18 The government needs to choose the kind of fiscal rules which best suits the Mongolian context, where (i)exploitation of the new mineral wealth as not startedyet, (ii)there is an existingDevelopmentFundwhere the windfall tax i s transferred(see Box 1.3); (iii)Future trend in internationalcopper prices is uncertain; (iv) Pressure to spend is high as a result of large needs and competingrequests, in particular in infrastructure; (v) absorptive capacity is limited as highlightedby the recent overheating ofthe economy. 1.18.1 Thefirst priority is to delink public spending decisions each year fiom the revenuesfiom the mining sector that year. This can be done by transferringall public revenues from the mining sector (windfall tax, corporate tax, royalties, and dividends) to a stabilization and saving fund. The creation of such a fund will also insulate the economy from the adverse effects of large and volatile mining revenues. This fund should replace the existing Development Fund, and it should be accompanied by governanceruleswhich follow good internationalpractice(see above andBox 1.2). 1.18.2 Secondly, the government needs to define how much of the fund should be transferred eachyear to the General Budgetfor spendingpurposes. Box 1.2 describes various existing fiscal rules used in different country settings. There might be more than one rule that fits Mongolia (e.g. price-based such as in Chile, or total wealthmermanent income approach such as in Azerbaijan).6It is recommendedthat the rule is defined (i)so that it smoothes available resources over time for sustainable spending and intergenerationalequity; (ii) it isolates public spending and the economy from copper price volatility; (iii)in line with the absorptive capacity of the economy and a moderate real exchange appreciationto maintain long-term competitiveness of tradables. 1.18.3 Thirdly, Mongolia should adopt Jiscal rules to keep the non-mineral deficit under control. A variety of fiscal rules exist, and there is more than one rule which can be applied to the Mongolian context. But some good principles apply for the fiscal rules to successfully achieve fiscal discipline in a sustainable manner: 1.19 Looking at international experience, a few good principles may be followed when deciding on fiscal rules (see Box 1.1 and 1.2). First, any fiscal rule alone, without political will to enforce it, is doomed to failure. In other words, fiscal rules can be regarded as a formal expression of the political will to maintain fiscal discipline. Second, a strong legal basis and popular buy-in is necessary to build the fund credibility over time and to avoid the fund being raided by subsequent governments. To this end, simple and transparent rules should be adopted. Third, if a saving and stabilization fund is created for mining revenues only, it should be accompanied by fiscal discipline in controlling the non-miningdeficit, to avoid that the creation of the fund be accompanied by a build-up of cumulative fiscal deficit, which would result in lower net savings. Fourth, these fiscal rules should be enforced within a medium term budget programming and macro forecasting, which is a prerequisite for well-informed policymaking debate. Fifth, a mechanism of mid-course correction for unanticipated deviations from target, unless they stem from cyclical fluctuations covered by escape clauses or can be offset by a recourse to a contingency fund. Finally, compliance with fiscal policy rule and proceduralrules must be subject to continuous monitoring preferably from an independent and competent authority in additionto the oversightby the media and civil society (Kopits (2007)7). On Azerbaijan's experience with managing oil revenue, see Budina and van Widjnbergen, 2008, "Fiscal Sustainability and Oil Wealth: Managing oil and gas volatility in Azerbaijan"", mimeo, available upon request. Source: George Kopits (2007) "Fiscal Responsibility Framework: International Experience and Implicationsfor Hungary", MNBOccasional Papers, 62. 8 Chapter I -Fiscal Trends 1.20 In the recent years, Parliamentary discretion over the annual budget has been increasing significantly.The PSMFL is silent on any limits on legislative changes, and provides Parliament will full discretion on making changes to the budget. For example, Parliament was able to increase domestic investment by 21 percent (which corresponds to a significant 11 percent increase of total expenditures) when discussing the 2008 budget at end 2007. In doing so, the legislature added a myriad of small regional projects. Good practice suggests that legislative discretion over budgetary expenditures should be restrained. In Thailand for example, the Constitution stipulates that the Parliament must not increase the amount of any budget item. Parliament can only reduce allocated amounts. If they do so, the government may then propose an additional budget in the amount equal to what has been cut. Looking at the Thai example of god practice, policymakers in Mongolia could consider rebalancing power betweenthe Executive and the Legislative in the area of fiscal and budgetary management, by limiting parliamentary discretion over the Budget. Box 1.1: What fiscal rulesaround the world? A fiscal policy rule consists of a permanent constraint on a broad performance indicator, usually expressedinterms of stock (public debt) or flow aggregates (government balance, borrowing, expenditures, or some componentthereof). Policy rules are also knownas numericalrules ofien set inproportionof GDP. In general, the stock of public sector liabilities (or net worth) is seen as a key summary indicator of a country's vulnerability. Financial markets tend to assess default risk on the outstanding debt of the public sector as a whole, rather thanjust the central government, given the implicit guaranteeprovidedby the central government on the liabilities of the rest of the public sector. More generally, to maintain or restore fiscal sustainability, a number of countries have introduced policy rules, first to reduce public debt, and then to stabilize it at a prudent ratio to GDP. InNew Zealand and the United Kingdom, the government is requiredto set a medium-termtarget or ceiling for the debt ratio, as well as an adequate floor for public net worth. In addition, to avoid free-rider behavior, in Brazil, a target debt ratio is set at each level of government. For similar reason, in the EuropeanUnion, member governments are obliged to reduce the gross debt ratio to 60 percent of GDP. Convergence to the debt ratio ceiling usually requires complying, either implicitly or explicitly (Brazil), with a minimumprimary surplus as an operational target. A more common rule is defined in reference to a comprehensiveflow indicator offiscal performance, such as the budget balance. There is wide variety of budget balance rules: maintenance of overall balance, current balance, primary balance, or non-oil balance. Alternatively, a numerical limit is set on the overall deficit (European Union, Peru, India) or a floor for the overall surplus (Chile, Sweden). The current balance rule, also calledthe `golden rule' (Brazil, India, Venezuela), i s commonly usedto prevent crowdingout much- needed public investment. The actual target or numerical limit (or floor) is specified by the circumstances of the given country, includingthe needfor simplicity, transparency, and ease oftechnicalimplementation. In some countries, the budget balance rule is accompanied by additional limits on total government expenditures (Bulgaria, Venezuela), primary outlays (Argentina, Ecuador, Peru, Sweden), interest payments (Colombia) and/or the wage bill (Brazil, Colombia) in order to contain the fastest growing components of fiscal imbalance and the ensuing distortions in the composition o f the budget. Further, setting expenditure targets in line with potentialGDP growth(Ecuador) can help support a neutral stance with respect to the cycle. Similarly, to ensure cyclical neutrality, the budget balance rule can be defined in terms of structural or cyclically-adjusted balance (Chile, Sweden, and United Kingdom) that allows for the operation of automatic stabilizers. A similar function is performed by a balanced budget requirement specified in a multiyear or medium-term context (New Zealand, Estonia, European Union) with scope not only for the operation of automatic stabilizers, but also for active countercyclical discretionary action. An alternative approach to encouragecountercyclicalaction (or to support the structural or medium-termbalancerule) requires depositing contingency reserves in a stabilization fund, generated from fiscal surpluses during economic booms, and allows withdrawals to finance deficits during recessions(Argentina,Chile, Ecuador, Estonia, Peru). Source: George Kopits (2007) "Fiscal Responsibility Framework: International Experience and Implicationsfor Hungary", MNB Occasional Papers, 62. 9 Mongolia: Public Exwenditure and Financial ManagementReview 1.21 A look at planned versus actual revenue reveals that revenues have been systematically underestimated in planned annual budgets. Figure 1.5 and Table 1A.1 in Annex 1A reports planned, amended, and actual budget figures between 2001 and 2007. As can be seen from Figure 1.5 since 2003, revenue projectionshave fallen below actual figures by 11to more than 50 percent. These forecasting errors can be attributed in part to difficulties in projecting the volatile mineral prices on which Mongolia's public revenue depends. Another factor has undoubtedlybeen improvementsintax administration. 1.22 Turning to expenditures,actualexpenditures have been higherthan planned up to 10 percent of initial budget. Since 2004, the practice of amending the annual budget in mid- course was introduced. Interestingly, as shown in Figure 1.5, expenditures have failed to materialize as budgeted in the amended budget the most recent years. Indeed, budgeted capital spending has been systematically under-spent by 11, 16 and 13 percent in 2005, 2006, and 2007 respectively. The same has been true for net lending (of which foreign-financed project constitutes the majority), particularly in 2005 and 2006.' The GFMIS greatly contributed to improved execution of public spending, better treasury management, and public savings. However, in 2006, the implementation of the new procurement law created delays in budget execution as budget entities had to adapt to the new legislation. Difficulties also arose in executing the expanded expenditures as planned, especially in the investment program, due to limitations in absorptivecapacity. As a result, expenditures in 2005 and 2006 were not executed as planned. Bothhigher than plannedrevenue and lower then plannedexpenditurescontributedto unforeseengovernment budget surpluses in2005,2006 and 2007. Figure 1.5: Forecasting revenue has become a more challengii !task 1.2 1.1 1.o 1 2001 2002 2003 2004 2006 2007 1 0.9 -+-Total revenues (AcluaVPbn) 4- -Totalrevenues(AmndedlRan) -Total expenditures(ActuaVRan) ---e- Totalexpendnures (AmendedlBan) Source: MOF * Netlendingis primarily composed of direct capital infusions of funds to public (and sometimes private) enterprises(for example when governmentsacquire equity) andof governmentcredit programs undertaken for policy purposes.Given its implicit subsidy element and the possibility that some of the loans will never be repaid, net lendingis usually treatedas a deficit creating item, as it cannot be defined as pure financial intermediation. 10 Chapter 1-Fiscal Trends Box 1.2: Commodity saving and stabilization funds: asset allocation and spending rules Commodity saving funds allow governments to generate a permanent stream of foreign currency that i s independentof mineralresources,while also separatingcommodity revenues from fiscal budget planning and stabilizing income through portfolio diversification. Typically, these funds have one of 2 (or both) purposes: (i)Stabilize or dampen effect of oil revenues on fiscal budget, (ii)Save for current and future generationsto ensure intergenerational equity. The main challenges of managing commodity revenues are relatedto the so-called "resource curse" syndrome: (i)Commodity revenues are volatile and finite; (ii) There is a risk for fiscal policy to become pro-cyclical and for the economy to follow boom andburst cycles; (iii)Changes inrelativeprices lead to a relative decline in the manufacturing sector, with import-driven consumption burst and inflationary pressures (the so called Dutch-disease); (iv) Rent-seekingbehavior as the commodity rents becomebigger; (v) Ensuringintergenerational equity. Importantconsiderations to take into account when designingthe fund include: When deciding asset allocation, it is importantto keep in mind that the objective is to convert a depleting asset into diversified portfolio to create financial wealth. Therefore, the asset allocation should generate sufficient returns over the long-run. Finally the bulk of the fund should remain outside the domestic economy to avoid domestic pressures on prices. Turning to the fund management, management can be given to the Central Bank alongside reserves, to a separate investment unit in central bank (e.g. Norges Bank), managedby the Ministry of Finance (e.g. previous setup of Alberta Investment Management), or a separate agency with a professional Board(e.g.New Zealand SuperannuationFund). Commonspending rules usedby commoditysaving & stabilizationfunds are: 1 Bird-in-Hand rule: spending is equal to the expected real return times the market value of the fund (Norway, Alaska) 1 Total wealth rule: spending is equal to the expectedreal returntimes the total wealth (market value o f the fund +oil inthe ground) (East-Timor, Mauritania, Sao Tome) 1 Hybrid rule: Inbetweenbird inhandand total wealth rules 1 Mixed rule: in between bird in hand (or total wealth) and constant real spending (University endowment funds) Other rules include: price-based saving rules consisting of saving revenues in excess of budgeted revenues based on reference price (Russia, Oman) which can be withdraw when falling short of revenues (Trinidad and Tobago, Chile). 1 Macro-economic rules, which target non-oil primary deficit to GDP (with or without debt reduction) (Norway). Do commodity funds work? David et al. (2001) find weak evidence that countries with an oil fund tend to have more cautious fiscal policies. But a recent IMF study* suggests that commodity funds help reduce macroeconomic volatility measures by broadmoney and prices. Governance, ownership and credibility For a stabilization and saving fund to succeed requires an on-going education campaign along with a governance structure which assures the population that the funds are, indeed, being saved and properly invested. The authorizing environment may be a constitutional amendment, presidential decree or law passed by parliament, with a public buy-in through educational campaigns, legislative hearings, and consultative documents for the public. Transparency is key, and the Extractive Industry Transparency Initiative (EITI) is instrumental is promotinggood practice inthat area. Who should be thefiduciary of the Fund? Most countries have decided to have on some mixed of government, central bank and civil representation. The question of whether there is an inherent conflict between the government's desire to spend and the needto save should be addressed by a clear guideline on the maximumamount of budgetary finance available from the fund each year. This should be also completed by an overallfiscal discipline in controlling the non-oil/mining deficit, as the creation of the fund can be accompanied by a build-up of cumulative fiscal deficits, resultingin lower net savings. *Looking beyondthe fiscal: do oil funds bringmacro-stability(2007, IMF Workingpaper, WB/07/96) Source: WorldBank Treasury Department. 11 Mongolia: Public Expenditure and Financial Management Review Box 1.3: Windfall tax and Development Fund In June 2006, Mongolia introduced a windfall tax to: (1) capture a higher share of the revenues accruing to copper and gold mining companies from high export prices and (2) mobilize resources into a special fund-the Development Fund. The tax, which is modeled after the Russian oil export tax, is imposed on sales of gold and copper ore and concentrate extracted in Mongolia under the following provisions: Copper concentrate: the tax is levied at a rate of 68 percent on the difference between actual copper prices on the London Metal Exchange and the sum of a base price (set at US$2,600 per ton) and smeltingcosts (estimatedto be US$1,580 per (metric) ton in2006). Gold: the tax will be levied at a rate of 68 percent on the difference between the Bank of Mongolia's gold price' and the base price (set at $500 per ounce of gold). The windfall tax is deductible for purposes of the regular income tax. The base prices-$2,600 per ton for copper and $500 per ounce for gold-are not indexed. Companies with stability agreements are in theory unaffectedby this tax. The largest gold mine-Boro-was exempt, as it has a stability agreement; the agreementwas renegotiatedin2007 so that the company would start payingthe tax. The Development Fund (DF) was established with the objective of managingthe fiscal revenue from the windfall tax on copper and gold production. Its main provisions are contained in the Law on the Development Fund, adopted in February 2007. The law governs the building up, spending, reporting on execution and oversight of the development fund. The following are the main features of the DF according to the legislationinplace as of end-2007: Revenue sources: According to the DF law, the DF resources are: (i)the budget revenue under the windfall tax law; (ii)up to 100 percent of the previous year's budget surplus, net of the windfall tax, as decided by Parliament; (iii)the unspent balance of the contingency fund; (iv) interest earnings from the DF account; (v) other revenues. Operations of the DF: The DF has been designed to both finance specific spending priorities and fulfill a budget stabilization objective. According to the DF law, one third of the windfall tax revenues should be spent on domestic investment; one third should be spent on social protectionprograms to support children and families; and one third should be saved to compensate for "revenue loss due to force majeure conditions." The fiscal surplus, net of all windfall taxes in the previous year, should be spent on public investment inthe current year. The specific allocation of DF resources must be proposed by the executive and approved by Parliament. The DF law does not explicitly state that the spending execution procedures applied to the DF must be the same as those under the regular budget. The DF law does not have explicit provisionsregarding asset liability management, apart from stating that the resources of the DF must be placed in a special account ofthe Treasury and its interest earnings should be placedinthe Treasury Single Account. Accounting and reporting: According to the DF law, the Ministry o f Finance must incorporate the operations of the DF into the government's annual financial statements in line with International Accounting Standards. For that purpose, a manager of the DF and the spending units that finance projects/activitieswith DF resources should prepare expenditure reports on a half-yearly and annual basis. The financial statementsof the DF must be audited in accordancewith provisionsof the Public Audit Law andbe subject to an audit opinion. Inaddition, year-end expenditure reports ofthe DF must be disclosed in daily newspapersalongwith the audit opinion within 20 days ofthe reportsbeingfinalized. Source: IMF and WorldBank Stafs 1.23 Mongolia is at low risk of debt distress over the medium term and, with its vast mineralwealth, the public debt burden is projected to continue to decline. As a result of the primary deficit reduction and accelerating economic growth, there has been a sharp decline in public debt indicators and a reduction in government borrowing requirements. The projected path of debt is sustainable, assuming no major negative terms of trade shock and only concessional 9See Joint IMFrWorld Bank Debt Sustainability Analysis (2008), 12 Chapter I -Fiscal Trends borrowing. With high economic growth, favorable export prices and favorable terms (as of end- 2007, 95 percent of Mongolia's public external debt is concessional), this modest increase in dollar terms led to a significant decline in the public debt-to-GDP ratio, from 93.6 (61.0 NPV terms) percent in200'1to 40 (25.1 inNPV terms) percent in2007. The Government is working on a debt management strategy which will articulate the medium-term priorities for domestic and externalborrowing. ECONOMICCOMPOSITION OFEXPENDITURES 1.24 In 2007, total public spending increased by 58 percent (keeping prices constant), which is equivalent to the cumulative increase in public spending between 2000 and 2006. Over the latter period, public expenditures increased at a lower rate than real economic growth, leadingto a moderate decline in public expenditure as a share of GDP (from 36 to 29 percent)", followed by a quick increase to 41 percent in 2007 (see Figure 1.1). 1.25 This increase in government spending was largely financed by the even larger expansion in revenues.Expenditureswere characterizedby a: . . Rise intransfers includingsocialsecurityfunds and social welfare transfers; Decline in interest payments on foreign debt, owing to a stable outstanding stock of domestic and externaldebt; . Relatively stable share of personnel and good and services expenditures (20 percent and 30 percentrespectively); and Recent effort to expand capital expenditure, accompanied by a net decline in net lending (from 13 to 7 percent oftotal spending). 1.26 Overall, the budget is becoming increasingly rigid, thereby somewhat limitingthe medium-term fiscal space. The term "fiscal space" refers here to discretionary expenditures which can be undertakenwithout impairingthe country's solvency. The fiscal space equals total expenditures less personnel expenditures, interest payments, subsidies, social transfers and transfers to local governments. More spending is allocated to fixed or difficult to adjust expenditures, such as personnel, transfers, and interest payments. Fiscal space has diminished from 55 percent to 49 percent oftotal public expenditures. l1 1.27 The share of personnel expenditures has remained stable at around 19 percent of total public spending. But steady annual growth let to a slight decrease of the overall wage bill as a share of GDP over the same period, from 8 to 6 percent of GDP. Since the beginningof the transition, a policy of across-the-board increases of public wages has been implemented, with annual increase rates varying from 7.5 to 35 percent. As a result, adjusted for inflation public wages have doubled between2000 and 2006. By comparison, privatewages have increasedby 60 over the same period.The average salary inthe public sector has now outpacedthe average salary inthe private sector (see note on civil service in Annex IC). 1.28 Explicit subsidies to public enterprises have remained stable at around 1 percent of total expenditures, but implicit subsidies are significant. Explicit subsidies include subsidies to energy and heating supply and transportation. Significant implicit subsidies exist in the form of net-lendingto public entities. Electricity, heat, water and/or sanitationtariffs are set too low to cover operating and investment costs. This makes the private sector uninterested in investingin service provision, and diverts government resourcesto coveringutility losses.Inturn, loWe usethe Consumer Price Index(CPI) as the deflator ofpublic spending. IIHeller (2005) "Understanding Fiscal Space", IMFDiscussionPaper, March, 13 Mongolia: Public Expenditure and Financial ManagementReview the cost o f new investments has been born, over the last 10 years, almost completely by donors; loans and grants have accounted for around 90 percent o f infrastructure investment.'* Box 1.4: Civil Service structure and changes and wage bill increases- Introducing differentiation Inthe approved2007 budget,the Governmenthadprovided for an increase in the wages by 30 percent. In the beginning of the fiscal year, wages were increased by 20 percent. The remaining 10 percent was retained. In July 2007, the government moved to increase salaries by the remaining 10 percent. However, the Parliament did not allow this additional increase, but mandated that the Government increase the average civil service wage up to tgs. 250,000, starting in October 1, 2007. Following this decision, the Government issued two resolutions (No.236 and No.237) to revise the civil service grades and classification, and revise the minimum wages for each grades. According to these new regulations, bonuses for support staff, especially for primary and secondary school teachers are incorporated into the basic salary. Prior to this revision, the civil service was classified into four categories: (i)political, (ii) administrative, (iii)special and, (iv) support staff. Resolution No.236, increased the categories to: (i) administrative, (ii)political, (iii)special staff, (iv) support staff in science sector, (v) support staff in tertiary and vocational education, (vi) support staff in primary and secondary education institutions, (vii) health sector support staff, (viii) support staff of cultural and art institutions, and (ix) support staff in other entities. The government also approved minimum and maximum salary for administrative, political and special civil servants but defined only the minimum salary for all types of support staff, In FY2008 the government has decided to increase the wage bill by between 20 - 30 percent, by providing differentiated increases to different positions. This is the first time the Government has decided to provide differentiated wage increases to different positions. Previously, the Government had provided across-the-boardincreases. The FY2008 proposal calls for decompression of the wage structure between public management and operational posts by doubling the wages of state high-ranking and equivalent positions, and staff at the court and prosecutor's office, as compared with other posts. The compression rates ofpublic administration posts would increase from 2.0 to 2.5, and for technical posts from 2.4 to 3.1, ~ .~ ~~~ - ~ ~~ ~ As part of the Budget process the Government is proposing to eliminate extra duty allowances and to consolidate the provision into the basic pay package. At the same time, the proposal would increase the additional pay for political posts for special work to 30 percent, taking into account specifics of the work and functions. However, various additional allowances for special civil servants for service term and rank havebeenretained. Source: WorldBank Staff 1.29 Since 2000, social transfers to households have increased substantiallywhile social insurance contributions have declined. Budgeted social insurance expenditures for 2007 are estimated to be about 11 percent o f 2007 GDP, having risen substantially from about 7.0 percent of GDP between 2000 and 2005. Much of the reason for the increase was the provision of the so- called Child Money program benefits; in 2007, these benefits amounted to about 11 percent o f social insurance expenditure^.'^ Old age pensions represent the largest single category o f social insurance expenditure, anticipated to be about 4.3 percent o f GDP during 2007 or about 48 percent o f all social insurance expenditures. Pensions rose from 3.75 percent o f GDP in 2001 to 4.26 percent in 2007. Taken together, old age, disability, survivors and military pensions represented about 65 percent o f all social insurance expenditures. Disability pensions were anticipated to be about 0.8 percent o f GDP during 2007 or about 9.5 percent o f all social insurance expenditures. ~~ I2 World Bank (2007) "Foundations for Sustainable Development: Rethinking the Delivery of Infrastructure Services in Mongolia". I3Other social insurance programs were introduced in 2006 such as for newly married couples, newborns, school lunches, as well as for large families. 14 Chapter I -Fiscal Trends 1.30 Recently, there has been more emphasis on capital spending, with capital expenditures accounting for one quarter oftotal expenditures (5 percent of GDP) on average since 2000, and up to one third of total expenditures (11 percent of GDP) in 2007. This recent increase reflects the priority on infrastructure investment inthe newly created development fundI4. Box 1.5: Mongolia Public Investment Program The 2008 Budget reportsthe following principleswhich shouldguide public investmentdecisions in Mongolia: . The Government Action Plan The completionof on-goingprojects m The concept of regionaldevelopment Incomegeneration by encouragingeconomic growth Youthrecreational andentertainment activities In addition, the proposed Budget includes an assessment of investment efficiency that looks at two main indicators:the share of investment in total gross domestic product (GDP) and the incremental capital- output ratio (ICOR). The latter is used to estimate the capacity of the economy to absorb planned public investment. The Public InvestmentProgram (PIP) focuses on domestically-financedprojects. The PIP consists of a list or projectsfor each ministry.Althoughthe Development Fundis kept separate from the GeneralBudget account, the 2008 Budget proposal and PIP discuss public investment projects from all domestic sources (Le. the General Budget, Development Fund, and Road Fund). The linkages with the overarching planning strategy are explicitly presentedin the 2008 Budget document, but the rationale(e.g. social and economic) for the selection of each ofthese projectsneedsfurther elaboration. It would also be usefiil to discuss inthe project appraisal the trade-offs that were involvedindecidingon these public investmentprograms, and the associated cost-benefit analyses. More importantly, an estimate of the yearly recurrent costs generated by new public investments should be discussed, since these will impact the future fiscal space of the Governmentand its medium term expenditure framework(MTEF). - - - ~~ ~~~~ - ___ Foreign fimded projects are presented separately. The proposed 2008 Government Budget document discusses progress in each of the projects, though without discussing their rationale, their consistency with the domestic PIP, and overallprioritization. Source: Mongolia 2008Annual Budget Proposal maintenance spending. Contrary to international best Figure 1.6: Official development assistance contribution is declining practices, the amount of 70% expenditures allocated to existing capital has declined as 60% a share of total capital spending 50% capital expenditures by the government, from 10 percent in 2000 to 4 percent in 40% +Share of ODA in total government revenues the 2007 amended budget. In 300h addition, there is a need for an I$.. Share of total ODA inGDP 2ooh explicit analysis and calculation for budgetary purposes of the 10% associated recurrent costs which 0% come with new investments, to 2001 2002 2003 2004 2005 20Og 2007 2007 clarify the actual cost of Source Development Assistance Committee (DAC) database MOF ~ ~~ l4By law, one third of spendingunder the fund shouldbe allocatedto infrastructure. 15 Mongolia: Public Expenditure and Financial ManagementReview 1.32 Efforts to recollect loans from third partiesfrom the general budget have led to a decline in net lending in total expenditures, while new foreign loans commitments show a steady increase. But the fast expansion of domestically-financedcapital outlays has diminished the significance of foreign-financed capital expenditure in the budget. Before 2006, foreign funding contributedto more than half of the budget; in 2007 and 2008, its share droppedto 15 and 28 percent respectively. 1.33 Net official development assistance (ODA) has remained steady since the beginningof the decade but as the economy grows and expands, its relativecontribution to Mongolia'sGDP and public revenues diminishes.Net disbursed annual ODA amounted to, on. average, TG 230-300 billion (or USD 200-250 million) since 2001, of which only a third is directed to the Government. Projects loans constitute most of the aid (77 percent on average between2001 and 2006) while program loans and grants account for the rest. Exceptfor program loans, all foreign lending is executedoutside the budget. FUNCTIONAL COMPOSITION OF EXPENDITURES 1.34 A comparison of the functional composition of Mongolia's public spending with internationalcomparatorsrevealsthe following issues: (i)Arelativelyhighlevelofspendinginthesocialsectors,includingineducation, health, social security and social welfare. Education and health expenditures are high relative to other countries with similar level of GDPs, at an average of 20 and 10 percent of total government spending respectively between 2000 and 2005. (i& recently, there hasbeena relativedeclinein spendingon education More and health with more emphasis on social transfer programs. With high economic growth and fast-growing government revenues and expenditures in 2006 and 2007, relatively less emphasis was put on health and education, in favor of social welfare and infrastructure spending. Hence, in 2006 and 2007, the shares of education and health were 13 and 8 percent respectively of the government's total spending. (E)A recent emphasis on infrastructure capital spending (transport, communication, energy and heating), with a tripling of infrastructure investment between 2006 (actual) and 2008 (as approved in the 2008 budget). This surge in budgeted capital spending raises doubt on whether it can be absorbed without generatingmuchinflation. 1.35 The share of government spending spent on education has recovered since 2000 after struggling throughout the transition period.Under the former centrally planned system education was a high priority and received the largest share of government resources. However, spending on education dropped sharply during the transition period. After reaching a low of 16 percent of the government budget in 1997, the share of the budget spent on educationhas begun to increase, reaching an average of 20 percent of total public expenditures between 2000 and 2005, a policy target announcedby the Government. More recently, the share has fallen again, to 16 percent in 2006 and 11 percent in 2007 planned budget, as a result of quickly increasing expenditures, and a relatively weaker emphasis on education than on social transfers and infrastructures. 16 Chapter 1-Fiscal Trends 1.36 Government health expenditures are also relatively high compared to countries with similar income levels. This reflects the relatively high level o f total health expenditures in the country, including private health care, and an average Figure 1.7a: On average, priorities are given to social commitment o f 10 percent o f sectors (social spending, health, education) ... government spending to health. Mining,manufactunng, and construction 1.37 Social assistance Housingand community amenities expenditures have dramatically Agnculture,forestry increased since 2000. Social Recreationalandsporting,cultural services assistance comprises social Energy,heating Transportand communication insurance (retirement, Other economicactivities unemployment or sickness), social Defense assistance allowances (child- Publicorder and safety related allowances) and active Generalpublic serviws labor market programs. Overall, Health total social assistance expenses Education 18 social secunty ana social welfare increased from 20 to 23 percent of 0% total government spending 5% 10% 15% 20% 25% Average share in total expenditures (2001-2007) between 2001 and 2007. l 7 2 r igure I.IU; -__I_ _-^ 1 LIL. ....DUL are iriereasingiy smiling iuwarus eveii D - . A I ___.-I-1- __I___ 1--I-. _LLA-*II_ I 1.38 Pensions are more social spending and infrastructure away from health increasingly burdening the and education government budget. Examining contributions and benefits together, the overall difference ' L 2006-2007 Generalpub1 I between contributions and benefits has grown from less than 2 percent o f total government expenditures in 2000 to 5.8 percent o f total government I I1 spending in the approved 2008 budget. The main reason for this is the decline in security contributions from about 4.8 Social ecunty a d social elfare I I I ',.A\\\ I I I percent o f GDP in 2000 to 3.6 -6% 4% 4% -2% 0% 2% 4% 6% 8% 10% percent o f GDP in 2007. Since (+)(-)Variation of share in total expenditure about 73 percent o f social Source: MOF insurance benefits are for long- term benefits such as pensions; disability and survivorship, this means that the Government subsidy o f social insurance has had to increase substantially since 2001 in order to support the shortfall between pay-as-you go contributions and long-term benefits. 1.39 The Child Money Program (CMP), designed in 2005 to rationalize social assistance benefits, has quickly become the main driver of a fast-expanding social assistance. The share o f social welfare expenditures in total spending has increased from 3.2 in 2001 to 10 percent in the 2008 planned budget, with the CMP accounting for 5.6 percent o f total expenditures. Except for child-related allowances, which have been recently consolidated into a single CCT, these programs are largely still in place. 1.40 The CMP was originally conceived as a targeted Conditional Cash Transfer (CCT) Program, but was quickly universalized. Created in 2005, the CMP grouped the many redundant child-related cash benefits inherited from the transition period into a targeted CCT 17 Mongolia: Public Expenditure and Financial Management Review program. CMP implementation has had many difficulties, but has had a significant impact on poverty reduction. 1.41 Capital spending shows greater emphasis on economic services, including infrastructure,agricultureand forestry, mining,manufacturingand construction,transport and telecommunication. Capital spending in economic services accounted for more than half of the total capital budget in 2007. Investment in infrastructure has been maintained throughout the decade at around 4 percent o f GDP or 13 percent of total spending, reaching a record level in 2007 at 19 percent o f total expenditures (which corresponds to almost 10 percent of GDP). While part of this greater focus is related to increased investment in the infrastructure and energy sectors, greater investment resources have also been devoted to building or upgrading local and central government infrastructure. Investment in other economic activities has increased to an average o f 12 percent o f total capital expenditures beginning in 2006. This reflects greater attention to investment in central government services and local government, which contributes a third to the increase. 1.42 More attention has also been drawn to public investment in the education and healthsectors as reflected in the 2007 and 2008 budget, with an average increase o f 30 percent per annum since 2001. Because o f the high rate of growth in total expenditures, this translates into a stable share o f total capital expenditures. 1.43 Provision of general public services has increased at an average rate of only 3 percent per annum since 2001, with ajump of 63 percent in 2007. As a share of expenditures, however, these increases translate into a marked decline, from 25 percent of total spending in 2001 to 16 percent in 2007. More attention has been drawn to capital expenditures over time, with capital expenses accounting for more than a third o f expenditures in 2006/7 but only 17 percent in 2001. LOCAL EXPENDITURES" 1.44 There has been a marked move towards fiscal centralization since 2003. The intergovernmental fiscal mechanism has its legal basis in the General Budget Law and the Public Sector Management and Finance Law (PSMFL, 2003). The former law regulates the revenue relations and the latter one the expenditure responsibilities. With the promulgation o f the PSMFL most o f the taxes that were previously shared with local governments were centralized, along with service delivery responsibility in all major sectors. Tax laws and rates are determined and administered by the central government after getting approved by the Parliament. Local governments have insignificant discretion in determining rates o f taxes assigned to themI6. Currently, the Value Added Tax (VAT) on domestic transactions is the only tax that is shared with aimags based on a local population pro rata formula. Income taxes, excise taxes, VAT on imports, and royalties on gold are centralized. Some local taxes, including property tax, non-gold resource usage fees, and some local levies are assigned to local governments. Local governments have very limited power neither to institute new taxes nor to adjust the tax rates. In 2004, aggregate local government spending accounted for only 8 percent o f total revenues and equivalent to 7 percent o f total expenditures. More recently in 2008, local governments were authorized to keep the personal income tax according to the amendment into the General Taxation Law and General Budget Law. l5See World Bank Mongolia Regional DevelopmentPolicy Note (2005). 16The General Taxation Law Article 18-3 states that Aimag Hurals may define at the margin the rates of 4 revenues: (1) License fees for the use of natural resources except minerals, (2) Payment for use o f natural plants, (3) Paymentfor use of other widely spread minerals, (4) Payment for use of mineral springs. 18 ChaDter I -Fiscal Trends 1.45 Local governments are currently responsible for limited service delivery. Article 52 o f the PSMFL specifies that local governments will deliver the following services, which are financed. from local government revenues: (i)measures on public hygiene; (ii)local environmental conservation and protection, gardening and maintenance; (iii)pest eradication and control; (iv) local road maintenance; (v) operation o f local water and sewerage system, (vi) flood prevention and soil protection; (vii) outputs associated with local public infrastructure facilities; (viii) fire prevention, protection, and mitigation; and (ix) measures to fight and prevent infectious livestock and animal diseases. Besides these, all other services are to be provided by the central government. Table lB.l in Annex 1B shows the division o f roles and responsibilities by sector between central and local governments in Mongolia. 1.46 Provision of education and healthcare services is done via agency contracts between central government and local governments. By law the central government is responsible for service delivery in the education and health sectors. The central government determines the overall policy and funding in each sector while service provision is done via agency contracts between the central government and the local government. Box 1.6 presents the institutional and intergovernmental fiscal framework for the educational sector. Box 1.6: Intergovernmental fiscal mechanism in the education sector The PSMFL has vested the responsibility of education provision with the central government. The national policy and the formula determining school financing are establishedby the Ministry of Education. This financing formula comprises a fixed cost component (energy, water, and heating) and a variable cost component (teacher salaries and other operational costs). Basedon the financing formula, schools propose a budgetaryrequirement to the respectiveaimags who, in turn, collates requirementsfor all schools within the aimag and submit the total requirement to the Ministry of Education. The variable cost component contained in the proposal is based on an assumption of the number of pupils who shall be enrolled at the respective school. The central government then decides how much resources will be provided at the aggregate level to the education sector. Once the sectoral resourceenvelope is determined, the Ministry of Education provides an allocation for each school. Experience shows that schools generally assume more pupils would be enrolled, while the central governmentprovides fewer resources than requested.Even after budgets for each school are fixed, schools can ask for amendments to their allocations by providing justifications. These amendment requests are transmitted to the central government via the aimags. The system for approving/disapproving amendments i s not clear and providesroom for patronage. The Ministry of Education enters into performance contracts with aimag Governors for delivery of education services in their respective jurisdictions. These contracts set out the curriculum that will be taught at local schools and the resources that will be provided by the Ministry of Education. At the aimag level there is an education sector coordinator who monitors the performance of schools within the respective aimag. The sector coordinator reports to the Ministry of Education in consultation with the Governor. Aimag Governors appoints school principals in consultation with the Ministry of Education.The principal is vested with the responsibility to run the school and to ensure that education is imparted to students. Shehe has the authority to appoint teachers and utilize the school budgets according to guidelines establishedby the central government. The principal reports to the aimag education sector coordinator on performance. Schools are classified as individual budget entities and as such have sub-accounts in the Treasury Single Account. The funds flow arrangement works as such that the Ministry of Finance, on advice from the Ministry of Education, credits each school sub-account with the approvedbudget. This credit is treated as funds provided through the aimag budget, but only notionally because aimags Governors do not have any control over these finds. All non-wage expenses are paid directly by the central government, upon submission of respective bills by principals to aimag Treasury offices. Only wage payments are made directly to the Principals. It is the responsibility ofthe centralgovernmentto ensuretimely provisionofresources for schools. Source: World Bank Mongolia Regional Development Policy Note (2005) 19 Mongolia: Public Expenditure and Financial Management Review 1.47 Consistent with reduction in assigned Figure 1.8: Local government expenditureswere reduced responsibility of local in accordance with the transfer of service delivery governments, there has been responsibilitiesto the Center corresponding reduction in 100% 90% local government spending. 