Report No. 37933-MD Republic of Moldova Improving Public Expenditure Efficiency for Growth and Poverty Reduction A Public Expenditure Review for the Republic of Moldova February 12, 2007 Poverty Reduction and Economic Management Unit Europe and Central Asia Region Document of the World Bank TABLE OF CONTENTS EXECUTIVE SUMMARY ................................................................................................. i . REDUCTION..................................................................................................................... 1 CREATINGFISCAL SPACE FORECONOMIC GROWTHAND POVERTY 1 A 1 B RECENT 2 C OVERALLFISCAL PERFORMANCE 4 D I STHE SIZEOFGOVERNMENTTO0LARGE? ... . INTRODUCTION ........................................................................................................ MACROECONOMIC DEVELOPMENTS ............................................................................. ........................................................... ............................................................. 6 E. IsTAXATION LIMITINGGROWTHPROSPECTS? ........................................................ 9 F. COMPOSITION OFEXPENDITVRES G PUBLIC FINANCIAL MANAGEMENT H CONCLUSION .. ........................................................................... 11 .......................................................................................................... ........................................................................ 16 17 AND MANAGEMENT INMOLDOVA ......................................................................... 2. INVESTINGFORECONOMIC GROWTH: PUBLIC INVESTMENTPLANNING19 A INTRODUCTION . ........................................................................................................ 19 3 SAVING THE ROADNETWORK ............................................................................ . 33 A THEDETERIORATIONOFMOLDOVA'S .. ROADASSETVALUE.................................... 33 B "SAVING MONEY" BYNOT MAINTAINING ROADS? c. .................................................... 33 34 D CURRENTSYSTEMOFFINANCINGFORROADS . ECONOMIC EFFECTSMOLDOVA'S OF DETERIORATEDROADS .................................. ......................................................... 37 4 MANAGING CIVIL SERVICEPAY REFORM., . ...................................................... 41 A. INTRODUCTION ........................................................................................................ 41 B. RECENTDEVELOPMENTS INPUBLIC SECTORPAY 41 C CIVIL SERVICE REMUNERATION . ................................................... 43 D. OPTIONS FORCIVIL SERVICE PAYREFORM SYSTEM................................................................. ............................................................. 46 5. INCREASINGTHE EFFICIENCY AND QUALITY OF HEALTH CARE SERVICES........................................................................................................................ 49 A. INTRODUCTION c. CURRENTSYSTEM HEALTH ........................................................................................................ 49 OF EXPENDITURES 50 D HEALTH . ........................................................ SECTORPERFORMANCE ............................................................................. 52 E. RECOMMENDATIONS ................................................................................................ 54 6 ENHANCINGTHE EFFICIENCYOFRESOURCEUSE INEDUCATION........... 57 . A. INTRODUCTION ........................................................................................................ 57 B COSTSOFEDUCATION ............................................................................................. C FINANCINGSYSTEMFORPUBLIC EDUCATION.......................................................... 58 60 D SCOPEFORIMPROVEMENTSINEFFICIENCY ............................................................. 61 E FISCAL SPACE? 64 F.. CONCLUSIONS.......................................................................................................... ... ........................................................................................................ 66 7 PROMOTINGTHE FISCAL SUSTAINABILITY OF THE PENSIONSYSTEM...69 . A INTRODUCTION ........................................................................................................ 69 B PENSION INCOME .. ..................................................................................................... 70 C. TOWARDS A UNIFIED CONTRIBUTORY SYSTEM ......................................................... 72 D SETTINGUPTHEPENSIONFUND'SREGULATIONSAND SUPERVISORYINSTITUTIONS 75 . E CONCLUSIONS .......................................................................................................... . 79 8. SOCIAL ASSISTANCE BENEFITS INMOLDOVA................................................ 81 A INTRODUCTION . ........................................................................................................ 8 1 B. SOCIAL ASSISTANCEEXPENDITURES 82 C OPTIONS FORREFORM ............................................................................................. D POLICY RECOMMENDATIONS................................................................................... .. ........................................................................ 88 E. CONCLUSION ............................................................................................................ 90 95 Index of Tables Table 1.1: Moldova: Macroeconomic Framework. 2001-2008 ..................................................... 3 Table 1.3: Revenues and Expenditures Relative to the Predicted InternationalNorm...................7 Table 1.2: Revenues and Expenditures inMoldova, 2001-2005 .................................................... 6 Table 1.4: General Government Revenue Trends......................................................................... 10 Table 1.5: General Government Expenditures - Functional Classification 12 Table 1.6: General Government Expenditures - Economic Classification .................................. .................................. Table 2.1: InvestmentBudget Allocation 2005 ............................................................................ 14 26 30 Table 3.2: Moldova - Losses to Roadusers due to inadequate Road maintenance...................... Table 3.1: Moldova-Road Sector Expenditures 1997-2006 (actual vs.needed) .................... Table 2.2: R-Coefficients: World Bank Projects.......................................................................... 35 Table 5.1: Composition o f Health Expenditures in2005 ............................................................. 36 Table 5.2: Selected Quality and Health Care Indicators............................................................... 50 53 Table 5.3: Efficiency Indicators of the MoldovanHospital Sector Table 6.1: Allocation of State Fundsto ATU budgets: Norms per Student, 2003-2008 (lei) .....54 .............................................. Table 6.2: Selected Education Indicatorsby level, general schools, 2005 ................................... 60 62 Table 7.1: Average Wage and Replacement Rates ....................................................................... 72 Table 7.2: Employment and Number o f Contributors li 73 Table 7.3 : Pension FundFinancial Results 2002-2005 ................................................................ ............................................................... Table 7.4: Medium-term Pension System Financial Balance....................................................... 76 77 retirement ....................................................................................................................................... Table 7.5: Medium-term Pension System Financial Balance Gradual increase o f pensions at 78 Table 8.2: Poverty Gap ................................................................................................................. Table 8.1: FGT Poverty Measures With and Without Social Assistance Benefits....................... 87 88 Index of Figures Figure i: Expenditures. Revenues and Deficit. ............................................................................ Figure ii: Large Governments Limit Economic Growth............................................................. ii Figure iii: Shrinking the Size o f Government ................................................................................ iv v Figure 1.3: Reducingthe Size o f Government by Constraining Expenditure Growth ...................9 Figure 1.2: The Size o f Government and Economic Growth ......................................................... Figure 1.1: Expenditures, Revenues and Fiscal Balance ................................................................ 5 8 Figure 1.4: Arrears on Public andPublicly GuaranteedExternal Debt (Percent of GDP) Figure 1.5: Revenues and Expenditures by Levels of Government. 2005.................................... ...........13 15 Figure 2.2: Public Investment Spending:...................................................................................... Figure 2.1: Public InvestmentinMoldova: ................................................................................. -21 21 Figure 2.3: Composition o fPublic Investment in......................................................................... Figure 3.1: Revenue Sources ........................................................................................................ 22 Figure 4.1: Wages and Salaries inPercent of Total Government Expenditure 2003 ...................42 37 Figure 4.2: Index of Monthly Salary ............................................................................................ ......................................................................... 43 45 Figure 6.1: Fiscal Space inEducation: Possible Saving & Extra Spending, 2009 ....................... Figure 5.1: Composition of Public Health Expenditures by Budgetary Account, ........................ Figure 4.3: Compression of Civil Service Wages 51 Figure 7.1: PensionBenefits Levels incomparison to the Average Wage................................... 66 70 ............................................................................... Figure 7.3: Elderly inpercent of Working-age Population........................................................... Figure 7.2: Pension Income at Retirement by 71 76 Figure 8.2: Distribution of Social Benefits by Income Quintile ................................................... Figure 8.1: Structure of Expenditure and Beneficiaries ............................................................... 85 86 Figure 8.3: ExpenditureForecast (Scenario 1) ............................................................................. 89 Index of Boxes Box 1: Issues inPublic ExpenditureManagement: Summary of Previous Findings...................18 Box 2: Moldova's Experience with PIPS, 1994-1997................................................................... 23 Box 3: InvestmentBudget Preparation: Some Examples ............................................................ 27 ................................................................................. Box 5: The Experience with Second Generation RoadFunds...................................................... Box 4: A Framework for Public Investment 30 Box 6: The Government Reform Strategy.................................................................................... 39 91 ACKNOWLEDGEMENTS The Public Expenditure Review (PER) has the objective of addressing key issues in public finance confronting Moldova and recommendingoptions for reform. Giventhe breadthof issues to be covered, the PER is being undertaken in a programmatic approach. This first phase explores expenditure allocation and efficiency in a number of key sectors - public investment, public administration, roads, health, education, pensions and social assistance. The Public Expenditure Review was prepared by a World Bank team lead by Lawrence Bouton and ErwinTiongson. The primary authors of the study are Lawrence Bouton (Chapter l), Erwin Tiongson and Iaroslav Baclajanschi (Chapter 2), Svetlana Proskurovska (Chapter 3), Andreas Schliessler (Chapter 4), James Cercone (Chapter 5), Martin Godfrey (Chapter 6), Paulette Caste1 (Chapter 7) and Paolo Verme (Chapter 8). Generous financial support to the study was given by the UK Government's Department for International Development (DfiD). Elena Nikulina prepared the update on public financial management issues for Chapter 1. Anca Dumitrescu provided useful inputs to Chapter 4 during the initial stages of the study. Hideki Mori provided valuable inputs and strong guidance throughout the course o f this study. Andrew Keith (DfiD) also participated in the overall preparation of the study and his contributions are gratefully acknowledged. Similarly, the PER benefitedfrom the comments received by Hans Lundquist (SIDA), Thomas Richardson (IMF), Jose Leandro, Andreas Papadopoulos, and Lucio-Mauro Vinhas-de-Souza (EC), Andrew Bird and Simon Groom (DfiD consultants). The team gratefully thanks the Government Counterpart Team, ledby the Minister of Finance, Mihail Pop, for guidance, cooperation, and support at all stages o f the PER preparation. The efforts of the Ministry of Finance, the Ministry of Economy, and other agencies in helping prepare this report are gratefully acknowledged. The efforts of the counterpart team were coordinated by Vasile Bulicanu. The team itself consisted of Natalia Rata, Marina Semeniuc,Veaceslav Serbin, Liudmila Sverdlic, Svetlana Purici, Gheorghe Sircu, Svetlana Bogatu, Lilia Gantea, Nicolae Ciobanu and Vasile Botica. The team benefited greatly from numerous discussions with the representatives o f the donor community, local think tanks, independent experts and numerous other stakeholders in Moldova. The advice and guidance of EdwardBrown, Country Manager, throughout the study i s also gratefully acknowledged. Ulrich Lachler (LCSPE) and Edward Mountfield (PRMED) were peer reviewers for the study. Viorica Strah and Olesea Vinaga were instrumental in coordinating the consultation process in Moldova. Susana Padilla provided able support with the production of the report. Asad Alam was the Sector Manager supervising the preparation of the study. Cheryl Gray was the Department Director, and Paul Berminghamwas the Country Director for Moldova. ImprovingPublicExpenditureEfficiencyfor Growthand PovertyReduction EXECUTIVE SUMMARY I.Introduction 1. The Government o f Moldova's Economic Growth and Poverty Reduction Strategy Paper (EGPRSP) lays out an ambitious plan for sustaining growth and poverty reduction and reshaping the government to meet the needs o f a market economy. The public expenditures envisaged under this ambitious plan, however, vastly exceed the domestic resources available to the Government. Additional foreign budgetary support may help alleviate some o f that resource constraint. Recognizing that the share o f tax revenues and expenditures to GDP in Moldova already greatly exceed comparable international levels, generating additional domestic tax resources risks crowding out the private sector and undermining growth prospects. This suggests that in order to finance higher order public expenditures priorities, the Government needs to create fiscal space from within the existing resource envelope. This will require inter and intra-sectoral reallocation o f expenditures and an increase in the efficiency o f public spending rather than increasing the relative size o f government. 2. Fiscal space can be generated in a number o f manners. Changes in tax policy and administration to broaden the tax base and increase revenue potential without distorting or undermining the private sector incentives to invest and save i s one option. On the expenditure side, fiscal space can be created by shifting resources from less efficient to more efficient uses. Onthe financing side, fiscal space can be created by mobilizinggrant aid, privatizing state assets, or by accelerating the reduction o f public debt so as to reduce the amount o f expenditures devoted to interest payments on debt. There i s also a strong inter-temporal dimension to fiscal space. The effective use o f public resources to improve human and physical capital will lead to increased productivity and income and, hence, expand the scope for private and public consumption opportunities inthe future. 3. This Public Expenditure Review (PER) concentrates on some key reforms that are needed to improve the functioning and efficiency o f public spending as well as create fiscal space for growth enhancing public expenditures. 11. ChangingtheParadigmfor Growth andPovertyReduction 4. Moldova's strong growth performance since 2000 has reversed a decade of economic decline and risingpoverty. The economy has grown by nearly 40 percent leading to a massive movement out o f income poverty. Starting in 2004, however, poverty reduction has stalled despite continued vigorous economic growth. As highlighted in the recent Country Economic Memorandum (CEM), this trend has heightened concerns that Moldova's growth paradigm is no longer reducing poverty. 5. Up until now, growth in Moldova has relied primarily on the expansion o f domestic household consumption and construction heled by massive inflows o f workers' remittances. Most o f this increased demand has been met by foreign production and, as a result, the external accounts have widened progressively. In order to put growth and poverty reduction on a sounder footing, a stronger and more robust domestic supply response will required. This will necessitate further improvements in the domestic investment climate in order to increase heretofore low domestic private investment and innovation. As with private investment, public investment has been very low with the result that public infrastructure has continued to deteriorate with negative implications for growth. 6. Economic management has been complicated by external price shocks - energy prices have increased rapidly with natural gas prices doubling since the beginning o f 2006. Inaddition to the negative impact on economic growth and the external accounts, these shocks have put pressures on the fiscal accounts. It has directly raised the cost o f electricity, natural gas and oil products procuredby the Government at all levels. Further, to prevent the reemergence o f the large quasi- fiscal deficits o f the 1990s, higher energy prices have been largely passed on to final consumers. This has created rising pressures for additional subsidies and transfers that complicate the management o f fiscal policy. From a policy perspective, to prevent monetary and exchange rate policy from bearing an excessive burden in achieving the needed macroeconomic adjustment, some support from fiscal policy will be required. 111. Fiscal Retrenchment andRecovery borrowing and domestic payment arrears. As a result, public external 3 50 ~ I debt increased sharply raising 40 concerns about debt sustainability. I 30 - 8. The 1998 regional financial 2o crisis forced the Government to IO -mi I activity, government revenues -10 I dropped sharply. With limited -20 1 i domestic and external source o f deficit financing, the regional crisis 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 source: MTEF and IMF Article IV Staffreport necessitated a sharp drop in expenditures. Most o f this fiscal retrenchment resulted in deep cuts in health, education and social protection spending. This forced fiscal adjustment resulted in a sizable improvement in the overall fiscal (cash) deficit o f the general government - moving from an average deficit o f 8.4 percent in 1996-1999 to 3.5 percent in 2000 thereby setting the stage for the markedly improved economic and fiscal performance experienced by during the recovery that began in2001, .. 11 9. The economic recovery, which began in earnest in 2001, has resulted in a significant upsurge infiscal revenues. By 2005, total revenues had recovered almost all o f the lost ground at the end o f the 199Os, growing to nearly 40 percent o f GDP. The expansion in fiscal revenues has given the Government some room to increase expenditures, particularly social spending. Further fiscal space has been gained by rapid reduction inpublic and publicly guaranteed debt as a share o f GDP as a result o f growth, exchange rate appreciation and better debt management. With access to budget financing remaining limited, the increase in expenditures has been constrained by revenue mobilization. With taxes very buoyant, starting in 2003, the fiscal accounts of the general government have moved into surplus. While government has maintained sound macroeconomic policies, the quality and equity in access to public services and the efficiency of resource utilization has remainedproblematic. Table i: Revenues and Expenditures inMoldova,2001-2005 (percent of GDP) 2001 2002 2003 2004 2005 Total Rewnues andGrants 29.2 29.6 34.0 35.4 39.5 Tax revenues 24.4 25.8 27.5 29.8 32.3 Direct Taxes 3.7 4.0 4.3 4.9 4.5 Indirect Taxes 13.9 14.6 16.0 16.1 18.5 ofwhich: VAT 7.9 9.0 10.1 10.7 12.6 Social and Health Fund Contributions 6.8 7.3 7.2 8.8 9.3 Non-taxrevenues 3.9 3.1 2.6 1.9 2.5 Grants 0.9 0.6 0.3 0.4 1.3 Other Revenues ... ... 3.6 3.2 3.4 Total Expenditures and NetLending 29.4 31.5 33.3 34.6 37.9 National Economy 1.7 1.8 3.6 4.2 5.3 Social sphere 16.6 19.5 20.9 22.0 24.0 Education 4.8 5.5 6.7 6.8 7.3 Social Fund expenditures 7.2 8.4 8.2 8.6 10.0 Health & Health Fund expenditures 2.8 3.5 3.9 4.2 4.3 Other social expenditures 1.7 2.0 2.2 2.4 2.3 Interest p a y m n t s 4.2 2.2 2.1 1.9 1.3 Other Expenditures 7.1 8.3 6.7 6.5 7.3 Balance o f Special Accounts and statistical discrepancy -0.1 0.6 -0.4 0.1 0.1 Owrall Fiscal Balance (cash) -0.4 -1.4 0.2 0.8 1.7 Primary fiscal Balance (cash) 3.8 0.7 2.3 2.7 3.0 M e m : Public Debt (YOo f GDP) 77.0 61.1 52.5 35.3 28.2 Sources: MTEF 2007-2009 and IMF, StaffReport for 2006 Article IVConsultations. IV. Key ChallengesConfrontingPublicExpenditurePolicy 10. Sustaining economic growth and poverty reduction will require a set o f policies that will encourage stronger private sector investment and innovation as well as better and more efficient use o f resources by the public sector. The key fiscal challenge will be finding the fiscal space to increase critical physical and human public investment without placing a greater burden on the private sector. This PER concentrates on some key reforms that are needed to achieve this through improving the efficiency o f public spending inMoldova. ... 111 Reducingthe Size of Government 0 Reduce size of Government (as a share of GDP) to avoid undermining private sector's incentives to save and invest. 0 Within a smaller fiscal envelope, maintain a prudent fiscal policy stance that also supports monetary and exchangeratepolicies. 11. As the experience in Moldova during the 1990s has demonstrated, the inability to keep government expenditures in line with the available resource envelope can contribute to macroeconomic volatility and instability. Large fiscal deficits, by increasing policy and business uncertainty and crowding out the private sector, impedes a country's ability to attract the private sector investment and innovation so critical for economic growth. Moldova's adherence to prudent macroeconomic policies since 2001 has been an important contributor to restoration o f economic stability and the resumption o f growth. The stance o f fiscal policy will need to remain tight inorder to support monetary and exchange rate policies. 12. While the fiscal deficit has been kept under control, the level o f expenditures and revenues have grown rapidly and are now too high given Moldova's income level and institutional development. It i s estimated that expenditures and revenues are 7 to 8 percentage points o f GDP higher than one would expect o f a country o f similar income levels. This suggests that the creation o f fiscal space by further increases in revenue mobilization i s not a desirable option. Instead, Moldova will need to focus its reform efforts on increasing the provision o f quality public goods and services by improving the efficiency o f tax revenues and expenditures. in Figure ii, in countries where 50 - public institutions are weaker, a I category that currently includes 0 25 - I O Moldova, the size of public 9 0 0 expenditures has a decidedly o - negative impact on growth. The *1 international empirical evidence 1 - 2 5 . I suggests that in countries where expenditures levels exceed 35 II percent o f GDP, an increase in -50 20 40 60 80 iv Figureiii: Shrinkingthe Size of Government 14. Reducing the size o f Government (percent of GDP) will require constraining the increase of expenditure relative to GDP growth. As shown in Figure iii,continued rapid 42 - increases in public consumption, as 40 - MTEF expenditureplan , ' .-..-..-.- . proposed by the MTEF expenditure plan, I implies that the size of government will 38 - continue to grow over the medium term. 36 - If, however, the growth in public ExpenditureGrowth l i m i t e d 1 expenditures were constrained to be less 34 // to Inflation + 2 percent than GDP growth, say 2 percent in real terms over the mediumterm, thenthe size 32 o f Government would decline over the 30 4 medium term and approach the threshold 2003 2004 2005 2006 2007 2008 2009 Sources: MTEFand World Bank Staff Estimates level o f 35 percent indicated above. ReorientingExDenditures towards Growth Promotion Reducepublic consumption spending and increase emphasis on growth-enhancingpublic investment in human capital and infrastructure. Introduce evidence-basedsetting of strategicpriorities in the investment budget. Project screening, selection, and appraisal should draw from appropriate cost-benefit analysis, with an initial emphasis on larger investmentprojects. 15. From the perspective o f expenditure allocation, the consumption orientation of the public spending has been crowding out growth-promoting public investment expenditures (Le. infrastructure and quality improvements inhealth and education). For a given level of spending, changing the composition o f overall expenditures in favor public investment will improve long- term growth prospects. In Moldova, the increase in public expenditures since 2001 has been not only increased as a share o f GDP, it has consumed a growing share of the budget - from 58 dominated by the increases in consumption spending on the social sectors - social spending has percent of the budgetin2001to nearly 63 percent in2005. 16. Inaddition to large transfers to households inthe form of pensionand social assistance spending, a significant portion of current expenditures i s devoted to staffing costs (i.e. wages broadly defined). The wage bill of public sector employees, including those in the MHIF, has grown sharply since 2003 and i s now close to 11 percent o f GDP or nearly 30 percent of total expenditures. The downward stickiness of nominal wages in the public sector and the lack o f instruments for rationally downsizing the public payrolls make it difficult to adjust the wage bill during economic downturns, undermining the flexibility of fiscal policy. In the absence of reforms to increase the productivity of public workers, the growth in public sector wages risks crowding out government operational and capital expenditures. Any increase inreal public sector wages will need to be offset by reductions in public employment so as to keep the public sector wage bill under control and avoid crowding out neededcapital spending. 17. For upgrading infrastructure and sustaining economic growth, Moldova needs to invest more. Deficiencies in public infrastructure represent a considerable constraint to economic growth inMoldova. Despite the demand for upgrading and expanding the public infrastructure in V Moldova, public capital spending has remained quite low in recent years-only about 2 percentage points o f GDP. The limited availability o f public resources highlightsthe importance o f strengthening the management o f public capital spending in order to ensure that the available resources are utilized effectively and efficiently. 