80% Figure 1.8 shows the share of 70% expenditures by local 60% governments, between 2001 50% and 2002 and between 2003 40% and 2007. As can be seen, there 30% was a significant reduction in 20% expenditures by local 10% 0% government after 2002, with a 2001-2002 2003 2007 compensating increase in m Central government spending expenditures by the central I!Local government spending (financedby localtax and VAT) Localgovernment spending(financedby central governmenttransfers) government. This shift in expenditure depicts the shift in , Source: MOF service delivery responsibility from the local government level to the central government. The reduction in local government revenues depicts the centralization of taxes. Local governments still receive transfers from the centralgovernment, yet these transfers representnow a minor portionoftotal public spending. 1.48 The current intergovernmental fiscal mechanismof centralized fiscal management has had many benefits in Mongolia: Macroeconomic stability has been maintained. By limiting the risk for deficit accumulation at the sub-national level, with central government controlling revenue collection, expenditure allocations, and deficit financing decisions, the authorities have succeeded inbringingdownthe aggregate fiscal deficit. Wagearrears have been controlled. Prior to 2003, local governments exercisedmore authority on expenditure allocation. It became apparent that there were significant arrears on wages and services, even though funds had been released to local governments. Because civil servants had to be paid, the central government had to clear the arrears. As fiscal control was centralized, the problem of arrears on wages was resolved. Timely and predictable budgetary transfers are being made to local governments. Once the national parliament approves the state budget specifying financial support to each aimag, the Ministry of Finance and aimag administrationdiscuss the latter's monthly cash-flowplans (prepared by aimags). After discussions, the monthly plan is approvedby MoF andtransfers are madeto aimags accordingto these monthlyplans. In addition, as funds are now directly sent from the central government to schools and hospitalsthrough aimagtreasury offices, there has beenan improvement in funds receipts by these entities.Analysis of fundtransfer from 2001 to 2004 also shows that the central government has been providing full funds to local authoritiesas per their approved budgets. 1.49 However, there is a need to clarity the methodology of, and increase local government participation in, the process for determining transfers to local governments, including VAT and mining revenue sharing arrangements. At the moment the central government determines transfers to local governments, including VAT share, with minimal 20 Chapter I -Fiscal Trends consultation with local government. In addition, there is no clear methodology applied by the central government in this regard. This results in significant fiscal uncertainty for local authorities, as they cannot predict what their resource envelope will be each year, thereby frustrating budget planning. To this end, a committee consisting of the Ministry of Finance and the local governments could be formed to decide on the methodology for determining the revenue-sharing formula. 21 CHAPTER 2. PUBLIC SECTOR FINANCIAL MANAGEMENT PROGRESSAND CHALLENGES Budget planningand execution The Public Sector Financial Management Law (PSFML) has been a major positive milestone for Mongolia's public finances. It is a robust and internally consistent legislative framework, and contains all the key elements o f a good public finance management system. Budget planning is underpinned by a clearly-defined legal framework, including the medium-term fiscal framework. It could also benefit from fuller integration o f sectoral priorities, ODA, and local preferences. The PIP does not: include foreign-financed projects, and does not provide the rationale for investment selectivity, thereby weakening the link between stated policy priorities at the sectoral and local level and realizedexpenditures. Mongolia has made significant headway in improving the budget execution system. Since the PSMFL was enacted, a functional Treasury Single Account has been established, a unified and consistent chart o f account was implemented, an integrated debt management information system and an integrated government financial management information system were deployed, a country-wide electronic payment settlement system at the Bank o f Mongolia was put in place, and a performance contracting framework has been operationalized across the government. Key internal controls surrounding the Government Financial Management Information System (GFMIS) are adequate, but they still need improvements. Additionally, these key internal controls should be reviewed periodically, at least yearly, to ensurethat they are still functioning as designed. . Procurement The new Procurement Law enacted inFebruary 2006 was a commendable improvement over previous legislation. However, a February 2007 amendment significantly altered the spirit o f the law by allowing direct contracting and largely restricting international competition by significantly increasing the threshold for international competitive bids. Audit . The overall legislative framework is adequate for the Mongolia National Audit Office to effectively audit the Government. The biggest challenge is ensuring that the agency has the capacity to conduct robust financial audits. Transparency In2006, the Government made significant progress indisseminating some key budget documents and information. However, the PSMFL remains silent on disclosure o fbudget information. 23 Mongolia: Public Expenditure and Financial Management Review KEYFUTURE PRIORITIES Budget planningand execution 1 Rationalize high level strategies and strengthen the links among different sectoral strategies, Official Development Assistance (ODA), Public Investment Program (PIP) and the Medium-Term Budget Framework (MTBF)by making the Public Investment Program the sole entry point for capital projects. > Suggested "how to": The key government policy documents should be: (i) the Government Action Plan (GAP); (ii) the Socio-Economic Guidelines that translate the GAP into annual and rolling action plan; (iii) the MTBF which essentially sets the corresponding me 3) budget framework; * 'stries that take into d then translates iture frameworks 1 Prepare a robust Public Investment Program with clarified selection criteria. > Suggested "how to": Constitute an inter-ministerial working group led by the Directorate General for fiscal policy department to: (i)determine the medium fiscal space taking into account the budget momentum (Le. medium-long term committed expenditures); (ii) investment needs across the sectors; and (iii)the government action plan. 1 Review the efficacy o f performance contracts at the Department levels as well as the use o f strategic businessplans at the Budget Entity Levels. 9 Suggested "how to": Commission a review of the progress and challenges with implementation of performance contracts under PSMFL, with a focus on efficacy o f each level o f contracting. There is need for such a review to do a cost o f compliance versus benefit from performance contracting for each level and present its findings to the Government for review and action. From the PER analysis it seems that there is limited real value o f performance contracting at the department-staff level because it is difficult (if not impossible) to assign individual staff contribution to collective departmental outputs. This is similar for budget entity level strategic business plans, as it is most times difficult to separate budget entity strategic plans from the parentministry action plans. 1 Strengthen local horizontal accountability-between local government and service delivery units- while maintaining the existing centralized budgetexecution framework. > Suggested "how to": Amend the PSMFL to include: (i) assessment reports o f service delivery annual units from the provincial government to the parent ministries; (ii) mandate publication o f performance reports o f budget entities at the provincial level; and (iii)mandate publication at the local level total budget allocations and disbursements to service delivery units from central governments. 1 Develop a Government-Development Partner action plan with steps and timelines for integrating ODA into country systems, and move towards more sector-wide approaches learning lessons from the successful donor coordination inthe education sector. > Suggested "how to": Constitute the MoF team by order o f the Minister responsible for working with Development Partners to develop a phased plan for closing all special project accounts held outside the TSA and to migrate them to GFMIS. This MoF team should by led by the Treasury Department with representation of all concerned departments. The government should pre-announce this intention to all development partners and determine specific dated milestones for this migration plan. 1 GFMIS controls around budgetary, system integrity, payment security, and system management should be improved upon and key internal controls should be reviewed, at least annually, to ensure they are still functioning as designed. > m e s t e d "how to": Review the systems audit report from the recent Mongolia National Audit Office 24 Chapter 2 -Public Sector Financial Management o f GFMIS andto develop respective regulations and system-fixes to address those concerns. Procurement . m Withdraw the February PPLM amendment. Improve the integration o fpublic procurement activities with the Government FM system. > Suggested "how to": Implement the Procurement Management, Monitoring and Information Svstem and inteaate it with the Government Financial Management and Information Svstem, m Develop professional standards for public procurement officers and design appropriate capacity buildingprograms for those that do not meet them. Audit Strengthen the capacity o f Mongolia National Audit Office (MNAO) to conduct robust financial audits o f GFMIS. > Suggested "how to": Parliament to provide sufficient annual budget for various training o f MNAO staff and for MNAO to continue working on systems audit capabilities. M N A O could consider (i) having dedicated system audit teams to receive necessary continued training to perform this task and, (ii)instilling periodic internal examination of MNAO staff for consideration of promotions or demotions to improve the overall quality o f MNAO. Transparency Propose clauses for relevant laws for public release o f budget information. > Suggested "how to": The Ministry o f Finance has established a dynamic portal website. Now the Minister for Finance should issue a Ministerial Instruction not all concerned departments reporting to the Ministry of Finance to: (i)list all laws, decrees, regulations, and guidelines that determine respective departmental operations and to publish them on this MOF dynamic portal; and (ii) require departments to submit compliance reports on the same. Establish a regular forum during which the Government will present MBTF policies and proposed State Budgets to media and civil society organizations, around the same time that the policies are presented to Parliament > Suggested "how to": As part o f the budget preparation and approval process, the Minister for Finance should plan and pre-announce a timetable for discussing the executive budget proposal with media and civil societj, 2-3 weeks before this proposal is submitted to the Parliament. This is to ensure that there i s appropriate time for modifying the budget proposal as necessarybasedon these consultations. 25 Mongolia: Public Expenditure and Financial Management Review INTRODUCTION 2.1 This chapter highlights key developments in the public financial management (PFM) of public finances in Mongolia since 2003, along with the current challenges. The chapter analyzes issues faced by authorities and recommends ways of addressing issues that would improvethe PFM of public finances. 2.2 This chapter covers four elements of public financial management: (i)the legislativeframework, (ii) public finance planningand execution, (iii)oversight and adequacy of audit functions, and (iv) disclosure and transparency in public finances. These elements relate to the entire work-flow of the budget system at the aggregate level. This chapter examines these elements at the macro level. Sector-specific issues are covered in subsequent chapters of this report. THELEGISLATIVE FRAMEWORK 2.3 The legislative underpinning of public finance is the Public Sector Management and FinanceLaw (PSMFL, 2003) and the annualBudget Law. This Law applies to all levels of organizations and to filly and partly (with government ownership of 51 percent and more) centrally- and locally-ownedlegal bodies. The PSMFL was promulgatedon January 1, 2003 and mandatedthe following key provisions: 2.4 Centralization of the intergovernmental fiscal management framework. Budget planningand executionauthority was shiftedaway from provincial authorities. Responsibilityfor policy planningwas vested with sectoral ministriesand responsibility for overall budget planning and execution, as well as revenue policy and administration, with the Ministry of Finance. Provincial authorities were tasked with performing general administration. The PSMFL vested the authority to borrow (domestically and internationally) with the Ministry of Finance and clearly set out a centralizedcash managementframework. a. Institutionalization of a medium term framework for budgeting. The PSMFL mandates the development of a medium term (three year) fiscal framework (MTFF) along with a deadline for submission ofthe fiscal frameworkto the Parliament(May 1). This MTFF should guide the formulation of the socio-economic guidelines and the State Budget. The Law also mandates that budget entities send multi-year budget requests to their respective sector ministries, and sets a clear timetable for preparation and submission of budget estimates to the Parliament.At the time of the submission of budget estimates, the Law also requires the government to submit a statement of fiscal risks, including risks which could resultfiom decisions already taken by the Government, and any contingent liabilities. b. Clarificationof the roles and responsibilities of each levelof Government, consistent with the intergovernmentalfiscal relations framework. PSMFL is very clear on the roles and responsibilities of the Parliament, portfolio ministries, and state administrative bodies responsible for finance and budget. This includes the administration of bank accounts into which receipts are deposited and from which payments are made. The Law also lays down prohibitions for state budget entities, consistent with the overall financial responsibilitiesthat are laid out in the Law (Article 14). PSMFLmandates a vertical accountabilitystructure at all levels inthe financial managementframework. c. Introduction of a comprehensive performance management framework across the public sector, at all levels.There is avertical system of performance contracts mandatedby the Law-individual staff with the divisioddepartment director, departmental director with 26 Chapter 2 -Public Sector Financial Management the state secretary, state secretary with the minister, and the portfolio minister to the Parliament.This accountabilitynetwork is emphasizedthroughoutthe performance contracts. The Law mandates that the performance agreement needs to be concluded within one month of the approval of the State Budget by the Parliament. Further, the Law mandates the preparation of Strategic Business Plans for executing the Budget. The description of the strategic business plans in the Law (Article 26) is quite detailed, as are the timelines for submission of these plans to respective ministers (July 1) and to the Ministry of Finance (August 15). A similar framework is laid out for State Owned Enterprises (SOE), which submit their respective plans to the State Property Committee. At the same time, the Law requires Parliamentary bodies to provide advice to ministers on policy matters including regulations.Box 2.1 shows how performance management is practicedinMongolia. d. Full accrual accounting mandate, with financial reporting to the Parliament: The Law mandates extensive internalreporting on budget executionto the Parliament including: (i)operatingstatement; (ii)balancesheet; (iii)statementofcashflows; (iv) statementof changes in net assets; (v) statement of commitments and contingent liabilities; (vi) statement of accountingpolicies; and (vii) budget executionreports compared with appropriations.The Law also mandates that Government submit the financial statements for the entire government to the MNAO by March 20 (fiscal year runs from January - December) each year. The Law also requires the MNAO to submit the audit opinion to the Parliament within one monthof receivingthe financial accounts from the Government. Article 40 stipulates that all state budget entities must prepare financial statement on a semi-annual and annual basis, includingbalance sheets conformingto internationalaccountingstandards. A review of the PSFML implementationprogress 2.5 The promulgation of the PSMFL has been a major positive milestone for Mongolia.The Law has established a robust legislativeframework for managingpublic finances. Overall, the legislative framework for public finance management in Mongolia is similar to that inNewZealand andAustralia. 2.6 Implementation of the PSMFL has been steady, with the Government making significant progress toward centralizing the intergovernmental fiscal framework and implementinga functional centralized cash managementsystem, with the provincialTreasury offices reportingvertically to the Ministry of Finance.Before 2003, the Mongolian public finance management system was fragmented and lacked operational fiduciary controls in budget execution. The Government was borrowingovernight to meet expenditure needs as there was no consolidated single account, and budget reporting was negligible. By 2007, the authorities had managed to implement a fully functional integrated budget execution system, addressed the problem of arrears, consolidated the government finances into a Treasury Single Account framework, and significantly increased budget reporting.The medium-termfiscal framework has also been established, and the Parliament uses this tool to approve aggregate macro-fiscal ceilings. The government has also institutedthe performance management framework across the public sector (discussed in Box 2.1). 27 Mongolia: Public Expenditure and Financial ManagementReview Box 2.1: Performance-basedManagement- Theory and Practice in Mongolia A performance-basedmanagementframework is a systematic approachto improving service delivery effectiveness. This is achieved by establishing strategic performancemetria for agenciesand departments; measuring performance by collecting, analyzing, reviewingand reportingperformance data; and using this informationto drive andrewardperformanceimprovement. Within a performance-basedmanagementsystem, all levels of a Departmental work together to define requirements, establish desired results, and agree on management methods for measuring and evaluating outcomes. To this end, performance-based management encourages innovation and drives long-term performance improvement, must be able to respond to unique situations, and should establish appropriate accountability mechanisms. Performance-based management also prompts partnership, trust, and transparencywhich in-turnencourages innovationand creativityat all organization levels. Performancemanagementsystems have beenused successfully in the private sector for many years. In the manufacturingand service sectors, where it is possible to identify the quality and quantity of outputs, performance management has meant that individuals and units are provided resources to then deliver certain units of outputs of a certain quality. Indicatorstransparently monitor the performance of individuals and units. At the end of the performanceassessment period, results are measuredagainst the agreed targets, and individuals who perform well get bonuses while those who underperformed are penalized. As outcomes are easily verifiable in these situations, such performance agreements have helped improve the effectiveness of staff andprocesses. In Mongolia, the government has instituted the performance-based management across the public sector since 2004. Strategic business plans and performance contracts are conducted from low-level management and staff to higher levels-Minister to Cabinet, State Secretary to Minister, Departmental Directorsto State Secretary, DepartmentDirectorto respective Division Directors and staff. At the beginning of each year, personnel sign performance agreements with their supervisors. At the end of the year, supervisors discuss and appraise individual staff based their performance. It is good practice to have such a performance management system operational at the levels of Departmental Directors and State Secretary, and above. However, it is unclear what the benefit is of performance agreements between Departmental Directors and staff. This is because: (i)It is generally difficult to distinguish individual staff responsibilitiesin departmental tasks. For example, ifthere is a team o f 4 staff tasked with developing the Debt Management Strategy, it is often difficult to enter into performance agreements with each of the staff for their part of the strategy. (ii) Financial incentives to reward good performance are often too small, and due to lack of staff HR control at the departmental level, it is not possible to take action against under-performers. (iii)During the year, staff is assignedtasks that were not originally envisaged, and therefore it is not possible to measure performance against pre-set targets. (iv) Developing performance contracts takes a significant amount of time and effort, especially at the lowest level, because most of the tasks are inter-dependent, so it is not possibleto accurately isolate the sub-tasks that are assignedto specific personnel. Therefore, it is recommendedthat the Government revisit the practice of performance agreements at the department-staff level and determine whether benefits of such performance agreements out weigh the costs. A review of the progress and challenges with implementation of performance contracts under PSMFL could be commissioned, with a focus on efficacy of each level of contracting.There is need for such a review to do a cost of compliance versus benefit from performance contractingfor each level and present its findings to the Government for review and action. From this analysis, it seems that there is limited real value o f performance contracting at the department-staff level because it is difficult (if not impossible) to assign individual staff contribution to collective departmental outputs. This is similar for budget entity level strategic business plans, as it is most times difficult to separate budget entity strategic plans from the parentministry action plans. Source: WorldBank Staff 2.7 The Government has managed to implementa functional modified accrualsystem (commitment accounting without accounting for fixed assets). With regards to full accrual accounting (commitment and fixed asset accounting), the government has passed regulations governing fixed asset accounting, but there has been little progress in developing fixed asset 28 Chapter 2 -Public Sector Financial Management valuation accounting, depreciationprocedures and recordingof tangible and non-tangible assets. This is the reason that Mongolia has not been able to develop a government-widebalance sheet, as stipulated in the PSMFL. Clearly, valuation of assets i s a skill-intensive task and it is understandablethat Mongolia has not been able to make significant headway in this area because of human capital scarcity on asset valuation. However, the Government has recently started pilotingfixed asset valuation. 2.8 Two sets of amendments to the PSMFL are broadly contemplated. The first set of amendments aims to increase the budget for local government community expenditures. The measure aims to provide resources for the provincial Governors' offices to meet emerging needs of constituents. For example, an amendment passedin mid 2008 gives local government authority to retain revenues from personal income tax. The 2009 proposedbudget provides new funding to local government for infrastructuremaintenance. A second set of amendments aims to: (i)drop provisions relating to output based budgeting and service delivery reporting; (ii) drop performance contracting. 2.9 In the next sections, this chapter considers progress on budget planning, execution, reportingand transparency, and related issues. These sections will highlight issues pertainingto PSMFL and will lead to recommendations for aligning some provisions of the Law with: (i) implementationcapacity; and (ii) cost-benefitanalysis. Box 2.2: A Summary of the Public Sector Management and FinanceLaw Key Provisions Key Provisions Status of Implementation Recommendations A. Centralizationof Implemented.Ministry of Finance is Ensuretransparency in thebudget Intergovernmentalfiscal management fully responsiblefor Treasury planningprocessso that local framework functionsand revenuepolicy and government level is aware of the administration. resourceplans and flows. B.MediumTermFiscalFramework Instituted.Governmentsubmitsthe Developrobust mediumterm sectoral (MTFF) adoptedto guide budget MTFF to Parliament in May. strategiesto inform sectoral planning Parliament discusses and approves allocationsas per the MTFF. aggregate macro-fiscalceilings. Governmentprepares annualbudget basedonthese approved ceilings C. Centralizationof policy Sectoralpolicy-makingis now under Enhancehorizontalaccountability formulation thejurisdiction of respective sector within vertical controlstructure. ministries.However, formal rolesfor local government in policy making are not clear. D.Performancemanagement The Governmenthas implemented a Roll-out strategic business plans introducedacrossthe public sector performancecontract system along alongfunctional lines, as with Strategic BusinessPlansfor recommendedin Box 2.1. each budgetentity. E.Full accrual accountingmandated The Governmenthas implemented Considerthe efficacy of budget modified accrualaccounting- entity-level balancesheets, in view of commitment accounting. but no fixed limited capacity. asset valuation is in place. PUBLICFINANCE PLANNING 2.10 Since the promulgation of the PSMFL, the government has made progress toward improvingthe budget planningprocess. This section analyzes developments inbudget planningin the mediumterm context, as well as inthe annual budgetprocess. 29 Mongolia: Public Expenditure and Financial Management Review Fiscal planning-medium-termto annual 2.11 By far, the most important and positive development was the successful introductionin 2003 of a three-year MTFF,which provides strategic guidance into the annual budget process. The MTFF consists of a medium-term macro-fiscal outlook with revenue projections, an aggregate expenditure forecast by economic classification and the corresponding domestic and foreign financing. Based on the macro-economic projection and structure of revenuepolicy, the Ministry of Financeestimatesdomestic revenue and collects information from development partners on ODA for the medium term. Using these estimates, the Ministry of Finance allocates expenditure ceilings, with the primary objective of maintaining macro-fiscal stability. Inrecent years, the MTFF has set an overall budget deficit cap of 3 - 4 percent of GDP annually. 2.12 The budget preparation timetable is clearly laid out in the PSMFL and the Government has followed this timetable consistently. The Government submits the MTFF to the Parliament at the beginningof May each year, as stipulated in the PSMFL. The Parliament reviews and approves the aggregate macro targets and the overall budget deficit target. Upon approval (at the end of the spring session) by the Parliament, the Ministry of Finance starts preparing the annual budget in consultation with sector ministries and provincial governments. Once the aggregate expenditure ceiling is set, ministries and agencies can plan individual budget requests. The Ministry of Finance cannot breach the overall budget deficit target set annually by the Parliament, unless Parliament amendsthe budget. The Minister of Finance has the authority to reallocate funds within the aggregate ceilings, but not to exceed the overall ceiling without prior approvalfrom Parliament. 2.13 Parliament reviews the MTFF and approves the aggregate expenditure ceiling, the overall deficit target, and the macro-economic parameters. After this approval, the Ministry of Finance commences discussions with sector ministries and provinces for finalizing the annual budget.The Ministry of Finance submits the draft budget proposalto the Cabinet in the middle of September. The Cabinet then considers the budget proposal, directs the Minster of Finance to make necessary amendments, and submits the revisedproposalto the Parliamenton October 1. The Parliamentthen considers the proposaland approves the budgetby December 1. 2.14 The Government is putting in place a performance framework into the budget cycle. Performance budget proposals from the Ministry of Education, Culture and Science and Ministry of Social Welfare and Labor were piloted in the 2008 budget, and extended to all line Ministries in the 2009 budget preparation. Under a performance budget framework, resources are no longer based exclusively on inputs (e.g. salaries, operations and maintenance, capital expenditures) but based on programs which are expected to deliver planned outputs (i.e. public goods and services). The following steps are requiredto quantify the output: i.Activitiesareidentifiedforeachoutput; ii.Anindicatorofperformanceisdefined; iii.Aquantitativeorqualitativetargetfortheperformanceindicatorsisset; iv. Expendituresare plannedfor each activity accordingto the desired output target; v. Unitcosts are estimated by dividing plannedexpenditures by the quantifiedtarget. ISource: Manualofproceduresfor the Medium-TermBudget Preparation,June 2008, Ministry of Finance, Mongolia 30 Chapter 2 -Public Sector Financial Management 2.15 The introduction of the MTFF has been a positive step. The Parliament is now able to discuss overall policy issues and approve the aggregate budget ceilings three months before the annual budget is discussed by Parliament. There are however, certain areas that need to be further strengthened to improve the efficacy and effectiveness o f the MTFF. 2.16 There is a need for robust sectoral strategies that would guide medium term sectoral resource allocations under the aggregate MTB level. The sectoral strategies are not adequately sequenced, prioritized or credible. Consequently, sectoral allocations are mostly incremental increases over the previous year's actual expenditures. There are currently master plans for education and the health sector, and the respective agencies are now converting these master plans into medium term work plans. Once this is accomplished, it will be possible for the MTBto naturally incorporate sectoral strategies into the medium term resource allocations. Inthe education sector, the government is piloting aggregate ceilings for categories o f expenditures, which allow sector ministriesto allocate funds into component programs. 2.17 High-level strategies should be harmonized to better inform medium term resource allocations. Currently, Mongolia is developing a long-term National Development Strategy and an MDG-based growth strategy; these will be added to the list o f high-level strategy documents, which includes the Socio-Economic Guidelines and the Government Action Plan. All o f these strategies aim to strategically inform budget allocations. However, developing this many strategies taxes government capacity and causes the MTB to be pulled in too many `strategic' directions. The result is that the status quo i s maintained, with the MTB providing incremental allocations to sectors. The OECD 2006 Survey on Monitoring the Paris Declaration also indicates that the links between policy priorities and the budget are weak. The key government policy documents should be: (i)the Government Action Plan; (ii)the Socio-Economic Guidelines that translate the GAP into annual and rolling action plan; (iii)the MTBF which essentially sets the corresponding medium term (N+3) budget framework; and (iv) focus on developing medium term expenditure frameworks for sector ministriesthat take into account the Government Action Plan, Socio-Economic Guidelines, medium term budget framework, and then translates them into detailed sector expenditure plans. Currently sectoral medium-term expenditure frameworks are missing. Box 2.3: Why a centralized public Budget management framework is appropriate in Mongolia Back in 2002, some of the most important recommendations for Mongolia as highlighted in the Bank PEFMR at that time were to regain control over public resources by "recentralizing provisions of all public goods", "introduce top-down budgetaryprocess", and "centralize policy-making role to line ministries". Since 2003 the budget management framework in Mongolia has been centralized. This means that the Ministry of Finance is responsible for collecting almost all revenues and providing approved budget appropriations directly to service delivery units. These service delivery units execute their budgetsbasedon approvedplans via the centralizedbut execution system (GFMIS) via respectiveprovincial treasury offices. Large budget entities and more that 250 districts (soums) have been connected in real time to GFMIS via district treasury offices. With regards to this centralized collection of revenues and management of expenditures, since 2003, the following improvementshavebeenachieved: - Resources are reaching service delivery units in a timely mariner and budget discipline has been -achieved;on Arrears goods and services, which were prevalent in the decentralized system prior to 2003, have -byCentral and large beeneliminated; government can transfer resources across provinces with ease and in real-time via the Treasury Single Account; Box 2.3 continuesI.. . 31 Mongolia: Public Expenditure and Financial ManagementReview ...I Box 2.3 - Government is able to determine its consolidated financial position daily and ensure all necessary fiduciary regulations are complied with becausecentralizedbudget execution systemhas this ability; The PEFMR analysis argues that Mongolia should maintain this centralized budget management system inorder to capitalize the above listed improvementsas well as on the following additional grounds: - Most of the population and correspondingpublic service needs are concentratedin few urban areas. In most of the districts transactions are fairly limited to justify the high cost of decentralizedbudget -execution. Managing the Treasury Single Account effectively is optimally done via a centralized budget execution system, managedby a vertical national treasury department structure that exists in Mongolia. A decentralizedbudget execution would imply a more complex and expensive technology as well as significance improvements in skills and technology at the local level to ensure that the TSA is effective managed, while benefits would be limited and upside risks of the system not being properly -maintainedinwould be greater. Already the current GFMIS, large budget entities and more than 250 districts are executing expenditures in real-time and this means that service delivery units have direct control over entering expenditure requests, while the respective treasury department only cross-checks the budget -availabilityand documents submitted prior to processingthe transaction; Ensuring that there is appropriate central government oversight and timely consolidation of government statements requires a more centralized budget execution system that can provide such reports inatimely manner. Source: WorldBank Staff 2.18 Local government participation in budget planning could be enhanced. The PSMFL has effectively centralized the intergovernmental fiscal relations framework. At the same time, policy making has also been centralized with the sector ministries, as demonstrated by the budget planning process.2 Budget entities (service delivery units) prepare their annual budget requests and submit them to provincial departments. Provincial departments consolidate these budget requests and send them up to the respective sector ministries. The sector ministry then consolidates all provincial budget requests and presents them to the Ministry o f Finance. The Ministry o f Finance then allocates funds to each budget entity and passes the funds through the respective sector. Funds are controlled by the head o f the budget entity and disbursed through the regionaltreasury offices. Local government3authorities have negligible input into the process and are often critical o f the central government's budget allocations on the basis that they do not reflect the needs o f provincial citizens. There is some validity to this argument, and it is recommended that the central government institute mechanisms to include more formal consultation with provincial authorities on the rationale of the budget plan. 2.19 A first step towards strengthening horizontal accountability between local government and service delivery units could be amending the PSMFL to include: (i)annual assessment reports o f service delivery units from the provincial government to the parent ministries; (ii) mandate publication o f performance reports of budget entities at the provincial level; and (iii) mandate publication at the local level total budget allocations and disbursements to service delivery unitsfrom central governments. There is no constitutionalrequirement for the proportionof budget distributionfrom the centralgovernmentto the local governments. For adescriptionof intergovernmental fiscal relationssince the PSMFL implementation,see the World BankPolicy note, "Mongolia RegionalDevelopment" (2005). Available at http:/lsiteresources.worldbank.org/MONGOLIAEXTNIResources/MNG_Policy_Note_Reg_Devt_Final_eng.pdf 3 Local Government comprisesprovincial (Aimag) and District(Soum) administrations, with the former being responsible for the latter. 32 Chapter 2 - Public Sector Financial Management Box 2.4: Budget preparation at the Municipality of Ulaanbaatar (MUB) The budget preparation process in MUB is driven by the Budget Law and the PSMFL. The responsibility for preparation and issuance of budget guidelines is with Department of Finance, Economy and Treasury (FE&T), MUB. The FE&T department receivesthe budget guidelines from MOF by June 20. The guidelines are essentially the policy statements that provide departments with a sense of the budgetary environment for the forthcoming year. They provide, inter alia, the details of all budget input formats, and they call for compliance. All budget entities in MUB must submit their strategies, investment plans, and financial statements by July 10. By October 1, FE&T consolidates the budget plans and the Parliamentapprovesthe submittedbudgetby December. An ongoing debate on budgeting in MUB revolves around budget decentralization from the Central Governmentto MUB.As statedearlier inthis chapter, budget planning is done by sector ministries and the Ministry of Finance, and this has caused problems for MUB. For example, sector ministriesare often not aware of the total number of people residing in Ulaanbaatar and therefore under-budget resources-ven though MUBhas this information readily available. There is little flexibility inreallocation of resources. In order to reallocate resources from one sector to another, MUB must go through a cumbersome process of obtaining the approval of the respective sector ministries. However, MUB does not seem to have the capacity for effective budget planning. Therefore, it is recommended that MUB capacity for budget preparation be enhanced, and that the central Government introduce formal consultative mechanisms to ensure that sector ministries to have adequate information to make effective allocations. Source: WorldBank Staff 2.20 At present, the Government has not started preparing a statement of fiscal risks, as mandated by the law (Article 32.2.3). Information on contingent liabilities remains incomplete and a clear analysis o f fiscal risks has not been prepared. Such a statement is important for the Parliament to consider when reviewing the budget proposal. However, this requires both capacity strengthening and formulation o f a framework for monitoring and recording contingent liabilities. 2.21 There is a need to improvecoordinationand integrationof ODA. ODA is governed by the Law on Coordination of Foreign Loans and Grant Aid, promulgated on June 13, 2003. This Law vests the authority to decide on all loans and grants to the Minister for Finance. In practice, foreign loans are channeled through the Ministry of Finance, and are fully recorded by the Debt Management Division at the National Treasury. Foreign grants, however, are not coordinated through the Ministry o f Finance because grant assistance is often discussed directly with respective sector ministries, and funds are released to the Ministry o f Finance once the relevant sector ministry approves the grant. There is no structured process under which the Ministry o f Finance can learn of and record such assistance. The consequence is that the medium term budget framework fails to capture all the information and cannot effectively allocate resources based on cumulative financial flows to respective sectors. 2.22 Although the Government has embarked on an ongoing medium term budget planning exercise, it i s generally not possible for development partners to: (i)align their support to the Government budget calendar because different donors have different fiscal years; and (ii)make multi-year commitments, because most donors work on single year appropriations. These problems result in fragmented planning between the recurrent budget, which is mostly financed by government's own resources, and the capital budget, which is largely financed by ODA (about 80 percent). There is a need for donors to align themselves to the Government budget calendar and requirements. 33 Mongolia: Public Expenditure and Financial Management Review Box 2.5: Good practices on budget deliberations The United States: After receipt o fthe President'sbudgetrequest on the first Monday inFebruary,the Congressional Budget Committees hold hearings to receive testimony from Administration officials, experts from academic and business communities, representatives of national organizations, members of Congress, and the general public. During the same period, the other committees in Congress review the sections of the President'sbudget submissionthat relate to programswithin their jurisdictions.Inaddition, during February, CBO sends annual reports on the budget and economic outlook to the Budget Committees.InMarch,CBO sends its report analyzingthe President'sbudgetrequestto the Appropriations and Budget Committees.At the same time, the Budget Committeesdraft the Budget Resolution,which is a formal and bindingagreement betweenthe two chambers on total spending, receipts and other aggregates, along with a division of spendinglevels among budget functions. When the Budget Committeescomplete their mark-up of the Budget Resolution, they report their respective resolutionsto the full Senate and full House.The BudgetAct sets April 15 as the date for completionofthis work. Australia: Afler the Government introduces the Appropriation Bills and presents the Budget Papers and, the process allows for one to two months of budget deliberation by the legislature. There is no committee stage in the House of Representatives,which has decisionauthority over the budget.The Senate EstimatesCommitteesscrutinize the Appropriation Bills and other budget documentation, particularlythe PortfolioBudget Statements. The basic function of the Committee is to require the presence of, and seek explanations from, the Ministers who formulate policy and Departmental officers who implement policy. Eachofthe EstimatesCommitteestakes responsibility for a number of agencies. The EstimatesCommittee process is generally finished intime for Parliamentto passthe Appropriation Bills beforethe endofJune. Thailand: Budget deliberations start with the first reading of the budget, which is called a vote-of- confidence in the government. If the vote is won-as i s normally the case-a Scrutiny Committee is selectedto examine the government's budget proposal. In the 105 days has available for considerationof the budget,committeemembersmeet every workingday. The Committeemeets inplenary inthe beginning of the process, in order to consider the overall work program for examiningthe budget proposal, and at the end when it makes its final recommendationsto the House. Source: OECDJournal, Volume 3. Number. 