18. Because public resources are limited, there i s a need to prioritize and plan government investment strategically. Over the last few years, the Government has significantly improved the budget preparation process and the strategic allocation of resources, mainly through the introduction o f the Medium Term Expenditure Framework (MTEF). The Government has also taken steps to gradually integrate all public resources into the MTEFand the annual state budget. Despite these achievements, a number of weaknesses remain. The portfolio of investments being implemented remains too large compared to the available financing. In addition, investments are being undertaken without adequate consideration being given to the requirements for financing o f downstream recurrent costs. For further rationalization o f the public investment portfolio, a number o f improvements are required in the identification, prioritization, preparation and appraisal o f projects. ReformingPublic SectorRemunerationSvstem e Implement planned wage increases within the context of broader public administration reform. Implementing a transparent and uniform remuneration system and consolidate multiple salary supplements into the basepay. 19. Public sector salaries consume close to a third o f total government resources. In2005, the wage bill accounted for some 29 percent o f total government spending, representing about 10.8 percent o f the GDP. Across sectors, there are some variations inthe share o f wages and salaries in total spending: for example, as a share o f total expenditures, spending on personnel i s much higher in the education and health sectors (over 50 percent o f total spending in those sectors). Moldova's public sector wage bill i s already among the highest compared to other countries in the region, including several advanced OECD economies. The wage bill, however, i s programmedto continue rising through 2008 to keep pace, inpart, with rising overall wage levels inthe economy. Inrecent years, the stable level ofpublic employment and risinglevel o f public wages has made it difficult to increase other essential spending, especially operation and maintenance expenditures. As public sector wages are projected to rise and outpace economic growth in the short-term, the rising wage bill will likely create significant fiscal and macroeconomic pressures inthe process. 20. Because the design o f the public sector wage system also has significant implications on the quality and efficient function o f government, the government has on its hands the delicate task o f striking the right balance between a public wage bill that i s fiscally affordable and a remuneration system that i s able to attract qualified and motivated workers to the public sector. Despite the recent, rapid increase inpublic sector wages, compensation levels inthe public sector are still very low when compared with the private sector. While civil servants are underpaid compared to the private sector at all levels, the relative under compensation o f middle and senior management is quite extreme. At the lower end o f senior management, for example, the private sector compensation i s more than six times the civil servant salary. In addition to low pay, the compression o f salaries inthe civil service remains quite high.As a result o f l o w and compressed compensation levels, the public sector in Moldova has had a very difficult time in attracting and retaining skilled employees. vi 21. Inadditionto lowremunerationlevels, the compositionofpublic sector compensations is very complex and lacking intransparency. Beyond the base salary level, public sector employees receive a number of additional payments including: a reward for taking on tasks of "major" importance; supplements for additional work resulting from the temporary absence o f a coworker; a length of service bonus; and a bonus for knowledge of additional languages. These bonus and allowance payments can be considerable when compared to the base salary. By way o f comparison, in other countries this type of compensation usually only amounts to 10-20 percent of the base salary. Although the system can provide higher compensation to qualified employees, it is completely lacking in transparency. This creates incentive problems (combined, to begin with, with low public salaries). 22. Increasing wage dispersion (or pay decompression) and consolidating multiple salary supplements into the base pay while keeping constant the overall wage bill, will be the key elements of a good reform program. The ongoing, government-led sectoral review, under a broader central public administration reform program, will be critical in identifyingopportunities for reducing non-essential staffing, streamlining government processes, and promoting public sector productivity RestorinPthe TransportNetworkfor SustainedEconomicGrowth Reform the existing road maintenance financing mechanism by considering the introduction of a second generation road maintenancefund to ensure a stable source of financingfor the transport sector. 0 Introduce competition in road maintenance by revising the contracting modalities for road maintenance worhx 23. The lack of road maintenance and rehabilitation inthe past 15 years has led to massive physical deterioration and to heavy loss of rQadnetwork asset value. The present asset value of the Moldovan road network is only about US$ 8.4 billion, far short of the US$ 12 billion it would have been, had the existingroadnetwork been adequately maintained. This means that the loss of road network asset value which resulted from insufficient maintenance and rehabilitation i s a shocking US$ 3.6 billion, equivalent to 1.4 times the entire GDP of Moldova in2004. 24. Despite the considerable loss of asset value, the road network i s still Moldova's single most important national asset, certainly much larger in value than all power plants, railway lines and the electricity grid combined. An urgent and important effort i s needed to save what i s left and to gradually improve the road network, without which sustained economic and social development will not be possible. Two important institutional reforms appear necessary: (i) the reformo f the existing road maintenance financing mechanism, and (ii) contracting modalities the for road maintenance works. There are several options for creating a stable funding mechanism for road maintenance in Moldova. There are, for example, good arguments for separating the financing o f road maintenance (such as through earmarked resources) from those of new road investments(through the general budget), although experience from other countries suggest that a number of supporting conditions need to be in place to enable this system to work. Meanwhile, the present system of awarding and supervisingroad maintenance contracts suffer a number of weakness, although less apparent in recent years given the very low level o f resources for road maintenance. Inparticular, the present system o f sole-source contracts with the state-controlled local district road maintenance firms has several deficiencies which may result in a lower efficiency in the use of resources. As transport spending i s projected to rise, these institutional features will become more critical inthe efficient management of transport resources. vii Increasingthe Efficiencvand Oualitv of HealthCare Ensure sustainable and equitablefinancing of health care by increasing collection rates and expanding insurance coverage in rural areas. 0 Promote quality and user satisfaction in health care byfurther consolidatingservices and introducing efficiency oriented measures at theprovider level. 25. Moldova has made significant progress in reforming the health care system - health insurance has been introduced, hospitals have been transformed from budgetary institutions to self-governing entities, and primary health care has been strengthened. Since the end o f the 1990s, expenditures on health have increased dramatically - from about 2.8 percent o f GDP to nearly 4.3 percent o f GDP in 2005. Inper capita terms, spending on health i s now at the highest levels experienced since independence. Nearly half o f these expenditures are on wages, with spending on drugs and equipment still remaining quite low. Further, a substantial portion o f the expenditures o f the National Health Insurance Company (75% in 2005) i s covered by transfers from the state budget. The sustainability o f the system i s further undermined by the weak participation inthe system by Moldova's most productive population - less than 50% o f the 25 to 44 age group participates inthe system. 26. All these factors helpedreverse the unfavorable trends experienced inthe 1990sand, as a result, a broad range o f health indicators have improved. While mortality and morbidity rates are still far from the EU average, Moldova i s making progress towards meeting MDG indicators on health. The introduction o f health insurance has also increased the financial protection o f the population with nearly 75 percent o f the population covered by insurance. Some gaps in insurance coverage remain - coverage in rural areas and small towns are still below those in Chisinau. Formal and informal out-of-pocket expenditures - mainly allocated to purchase o f drugs - still represent a large share o f overall health expenditures (41 percent in 2005). Expanding the drug benefit package to reduce these out o f pocket expenditures will require, among other things, increasing collectionrates. 27. Despite the achievements in this sector, the hospital sector continues to be burdened by overcapacity. About 40 percent o f inpatient health care facilities are outside the control o f the Ministry o f Health and Social Protection and, hence, have not been subject to recent reform effort. Poor performance i s observed in a number o f key productivity (i.e. discharges per physician) and cost-containment indicators. While hospital efficiency has improved, these indicators are still among the lowest in Europe. Efficiency gains can be achieved by improvements in utility management increases in ambulatory surgery, outsourcing o f services (such as laundry) and further hospital consolidation. While public health expenditures have increased, hospital activity has decreased. This suggests the need to control expenses through improved resource allocation mechanisms and to emphasize the quality rather than quantity o f spending. Clinical practices will need to be modified to promote more cost-effective care and quality based initiatives. ... V l l l Enhancingthe Effciencvof ResourceUse inEducation 0 Consolidate school networks to promote the more eficient use of resources in the education sector. Such a program could create fiscal savings of up to 0.5 percent of GDP. 0 Consider the best use offiscal savings: invest in quality-enhancing inputs within the education sector, or reallocate resources to another sector. 28. With expenditures on education over 7 percent of GDP - levels broadly comparable to other countries in the ECA (Europe and Central Asia) region - issues related to the allocation o f resources within the sector as well as those of technical efficiency are dominant. Because salary levels in education are among the lowest in the economy, the Government has emphasized the need to pay higher wages to attract better skilled teachers, with the goal of improving the quality o f education. However, funds allocated to schools are on the basis of norms established by the Ministries o f Finance and Education. As such, the flexibility of local governments and school managers inusingthe resources i s severely limited. Indeed, higher wages will only be achievable by increasing the efficiency of spending inthe sector. 29. A number of areas offer opportunities for achieving increased efficiency. First, many schools have very low numbers of students, small classes and low studentlteacher ratios. These figures are low even by international, OECD standards. Moreover, there are wide variations between schools within and across judets. If studendteacher and studentlnon-teacher ratios were reduced to OECD levels, by closing down or relocating secondary classes from schools that are below an efficient size (while allowing pre-schools and primary classes to remain where they are, with multi-grade teaching if necessary) and the abolition of strict norms for the number of teaching and non-teaching staff in autonomous schools, fiscal savings totaling 277 million lei would result. Some offsetting increases in expenditure would of course be necessary to finance separation packages, retraining, and transport costs. The net savings, amounting to about [ 1, could be applied to salary increases and quality improvement in the education sector or transferred to other sectors. Second, the extent to which decentralization o f authority to autonomous schools will help address the allocation of inefficient education expenditures will be explored. The proper sequencing of reforms i s of course an issue, to create the right incentives for merging schools while decentralizing the authority over school budgets. Promotinvthe Fiscal Sustainabilitv of Pensions 0 The Government strategy to unijj the contributory system should increase revenues over time. New regulations should be introduced gradually in the case of farmers, because many arepoor and vulnerable; they can be implemented more quickly in the case of self- employed workers and wage employed workers in the agricultural sector. Develop and strengthen the NSIH reserves management capacity and monitoring institutional agencies. 30. Over the last four years, the National Social Insurance House (NSIH) has enjoyed a favorable economic and demographic environment. Although the NSIH experienced a deficit at the end of 2005, the pension system i s currently on sounder financial footing with accumulated reserves o f about 1.6 percent of GDP. Simulations also suggest that low pension benefits and favorable demographics trends should effectively allow N S M to keep producing surplus and ix accumulating reserves in the medium-term. However, as the dependency ratio deteriorates, the long-term financial viability o f the system will depend on how the reserves are managed and secured for the next generation. Setting up the regulations and supervisory institutions for this purpose i s critical, given that moral hazard and agency problems typically lead to situations where funds mismanaged. While there are no obvious solutions to these problems, it i s usually advisable convene a board o f trustees that represents diverse groups and diverse interests. The most successful experiences are associated with pension fundtripartite boards. 31. The pension system currently faces two additional challenges: First, pension incomes at retirement are very low incomparisonto current wages. Second, contributors' participation inthe pension system i s largely subsidized. Inthe first case, the strategy to upgrade the system has not yet been presented. In the second case, the government has developed a strategy towards the unification o f the contributory system and the reduction o f subsidies. Because poverty rates are extremely high among farmers, a gradual phasing in o f new regulations should be considered. Self-employed workers, more than three quarters o f whom are subsidized, have the lowest poverty rate. Because leakages are likely in this group, a more rapid enforcement of new regulations, including the payment o f the desired minimumcontribution, i s desirable. MaximizinPthe Benefitsfrom SocialAssistance 0 Promote the targeting eficiency of social assistance. Simulations suggest as much as a quarter of social benefits or 0.3percent of GDP accrue to the richest households. 0 Create the legal, administrative, and institutional bases for transitioning the social benefits systems into a means testedprogram. 32. By international standards, total government spending on social assistance benefits in Moldova i s about average. In 2004 such spending accounted for about 1.7 percent o f GDP and 4.8 percent o f total government expenditure. While spending in this sector i s not particularly large, there are significant inefficiencies in this expenditure area when viewed from the perspective o f poverty reduction. Social assistance benefits are, in general, not an importance source o f income for households and, among poor households, such benefits are only marginally important. Almost all spending i s based on categorical classifications, with the disabled, the war veterans and families with children constituting the largest groups o f beneficiaries. The poor as a separate category appear only once across various types o f benefits (under "material assistance') and the share o f beneficiaries i s very small. There i s little means testing in the allocation o f benefits: while the system i s meant to support the poor and vulnerable, in practice it supports selected categories o f beneficiaries, irrespective o f income. As a result, the benefits system i s quite regressive, although there has been some improvement in recent years. In2004, data from the Household budget survey indicates that the poorest households received only about 8 percent of all benefits while the richesthouseholds received24 percent. 33. The ongoing government reform program-which includes the implementation o f a pilot program for means-testing-represents a good platform for improving the targeting efficiency o f the social benefits system. However, a number o f accompanying steps and policy interventions will be critical. These include the following: the revision o f the legal framework underlying a means-tested social benefits systems; clarifying the sequencing o f reform and the categories o f benefits that are to be phased out as the systems transitions into a means-tested program; creating an administrative division between the social insurance and social assistance functions o f the present system; and enhancing the institutional capacity to manage, monitor, and evaluate the state provision o f social benefits. A Conclusion 34. The increase in fiscal revenues that accompanied Moldova's economic recovery has given it substantial room to increase Government expenditures. Despite devoting nearly 40 percent o f GDP to public expenditures, and even accounting for the prospect o f increased external grant assistance, there are a still a host o f unmet demands on the budget. Roads are in dreadful shape, higher natural gas prices are increasing the demand on social safety nets, public sector wages and pensions are still very low and poverty - particularlyrural poverty - i s high. Almost all o f fiscal space achieved during the recovery, however, has gone to increase public consumption - particularly, public sector wages, transfers and subsidies - expenditure items that have been severely eroding inthe transition process. As a result, critical infrastructure needs are largely unmet and state assets, such as roads, hospitals and schools, continue to deteriorate at an alarming rate. 35. Moldova must confront the reality that the size o f Government has grown too large. The tax burdenneeded to support highlevels o f government consumption spending negatively affects private saving and investment decisions and may limit Moldova's future growth prospects. This burden i s compounded by the fact that the public sector's contribution to investment i s critically low. As long as public expenditures are increased at a slower rate that GDP, the size and burden o f government will gradually decline. 36. IfMoldovais to accelerate growth andpovertyreduction, itmustcurtail excessive public consumption spending and increase spending on public investment while at the same time reducing the overall footprint o f Government. Cutting wasteful spending in the education and health sectors, for example, frees resources for more investment spending in those sectors. Increased spending on roads will put a halt to the deterioration o f this vital infrastructure, and support increased private sector activity and greater international competitiveness. Improving the targeting o f social assistance allows these programs to more effectively protect vulnerable groups by freeing up these misdirectedresources. Completingthe reforms to the pension system reduces the accumulation o f future unfunded liabilities and puts public finances on a sounder footing. The reform to public financial management currently underway provides an opportunity to strengthen the linkages between spending prioritization and budget management. This process of fiscal adjustment i s essential to put economic growth on a solid a sustainable trajectory. xi 1. CREATINGFISCALSPACE FORECONOMIC GROWTHAND POVERTY REDUCTION A. INTRODUCTION 1.1 The Government o f Moldova's Economic Growth and Poverty Reduction Strategy Paper (EGPRSP) lays out an ambitious plan for sustaining growth and poverty reduction and reshaping the government to meet the needs o f a market economy. The public expenditure program envisaged by the EGPRSP, however, vastly exceeds the domestic resources available to the Government.' Foreign budgetary support will help alleviate some of that resource constraint. Recognizing that the share o f tax revenues and expenditures to GDP in Moldova already greatly exceed international norms, generating additional domestic tax resources risks crowding out the private sector and undermining growth prospects. This suggests that in order to finance higher order public expenditures priorities, the Government needs to create fiscal space from within the existing resource envelope. This will require inter and intra-sectoral reallocation o f expenditures and an increase in the efficiency o f public spending rather than increasing the relative size o f government. 1.2 The ability o f Government to increase expenditures or reduce taxation without endangering the financial sustainability o f the public sector i s limited. Indeed, from Moldova's experience during the 199Os, the negative impact o f large fiscal deficits on macroeconomic stability i s well known. Even if,however, fiscal discipline i s maintained, disproportionately high levels o f government expenditures necessitate a high degree o f taxation and, hence, may limit Moldova's future growth prospects. If the Government is to more effectively utilize public resources to enhance economic growth and poverty reduction prospects, it must re-orient the budget towards growth-enhancing expenditures (i.e., public investment). In order to undertaken this re-orientation in a fiscally sustainable manner, increases inpublic consumption (i.e., public sector wages, transfers and subsidies) will need to be controlled and expenditure efficiency will need to be enhanced. 1.3 Fiscal space for increasing public investment spending can be generated in a number o f manners. Changes intax policy and administration to broaden the tax base and increase revenue potential without distorting or undermining the private sector incentives to invest and save i s one option. However, this option i s constrained by the need to reduce Moldova's already high tax burden. This suggests that the burden o f creating this fiscal space has to be on the expenditures side where fiscal space can be created by shifting resources from less efficient to more efficient uses. It should be noted that there is also a strong inter-temporal dimension to fiscal space. The effective use o f public resources to improve human and physical capital will lead to increased productivity and income and, hence, expand the scope for private and public consumption opportunities inthe future. ' The Government's Medium Term Expenditure Framework covers only 30 percent of the expenditures included inthe EGPRSP. Since the EGPRSP does not include cost estimates for a large number of investment projects, such as rehabilitatingthe road and rail network, this budgetary shortfall is likely even greater. 1.4 This Public ExpenditureReview (PER) concentrates on some key reforms that are needed to improve the functioning and efficiency o f public spending. This overview chapter presents a synopsis o f the emerging macroeconomic challenges and summary o f recent developments with the Government's expenditure framework. One o f the key fiscal challenges will be finding the fiscal space to increase capital expenditures. Increasing public investment without improving capital budgeting, however, will result inthe squandering o f scarce resources. Chapter 2 explores how a better capital budgeting process would, in and o f itself, create fiscal space since improvements in the planning, evaluation and allocation o f the capital budget would channel more resources towards more productive investments. Critical to Moldova's growth prospects i s increasing public investment in the maintenance and rehabilitation o f the road infrastructure network. Chapter 3 details the negative implications o f consumption orientation o f the budget on the quality o f the road network in Moldova and makes recommendations to improve financing mechanism and contracting modalities that will enable an increase in domestic funding o f road works. Inan effort to increase the capacity o f Government to conduct economic policy, public sector wages have been increased inthe hopes o f attracting and retaining skilled workers. Chapter 4 examines the growing pressures this has placed on the budget and explores the need to reform the civil service pay system to improve incentives in public administration. Chapters 5 and 6 explore how fiscal space for human capital investment can be achieved through the more efficient use o f resources within the health and education sectors. With poverty trends stagnating and the pressures on the budget to address the impact o f higher energy prices growing, Chapter 7 examines the long-run sustainability o f the pension system to determine the extent to which it poses a risk to future fiscal space. Finally, Chapter 8 addresses the significant inefficiencies in the Government's social assistance programs. B. RECENT MACROECONOMIC DEVELOPMENTS 1.5 Since 2000, Moldova's strong growth performance has reversed a decade o f economic decline and rising poverty. The economy has grown by nearly 40 percent since 2000 with average real growth of around 7 percent per annum. This economic recovery has ledto a massive movement out o f income poverty. Between 1999 and 2003, over 37 percent o f the population moved out o f poverty, the largest absolute decline in poverty in the Europe and Central Asia region over this period. The decline in absolute poverty was broad-based and poverty fell across all locations. Inequality, as measured by the Gini coefficient, has been on a slight downward trend since the late1990s. Starting in 2004, however, poverty reduction has stalled despite continued vigorous economic growth. This latest trend in poverty, the increase in poverty rates seen in 2005, heightens concerns that rapid economic growth i s no longer reducingpoverty. 1.6 The major source o f growth in Moldova has been domestic household consumption and construction fueled by massive inflows o f workers' remittances - officially recorded at nearly 30 percent o f GDP in 2005. While growing slowly as a share o f GDP, private domestic investment has averaged only 17 percent o f GDP since the beginning o f recovery. During the economic recovery, the increase in domestic output has relied more on increased capacity utilization rather than significant additions to capacity. While remittance inflows are likely to continue to support strong consumption growth inthe short run, over the medium-term Moldova will need to generate a stronger and more robust domestic supply response. Indeed, as the opportunities reaped from labor shedding and increased capacity utilization are exhausted, future gains in productivity will require increasing domestic investment and innovation. Similarly, public investment has remainedvery low with the result that key public infrastructure, particularly roads, has continued to deteriorate with negative implications for growth. Findingthe fiscal space for increases in 2 growth-enhancing public investment will be an important part o f the strategy to change Moldova's underlying growth paradigm. Table 1.1: Moldova: Macroeconomic Framework, 2001-2008 Actual IEstimatedI Projected 2001 2002 2003 2004 I 2005 I2006 2007 2008 RealGDP growth (%) 6.1 7.8 6.6 7.4 7.1 3.0 3.0 5.0 CPI Inflation (average, %) 9.8 5.3 11.7 12.5 11.9 12.3 10.5 8.9 Exchangerate (average, MDU$) 12.9 13.6 13.9 12.3 12.6 Real effectiveexchangerate, % -1.2 -6.1 -5.4 12.8 -1.3 1.8 -6.5 -2.9 Revenues and grants, % of GDP 29.2 29.6 34.0 35.4 39.5 40.3 42.4 41 Expendituresand net lending, % of GDP 29.4 31.5 33.3 34.6 37.9 40.3 44.4 43.2 Primary balance(cash), % of GDP 3.8 0.7 2.3 2.7 3.0 1.3 1.0 0.7 Overallbalance(cash), % of GDP -0.4 -1.4 0.2 0.8 1.7 0.0 -0.5 -0.5 Currentaccountbalance, % of GDP -2.5 -4.6 -7.1 -2.0 -8.3 -10.4 -7.2 -4.2 Exports G&S growth (percent) 14.7 19.1 20.9 27.4 13.3 2.1 14.6 12.6 Imports G&S growth (percent) 12.0 18.9 33.5 22.8 29.1 18.1 11.1 9.0 Gross official reserves(months ofimports) 2.1 1.9 1.7 2.1 2.2 2.3 2.4 2.7 Total Externaldebt/GDP, % 105.1 100.8 88.7 63.8 54.7 51.6 55.1 53.7 of which, public andpublicly guaranteed (PPG) 77.0 61.1 52.5 35.6 28.7 24.1 24.8 23.1 ExternalArrears (millions $US) 17.5 42.9 86.1 50.6 56.3 0.0 0.0 0.0 PPGdebt serviceto exports of GNFS, % 24.6 27.9 19.8 21.0 20.3 16.7 14.4 14.1 PPGdebt serviceto revenues, % 59.5 52.8 34.7 25.2 17.7 14.9 13.7 12.5 Sources: Moldovanauthorities (MTEF), IMF Article IV StaffReport, December2006. Fiscalaccounts inthe table are for thoseof the generalgovernment. 