2/ Volume 5. Number 3/ and Parliament of Australia, retrieved at http://www.aph.gov.au/Librarv/aubs/ra/2002-03/03RP06.htm#overview Box 2.6: Wells in Mongolia To be or not to be - The government has a policy to ensure that there are two wells per district. Basedon informationon well constructionby developmentpartners across districts in Mongolia, the Government makes a budgetary allocationto districtsthat do notreceive support from developmentpartnersfor constructingwells. In 2006, the government reviewed the program and found that there were districts where there were more than two wells, and districts where there were no wells. This was a puzzlingoutcome, because the government had carefully allocated funds to make sure that all districts had funds, either from the State Budget or from development partners. Government analysis revealed that during the year some development partners hadchangedthe districts they were supportingwith well construction. Because there was no feedback to the Ministry of Finance,the government hadcontinuedmaking allocations accordingto the originalplans. Source: WorldBank Staff 2.23 The centrality of the MTFF in informing budgeting should be enhanced. At the moment, new policy initiatives are approved directly by the Parliamenteven when they have not been vetted through the MTFF. This weakens the sanctity of the MTFF process and provides incentives for by-passingthe formal MTFF process. Although it is the prerogative of Parliament members to propose and approve spending initiatives, it is prudent to have a framework in place that allows the technical managers of government to review those policy measures and inform 34 Chapter 2 -Public Sector Financial Management decision makers about the implications o f such proposals. Additionally, it is critical that all proposals be reflected in the MTFF, as this is the main instrument for allocating scarce resources through strategic decision making. 2.24 Revenue projection could be improved ...During the last five years, actual revenue collection has exceeded projections by an average o f 13 percent annually. This has largely been due to higher than expected mineral production and commodity prices. However, the government should be able to factor these developments into its revenue forecasting model. Sound revenue modeling generally aims to project revenues such that actual collections are within five percent o f budget estimates. 2.25 ...Ascould expenditure estimation.Infour o fthe past five years, the government has had to present mid-year budget amendments. Most o f the reasons for budget amendments have been: (a) increasing civil service salaries; (b) under-projection o f new social transfers such as the marriage allowance; and (c) approval by Parliament o f new expenditure programs during the fiscal year. According to the data on FY2006, the main reasons for amendments were increases in basic salary, pension and benefit insurance, expenditures on goods and services, and an underestimate o f social transfers. Also, many new programs were approved after the first budget appropriation was adopted by the Parliament. Mid-year budget amendments: (i)weaken the credibility o f the annual budget process; (ii)focus the limited capacity o f the Government on immediate financial issues, rather than on proper planning for the following year budget estimates; and (iii)provide incentives for agencies to have programs approved by Parliament at artificially low levels, and then to request increases later, as the magnitude o f resources needed becomes apparent. There is significant need to improve budget forecasting ability, on both the revenue and expenditure sides. Fiscal planning the PublicInvestmentPlan - 2.26 The Public Investment Plan (PIP) prepared by the Ministry of Finance guides medium term capital budgeting.The PIP is attached to the MTFF document for consideration by the Parliament. Notionally, the PIP is a strategic capital budget planning tool that brings together capital expenditures financed by the State Budget and ODA into a coherent, consistent and prioritized framework. Over the last three or four years, ODA has financed an average o f 80 percent o f the PIP annually. Since end-2006, the responsibility for preparing the PIP has been vested with the Fiscal Policy Department o f the Ministry o f Finance. 2.27 Preparing the PIP in conjunction with the MBTF and vesting both of these responsibilitieswithin the same entity lays the institutionalfoundation for ensuringthat: (i) the PIP and MBTF are well integratedand consistent; (ii)adequatemaintenancecosts have been factored into the recurrent budget for maintaining assets. At present, the PIP is effectively a list o f projects to be financed by the State Budget and ODA. There is little strategic prioritization or formal linkage with sector strategies. Cross sector collaboration on PIP formulation is weak, capacity to analyze project cost-benefits is inadequate, and ODA does not seem well-aligned with overall Government strategy. Based on discussions with authorities, the following issues have been identified for improvements in order to enhance the consistency and effectiveness o f the PIP. 2.28 Enhancing capacity for conducting cost-benefit analysis and prioritization of the PIP is necessary. In order to determine the financial viability o f capital investments, there is a need for trained staff to analyze the cost-benefit trade-offs for capital investment, and to ensure net positive financial results in a sustainable framework. At the moment, there are only about five staff members at the Ministry o f Finance tasked with preparing the PIP. This seems inadequate 35 Mongolia: Public Expenditure and Financial ManagementReview for preparing the PIP. Therefore, it would be advisable for the Government to conduct a needs assessment to determine the staff and skills needed for preparing a robust PIP. 2.29 Better inter-ministerial coordination on strategic prioritization of investment expenditures is also crucial. Currently, sector ministries submit lists o f projects that are needed at the sector level to the Ministry o f Finance, and dialogue simultaneously with donors to obtain resources for financing these projects. The Ministry o f Finance does not have the capacity to analyze all o f the requests and there is no institutional mechanism for sector ministries, the Ministry o f Finance, and development partners to come together to discuss the proposed capital expenditure plans. This coordination issue i s further exacerbated by the fact that donors often provide grant assistance directly to the sector ministries, without the formal signature o f the Minister o f Finance. This is in contravention o f the Law on Coordination o f Foreign Loans and Grant Aid, which stipulates that all ODA needs to be formally signed and approved by the Minister o f Finance. Although the Ministry o f Finance has a fully functional debt management and recording system, the same does not exist for grant assistance. 2.30 Making the PIP the sole entry point for capital projects ... At present capital projects are financed even when they have not previously been included in the PIP. This weakens the credibility o f the entire PIP exercise and provides incentives for agencies to by-pass the PIP process. The Government should consider making it mandatory for projects (external or domestic) to be included in the PIP for the respective year in order to receive financing. 2.3 1 ...and improving data on ODA by sector and category of expenditure. The first step towards improving PIP planning is to have a systematic way o f recording ODA. The current situation results in an overlap o f expenditures and a lack o f accountability for results. It is recommended that the government consider establishing a forum for discussing ODA management issues with donors, and also establishing an inter-ministerial working group to discuss the proposed capital investments in each sector, keeping in view the costs and benefits o f projects, sector development strategy, and financing options. Box 2.7 presents a brief description o f how a PIP i s prepared in Chile. Box 2.7: Good Practice on investment program - Chile Chile has accumulated over three decades of successful experience in the systematic appraisal of public investment. Modem project appraisal for public investment started in 1975, with the establishment ofthe National System of Investment(SNI) at the Ministry of Planning (MOP), currently administeredjointly with the Ministry of Finance (MOF). As to the appraisal, every public-investment project is subject to the same cost-benefit analysis, under a set of clearly specified methodologiespublished by the MOP. The law mandates that the capital budget sent by the MOF to Congress can only included projects within the SNI, which have also been favorably assessed by the MOP. This process screens out `white elephants,' and also rules out the possibility of investment projects entering the budget at the congressional stage-the only entry door is the SNI. The SNI, thus, is a main determinant for a successful public investment in Chile. It refers to a set of norms, techniquesand procedureswhich govern the public investmentprocess. The objective of the SNI is to improve the quality of the public investment, by selecting the projects with the largest net present value-and thereby increasingthe country's net worth. The maintools of the SNI are: (1) Methodologies for preparing and assessing projects from a coherent socio-economic perspective- including a specified set of shadow prices, and a social discount rate (currently 8 percent). (2) Continuoustraining of officials to develop adequate technical capacity at all levels of government and inall regions. (3) A publicly available integrated data bank of projects (BIP), enabling efficient and coordinated public- finance managementand well-informed policymaking. Source: The WorldBank, Appraisal of Public Investment, December 8, 2006. 36 Chapter 2 -Public Sector Financial Management 2.32 Ensuring there is appropriate funding in the budget to finance recurrent cost of investmentprojects. Since 2003, there has been an increase in investmentwhile expenditure on capital maintenance has been reduced (see Figure2.1). This is a source of significant concern, as capital investments deteriorate faster if they are not maintained Figure 2.1: Mongolia total investmentand adequately. In the absence of a capital maintenance methodology on depreciating assets, adequate provision for Mongolia: Total Investment and Capital Maintenance, FY03 FY06 - maintenance becomes yet more 180,000 critical. Maintenance costs should be explicitly presented 150,000 with each investment project, so 120,000 that decision makers can consider the maintenance costs 90,000 of projects during the approval 60,000 process. 30,000 2.33 A first step towards 0 strengthening the PIP could 1 FY2003 FY2004 FY2005 FY2006 be to constitute an inter- ministerial working group led +Total Investment +Capital Maintenance by the Directorate General for fiscal policy department to: (i) Source: MOF determine the medium fiscal space taking into account the budget momentum (i.e. medium-long term committed expenditures); (ii)investment needs across the sectors; and (iii)the government action plan.This inter-ministerialcommittee should be tasked to prepare the medium term policy consistent PIP that would be discussed at cabinet and the parliament. This PIP once approved would be implemented over the mediumterm with annual reviewof progress and challenges.And when the Government Action Plan changes, the PIP would change appropriately.With this requirement in mind, the government should commission a needs assessment for capacity and financial resource requirement for such a task to be done. EXECUTION OFPUBLICFINANCES 2.34 Before 2003, Mongolia's budget execution system was fragmented and lacked any effective control. The Treasury Department operated a basic, locally-developedbankinginterface software calledBUMIS.There were more thantwo thousand bank accounts, operated by different agencies, and it was not possible to determine the Government's consolidated cashpositionat any point. Budgetreports were consolidated with alagofmore than 3 months 2.35 Mongolia has made significant headway in improving the budget execution system. The PSMFL mandated the Minister for Finance to establish centralizedcontrol over all government bank accounts and to taken on the responsibility for centralized cash management. Since 2003 the Government ofMongoliahas beenable to: Establish a functional Treasury Single Account framework, which entailed closing more than 2000 individual government Bank accounts and consolidatingthem under one umbrellaaccount, managedandcontrolledby the Ministry of Finance; 0 Implement unified IMF Government Financial Statistics, a consistent Chart of Accounts and budget nomenclature.This entailed revising the fragmented accounting 37 Mongolia: Public Expenditure and Financial Management Review structuring to fully integrated one and effectively training all budget agencies on its effective usage; 0 Operationalizean integrateddebt management informationsystem.The Government established a dedicated Debt Management Division (DMD) within the Treasury Department to managepublic debt. The DMD has establisheda consistent debt recording and monitoring system. Consequently, the discrepancies in debt information between creditorshave beenreconciledadequately. 0 Deploy a country-wide electronic payment settlement system at the Bank of Mongolia. This payment system manages the Treasury Single Account across the country in real time and ensures smoothtransfer of funds to provincesand municipalities. 0 Institute a fully integrated government financial management information system (GFMIS) to execute government expenditures real time. This system reconciles expenditures made at the Treasury with the funds releasedby the Bank of Mongolia on a daily basis. The system also has undeletable audit trains for each transaction. For additional details, please refer to Annex 2B which describes the findings of the GFMIS compliance check and internalcontrol system review, together with recommendations for improvements. 2.36 The aforementioned reforms in budget execution are commendable in every respect.It has taken many countries with higher capacity levelsa much longer time to implement such reforms.Table 2.1 shows how Mongoliafairs in comparison with other East Asian countries with regardsto core public finance managementsystems. Table 2.1: Regional public finance managementsystem component comparison PFMComponent Elements Mongolia Cambodia Vietnam Indonesia Philippines OperationalTreasury Single Account System Yes No No No No IntegratedChart of Accounts consistent with International accounting Standards Yes No Yes No No IntegratedBudget and Treasury System Linked to Yes No No No No the Central Bank UpdatedDebt Management System Providing Input Yes .... Yes Yes .... to the IntegratedTreasury and Budget System Treasury System Linked with the Revenue Yes No No No No Collection Agency(s) Source: WorldBank Staff 2.37 The results of these reforms have been: 0 Arrears on goods and services are under control.Beforethe Treasury SingleAccount and GFMIS were established, wages and salaries were often in arrears for more than two months. Since these reforms, the government has remainedcurrent on such payments; 0 Governmentconsolidated TSA balances are in excess of US$65 million. Before the TSA was introducedin 2003, the government hadto resortto borrowingfrom the central bank to meet its daily expenditure needs because government cash holdings were fragmented across many accounts; 0 Fundsare reaching budget entities on time and as per appropriations; Improved controls over budget execution have reduced leakages, and contributedto the positive budgetbalance; 38 Chapter 2 - Public Sector Financial Management Mongolia is now able to providefiscal reports consistently, accurately, and on time, consistent with the IMF Government Financial Statistics. 2.38 A GFMIS compliance check concluded that the business processes and internal controls for GFMIS were adequate for financial management purposes. With the implementationof GFMIS on January 1,2005, there was a needto conduct a compliance check to examine the extent to which the business processes and internal controls in the FreeBalance e- Financials software were working in practice. The detailed findings are reported in Annex 2B. This will assist the Ministry of Finance (MOF) to fine-tune GFMIS functionalities. The internal control review relates to ensuring that all underlying funds and information flow processes in GFMIS are also working. 2.39 GFMIS controls around budgetary, system integrity, payment security, and system management should be improved upon and key internal controls should be reviewed, at least annually, to ensure they are still functioningas designed. To do so, the systems audit report from the recent Mongolia National Audit Office of GFMIS should be reviewed and respective regulationsand system-fixes should be developed to addressthose concerns. 2.40 These findings may increase external donor confidence in using GFMIS for their portfolio management.By usingGFMIS for donor portfolio management,the donor community would address one of the key areas from the Paris Declaration on Aid Effectiveness (March 2005), "Strengthening andUsingCountry Systems." Box 2.8 below discusses this subject further. 2.41 Donor funds are essentially managed outside of the Government's own PFM systems, resulting in fragmentation of donor programs, disconnect in project accountingwithin the national budget, and over-burdened staff. This practice also undermines public expenditure management systems. In addition, project management/implementing units (PMUs or PIUS), which are outside of the government structure and funded by projects, have been created to managethese donor projects.Not only do these unitsrequireadditionalresources, they also divert highly skilled talent from the government, since PIUs pay higher salaries than the civil service. Reliance on GFMIS for donor portfolio managementwould mean that instead of openingspecial accounts for each project at commercial banks, donors would disburse funds directly to the Treasury Single Account at the Bank of Mongolia. GFMIS would be used to execute eligible project expenditures, account for the transactions, and generate project related reports on funds usage directly from the system. Audits would also be conducted by the MNAO as part of its overallaudit of government accounts. 2.42 The Government could develop a Government-DevelopmentPartner action plan with steps and timelines for integratingODA into countrysystems, and move towards more sector-wide approaches learninglessonsfrom the successful donor coordinationinthe education sector. To this end, A MoF working group could be constituted by order o f the Minister responsible for working with Development Partners to develop a phased plan for closing all special project accounts held outside the TSA and to migrate them to GFMIS. This working group should by led by the Treasury Department with representation of all concerned departments. The government should pre-announce this intentionto all development partners and determine specific dated milestones for this migrationplan. 39 Mongolia: Public Expenditure and Financial ManagementReview Box 2.8: Use of country financial management systems In the World Bank, use of country systems generally refers to the Bank's reliance on the borrowing country's systems for the safeguards and fiduciary aspects of operations it supports. Use of country F M systems can be defined as the use of a country's normal institutional arrangements and processes for financial management. The main components of a country's fmancial management system include: (1) Budget preparation; (2) Budget management (execution); (3) Accounting; (4) Treasury management; (5) Debt management; (6) Internal control framework; (7) Management reporting; and (8) Auditing. For treasury management, using country systems involves using the financial systems of the government - regular systems for banking, payment authorization and payments. For internal controls, it involves using the country's regular rules, procedures, manuals and internal control framework. The Bank's merational Policy (OP) 10.02, Financial Management,provides that Bank-supportedoperations should use the country's own PFM system, if it is acceptable, as the default option for financial management. Moreover, Bank managementhas committed to making Mher progress on the Paris Declaration targets for aid effectiveness, which include a reduction in aid flows outside partner countries' PFM systems. There are many benefits to using country systems. Country systems: facilitate the comprehensiveness of resources administration; avoid isolating projects from other governmentprograms; enhance capacitybuilding; 0 increase country ownership and project sustainability; facilitate harmonization and simplification; and reducecosts for Borrower and Bank. Reliance on GFMIS for donor portfolio management would mean that instead of opening special or designated accounts for each project, donors would disburse funds directly to the Treasury Single Account at the Bank of Mongolia. GFMIS would be used to execute eligible project expenditures and reports generated directly from the system on funds usage. In addition to the execution and reporting function, the auditswould then be conducted by the MNAO as part of its overall audit of governmentaccounts. Source: WorldBank Staff 2.43 Mongolia has implemented the first generation of reforms to improve budget execution. The next challenges are to: Integrate ODA execution into the Government budget execution processes. At the moment, most ODA i s executed through extra-budgetary accounts set up outside the government's TSA. This practice fragments budget execution systems and dilutes accountability for and reporting on the resources provided by development partners. It is recommended that the Government discuss with development partners an action plan for migrating O D A into country systems. This would then set out a roadmap for the Government and donors. 0 Roll out budget execution responsibilities to budget entities. Currently, the primary responsibility for budget execution is with the National Treasury. Budget entities prepare expenditure requests and related forms, and then have to send all the documents to the Treasury Department for processing within the Government Financial Management Information System. This adds to transaction costs for budget entities and also overloads the staff at the Treasury department, who have to enter transactions into the system. Rolling out budget execution responsibilities to budget entities would be consistent with sound principles o f budget execution. The Government i s already in the process o f rolling out GFMIS to large budget entities and districts, and this will allow: (i)budget entities greater control over spending; (ii)Treasury department to take on a more active oversight role and focus on issues pertainingto cash management. 40 Chapter 2 -Public Sector Financial Management PROCUREMENT4 2.44 Mongolia has made significant efforts in improving public procurement. In 2000, Mongolia promulgated a procurement law. This law centralized procurement and tightened fiduciary controls. However, the centralizationof procurement increased procurement lead time (causingdelays) and budget entities felt disenfranchised at the loss of control over funds allocated to them by Parliament. The main provisions of the law were generally consistent with good international practice (as compared with the UNCITRAL Model Law on Procurement), but the law contained too much detail for basic enabling legislation. The PPLM at that time included proceduraldetails of procurement practicethat would normally benefit from periodic amendment to reflect experience gained and changes in dynamic factors such as exchange rates and the country's rate of economic growth. Such proceduralamendmentswould require the PPLM to be regularly amended which would then subsequently require Parliamentary approval. Regular amendmentto the PPLM couldthereforeproveto be both burdensomeand time consumingto the GoMa5In order to improve procurement efficiency, while maintaining necessary fiduciary controls over public procurement, the Parliament passed the revised Procurement Law in December2005. 2.45 A new PPLM enacted in February 2006 was a commendable improvement over the previous law. The law was more fully aligned to best international practice, as many of the procedural details were removed. The PPLM also dictated that the PPCD issue supplementary regulations, guidelines and other documentationand introduceda complaints review mechanism. Under the new law, the Procurement Policy and Coordination Department(PPCD) of MOF has received a mandate for advising procurement entities on effective practices, overall monitoring and reporting of public procurement performance. The new law has introduced many improvementsinthe legal and regulatoryframework of public procurement.For example, the law decentralized procurement functions to line agencies, with PPCD's prior review of major transactions and complaint review mechanisms, and set requirements for reporting by line agencies to PPCD and MOF to Parliamenton public procurementperformance.To implementthe new law, PPCD has provided training on the new law to all line agencies and issued several implementation regulations in the past year. However, despite these efforts, there have been indications that the Government has had difficulties in implementing the law because of weaknesses inthe implementationarrangements. 2.46 A February 2007 amendment to the PPLM significantly altered the spirit of the law. The most important changes to the law (i)allowed direct contracting in certain sectors of government funded investment and (ii) significantly increasedthe thresholdbelow which foreign companies could be excludedfrom participatingin bidding for contracts. The introductionof this amendment deprived the Government of much-needed international competition in areas where there is limited domestic capacity and where there is a risk of creating a monopoly, which may result in higher prices and a lack of transparency. Specific commentary on the February 2007 Amendment and suggestedfurther improvements to the PPLMcan be found inAnnex 2C. 2.47 The institutional framework and management capacity for procurement are weak. The decentralizedapproach to procurement allows entities at the ministry, provincial and project level to conduct and manage their own procurement activities. The institutional This section draws both from a 2007 World Bank procurement survey and the 2007 OECD-DECjoint venture for Procurement - Results of the Benchmarking and Assessment of the Mongolian Public Procurement System. See the WorldBank ProcurementAssessmentReport(CPAR)conducted in2003 o f the Mongolian Public Procurement System. 41 Mongolia: Public Expenditure and Financial Management Review framework does not currently provide for the adequate control and monitoring of procurement activities and there is a lack of integration of these activities into the financial management system. 2.48 . Specific areas ofweaknesses are identified as follows: Procurement plans are prepared after the approval of the budget and thus are not used . as a means of calculatingthe requiredfunds but as a means of spendingthe budget after it is madeavailable; The current system for collecting procurement information and integrating it with financial information is improving, but still incomplete.There are procedures in place for procurement planning and reporting of procuring entities (MoF order dated 5 January 2007). The Government is in the process of developing tan integrated procurement management, monitoring and information system (PMMIS) which is ....expected to provide central and local governments with procurement management and monitoringtools, at projectand contract level. Safeguards in the financial management system are not adequate to ensure that procurement actions are only initiatedwhen funds are available; Completion reports are available but not yet integrated into a financial management system; There is a need for improvedreportingmechanisms and capacity development; There are no quality control or performance evaluation methods for procurement processes. 2.49 Lookingat procurementoperationsand market practices, there is a lack of formal qualification requirements for procurementofficers in the public procurementsystem. Very little progress has been made on analyzing the gap between the qualification and skill levels required and the qualificationand skill levels that currently exist within the public procurement entities. 2.50 There is also a lack of formal mechanisms for dialogue between the public and private sector regarding procurement. Evidence provided by the private sector suggests that public procurement entities are perceivedas weak in terms of planning, budgetingcapability and technical expertise.6The privatesector believes that this is contributingto ineffective competition for public sector contracts. 2.5 1 Turning to the accountabilityand transparency of the public procurementsystem, a robust audit framework is in place but control mechanisms are weak. Public procurement operations are audited by both the State Inspection Agency and the National Audit Office. Inspectionsand audits are carried out annually on a systematic basis. Weaknesses in controls are primarily attributedto lack of clearly defined procurement audit proceduresand inspectorshaving limited knowledge and experience with procurement best practice. 2.52 Public disclosure of procurement information is an essential component of demonstrating transparency and accountability and combating corruption. The PPCD has created a dedicated procurement website for disclosing informationon contract awards, bidding procedures, bidding documents and guidelines, but it is still in its infancy and is under-utilized. Furthermore, due to a lack of monitoringand enforcementcapability, not all invitationsto bid and contract award details that requirenotification on the PPCDwebsite are submitted for publication 2.53 In light of these findings,the followingrecommendations are suggested: See OECD-DECJoint Venture for Procurement- Results of the Benchmarkingand Assessment of the MongolianPublicProcurementSystem. 42 Chapter 2 -Public Sector Financial Management Consider withdrawing the February 2007 PPLM Amendment and incorporate further changesto the PPLMwhich would improve its clarity. Review ways to improvethe coordinationbetween budget departments andthe MOF to . ensure better integration of public procurement activities with the GoM's financial management systems. . Establishan appropriateprofessionalprofile for public procurement officers interms of qualifications, skill requirements, knowledge levels and status and provide access to education andtrainingto enable procurement officersto attainthe requiredprofile. Developaudit manualsand procedures for use by boththe State InspectionAgency and . the National Audit Office for the conduct of various audits (Le., compliance, performance and efficiency). Enforce the utilization of the PPCD website for disclosure of public procurement informationandthe publicationof invitationsto bid. INTERNALAUDIT 2.54 Internal audit in the Mongolian government primarily addresses compliance with rules and regulations. Audit inspectors' focus is primarily regulatory compliance (imposing penalties and giving instructions for reimbursement of money that has been wrongly applied) with no emphasis on the efficient or effective use of public funds. In light of the current lack of capacity within the external audit unit (see below), the GoM should place priority on building up the external audit capacity and then exploring ways to improve the GoM's internal audit capability. In the meantime, it is recommended that GoM should adopt control systems-based auditing, starting with the assessment of internal controls of the bodies which they inspect. Control systems-basedaudit should replace compliance inspectionas the dominant internalaudit methodology across the government. 2.55 The structure of the Government's internal inspection is not adequateas portfolio levels are not systematicallycovered. Although the PSMFL requires portfolio management at the line ministry and aimag levels, there is no formal internal audit structure in place within the portfolio level, which produces irregular oversight of portfolio financial management. Thus, the GOM should consider the establishment of portfolio level internal audit when adopting control systems-basedauditing. EXTERNAL AUDIT 2.56 Overall, the legislative framework is adequate for the MNAO to effectively audit the Government. The legislative underpinnings of external audit are containe'din the PSMFL Article 38 and further elaborated in the 1995 Audit Law. The responsibility for conducting external audit has been assigned to the MNAO. The Law mandates that the Government submit the financial accounts no later than six weeks after the end of the fiscal year. The main challenge for external audit in Mongolia is to build MNAO's staff capacity for conductingrobust financial audits of financial statements generated by the GFMIS. For instance, the possibility of migrating donor funds into GFMIS would requirethat MNAO be able to audit them. Audits of donor funds would no longer be conducted by external audit firms, as these outside firms would not be able to access government internal information system. The next level of capacity would be to develop skills for performance audits. However, in view of the limited number of staff, current skills mix, andthe critical fiduciary need, it is recommendedthat priority be given to strengthening skills for improvedfinancialaudit of GFMIS . 43 Mongolia: Public Expenditure and Financial Management Review 2.57 In order for the MNAO to be able to effectively conduct financial audits of the consolidated government account, the MNAO must have the capacity to conduct financial audits of GFMIS.At the moment, as GFMIS is a fairly new system, the MNAO is inthe process of building capacity to conduct GFMIS audits. Conductingfinancial audits within an information system environment requires skill-sets and audit methodologiesthat are currently not available within MNAOq7Furthermore, once the audit report is submitted to the Parliament, it would be advisable to publishthe report, with the audit opinion. 2.58 To strengthen MNAO capacity to conduct robust financial audits of GFMI, Parliament should provide sufficient annual budget for various training of MNAO staff and for MNAO to continue working on systems audit capabilities. MNAO could consider (i) having dedicated system audit teams to receive necessary continuedtraining to performthis task and, (ii)instilling periodic internal examination of MNAO staff for consideration of promotions or demotions to improvethe overallquality of MNAO. DISCLOSURE TRANSPARENCYPUBLICFINANCES AND OF 2.59 Important internal reportingrequirementshave been fulfilled, which is extremely useful for improving internal cash management and budget monitoring. Significant internal reporting requirements are mandated by the PSMFL (Articles 39-42). The reporting responsibilities of portfolio ministers, general managers of budget entities, and SOE's are extensive, but are mostly limited to internal consumption. The reforms to key budget execution systems have also made it possible for the Ministry of Finance to report the daily reconciledcash positionofthe government. Box 2.9: OperationalEvaluationCountry Departmentbest practices on budget transparency OECD Member countries are at the forefront of budget transparency practices. At its 1999 annual meeting, the OECD Working Party of Senior Budget Officials asked the Secretariat to draw up a set of Best Practices in this area based on member countries' experiences. The Best Practices are designed as a reference tool for member and non- member countries to use in order to increase the degree ofbudget transparency in their respectivecountries. According to the OECD Best Practices on Budget Transparency, budget transparency is defined as the full disclosure of all relevant fiscal information in a timely and systematic manner. The best practices are divided into three parts. Part 1 lists the principle budget reports that government should produce and their general content. Part 2 describes specific disclosures to be contained in the reports, including both financial and non-financial performance information. Part 3 highlights practice for ensuringthe quality and integrity of reports. OECD recommendation for best practice on budget transparency is to publicly disclose: (i) the pre-budget statement; (ii)executive budget proposal; (iii)the approved budget; (iv) in-year budget execution reports; (v) mid-year budget review and assessment; (vi) end of fiscal year budget report; (vii) report of the auditor general; and (viii) citizens budget, and the pre-election report and the longterm economic projection. Sources: OECDBudget Journal 2.60 The PSMFL,however, remains silent on disclosure of budget information beyond the Parliament. There is no legal requirement to provide information on public financial management to the public. Prior to 2006, disclosure of budget informationwas limited solely to the publication of the high level budget. Hardly any information was provided to the public. Mongolia scored very low on fiscal transparency indicators, such as the Open Budget Initiative, which were basedon availableinformationas of end-2005. 7 The World Bank is currently assisting MNAO on this issue and has provided an Institutional DevelopmentFund(IDF) trust fund to help MNAO address the matter. Under this trust fund, MNAO will work with specialistsinthe system'saudit arena to learn and buildthe necessary system's audit capacity. 44 Chapter 2 - Public Sector Financial Management 2.61 The Government could propose clauses in relevant laws for public release of budget information.Inaddition, the Minister for Finance could issue a Ministerial Instructionto all concerned departments reporting to the Ministry of Finance to: (i)list all laws, decrees, regulations, and guidelines that determine respective departmental operations andto publishthem on this MOF dynamic portal; and (ii)require departments to submit compliance reports on the same. 2.62 There are seven key budget documents that OECD Best Practices for Fiscal Transparency highlights as the most important for engendering accountability. These include: (i)the pre-Budget Statement, (ii)the National Budget; (iii)monthly budget execution reports; (iv) the mid-year budget assessment and economic review; (v) the end of fiscal year report; (vi) the report of the Auditor General; (vii) the citizens budget, pre-electionreport, long- term fiscal outlook. While considering the availability of these documents in Mongolia, this review considered two additional dimensions of national budget reporting: (i)the executive's budget proposal; and (ii)the enacted budget. This division is useful since citizens must be able to access the budget before it is approved in order to provideinput into budgetary policy making. 2.63 In 2006, the government made progress in disseminating some key budget documents and information relating to budget planning and execution. These key budget documents are reviewedbelow. 2.64 The Pre-Budget Statement. This document is prepared by the Ministry of Finance, and approved as a parliament resolution. By law, this document should be submitted to the Parliament by May 1 and approved during the spring session of the parliament. The budget framework statement should provide the basis for determining the general socio-economic development directives and the state budget. The fulfillment of the budget framework is the responsibilityof general budget governors. 2.65 In its approved form, the budget framework statement is a brief document with macroeconomic indicatorsshowing the limits that should be used in preparingand approvingthe budget for the next three years, along with the medium term objectives of the government. The draft budget framework is a much more detailed document, with current macroeconomic analysis and budget projectionsfor three years ahead, includingbothrevenues and expenditures. 2.66 For example, the 2008 budget framework was discussed by the Cabinet on April 24, 2007, This document is availableon the Parliamentof Mongoliawebsite, and on request from the Ministry of Finance. However, the document is not yet available on the Ministry of Finance's own website. 2.67 Executive's budget proposal. This is arguably the most important document for the public. For citizens to effectively participate in the budget process, and to influence budget outcomes, they should be able to access information about budgets before the budgets are approved. The executive's budget proposal is prepared by the Ministry of Finance, discussed by the cabinet, and must be submittedto the Parliamentbefore October 1". In 2006, the proposal for the 2007 budget was made available to public through a government run website (www.open- government.mn) in early November. It is important that this document was released at the same time it was submitted to the parliament, since by November the parliamentwill usually have had two rounds of discussions, and made most ofthe importantchanges. 2.68 Furthermore, the document that is posted on the web is in PDF format, making it difficult for analysts and others to use the numbers for alternative analyses. Therefore, it is recommendedthat the government use spreadsheet or other tab-delimitedformats to disseminate quantitativebudget information. 45 Mongolia: Public Expenditure and Financial Management Review 2.69 The budget proposal is accompaniedby the draft general directives for socio-economic development for the corresponding year, which analyzes performance for the current year andthe projections for both the coming budget year and the medium term for most important macroeconomic and social indicators. However, the linkage between this document and the annual budget proposal is not explicit or easily traceable. 2.70 The Approved Budget. The budget proposal submitted by the executive is discussed by the parliament and promulgatedas a law by December 1". The law sets the lower bounds for revenues to be collectedto the state budget, and upper bounds for expenditures, broken down by administrative unit. It also specifies the amount of investments to be made, inter-governmental transfers to localgovernments and revenuesto betransferred from localto centralbudgets. 2.71 The enacted budget is available at the Parliament website (www.parl.gov.mn) and through the State Information Bulletin, the official publisher of laws and other official documents. However, the information is highly aggregated and thus not particularly helpful for citizens or analysts in monitoringor detailed analysis. Also, the document includes only the state budget, so it is not usefulfor analyzinglocalbudgets. 2.72 Monthly Budget Execution Reports. Monthly reports are important in terms of showing progress in the actual implementationof enacted budgets. They also serve as a basis for any mid-year amendments that may be necessary. The monthly reports are prepared by the Ministry of Finance, and availableat their website and from the Monthly StatisticsBulletin of the NationalStatisticalOffice. 2.73 Contents of the monthly reports have become richer in the last year. Recent reports show revenues for the general budget, including both state and local budgets, by tax types, expenditures by economic classification, in accumulated terms from the beginningof the year. It also shows local budgets, with aggregates for each province and intergovernmentaltransfers. The most recent report, for May 2007, also contains a brief narrative report explaining the main indicators in the quantitative report, making the monthly reports even more in compliance with the OECD Best Practices. In order to further improve reporting, the government could consider the possibilityof providingmonthlyreports with functional and administrativebreakdown. 2.74 Mid-year budget assessment and economic review. Currently, the mid-year report that is presented to the public is similar to the other monthly budget reports. It is recommended that the Government improve the mid-year report in terms of both content and review process. Specifically, the government could officially present mid-year performance to both the Parliament and the public. This would allow the Parliament to decide whether in-year amendments are requiredor not. Also, it would be advisable to include changes over the first half of the year to economic assumptions that were used to forecast the annual budget in mid-year reports. These changes should be presented with a comprehensive discussion of the Government's financialandnon-financialassets and liabilities. 2.75 End of fiscal year report. Preparation and discussion of year-end reports are covered by the PSFML. However, the report itself is not available to the public. Therefore, we recommend that the Ministry of Finance release the end of fiscal year report to public as soon as it submits the reportto the State Audit Office andthe parliament. 2.76 For the government budget of 2006, some details of the report were availablefrom the parliament website in the form of analysis of the parliament standing committee members, to make it easier for the parliament members to discuss the report. The parliament releases a very short approval of the budget performance for the given year, with aggregated numbers. But as mentioned above, for citizens and researchers, the publicationof the entire report is the first step to ensuretransparency and accountability. 