1.7 While strong economic performance in Moldova's main trading partners has encouraged a rapid increase in exports (averaging 19 percent over 2001-2009, the remittance-fueled increase in consumption has pushed up import growth even faster (averaging 23 percent over the same period). As a result, the trade deficit has continued to worsen, exceeding 40 percent o f GDP in 2005. With large inflows o f workers' remittances (31 percent o f GDP in 2005), the current account deficit has been more modest (averaging slightly over 4 percent o f GDP since 2000). Over the longer run, with underlying pressures on the real exchange rate to appreciate, the increased international competitiveness needed to close this large trade gap will require an increase in domestic productivity. This will require significantly more domestic private and public investment, as well as foreign direct investment, than i s currently undertaken. 1.8 Monetary policy has strived to achieve a number o f objectives, including maintaining a competitive exchange rate, keeping inflation low and accumulating a precautionary level o f international reserves. The large inflow o f remittances has contributed to appreciation pressure on the real and nominal exchange rate underscoring the need for more rapid improvement in domestic productivity. The National Bank o f Moldova (NBM) has actively intervened in the the monetary impact o f these interventions - and the inflationary pressures - the NBMincreased foreign exchange market and has been rapidly accumulating foreign exchange reserves. T o rein in its sterilization efforts and slowed reserve money growth. 1.9 Economic management has been complicated by recent increases in natural gas prices. Since Moldova imports nearly all o f its energy needs, the recent increases in energy prices have had a severe impact on Moldova. For most o f the recovery period, Moldova paid significantly 3 lower prices for natural gas than the rest o f Europe - $80 vs. $235/thousand cubic meters (mcm). - by mid-2006 - with further increases inprices a possibility. The impact o fthis increase has been In2006, Russia substantially increasedthe price ofnaturalgas paidbyMoldova to $160/mcm to put additional pressures on the underlying trade and fiscal deficits. It has directly raised the cost o f electricity, natural gas and oil products procured by the Government at all levels adding pressures to the budget. Further, to prevent the reemergence o f the large quasi-fiscal deficits o f the 199Os, the Government has largely passed these higher energy prices on to final consumers. This has created rising pressures for additional subsidies and transfers that complicate the management o f fiscal policy. Given that these energy price shocks have a high degree of permanence, the economy will need to adjust to the new reality o f higher natural gas prices. From a policy perspective, preventing monetary and exchange rate policy from bearing an excessive burden o f achieving the needed macroeconomic adjustment, some support from fiscal policy will be required. 1.10 The economic recovery has also greatly improved Moldova's external debt outlook. Combined with the impact o f U S dollar appreciation, net repayment to creditors and favorable rescheduling operations, the ratio o f total external debt to GDP fell from 106 percent o f GDP in 2001 to around 55 percent in 2005. At the end o f 2005, Public and Publicly Guaranteed (PPG) debt amounted 28 percent o f GDP (down from 65 percent o f GDP in 2001). Debt to multilateral creditors, primarily the World Bank (International Bank for Reconstruction and Development (IBRD) and International Development Association (IDA)) and IMF (International Monetary Fund)amounted to nearly two-thirds o ftotal PPGdebt. Debt to bilateral creditors, primarily Paris Clubcreditors (and chiefly Russia) comprised the remainder o f the total. Compared to 2001, debt to commercial creditors has declined to very small levels (less than 4 percent o f total debt). In Net Present Value (NPV), PPG external debt i s around 24 percent o f GDP or 50 percent of exports. The NPV o f PPG debt to revenues ratio i s around 62 percent (88 percent including domestic currency public debt) while the ratio o f debt service to revenues i s around 9 percent. Consistent with the authorities' medium-term fiscal program, supported by the IMF's PRGF, a further decline inthe debt-to-GDP ratio i s projected, with the NPV o f public sector debt-to-GDP ratio declining to 10 percent by 2025. It i s largely as a result o f the fiscal adjustments achieved since 2001 that Moldova has overcome what was previously thought to be a structural debt sustainability problem. c. OVERALL FISCAL PERFORMANCE was mixed. As was highlighted in the previous public expenditure review - World Bank (2003) 1.11 Fiscal Deterioration during the Transition. Up to 1999, fiscal performance inMoldova Moldova: Public Economic Management Review (PEMR) - poor economic performance combined with weak budget processes, low tax enforcement capacity and poor governance resulted in severe and unsustainable fiscal imbalances during the 1990s. While on a declining trend for most o f this period (see Figure 1.1), fiscal revenues and grants ranged between 35 and 40 percent o f GDP. However, fiscal expenditures, at between 45 and 50 percent o f GDP, were much higher. The resulting fiscal deficits, which averaged more than 6 percent o f poor economic performance combined with weak budget processes, low tax enforcement capacity and poor governance resulted in severe and unsustainable fiscal imbalances during the 1990s. While on a declining trend for most o f this period (see Figure 1.1), fiscal revenues and grants ranged between 35 and 40 percent o f GDP. However, fiscal expenditures, at between 45 and 50 percent o f GDP, were muchhigher. The resulting fiscal deficits, which averaged more than 6 percent o f GDP, were financed through a combination o f external financing, domestic borrowing and domestic payment arrears - mostly on social expenditures, including pensions. 4 With only a relatively small portion o f external debt contracted on Figure1.1: Expenditures,Revenuesand FiscalBalance concessional terms, public and publicly guaranteed (PPG) external 50 _- debt rose sharply from almost zero in the early 1990s to peak at 81 percent 4 0 - . .----. o f GDP in 2000 making Moldova one 3 0 - Fiscal Revenues o f the most heavily indebted countries in the region and raising concerns II 2 0 - ~ about debt sustainability. 10 - j 1.12 The 1998 regional financial - crisis forced the Government to m ' I ' i ' I ' I ' = ' I I' I tighten fiscal policy dramatically. - OverallCashBalance II With the negative shock to economic I ~ activity, government revenues -20 - dropped sharply. As can be seen in 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Figure 1.1, government revenues fell Source: MTEF and IMF Staff Report 1.13 The fiscal retrenchment necessitatedby the regional crisis resulted in a sharp drop in expenditures. Between 1997 and 1999, expenditures dropped by more than 14 percentage points o f GDP. With spending on health, education and social protection comprising the largest items in the budget, spending in these sectors fell steeply, both as a share o f GDP and in real terms. This "forced" fiscal adjustment resulted in a sizable improvement in the overall fiscal (cash) deficit o f the general government - moving from an average deficit o f 8.4 percent in 1996-1999 to 3.5 percent in 2000 thereby setting the stage for the markedly improved economic and fiscal performance experienced duringthe economy recovery. 1.14 Economic Recovery and the Creation of Fiscal Space. The economic recovery, which began inearnest in2001, resulted ina significant upsurge in fiscal revenues and a reduction inthe external debt servicing burden. By 2005, total revenues had recovered almost all o f the lost ground at the end o f the 199Os, growing to nearly 40 percent o f GDP. Interest and amortization payments, which had consumed nearly 30 percent o f the general government budget in 2001, were cut sharp as a result o f limited external borrowing, rapid economic growth, exchange rate appreciation and improved debt management. With access to budget financing remaining limited, these developments gave the Government the significant fiscal space to increase public expenditures. Total expenditures have increased by nearly 9 percentage points o f GDP since 2001. With the brunt o f the fiscal retrenchment falling on social programs, however, the government placed a heavy emphasis on increasing social spending during the economic recovery. As a result, to the detriment o f Moldova's future growth prospects, very little o f the fiscal space achieved during the recovery has been used to increase growth-enhancing public investment. 5 Table 1.2: Revenues andExpendituresin Moldova, 2001-2005 (percentof GDP) 2001 2002 2003 2004 2005 rotal Rewnues andGrants 29.2 29.6 34.0 35.4 39.5 Taxrevenues 24.4 25.8 27.5 29.8 32.3 Direct Taxes 3.7 4.0 4.3 4.9 4.5 Indirect Taxes 13.9 14.6 16.0 16.1 18.5 ofwhich: VAT 7.9 9.0 10.1 10.7 12.6 Social and Health FundContributions 6.8 7.3 7.2 8.8 9.3 Non-tax revenues 3.9 3.1 2.6 1.9 2.5 Grants 0.9 0.6 0.3 0.4 1.3 Other Revenues ... ... 3.6 3.2 3.4 rotal Drpenditures andNet Lending 29.4 31.5 33.3 34.6 37.9 NationalEconomy 1.7 1.8 3.6 4.2 5.3 Social sphere 16.6 19.5 20.9 22.0 24.0 Education 4.8 5.5 6.7 6.8 7.3 Social Fund expenditures 7.2 8.4 8.2 8.6 10.0 Health & Health Fundexpenditures 2.8 3.5 3.9 4.2 4.3 Other social expenditures 1.7 2.0 2.2 2.4 2.3 Interest payments 4.2 2.2 2.1 1.9 1.3 Other Ekpenditures 7.1 8.3 6.7 6.5 7.3 Balance o f Special Accounts and statistical discrepancy -0.1 0.6 -0.4 0.1 0.1 herallFiscal Balance (cash) -0.4 -1.4 0.2 0.8 1.7 ' r i m y Fiscal Balance (cash) 3.8 0.7 2.3 2.7 3.0 iources: MTEF 2007-2009and IMF,StaffReport for 2006 Article IVConsultations. 1.15 While the Government neglected the need for increased public investment, it has been very successful in improving o f the overall fiscal policy stance. Duringthe recovery the increase in public expenditures has been kept in line with the increase in revenues. Indeed, starting in 2003, the fiscal accounts o f the general government have even moved into surplus. On average, during2001-2005 the overall fiscal balance has been slightly positive (0.2 percent o f GDP) while the primary balance has been decidedly so (2.5 percent o f GDP). Since the adverse effects of fiscal deficit on economic growth is well documented intheoretical and empirical studies, it i s not surprisingthat Moldova's adherence to prudent macroeconomic policies since 2001 has been an important contributor to restorationo f economic stability and the resumption o f growth. D. I STHE SIZE OFGOVERNMENT TOO LARGE? 1.16 Inaddition to the fiscal stance, the recent empirical literature also suggests that the size o f government itself may have important implications for economic growth. Even though overall fiscal discipline is being maintained, growing and disproportionately high levels o f government expenditures raises some concern about the impact o f fiscal policy on Moldova's future growth prospects. Large public expenditures require high levels o f taxation. High tax rates are, in turn, thought to reduce the rate o f saving and investment as well as distort the incentive to work. Inthe case of Moldova, the tax policy need to finance government expenditures may also contribute to the incentive to migrate, particularly among highly skilled workers. 6 Table 1.3: Revenuesand ExpendituresRelativeto the PredictedInternationalNorm (percent of GDP) Over Over Revenues (under)- Expenditures (under)- Actual Predicted Performance Actual Predicted Performance Armenia 13.0 23.6 -10.6 20.9 31.6 -10.6 Azerbaijan 15.0 23.6 -8.6 27.9 31.6 -3.7 Belarus 39.0 25.8 13.2 44.6 34.0 10.6 Estonia 32.0 29.4 2.6 37.4 31.1 -0.3 Georgia 14.0 22.4 -8.4 19.4 30.0 -10.6 Kazakhstan 23.0 26.2 -3.2 22.4 34.3 -11.9 Kyrgyz Republic 18.0 20.9 -2.9 28.4 28.4 -0.1 Latvia 28.0 28.4 -0.4 35.9 36.6 -0.6 Lithuania 26.0 28.9 -2.9 31.9 37.2 -5.2 Russia 35.0 27.6 1.4 21.4 35.1 -8.3 Tajikistan 15.0 19.3 -4.3 20.1 26.6 -5.9 Ukraine 31.0 25.4 5.6 39.4 33.6 5.8 Uzbekistan 31.0 20.9 10.1 31.8 28.3 3.5 Moldova 28.0 20.5 7.5 34.5 28.0 6.5 Sources:OECD and World Bank database and staff estimates. Note:Basedon regressionframeworkusedby Lorie (2003). Includes65 advanced, transition, and middle- income countries. Datarefer to 2004. Double-logregression of expenditureto GDP ratio on PPP incomeper capitaand regionaldummies. 1.17 Compared to other countries with similar income levels, Moldova spends and taxes too much. Tablel.2 presents cross country data for 2004 for transition economies. In terms of the share actual tax revenues mobilized and expenditures in GDP, Moldova falls somewhere the middle of the group of transition countries. However, after taking into account Moldova's low level of income, the predicted share of revenues and expenditures that would be expected in Moldova i s among the lowest in the region (exceeding only Tajikistan). Thus, relative to this predicted"international norm", Moldova is over performing on boththe revenue and expenditure front.* Revenues are nearly 8 percentage points of GDP higher than one would expect of a country of similar income levels, while expenditures are nearly 7 percentage points higher. 1.18 What i s the impact of the size o f government on economic growth? The ability of an economy to support a large public sector i s not only related to the size o f government, it i s also relatedto the quality o f public programs. Effective public institutions and good governance may mitigate the impact of a large public sector on growth. As shown in Figure 1.2, there i s no negative relationship between the size of Government and economic growth in countries with strong public institution^.^ In countries where public institutions are weaker, a category that currently includes Moldova, the size of the public sector has a decidedly negative impact on The econometric comparison i s based on a large group o f developed and developing countries and takes account o f income differentials as a way to assess whether a country is over or under performing relative to some "international norm" (Le. the predicted level o f revenues or expenditures based on, among other things, per capita income in PPP terms). The analysis uses the Government Effectiveness indicator complied by the World Bank. See D. Kaufman, A. Kray and M.Mastruzzi, "Governance Matters V: Governance indicators for 1996-2005", September 2006. 7 growth. Further, this impact i s even more acute as the size o f government moves beyond a certain threshold. The empirical evidence suggests that in countries where expenditures levels exceed 35 percent o f GDP, an increase in expenditures by 1 percentage point reduce growth by 0.3-0.4 percent per year. This suggests that the current size o f Government in Moldova - 38 percent o f GDP in2005 -i s reducing growth by an estimated 0.9- 1.1percent per annum. Figure 1.2: The Size of Government and Economic Growth I MoreEffediveGoverrments I- f 1.19 How, then, does Moldova reduce the size o f Government? A key element o f this fiscal adjustment will be improvements to the business and fiscal environment that encourage private sector saving, investment and innovation and, hence, private sector led growth. In addition to reducing the administrative and regulatory burdens imposed on private entrepreneurs, the Government needs to support this growth effort by changing the composition o f expenditures to emphasize growth-inducing public investment in human capital, infrastructure and economic services. Of course, reforming public expenditure policy in order to increasing the efficiency and quality o f public spending i s also critical and this topic i s the main theme explored in the subsequent chapters o f this report. 1.20 Reducing the size o f Government will, however, also require constraining the increase o f expenditures relative to GDP growth. As shown in Figure 1-3, continued rapid increases in public consumption, as proposed by the MTEF expenditure plan, suggest that the size o f government will continue to grow over the medium term.4 If, however, the growth in public expenditures were constrained to 2 percent in real terms, then the size o f Government would decline over the medium term and approach the threshold level o f 35 percent indicated above. This analysis utilizes the expenditure plan envisaged under the MTEF, but alters the growth assumption so that economic growth will slow to 3-4 percent rather than remain at level of 6 percent assumed by the MTEF. 8 44 - 42 MTEF expenditure plan .- c - - _- -- - - 40 . - - c e I - . c 2003 2004 2005 2006 2007 2008 2009 Sources: MTEFand World Bank Staff Estimates E. I sTAXATION LIMITINGGROWTH PROSPECTS? 1.21 With the economic recovery, total general government revenues and grants have increased by nearly 8 percentage points o f GDP to reach 40 percent of GDP. Given Moldova's status as a low income country, the ability to mobilize this sizable amount o f tax resources (i.e. over perform relative to the international norm) i s a notable achievement. At the same time, however, it must be reiterated that this relatively large tax burden has the potential of discouraging all important private sector investment and innovation. The creation of fiscal space through further increases in taxation, therefore, i s not a desirable option. The Government has already taken steps some important steps to reduce the burden o f taxation, cutting personal income and corporate profit taxes. It has also strengthened tax administration- concentrating mainly on improving collection o f indirect taxes. This PER concentrates on public expenditures. A separate note will focus on issues relatedto taxation and tax administration. For completeness, however, this section provides a brief overview of the revenue side o f the budget. 1.22 The bulk of revenues come from tax revenues (as opposed to non-tax revenues) and have accounted from almost all the increase inrevenues since 2000 (see Table 1.3). The growth intax revenues has been the reflection of the rapid growth in GDP and foreign trade as well as the increaseinnominal wages. 1.23 Since 2000, direct taxation has increased as a share of GDP, and comprising about one- third of total revenues and grants. Taxation of personal income and corporate profit, however, has not been a major contributor to overall tax revenue mobilization. Motivated by a desire to enhance the business environment, and broaden the tax base as well as by a need to avoid losing competitive ground to neighboring countries on tax competitions grounds, the government has in the last few years reducedpersonal income and corporate profit tax rates. The corporate tax rate has beenreduced steadily from 28 percent in 2001 to 15 percent in 2006. Over the same period, the highest personal income tax rate hasbeencut from 28 to 20 percent. The trend toward lower 9 tax rates, in Moldova and elsewhere in the region, has not generally resulted in lower tax revenues leading some to suggest that this i s evidence o f a "Laffer curve" effect. While rate reductions may be conducive to growth, it heightens the need for improvements in tax administrationenforcement mechanisms and simplifications o f the tax structure. Table 1.4: GeneralGovernment RevenueTrends (percent of GDP) 2000 2001 2002 2003 2004 2005 Total Revenues and Grants 34.8 32.1 33.2 35.3 35.6 40.0 Revenues 30.6 28.3 29.0 30.6 32.1 35.6 Tax revenues 25.8 25.2 26.6 28.1 30.6 33.0 Direct Taxation 9.7 10.5 11.3 11.5 13.8 13.1 Profit tax 1.7 1.8 1.9 2.1 2.4 1.4 Income tax 1.7 1.8 2.1 2.3 2.5 2.4 Social Fundcontributions 6.3 6.9 7.3 7.2 7.9 8.2 Health Fund contributions 1.o 1.2 Indirect Taxation 16.1 14.7 15.3 16.6 16.8 19.9 VAT 8.4 7.9 9.0 10.1 10.7 12.6 Excise 4.1 3.6 2.9 3.2 2.8 3.2 Taxes on foreign trade 1.4 1.2 1.5 1.7 1.5 1.9 Other taxes 2.2 2.0 1.9 1.5 1.7 2.3 Non-Tax Revenue 4.8 3.1 2.4 2.5 1.5 2.6 Special resources and funds 3.4 3.1 3.6 4.4 3.3 3.3 Grants 0.8 0.8 0.6 0.3 0.4 1.3 Source: MTEF and IMF Article IV staff report. 1.24 Most o f the revenues from direct taxation are in the form o f social and health fund contributions. The increase in this source o f taxation i s the result o f the impact o f highnominal wage growth (140 percent since 2001) on the collection o f social insurance fund contributions as well as the introduction o f mandatory health insurance contributions in 2004. The bulk o f these contributions are paid by the employer. In the last few years, however, there has been an effort shift some o f this taxation onto the employee's shoulders. In2001, the social insurance premium was 32 percent, o f which the employer paid 31percent. By 2006, this premium had been reduced to 29 percent with the employer paying 26 percent. This trend i s expected to continue over the medium term with the employer share dropping to 23 percent by 2009. The reduction in the premium. This premium, currently at 4 percent - split equally between employer and employee - social insurance premium also reflects the introduction in 2004 o f a mandatory health insurance i s expected to increase to 5 percent in order to expand the available package o f health services. 1.25 One o f the main features o f Moldova's tax system i s the high degree o f indirect taxation as well as the high taxation o f labor. Social and health insurance premiums, combined with the personal income tax, represent a hightax wedge on labor income. As i s well known, high labor income taxation impedes employment generation, particularly in the formal labor market. 10 Improved tax administration combined with the longer-term potential for the social and health funds to generate surpluses introduces the possibility o freducingthe taxation of labor. Chapter 7 discusses this issue, and the implied tradeoffs, in the context o f the financial sustainability o f the Moldovanpension system. 1.26 Indirect taxation still contributes the most to tax revenues (nearly 50 percent). As evident from Table 1.3, the composition o f this taxation has been shiftingtowards the VAT (Value Added Tax) which now collects (in net terms) nearly 13 percent o f GDP. The standard VAT rate in Moldova i s 20 percent with the number o f goods and services (mainly in the agricultural sector) with zero rate VAT having been reduced inthe last few years. The steep growth VAT collection (nearly 5 percentage points o f GDP since 2000) i s a reflection o f the very rapid growth in imports. While duties collected on foreign trade are relatively insignificant, nearly 90 percent o f VAT revenue is collected by Customs on imported goods. The VAT now constitutes nearly 38 percent o f total tax revenue. 1.27 The appropriate balance between direct and indirect taxation inMoldova will be explored ina separate note. For now, however, suffice it to say that the Moldova needs to reduce the tax burden on the economy. The fiscal space for lower taxation will come from sustained fiscal discipline (thereby enabling a further reduction in the debt burden), better control and efficiency o f primary expenditures, broadeningthe tax base and reducingthe size o f the informal sector. F. COMPOSITIONOFEXPENDITURES 1.28 The increase in fiscal revenues over the recovery has given Moldova the room to increase public spending. While the share o f public expenditures in GDP i s still below the levels achieved duringthe mid-1990s (reflecting many inherited expenditures programs from Moldova's past), the current level o f public spending in Moldova i s high. Further, the consumption orientation o f the public spending has been crowding out growth promoting public investment expenditures (Le. infrastructure and quality improvements inhealth and education). For a given level o f spending, changing the composition o f overall expenditures in favor public investment will contribute to improved long-term growth prospects. At the same time, as the various sector- specific chapters o f this PER will attest, the efficiency with which public resources are utilized must also improve. Given the needto containthe size o f government so as to avoid crowding out private sector activity, the fiscal space needed to increase these growth-promoting expenditures can be only be obtained by containing growth in public consumption and improving the efficiency o f these expenditures. 1.29 This section presents three different ways o f looking at the composition o f Government expenditures: by functional classification, economic classificationand various parts o f the general government. 1-30 Functional Classzjication of Expenditures. The increase in public expenditures since 2001 has been dominated by the social sectors - nearly 90 percent o f the increase in public expenditures has been in those sectors (Table 1.4). As a result, social spending has not only increased as a share o f GDP, it has consumed a growing share o f the budget - from 58 percent of the budget in2001to nearly 63 percent in2005. Overwhelmingly, the increase insocial spending has been concentrated on social assistance and protection, mainly inthe form o f higher pensions. Chapter 5 examines health expenditures in greater detail, while Chapters 6, 7 and 8 look at education, pensions and social assistance spending. 11 Table 1.5: GeneralGovernmentExpenditures-FunctionalClassification (Percentof GDP) 2001 2002 2003 2004 2005 Total Expenditures 30.6 34.4 33.3 35.2 38.0 General Public Services, External Activities and Judicial authorities 2.0 2.3 2.3 2.3 2.5 Defense, Public Order and Safety 2.1 3.1 2.7 2.5 2.5 Social Expenditures 17.8 21.5 20.8 22.0 24.0 Education 5.7 6.9 6.7 6.8 7.3 Recreation, Culture and Religion 0.5 0.7 0.7 0.8 0.9 Health 3.2 4.1 4.0 4.2 4.3 Social Insurance and Assistance 8.3 9.8 9.4 10.2 11.5 Economic, Environment and Science 2.2 3.8 4.0 4.8 6.0 Other expenditures 1.7 1.3 0.8 0.7 0.9 Interest payments 4.3 2.2 2.1 2.4 1.3 Net lending -0.1 -0.3 - ._ -0.2 -.- -0.2 .~ -0.1 .~ Source: MTEFand IMF, Article IV Staff Report 1.31 So-called economic expenditures - predominately spending on agriculture - have also increased significant. As noted inWorld Bank's Agricultural Policy Note on Public Expenditures (2006), however, the level o f public spending on agricultural is low when compared to the importance o f the sector in the economy and spending in other countries. There i s ample public goods such as research and development, rural infrastructure and education - on pro-poor international evidence o f the strong positive impact o f agricultural expenditures - particularly on growth. In Moldova, however, this type of spending has been negligible with most the spending going toward product and input subsidies for larger operators. Under the reform program supported by the Poverty Reduction Support Credit program, the Government has committed itself to reducing wasteful recurrent agriculture subsidies and increasing growth-enhancing spending inthe sector. 1.32 The share o f expenditures devoted to interest payments on debt has declined significantly as a result o f the accumulation o f primary fiscal surpluses since 2001 as well as limited external borrowing. In 2001, interest payments consumed 14 percent of total expenditures. Adding amortization payments to this figure, total debt service paid by the budget was nearly twice that figure. By 2005, however, the burden o f servicing current and past debts had fallen considerably. While rapid growth and the appreciation o f the exchange contributed to this decline, the Government also created fiscal space for increased social spending by running arrears on some external debts owed to bilateral creditors. 