46 Chapter 2 -Public Sector Financial Management 2.77 Report of the Auditor General. The budget performance and annual financial report of the government are audited by the MNAO, in accordance with the PSFML requirements. The standard used in external audits meets the International Organization of Supreme Audit Institutions(JNTOSAI) standardfor external government auditing.The audit reports are available from the website of the State Audit Office, but are not widely accessed by the public since there is an IT shortage in Mongolia. Also, audits are done for specific components of budget expenditures. For example, there is a separate report on investments made from the state budget. 2.78 The audit reports are discussed by the parliament in order to approve the final budget performance. However, it is not clear whether actions are taken based on the audit findings and recommendations therein. There is no legal requirement for the government to implement the recommendations of the auditor's report. It is under the purview of the Budget Policy committee to ask the government about any follow-up to the recommendations. However, it is not clear to what extentthis is practicedin Mongolia. 2.79 Citizen's budget, pre-election report and long-term report. For the general public, budgets are difficult to understand. In some countries, governments issue "citizen's budgets," which present information in easily understandable manner. There is no such presentation in Mongolia, so the government might consider issuing one, at the national, ministerial and local government levels. Also, there is a legal requirement for a pre-electionbudget report, at least one month in advance of the elections. The report was done in 2004, but the presentation could be made more clear and intelligible, and accessibleto non-professionals.Otherwise, the pre-election report cannot serve its purpose, which is to foster an informed electorate and stimulate public debate. Long term reports, that discuss the sustainability of government policies in the longterm, are not prepared or released by the government. These would be useful for assessing whether current policies are enforceable or not, and for making policy changes. The government should consider issuingsuch reports, at least every five years. 2.80 Local level budgets. During the preparationof this report, the Open Society Forum consideredthe availability of informationon localbudgets, both in the capitalcity and provinces: The general conclusion was that that there is no centralized point where disaggregated data and informationon local budgets is presented. 2.81 Usually, budget information is available from the finance division of the governor's offices in the aimags. However, there is no proactiveinformationdissemination.Especially at the soum level, there is very little information on budgets, both for the local administration and for schools and hospitals that are financed directly from the center. There is also no budget for disseminationto citizens. Given the centralizednature of the budget process, local governments have few incentivesto promote budgets, or havethem discussedby localcitizens. 2.82 There is a need for the Ministry of Finance to run a consolidated website with more up-to-date and aggregated information on the budget.The budget information should show both local budgets and budgets of local budget entities that are financed from the center directly. This would provide citizens with a one-stop portal for obtaining information and allow them to participate in budget dialogue. 2.83 Overall, there is a need for the Governmentto strengthen the dialogue on public expenditures with civil society. A way forward could be to establish a regular forum during which the Government will presentMBTF policiesandproposedState Budgetsto media and civil society organizations, around the same time that the policies are presentedto Parliament.As part of the budget preparation and approval process, the Minister for Finance could plan and pre- announce a timetable for discussingthe executive budget proposal with media and civil society 2- 47 Mongolia: Public Expenditure and Financial .Management Review 3 weeks before this proposal i s submitted to the Parliament. This is to ensure that there is appropriatetime for modifyingthe budget proposalas necessarybasedonthese consultations 48 CHAPTER 3. EDUCATION PROGRESSAND CHALLENGES high among comparators in terms o f education expenditures, enrollment and ere are important disparities in student completion rates and performance by area Internal rural to urban migration challenges the survival o small rural schools and puts pressure on urbanschools. These issuesare complicatedby a large and sparsely populated territory with highcosts o f education service delivery. D Domestically-funded capital expenditures are low, but this low level tends to be offset by donor financing, which i s focused on new investments. Donor coordination has been improving with the establishment o f the Education Donors Consultative Mechanism (EDCM) and the endorsement o f Mongolia's Education Sector Master Plan (2006-201 5). The budget gap identified inthe Master Plan, which is due in part to the expansion o f the general education cycle to 12 years, is being addressed by the Fast TrackInitiative's CatalyticFund. The 2002 Public Sector Management and Finance Law (PSMFL) has improved the predictability o f school financing and fiscal discipline at the school level. Yet it has also introduced more rigidity into the system. Schools are limited in their use of saving. Performance-based contracts are used by schools, but fail to evaluate teachers on measurable outcomes. Allocations for fixed costs do not encourage savings, while differences in infrastructure make it difficult to estimate needs. Aside fiom fixed utility costs, Since 2008, variable costs are allocated based on the nationwide location and type o f school andteachers' salaries allocation are calculated separately from other variable costs. Lack o f budget planning skills among local authorities and school management limitsthe efficiency gains from the per-capita funding formula. B Demand-side financing programs benefit the poor and have student performance, butthe coverage is limited and there are significant leakages KEYFUTURE PRIORITIES 1. Improve annual budget planning instead o f introducing more flexibility in budget execution. Cross- country experience suggests that carrying over savings is difficult to implement. 9 Suggested"how to": train localstaffandcouncils for budgetplanningandfinancialmanagement. ii. Improvespendingefficiencybyreducingheatingcosts. > Suggested "how to": increase capital maintenance funding; improve heating infrastructure (boiler, pipes, insulation, heat controllers), install meters, and tie funds to consumption and provide incentives to save. iii.Toaddresshighrural-urbandisparitiesandpromoteequity,promotepolicieswhichencouragemulti- grade and multi-subject teaching inrural schools. 9 Sumested"how to": consider a campaign to promotemulti-grade teaching insmall schools with low enrollments, by using successful experiences from other countries. In the short term, schools that transition to a multi-grade model could be given a financial incentive-in the form o f a school grant. iv. In financial aid program, use a combination of poverty targeting (through proxy means testing) and merit to allocate tuition aid for higher education. Use a proxy means test to improve poverty targeting 49 Mongolia: Public Expenditureand Financial Management Review o f demand-side financing programs in general education (school supplies and textbooks). Build a stronger system o f monitoringand impact evaluation. > Suggested "how to": Commission a feasibility study to assess the current state training fund and examine alternative demand-side financing mechanisms (grants, loans) to improve access to higher education in a cost effective way. Facilitate access to student loans for those who do not qualify for grants and improve the design o fthese loans. 50 Chapter 3 - Education OUTCOMES, ACHIEVEMENTS AND REMAININGCHALLENGES Outcomes 3.1 Overall, Mongolia ranks high among countries in the East Asia region in terms of enrollment rates and the percentage of students completing basic education. In 2006, net enrollment rates were 92.8 for primary, 65.3 for lower secondary, 71.7 for upper secondary and 34.15 for higher education.' Secondary and higher education gross enrollment rates are high relative to other countries in the East Asia and Pacific region, while Figure 3.1': Mongolia basic education ranks high in primary enrollment rates are East-Asia but not so high compared to Central Asia This may be due to reluctance of parents in the countryside to send Completion of Basic Education (%) their young children to schools in district and provincial centers, which are far from home. Mongolia does well relative to other low income countries, which have an average net enrollment rate of 78.2 for primary, 38.3 for secondary and 8.7 for higher ed~cation.~Compared with other East Asian countries, Mongolia has also been successful Cambodia Indonesia WetnamThailand Mongalia Armenia Tajikistan in ensuring that a large percentage Source: WorldBank Staff o f students complete a basic education, although not as successful as some of the Central Asian countries (See Figure3.1). 3.2 Yet enrollment figures mask important disparities in completion rates by area, wealth and, to a lesser extent, gender. Children attending schools in rural areas (bagh and soum schools) are less likely to complete lower and upper secondary then students in urban areas (aimag centers and UB schools) (See Table 3.1). This can partially be explained by the need for many children in rural areas to help support their families with animal husbandry. The consolidation of upper secondary grades into provincial capitals has also had an impact on rural students transitioning to this level. The GOM recommends that schools transfer at least 70 percent of grade 9 students to grade 10, but in practice many overcrowded schools have taken this to be the upper limit. In terms of gender, boys' dropout rate after completing lower secondary is significantly larger than girls. Disparities are even larger across consumption quintiles, with a higher dropout incidence among children from poorer households. 'HIES 2006 Gross enrollment rates for primary, secondaryand higher education inMongolia for 2005 were 93.3, 91.8 and 43.2, respectively, while for the EastAsia and Pacific region these figures were 111.I,71.8 and 20.4. World Bank Edstats 2004. Note figure for higher education in low income countries is gross enrollment. 5 1 Mongolia: Public Expenditure and Financial Management Review Table 3.1: There are important spatial and income disparities in completionrates :YOenteringprimary that complete eachlevel) Children aged 8-25 Total Urban Rural Boys Girls Poorest Richest Primary 93.95 96.74 89.36 91.69 96.13 90.83 96.60 Lower secondary 81.04 87.41 70.00 75.98 85.72 72.93 90.61 Upper Secondary 54.87 60.01 45.73 48.73 60.35 42.59 69.93 Diplomas 41.91 47.20 31.56 38.31 45.49 39.59 5 1.47 University 9.43 10.38 7.79 10.36 9.20 7.48 I 12.60 Source: HIES 2006 Notes: Figures are Kaplan Meier estimates, which account for censoring (i.e. individuals still attending school at the time of the survey only contribute to the estimation sample up to last level currently attended). Survival rates indicate the cumulative probability of making the transition to each level. 3.3 Learning achievement is low overall and there are significant differences in the performanceof students by location (Figure 3.2). Based on the results from the 2005 National Assessment o f Students' Achievement (NASA), which tested eighth grade students in both math and civics, the average percent correct for students in mathematics and civics was 50 and 47 percent, respectively - low by international comparison. NASA is based on the adaptation to the local context o f two international tests, TIMSS and CIVED.4 In mathematics, students in the capital, Ulaanbaatar (UB), scored above the mean, while students in provincial centers (aimags) and district schools (soums) scored well below the mean. The differences are not quite as sharp in civics, but students in UB and aimag centers outperformed students in soums. The differences are most significant between students in soum and aimags and UB schools. Figure 3.2: Students' achievement by location 1 Almag Cente d Civics R Math - 1I -0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 Source: Education Evaluation Center data 2005 Note: Z scores are usedfor both math and civics, The z score reveals how many units ofthe standard deviation a case is above or below the mean. Math: T-test of differences signijkant at 95 percentfor UB and aimag centers and 99percent for UB and soum/baghs and 99percent for aimag centers and soums. Civics: T-test of differences significant at 99 percent for UB and soum/baghs and 99 percent for aimag centers and soums. 4 Civic education included domains in politics and democracy, Human rights and freedom, State structure and Mongolian history. The number o f tasks that were kept from TIMSS and CIVED is small to make a meaningful international comparison. See Education and Evaluation Center (2006) for details. 52 Chapter 3 -Education Achievements 3.4 There have been significant improvements in the education policy and planning process through the development of the Education Sector Master Plan (2006-2015). The EducationSector Master Plan (ESMP-2) puts forth the sector's main goals and was developed by MECS in consultation with a broad cross-section of education stakeholders. A local Education Donors Consultative Mechanism (EDCM) was established to promote increased alignment of donor projects and programs in support of ESMP-2 and consists of key development partners in education. A weakness ofthe ESMP-2,however, is that while it highlightsmany important issues in the sector, it does not consider carefully enough the strategic choices that policymakers are likely to face when constrained by budget limitations.Thus, it is importantfor MECS to develop multi-year action plans that prioritize activities in the sector. These action plans should be reviewedannually by MECS together with the EDCM. An annual action plan was developed for 2007 as part of Mongolia's recent accession to the Fast Track Initiative (FTI) and receiptof grant resources from the FTI's Catalytic Fund. 3.5 The government has begun an important structural reform by adding an 1lthyear to the primary-secondary education cycle in 2004-2005 and plans to add a 12th year beginning in 2008-2009. The reform addresses the very short (10-year) primary-plus-secondary cycle that existed in Mongolia prior to 2004-2005. Although the content which many countries begin to teach in grade 1, was covered in the last years of pre-school education, many children, especially the most disadvantaged, did not have access to pre-school, which was never made "compulsory," making these children doubly disadvantaged. The expansion of this cycle will be more aligned with international standards for pre-tertiary education. Even with a late start age, grade 1 has traditionally had the highest dropout rate of basic education. The government will face significant challenges to get all six and seven year-olds enrolled in the future and even greater difficulties in keepingthem there. 3.6 The private sector plays a key role in facilitating the access and improvement of education, but unregulated development may lead to lower quality. The number of private schools in the basic education sector has steadily increased over the past few years and now accounts for 20 percent of all schools. However, in terms of overall enrollment, the share of students attendingprivate schools is relatively small (3 percent of all general education students). The government partially subsidizes private schools by providingthem with a per-studentamount to help cover teacher salaries. At the tertiary level, the lack of a well-developed accreditation system has resultedin many low quality institutions.The 1991EducationLaw established a legal basis for privateuniversitiesand since then they have expandedrapidly (73 percent of institutions and 34 percent of students). The current challenge for the government is to develop a quality assurance and accreditationsystem that can ensure the quality of private institution^.^ And while 50 percent of institutions are now accredited,just 15 percent of these were privateinstitutions (as of 2003 for afive year period). Challenges 3.7 Mongolia's low population density, large territory and harsh winters make it difficult to deliver educational services and translate into high costs. Mongolia's educational expenditure-as a percentage of total public expenditure- is among the highest of the transition See Asian Development Bank and the Ministry of Education, Culture and Science (2005) for details. Major criteria for accreditation focus on whether an institution has a clear and publicly stated mission, has made progress towards achievingthe missionand purposes, has establishedthe extent to which resources o f the institution are directedand organizedtoward achievingeducationalobjectives, andhas demonstrated the integrityandcommitment to missionaccomplishment. 53 Mongolia: Public Expenditure and Financial Management Review countries. This is, in part due, to the relatively high cost o f delivering education services in the country. For instance, because o f its sparse population and its nomadic culture, school dormitories play a key role in enabling rural children to access school. Inaddition, harsh winters also result in high expenditures for heating. Nearly 15 percent of total education expenditures in 2005 went to finance heating costs. Heating sources vary considerably by school location making it difficult to set national standards for heating costs and apply them consistently across schools. 3.8 Internal migration to urban areas challenges the survival of small rural schools and puts pressure on urban schools. Migration from the rural areas to the provincial centers and cities i s a large and growing phenomenon in Mongolia. As a result, Mongolia is struggling to maintain school facilities in remote and rural areas, while at the same time face difficulty in meeting rising demand for education in urban areas. Migration results in falling student numbers on the sending end and rising student numbers on the receiving end, and both can result in inefficiencies for the system. The pupil-teacher ratio (PTR) in primary education in rural areas is 31:l compared with 34:1 in Ulaanbaatar, on average. Declining school quality in rural areas appears to be contributing to migration to urban areas6 PUBLIC FINANCINGOF EDUCATION Education expenditures 3.9 Education's share of government spending has recovered in recent years, after struggling throughout the transition period. Under the former, centrally planned system, education was given a high priority and received the largest share o f Figure 3.3: Mongolia spends relatively more in government resources. After education than others east and central Asia transition reaching a low o f 16 percent o f the countries government budget in 1997, education's share of the budget has begun to increase. Education expenditures followed GDP growth I Cambodia 153 i I more closely increasing nearly 7 11 6 percent from 2004 to 2005 to 20 Lao PDR percent o f total public expenditure. 169 ~ The Government o f Mongolia's Tajikistan announced policy in regard to Indonesia education spending is that the ~ education's share o f total Philippines government budget should be 20 percent. 2007 figures, however, fall Azerbaijan below that threshold down to 11 percent, Mongolia 3.10 Mongolia spends more on 0 5 10 15 20 education than most low income 0 % of GDP I% of Total Public Expenditures countries. In terms o f both expenditure as a percent o f GDP Source UNECOInstitute of Statistics -2004 data and as a percent o f total public expenditures, Mongolia spends more relative to other low income countries (Figure 3.4). Total 6 A 2005 report by Save the Children found that one-third of households in their survey reported their children's schooling, after seeking employment, as the most important reason for migrating. 54 Chauter 3 - Education spending on education as a percent o f total public expenditures was 19.1 in 2004 compared with an average o f 16.2 for other developing countries in the Asia Pacific region (Figure 3.4).7 However, as previously mentioned, education services are more costly to deliver in Mongolia due to its low population density and harsh winters. towards recurrent costs. Public spending on education covers both Figure 3.4: Sub-sector shares of education capital and recurrent expenditures. expenditures Since 2000, the average share o f capital expenditures for the overall 6oo% education sector was just 3 percent 5o o% o f total education expenditures, 4 o o o ~ Given Mongolia's harsh climate and 3o o% low population density, capital 2o o% expenditures are important in ,o o% maintaining school facilities that 0 0% ensure children's enrollment in &6 B6 8 .rc 00 * 200 !!* fa 150 Y Ft 100 ac 50 0 0 5,000 10,000 15,000 20,000 0 5,000 10,000 15,000 20,000 GDP p.c. (constant lnternatlonal US$) GDP p.c. (constant lnternatlonal US$) Source: Datafrom WorldDevelopment Indicators (2007) 4S Mongolia has made Figure4.2: Socioeconomicand geographicdisparities in important achievements in reducing mortality mortality, disparities in mortality rates across socio-economic groups and geographic areas remain important (Figure 4.2). For example, I mortality rates are five times higher among the poorest 20 percent of the population than the richest 20 percent. Although not unusual by international standards, these disparities point to persistent challenges in reaching vulnerable and high-risk groups with - - - I B r informationand services. 0 p gc IDi g 3 Z E 6i . 'e 1 4'6 3 One contributing factor to the 8 a in infant and is Source: Data on infant mortality rates (ZM; and under-jive mortality rates (U5MR)from 2005 Middle likely to be the high rate of zncome Countries, vaccination achieved in Mongolia. Indeed, unlike most other low- and middle-income countries, Mongolia has succeeded in achieving near universal immunization coverage (Figure 4.3). Mongolia has also seen improvements in child nutrition in recent years, although stunting remains high at nearly 20 percent. Maternal mortality is also a persistent problem, although the mortality ratio is low by relativeto other countries at similar levels of economic development (Figure 4.4). Infant mortality rate declined from 78 to 39 between 1990 and 2005, while under-five mortality rate declined from 108 to 49. This corresponds to annual rates of reduction of 4.6 percent and 5.3 percent respectively. On the basis of these data, Mongolia is on track to meet the MDGtargets related to child mortality (which require a 4.3 percent annual reduction). 74 Chapter 4 Health - Figure4.3: Immunizationrates-a successstory ;&:&an* * ;** * :*** * s: : * * ** * 1, + 3:8 *Capbop ** ** *Kazakhstan * * : * 0 *** * .** *** *** *'b * * * * * * ** * * * 0 5,000 10,000 15,000 20,000 0.00 5,000.00 10,000.00 15,000.00 20,000.00 GDP pc (PPP) GDP pc (PPP) Source: Datafrom WDI (2007) Figure4.4: Maternal mortalityand attended delivery 1200 :* l*r * *4e * * *: ** +Cambodia * 0 5,000 10,000 15,000 20,000 0 5,000 10,000 15,000 20,000 GDP p.c. (international US)) GDP pc (PPP) Source: Datafrom World Development Indicators (2007) 4.7 While infant and child mortality rates appear to be declining, adult mortality rates are on the rise.Non-communicable diseases, such as heart disease, stroke, cancers, hypertension, diabetes, and chronic lung disease, as well as injuries, are now accounting for a growing share o f mortality (Table 4.1). In some cases, incidence has been increasing rapidly. For example, the incidence o f circulatory disease has increased from 162.30 in 1990 to 449.52 in 2002. These changes reflect behavioural patterns and trends, such as high smoking prevalence, increased fats, sugars and calories in the diet, high levels o f alcohol consumption, and decreased physical activity (Figure 4.5). New challenges call for new skills, programs, and capabilities in the health system. Yet, similarly to many other countries in the region, Mongolia's public health system has been heavily oriented towards communicable diseases, and the required health system changes comprise an important agenda for the future. 75 Mongolia: Public Expenditureand Financial Management Review Table 4.1: Ten leadingcauses of mortality (rate per 100,000 in 2004) Total Male Female 21 Diseases of circulatorv svstem 230.6 264.1 200.8 Tumors and neoplasms 121.6 138.9 106.2 3 Iniuriesand poisoning and certain other conseauenceso f 103.4 171.1 39.7 4 Diseases of the digestive svstem 48.2 55.5 41.6 65 Diseases of the respiratory system 30.3 37.2 24.1 Certain diseases originating from the perinatal period 18.7 22.2 15.6 78 Infectious and parasitic diseases 15.0 18.5 11.9 Diseases of the genitourinary system 13.5 15.7 11.5 9 Diseases of the nervous svstem 8.8 10.4 7.3 10 Congenital malformations. deformations and chromosomal 5.5 6.7 4.4 Source: MOH 2004. Health Indicators 1 . I * * . * * * * *&erb$ijan* 0 0.00 5,000.00 10.000.00 15,000.00 20,000.00 0.00 5,000.00 10,000.00 15,000.00 20,000 00 GDP pc (PPP) GDP pc (PPP) I Indontsm I ,end Chile Philippines M a l a y d * * uzbeklsta%r~Lanka '* * 8,az,i ** South Africa 10 * * j 0 0 5.000 10,000 15,000 20,000 25,000 30,000 GDP pc (PPP) Source: Data on obesityfrom WHO Global InfoBase Online; data on alcohol consumptionfrom WHO Global Health Atlas for alcohol consumption data; data on smokingfrom WDI2006; GDP datafrom IMF WorldEconomic Outlook Database. 4.8 While chronic disease has emerged as a growing health threat, communicable diseases remain important. The greatest threat comes from tuberculosis (TB), with a doubling 76 Chaater 4 - Health of incidence (from 79 to 141 per 100,000) between 1990and 2001.4HIV prevalence is currently low, but a number of factors indicate risksthat the disease may spread inthe future.5 4.9 Despite significant communicable and chronic diseasechallenges, the picture on the preventiveside is a mixedone. Coverageof immunizationsandmaternityservicesremains high, and this has undoubtedly contributedto relatively good health outcomes in some areas. On the other hand, there is a clear need to strengthen primary and secondary prevention of chronic disease, as well as of some communicable disease, suchas TB and STIs. 4.10 Available (incomplete) data on access curative healthcare services suggest that high levels of out of pocket spending thereby constituting an important financial burden for households.Dataon this are quite limitedin Mongolia.For example, available household survey data have not been analyzed systematically to determine levels and trends in the incidence of catastrophic health expenditures or the benefit incidence of government spending on healthS6 According to the 2002/03 Household Income ExpenditureSurvey, total per capita out-of-pocket spending by households amounted to 1,919 Tugrik, or Figure4.6: Out-of-pocket spendingby householdsin Mongolia 5.2 percent of total consumption. With the exception of Vietnam, this is 90% the highest share of spending 80% in household consumption of 70% all the 15 East Asian 60% countries/territories considered by the EQUITAP 50% study.7 Hence, although data 40% on catastrophic spending are 30% not available, the relatively 20% high levels of out-of-pocket 10% spending in consumptionraise concerns about the potentially 0% National Poorest Q2 4 3 Q4 Richest impoverishing effects of Outpatient services .Medicines OHospitai stays OReproductive health healthcare spending. Source: Datafrom 2002/03 HIEYLSMS 4.11 Most of out-of-pocket spending (90 percent) is on - . - outpatient and medicines, with the remainder being spent on inpatient services (Figure 4.6). For the sample as a whole, the split betweenspendingon outpatient services and drugs is almost even. However, richer households spend considerably more on outpatient care, while the data suggest that poorer householdsare more likely to self-medicate. This rise is due in part to improved detection and reporting that has gone hand-in-hand with the introduction of Observed Treatment Short Course Program (DOTS) in 1995. This has also contributed to an increased cure rate, from 3 1.4 percent in 1990 to 83.4 percent in 2002. For one thing, Mongolia has seen a rapid increase in STIs in recent years, with growing incidence of gonococcal infection, syphilis, and other STIs. Despite these trends, studies have found relatively low condom use in encounters with commercial sex workers among male clients STI clinics. Injecting drug users comprise another important risk froups about which little is currently known. Data on these issues are available from a wide range of other countries in East Asia. For details, see working papers 'onFor www.equitap.org. details, see "Paying out-of-pocket for health care in Asia: Catastrophic and poverty impact" by Eddy van Doorslaer et al., EQUITAP Project: Working Paper # 2. 77 Mongolia: Public Expenditure and Financial ManagementReview FINANCING HEALTHCARE OF Health financing in Mongolia: the big picture 4.12 According to National Health Accounts (NHA) estimates', per capita expenditures on health were approximately US$40 (official exchange rates) in 2004 (Table 4.2). Government expenditures accounted for approximately63 percent of total healthexpenditures. This includes both expenditures financed from general tax revenues and expenditures financed through the Health Insurance Fund (HIF). While the share of government expenditures in total health expenditures declined in the period 1998 to 2004, the share of private spendingincreasedfrom 23 percent to 37 percent, with most of this taking the form of out-of-pocket payments by 2004. While most health expenditures are ultimately financed by households-eitherin the form of taxes or direct payments--external financing is also an important, albeit fluctuating, source of financing. During the period for which comparable data are available, the share of external financing has ranged from 3.2 percent to over 22 per~ent.~The significant share of external financingraises importantchallenges relatedto strategic prioritization in resource allocation and effectivepublic expendituremanagement.lo Table 4.2: Health expenditures and financinginMongolia 1998 1999 2000 2001 2002 2003 2004 Total Health Exp. (US%,official exch. rate) 20.7 19.9 26.4 29.6 29.1 33.0 40.3 of which financing agent) General government 77.1% 73.3% 65.9% 66.3% 69.8% 63.8% 62.9% Government spending(excl. SHI) 45.0% 54.1% 50.0% 45.4% 43.7% 39.7% 39.1% SHI 32.1% 19.1% 15.9% 20.9% 26.1% 24.1% 23.8% Private spending 22.9% 26.7% 34.1% 33.7% 30.2% 36.2% 37.1% OOP 8.8% 14.9% 14.5% 14.1% 25.5% 33.0% 34.0% Other private (incl. NGOs) 14.1% 11.9% 19.6% 19.6% 4.6% 3.2% 3.1% of which financing source) External financing 10.6% 18.4% 22.5% 21.1% 12.8% 3.2% NIA Source: WHONational Health Accounts 4.13 Taken together, total health expenditures in Mongolia are high relative to other countries at similar levelsof GDP, and at 6.7 percent, the share of total healthspending in GDP is higher thanmostother low and middleincomecountries (Figure 4.7). * Data are from WHO NHA (2006) data base (see htt~://www.who.intlnha/countrv/mn~en/),and refer to the April 2006 update. Further updates of the NHA data have taken place since, but are not reflected in this chapter. The NHA exercise in Mongolia has been documented in the Ministry of Health and World Bank report "Mongolian National Health Accounts" (2005), and this chapter will not discuss the data sources or estimation methods. It should be noted that the 2005 report only includes data up to 2002. Subsequent updates reported by WHO have been developed by the Ministry of Health(see details of WHO web page), but the data and approach on which these updates are based are not documented. Although NHA data are compiled by national institutions using official data (survey data in the case of out-of-pocket expenditures), there are often minor discrepancies with budget data and other sources. Given that discrepancies are relatively minor, and given the complexity of the tasks, this chapter will not seek to reconcile the NHA data with the budget data presented in the chapter. These fluctuations are likely to be explained at least in part by the way in which external financing to the health sector i s recorded. For example, insofar as the data reflect financing agreementsor disbursements, these will be `lumpier' than the actual use of externalfunds. A detailed assessment of externalfinancing is beyondthe scope of this study. lo These general issues are discussed elsewhere in the report. See also Ministry of Finance and UNDP (2005). "Relationship between Socio-Economic Guidelines and Budgeting." Ulaanbaatar, Mongolia. 78 Chapter 4 -Health Figure4.7: Total healthexpenditureand GDP (2003) 12.0 Cambodia - 10.0 u 0 A i6.0 g { -J 4.0 I- 2.0 0.0 0 1,000 2,000 3,000 4,000 5,000 6.000 7,000 8.000 9,000 10,000 0 1,000 2,000 3,000 4,000 5,000 6.000 7,000 8.000 GDP P.C. (current international US$) GDP p.c. (current Internatlonal US$) Source: Datafrom World Development Indicators (2006) and WHO Statistical Information System (2006). 4.14 From an internationalperspective, Mongolia also stands out as having a high level of government spending on health as a share of GDP (Figure 4,8a), and an unusuallyhigh share of total healthexpenditures accounted for by government (64 percent in 2003) (Figure4.8b). This reflectsbotha relativelyhigh levelof overall government spending, and a commitmentof over 10 percentof general government expenditures to healthin 2003. Figure4.8: Publicspendingon health . e t 1' . . ..*. . . * .* 0.0 0.04 0 5,000 10,000 15,000 20,000 0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000 10,000 GDP P.C. (current international US$) GDP p.c. (current Internatlonal US)) Source: Datafrom World Development Indicators (2006) and WHO Statistical Information System (2006). 4.15 Turning to official budget data, a number of trends and patterns in government expenditure on health are discernable (Table 4.3)." First, in real terms, government health expenditures increased at a relatively slow rate (6.3 percent p.a., or 5 percent on a per capita basis) between 2001 and 2006. Second, reflecting patterns in overall government spending, the trend in government health expenditure has beenerratic, with significant increases in some years and declines in other years. Third, government expenditure on health has grown at a slower rate than GDP in the period2001-2005 and, as a result, health expenditures have been declining as a share of GDP. This seems to be primarily due to a decline in overall government expenditure as a I] The data in Table 4.3 is basedon the sectoral classification of expenditures.There are significant differences between sectoral and institutional (Ministry of Health) allocations prior to 2003 (when health sector expenditures were highly decentralized), but only slight differences thereafter. As the sectoral expenditures include both central and local components, this series is more appropriate for assessingexpenditure trends. 79 Mongolia: Public Expenditure and Financial Management Review share o f GDP, rather than a reduced share o f health expenditures in overall government expenditures. 4.16 On the financing side, there has been little change in the composition between 2001 and 2005 (Table 4.4). Budgetary financing remains predominant, reaching 71 percent in 2006. The Health Insurance Fund (HIF) accounts for another 24 percent, with the remainder financed from facility revenues or balances. Table 4.3: Governmentexpenditureon healthat a glance Actual Plan 2001 2002 2003 2004 2005 2006 GDP 1,115,641.3 1,240,786.8 1,461,169.0 1,910,880.9 2,266,505.2 NIA Population (million) 2.42 2.45 2.48 2.5 1 2.55 2.58 Tot. govt. exp and net lending (GE) 491,257.6 550,48 1.2 615,771.3 752,486.4 764,597. I 1,209,240.9 Tot. govt. curr. exp. (GE curr) 366,700.3 415,309.0 434,829.4 538,699.2 600,288.8 876,974.4 Govt. health exp. (GHE) 53,096 57,964 58,128 73,243 80,152 99,995 Current 96.3% 95.0% 94.2% 95.2% 95.7% 93.5% Capital 3.7% 5.0% 5.8% 4.8% 4.3% 6.5% Central share 36.4% 38.1% 99.6% 99.7% 99.7% 99.8% Local share 63.6% 61.9% 0.4% 0.3% 0.3% 0.2% GHE p.c. 21,928.2 23,673 .O 23,442.7 29,126.2 31,382.8 38,757.9 GHE p.c. (const. 2000 prices) 20,322.7 21,560.1 20,402.7 22,862.0 22,480.5 25,947.2 GHE p.c. (US$, curr. official rate) 20.0 21.3 20.4 24.6 26.0 32.7 GHE p.c. (US$, curr. PPP) 74.6 76.7 69.4 75.0 74.5 NIA GHE as YOof GDP 4.8% 4.7% 4.0% 3.8% 3.5% N/A GHE as YOof GE 10.8% 10.5% 9.4% 9.7% 10.5% 8.3% GHE cum. as YOof GE curr. 13.9% 13.3% 12.6% 12.9% 12.8% 10.7% GE cum. as YOof GDP 32.9% 33.5% 29.8% 28.2% 26.5% NIA Annual rate of change (constant prices) GDP 9.3% 12.5% 17.9% 8.2% NIA GE curr. 11.3% 0.1Yo 11.7% 1.7% 36.5% GHE 7.3% -4.2% 13.6% -0.1% 16.6% GHE p.c. (const. 2000 prices) 6.1% -5.4% 12.1% -1.7% 15.4% All amounts in million Tug. unless otherwise specified. Note: Data on GHE is based on the sectoral classification of expenditure. Since 2003, health sector spending has been close to MOH spending. As a rule, MOHspending is 3 3 % higher than sectoral spending due to higher allocationsfor heating, centralized measures, and subsidies and transfers. Table 4.4: Sourcesand trends in government healthfinancing Actual Plan 2001 2002 2003 2004 2005 2006 Financing resource to cover expenses 100% 100% 100% 100% 100% 100% Government budget 68% 64% 65% 68% 66% 71Yo Health InsuranceFund 26% 29% 28% 26% 28% 24% Core activities revenue 5yo 4% 5yo 4yo 4yo 3yo Non-core activities revenu 1Yo 3?"o 2Yo 2Yo 1Yo 1 Yo Beginning balance 0% 0Yo 0Yo 0% 1YO 0Yo All amounts in million Tug. unless otherwise speci9ed. Source: MOH 80 Chapter 4 - Health Improvingthe allocationof governmenthealthexpenditures 4.17 The share of salaries and other personnel costs in overall spending is high and increasing. In the last six years (2001-2006), the share o f salaries and related expenses has increased from 28.3 percent to 35.1 percent for the health sector as a whole (Table 4.5). This expansion has taken place at the expense o f expenditures on goods and services, in particular medicines and vaccines. For the health sector as a whole, personnel related expenditures account for 35 percent of the total. However, as can be seen from Table 4.6, the percentage i s considerably higher for hospitals and other service providers, where the displacement of expenditures on goods and services in some cases has been more dramatic. Although the implications o f these allocation patterns require further study, they clearly entail risks that non- personnel expenditures are insufficient to ensure adequate maintenance and material inputs, with consequences for both efficiency and quality o f service delivery. Table 4.5: Government healthexpenditure by lineitem 2001 2002 2003 2004 2005 2006* Total expenditure (million Tog.) 53,096 57,964 58,128 73,243 80,152 99,995 Total current expenditure 96.3% 95.0% 94.2% 95.2% 95.7% 93.5% Salary, wages and supplementary 28.3% 27.9% 30.6% 31.2% 31.O% 35.1% Employers' insurance contribution (pension & health 8.0% 7.5% 8.8% 8.2% 8.2% 9.3% Expenditure on other goods and services 59.2% 58.7% 54.6% 55.0% 55.5% 48.2% Medicines & vaccines 18.0% 15.7% 17.2% 16.4% 15.6% 12.6% Heating 10.5% 9.2% 9.2% 7.9% 8.2% 8.0% Food expenses 5.0% 4.7% 5.0% 4.7% 4.5% 3.7% Fuel and transportation expenses 3.7% 3.4% 3.5% 3.5% 4.2% 3.7% Other 22.0% 25.7% 19.8% 22.5% 23.1% 20.3% Subsidies and transfers 0.8% 0.8% 0.8% 0.7% 1 .O% 0.8% Capital expenditures 3.7% 5.0% 5.8% 4.8% 4.3% 6.5% *All amounts actual, except 2006, which refers to approved budget allocation Source: MOH Table 4.6: Line item compositionfor differenthealthsector institutions 2003 2004 2005 2006* 2007" Clinical hospitals Salary and wages 32.4% 35.7% 33.5% 39.4% 45.4% Exp. on other goods and services 66.6% 63.6% 65.9% 60.0% 54.1% Other exp. 1.O% 0.8% 0.6% 0.6% 0.6% District hospitals Salary and wages 38.6% 41.O% 42.1% 50.3% 56.9% Exp. on other goods and services 60.3% 57.6% 56.8% 48.7% 42.3% Other exp. 1.O% 1.4% 1.1% 1.O% 0.8% Aimag and city hospitals Salary and wages 35.1% 38.1% 39.2% 45.1% 49.6% Exp. on other goods and services 63.0% 60.1% 60.4% 54.7% 50.2% Other exp. 1.9% 1.8% 0.4% 0.2% 0.2% Soum and inter-soum hospitals Salary and wages 44.4% 46.6% 47.6% 57.8% 60.3% Exp. on other goods and services 51.3% 50.1% 51.8% 42.1% 39.5% Other exp. 4.3% 3.3% 0.6% 0.2% 0.3% Other entities Salary and wages 20.6% 20.0% 20.7% 23.2% 24.2% Exp. on other goods and services 45.3% 52.2% 50.7% 45.2% 41.5% Other exp. 34.1% 27.8% 28.6% 31.7% 34.4% *All amounts actual, except 2006, which refers to approved budget allocation Source: MOH 81 Mongolia: Public Expenditure and Financial Management Review 4.18 The allocation of government expenditures is heavily oriented towards secondary and tertiary level providers.In 2006, more than 50 percent o f the MOHrecurrent budget went to secondary and tertiary hospitals (Figure 4.9). The second largest component was primary care providers (28 percent). Referral hospitals are inherently expensive to operate, and the high share o f spending accounted for by these facilities is not unusual by international standards. That does not mean, however, that there is not a case for re-allocating resources away from referral-level facilities. Indeed, as we will see below, referral hospitals currently provide a large proportion o f outpatient services and a large volume on simple (and possibly unnecessary) inpatient care. If some o f these services can be re-directed towards primary care providers and social services, it may be possible to reduce funding for higher level facilities (or, conversely, to improve the quality o f services with the same level o f funding). Figure4.9: Ministry of Health budgetby institutionalclassification(2006) Composition of MOH budget allocation, 2006 Sport-related Other expenditures7 6% Administration Public health institutions and Secondary and tertiary hospitals 51O/o Source MOH Secondary and tertiary hospitals include clinical hospitals, district hospitals, aimag and city hospitals, and other referral hospitals Primary providers andpharmacies include soum and inter-soum hospitals,family group practices, andpharmacies (reimbursed by HIF) The largest component of 'Other ' is the government subsidized health insurance contribution for some households 4.19 Soum and inter-soum hospitalsconsumea largeshare of the budget. Another striking feature o f the composition o f the MOH recurrent budget is the large share (18 percent) o f the overall budget that is allocated to the 333 soum and inter-soum hospitals that are spread around the country (Table 4.7). Given the relatively small share o f the population that is served by these facilities, and given their limited capabilities, the provision o f services to the rural population comes at a high costs. With Mongolia's geography and the disperse population, this is to some extent inevitable. However, it is likely that reorganization and restructuring o f service delivery in rural areas could achieve substantial cost-savings without adverse consequences for access to and quality o f services. This could include, for example, a reduction o f staffing and service range in soum hospitals, combined with a more effective referral system or outreach by aimag level facilities to provide specialist services. The feasibility and cost-effectiveness o f different options along these lines i s beyondthe scope o fthis study but merits attention in future work. 82 Chapter 4 - Health Table 4.7: Allocations to differenttypes of healthcare facilities(2006) Hospitals and clinics sub-total Tot. budget Num. institutions Average budget Aimag and city hospitals 23,743.0 28.0 848.0 Other hospitals 9,274.8 8.0 1,159.4 District hospitals 8,487.6 12.0 707.3 Clinical hospitals (1,2,3) 7,570.8 3.0 2,523.6 Soum and inter-soum hospitals 18,989.2 333.0 57.0 Familygroup practices 4,409.7 23 1.O 19.1 Private pharmacies and clinics (HIF funding) 3,529.5 47 1.O 7.5 Note: Other hospitals include mental disease, infectious disease hospital, maternity and infant research center hospital, infant clinical hospital, higly contagious diseas station, traumatology and reconstructive clinic. emergeny hospital, andpsychiutric clinical hospital Source' MOH 4.20 There are significant geographic disparities in expenditures. From a territorial perspective, nearly 61 percent o f the MOH budget expenditures in 2006 are spent at aimag level." When looked at from a per capita perspective, there are substantial disparities across different aimags, ranging from nearly 60,000 MNT. In Gobi, to less than 20,000 MNT. In UB, with a weak relationship between the per capita health budget and gross regional product per capita (Figure 4.10). Although current expenditure patterns look somewhat inequitable, there are clearly a number o f complex factors in play. For example, the unit cost o f providing services to a highly concentrated population, such as in UB, than in rural areas where the population is dispersed and inaccessible. Given that government equity objectives tend to relate to health outcomes and access to service rather than spending per se, the disparities described in the graph do not provide a sufficient case for re-allocating government spending in their own right. Rather, the reasons for the observed disparities in spending, as well as the determinants o f unit costs in different contexts, need to be better understood before any specific recommendations can be made in respect o f the allocation o fresources. Figure4.10: Government healthexpenditures at aimag level (a) 2006 MOH allocatlon to airnags p.c. 70,000 1 j: j 60.000 50.000 GobiAnal B 20,000 1 0 4 300 400 500 600 700 800 0 10,000 20,000 30,000 40,000 50.000 60,000 Aimag GDP ps. Source: MOH and Regional GDP data from 2006 Statistical Yearbook. For expositional purposes, panel (b) does not show datafor UB and Orkhon, where GDPp.c, is substantially higher than 800,000 Tugriks. Remaining expenditures are accounted for by central level institutions and facilities. Many of the activities and servicesprovidedby these institutionsandfacilities benefitspecific aimagpopulations,but attribution is difficult. 