12 Figure 1.4: Arrears on Publlc and Publicly GuaranteedExternalDebt period As shownMoldova in Figure 1.4, over this accumulated (Percent of GDP) significant external payment arrears. Starting in 2004-2005, part o f these 4 5 % arrears were cleared as a result o f 40% restructuring deals with bilateral 3 5% creditors (Romanian and Turkey), 3 0% cancellation o f state guarantees on 2s% some EBRD loans (Giurgiulesti and 20% Vininvest) and commercial creditors (Gazprom and Hewlet-Packard). In I5% M a y 2006, a rescheduling agreement I0% was reached with remaining Paris O s % Club creditors that, in addition to O0% restructuring future debt service ZOO0 2001 2002 zw3 2004 2005 payments, has also resulted in the Source: IMFArticle IV Staff Report clearance o f these remaining arrears. 1.33 Economic Classification of Expenditures. Over ninety percent o f public expenditure i s on recurrent spending. The single largest category i s transfers to households. Overwhelmingly, this i s comprised o f pension and social assistance spending. Chapters 7 and 8 explore these expenditure programs ingreater detail. The increase in spending on Goods and Services reflects the introduction o f health insurance and transfers to the MHIF,described inChapter 5 1.34 The rapid increase inpublic consumptionhas crowded out capital spending. As shown in Table 1.5, public sector capital expenditures appear to be a relatively significant and growing budget i s complicated by a number o f measurement issues - such as the inclusion o f maintenance share of GDP. As shown in Chapter 2, however, the presentation o f capital expenditures in the expenditures or spending on repairs that should be treated as current spending - that tend to exaggerate the extent o f public sector capital expenditures. Ifone makes an effort to exclude some o f this current spending from the investment budget, it i s apparent that public investment, particular investment in infrastructure, i s still very low (see alternative measure at bottom o f Table 1.5). Further, a significant share o f capital spending comes from foreign finance investment project which may not represent a sustainable source for needed capital spending. As already noted, the lack o f investment in public infrastructure i s detrimental to Moldova's future growth prospects. Chapter 3 describes the inter-temporal trade-offs and economic costs implied bythis limited investment effort as they pertainto roadinfrastructure. 1.35 The second largest portion o f current expenditures i s devoted to staffing costs (i.e. wages broadly defined). When the Government created the MHIF in 2004, healthcare workers previously recorded at public sector employees were moved "off the books" since the facilities operated by the MHIF were considered financially self-sufficient entities. One o f the implications o f this treatment is that, starting in2004, the staff cost o f most healthcare workers i s no longer included in the public sector wage bill. Instead, these expenditures show up indirectly in Goods and Services, among other expenditure categories. The discrete downward jump in wages evident in Table 1.5, therefore, belies the growing pressures o f public sector wages on the government. As noted inTable 1.5 (alternative measure), the wage bill o fpublic sector, broadly defined, i s closer to 11percent o f GDP, has grown sharply since 2003 and now comprises nearly 30 percent o f total public expenditures. As international experience suggests, the downward stickmess o f nominal wages in the public sector and the lack o f instruments for rationally 13 downsizing the public payrolls make it difficult to adjust the wage bill during economic downturns. Further, in the absence o f reforms to increase the productivity o f public workers, planned increases in public sector wages risks further crowding out government operational and capital expenditures. Chapter 4 examines the public sector wages in greater detail and underscores the need for reformto the pay system. Table 1.6: GeneralGovernment Expenditures-EconomicClassification (percentof GDP) 2002 2003 2004 2005 T ota1 Expe n ditures 31.5 33.3 34.6 37.9 C u r r e n t E x p e n d i t u r e s 26.9 28.9 29.8 3 1 . 7 W a g e s 9.5 9.6 7.8 8 .o G o o d s a n d services 6.9 6.2 7.8 8.1 Interest p a y m e n t s 2.2 2.1 1.9 1.3 Transfers 10.5 10.3 11.6 14.2 Transfers to e c o n o m y 1.4 1.6 2 .o 2.8 Transfers to households 9.1 8.8 9.7 11.5 O t h e r c u r r e n t expenditure -2.1 0.7 0.6 0 .o C a p i t a l expenditure 5 .o 4.6 5 .o 6.4 N e t c r e d i t -0.3 -0.2 -0.2 -0.1 M e m o : A l t e r n a t i v e measures P u b l i c Sector S t a f f i n g Costs 9 .5 9.6 10.4 10.8 P u b l i c I n v e s t m e n t 2.9 2.8 3 .O 3.7 S o u r c e s : MTEF and I M F Article 1V StaffReport 1.36 Intergovernmental Composition of Expenditures and Revenues. Total revenues and expenditures o f the General Government are distributed organizationally between the State Budget, the Budgets o f the Administrative Territorial Units (ATUs or local governments), the (MHIF). It is not surprisingthat the Central Government - the State Budget - with its ability to State Social Insurance Fund (SSIF) and resources o f the Mandatory Health Insurance Fund raised taxes on a national or cross regional level collects the bulk o f fiscal revenues (62 percent). Local governments, on the other hand, have been much more limited in their ability to mobilize the revenues have need substantial transfers from the state budget fund their expenditure requirements. In2005, nearly 40 percent o f ATU expenditures were financed by transfers from other budgets, primarily the State Budget. The SSIF has seen expenditures - primarilypensions - increase rapidly, reaching 10 percent o f GDP by 2005. The SSIF has also been financing expenditures by drawing down accumulated savings (Le. deposits with Banks). Mandatory Health Insurance was introducedin2004, funded by a 4 percent payroll tax (split equally between employer and employee) as well tax on other forms of remuneration. Spending by the MHIF has increased rapidly and by 2005 amounted to around 4.3 percent of GDP. (around 75% of total MHIFexpenditures constituted transfers fromthe statebudget). 14 cfldfturesby LewisafG c, 2005 cent of Tatat) 1.37 l n t e r ~ g ~ ~ e ~ ~fiscal ~ dael c e n t ~ ~ i ~ ~efforts in h ~ o l d o ~have been rather e n i o n ~ a spasmodic^ going t h r o u g ~tivo c o n ~ a d ~ ~ d~ ~~ ~n c~ a ~In~ 1998,s the old Soviet local o n . g o s f e ~ m e system of 40 rayons was ~ r a n ~ f ointo~ai smaller system of 10judcts. W i l e this n ~ ~ e ~ initial effort put in place the ~~undations a s e n ~ l~~ b~ ~~s~l a tfi rv ae ~ ~ e ~it~wasr kf , o ~ by~ ~ ~ e ~ of ~ the absence of c o m p l e m e n ~ ~e ~ u l a ~andnpoor i ~ i p l e m e n t a ~ ~As~tli result of the perceived i o o ~ ~ n e f f e c ~ ~ ~ ofndecen ~ e s s d m 2003, lcading to the rcturrt of the rayan a d ~ ~ n ~ ~and the a t ~ o ~ of judets. ~Completing ~ s ~ e ~ ~ g o ~ ~ e r n ~ i fiscal ~~l ~ c e n ~ a ~ i efforts~ inn~ ~ l d hgvev been rather s p a s ~ ~ d ~going e n ~ z a t ~ ~ a c , through two ~ ~ n ~ a d j~e~~d~i nigc a ~ ~Ino1998,. the old sovtet local ~ o ~ ~ system of 40e ~ ~ f n ~ e ~ m rayons was ~ a n s f o ~ i einto a smaller system of 10judets, While this initial effort put in pIace d the founda~ionsof a sensible tegisiafive f r a ~ ~ e ~it owas~ f~~ l ~ o ~by~ thed absence of r e ~ o n ~ ~ ~ e mree n~ t~al a~ tand~poor i I ~ p ~ e ~ ~ e n ~Asa a~ result,of the perceived i n e f f e c ~ i ~ ~ e ~ e s s i o ~ o n ersed in 2003, leading to the return ofthe o ~ p ~ fiscal~ ~~~ nc ~ ~ ~ ~ randlrefaz a ~ ~ o n e a ~ cononiic d e ~ ~ e ~ ~ ~ n i e ~ t ova. Al~h~ug~i public g many e this represents more policy ~ ~ ~ ~ s i o n1,ocat ~ m ~ao~~ ~~ ~e ~ are~ stilln highly ~ependenton the center wtth r ~. e ~ s ac~ountabi~i~y~ e ~ a i nhierarch~caland vertical, rather than directed at local c ~ n $ ~ ~ ~ Asea~ t ~ , i ~ s u result, the g o ~ e r n ~ ~ attempts to rati~na~izethe state and ~mprovethe e n ~ ' s ~ c of~ e ~ ~ ~ expend~tur~shave been h a ~ ~ e r e d . 15 decentralization requires a harmonic system o f proper revenue and expenditure assignments, an appropriate intergovernmental transfer mechanism to address vertical imbalances and horizontal fiscal disparities, as well as enhanced local autonomy and accountability. The examination of this topic will be undertaken inthe future stage o f the PERwork inMoldova. G. PUBLIC FINANCIAL MANAGEMENT 1.39 While this report doesn't undertake an in-depth examination o f public financial management, it i s important to note that the Government has achieved impressive progress in consolidating the national budget formulation process. The introduction o f the MTEF as a strategic planning framework for the regular budget cycle was the main vehicle for improvement. The MTEF has been developing gradually since 2002 as a comprehensive analytical framework. For a given set o f macroeconomic assumptions, the MTEF provides decision makers with a projection o f the expected national public budget (including the state budget, local budgets, state social insurance budget, funds for state health insurance) outcomes for a three year time horizon. The MTEF i s currently updated annually at the initial stage o f budget formulation and accompanies the annual budget laws presentation to the Parliament. The comprehensiveness of the MTEF approach has promoted adjustments in the coverage o f the annual state budget law to incorporate extra-budgetary funds, as well as the donor funded investment projects. The Government i s working on further improving the strategic policy focus o f the budget through developing stronger linkages between the MTEFand strategic national policy documents (Le. the EGPRSP and EU (European Union)-Moldova Action Plan). An important area for further improvement i s the development o f the capacity at the line ministries' level to link expenditure planning to sectoral policy priorities. 1.40 Progress has also been achieved in strengthening o f the national treasury system and preparing for its modernization. Controls over budget execution have been significantly tightened through the gradual expansion o f the treasury coverage that includes by now the budgets o f territorial administrative units, extra budgetary funds and means, and revenues o f the social and health insurance budgets. The budget execution process i s expected to be modernized in the next few years on the basis o f internationally recognized methodologies simultaneously with the replacement o f the existing patchy semi-automated treasury system with the modern FMIS. Further consolidation o f the budget executionprocess and improvements in cash management are planned through implementation of the Treasury Single Account (TSA) concept. At the same time, the Government i s preparing for important changes in public sector accounting and reporting. The new integrated budget classification and chart o f accounts system i s being developed on the basis o f the GFS2001 standards to be launched with the new Financial Management Information System (FMIS). Multiple versions o f the chart o f accounts currently in use by different levels o f budget entities will be replaced with a unique chart of accounts compliant with cash based InternationalPublic Sector Accounting Standards (IPSAS). 1.41 The development o f the public sector auditing capacity has also been receiving growing Government attention recently. The concept for the development o f the modern Government internal audit function in line with EU requirements was recently developed with support from the PFM project. At the same time, Moldova's supreme audit body, the Court o f Accounts (COA) has indicated a willingness to develop into a modem external audit institution along the EUcounterparts through launchingits medium-term strategic development plan. 16 H.CONCLUSION 1.42 The increase in fiscal revenues that accompanied Moldova's economic recovery has given it substantial room to increase Government expenditures. Despite devoting nearly 40 percent of GDP to public expenditures, and even accounting for the prospect o f increased external grant assistance, there are a still a host of unmet demands on the budget. Roads are in dreadful wages and pensions are still very low and poverty - particularly rural poverty - i s high. Almost shape, higher natural gas prices are increasing the demand on social safety nets, public sector all of fiscal space achieved during the recovery, however, has gone to increase public consumption- particularly, public sector wages, transfers and subsidies - expenditure items that have been severely eroding in the transition process. As a result, critical infrastructure needs are largely neglected and state assets, such as roads, hospitals and schools, continue to deteriorate at an alarming rate. 1.43 Moldova must confront the reality that the size o f Government has grown too large. The tax burden needed to support high levels of government consumption spending negatively affect private saving and investment decisions and may limit Moldova's future growth prospects. This burden is compounded by the fact that the public sector's contribution to investment is critically low. As long as public expenditures are increased at a slower rate that GDP, the size and burden of government will gradually decline. 1.44 IfMoldovais to accelerategrowth andpovertyreduction, itmustcurtailexcessivepublic consumption spending and increase spending on public investment while at the same time reducing the overall footprint o f Government. Cutting wasteful spending in the education and health sectors, for example, frees resources for more investment spending in those sectors. Increased spending on roads will put a halt to the deterioration o f this vital infrastructure, and support increased private sector activity and greater international competitiveness. Improving the targeting o f social assistanceallows these programs to more effectively protect vulnerable groups by freeingup these misdirectedresources. Completing the reforms to the pension systemreduces the accumulation of future unfunded liabilities and puts public finances on a sounder footing. The reform to public financial management currently underway provides an opportunity to strengthenthe linkages between spendingprioritization and budget management. This process o f fiscal adjustment i s essential to put economic growth on a solid and sustainable trajectory. 