83 Mongolia: Public Expenditure and Financial Management Review What future for social healthinsurance? 4.21 Social Health Insurance (SHI), sometimes referred to as Citizen's Health Insurance, was introduced in Mongolia in 1994, and now comprises one o f the five main social insurance programs in the c o ~ n t r y . 'SHI was introduced in a time o f fiscal decline, and a key aim o f the ~ program was to mobilize additional funds for the health system and to reduce the reliance on budgetary financing. The SHI system was also expected to help improve service delivery performance through strengthened accountability and the introduction of market mechanisms in the health sector. Box 4.1: Key Features of Social Health Insurance in Mongolia HIF was initially administered by a non-government enterprise, but has since 1996 been managed by the State Social Insurance General Office (SSIGO) of the MOSWL. SHI policy and supervision is the responsibility of the HealthInsurance Sub-council, chairedby the State Secretaryofthe MOH. SHI is compulsory, except for foreign residents.Coveragei s currently around76 percent ofthe population. Revenues are based on (i)payroll contributions (6 percent, shared equally by the employer and employee) for individuals in the formal labor market; (ii)flat-rate contributions of 500 MNT. For informal sector workers, students, self-employed, and the unemployed; and (iii) full government subsidy of 300 MNT. for vulnerable groups (childrenunder 16, pensioners, parents with children under 2 years old, military service, social welfare beneficiaries, convicts). In 2005, about 65 percent of beneficiaries are subsidized by government, 20 percent make payroll contributions, and 15 percentmake flat rate contributions. Benefits under SHI are uniform for all participants. Benefits currently include cover for inpatient and outpatient services from secondary and tertiary level facilities that are not financed through the budget. In order to be entitled to benefits, the insuredmust pay contributions for 12 consecutive months. The degree of cost sharing by the insured patient in Mongolia depends on the type of hospitals; 10 percent insecondary hospitals and 15 percent intertiary hospitals. Exemptions apply for children under 16, students under 18, pensionerswith no other income, and mothers of children under 2. Source: MEH and various sources Figure 4.11: Coverage of Social Health Insurance SHI coverage by beneficiary category loo% 1 0Flat-rate contribution 90% 80% 70% 60% 50% 40% 30% 20% 10% w w 0% 4 I Source: Datafrom MOSWL 4.22 Over the last few years, approximately 25 percent of overall government health expenditure has been financed through the SHI. Prior to 2006, the arrangement was that budgetary expenditures (financed from general tax revenues) were used to cover the 'fixed costs' ''The other programsare: (i) retirementpensionfund; (ii) social benefits fund; (iii) occupational injury fund; and (iv) unemployment benefit. 84 Chapter 4 Health - of the operating expenses of facilities, as well as a core package of services that were provided for free regardless of insurance status (see Table 4.8). SHI coverage entitled participants to receive services and drugs outside the government financed `core package' for free or with minor co- payments. This included inpatient and outpatient services provided both by hospitals and primary care providers (soum hospitals and FGPs). Ineffect, the HIF reimbursed facilities for the variable costs associatedwith these drugs and services. Individuals without health insurance had to pay for drugs and services outside the core page through out-of-pocket payments, raising concerns about access for the uninsuredgroup. Table 4.8: Financingarrangementbefore and after reform Before2006 reform After 2006 reform Budget financing Medical emergency and ambulance Compulsory services providedby service FGPs, soum hospitals, and inter- TB, cancer, and mental disorders soum hospitals (maternitycare, Plannedimmunizations child health care, communicable Sanitation and disinfection disease prevention, care and Healthcare related to childbirth treatment, NCD(cardiovascular, Healthcare associatedwith natural mental disorder, cancer, care for disasters elderly, etc.), emergency and first Fixedcosts of state ownedhealth aid, and social health) facilities and FGPs (in practice, budget Some additional services provided financed approximately 40 percentof by higher level (referral) health FGPs, with the remaining60 percent institutions(diagnosis and comingfrom the HIF) treatment of cancer, infectious disease, mentaldisorders; obstetric and infant care; andresuscitation, intensive care, and ambulance diagnosis andtreat `Other services' providedby all health institutions(emergency and disaster response,research,expert training, social healthcareservices, mineralwater and sanatorium treatment). Full or partial financing Inpatientservices (variable costs) 0 Inpatientand outpatient services from HIF Outpatient services, includingFGP not includedin budget financed (variable costs) service packages(see above), Essentialdrug list including: Neurologicalsystem disorders Internaldiseases Ear, nose andthroat disorders, skin problems, muscle andjoint disorders Non-emergencytrauma and surgicaloperations 4.23 These financing arrangements have changed considerably as a result of amendments of the Health Act and the Health InsuranceLaw in 2006, and with the passing of other recent healthsector legislation.The aim ofthese reforms has beento broaden access to primary care (Le. to ensure equivalent services and benefits for the insured and the uninsured), and to create better incentives for health care providers. Key elements of the reform include the following: 85 Mongolia: Public Expenditure and Financial ManagementReview i. shiftfrom mixedbudgetandhealthinsurancefinancingto full budgetaryfinancing for primary care (FGPs and soum hospitals); ii. shift from budgetary financingfor fixed costs and health insurance financing for variable costs for secondary and tertiary hospitals, towards a model where the financing split is determinedby type (or `package') of service; iii.change in the capitation rate (payment per `covered' individual) for FGPs and introductionof capitation-basedfinancingfor variable costs of soum hospitals; iv. introductionof case-basedpayment (fixed paymentrate for differenttypes of `cases') for most services providedby secondary andtertiary hospitals;and v. provisions for state-owned health facilities to use financial surpluses for the improvementof working environment and socialprotectionof their healthpersonnel. 4.24 These reforms are likely to bring a number of challenges. What will be the future financing requirement for the budget and HIF re~pectively?'~How should payment rates for providers be set under the new payment systems? What capabilities and informationsystems are required on the government and provider side to manage new payment systems? Some of these questions are addressed in more detail below.However, before doing so, it is worth steppingback from the current system and asking whether the SHIprogram is achievingits intended aims, and what options the government has for improvingthe overallfinancing arrangements inthe future. 4.25 Most healthsystems have similar goals, whether healthcare is financed through tax revenues, S H I contributions,or a combination of the two. One goal is to reduce financial risk associated with health care expenditures by pooling resources for health care. Most countries want also want to achieve a `fair' distributionof the financing burden, typically meaning that the better off pay a larger, or at least a similar, share of their income on health as worse-off households. Effective risk pooling, combined with a fair distribution of the overall financing burden, lowers the financial barriers to seeking care, and hence serves to promote equity in the use of healthservices-another important goal inmost healthsystems. 4.26 The key difference between tax financing and S H I is that SHI contributions are earmarked for health and entitlement to benefits tends to be linked to contributions. In countries where the labor market is largely formalized, universalcoverage of healthcare benefits can, at least in principle, be achieved either through tax financing or income-related SHI contribution^.'^ In contexts where informal sector employment is important, it i s more difficult to achieve universal cover through contribution-based financing. This is due to the inherent difficulty in assessing the income for informal sector workers, and hence to relate health insurance contributions to income. If the SHI scheme cannot levy income-relatedcontributions for informal sector workers, it faces a tension betweenthe need to mobilize resources on the one hand and the goal of fair financing on the other-if flat-rate contributions are set too low, the scheme does not generate sufficient revenues for meaningfulbenefits, but if contributions are too high, the worse-off cannot afford to participate. This means that low- and middle-income countries with SHIsystems tend to go down one of two possible routes. The most common route is offer relatively generous benefitsto a small share of the population (formal sector and better- off informal sector workers). In these countries, general revenues are then used to finance a less l4For example, recentreforms have establishedthat the health insurancefund will no longer be responsiblefor primary care, but it will carry a larger share of hospitalcost. Most likely, the upshot will be an increase in expenditures of the health insurance fund. This increase will have to be financed either through increased contributions, increased budgetarytransfers to the health insurancefund, or a combination of the two. IsIn practice, many countries with SHI system also rely extensively on general revenues, either to finance health insurancecover for specific populationgroups, or to keep payrolltaxes low. For details, see Wagstaff, A. 2007. "Social Health InsuranceReexamined." World Bank PolicyResearchWorking Paper. 86 Chapter 4 Health - generous package of services for the uninsured population.Alternatively, the government could seek to achieve universal coverage through a SHImodel by financingthe contributions in part or in full for informal sector workers andthe poor.This is the route recentlytaken by Thailand. The challenge with this option is that it tends impose a significant demand on government resources. Moreover, by paying the contribution for informal sector workers and the poor from general revenues, the government risks creating incentives for informalization-why register your workers and pay SHI contributions if they can get similar benefits without making any contribution? 4.27 I n Mongolia, it is questionablewhether the S H I has succeeded in its intendedgoal of generatingadditional funding for the health system. After all, the reason why people may be more willing to contributeto a SHIscheme than paying general taxes is that they see a clear link between contributions and benefits. If that link is broken, as is the case in Mongolia, it is not obvious that the resources could not be mobilized as easily, and with similar or lesser distortionary effects, through the tax system. Mongolia has achieved high levels of enrolment in the SHI scheme. However, the share of enrollees making payroll contributions accounts for only 20 percent of members (Figure 4.12a). In 2005, this group accounted for 84 percent of total SHI financing. This may not be how the SHI scheme was originally conceived, but while per capita payroll contributionshave increased steadily since 1994 (more or less in pace with GDP), flat rate contributions and government subsidies have stayed more or less flat (Figure 4.12b). The group makingpayroll contributions use slightly more (or more expensive) services than the other members, but not nearly at a level commensurate with their contributions.As a result, this group i s heavily cross-subsidizingother SHI members. This situation has contributedto frustration and lack of trust in the SHI scheme among formal sector workers, who feel that they are payingtwice for healthcare, and still do not see their expectations met. Figure 4.12: S H I contributionsand t nefitsin Mongolia (a) (b) 9%1 __ -__ _____ _._.-- . __ 450 -Payroll contribution --Fiat-rate Contribution ;; 80 0% -Government - subsidy 400 GDPpc D 4 70 0% 3502 m 60 0% I $5o,ooo 3002 u i 50 0% i 250 1 es40,030 P 40 0% 200 30,WO 2" zr 30 0% 150 20 0% 100 li : n 10 0% 50 0 0% 0 m . . o O - N m Enrollees Contributions HIF expenditures m 2 2mn 2mr 2mm2 mmR Ro Zo Ro * ZoY Ec I subsidy8Payrollcontribution Government Flat-rate contribution Source: Datafrom MOSWL. Data in panel (a) arefor 2005, 4.28 The current design of the S H I system also raises questions about equity in access to services. Despite being compulsory, and despite the presence of government subsidies, over 20 percent of Mongolia's populationcurrently lack health insurance coverage. Inthe past, this meant that they faced restrictions in access to both primary care and specialist and inpatient services at referral level. This changed with the 2006 reforms, whereby primary care will be fully tax financed. However, as noted, these reforms also altered the role of budgetary finance for higher level facilities, shifting it away from covering 'fixed costs' towards a specific set of services. This is likely to result in higher prices for services that are not included in the budget financed 87 Mongolia: Public Expenditure and Financial ManagementReview packages (prices will now in principle include both fixed and variable costs). Hence, the reforms may restrict access to some hospital services for the uninsured. Socioeconomic inequalities in access to health care are, of course, to some extent inevitable, and the extent to which the health system should seek to reduce those inequalities is ultimately a political decision.For now, it will be important to monitor the consequences of the recent reforms on use of services and the financial burden of out-of-pocket payments for different groups, and to promote an open debate on whether the outcomes are compatiblewith the prioritiesand values ofthe population. 4.29 So far, the S H I system has not contributed in any significant way to improvements in service delivery in Mongolia. Ifanything, the fragmentationof financing sources has resulted in multiple funding flows to health care facilities, sometimes generating inconsistent or even conflicting incentives. It is often alleged that havingthe MOH both finance and manage facilities will create an inefficient and bloatedhealth sector-it is assumed that the MOH will spend more effort looking after the interests of health sector personnel than the population that they are supposed to serve. In principle, the introduction of SHI can help establish a clearer separation between financing and provision, and may in that way help strengthen the accountability for performance.'6For example, the SHI agency may be more active in promoting value-for-money through a more appropriate selection of services to cover, better mechanisms for paying providers, clearer performance standards, and so forth-functions often referred to as `purchasing'. Optionsfor the future 4.30 One option would be to repeal the S H I legislation and dismantle the S H I system. This would unify government healthexpenditure in a single agency. This, inturn, would facilitate the development of a core package of benefits that would be available to all. With a single purchaser, it would also be easier to develop payment systems that create coherent incentivesfor providers. If the government were to go down this route, it would have to put in place credible plans for makinga tax financed systemwork effectively.This would include ensuringa sufficient financing base (convertingpayroll contributionsto a regular payroll tax and, perhaps, over time increasing the reliance on other taxes). It would also include strengthening the capacity of the MOH in key areas. The government would also have to think through how the expectations of the better off for services not financed by the government can best be met. Most likely, this would haveto entail some form of government regulated or operatedvoluntary healthinsurancescheme. 4.31 I f the SHI scheme is retained, one needs to clarify the vision for the respective financing mechanisms. I s the goal of the SHI scheme to provide supplementary benefits for the better offwhile the tax revenues are usedto finance universalaccess to a core package of services or should the SHI be the mechanism for financing health services for all? The former model recognizes that there are limits to what the government can finance in the health sector. It opts to finance a core set of services through the budget on the basis of universal entitlement, while acknowledgingthat many formal sector employees and some informal sector workers have both the means and an interest in purchasing more extensive cover. The SHI system can cater for this group by poolingtheir resources and providing deeper and more extensive cover than offered by the budget-financed package. In many respects, this is not an ideal system-it effectively creates a two-tier system and separate purchasers of care. Nonetheless, the demand for complementary cover is a reality that the government will have to deal with in some way, and providing some of that cover througha government-runSHIscheme has many merits. Inany event, budget subsidies l6In practice, reformsto improve payment systems, monitor provider performance, ensureaccountability in the use of finds, and ensure appropriate levels of provider autonomy can be-and havebeen-implemented in both SHI and tax financed systems. Of course, in some cases, the introduction of SHI can be the only politically feasible way to change institutions and incentives in the health sector. 88 Chapter 4 - Health to the SHI should arguably be phased out, and the extent o f cross-subsidization by members that are making payroll contributions should be reduced. If,on the other hand, the SHI scheme is seen as an instrument for achieving universal cover, ways must be found to make coverage truly universal, and for converting the SHI scheme into the primary purchaser o f health services. 4.32 Recent reformsseem to indicatethat Mongolia is choosing neither of these options. They leave in place a system where a minority (15 percent) o f the population pay payroll contributions and subsidize some hospital services for other SHI members, but that still leaves nearly a quarter o f the population having to finance these same hospital services out-of-pocket. If these services are deemed part o f a core package o f services that should be universally accessible, should they not also be subsidized for the population that is currently not covered by SHI? Conversely, if these services are not part o f core services, why are payroll contributions used to subsidize access for some but not for others? 4.33 If the S H I is developed as a model for providinguniversalcoverage, it is important to be realistic about the scope for contribution based financing in the short- to medium- term. There may be some scope for increasing contributions. The government could, for example, increase payroll contributions. However, this is likely increase frustration among formal sector workers and may contribute to informalization in the labor market. The government could also try to increase contributions from the informal sector. However, as noted, increased flat rate contributions raise equity concerns, and are also likely to make enrollment more difficult to enforce. Hence, even with a SHI system, general revenues are likely to remain the predominant financing source for some time. Moving forward, it is also important recognize that a system that relies on payroll contributions for formal sector workers and low flat-rate contributions for informal sector workers may create adverse incentives in the labor market. Individuals face a clear disincentive for joining and staying in formal employment ifthey can enjoy similar benefits as self-employed or informal sector workers (or even by leaving the scheme entirely) with significantly lower financial contributions. By the same token, the system provides a disincentive for individuals that are enjoying government subsidized cover from getting work or otherwise change his or her status. These incentive effects have been shown to be important in some countries, but there is currently little information on the extent o f `informalization' and other incentive effects in Mongolia. Improving purchasing 4.34 At the moment, purchasing functions in Mongolia are not being performed adequately. There is a need to strengthen purchasing. This means that the agencies that hold funds for health care-currently the MOH and the SSIGO-need to become more active in safeguarding the interest o f the beneficiaries. This includes making sure that funds are allocated to the drugs and services that will provide the greatest benefits to patients, establishing methods o f financing or paying facilities that promote efficiency, quality, and cost control, establish and monitor clear performance standards, hold providers accountable for performance, and so forth. 4.35 To addressweaknesses in purchasing,one needs to address a mix of policy, capacity and institutionalissues.For example, the practice o f financing hospitals through a combination o f detailed line-item budgets (state budget) and fixed payments per episode (HIF) was established through decrees and legislation. The methods for hospital financing have now been changed, although significant implementation challenges remain. Weaknesses in purchasing are also the consequence o f limited capacity. Neither the MOH nor the SSIGO currently have the systems or skills required to manage a more sophisticated purchasing process. Finally, institutional fragmentation and coordination failures between MOLSW, MOH, and MOF, have stood in the way o f improving purchasing. For example, hospitals currently have separate contracts with 89 MonPolia: Public ExDenditure and Financial ManaPement Review SSIGO and MOH. Moreover, both the SSIGO and the MOH largely reimburse hospitals, and to not monitor quality in a coherent way." The current split is not only efficient, but also undermines coherency in the incentivesfor providers and makes effective prioritization difficult. It makes little sense for scarce capacity to be spread across different `purchasing' agencies. 4.36 A key step for the future will hencebe institutionalstrengtheningthrough improved skills, better management information systems (on both the purchaser and hospital side), and strengthened management systems and procedures. Important policy reforms have already beentaken in the area of providerpayment. However, systems and capacity to implement these reforms, and to perform complementary functions, are still lacking. Achieving this will dependon strong inter-agencycoordination,or possibly even integrationof purchasing in a single agency. THEDELIVERY OFHEALTHSERVICES 4.37 Many features of the socialist period service delivery system were still in place in 1999 in Mongolia.At that time, the healthcare delivery system was heavily orientedtowards the hospital sector at central, city, and aimag level, while primary care was provided polyclinics in urban areas, and aimag clinics (associatedwith the aimag hospital), soum hospitals, andfeldshers in rural areas. There has been some restructuringin the hospital sector, with a reduction in bed numbers from 10 to 7.4 per 1,000 people between 1994 and 1999, but little change since then (Table 4.9). Table 4.9: Bedsand facility infrastructureinMongolia 1999 2000 2001 2002 2003 2004 2005 Total hospital beds 17,877 17,974 18,090 18,616 18,180 18,371 18,285 Beds per 1,000 pop. 7.4 7.6 7.5 7.6 7.3 7.3 7.2 State Health Institutions 396 396 373 373 385 374 365 Tertiary level hospitals and centers 19 19 17 17 19 17 17 Aimags and district hospitals 30 30 33 33 45 30 30 Intersoum hospitals 13 13 19 31 31 31 Soum hospitals 347 334 310 304 290 296 287 Private Health Institutions 448 565 658 769 842 807 911 Private hospitals and clinics 448 466 480 536 611 577 683 Family Group Practices NIA 99 178 233 23 1 230 228 Drug stores 628 619 721 770 724 770 735 Total 844 961 1,03 1 1,142 1,227 1,181 1,276 Note: From 1999, 2000. and 2005 Statistical Yearbooks 4.38 Since the beginning of the decade, several changes have taken place. At primary level, the number of soum hospitalshas declined by nearly a third, to 287 in 2005. At the same time, inter-soumhospitalshave beenestablished as centers providing simple referralcare in areas far from aimag centers. The government has also established Family Group Practices (FGPs), organized as private entities providing primary care services on the basis of capitation-based contracts. Since their introduction in 1999, over 200 FGPs have been established. The aim of the reformwas to shift the emphasis of the healthcare system from hospitalcare to primary care and from curative to preventive services. However, the FGP reforms have not yet achieved their intended purpose. Perhapsthe most important shift on the delivery side has beenthe expansion of l7 Health Department of State Inspection Agency is responsible for quality monitoring and licensing of hospitals in Mongolia. This regulatory function is important. However, as financiers of care, the MOH and SSIGO, are in a better position to monitor quality on a routine basis, and to take corrective measures when necessary. This includes care providedby private hospitals and clinics. 90 Chauter 4 Health - and other public health functions, such as 120 - epidemiological surveillance, 100 . I environmental health, infectious disease k.rtr1j.n control, behavior change interventions, 89 : 8 0 . KazIkhltln .. :....,I. . . preparedness for public healthIg 6 0 . I UzD.k,.t.n emergencies, and so forth, are also . YIyrqyz RoPUbIIC t. Arqantln. critical. Duringthe Semashko system, the L o . C 0 ' ' 511 Lank. Sanitary Epidemiological Services (SES) ** C+%d &y 1 . 20. .* was responsible for public health. Since .&"&p% ; C'"b9"O*p .:e. .. 0 7 I 91 Mongolia: Public Expenditure and Financial Management Review (SSIA) as a separate entity responsible for coordination of various inspections that previously were under different public sector entities, including the SES. A number of other specialized agencies, such as the Cancer Center, the Blood Center, and the Infectious Disease Research Center remain. This system faces considerable challenges. Efficiency of service delivery 4.41 One of the most notable features of the Mongolian service delivery system is the extensive infrastructure and large workforce. As noted, these features of the system have undoubtedly contributed to some of its achievements, but are also the source of important inefficiencies. For example, notwithstanding reductions in bed numbers during the 1990s, Mongolia still stands out in an internationalcontext as having an unusuallyhigh number of beds per population (Figure 4.13). Similarly, on the health workforce side, Mongolia has a large number of health workers (doctors and nurses) relative to population (Figure 4.14a), and a high physician to nurse ratio (Figure 4.14b). Moreover, there are notable differences across aimags in the density of healthprofessionals, with aimags that have a high number of physiciansalso being more likely to have a highnumber of nurses (Figure4.15). Figure 4.14: Health workforce high as well 2.50 Colombia* -c 2.00 . 0 2 China f 2 . O 0 'G 1.00 m n 1 ., 1.50 * w c .. 0.50 0 50 ' Tp@nd * ' 8- .Samb&aneaa ,. $* * 0.00 ~ 000 d w : I 0 5,000 10.000 15,000 20,000 0 5,000 10,000 15,000 20,000 GDP pc (PPP) GDP pc (PPP) NOTE: Graphexcluder three countreel in the relevantinwrnc range(urvguav, Argentina, and Lebanon). These wuntresall havephyslclan/nurre ratlm In excess of 2.5 Source Datafrom 2006 WorldHealth Report (WHO) Figure 4.15: Human resourcesat aimag level Ulaanbaata Orkhon Darkhan- Govisumber Oornogovl Govi-Altal Dornod Dundgovi uvs Selenge Bulgan Khentil Sukhbaatar Omnogovl Khovd TOY Bayankhon Khovsgol - Arkhangai Zavkhan Ovorkhang Bayan-01g11 .. 0 1 2 3 4 5 6 7 8 9 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 Nurmcs+doctors per 1,000 pop. Doctors per 1,000 pop. Source: Datafrom 2005 Statistical Yearbook 92 Chapter 4 Health - 4.42 Extensive physical and human resources are accompanied by high levels of service utilization and there appear to be substantial differences in productivity across different aimags. International data on the utilization or inpatient services are not routinely available. However, on the basis of a limited set of comparator countries, Mongolia, with nearly 0.25 inpatient discharges per capita per year, stands out in comparison with other countries for which data are available (Figure 4.16). Moreover, there are substantial disparities in hospitalization rates across aimags, well beyond what would be expected on the basis of epidemiological differences (Figure 4.17a). It is also striking to note that, contrary to what might be expected, aimags with a large health workforce relative to population are do not have higher hospitalization rates (Figure 4.17b). Figure4.16: Use of inpatientservices in selected countries Mexlco China Turkey Canada Japan Spain Korea Poland Switzerland Australla Denmark Germany Mongolla Hungary France 0.00 0.05 0.10 0 15 0 20 0.25 0.30 Inpatient discharges per caplta Source. Datafrom OECD Health Data 2005; China Health Statistical Yearbook2005; Mongolia Statistical Yearbook. Data arefor different years (2001-2004). 4.43 Available data suggests that there is considerable scope for improving efficiency without adverse consequences for either health outcomes or equity in access to services, even though international and inter-aimag comparisons of bed and health worker density are revealing, but need to be treated with caution, as political priorities very across countries, and needs and circumstances vary across both countries and aimags. Figure4.17: Use of inpatientservices at aimaglevel (a) (b) .. .... 350.0 daanoaatar 7 Govi-A ta r v s d Arkhangal 0 Eayankhongor 0 Dornogovi 300.0 Khovd Zavkhan 6*: - Oornod n Eayan-Olgli Khentil % 250.0 - Khovsgol Selenge VI Dundgovi 0 I Darkhan-Uul m Bulgan 200.0 - Omnogovi 5ukhbaatar Ovorkhangai =.-f Orkhon TOY Orkhon 150.0 4 I 0 50 100 150 200 250 300 350 400 3 4 5 6 7 8 9 Hospltallzatlons per 1000 pop. Doctors and nurses per 1,000 pop. Source: Datafrom 2005 Statistical Yearbook 93 Mongolia: Public Expenditure and Financial Management Review 4.44 Improvements in efficiency require a rationalizationof facility infrastructure, a re- structuring of the health workforce, and systems and incentives for reducing unnecessary inpatient admissions. This situation is not dissimilar to that faced by many eastern and central European countries in the early 1990s. Experiences from reforms in these countries provide important lessons for Mongolia. They suggest that reductions in bed numbers and provider payment reforms-the measuresthat have received most attentionin Mongolia so far-tend to be ineffective in reducinghospital capacity on their own. Many countries introducing new provider payment mechanisms (such as diagnosis-relatedgroup-based system) in the hope of driving a restructuringof the hospitalsector achieved a reductionin length of stay but also to saw a rise in the number of admissions as hospitals compensated for the lower payments they were receiving for each admission. Alternatively, hospitals faced with reductions in funding may allow their capital stock to deteriorate or start running deficits. Successful cases of hospital reform suggest that closing or mergingof facilities is requiredin order to achieve significant savings. Ratherthan merely relying on financing as a lever to drive such changes, careful planning is required to determine how a reformedhospital system can best meet the healthneeds of the population, and how to manage the restructuring process. Effective hospital rationalization will also need to safeguard the interest of key stakeholders. This includes health professionals, for whom a rationalization is likely to have far-reaching consequences, but also patients who have tended to rely on the hospitalsector for long-termand socialcare. 4.45 There is a broad consensus that hospital rationalizationis needed in Mongolia and the main bottleneckfor reform appears to be the lack of an inclusive processof consultation and consensus building. Over the last decade, a number of reports have been prepared that explore different scenarios for hospitalrationalization,suggesting that significant savings can be had. Hence, at a technical level, a lot of the ground-work for hospitalreformhas been done. Yet, so far, progress inimplementingreformshas beenslow andbuilding consensus in needed. Betterpaymentmethods for hospitals 4.46 To some degree, current inefficienciesare the product of how hospitals have been financed, combined with past investments in health system infrastructure and workforce. Until recently, the payment system was based on a fixed payment per case from the HIF, combined with a highly detailed line-item allocation from the government budget based on both historical allocationsand bed occupancy rates.18This system provided incentivesfor hospitals to have more beds, increase admissions, and keep the beds occupied, but did not stimulate cost control, quality improvement, or efficiency. 4.47 Recognizing these problems, the government introduced reforms in 2006 to clarify the role of budgetary finance, and to move towards a system of case-basedpayment from the HIF." There is ample internationalexperience to suggest that a move away from line-item budgetingor fee-for-service payment towards case-based payment can help improve hospital efficiency, so these reforms are undoubtedly a step in the right direction. However, the introduction of case- basedpayment in hospitals is rarelywithout its problems. 4.48 Provider payment reforms need to be accompanied by strong systems and capacity and gradually. The development and introductionof case-based payment systems is a time and capacity intensive exercise, often taking many years. There needs to be adequate preparatory The HIFprovidedprospective funds on a monthly basis, based on expected admissions. l 9Resolution 436/451/127 of 29 Dec. 2006 establishes payment rates from the HIF for both inpatient and outpatient care, comprised on basic tariff rates for inpatient episodes, day-care episodes, and outpatient visits for different types of institutions, and factors (reflecting relative cost) for 17 different types of inpatient and day treatment services and two categories of outpatient services 94 Chauter 4 - Health work on issues such as case classification, costing and rate setting, and information systems.20 Moreover, there needs to be strong systems and capacity to manage some of the risks associated with case-based payment. For example, many countries have seen an increase in the volume of services following payment reform, and have been forced to introduce volume or budget caps in order to control costs. Payment by case also introduces incentivesto skimp on quality and turn away complex (and expensive) cases. And there is always a risk that hospitals `game' the payment systems by classifying cases in ways that increase their revenues. Giventhat it will take time to develop the requisite capacity and systems in the MOH, MOLSW, and SSIGO, there is a strong case for proceeding with reforms on a gradual basis, starting with careful piloting and evaluationinthe first stage. 4.49 The new reforms are a challenge to current financial management practices. One key aspect of recent reforms is the introduction of `output-based' allocation of resources to providers. In the case of the hospital sector, the volume of different categories of patient cases will be used to allocate health insurance funds will health insurance funds (hospitals will also receive budgetary financing, but the allocationcriteria to be used is still unclear). These changes have the potential of making the allocation of resources more rational. Nonetheless, they also raise a number of issuesrelatedto current financial managementand civil service legislation: (i)The introductionof case-basedpaymentwill requirea shift away from the current practice of line-item budgeting. The PSFML allows for such a shift, but it would run counter to current practice. The shift will require intensive training in line-ministries finance departments. (ii)The allocation of resources on the basis of case mix and volume may result in substantial changesrelativeto current budgets.This may be appropriate, but the transition will need to be carefully managed in order to avoid service disruptions and other problems. This could, for example, take the form of a system of `bridge financing' to managethe transition. (iii)The logic of the payment reforms is that providers should be left to decide on the most efficient and appropriate way to meet patients' needs. In order for this to work, hospitals should be empoweredto make key decisions that influence cost and efficiency. At the moment, civil service legislationand restrictions on investmentdecisions make it difficult for hospitalsto make adjustments to their workforce or other aspects of capacity. (iv)In order for new payment methods to impact on performance, providers that are efficient and attract patients must be suitably rewarded. Different incentivesschemes can be envisagedto do so. For instance, inAustraliaand Singapore, a productivitycomponent has been created, which allows drawing from competitive fund, providedthat the public sector unit is able to show high performance. In such case, what is needed is a good Monitoring and Evaluation (M&E) framework which can judge performance of service delivery within a sector. 4.50 Hence, the success of recent provider payment reforms depend in part on the development of systems and capacity within the MOH and the MOSWL and the Civil Service Council (CSC), and on improved coordination between the two ministries and CSC. Capacity buildingthrough training is part of the solution, so that the shift to the new system can take place over time. However, equally important, the reforms are unlikely to have their intended impact 2o Some decisions have already been taken in regard to these issues in Mongolia. For example, joint resolution 436/451/127 of the Minister of Health, Minister of Financeand Minister of Social Welfare and Labor on December 29 of 2006 establish paymentrates for different categories of providers and cases. However, these policies were developed with very limited backgroundwork, and it would be risky to proceedwith implementation on this basis. 95 Mongolia: PublicExpenditureandFinancialManagementReview unless the MOF and other relevant central ministries and agencies do not help address some of the risksand challengesoutlinedabove. Improving primary healthcare services 4.5 1 While the hospital sector consumes a large share of government health expenditures, improvements in the primary care system are also fundamentalto enhance overall efficiency in the service delivery system. In order to achieve these aims, there is a need to further re-orient soum hospitals towards preventionand primary care, make the FGP system work better, and put in place an effective gate-keepingsystem that reduces the proportionof non-specialistoutpatient care providedby referral level facilities. The shift to full budget finance for these facilities has the potentialto facilitate these changes, but other steps are also needed. 4.52 Enhancing capacity and trust in the primary care system can be done through enhanced health provider compensation and payment methods. It has often been noted that primary health care providers, both FGP and staff in soum hospitals, have limited technical capacity. This is an important factor in undermining public trust in the primary care system. Improvingcapacity is partly a questionof on-the-job training. However, inthe longrun, it will be more important to ensure that the level of income and professional status of primary care providers are sufficiently high to attract staff. For example, FGP doctors currently have an income of around 110,000-120,000 Tugrik per month, comparedto 160,000-170,000 Tugrik for physicians in aimag hospitals. In order to improve compensation, the system for financing primary care facilities will require reform. In the past, soum hospitals were financed through a detailed line-item budget from the MOH, while FGPs receivedcapitationpayments from boththe state budget and the HIF-Le. a fixed amount for each person registeredwith the facility.2' Since 2006, both soum hospitals are to be fully financed from the state budget, and both types of facilities will be paid on a capitationbasis. This reform has many potential benefits.However, as was the case with the introductionof case-based payment for hospitals, it needs to be carefully managed. 4.53 Afirst challenge is to getpayment rates right. As noted, capitationpaymentrates must be sufficiently high to motivate health workers at the primary level, and to attract young and qualified doctors and nurses to the profession.All available information suggests that this is not the case at the moment. A second challenge is to get provider incentives right. Capitation paymentscreate incentivesfor cost-controland efficiency, but can also result in 'under-servicing' of patients4.g. doctors providing fewer services than clinically appropriate, or simply referring patientsto hospitalsto higher levelrather than treatingthem themselves (this saves money for the doctor, but leaves revenues unchanged). One important measure to mitigate this risk is to put in place strong systems to monitor quality, assess the appropriateness of referrals, enforce gate- keeping arrangements, and so forth. For the most part, these systems are currently missing in Mongolia, and will clearly need attentioninthe future.A third challenges is that it is also unusual for primary careproviders to bepaidpurely on capitation basis. Indeed, capitationpaymentsare typically combinedwith fee-for-servicepayment or performancebonuses in order to incentivize the delivery of priority public health interventions (e.g. immunizations, screening for sexually transmitted diseases, healthpromotion,etc.). 4.54 Capitation payment also introduces special challenges in the case of soum hospitals, which have previously been financed through line-item budgets. Consider, for example, what would happen if a soum experiences rapid out-migration. In this case, capitation-based funding would lead in a reductionin financing for the soum hospital over time. The reduction in funding i s arguably appropriate, and should help drive a gradual scaling down of staff and capacity. *'Currently,FGPs receive 2,450 Tugrik per capita per year in Ulaanbaatar, and 3,400 Tugrik for other areas. 96 Chapter 4 - Health However, this assumes that the aimag health department or hospital director being able to make the necessary staffing or investment decisions, even though they may be politically unpopular.If these difficult decisions are not made, the fundingmodelwill leadto an ever-growingshare of the budget being absorbed by staff costs in soums with declining populations. Soums also vary substantiallyin populationdensity, geography, health needs, and so forth. These differences can have significant implications for the cost of service delivery, and these differences need to somehow be reflectedinpaymentrates. 4.55 In sum, the recent payment reforms have the potential for improving primary health care in Mongolia but also bring significant risks. The reforms will take a longtime to implement, andthe MOHneedsto make substantial investments in systems in capacityto manage against some of the risksthat are inherent in the reforms. It is also importantto recognizethat, as in the case with hospital rationalization,financial incentives are unlikely to sufficient on their own to drive a rationalizationprocess. Carefulplanning is required in order to identify medium- and long-termneeds, develop options for meeting these needs through a more efficient delivery system, andto managethe transition. 97 CHAPTER 5. SOCIAL PROTECTION PROGRESSAND CHALLENGES Social assistance . The 2003 Social Security Sector Master Plan aimed at consolidating and rationalizing the system which was overly complex, and improving cost-efficiency. The Child Money Program (CMP) responded to these objectives and consolidated all child-related allowance programs. The CMP was successful in conditioning benefits on positive social behaviour. For example, the program had a large impact on high school enrolment among the poor. The move to a proxy-mean test was an outstanding . improvement. Nevertheless, the impact among the poor was made at highcost because o fmis-targeting to non-poor. The CMP was universalized in July 2006 and the benefit increased to MNT 136,000 (US$ 116) per child as o f January 2008. The CMP income is predicted to account for 34 and 20 percent o f total household consumption among the poorest and second poorest quintiles. The large size o f the program raisedconcerns about its fiscal sustainability and pro-cyclical nature. Social insurance .. Examining contributions and benefits together, the overall deficit between contributions and benefits has grown from less than one percent o f GDP in2001 to almost three percent o f GDP for 2007. The adoption o f an NDC scheme in 1999 improved the inter-generational sustainability, equity and predictability o f the Pension Insurance Fund. But it did not fundamentally accommodate the special . needs o f herders, the informal sector, and those formal sector workers with insufficient lifetime contributions to support a pension or who otherwise fall into poverty duringretirement. Problems include: (i) The minimumpension framework is too generous and creates weak incentives. (ii) retirementageistoolowtoensureeitherfinancialsustainabilityormeaningfulareplacement The rate. (iii)Actual indexation has resultedin many cases in a real depreciation o f the real benefit value. (iv) Disability benefits are sufficiently generous to pose a fiscal burden and weak incentives under the defined-benefit scheme and are inappropriately financed inthe cas the NDC scheme. KEYFUTURE PRIORITIES Social assistance i. Inorder to maximize its social impactat the least cost, the CMP should returnto cost-effectively targeting the poor and make benefits conditional to human capital investments. Preparation work should focus on refining targeting, transparency and credibility o f the targeting criteria reaching the poorest, ex-ante analysis on the benefit rate, and implementation arrangements. 9 Suggested"how to": (i)Commission anevidence-based studyto assess ofthe mainsocial risksfaced by different groups and identify who is most at risk in each case; (ii)Work out a time plan for the gradual re-targeting o f CMP to the very poor. This can be done by setting increasingly lower poverty lines (i.e. cutoffs inthe distribution o f the proxy means measure that are used to identify the poor) over time so that an increasingly smaller percentage o f families get the benefit untilonly very poor families get the benefit; (iii) Organize consultations, with participation from the civil society, on the targeting tool and to gain public support for the gradual re-targeting o f CMP to the poor based on the evidence presented inthis report and others; (iv) Train local social workers, put in place mechanisms to make 99 Mongolia: Public Exoenditure and Financial Management Review them accountable, including a complain mechanism, and set up an individual-level registry to track beneficiaries. Social insurance 11. Review the societal objectives o f the combined public and private pension and social security system to provide: (i)a minimum living subsistence to the elderly, permanently disabled and survivor/dependents; and (ii) mandatory consumption smoothing for most formal sector workers and voluntary contractual savings for informal sector workers. > Suggested "how to": (i)Convene a Pensions Policy Reform Steering Group and appoint a Technical Secretariat to undertake technical work supporting option s and decision posed to the steering Group; (ii)Prepare a White Paper on Pension Reform Objectives and Strategy which spells out the Government's key objectives for the public pension system and the principles behindreforms to policy parametersandinstitutional configuration. iii.Retaintheexistingdesigncharacteristicsofthecurrentsocialinsurancesystembuttostrengthenfiscal sustainability, affordability, predictability and equity o f benefits, revise the provisions for retirement age, benefit indexation, minimum pension provisions, the wage base for determining benefits, the formulation andfinancing o fthe disability benefit., > Suggested "how to": Engage in a process o f reviewing the impact o f parametric reforms on pension system sustainability as well as on benefit entitlements by simulating various reform option using actuarial projection tools. 100 Chapter 5 -Social Protection OVERVIEW OF SOCIALPROTECTION SPENDING 5.1 Since 2001, social protection expenditures have increased substantially. Budgeted social protection expenditures for 2007 are estimated to be about 8.9 percent o f 2007 GDP, having risen substantially from about 7.0 percent o f GDP between 2001 and 2005. 5.2 Both social insurance and social assistance expenditures have expanded.' On the social assistance side, much o f the reason for the increase during this period has been the provision o f the so-called Child Money Program benefits since 2005 which in 2007 amounted to about 23 percent o f social protection expenditures. On the social insurance side, old age pensions represent the largest single category of social insurance expenditure, o f about 4.3 percent of GDP in 2007 or about 38 percent o f all social insurance expenditures, having risen from 3.3 percent of GDP in 2002 to 4.3 percent in 2007 (See Table 5.1 ). Taken together, old age, disability, survivors and military pensions represented about half o f all social protection expenditures. Disability pensions amounted to 0.8 percent o f GDP during 2007 or about 7.2 percent o f all social insurance expenditures. Table 5.1 Social protectionexpenditures Billions TG % total expenditures %GDP -2007 2008 2002 - - - - - 2002 2007 2008 - 2002 2007 2008 Pensionand benefit 73 249 366 13.3% 14.2% 14.3% 5.2% 5.5% 6.7% Old age pension 47 161 236 8.5% 9.2% 9.2% 3.3% 3.5% 4.3% Disability pension 7 31 45 1.4% 1.8% 1.8% 0.5% 0.7% 0.8% Survivors' pension 8 24 31 1.4% 1.4% 1.2% 0.6% 0.5% 0.6% Military pension 4 14 22 0.8% 0.8% 0.9% 0.3% 0.3% 0.4% Sickness benefits 1 2 2 0.2% 0.1% 0.1% 0.1% 0.0% 0.0% Funeralbenefit 2 2 2 0.3% 0.1% 0.1% 0.1% 0.0% 0.0% Pregnancyand maternity benefit 1 3 6 0.2% 0.2% 0.2% 0.1% 0.1% 0.1% Work accident'disease pension & benefit 2 I O 18 0.4% 0.6% 0.7% 0.2% 0.2% 0.3% Unemploymentbenefit 1 2 3 0.1% 0.1% 0.1% 0.0% 0.0% 0.0% Other 0 0 0 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% SocialWelfare Expenses 17 237 255 3.2% 13.6% 10.0% 1.2% 5.2% 4.7% Child Money Program 0 143 142 0.0% 8.1% 5.6% 0.0% 3.1% 2.6% Welfare pension 4 16 24 0.8% 0.9% 0.9% 0.3% 0.3% 0.4% Pregnancyand maternityleavebenefit 4 17 19 0.8% 1.0% 0.7% 0.3% 0.4% 0.3% Newly weds allowance 0 31 18 0.0% 1.7% 0.7% 0.0% 0.7% 0.3% Conditionalcash allowance (lunch program) 0 14 12 0.0% 0.8% 0.5% 0.0% 0.3% 0.2% Prizedmothers medal allowance 0 I O 12 0.0% 0.6% 0.5% 0.0% 0.2% 0.2% Supportto the elders and veterans 2 2 9 0.4% 0.1% 0.3% 0.2% 0.0% 0.2% Newlybornallowance 0 6 7 0.0% 0.3% 0.3% 0.0% 0.1% 0.1% Other 7 143 156 1.2% 8.1% 6.1% 0.5% 3.1% 2.9% Total 91 487 621 16% 28% 24% 6.4% 10.7% 11.4% 2008: Plannedbudget Source: MOF and MSWL 5.3 This chapter aims to provide an analysis of the current social protection spending system in Mongolia and suggest possible ways forward for reforms. The social assistance section provides an evolving picture o f the social assistance system in Mongolia and assesses the system's ability to address the needs o f the poor and vulnerable in a cost-effective way while 1 Social insuranceis a contribution-basedpensionscheme and social assistance is a general revenuefunded targetedassistanceprogram. 101 Mongolia: Public Expenditure and Financial Management Review providing the right incentives. The analysis is mainly focused on the Child Money Program (CMP) as the central piece o f the social assistance system in recent years. The analysis is driven by the principles o f a modern social safety net system, namely that: (1) it addresses the main risks faced by the different demographic groups o fthe population (children, youth, adults and the elderly); (2) it reaches those who are at risk (Le. the poor and vulnerable), and only those; (3) it tries to prevent risks, not just mitigate and cope with them; and (4) it induces individuals and households to take the necessary actions to reduce exposure to risk (`corresponsibility'). Social assistance programs should avoid welfare-dependency and incentivize (or at least not discourage) work. 5.4 The social insurance section identifies weaknesses in the current social security system and suggests policy and institutional reform options for consideration by the authorities. It reviews the recent historical development o f the Mongolian Pension System; describes current parameters; examines the recent evolution o f public expenditures on social insurance and reviews key reform issues and suggests policy and institutional reform options. SOCIAL ASSISTANCE Key milestonesand overall trends 5.5 In its transition to a market economy, Mongolia adopted a social assistance system to address the needsof peoplewho lost their traditionalsafety net-jobs and pensions-and were not covered by social insurance.The public safety net system included two main types o f programs: social insurance (retirement, unemployment, or sickness); social assistance (specific benefits to protect vulnerable groups, including the disabled and child allowances); and active labor market programs (ALMP) (job mediation services, entrepreneurship services, job training, public works and wage subsidies). Social insurance is only available to workers contributing to the social security system (40 percent o f workers in 2006-HIES 2006). Up until the 2006 amendments to the Labor Law, ALMP have been limited to the unemployed who are registered at the Labor and Social Welfare Assistance Divisions (54 percent o f the total unemployed, or about 30,000 people-LSMS 2002-03). The main legal framework for social assistance benefits was provided by the 1995 Social Welfare Law and the 2001 Employment Promotion Law. 5.6 The 2003 Social Security Sector Master Plan (SSMP) had the right ideas but no clear implementation strategy for consolidating and rationalizing the system to achieve a coherent social assistancestrategy. The SSMP aimed to consolidate and rationalize the system and improve cost-efficiency (mainly through improved targeting). In particular, the main areas o f reform o f social assistance identified in the master plan were: (a) to rationalize the number and type o f benefits; (b) to clarify the eligibility criteria o f the different benefits to improve quality and accessibility o f the sector's services; (c) to decentralize the management o f social assistance services; (d) to professionalize the social welfare officers in charge o f implementing social assistance; and (e) to promote community participation in specific care-provider institutions. 5.7 The introductionof the ChildMoneyProgram(CMP) in January 2005 respondedin part to the strategy laid out in the SSMP, in particular the need to consolidate programs and improvetargeting.The CMP was originally designed as a conditional cash transfer program (CCT) targeted to poor households with children aged 0-18. It introduced two important innovations in the social assistance system: it conditionedthe benefit receipt to specific behaviors on the part o f the households; and it introduced a proxy-means test for the identification o f the poor. Cash transfers were conditioned on following mandatory immunizations, living with parents, not being engaged in intolerable forms o f child labor and, for children between the ages o f 8 and 18, being enrolled in school. 102 ChaDter 5 -Social Protection 5.8 The CMP introduced the principle of beneficiaries being selected on the basis of need through an objective set of indicators, replacingthe old system of subjective assessment of poverty by local authorities. Although the objectives were not clearly spelled out, the basic design of CMP mimics that of successful CCT elsewhere, which have the dual goal of reducing current poverty (by providing monetary transfers to poor families) and reducing future poverty through improvedhumancapital(by conditioningbenefitson human capital investments). 5.9 The amendments to the SocialWelfare Law of July 2006 universalizedthe CMP and introduced other large allowances-all facilitated by large revenues from mining.The CMP was universalizedto all families with children in July 2006 and the benefit was increased in January 2007 from 36,000 MNT per child a year (3,000 MNT per month) to 136,000 MNT. The CMP was moved out of the budget into the newly established Development Fund (DF), establishedwith the objectiveof managingthe fiscal revenue from the windfall tax on copper and gold production. Also, two large one time allowances were introduced-newly weds (500,000 MNT) and newborns (100,000 MNT). The July 2006 amendments also consolidated all child- relatedallowances (other than CMP) into a single CCT of 20,000 MNT per month. 5.10 Social assistance expenditures have been increasing continuously since 1997, but they skyrocketed after introduction of the CMP in 2005 and the subsequent changes, resulting in large increases in the share of social assistance in the total budget and GDP. Between 1997 and 2004, social assistance increased by 127 percent in real terms. The introductionof the CMP in 2005 increased expenditures by 61 percent, and further increased by 54 percent with universalizationof CMP in 2006. The January 2007 benefit raise has increased projected expenditures by 173 percent. Over that period, the share of social assistance in the budget and GDP went from 2.2 percent and 0.7 percent in 1997, respectively, to a projected 11.9 percent and 5.2 percent in 2007, respectively. Figure 5.1: A dramatic increase in social Figure 5.2: Which has translated into large assistance particularly since 2005 shares of the budget and GDP 1 (real index of wcial assistance 1997-100; projectedfor 2007) 14.0 12.0 10.0 8 0 6 0 ' 4 0 ~ 2 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 +-%ofbudget+%of GDP Source: Ministry of FinanceandNational Statistical Once (nS0).Figuresfor 2007 are based onprojected budget, integrating CMP,and GDP. 5.1 1 Soon after its introduction in January 2005, the CMP quicMy became the largest social assistance program, accounting for 67 percent of beneficiaries and 40 percent of the budget in 2005. By April-June of 2006 (right before the universalization of the program), 68 percent of families with childrenwere benefitingfrom the program, which covered 29 percent of all children aged 0-18 in the country. In fact, in 2006 the CMP represented the main source of welfare payments to households together with the traditional main source of welfare payments (pensions) (see Figure5.4 and 5.5). It accounted for 44 percent of welfare paymentsto an average 103 Mongolia: Public Expenditure and Financial Management Review household42 percent among households receiving CMP benefit-and 8 percent o f their income. Social assistance inthe pre-CMP era-the right idea but inadequate system 5.12 The social assistance system that emerged from the transition had (and still has) three basic types of programs: social welfare pensions, social assistance allowances (child- related allowances) and social welfare services and concessions. Except for child-related allowances, which have been recently consolidated into a single CCT, these programs are largely still in place: (i)Social welfare pensions are long-term cash benefits given to very poor elderly (as assessed by the local governor), people with permanent disabilities, and survivors for individuals, provided they are not covered by social insurance. In 2004 (before the CMP was introduced), social welfare pensions accounted for 11 percent o f beneficiaries and 29 percent o f the total social assistance budget. In 2006, there were 45,449 beneficiaries receiving an average monthly benefit o f 26,500 MNT, accounting for about 1 percent o f the total consolidated government budget. (ii)Social assistance allowances are child-related cash benefits given to families with children. Before the July 2006 amendments to the Welfare Law, they included allowances for (1) pregnancy and delivery (short term benefits given to all women not covered by social insurance); (2) children in large families (long term benefits for all families with 4 or more children); (3) child care (short term benefits given to care giver staying at home); (4) infant care (lump sum benefit given to very poor families, as assessed by local governors); (5) Twins and other multiple births (lump sum benefits to all families with twins and other multiple births); (6) adoption (long term benefit to all families with adopted children). In 2004, child related allowances were the single most important type social assistance program, accounting for 43 percent o f beneficiaries and 51 percent o f the budget, being child care the most important program. In July 2006 these allowances were consolidated into a single CCT o f 20,000 MNT per month, reaching 24,287 beneficiaries in 2006, and accounting for 0.7 percent o f the total budget. (iii)Social welfare sewices and concessions are in-kind and indirect cash benefits for the poor (as assessed by local governors), disabled, poor elderly, war veterans. They include nursing home services, other social welfare services (e.g. meals, bathing, cleaning, assistance in fuel collection), and concessions (e.g. heating, disability gear, travel, house rent, treatment in a sanatorium, prosthetic device). In 2004 these benefits accounted for 46 o f beneficiaries but only 19 percent o f the social assistance budget. The July 2006 amendments to the Social Welfare Law introduced Community-Based Integrated Social Welfare Services, reaching 6,3 14 beneficiaries in 2006. Social welfare services and concessions accounted for 0.9 percent o f the total budget. 5.13 The social assistancesystem designwas complex (particularly social welfare services and concessions) and had many redundancies (e.g. benefits for families with children), resultingin a lack of internal coherence in terms of objectives and target populations. The system focuses on poor and vulnerable children and elderly, but poor and vulnerable youth and adults are largely not (directly) covered. Herders are not specifically targeted despite being a very vulnerable group. Also, many benefits are not poverty targeted (all child-related allowances except infant care are not poverty targeted), and among those targeted, the poverty status is subjectively assessed by local governors. Finally, none o f the benefits are conditional on behaviors that reduce welfare-dependency o f beneficiaries. 104 Chapter 5 -Social Protection 5.14 An analysis of the efficiency and effectiveness of the system based on the LSMS 2002-03 revealed a largeleakage to non-poor households,largeexclusionof poor households from social assistance, and a small impact on poverty reductionn2Some 57 percent o f total households received social assistance in one form or an~ther.~ As many as 70 percent of beneficiaries were non-poor households (inclusion error), while about 40 percent o f poor households-25 percent o f households in the poorest quintile-did not receive any assistance at all (exclusion The poverty rate without these social assistance benefits is only 10 percentage points higher than the poverty rate with social assistance benefits included. This small difference was due to the small average size o f benefits (15,000 MNT per household per month) relative to the amount necessary to cover basic food needs, and bad poverty targeting. Nonetheless, this amount was important for the extreme poor for whom social assistance accounted for 24 percent to their household consumption on average. The targeted CMP era-a good start with great potential 5.15 The basic design of the CMP brought about important innovations to the social assistance system: the conditionality o f benefits on positive social behavior and objective poverty targeting. The basic design provided an incentive for poor families to invest in the human capital o f their children-reducing structural long-term poverty-while covering their transitory needs through the monetary transfer. The CMP was also meant to replace or consolidate other benefits to poor families with children. But the program was introduced with little preparation and without any monitoring and evaluation strategy built in, which resulted in design and implementationproblems and insufficient informationto make timely program refinements. 5.16 The CMP targeting formula had a large impact on high school enrollment among the poor, particularlythe poorest of the poor.But because o fthe large leakage to the non-poor, for whom the program had no impact, this large benefit was achieved at a high cost in terms o f efficiency. Some o fthe implementation issues are described below: The Government put a lot o f effort into the proxy means test but its design could have benefited from further refinements and piloting. The benefztper child was calculated based on the budget allocated for the program and the estimated number o f beneficiaries, rather than calibrated to some relevant reference ( e g direct costs o f schooling, poverty line). Although the selection of CMP benefzciaries was supposed to be done by local social welfare workers on the basis o f a proxy means test, in practice this selection was preceded by the discretionary screening o f local governors. Social welfare officers received little training to do the selection. The proxy means testing and scoring was supposed to be done through a computer program in Excel, but many local welfare offices did not have computers. There was no beneficiary database (to track benefits received), nor a system to monitor the conditionality o f benefits (school attendance, immunizations and child labor) and record changes in family composition (newborns, deceased children, children that turn 18 years o f age). Program officers were supposed to keep records o f all applicants but in * World Bank (2006). Mongolia Poverty Assessment, World Bank, Washington DC. 3 Social assistance in the LSMS 2002-03 consists of child allowances, maternity allowances, disability pensions, survivor's pensions, special pensions, and transfers from local government, NGOs, and charitable organizations 4 Poverty is defined as consumption below 24,743 Tugriks per person per month, which is the basket of 2100-calorie food needs and non-food basic needs. 105 Mongolia: Public Expenditure and Financial Management Review practice only kept those of beneficiaries. The CMP data available to the government is limitedto aggregate information about the total number of beneficiaries and benefitsdisbursed per month at the Aimag level. 0 School social workers and social welfare officers were given a vague mandate to check compliance but had little incentiveto reportnon-compliance. Ex-ante CMPanalysis- good start with roomfor improvement 5.17 Moving to a proxy-means targeting system was an outstandingimprovement from the previous Mongolian social assistance system, but methodological issues existed. The main weakness of the PRR is that it is constructed based on bivariate, rather than multivariate, relationships between poverty and the proxy welfare indicators.5The indicators were selected independently from their weights (i.e. importance in explaining poverty), andthe weight for each indicator is computed independently from the others (Le. not accounting for the relative contribution of each indicatorto explainingpoverty). 5.18 A study of the targeting properties of the PRR and alternative methodologies finds that Figure 5.3: There is room for improving the Child there is room for irnproGing the CMP Money Program formula particularly to includethe formula, particularly among the extreme poor extreme poor. In particular, the analysis revealed that the straight 50 application of the PRR formula would 40 result in 38 percent of non poor 30 households receiving the CMP while 20 42 percent of the poor (and 27 percent 10 of the extreme poor) would be 1 0 excluded from it (Figure 5.3).6 An alternative method could be based on Poverty 1 Extremepoverty Poverty 1 E&eme poverty a multivariate model of per capita CMP Atemafive formula consumptionwith a set of household- level variables that best explain consumption, estimated for UB, aimag I centers, soum centers and the Source: Simulations based on the LSMS 2002-03 countryside separately. This alternative methodology reduces the overall inclusion and exclusion errors relative to the PRR method, and clearly outperforms the PRR method in terms of identifying the extreme poor by reducingthe exclusion of the extreme poor to 12 percent (Figure 5.3).' In addition, this method uses only half of the variables used by the PRRmethod, which would simplify the implementationofthe formula. 'The CMP proxy-means formula, the poverty risk ratio (PRR), was derived from bivariate relationships betweenpoverty and a series of indicators that are correlatedwith it. The PRR is ultimately a weighted sum of these proxy welfare indicators, where weights measure how overrepresentedthe poor are along these different dimensions of welfare.' A household was eligible for the CMP if the weighted sum of these householdcharacteristicsdivided by the number of variables inthe formula is larger than 1. 6It was not possible, however,to do aperfectreplication ofthe PRR usingthe LSMSdata becausesome of the variables inthe PRR are either not inthe survey or are included in a slightly modified version. 'Apossible explanation for the small effect on overall exclusion of the poor relative to the exclusion of the very poor is that many of the poor may transitory (Le. right below the poverty line in a given year, right above it inthe next) rather than chronic. 106 Chapter 5 -Social Protection 5.19 The ideal benefit rate is such that it isjust enough to lift beneficiaryhouseholdsout of poverty and to induce them to send or keep their childrenin school. As noted earlier, the benefit rate was calculated not on the basis of need but rather on the budget allocated to the program. The size of benefits, however, does affect the eventual impact o f the program on poverty and school enrollment (effectiveness) and the cost at which this impact is achieved (efficiency). It is often hard to find the optimal benefit rate, but it needs to be determined on the basis o f need. The size o f the benefit (3,000 MNT per month per child) compares well to the direct costs o f education but it is only one-eighth o f the poverty line. According to the LSMS 2003-03, the average cost o f education was around 3,000 MNT per child for all grades up to the last two grades in upper secondary, where costs increased to 4,000 MNT and 6,000 MNT respectively. Ex-post CMPassessment -large leakage to the non-poor and conditionalitynot monitoreds 5.20 Soon after its introduction in January 2005, the CMP quickly became the largest social assistance program, accounting for 67 percent o f beneficiaries and 40 percent of the budget in 2005. By April-June o f 2006 (right before the universalization o f the program), 68 percent o f families with children were benefiting from the program, covering 29 percent of all children aged 0-18 in the country. In fact, in 2006 the CMP represented the main source o f welfare payments to households together with the traditional main source o f welfare payments (pensions) (see Figure 5.9 and 5.10). It accounted for 44 percent o f welfare payments to an average household-62 percent among households receiving CMP benefit-and 8 percent o f their income. 5.2 1 Conditionalitywas not followed, and beneficiariescontinued receiving benefits after they had dropped out of school. The efficiency and effectiveness o fthe program depends in part on whether the conditionality was actually monitored and enforced. None o f the beneficiaries in the HIES 2006 sample was over 18 years o f age, and none o f the 7-18 year old beneficiaries was engaged in productive activities. But all children, except one, who have dropped out o f school continued receiving benefits in April-June 2006-6.7 percent o f beneficiaries were out of school by that point. This affects the impact o f the program (by reducing the value o f the conditionality) and efficiency (by diverting benefits to those who should not still be getting them). 5.22 The proxy means test was not followed, and most non-eligible households received benefits. Even if the CMP proxy means test were to accurately identify the poor and the poor only, if it is not followed in practice the efficiency and impact o f the program are reduced. A straight application o f the proxy means test yields a 70 percent o f inclusion error (i.e. households that were not eligible to receive benefits according to the proxy means test but did receive them) and a 14 percent exclusion error (households that were eligible to receive CMP benefits but did not receive them).g About 47 o f families in the richest 2 proxy means quintiles (i.e. the two 8 The World Bank also worked together with MSWL, NSO, UNDP and ADB to collect additional data on the CMP to be able to assess implementation, targeting and benefit incidence as well as the impact of the program on poverty and schooling. To this end, a module on the CMP was includedin the April-June 2006 round of the HIES. The analysis below is basedon these data and complements the work that UNICEF has just completed using the same data (UNICEF (2007), "Child Benefits and Poverty Reduction: Evidence from Mongolia's Child Money Program"). This section provides an assessment of the CMP in terms of intermediate outcomes, Le. outcomesneededfor the program to be cost-effective. 9The HIES 2006 has informationon all the indicatorsthat were used to compute the proxy means score. A straight applicationof the proxy means test consists on identifying beneficiaries as those with a pro means score greater than 1. 107 Mongolia: Public Expenditure and Financial Management Review groups with the lowest PRR scores) received benefits while Figure5.4: A large leakage of Child MoneyProgram 16 percent of families in the benefitsto non-eligible households poorest 2 quintiles were 100 excluded (Figure 5.4). 80 5.23 As outlined above, a 60 number of factors contributed to the poor application of the 40 targeting formula. Local 20 governors, households and 0 program officers influenced the outcome of the targeting Lowest 2 3 4 Highest formula-the indicators and score score weights used inthe proxy means % of beneficiaries in quintile I% of total CMP beneficiaries test were common knowledge. Targeting mistakes were also Source: HIES 2006 the iesullof the limited capacity and trainingof program officers. In some cases, for example, all of the households that had been prescreened by local governors were selected into the program automatically without going through the proxy means test. The lack of accountability of social welfare workers and the tools to support them (records of all applicants, a system to verify the correct application of program rules and a complaint mechanism) also contributed to the poor applicationofthe targetingformula. 5.24 The benefits excluded some of the poor and it included most of the non poor. The CMP formula had a lot of room for improvement but, most importantly, it was poorly applied. - _ _ This resulted in- large inclusion errors. Figure5.5: A large leakage of the ChildMoney Using the population-weightedaverage of Programbenefitsto non-poor households the Minimum SubsistenceLevel (MSL) in 2006 (25,620 MNT) as the poverty line, 100 19 percent of the poor were excluded from 80 the program and 61 percent of the non poor were included. Disaggregating by 60 quintile of household per capita 40 consumption (Figure 5.5), 19 percent of 20 the households inthe poorest two quintiles were excluded from the program, while 55 0 percent of those in the richest two Poorest 2 3 4 Richest quintiles received benefits. And the % beneficiaries I% of CMP beneficiaries poorest two quintiles only account for 50 percent oftotal beneficiaryhouseholds. So Source: HIES 2006 a lot of benefits went to householdthat (as shown in the next section) did not need them, resultingin large cost-inefficiencies. CMPCost-effectiveness: Large impact on enrollmentamongthepoor but at a high cost 5.25 This section looks at the effectiveness of CMP in terms of two goals: reduce current povertyand reduce future poverty by increasingschool enrollment. The analysis of the impact of the CMP on enrollment tries to answer the question: What would have been the proportion of 108 Chapter 5 -Social Protection CMP beneficiary enrolled in school enrollment had they not received CMP benefits? The analysis suggests the following results: lo There is only room for improving enrollment in high school, where CMP beneficiaries seem to have done better (Figure 5.6). Looking at unconditional differences in enrollment rates between CMP beneficiaries and non beneficiaries reveals that: (i)enrollment rates for groups (1) and (2) are close to universal, thus with little room for improvement; (ii)there is room for improving enrollment in high school (around 14 percent o f 16-18 year olds that completed lower secondary are not enrolled in school); and (iii)CMP beneficiaries are about 13 percentage points more likely to be in school than non-beneficiaries among 16-18 year olds who have completed lower secondary. This difference in high-school enrollment is due to the CMP, and the impact is concentratedamong the poor only. After factoring out the different characteristics o f CMP beneficiaries and non beneficiaries, the CMP i s found to be associated with a 16 percentage point increase in the probability o f being enrolled among 16-18 year olds who have completed lower secondary. The difference is somewhat larger than the difference in unconditional means because CMP beneficiaries have characteristics that make them both ` more likely to be selected as beneficiaries and less likely to be in school (Le. they are poorer, have a higher proxy means test score and have more children). And the analysis by quintile o f household per capita consumption shows a large effect among the poor (35 percentage points among the poorest quintile) but no significant effect among the non poor (the richest two quintiles) (Figure 5.7). Most of the impact of CMP on enrollment is likely to have come from the income effect of the cash transfer (rather than the inducement effect o f the conditionality) since the enrollment conditionality was not enforced. It thereby indicates the importance o f affordability in explaining high school enrollment in Mongolia." This result also shows that the rationing rule used to limit student access into upper secondary school may not have been binding.I2 Since entrance is determined by performance in the state examination, and performance o f students in rural school is significantly lower, students from rural schools are disproportionately `screened' out o f the system.13 But most o f the impact o f CMP on enrollment comes from the poor, particularly the poorest, who tend to be located in rural areas. It is important, however, to revisit the rationing policy to avoid bottlenecks. Finally, lo Data from non-beneficiary children are used to identify the difference in enrollment rates between beneficiary and non-beneficiary children that is due to the program, and the program only, controlling for differences in household characteristics (probit model estimated by maximum likelihood). Ideally, this model would be estimated on the pool of 7-18 year old children who were in school when the program was introduced (January 2005),1° looking at whether they are still in school by April-June of 2006. The HIES 2006 has information on whether a childever received CMP benefits (and when it starteddoing so), but the information on when children dropped out of school is only available for CMP beneficiaries that lost their benefit as aresult of that." Inthe absence ofthe necessary information we estimate the model for groups of `comparable' children in terms of the time exposure to the risk of dropping out of school: (1) Primary school-age children (8-1 I)," (2) Lower secondary school-agechildren (12-15) who had completed primary school at the time of the survey; and (3) Upper secondary school-age children (16-18) who had completed lower secondary school at the time ofthe survey. 1 1It is still possible, however, that households may have perceivedthe conditionality as binding even when it was not enforced. But this inducement effect would only be temporary, as households learn about the lack of enforcement ofthe conditionality. l2 Only around 70 percent of student completing lower secondary can go into upper secondary, where entrance is determined on the basis ofperformanceinthe state examination. l3 World Bank. 2006e. Mongolia Poverty Assessment, World Bank, Washington DC. 109 Mongolia: Public Exuenditure and Financial ManapementReview another reason that the CMP had a significant impact on enrollment is that the cash benefit compared well with the direct cost o f schooling inhighschool. 0 The CMP had little impact on poverty. The analysis of household per capita consumption with and without CMP income shows that the program appears to be associated with small changes in poverty-a 4 percent reduction in the overall poverty rate and a 1.6 percent reduction in extreme ~0verty.l~ This is not surprising given that the benefit was only one- eighth of the poverty line in2002-03. Figure5.6: There is only room for Figure5.7: Child MoneyProgramhasa improvementinhighschoolwhere Child largeimpacton highschoolenrollment MoneyProgrambeneficiaries seem to have amongthe poor but noeffect on the non-poor done better 1 (X ineachgroup enrolkdinschool) 1 (Percentagepoint differenceinenrollmentbyquintile) Primary lower secondary Highschool Poorest 2 3 4 Richest Source: HIES2006 5.26 Overall, the CMP had a small impact on poverty and a large impact on enrollment but at a very high cost. This is because of mistargeting: most non-poor households were receiving benefits, benefits they do not need to either stay out o f poverty (because they are not poor) or to keep'their children in schools (their children would have stayed in school regardless). For non-poor households the CMP benefit is basically a consumption subsidy, and there is no social or economic rationale for the Government to be subsidizing the general consumption of the non-poor. The universalCMP-a costly Figure5.8: Child MoneyProgramaccountsfor a largeshare of the budgetand GDP and unsustainablespendingof public resources 5.27 The amendments to the Social Welfare Law of July 2006 universalized the CMP and introduced two large untargeted " I---- allowances (newborns and newly % budget % GDP weds), confirming a clear trend in social assistance to grant generous, universal and unconditional benefits to the population. This trend has been Source: MOF; Figuresfor 2007 are based on projections for the facilitated by large revenues coming budget and GDP. l4UNICEF(2007). This is a preliminaryresult as this does not result from an impact analysis: it compares householdper capita consumption with and without CMP income. 110 Chapter 5 -Social Protection from mining. The subsidies for newly weds and newborns are one time payments (500,000 MNT and 100,000 MNT, respectively) to all young couples marrying for the first time and all households with a newborns, respectively. They account for 0.5 percent and 0.3 percent of the planned budget in2007, respectively. The CMP was universalizedto all families with children in July 2006, and the benefit was increased in January 2007 from 36,000 MNT per child a year (3,000 MNT per month) to 136,000 MNT. CMP is now in 2007 predicted to account for 3.5 percent o f GDP (Figure 5.8). 5.28 CMP income is predicted to account for 34 percent and 20 percent of the total household consumption expenditures among the poorest and the second poorest quintiles. The universalization and subsequent benefit increase o f the CMP are predicted to have made CMP the main source o f welfare payments to households, covering 89 percent o f households, accounting for 73 percent o f welfare payments to an average welfare-recipient household, and a sizable contribution to income among CMP beneficiaries (14 percent) (Figure 5.9 and 5.10). Figure5.9: Pensionshave beenthe main Figure5.10: as ChildMoneyProgramis source ofwelfare payments butincreasingly introducedand later becomesthe main less so (YO) source ofwelfare payments(YO) I " I I I I %households %soda1welfare %income % households %socialwelfare % income 2006 IPjterCNPchanges inJul.2006 I I Note: "Percentage of households" refirs to thepercentage of households withpensiodCMP benefits; "percentage of social werare '* refirs to the share OfpensiodCMP income in total social werfarepayments; and "percentage income" refers to the share ofpensiodCMP income in total household income among beneficiaries ofpensions/CMP. Sources: LSMS2002-03,HIES 2006. 5.29 The targeted CMP had a largeimpacton enrollment among the poor but problems in the design and, particularly, the implementation of the program reduced potential benefits and increased efficiency costs. In particular, the targeting formula could have been improved but, most importantly, it was largely not followed in practice because very little attention was paid to implementation arrangements (e.g. training and capacity of social welfare offices). As a result the CMP reach most non poor households even before it was universalized. The conditionality on positive social behavior was not (and is not) enforced, thus losing the potential positive effect o f the conditionality on, for example, school enrollment. CMP benefits were not enough to lift poor households out o f poverty. Because the program granted benefits to most non poor households for whom the benefits did not make a difference interms o f poverty or school enrollment, the large impact on school enrollment was attained at a very highcost interms o f efficiency. 5.30 The impact of the universalized CMP on enrollment or povertyshould be assessed with respectto whether the same impactcould have beenachievedwith a smaller budgetary cost. Clearly, whatever the impact of the universalized CMP, it is more costly than targeted conditional program. The universal benefit increase is likely to reduce poverty, but at the expense , 111 Mongolia: Public Expenditure and Financial Management Review of makingthe CMP even more cost inefficient, fiscally unsustainable, and politically risky. The comparison of simulated household consumption expenditures with universal CMP benefits at 136,000 MNT a year per child and expenditureswithout CMP incomeshow a 15 percentage point reduction in the overall poverty rate and a 5 percentage point reduction in extreme poverty.'* There is also the potential risk that the CMP may create some offsetting work disincentiveeffects given the large predicted share of CMP income in household consumption among the poor (30 percent among the poorest quintile). It is also unclear how much additional gain the benefit increase will bring on enrollment among the poor. More analysis needs to be conducted on the impact of the current programon poverty and enrollment.But it is clear that the cost-inefficiency from grantingbenefitsto nonpoor households is now far higherthanbeforeuniversalization. 5.3 1 The CMP universalizationhas beenjustified by an underlying fertility objective, but the CMP is not designed to attain this objective.For instance, the current CMP may very well eventually induce a small number of poor families to have more children at the cost of giving benefits to all the families whose fertility decisions will remain unchanged. Any social program needsto have a clear set of objectivesand specify the meansto achieve those objectivesin a cost- effective way. Any effort to increase fertility needsto be preceded by an ex-ante analysis of what type of intervention is most likely to have the largest impact on fertility at the lowest cost, and who is most likely to be affectedby it. 5.32 Last but not least, the large size of the program raises concerns on its fiscal sustainability given the high exposure of Mongolia to fluctuations in international prices of metals.While there is a political cost of adjustingthe targetingand benefitsof the program, there i s morerisk to increase socialspending at good times and be constrainedto reduce it at badtimes from an economic, social, and even political perspective. A possibleway forward 5.33 In order to maximize its social impact at the least cost, the CMP should go back to the basic principleson which it was originally designed: targeted to the poor and conditional on human capital investments, and thus aimed at reducingpoverty and increasinghumancapital. Targeting can be refinedalongthe linesof the multivariate proxy means test suggestedearlier or other alternativemethods (including geographic targeting or a combination of geographic and householdtargeting). In order to minimize the targeting errors of the multivariateproxy means test, the programneeds to either be focused on the very poor or use a bracket around the poverty line as a cutoff criterion(rather thanthe poverty line itself). The political feasibility of such a reform depends to great extent on the transparency and credibility of the targetingcriteriareachingthe poorest. There shouldnot be any prescreening of households by local governors. A gradual implementation of targeting reforms has also provento facilitate the process. Ex-ante analysis on the bene$t rate is needed that would have the largest impact on poverty and human capital at the lowestcost while beingfiscally sustainable. The right implementation arrangements need to be in place. Social welfare workers need to be adequately trained and have the resources (including computers) to be able to apply the formula and manage the programmore generally. They should be accountable for their work and should keep all the records concerning the selection of beneficiaries. A complaint mechanismshould be availableto households. 15 Calculation based on April-June2006 HIES data. See UNICEF (2007) for details. 112 Chapter 5 -Social Protection Inorder to maximize the value o f conditionality, there needs to be an ex-ante analysis of on the focus o f conditionality as well as a proper system for monitoring and enforcing these conditions. The decision on what to condition is really about identifying the human capital objectives on which the program can realistically have an impact given the cost o f monitoring the conditions. As noted earlier, because there is little room for improving enrollment at younger ages, it may make more sense to condition on school attendance. In order to monitor conditionality, beneficiary households can be required to present, at the time benefits are collected, written evidence from the child's teacher that he or she is enrolled (or from the doctor or nurse in the case o f immunizations). This evidence should be filed and available upon inspection o f the local social welfare office by the provincial and national authorities. And random inspections o f schools can also be conducted periodically. These visits to collect benefits can also be used to collect additional information from beneficiaries in order to update their eligibility status. The basic information on program beneficiaries should fed into a central database at M S W L to be able to (i)check whether the selection o f beneficiaries was done according to program rules; (ii)monitoring o f program outcomes and conditionality; and (iii)start a national database o f the poor and vulnerable and the social programs they benefit from to help identify gaps as well as redundancies inthe system. To ensure proper payments, an individual-level registry needs to be set up to track which beneficiaries collected the benefit and when they did it. Such a register constitutes not only an "invoice" o f the transaction but also allows to follow-up payment histories, and frequency. M S W L should also consider the use o f direct deposits, as with pensions, whereby beneficiaries could hold the money in an Agriculture Bank account and have access to other benefits and services (e.g. credit, interest payments on their deposits, etc). This system allowed the bank to keep a registry o f pensioners and to reduce double-payments and corruption. The poor are likely to be more isolated and have higher costs for registration and periodic collection o f the benefits. A direct system would then provide them with more flexibility (and a good saving instrument) while compensating them for some o f the transaction costs. International experience shows that another key pillar o f a good accountability system is a good impact evaluation o f program results. The purpose o f the impact evaluation (IE) i s to assess the effectiveness o f the project against its stated objectives (poverty reduction and increased human capital). IE analysis can answer a variety o f policy relevant questions, including: Didthe program work? Why did (or didn't) it work? Who benefited the most? Did the benefits from the programjustify its costs? The program design also needs to be flexible to allow the findings and recommendations o f IE to feed back into the program design and implementation.I6 5.34 CCT programs like the CMP are not panaceas and must operate as part of a well functioningand coherent system of social policies. The SSMP was a good strategic orientation but did not provide a clear implementationstrategy for consolidatingand rationalizing the system to achieve a coherent social assistance strategy. The very first step in weaving a safety net is to evaluate the main social risks faced by different groups and identify who is most at risk in each case (i.e. the poor and vulnerable). The second step is to identify the most cost effective policy instruments to address those risks. While the CMP may lift many households out o f poverty, it l6The CCT program inMexico (Oportunidudes) is a good illustration. The data collected on the CMP from the HIES and the subsequent analysis of the effectiveness and effectiveness of the targeted CMP by this study and others is a good example of what impact evaluation can do. This type of effort should be continued in the hture with the inclusion of a module on the CMP (and other relevant social programs) in the new roundsofthe integratedLSMS-HIES 113 Mongolia: Public Expenditure and Financial Management Review may not be sufficient to keep these households out o f poverty in a sustainable way because it does not adequately address the high rates o f joblessness and job informality in Mongolia4ther policies to enhance productive employment opportunities are needed. Also, as an incentive mechanism to invest in human capital, CCTs must rely on a good supply o f health and education services. THEMAKINGOFA MODERNSOCIALASSISTANCE SYSTEM INMONGOLIA: A ROADMAP 5.35 Ultimate goal: to design and implement a coherent and well-coordinated system o f social assistance programs that address, in a cost-effective way, the needs o f the poor and vulnerable proactively (i.e. preventing risks, not just mitigating and coping with them) with `corresponsibility' on the part o f beneficiaries (by conditioning the receipt o f benefits to positive social behavior), and inducing their re-activation (Le. avoid welfare-dependency and work disincentive effects). 5.36 Main driver of reform: An improved CMP that goes back to the basic principles o f targeting to the poor and conditionality on human capital investments can be the central piece o f this modern social assistance system. Box 5.1: The making of a modern social assistancesystem in Mongolia: A roadmap Ultimate goal: to design and implement a coherent and well-coordinated system of social assistance programs that address, in a cost-effective way, the needs of the poor and vulnerable proactively (Le. preventing risks, not just mitigating and coping with them) with `corresponsibility' on the part of beneficiaries (by conditioning the receipt of benefits to positive social behavior), and inducing their re- activation (Le. avoid welfare-dependency and work disincentive effects). Main driver of reform: An improved CMP that goes back to the basic principles of targeting to the poor and conditionality on human capital investments can be the central piece of this modern social assistance system. Recommended policies over the next year: laying out the ground work 1. Commission an evidence-basedstudy to assess of the main social risks faced by different groups and identify who is most at risk in each case 2. Refine targeting tool (not currently in use) based on a household-basedmultivariate proxy means test, geographic targeting or a combination ofthe two. 3. Work out a time plan for the gradual re-targeting of CMP to the very poor. This can be done by setting increasingly lower poverty lines (Le. cutoffs in the distribution of the proxy means measure that are used to identify the poor) over time so that an increasingly smaller percentage of families get the benefit untilonly very poor families get the benefit. 4. Organize consultations, with participation from the civil society, on the targeting tool and to gain public support for the gradual re-targeting of CMP to the poor based on the evidence presented in this report and others. 5. Evidence-based analysis of what to condition the receipt of CMP benefits on (e.g. school attendancerather than enrollment) and how to monitor compliance (e.g. evidence of compliance is presented at the time benefits are collected from the bank). 6. Laying out the ground work to strengthenimplementation of CMP: a. Training of local social welfare workers and resources (e.g. computers) b. Mechanismsto make social welfare workers accountable c. Complaint mechanismshouldbe availableto households d. An individual-level registry needs to be set up to track which beneficiaries collected the benefit and when they did it. The government could also consider using direct deposits, as with pensions. Box 5.1 continued... 114 Chapter 5 -Social Protection .../Box 5.1 Recommended policies over the medium .term: gradual implementation and learning 1. Start implementingthe gradualre-targetingof the CMP (e.g.from 100%o f beneficiaryfamilies to two thirds and so on) 2. Develop a CMP beneficiary database to monitoringthe program. This can also be the first step towards developing a database of the poor and vulnerable and the social programs they benefit fromto help identifygaps as well as redundanciesinthe system. 3. Assess implementationofthe CMP, its targeting and evaluate its impact on poverty and schooling by including a module on CMP in the regular HIES-LSMS.Use resultsto refine the design and implementationofthe programas needed. Recommended policies over the long run 1. Fullimplementationof the CMP targetedto the very poor 2. Database of beneficiaries of social assistance programs; public information on social programs (objectives,targeting, coverage, costs etc.) Source: WorldBank Staff PENSIONS Overview and characteristics of the pensionsystem 5.37 With the process of economic liberalizationinthe early 1990s, Mongolia inheriteda non- contributory public pension scheme with virtually universal coverage and both generous and unsustainable benefits. In 1994, pensions and other social insurance legislation was enacted which modified the existing legal provisions and broadly established the administrative framework. The 1994 legislation set out contribution requirements for social insurance which remaintoday, and also revisedthe qualifying conditionsand benefit parameters(see Table 5.2). A nationalhealth insurance schemewas added in 2001. 5.38 The Notional Defined Contribution (NDC) scheme was established in 1999 in an effort to reduce the fiscal burden and better align contributionsand benefits. The scheme appliedto all cohorts born after January 1, 1960.The Government envisionedthe introductionof the NDC scheme as an interim step designed to move the system to a partially-fundeddefined- contribution scheme in the context of largely undeveloped capital markets and limited governmentalfiscal resources. The policy documents of the period set as a goal that by 2005, 3 percent of the 19 percent total contribution rate would be actively managed and invested in securities and, over time, more funding would be permitted. Since 1999, disbursements to retirees under the earlier scheme have exceeded contributions so that, as of May 2007, no funds were actively managed, nor is it likely that movement from notional to partially-fundedaccounts will happenany time soon. 5.39 The recent decrease in the overall contribution rate for social insurance from 30 to 20 percent will increase formal labor competitiveness by lowering its cost, and also will lower the incentivesfor underreporting.The decrease inthe contributionrate however also has the effect of loweringthe revenues generatedwhich can be usedto pay current and future retirees. Moreover, decreases in the pensioninsurance contributionrate will have the effect of reducingthe notional account accumulationand thereforethe pensionbenefit.The previous 19 percent contribution rate for pensions insurance and 29-31 percent rate for all social insurance was a very heavy tax on formal labor. The combined contributionrate includes contributions to short-term, work injury, unemployment and health (See Table 5.2). Under the NDC scheme, 4 percent is to be used to fundthe costs of administeringthe system and operationalcosts. 115 Mongolia: Public Expenditure and Financial Management Review Table 5.2: Social insurancecontribution rates (inpercent o fcovered wages) Self-Employed EmDlover EmDlovee - Total (voluntary Pensions 7.0 7.0 14.0 9.5 Short-term benefits 1.o 1.0 2.0 1.o Work injury benefits 1 .O-3.0 1.0-3.O 1.o Unemployment Bene6ts 0.5 0.5 1.o Health Insurance 2.0 2.0 4.0 800 Tg/month Total 10.5-13.O 10.5 21.0-23.5 11.5+ 800Tg/month Source: Social Insurance Legislation, 2001 and 2007 amendment (November) 5.40 The entry replacement under the defined-benefit scheme depends heavily on nominalwage growth during the income averaging period although in most cases represents a greater benefit than under the NDC scheme. Old-age pension benefits for members o f the defined-benefit scheme are based on an accrual rate o f 2.25 percent per year for the first 20 years o f eligible service and 1.5 percent for each year o f service after that (see Annex 5A). Under the NDC Scheme, the benefit is determined based on 15 percent o f the 19 percent contribution rate credited to an individual account, the rate o f return applied to the notional balance (the three-year rolling average rate o f growth in average reported wages), and life expectancy at retirement using unisex retirement tables.I7 For an average full-term worker under the defined-benefit scheme, calculations in 2003 suggest that the entry replacement rate would be about 46 percent under the defined-benefit scheme while under the N D C scheme it would be about 28 percent. '*With over 70 percent o f retirees estimated to be receiving the minimumpension and with substantial growth in the minimum wage, the average old-age pension has been increasing. In 2003 the average old- age pension as a percent o f average gross wage was 39 percent." 5.41 Coverage has fallen substantially since 1995, with an estimated 368,000 active members who contributed to the pension's insurance fund at end-2005 or about 37 percent o f the economically active population in the same year having fallen from almost 55 percent in 1995 (see Figure 5.11). Herders, who in 2002 represented 43 percent o f the labor force, and the self employed, who represented 15 percent o f the labor force, are both exempted from making contributions." An estimated 29,000 herders and members o f the informal sector made voluntary contributions during 2006, about eight percent o f the total. Although some o f the decline in participation can be attributed to changes in the labor force composition, this limited voluntary participation along with suspected wage underreporting suggests that the social insurance and pension insurance schemes have weak credibility.'1 l7 At this writing, no workers born after January 1, 1960 have reached the regular or early retirement age and are unlikely until at least 2010, although it was technically possible for someone to reach early retirement age of 45 in2005. l8 This assumes nominal wage growth of 10 percent per year and 30 continuous years of service from age 30 to age 60 and a life expectancy of 16.3 years at retirement. See Patrick Wiese and Michael Cohen, An Evaluation of Pension Reform Optionsfor Mongolia, and Wiese, Patrick, An Evaluation of Pension Reform Optionsfor Mongolia, December25, 2006. "See Wiese and Cohen, op. cit. *' Ibid..IMF The reported that in 2003 the average monthly wage on which PIF contributions were paid was about MNT 65,000 while the average wages in budget institutions and public enterprises reported by the NSO were MNT 71,000 and 94,000, respectively. IMF, Optionsfor Expenditure Savings and Efficiency Improvements (mimeo) 116 Chapter 5 -Social Protection Table 5.3: Social insurancecontributors,beneficiaries,economically active population - - - - - - - Individuals 1995 2000 2001 2002 2003 2004 2005 Numberof insuredpeoplecoveredby pension, unempl. & ins.against employment injury and occupational diseases 409,100 381,400 392,700 308,100 357,600 366,000 367,900 Compulsoryinsured 395,400 363,900 370,200 285,800 332,000 336,800 338,900 Voluntary insured 13,700 17,500 22,500 22,300 25,600 29,200 29,000 Number of insuredpeoplecoveredby healthinsurance 1,992,400 2,030,200 2,019,700 1,963,200 1,984,700 1,990,000 1,959,400 Total of beneficiaries 284,277 254,394 259,607 262,948 268,638 277,179 283,65 1 Pensionbeneficiaries 222,816 179,668 181,382 182,137 183,968 186,618 190,939 Population,laborforce Population/in thousands/ 2,25 1,300 2,407,500 2,442,500 2,475,400 2,504,000 2,533,100 2,562,400 Economicactivepopulation/in thousands/ 812,700 847,600 872,600 901,700 959,800 986,100 1,001,200 Employed 767,600 809,000 832,300 870,800 926,500 950,500 968,300 Coverage 50.3% 45.0% 45.0% 34.2% 37.3% 37.1% 36.7% PensionSystemDependencyRate 54.5% 47.1% 46.2% 59.1% 51.4% 51.0% 51.9% Source: Statistical Yearbook of Social Insurance, 2006. 5.42 The ratio o f retirees and beneficiaries to contributors, or pension system dependency ratio has improved from 1995 to 2005, dropping from about 51 percent to about 37 percent (see Figure 5.1 1) To the extent that this continues, this suggests very modest demographic pressure on the system's finances. Figure5.11: Coverage and pensionsystemdependency ratios (percent) I \ / \ WS5 1998 1997 1891 $SOD ZWO 1WI 2002 ZW3 ZOO1 ZOp6 I-+Coverageof EwnomioalM Active Population +Pension System Depndency Ratio1 Source: SSIGO Statistical Yearbook, 2006. 117 Mongolia: Public Expenditure and Financial ManagementReview Figure 5.12: Pensioncontributors, beneficiariesand economically active population (Individuals) 3owow 25WQW 2 OW OW Source: SSIGOStatistical Yearbook, 2006. 5.43 Basic social assistance provided for the elderly through the Social Welfare Pension Benefit is very small with about 2,670 of the elderly poor assisted under this program, a number that has dropped each year since 1997 (See Table 5.4). This means-tested benefit is small, at 26,500 Tugriks per monthor about US$25per monthor about 44 percent of the Minimum Living Standard (See Table 5.5). Other forms of social assistance are also provided to the severely disabled not receiving other pension benefits; to those disabled since birth; incapacitated survivors; and other handicapped individuals including those that are blind, deaf, dumb and dwarves. Combined, all forms of social assistance have assisted about 45,000 individuals, with the greatest number of beneficiariesbeingthose disabledsince birth, who in 2005 made up about 22,000 individualsor about half of all the beneficiariesunderthe program. Table 5.4: Beneficiariesof social assistance (Individuals inthousands) -199819992ooO20012002200320042005 1997 Elderlyvery poor 6.40 6.20 6.23 4.72 4.27 3.73 3.29 3.05 2.67 Disabled(70% incapacitated) 3.30 3.40 4.64 5.42 5.89 7.05 7.69 8.80 9.44 Disabledh mbirth(%YOincapacitated) 10.30 11.20 12.70 14.28 16.42 18.11 19.61 21.66 21.99 Incapacitateddependentswhose breadwinnersdied 2.30 3.00 3.88 4.91 5.87 6.96 7.55 8.39 8.80 Fullyblirad,dumbor deafpeopleandd w e s 0.60 0.80 0.97 1.09 1.27 1.31 1.45 1.58 1.57 Incapacitatedas aresultofrescuing 0.03 0.01 0.02 0.04 0.01 0.01 0.00 0.01 Incapacitatedfamilymembersofthose diedinrescuing 0.01 0.02 0.02 0.02 0.01 0.00 0.00 Very p r single mothers(45+) andfathers (Sot) 0.13 0.14 0.13 0.15 0.10 0.13 0.16 0.15 0.21 Total 23.06 24.76 28.58 30.62 33.84 37.30 39.75 43.63 44.68 -199819992ooO20012002200320042005 Percent 1997 Elderlyvery poor 28% 25% 22% 15% 13% 10% 8% 7% 6% Disabled(7G% incapacitated) 14% 14% 16% 18% 17% 19% 19% 20% 21% Disabledh mbirth(50% incapacitated) 45% 45% 44% 47% 49% 49% 490/0 50% 49% Incapacitateddependentswhose breadwinnersdied 10% 12% 14% 16% 17% 19% 19% 19% 20% Fullyblind,dumbor deafpeopleandd w e s 3% 3% 3% 4% 4% 4% 4% 4% 4% Incapacitatedas a resultofrescuing 0% 0% 0% 0% 0% 0% 0% 0% Incapacitatedfamilymembersofthosediedinrescuing 0% 0% OYO 0% 0% 0% Very poor single mothers(45+) andfathers (5ot) 1% 1% 0% 0% 0% o?! G?! o?!0% 0% Total loo?? 100% 100% 100% 100YO 100% 100% 100% 100% Source: State Social Welfare Office Annual Report (2006). 118 ChaDter 5 -Social Protection Table 5.5: Benefits providedby social assistance (Millions ofTugriks) 1999 2000 2001 2002 2003 2004 2005 2006 2006111- 2007 al Social welfare pension 10,000 10,000 10,000 12,000 14,400 14,400 14,400 20,480 26,500 Social assistance allowances Allowances for pregnancy and delivery 48,000 72,000 99,000 99,000 99,000 99,000 99,000 99,000 240,000 Child care 9,000 9,000 9,000 10,350 10,350 12,420 12,420 12,420 - Child support for large families, per child 2,000 6,000 36,000 36,000 36,000 Infant care 12,000 18,000 24,750 24,750 24,750 24,750 24,750 24,750 Allowance for twins, per child 20,000 20,000 20,000 20,000 -- Allowance for adoption 96,000 144,000 198,000 198,000 198,000 198,000 198,000 198,000 198,000 Allowances for veterans & awardees Allowance for war veterans and spouses 24,000 36,000 49,500 49,500 49,500 99,000 99,000 99,000 200,000 Allowance for partisans and award 12,000 18,000 24,750 24,750 24,750 24,750 24,750 24,750 200,000 Allowance for national award 48,000 72,000 99,000 99,000 99,000 99,000 99,000 99,000 200,000 Minimum standards (Tugutshonth) Minimumwage level 12,000 18,000 24,750 30,000 30,000 40,000 53,600 53,600 69,000 MinimumLiving Standard 17,600 19,600 23,600 25,300 25,300 26,500 30,000 42,800 60,100 al budgeted. Sources: "Main Indicators of Social AssistanceProgrammesfrom Social We(fare Fund"; National Statistical Oflce, Statistical Yearbook2006. 5.44 Historical household survey data suggests that mandatory pension benefits provide an important source of income for rural and urban households alike (see Figure 5.13). Approximately 7-12 percent of Urban household income, for example, has been reported as received from pensions each year since 1990 (see Figure 5.13). Pension income has contributed somewhat less to rural incomes as such individuals have not been able to benefit from such formal sector programs. Such pension income still formed 6-11 percent of household incomes each year since 1990. Figure5.13: Household incomecomposition (inpercentoftotal householdincome) Urban Rural I I+ hsgaami- - - - D O P * O M ~ W - M w m h r m ~ * n g l o m t acw I- nspsasmmm- ~ m m a m ~ o ~ ammarrmm.mnmacwRa- ~ - - omu 119 Mongolia: Public Expenditureand Financial Management Review Insurance General Office (SSIGO). The current management structure for Social Security in Mongolia is that the Ministry of Social Welfare and Labour is responsible for setting out policy parameters for social security and the Social Insurance Office (SIO) is responsible for carrying out that policy. Within the S I 0 are the State Social Insurance General Office (SSIGO) and local offices and units. The PIF is responsible for administering the defined-benefit and notional defined-contributionschemes described above as well as the non-contributorypension insurance scheme for uniformedpersonnel. In addition, the Social Welfare Fund(SWF) providesnon-work related sickness, maternity, and funeral insurance. The Work Injury Fund (WIF) provides sick leave, disability, and survivor insurance for work-related injuries. The Unemployment Insurance Fund (UIF) provides unemployment insurance. Finally, the Employment Promotion Fund (EPF) established in2001 providesservices to the rest ofthe unemployed. Revenuesand expenditures of social insurance 5.46 Social insurance expenditureshave significantly expanded since the beginningof the decade. Pensionsand benefitshave increasedalmost fivefold since 2002 until 2008, or by almost one percentage point of GDP, and are projectedto reach 6.7 percent in 2008. Old age pensions represent the largest single category of social insurance expenditure, anticipated to be about 48 percent of all social insurance expenditures havingrisen from 3.5 percent of GDP in 2001 to 4.3 percent in 2008 (see Table 5.1 ). Taken together, old age, disability, survivors and military pensions represented about 65 percent of all social insurance expenditures. Disability pensions were anticipatedto be about 0.8 percent of GDP during 2008 or about 9.5 percent of all social insurance expenditures. 5.47 As a revenue source, security contributionshave declined from about 4.8 percentof GDP in 2000 to 3.6 percent of GDP in 2007 while at the same time income tax revenues more than doubled from 5.8 percent of GDP to 12.1percentof GDP (see 5.14). Socialsecurity contributions therefore have gradually lessenedinimportance as a revenue source. v Figure5.14: Compositionof total government revenues 35%1 30% 20%-I 10% // 0% 4 2001 Zoo2 2W3 2W4 2W5 2w6 2W7 P.m.111of GOP 1 --clnwrne Tax Revenue+ Social Security Contributions-(I-OtherTax Revenues+ TotalI Source: Ministry of Finance: Central Government Expenditures: 2001-2007. 120 Chauter 5 -Social Protection 5.48 Examining contributions and benefits together, the overall difference or deficit between contributions and benefits has grown from less than one percent of GDP in 2001 to almost three percent of GDP budgeted for 2007 (see Figure 5.15). As suggested, since about 73 percent of social insurance benefits are for Figure 5.15: Social insurancecontributionsand benefits (% long-term benefits such as of GDP) pensions, disability and I $y, _. ... .._... . ...-. . - .... . . .- ..-.. .....- survivorship, this means - .. ... that the Government subsidy of the social insurance system has hadto increase substantially since 2001 primarily to support the shortfall between pay- as-you-go contributions and long-term benefits. Since a majority of pension beneficiaries receive the minimum pension, pension expenditures are closely linked to changes in the minimum wage, which increased 20 percent, 42 I t Social Security ConbibubnsReceived+-Soaal lnsuramBenefitsPaid-A- Net] percent and 26 percent in Source: Ministry of Finance 2004, 2005 and 2006 respectively (see Table 5.6). Although average wages also increased substantially during this period, the minimum wage grew from 30.7 percent of average wages in 2003 to 42 percent in 2006. The projected long-termdeficit of the PensionInsuranceFund is about 3 percent of GDP. Table 5.6: Average wage, minimumwage and minimum livingstandard (Tugriksper month,percent) 2 0 0 3 -2004 - - 2 0 0 5 2006 Average Wages and Salaries 81,500 93,100 101,200 127,700 Legal M i n i m u m W a g e 25,000 30,000 42,500 53,600 M i n i m u m L i v i n g Standard (Ulaanbataar) 25,300 26,500 30,000 42,800 Average Insurable Earnings o f Contributor n.a. n.a. n.a. n.a. M a x i m u m Insurable Earnings 300,000 400,000 425,000 530,000 Percent o f Average Wages a n d Salaries Legal M i n i m u m W a g e 30.7% 32.2% 42.0% 42.0% M i n i m u m L i v i n g Standard (Ulaanbataar) 3 1 .O% 28.5% 29.6% 33.5% M a x i m u m Insurable Earnings 368% 430% 420% 415% A n n u a l C h a n g e Average Wages and Salaries 14.3% 1 4 . 2 % 8.7% 2 6 . 2 % Legal M i n i m u m Wage 20.0% 41.7% 26.1% M i n i m u m L i v i n g Standard (Ulaanbataar) 4 . 7 % 13.2% 42.7% M a x i m u m Insurable Earnings 33.3% 6 . 3 % 2 4 . 7 % Average A n n u a l Inflation 5 . 1 % 8.2% 12.7% 5.1% Source: National Statistical Office. Statistical Yearbook 2006 121 Mongolia: Public Expenditure and Financial Management Review K e y pension policy issues 5.49 The 2 main pensions policy issues are that first, the sustainability of the Pension Insurance Fund is not ensured over the long run, despite the establishment of the NDC scheme which reduced long-run deficits of over 9 percent of GDP to projected deficits of 3 percent of GDP.22Second, the scheme is left with substantial equity and predictability concerns which are also a cause for financial un-sustainability.There are problems of equity of treatment between different workers, of predictabilityin the benefit parameters sufficient to ensure public credibility and compliance, and of financial un-affordability to the Government, employers and employees alike.23Inwhat follows, the mainparametric characteristics of the pension system are reviewedas well as the current social pension and voluntary vehicle system in order to highlight existing issues.Recommendations for reformsare thereafter suggested. . Pension Parameters Review The minimum pension framework is substantially too generous, creates very weak incentives both for the defined-benefit and NDC schemes and is the largest contributor to the fiscal burden of Pension In~urance.2~ estimated 70 percent of An the recipientsof old-age pensionsunder the defined-benefitscheme receive the minimum pensionequal to 75 percent of the minimum wage. For workers in the NDC Scheme, the minimum pension guarantee is sufficiently high relativeto the distributionof wages that many of the retirees under the scheme will be entitled to the minimum creating poor incentives as well as a mismatch between cumulative contributions and benefits. In addition to creating substantial incentivesfor underreportingof wages for a majority of wage-earners, the level of the minimum pension relative to the anticipated regular pension benefit also increases pension expenditures above revenues thus compromising the system's financial ~ustainability.~~In addition, the minimum pension benefit formula is a function of the average reported wages and not minimum subsistence. Finally, the financing formula provided for the NDC scheme (a portion of 4 percent of covered wages) is entirely insufficient to finance the anticipatedcosts of the difference between . the old-age pension generated by the Notional account accumulationand the minimum pensionguarantee inthe future. The retirement ages for regular retirees (aged 55 for women and 60 for men) and those in special occupations (as low as age 45) are too low to enable a financial balance in the defined benefit scheme and a meaningful replacement rate at retirement in the NDC scheme. Further, these low retirement ages will create a financial burden to the Government for some time to come as a result o f deficits in the . defined-benefit scheme and as a result of the minimum pension benefit for early retirees in the NDC scheme. Although current legislationsuggests that benefitadjustmentsshould be in linewith inflation, the actual indexation has in many cases resulted in a depreciation of the 22These figures are citedinIMF, Optionsfor ExpenditureSavings and Eflciency Improvements. 23 Itshouldbe notedthat these policy issues havebeenraisedby earlier observers of the Mongolianpension system and in Government strategy documents. See Strategy (2001), Fox (2002), Thompson (2003), IMF (2005),andWiese andCohen (2003)and Wiese (2006). 24 See Wiese andCohen, op. cit.. 25 Underthe assumption of a normaldistribution of wages, Wiese, ibid, projects long-runexpendituresas a percent of coveredwages as 18.5 percent verses the 15.0 percent of coveredwages allocated to the NDC benefit. 122 Chapter 5 -Social Protection real benefit value?6 This creates uncertainty for pensioners and is fundamentally at odds with the objectiveof providingold-age income security. Figure5.16: Life expectancy and normalretirementage 95, 65 60 55 5 0 . 1 , I I , , , I I I I , , , I I I I I I , I , I I I I ~ I I , I I , , , , , , I , I 50 51 52 53 54 55 58 57 58 59 BO 61 82 83 84 85 88 67 88 89 70 71 72 73 74 75 78 77 78 79 80 81 82 83 84 85 88 87 88 88 80 81 A w I+- Male Life Expectancy +Female Life Expectancy +- Male Normal Retirement Age --a Female Normal Retirement I Source: Patrick Wiese,Mortaliw andAverage Lifetime Data, in Wtese, 2003 The current benefit formula for the defined-benefit scheme is too short a period to be equitable across different income groups.*' It uses the best 5 year average covered wage base as the basis for determining a retirement benefit. In those cases such as workers retiring in the early 1990s during and after high inflation levels, the 5 year average for determining the reference wage for benefits substantially reduced the effective replacementrate at retirement.Further, it created substantially different benefit levels for different workers subject to different inflation levels during their income . averaging period. Many individuals who were not near the minimum pension at retirement, soon found themselves subject to the minimumyears later. The invalidity benefit designis sufficientlygenerous to both pose a fiscal burdenand weaken incentives under the defined-benefit scheme and, in the case of the NDC scheme, is inappropriatelyfinanced. Those workers with 20 years of service who have been determined by medical examiners to have lost 50 percent of capacity receive the same accrual rate for service as they would have received for an old-age pension. Even more important, since the minimum invalidity pension is equal to the minimum retirement full pension (75 percent of the minimum wage); this suggests that all those meetingthe qualifying criteriawill receive the minimumpension for life. Inthe case of the NDC scheme, the worker losing more than 50 percent of capacity is entitled to an annuitizedbenefit for life equal to an average of the last three years' wages, regardless of age. There do not appearto be any actuarialprojectionsindicatingwhat proportionof the 4 percent of covered wages earmarked for this and other benefits would be earmarked towards providingsuch a benefitor indeedifsuch an earmarkwould be sufficient. 26For example, the IMF reportedthat pensionswere increasedby between 15 and20 percentinOctober 2002, and againby 15-42percent in January 2004, although inflation in 2004 was only 4.7 percent. See IMF, op. cit. 27Many schemes also adjust or "valorize" according to real coveredwage growth insteadof inflation. In the case of Mongolia, this would be too costly to be affordable. 123 Mongolia: Public Expenditure and Financial ManagementReview . The defined-benefit and NDC schemes offer very different survivorship benefits. Benefitsunder the defined-benefitscheme are providedbasedon the number of survivors but could be improvedby providing 100 percent of the accrued retirement benefit (with . actuarialreductionsprior to retirementage) divided between eligible survivors.The NDC scheme links benefitsto the workers' wages and not benefit accruals. The lack of a transition mechanism in the 1999 law establishing the NDC scheme will result in a substantial reduction in the replacement rate between those born in 1959 and those born in 1960. Such a dramatic adjustment in the benefits of a mandatory public pension scheme between those born in 1959 and those born in 1960 clearly will create substantial pressure for a benefit top-up for those younger cohorts. Several excellentpolicy options have been suggested by earlier analysts.28 Social Pensionsand Voluntary Savings Vehicles 5.50 Although the adoption of an NDC scheme in 1999 improved the inter-generational sustainability, equity and predictability of the Pension Insurance Fund in Mongolia, the design did not fundamentally accommodate the special needs of herders, the informal sector, and those formal sector workers with insufficient lifetime contributionsto support a pension or otherwise fall into poverty during retirement. Since the early 1990s, coverage of Mongolia's herder populationhas fallen dramaticallyand the self-employedand informal sectors have grown in size. There is considerable concern among the authorities over the limited coverage of the formal contributory schemes over these populations and the limited options for improving such coverage. 5.5 1 International experience suggests that merely focusing on improved compliance is insufficientto improve coverage and, instead, that the relative incentives for participation need to be changed. More generally, minimum subsistence by the elderly is almost entirely linked to 15 or 20 years of contributory work experience. Given the nature of the Mongolian economy, it is unlikely that the majority of workers can or will meet such a requirement for some years to come. As currently designed and implemented, the Social Welfare Pension BeneJt only provides benefitsfor the very poor and is therefore inadequate as a basic support mechanism for the elderly inpoverty. Optionsfor Reform 5.52 The review ofthe design and implementationof Social Insurance inMongolia with a core focus on the PensionInsurance Fundsuggeststhat: 5.53 First, the societal objectives of the combined public and private pension and social security system could be further clarified for these respective schemes. Assuming that the objective is to provide: (i)a minimum living subsistenceto the elderly, permanently disabled and survivor/dependents; (ii)mandatory consumption smoothing for most formal sector workers; and (iii)voluntary contractual savings for informal sector workers, the following reforms are proposed: 28Inparticular,see LaurenceThompson,Analysis of Social Security Options, 2003, Wiese andCohen, op. cit. and IMF op. cit. 124 Chavter 5 -Social Protection . A non-contributory universalsocial pension can be designed in such a way as to encourage broader participation in the contributory pension scheme, extend basic assistance to a wider scope of the elderly poor, and serve as an alternative to the current minimum pension g~arantee.~' The objective of a social pension would be to provide a minimum living subsistence (MLS)to the elderly, disabled and survivors. Cost estimates depend upon the size and design of the social pension and the efficiency of its deli~ery.~'Of course, the cost of a social pension of, for example, 1.5 percent of GDP would be more than offset over time by the cost savings realized by replacing the minimum pension with a social pension. It is also possible to realize additional benefits through the social pensions. For example, inter-generational resource transfers within . families, such as support of grandchildren's health and education, could be partially offsetby the support received by the social pension. Changes to the voluntary contributory pension scheme for herders and the self- employed could modestly increase savings mobilization while at the same time utilize the existing NDC framework. In particular, an enactment of provisions for the early withdrawal of partial savings accumulations in the case of vulnerabilities such as drought, loss of income, health, disability or death could make the voluntary savings scheme more attractive. In addition, those contributors to the NDC mandatory notional accounts could also be enticed to contributeadditional savings, also with the incentive of early withdrawal under specified conditions. The notional interest rate (average 3-year growth in covered wages) on these savings would be attractive provided individuals would be confident that they could have access to their funds quickly and efficiently. Additional efforts would also be needed to ensure prompt, accurate and accountable disclosure of account informationand efficient processingof withdrawal requests. 5.54 Second, in order to strengthen fiscal sustainability, affordability, predictabilityand equity of benefits, the existing design characteristics of the current system could be retained while revising some of the parameter provisions. A universalmeans-testedsocial pension could also be considered which would provide for minimal subsistence for the elderly poor and could replace current minimum pension guarantees. Some specific suggestions to revise some of the pension parametersfollow: . Three types of increases in the retirement age could be considered to restore financial balance: (i) phasingout early retirement for workers in hazardous occupations, working underground or inhighheat, as well as specialtreatment for women who have raised4 or more children; (ii)gradually unifying retirement age for women with that of men at 60 . which could be accomplished; (iii) increasing in the medium run the overall retirement age in accordancewith a careful assessmentof life expectancy at retirement. Ifthe objectiveofthe minimumpension is to keepthosewho havemetthe vesting period out of poverty during retirement, then the linkage to the minimum wage (as opposed to the Minimum Living Standard (MLS)) has mismatched the instrument from the core objective. The minimum pension could, inthe short-run, be delinkedfrom the minimum wage, linkedto the MLS and graduallyreduced, and inthe medium-term,a means-tested universal social pension could replace it (see below). Using the MLS as the anchor for a 29 It should be noted that the social pension concept presented here is very similar to the demogrant suggestedby Wiese and Cohen(2003) andWiese (2006). 30 Wiese and Cohen estimatedthe cost o f a demogrant equal to 12.5 percent o f the average gross wage to rise from 0.02 percent o f GDP in the first year to 1.55 percent of GDP in 50 years, assumingthat coverage rates remain low. If coverage rates in the contributory scheme increase thereby pre-fundingmore of the benefit,thenthe cost couldbe reduced. See Wiese andCohen, op.cit.. 125 Mongolia: Public Expenditure and Financial Management Review minimum pension guarantee requires that the MLS be objectively determined and routinely and rigorously revised in accordance with surveys of the costs of a minimum consumptions basket. The dramatic increases in the MLS from 2003 to 2007 suggested in Table 5.7 are pausefor concern. Table 5.7: Minimum livingstandard (Tugrik per month) ~~~~~ West 19,500 20,200 23,200 37,000 54,700 East 20,500 21,200 25,000 34,800 51,800 Central 19,800 20,600 22,300 39,000 56,700 Khangai 19,900 20,600 22,600 38,300 54,600 Ulaanbaatar 25,300 26,500 30,000 42,800 60,100 Cumulative Annual Change Change West 3.6% 14.9% 59.5% 47.8% 180.5% East 3.4% 17.9% 39.2% 48 9% 152 7% Central 4.0% 8.3% 74.9% 45.4% 186.4% Khangai 3.5% 9.7% 69.5% 42.6% 174.4% Ulaanbaatai 4.7% 13.2% 42.7% 40.4% 137.5% Average Annual Inflation 5.1% 8.2% 12.7% 5 . 1 % . Source: National Statistical Organization, Statistical Yearbook 2006. The depreciationof real benefitvalue can be remedied through the indexation of all new social insurance benefits to a weighted average increase in the consumer price index. Retroactive adjustments are also needed, if possible to increase the benefit levels for those who have received ad-hoc increases which havenot maintainedthe realvalue ofthe benefit. w Increasing the income averaging period and indexing the income would improve the equity and predictabilityof retirementbenefits. w Looking at invalidity as an insurable risk that is appropriate for a pooling mechanism through the Pension Insurance Fund; two options can be considered for reform: (i) a lump-sum benefit which can be supported by an earmarked contribution. (ii) An annuitized benefit based on the accrued years of service and with an actuarial reduction factor imposed in order to reflectthe longer benefitperiod.Financingeither benefit under the NDC scheme should happen through an earmarked contribution allocated to a disability reserve fund that would manage the relevant contribution, investment and disbursement flows. The lump-sum and annuitized benefits would need to be revised basedon an annual actuarialvaluation. Survivors ' benefits could be provided based on the principle that the accrued retirement benefit is basedon the notionalaccount accumulationdividedbetween eligible survivors. This option, however offers no co-insurance. In addition, a lump-sum benefit could be . calculated based on a desired contribution (say 1 percent of covered wages) and the lump-sumbenefitprovidedto the survivors basedon the current premium. Inthe transition to the NDC system, a retroactivereductionin the accrual rate for certain cohorts and a substantial reductioninthe accrual rate for all cohorts inthe defined-benefit scheme are suggested. The latter change will have a limited impact on the entry replacement rate. As suggested by Thompson (2003), certain cohorts could also have 126 Chapter 5 -Social Protection benefits determined by a combination of accruals under the old scheme and under the new ~cherne.~' 3'Suppose that the marginal accrual rate of 1.5 percentlyear was cut to 0.75 percent beginning in 2009. A 60 year old worker born in 1959 with 30 years of contributions would retire in 2019 with an entry replacement rate of 52.5 percent, assumingthat the last 10 years accrued at the lower rate and no nominal wage growth. The same worker with 30 years of contributionsborn in 1960 is projectedto have only a replacementrate of only about 28 percent. Similarly, if workers born before 1960were to receive account accumulations under the NDC scheme beginningin 2008, then the impact on their replacement rate would also be modest and the transition betweencohorts still abrupt. 127 CHAPTER6. AGRICULTURE PROGRESSAND CHALLENGES rn The agricultural (livestock) sector is a traditional core pillar o f the Mongolian economy. It is the largest employer in the economy, despite a declining share in GDP. It i s prone to exogenous shocks which can have a dramatic impact on the economy. rn The government stated policy objectives in the agricultural sector are in line with a market-oriented economy, but inpractice, public expenditures tend to favor direct interventions to influence the market, rather than focusing on correcting market failures and providing public goods. rn Public spending on agriculture is relatively low, though has increased rapidly in recent years and the proportion o f capital expenditure is growing. Donor funding contributes for a significant share and even though its share is falling, greater coordination i s required to align this with government spending andpriorities. Central-local accountability is weakened by informational asymmetry, lack o f incentives at the local level to implement centrally planned programs in a cost-efficient way, and low monitoring and majority o f agricultural expenditure is at the soum level vel is low, and in some soums there are no agricultural officers. There rn capital expenditure and not supported the capacity policy analysis. The result has been that the now significant outlays o f capital investment (especially into wells for livestock and for incentives for crop irrigation) have not been based on sound policy analysis. KEYFUTURE PRIORITIES 1. Strengthen coordination (inter-agency and government-donor) to align investment with policy. 9 Suggested"how to": The MOFALIcould formalize the policy analysis andaidcoordination division and provide an operational budget in 2009'. This division would have a mandate that includes inter- agency anddonor coordination. ii. Improvinglocalinvolvementinbudgeting P Suggested "how to": In the short term, the MOFALI could set out a ne * ` for the delivery o f public services for ulture sector, with an investment planof how rt local capacity. In the medium term, icultural officers need to be recognized as tical agents in the identification and implementation of agricultural programs, in conjunction with local communities o f herders andfarmers. iii.Improve policy analysis capacity central and local level by training to allow for better monitoring and evaluation o fproject I Update: as o f November 2009, a new Economic Policy and Aid Coordination division has been proposed and is under consideration. 129 Mongolia: Public Expenditure and Financial ManagementReview 9 Suggested "how to": MOFALI could increase staffig levels (currently only three staff are in the Department for Information, Monitoring and Evaluation) and invest in training programs to increase technical capacity inpolicy and expenditure analysis. iv. Improvethe finctioning o fmarkets byprovidingpublic goods and correcting market failures. > Suggested "how to": spending programs that help limitingrisks and uncertainty for herders should be considered such as improving access to information, providing training, education, and extension services. 130 Chapter 6 Agriculture - BACKGROUND, ACHIEVEMENT AND CHALLENGES 6.1 The agriculture sector, and in particular the livestock sector, has traditionally been a core pillar of the Mongolian economy, and although its share of the overall economy is dropping, it will continue to be a major sector in the future. The livestock sector (contributing to around 90 percent of agriculture in Mongolia) provides rural households with an important source of income, jobs and food security, and is a means for investingand storing wealth. The crop sector, important in pre-transition days when Mongolia was self sufficient in wheat and vegetables, has declined dramatically since the early 1990s when state farms were disbanded. Over the past several years, the agriculture sector has witnessed a sharp decline in its contribution to the overall economy. From 1993 to present, the share of agriculture in GDP dropped from 40 percent to 20 percent. However, despite its decline in contribution to GDP, the agricultural sector continues to be the largest employer in the economy. Around half of Mongolia's 2.5 million people are rural, and the mainstay of the rural (non-mining) economy remains the agriculture sector. 6.2 The agricultural sector is highly prone to exogenous shocks, which can have a dramatic effect on the sector and the wider economy. The agriculture sector i s subject to considerable environmental risk. The crop sector today is largely rain-fed, and with short growing seasons and highfrequency of droughts, production is prone to failure without irrigation. The livestock sector also has significant production risks, due to the dependence on open pastureland for the nutritional intake of the majority of animals. The country is prone to extreme climatic events that can cause high rates of livestock mortality, jeopardizing rural livelihoods. In particular, the frequent droughts and severe winterdsprings (known as dzud)can devastate herd numbers. Even though the succession of dzud between 1999 and 2002 were exceptional (with over 10 million livestock lost), these were by no means isolated Figure 6.1: Agriculture GDP growth events. The impact on agricultural GDP and on Agriculture GDP growth 1996-2005 (%) the overall economy is significant (see Figure 6.1)' 20 0 as shown on the table 15 0 below: GDP growth was 10 0 negative between 2000 to 5 0 2002 as a result o f the 0 0 livestock sector. The cause -5 0 of these disasters is partly -10 0 exogenous, as a result o f -15 0 extreme weather, but also -20 0 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 man-made. Poor management of livestock Source: Mongolia Statistical Yearbook by inexperienced herders was a contributing factor to the high losses in the 1999/2000 and 2000/2001 dzud, and overstocking of livestock, leading to pasture degradation and consequently the poor nutritional quality of livestock, also increasesmortality. 6.3 Given the size of the agricultural sector, the fact that a large share of population, and especially the poorest, depends on the sector and the sector's vulnerability to shocks, special attention from policymakers is required. The Government has an important role to play in a market economy: it has to produce public goods and it has to reduce market failures (see Annex 7A for definition of terms in the context of the agricultural sector in Mongolia). The response of the sector to the market reforms depends on rural infrastructure, technology (research and 131 Mongolia: Public Expenditure and Financial Management Review extension) and markets (including rural finance). Where these are absent, public investment will be necessary. Market failure is of special concern in agricultural markets in Mongolia.Transition in any country can only be successful if agriculture adjusts successfully to the new conditions. The landmarket, the market for agriculturalfinance andthe know-how of people employed inthe sector are the most important determinants of successful agriculturaltransition. For agricultureto improve efficiency policies should focus on mitigating the inherent and state-created market failure. Therefore, public expenditure in the agricultural sector can only be effective and efficient ifit is focusedonproductionofpublic goods andcontributesto mitigatingmarket failure. 6.4 for the agricultural sector are in linewith a market oriented policy ...Governmentobjectivesof The overallobjectives food and agriculturalpolicy of Mongoliaandobjectivesfor the livestocksector are stated inAnnex of Resolution#29 of 2003 of the State IhHural of Mongolia. The main two main directions mentioned in the report and which will be used for assessing the adequatenessof measuresare stated as follows: i) The activities for implementationshould be basedon principles of the rationaluse of natural, economic, financial and human resources, intensification o f production and enhancement of efficiency, product quality and competitiveness in providing safe products and raw materials and ensuring sustainable development of food, livestock and crop productionsectors. ii) The main model of the agriculture development of Mongolia should be efficient, reliable, pasturehtensified livestock and crop production based on all forms of ownership and self-sustaining farms adaptable to changes in the climate and environment. 6.5 ...but in practicepoliciesand expenditurestend to favor directinterventionsinstead of focusing on the productionof public goods or mitigationof market failure. In contrast to the stated policy, the Government often intervenes in the market by supporting individual producers (see below for an analysis of the efficiency of public investments). There is ample international experience which proves that private investment supported by subsidized loans or discounted input prices generally lead to low or even negative profitability. (i)Direct support suppresses the development of markets and impedes the removal of the main obstacles to productiongrowth, which is badly functioningmarkets and an inadequatesupply of public goods. (ii)Support for one group also implies indirecttaxationofanothergroup. Forexample, ifdairy farmers are supported directly by subsidized loans or discounted equipment, taxpayers will have eventuallyto pay for it. It is highly questionable whether the impliedtransfer effect is in line with the overall objective of reducingpoverty in the country.(iii)Favoringa specific group of people requires objectivecriteria for selection. It can be assumedthat the number of peoplewho demand support will be higher than the funds available for support. It is often the case; however, that adequate information is not available for accurate decision-making. Those who demand support are inclined to provide information to their favor and those who have to select do not have enough informationaboutthe profitability of individual activitieswhich might be supported. 6.6 Public expenditure in the Mongolia agriculture sector needs move away from subsidies and bail-outs, towards addressing market failures and producing public goods. Generally, subsidies are inefficient and can lead to the wasteful use of resources. Subsidies can play a role, however, in overcoming temporary market failures, offsetting some of the fixed costs of infrastructureand reducingrisk. Ifused, they should be "market-smart" and contributeto, rather than undermine, the development of markets. If possible, they should be targeted to poorer households, and finally, they should betemporary with a clear programfor phasingout. 132 ChaDter 6 -Agriculture PUBLIC EXPENDITURES 6.7 The share of agricultural spending has declined over time, as have donor resources since they peaked in 2005. Figure 6.2 provides informationon total Government expenditures, expenditures for agriculture on the nationallevel, andthe share of agricultureintotal spending. It is noteworthy that the share of spending for agriculture has declined over time. In 2007, 1.3 percent of total government expenditure was allocated to agriculture. Figure 6.4 also includes informationabout donor expenditures in the agricultural sector. The yearly expenditure data are estimates, as only the total amount over the lifespan of each donor project is available. Donors have spent significant amounts on Mongolian agriculture and some of these funds have most likely substituted for Government expenditures. Figure6.2: Spendingfor agriculture:Foreignassistancecontributes a significant amount -share of agriculture in 20 5% z5 C public spending I15 C v 0 4% g Total expenditure - 5 2 3% .g agriculture and 10 2% 2 B 2 forestry 5 -m +Donor supported 1% projects (agriculture) 0 0% 8 I Source: Ministry of Food and Agriculture (MOFALI) 6.8 When compared against other similar economies,the share of agriculturalspending countries of similar level of development. It is surprising that the 4% share of spending for the agricultural sector in Mongolia is small in 6B 3% international comparisons. It should 8 2% be no surprise that transition countries which are called `fast performers' sYg 1% (Poland, Czech Republic, Hungary and Slovakia) spent a higher share of s total public expenditure on agriculture. Moreover, most other countries shield their agricultural 133 Mongolia: Public Expenditure and Financial Management Review exports o f raw cashmere,' leading to a negative external protection for agriculture and the livestock sector inparticular. Of course, low expenditures do not necessarily indicate that a policy is efficient. The question is whether a policy with negative rates of external protection for the livestock sector is able to provide the necessary public goods for the sector which are needed for a quick transition o f the sector to a market economy environment. 6.9 The allocation of current and capital expenditure has changed significantly over time, with an increased volume and share of capital expenditure in recent years (see Figure 6.4 and Table 6.1). The share o f current expenditure was 91 percent in 2000, but went down to 62 percent in 2006 and is expected to drop further, to 55 percent. This development most likely expresses the Government's will to improve future well-being by investing today, and is largely attributable to the new program for investment in the rehabilitation o f wells. However, it remains to be seen whether current expenditures are sufficient to fulfill all Government tasks adequately. Figure 6.4: Expenditureof the Ministry of Foodand Agriculture from 2000 to 2007 14 .- 5 12 I O capital expenditure ~ + z 8 current expenditure 1 2 6 __ 4 2 0 I 2000 2001 2002 2003 2004 2005 2006 2007 I Source: MOFALI 6.10 Vaccination and disease control account for the largest share of current expenditure, while wells and crop production incentive programs are the largest contributors to capital expenditure. Table 6.1 provides information on the structure o f expenditure. Current expenditure for 2007 will be 62.5 percentage and investment only 32.5 percent. However, some o f the current expenditure items have a clear investment character such as measures which improve the quality o f livestock or enlarge the processing capacities o f milk. Current expenditures are mostly focused on vaccination and medicine; these items make up more than 50 percent o f current expenditures. An important activity is storage o f meat (12.0 percent o f current expenditures). The expenditure for intensive livestock farming ranks third with 9 percent o f public expenditures. The milk program and the improvement o f livestock quality rank fourth and fifth with 4.6 and 4.0 percent o f current expenditure. Investment in the livestock sector is focused on building new wells (28.4 percent o f total investment) and rehabilitation o f existing wells with 25.1 percent. The share o f expenditures spent on the livestock sector i s at least 71 2 The tax on cashmere export leads to lower producer prices for cashmere producers even if only a small share of total production is officially exported and taxed. According to estimates by a World Bank team the export tax reduced income of herder households by as much as US$236 in 1996. See The World Bank, 2003, From Goats to Coats. Institutional reform in Mongolia's Cashmere Sector. Report no. 26240-MOG, p.40. 134 Chapter 6 Agriculture - percent. Some expenditure items, such as `food security,' `organic farming,' green revolution,' and `internal current expenditure' are also partly in favor o fthe livestock sector. Table 6.1: Expendituresfor agriculture 2000 to 2007 (tog million) - 2000 2001 2002 2003 2004 2005 2006 2007 1) Current Expenditure 4,204.3 8,110.3 10,285.6 8,508.9 6,979.7 6,542.5 8,610.9 8,800.4 Measures o f precaution 2) 2,060.9 2,897.3 3,034.6 3,154.0 2,447.6 2,684.7 2,619.8 3,256.6 Vaccination against Food animal disease 460.0 1,866.0 2,341.0 2,049.7 1,710.4 1,996.1 2,000.0 Measures o f precaution against bird flu 3) 4,765.4 105.3 Plant protection 402.0 565.0 849.0 645.0 950.0 649.5 647.7 1,000.0 Purchase of goods for State reserve 1,100.0 1,100.0 2,960.0 1,925.2 Livestock breeding 13.0 30.0 35.0 35.0 35.0 35.5 33.9 35.5 Green revolution 150.0 180.0 66.0 70.0 150.0 120.0 120.0 200.0 Elite livestock male breeding 180.0 140.0 140.0 National program defence from dzud 1.397.0 1,000.0 427.0 700.0 99.8 Iodine Deficiency disorder program 13.0 10.0 15.0 15.0 Food supply, food security and nutrition 25.0 25.0 24.6 25.0 Dairy production national program 463.0 Organic food program 20.0 Cooperative development program 10.0 35.0 35.0 35.0 35.0 Improvement of livestock quality program 400.0 Intensive livestock program 50.0 294.5 289.5 900.0 Agricultural extension trainning 12.5 Search of ground water 18.4 75.0 Veterinary service 714.2 447.4 Premium for best herders and farmers 25.0 Internal current expenditure /IFAD, ADB 335 0 Capital Expenditure 402.0 394.0 515.0 450.0 1,680.0 1,475.5 5,241.7 7,191.1 Well rehabilitation 4 0 2 0 3 9 4 0 5 1 5 0 4 5 0 0 8 8 0 0 7 5 7 5 3,5150 New wells Irrigated crop production 8 0 0 0 7 1 8 0 8 7 0 0 2,0000 Food storage/ meat reserve 700 0 1,2000 Repair buildings of ag. depts. in aimags 1567 1670 Equipment for ag. depts. in aimags 132 0 New buildings o f ag. depts. in aimags 293 6 Investment o f Parliament members 193 5 Total expenditure 4,606.3 8,504.3 10,800.6 8,958.9 8,659.7 8,018.0 13852.6 15,991.5 Share o f current expenditure in total expenditure in % 9 1 3 8 7 9 9 5 2 9 5 0 8 0 7 8 1 6 6 2 2 55 0 1) approved 2) 2007: incl. Chronic disease 3) included expenses chronic disease and vaccination against animal disease - Source: MOFALI 6.11 At the local level, Government programs and policies are poorly implemented. Aimags and soums are mainly in charge o f implementing agricultural policy measures and complementing policies with local measures. The aimag Food and Agricultural Office operates as an agency budgeted from the MOFALI. A survey of four aimags looked at expenditure categories in rural soums and found that recurrent costs made up an average of 96 percent of the total b~dget.~ Most o f this went to salaries and to goods and services -mainly heating and fuel- leaving very little for other expenditure^.^ Under-budgeted programs lead in turn to poor implementation of projects in agriculture at the local level, suggesting the need for improved budget planning. 6.12 Central-local lack of accountability, poor local ownership of central government programs and imperfect information contribute to poor implementation of programs. Because the budget is paid by the Central Government, the central Government may not have the best information about the needs at the local level. However, the local government may not feel Informationon local spending inagricultureis not easily available.The main and only visible expenditure for livestockare for veterinary services and investment inwells. Center of Mongol Management(2004) Fiscal decentralization and localpublicflnances in Mongolia 135 Mongolia: Public Expendilure and Financial ManagementReview accountable for performing the tasks at minimal costs (see below). This issue is common to other sectors, as it pertains to the lack of horizontal accountability for vertical reporting structures of the intergovernmental fiscal system(see chapter 2). THEBUDGETCYCLE 6.13 The actual budget planningprocess does not provide the right incentiveat the local level to reveal an accurate and realistic estimate of needs. The procedure for deciding on the agricultural budget follows the general procedure for deciding on the budget, with line Ministries making recommendations to the Ministry of Finance. The Parliament approves the budget after the Cabinet of Ministers has agreed to it. The main issue is the lack o f accurate information received at the central level. The MOFALI collects information from the aimag and the aimag in turn receives the information from the soums. There are incentives for the aimag and soums to exaggeratetheir needs, rather than to provide an accurate picture of what should be requested(see Box 6.1 for an example). Box 6.1. Inefficiencies in Budget formation and execution: the example of new wells Soums make the initial budget request for a new well. The request is likely to be over-estimated, as soums are better off if they can improve their well systems and they do not bear the costs of requesting more than what is highly required. When submitting their requests, soums do not have to deliver a reasonableproject appraisal which would include economic and financial returns calculations.Nor do they have to show evidenceof externalities associatedwith the building of a well, such as pasturedegradationor the impact on poverty reduction. In turn, the MOFALIrequest for wells is far out of the range of the set ceilings, which underminesthe credibility of the request and leaves it up to MOF to cut spending. The proposal submitted to the Parliament contains a lower number of projects than originally requested, but still contains an allocation system to the aimags and soums. The MOFALI has the task of allocating the budget to individual aimags. To this end, six criteria are used for wells funds allocation. These are the current number of wells, number of animals, past expenditure for rehabilitation and construction of new wells, water availability, capacity of pasture, number of rehabilitated wells. The proposed allocation of funds does not include any information about economic or financial return or expectedoutcomesonpovertyreductionor economic development. Once approved, the lowest administrative levels (soum, bag) are notified that they are entitled to a number of new wells and rehabilitated ones. But they are not responsible for executing the funds. The aimag offers a tender and generally 4 to 5 companies apply. At the time of submitting the tender the companies do not know exactly where a new well may be built. Evenifthe locationhas been identifiedby the aimag changes may occur, because of technical issues (costs may be too high at the foreseen location due to earth conditions or lack of water at low levels) or administrative intervention. Anecdotal evidence suggests that soum administrators may have leeway in deciding on the exact location, leaving room for patronage. This step of the procedure has three significant shortcomings: First, the aimag is unlikely to seek to minimizethe costs. Ifa company actually needs less money than foreseen in the budget the savings will be at the central budget, because the center pays all bills. Second, it may be that the small number of bidders and the known available funds lead to collusion among bidders. Third, work monitoring has to be carried out by the lowest administrativeunit, which is adequate. However, as the lowest units do not save money for themselvesthey may have incentives to accept somewhatoverpricedservices. 6.14 Lack of capacity at the central level and information asymmetry limit the ability to assess and monitor policies and programs. The Ministry is hampered by not having a strong analytical unit.The Information, Monitoring and Evaluation Department of MOFALI is in charge o f monitoring and evaluating the expenditure requests from aimags and assessing the effectiveness o f public expenditure in the sector. However, the agency's annual report is mainly based on the reports of the lower administrative bodies (soums) and does not include policy evaluation usingwidely accepted criteria such as relevance, effectiveness, impact, efficiency and sustainability of projects. Hence, these reports are not sufficient to inform policy makers about 136 Chapter 6 -Agriculture the performanceof the projects.It is not knownwhether the whole programof well rehabilitation and constructionor other specific projectscontributeto an increase inthe overall living standards of Mongolian people. It is often not known whether the envisaged beneficiariesactually gain in the medium-termfrom the project. EFFICIENCY 6.15 Inthis section, the efficiencyofthe mainpublic programs for agriculture is analyzed.The main programs are: (i)well rehabilitation and building of new wells; (ii) irrigation and land reclamation for crop production; (iii)vaccinations and animal disease response; (iv) the IntensifiedLivestockDevelopmentProgram; (v) meat storage; (vi) milk program(National White RevolutionProgram); and (vii) cooperative development. (i)The rationale of the well rehabilitation and construction program needs to be carefully reviewed. Government investment in wells can be justified, though only after an evaluation of the investment on environmental and economic grounds. The government has launched an ambitious program to rehabilitate and construct new wells, the economic justification of which can be debated. Government investment in the constructionof wells is onlyjustified from an economic point of view ifthe rehabilitationand/or constructionof new wells exert externalities. Wells in Mongolia open up new areas of pasture, which are accessible to all herders in the area, but can also degrade pasture land by increasingstocking concentrations.However, in this case the externality is not based on the wells, but on the behavior of herders.Hence, the Government should exert incentivesfor herdersto use pasture appropriately.Ifthere are ways to provide incentivesto herders for co-operation in changing the grazingpatterns subsidizingthe constructionand rehabilitationof wells can beju~tified.~ (ii)TheFoodSupplyandSafeFoodProgramobjectiveistoproduceatleast80percent of food consumption domestically, yet the economic rationale of this objective is not clear. Policy measures to increase overall productionof cereals, meat and milk among other products are implemented according to this objective. However, the premise on which the policy is based, that an increase in domestic production will improve food security, can be questioned in a market economy, where the production patterns should reflect some comparativeadvantages.For instance, it is not clear that Mongolia would have a comparative advantage inthe large scale productionof cereals, milk, pigs, or poultry. Support to increasingfood supply, including for irrigation works, is now the largest capital expenditure in the Ministry's budget. The previously stated objective of targeting 80% food self sufficiency has, in 2008, been increased to achieving 100 percent self-sufficiency in major commodities by the year 2012, includingwheat, vegetables and milk. This objective is almost certainly unobtainable,and is likely in part driven by the recent spike in global food prices. Government is intervening in several ways to stimulate food supply, including subsidizing the importation of machinery (mostly tractors) and fertilizers and subsidizing long-term credit for agriculturalbusinesses. It is too early to say whether these interventions will achieve the stated policy objective and concerns persist on the possible underminingof local input markets and whether subsidies are beingcaptured by non-agriculturalentities and activities. 5Two pieces of legislation (an Amendment to the Land Law and a Pastureland Bill) are currently being preparedthat couldprovideherder groups possessionrights over pasture land. Ifenacted and enforced, this would underminethe economic argument for governmentsubsidies for wells, as it would be expected that herder groups with possession rights to surrounding pastureland to invest their own resources in the rehabilitationandmaintenanceofwells. 137 Mongolia: Public Expenditure and Financial Management Review A more reasoned approach to supporting food production should be adopted. While a return to food production surpluses is not feasible without very high subsidies (as was the case in pre-transition crop production), there i s potential for increasing food supply to a lesser extent. There has been a recent expansion o f small scale irrigation schemes mostly for vegetable production, which can provide a useful alternative income source for rural communities. Government can support this by provision o f technical services. For larger scale irrigation schemes, especially for wheat production, the government needs to be aware o f the trade-offs between the largely political motive for reducing dependence on imports and the economic rationale for the significant investments (both capital and recurrent) to support the sector. (iii)While Government spending for vaccinations and animal disease response is warranted, the scope and effectivenessof the interventionsneedto be assessed carefully. Outlays for veterinary services are justified to the extent that they prevent the spread o f infectious diseases, which can restrict exports o f livestock products. Budget outlays for veterinary services (medicine and vaccination) are about one third o f current expenditures. These outlays partly finance those activities which the Government performed before privatization. The budget for veterinary services is spent on infectious, parasitic, Food and Mouth Disease (FMD), and disinfection and sanitization efforts. The Government pays for vaccines, medicine and veterinarians' service fees for infectious diseases, but for parasitic diseases, herders are responsible for veterinary service fees. Herders also need to pay for the disinfection and sanitization o f livestock. In principle the approach i s adequate, as Government spending contributes to overall welfare in the society. The open questions are: a) i s there scope for improvement o f the system that would make outlays more efficient? and b) i s there a need to increase outlays in order to increase effectiveness? (iv) The Intensive Livestock Development Program covers a broad set of activities aimed at the intensificationand commercializationof the livestock sector, especially in peri-urban settings. The program is scheduled for two phases, one from 2003 to 2008 and the other from 2009 to 2015. The government is prepared to finance part o f the program out o f own resources, but succeeded in motivating donors to contribute. It started in 2004 with only Tg 50 mln, grew to Tg 289 in 2006 and is planed to total Tg 900 mln in 2007. Expenditure for `Intensive livestock farming' was 9 percent o f current expenditure and 5.6 percent o f total expenditure o f MOFA in 2007. `Improvement o f livestock quality' is the second largest subprogram, and will amount to 4.0 percent o f current expenditure in 2007; The program addresses important issues, but direct intervention by the Government into the sector i s not the most cost-effective approach. Intensity o f livestock farming has declined since 1990; moreover, migration from rural to urban regions has reduced the need for subsistence farms and increased the demand o f the growing urban population. The program has two different sets o f approaches: indirect support to herder groups or pasture management, and direct market intervention by the Government. The first set o f measures adequately addresses the cause o f many problems in the Mongolian livestock sector. The Government might even consider putting more effort into these activities in order to make them more effective.6 'The second set o f measures directly affects farm production and company level. The Government provides, among other things, breeding animals, tractors, equipment and credit at reduced prices. While these measures increase the yield per animal (milk per cow, meat per animal, value o f animal products per producer), their cost- effectiveness is questionable. Indeed, the income transfer effect o f the programs is not under See for the numeration of some shortcomings of present support Mau, G. and G. Chantsallkham, 2006, Herder Group Evaluation. Ulaan Baatar. 7 See also the Final Report of Pasture Demonstration Area Project, UlaanBaatar (2006) and Center for Policy Research (2006) Community mobilization for pasturelandmanagement. 138 Chapter 6 - Agriculture question, but it is not clear that the beneficiariesare the ones in need or the poorest people. Neither is it clear that any growth or multiplier effect materializes out of these measures. Moreover, the measures most likely do not direct investment to the most efficient use. Decisionsare rarely made based on strong economic ground, as government officials do not have enough informationto allocate machinery, tractors and equipment to the most efficient farmer. A more effective approach would be to remove the main constraints to growth of intensive livestock. More generally, research could be conducted to inform the design of a planned program. It is likely that lack of information and knowledge and badly functioning product and input markets (including credit) are the main constraints. More generally, the Intensified Livestock Development Program did not benefit from informed analysis that might enhance the cost-effectivenessofthe program. (v) Meat storage expenditures are significant but have not had the desired impact and might be done more efficientlyby the privatesector. The share in 2007 was 7.5 percent of agricultural expenditure and 12 percent of current expenditure. The justification for public intervention is to stabilize markets. Indeed, there may be a need to intervene in markets as seasonality of meat supply is significant and consumption much less so. If markets were to functionwell, private businessmen would buy at times of low prices and sell at times of high process. Ifthe private sector has not overtaken this business it may well be due to the high risk. Of course, at the time of purchase the selling price in the future is not known. Lack of market information is one important determinant for the inactivity of the private sector in storage of meat. However, even with a reasonable market information system the remaining risk might be too high for private traders. There is a strong presumption of market failure. The present systemhas the drawback that the risk for privatetraders may be even enhanced if the activity of the state is not predictable. For instance, there seems to be no information about the prices at which the Government may buy and sell the meat. So the risk for private traders, slaughter houses and meat processors may be even enhanced and the evolution of a privatemarket for meat storage is suppressed. (vi) Public support in the milk sector (National White Revolution Program) should focus on improving the legal and regulatory framework and avoid subsidizing the sector. The government has tried to intervene in the dairy sector based on the observation that there is a significant need for imported milk and that milk produced domestically and delivered to consumers does not meet always hygienic and food safety standards. Ensuring food safety is indeed a legitimate reason for public intervention, but increasing milk productionby supportingdairy farmers will not treat the cause of the problem. Froma private perspective, the expansion of the milk sector is profitable due to the public support, but from an economic point of view it may not be profitable, and may not result in an overall benefit for society. Most of the capital requirements should be provided by the private sector and bank lending. But the Government has a key role to play, particularly in improving the relevant legal environmentand in supportingtrials, demonstrations and informationtransfer. The focus should be on strengthening the vertical linkages across the private agents in the value chain.' (vii) The Cooperative Development Program supports cooperation among herders. Plans for the program includedsignificant amounts of trainingfor cooperatives and soft loans for the establishment of secondary cooperatives with processing and marketing facilities in the four western provinces.Underthe cooperative development component, almost MNT 400 million in soft loans was providedfor various activities of primary cooperatives and to some 8The FA0project onpasturemanagementapplies a holistic approachwhich is commendable. 139 Mongolia: Public Expenditure and Financial ManagementReview individuals. Repayment currently amounts to only 32 percent of the amount due. Cooperatives were also expected to make much use of the credit line to cover their operational expenses, but so far their share in total lending has been only 2.35 percent, or about MNT 390 million. Cooperative training was not provided to the extent foreseen in the project design. This is an area where the Government can play an active role, as cooperationdoes not always spontaneously emerge from market forces. Cooperatives can significantly improve income and employment opportunities of their members when they focus on, for example, providing better processing and marketing facilities for the members' livestock products. However, expectations have not been met in all respects. Learning lessons from past project performance and building on this experience is required. For instance, cooperation among herderscouldbe further promoted insteadof direct supportto cooperatives. Important areas for future cooperative development include: i)business development for cooperativesthrough special trainingand business consultations; (ii) external supervisionand enforcement of membership promotion of cooperative operations through audits; (iii)more involvement of provincial cooperative promotion centers in loan appraisal, monitoring, and collection; (iv) possible development by Project FinancialIntermediary(PFI) of specific loan products for cooperatives supported by capacity building to PFIand cooperative staff. (viii) The rationale for Government intervention in reserve management to protect humans and animals against emergencies is well justified. In addition, this role is of particular significance for Mongolia, where weather conditions are less stable and predictable than in other countries.The reserve system, includinghay and fodder for disaster preparation and recovery, is operated under the NationalEmergency Management Agency. A law which became effective in 2004 defines the tasks for preparing for an emergency. As part of the program, training and education are providedto prepare administratorsandthe privatesector. In addition, the Government and local governors are responsible for managing reserves. Informationon what kind of reserves andhow much is heldat the local level is, however, not always available, which makes it difficult to determine whether the system is appropriate or not. DONOR SUPPORT 6.16 Donors play an important role in the agricultural sector (particularly the livestock sector), though support has declined in recent years. The importance of donor projects is highlightedby the share of money spent as a percent of the agricultural budget. Unfortunately, there is no official data which provides information on annual expenditures of the individual donors. Hence the data in Figure 6.2 are based on averages: the total expenditure over the life spanof an individual projectwas dividedby the number ofyears of its life span. The total amount of donor's expenditure amounted to only 7 percent of public expenditure for agriculture in 1996 but grew to 145 percent in 2005, declining over the following two years and forecast to make up 77 percent in 2007. 6.17 Effective coordination and a better insertion of donor projects into government policies would improve long-term outcomes in agriculture. Links between donor-funded projects and government strategy are weak. All donor projects have a capacity building component which is building in some sustainability. However, the mainvalue of donor projects is their role as pilot projects.Ideally, the findings will allow the Government to base the design and implementation of policies on improved information. But in practice, the link between donor projectsand Government policies is weak. The Government is involved in the project design, but the findings do not seem to be exploitedas well as they could be. One critical challenge for better 140 Chapter 6 Agriculture - use of results is the absence of a policy analysis unit in the MOFALI. Such a unit could systematicallyanalyzethe impact of projectsandfeed the findings into Government policy. As of November 2008, the Government is proposing structural changes to MOFALI, which would include the establishment of a new Economic Policy and Aid Coordination Division and, for pastureland management, an Agriculture Common Lands Relations Division. Moreover, proposals have been tabled to strengthen the policy analysis and monitoring capacity of the Ministry by increasing staffinglevels.These are positivedevelopments. OPTIONSFORREFORMS-RETHINKINGTHE ROLEOFTHE GOVERNMENT 6.18 Publicexpenditures in agricultureneedto adjust to the reality of a marketeconomy. In a market economy, market forces play a major role in coordinatingindividual decisions. The Government's main responsibilitiesare to produce public goods and reducemarket failure. Inthis context, and in order to maximizepublic interventioncost-effectiveness, two conditions should be met: first, the decision maker must have the best possible information about the effects of alternativedecisions and second, the decisionmaker must be accountable. 6.19 The Government could focus on improving the functioning of markets and addressing failures. Mongolia relies heavily on self-enforcing markets. The large share of the informal sector (between 30 and 40 percent of GDP according to estimates) in the economy indicates the importance of these markets.' Herders often sell their products to traders in exchange for other goods or sell milk or meat directly to consumers on the informal markets. Self-enforcing markets will be functional if parties involved in the exchange have an obvious advantage and the exchange does not includeany uncertaintyor risk, which is not the case in the Mongolia livestock sector. 6.20 Informationfor policy makers needsto be improvedthroughthe conduct of regular ex ante and ex post policy assessment in order to improve the efficiency of public expenditure for the agriculture sector in Mongolia. Often, information is incomplete. However, any decision maker should try to get the most accurate information about expected policy outcomes and associated costs, and to analyze ex-post realized outcomes. Hence, policy analysis is needed ex ante and ex post (see Box 6.2). Policy analysis should document the expected impacts of policies and their efficiency. But because of a lack of policy analysis capacity, policymakers do not benefit from the most accurate and comprehensive information when making a decision in the agriculture sector, thereby weakening the quality of decision- making. 6.21 More decentralized budget planning would improve decision making. Public expenditure efficiency can be enhanced if it reflects the preferences of involved stakeholders. Those who are affected by a specific expenditure should be able to raise their voices. A more bottom-upapproach in budget planningwould help to ensure that local needs and preferences are taken into consideration. But good decisions can only be expected if decision-makers are financially accountable. At present, the local levels can articulate requests for the most important investments, such as wells, but they do not have to take into account the costs involved. Real grass roots participation and accountability in the budget process needs to develop over time. Improvements in the efficiency of public expenditure requires improving bottom-up budgeting andprioritization, basedon expenditure ceilingsprovidedby the center. 9 USAID, Open Society Forum and IRIS, "The size and Character of the informal sector and its shadow economy inMongolia" (2005) 141 Mongolia: Public Expenditure and Financial Managemenf Review Box 6.2: Analytical requirementsin the Ministry of Food and Agriculture The overall evaluation of expenditures in the agricultural sector is constrainedby limited capacity in the Ministry to conduct policy and program analysis. The Department for Information, Monitoring and Evaluation has only three staff and reporting is limited to a management performance audit, with no economic assessment. A good program evaluationwould respondto the following questions: 0 What is the relevance of the program? I s the Government doing the right thing?Do the objectives of the project or programmatchthe needs ofthe target groups? 0 I s the program effective? Does the projech'program achieve the objectives? To what extent are the desireddirect results achieved and what other indirectresults arise? 0 What is the impact? Does the project or program contribute to the achievement of overarching development results?To what extent does the project or program contribute to achieving the desired overarching objectivesand produceother indirectdevelopmentresults? 0 I s the government activity cost efficient? What i s the profitability of the program or project? 0 Sustainability: Is the outcome of the activity durable without further governmental support in cases where activity is plannedas temporal measure? 6.22 The Government should support improving livestock productivity by supporting education, training, extension services and information availability. Productivity o f animals and enterprises along the value chain o f livestock seems to be low in international comparison. Two factors seem to be more significant than others: lack o f expertise o f herders and lack o f information. There are a number o f herders who started the business in the 1990s without having specific knowledge about raising livestock. Understandably, the productivity o f their animals is low. Moreover, they lack pasture-management know-how. The Government can provide education, professional training, extension, and information to improve their capacity and skills. International comparisons prove that a well designed extension service can yield high economic returns. Interviews with successful farmers reveal that their success was related to their exposure to international good practice and expertise. The MOFALI should place priority on setting up an extension service and facilities and staff for professional training. 6.23 There are four possible paths to improve market functioning that are worth investigating: market information systems; grading and classification systems; support to the rural finance market; support to feed and fodder markets. 0 A market information system could inform on prices in different regions and make price forecasts. Less uncertainty about prices would make production for the market less risky. Herders would be more likely to move from subsistence agriculture to marketable production. Moreover, herders would be able to take expected prices into account in their decision to sell immediately or wait. One positive side-effect o f the system would be the increase in competition among traders and improved choices o f herders to whom to sell. Consequently, prices for herders would likely be higher. ii) The Government could contribute to a grading and classification system for specific agricultural and food products. Without collective action it is very difficult to set-up a grading and classification system. And without such a system the market tends to concentrate on low quality products which are associated with lower production costs. A buyer will pay higher prices only if he can check that the quality is superior. Under these circumstances, a market for high quality products can be facilitated if an impartial grading and classification system is put in place and if regular inspections take place. iii) TheGovernmentcouldfacilitateimprovementsinaccesstofinancialservicesin rural areas. Access to finance is important for diversifying the agricultural sector and for the development o f value chains. Herders continue to have limited excess to 142 Chapter 6 -Agriculture credit even though they can use animals for collateral.Furthermore, lending rates in Mongolia are relatively high compared to other similar economies and there is limited availability of long term funds. In the past, Government has undertaken measures to inject liquidity into rural areas through the banking and non-banking sectors (for instance under its Micro-finance Development Fund). However, more could be done to facilitate the development of new lendingproducts targeted toward rural areas, expand the availability of data on rural credit with enhancementsto the credit bureau, improvethe capacity of loan officers in reviewingcredit proposals and provide assistance to herders and other rural areas residents in developing sound business plans and credit applications. iv) The risk of producing livestockwould be lower if feed and fodder markets were developed. At present, herders are more than necessary hit by dzuds as they have little access to fodder, despite the state reserves. Moreover, the emergenceof a fodder markethas been suppressed, partly due to the lack of transparency of the activities of the NationalEmergencyManagementAgency andthe reserve activities ofthe aimags and soums. 143