17 Box 1: Issues in Public Expenditure Management: Summary of Previous Findings ThisPublic Expenditure Review buildson a knowledge base o fissues inpublic finance inMoldova that has been accumulated over time through various studies and in the course of public-finance related project preparation and implementation. The scope and coverage o f the PER is thus driven in large part by the availability o f detailed information on aspects o f public finance inMoldova. Inparticular, there have been several recent efforts to examine major issues inpublic expenditure management. We summarize some o f their main findings below: The World Bank's 2003 Public Economic Management Review (PEMR) conducted a comprehensive review o f the budget process in Moldova, including strategic decision-making, budget coverage, budget formulation and execution. The report suggestedmeasures for improving the comprehensiveness o f budget coverage by, among other things, incorporating externally financed projects and extra-budgetary funds. In addition, the report suggested improvements to the quality o f macroeconomic forecasts and revenue projects underpinning the budget process. I t also suggested broadening the coverage o f the expenditure commitment control system and procedures to strengthen budget execution. To support budget execution and monitoring further, the report encouraged the production o f budget reports that comprehensive, timely and reliable. Similarly, the 2003 Country Financial Management Accountability Assessment (CFAA) found the financial accountability framework in Moldova to be weak, characterized by a fragmented budget framework process, inefficient cashmanagement, weak internal control and internal and external audit. The IMF has also been actively engaged inexpenditure management issues inMoldova. The 2004 Report on the Observance of Standards and Codes-Fiscal Transparency Module assessed practices in fiscal transparency in Moldova and noted the lack o f integration o f extra-budgetary and donor-financed funds into the central government budget, the lack o f coordination between state and local governments, weaknesses in the preparation and monitoring o f the budget on a general government basis. The 2006 Report on Observance of Standards and Codes examined, among other things, the availability o f fiscal data, its coverage, periodicity, and timeliness. I t noted weaknesses in available fiscal data, including the absence o f published information on expenditure by economic classification. Neither are consolidated central government data disseminated. Another recent paper (2005) notes the absence o f comprehensive monthly budget reports. It also assesses the treasury system and noted weaknesses in cash management driven inlarge parts by the resources and funds not fully integratedinto the treasury system. The recently completed Public Expenditure and Financial Accountability (PEFA) Assessment documents many o f the strengths, as well as remaining weaknesses, in the public finance management system o f Moldova. Inparticular, the assessment noted the improved credibility o f the budget and gave the system high marks in transparency and public access to key fiscal information. A t the same time, the assessment observed the limited effectiveness o f payroll controls, the lack o f a consolidated overview o f fiscal risk (including the financial performance o f state-owned enterprises andjoint stock companies), the weaknesses in financial and performance audit (including the lack of legislative scrutiny and involvement), and challenges infully incorporating donor support into budget processes and management. Many o f the recent improvements in the budget process (also discussed in the main text) reflect key recommendations from these earlier studies. An ongoing World Bank-supported Public Finance Management project has also been working with the Government to improve the budget process. Sources: World Bank (2003) Public Economic Management Review; WB (2006) Public Financial Management Project: Project Appraisal Document; IMF (2006) Report on Observance of Standards and Codes; IMF (2005) "Fiscal Reforms" inRepublic of Moldova: Selected Issues; IMF (2004) Report on the Observance of Standards and Codes-Fiscal Transparency Module; and Republic of Moldova: Public Expenditure and Financial Accountability Assessment (May 2006). 18 2. INVESTINGFORECONOMIC GROWTH: PUBLIC INVESTMENT PLANNINGAND MANAGEMENT IN MOLDOVA A. INTRODUCTION 2.1 Deficiencies in public infrastructure represent a considerable constraint to economic growth inMoldova. Despite the demand for upgrading and expandingthe public infrastructure in Moldova, public investment has remained quite low in recent years-only about 2 percentage points o f GDP. This underscores the need to increase the amount o f public investment spending. The limited availability o f public resources also highlights the importance o f strengthening the management of public investment in order to ensure that the available resources are utilized effectively and efficiently. 2.2 Because public resources are limited, there is a need to prioritize and plan government investment strategically. Over the last few years, the Government has significantly improved the budget preparation process and the strategic allocation o f resources, mainly through the introduction o f the Medium Term Expenditure Framework (MTEF). The Government has also taken steps to integrate the investment budget into the MTEF and the annual state budget. The latest MTEF (2007-2009), for example, includes an investment annex reviewing actual spending inrecent years andproposinganinvestment budgetthrough 2009. 2.3 Despite these achievements, a number o f weaknesses remain. Public investmentplanning and budgeting still remains fragn~ented.~Furthermore, the portfolio o f investments being implemented remains too large compared to the available financing. Inaddition, investments are being undertaken without adequate consideration being given to the requirements for financing o f downstream recurrent costs. For further rationalization o f the public investment portfolio, a number of improvements are required in the identification, prioritization, preparation and appraisal o f projects. Nascent steps towards integrating externally financed investment expenditures needto be further strengthened. 2.4 This chapter aims to clarify weaknesses in the investment budget process and identify opportunities for increased efficiency. Inparticular, this chapter o f the PFR seeks to describe the current public investment program in Moldova and examine the institutional features of public investment budgeting. First, it will document the present level and sectoral composition o f ongoing investment projects. In addition, this section will examine the investment program's current and projected sources o f financing. Second, it will examine the budget preparation process, including the links between investment expenditures and budget allowances for future recurrent costs and debt service, the budget time horizon, the integration o f investment projects into the annual budget (or the medium-term expenditure framework), and the measures in places to ensure investment projects are economically and technically sound. This section will also look at arrangements for managingproject implementation, monitoring, and ex post evaluation. 11. Level and Compositionof PublicInvestment 'There have been indications that both foreign and domestically financed investments will now be covered within the Ministryof Financebya single agency, but this remains unclear. 19 The Measurement of Investment ExpenditureinMoldova6 2.5 There are three major issues in the measurement o f investment expenditure for Moldova to conform to international practice. First, investment expenditures inMoldova have too narrow a definition, which needs to be broadened to correspond to best international practices. Second, with respect to budget presentation and budget preparation, investment budgeting remains fragmented, with local government investment spending, for example, often functioning as a residual. Finally, repair or maintenance spending, which are, in principle, recurrent spending, are often classified under capital or investment spending. This i s somewhat related to the broader discussion on whether "public investment" should include not just the acquisition o f fixed assets but also the development o f institutional capacity, the development o f management systems and procedures, and others. In this chapter, we generally use the term "public investment" in the narrower sense o f capital spending, or the acquisition o f fixed assets, and excluding maintenance spending, the training or the upgrading o f systems andprocedures.7 2.6 "Objects" as units of measurement. In Moldova, public investment i s typically understood in term o f small, separate "objects." In contrast, based on international practice, investment units are typically understood interms o f "projects." A project may consist o f distinct components (conceptually similar to `objects' in the Moldovan context) necessary to achieve a project's intended objectives; for example, a project that aims to provide additional secondary school facilities in a certain geographic area consists o f several components including the construction o f such facilities, the provision of the necessary equipment for these facilities, the connection o f these facilities to network sources o f energy, and so on. While such components may be financed separately or constructed by separate contractors, it is typically much more convenient for analytical and budgeting purposes to think o f these outputs as components o f a single project. Projects, inturn, are typically grouped under broader programs to facilitate budget analysis and clarify the cross-sectoral allocation o fresources. 2.7 The fragmentation of the investment budget. Until recently, externally-financed investment projects as well as projects financed by "special fund and resources" were not clearly identified nor integrated into the annual budget. Domestically-financed capital spending, on the other hand, was identified as a single item. Local government investment expenditures continue to be reported separately. The fragmented presentation o f investment budgets, and the lack o f an integrated investment budgets reflect a more fundamental fragmentation inthe budget preparation system, as described inthe next section. 2.8 The classifcation of maintenance spending. For analytical purposes, maintenance expenditures or expenditures on standard repairs are typically classifiedunder "current spending," following international practice or the Government Finance Statistics (GFS) classification system. In Moldova, such routine maintenance spending has sometimes been classified under capital spending, leadingto the overestimate o f such spending. The information provided in this section i s drawn in large part from Groom (2005) and confirmed by the consultations with national authorities. 7 Note that some o f the foreign-financed expenditures may still include some training and upgrading o f management systems. 20 Trends inInvestmentExpenditure:Leveland Composition I Capitalspending Currentspendind defines "public investment." Measuring public I." I investment as broadly as possible, including 6 0 5 0 expenditures that are typically understood in 4 0 international practice as current spending (e.g., 3 0 spending on the maintenance o f existing assets), 2 0 1 0 suggests that public investment has averaged 0 0 about 5 percent o f GDP in recent years. 2002 2003 2004 2005 Sourcer:WEF (2006) and World Bank staffestimates Meanwhile, a very recent attempt by the ' government to capital spending, excluding maintenance and other recurrent expenditures, over the last few years suggests that between 2002 and 2005, average capital spending was about 3 percent o f GDP (Figure 2.1). 2.10 Measurement issues notwithstanding, a few things stand out from existing data on public investment. The include the following: Figure2.2: PublicInvestmentSpending: 2.11 First, capital spending has been EconomiesinTransition2004 (In percent of GDP) relatively low. Compared to other countries in the region, capital spending in Moldova i s significantly lower, on average, in percent o f GDP (Figure 2.2). Based on cross-country (In percentof GDP) 6.0 data for 2004 using comparable data across 5.0 transition economies, capital spending inother 4.0 countries in the region i s close to 4-5 percent 3.0 o f GDP. Many o f the EUaccession candidates 2.0 and CIS countries, in fact, spend over 5 1.0 percent of GDP. 0.0 New EU Accession Western CIS Moldova 2.12 Second, infrastructure spending has memberstates candidates Balkans been low and has decreased as a share o f total Souw World B d (IW6).MoldonMIW. mod~ Y R a m m m spending over time (Figure 2.3). Public spending on infrastructure currently amounts to about 2 percent o f GDP. As a share o f total public investment spending, infrastructure decreased from 72.7 percent in 2002 to 56.2 percent in 2005 (Figure 2,3).* As defined in the MTEF, "infrastructure" includes public investment in agriculture, environmental protection, industry and construction; transports, road services, communications and information; utilities and residential services, fuel and power industry,among others. * Despite this decrease, the estimated levels of infrastructurespending may still be exaggeratedas maintenance spendingand capital spending are often blurredin the budget process. With respect to externally-financedpublic investment, much of this spendingreportedlyaccrues toward capacitybuilding activities and efforts to introduce newmanagement systems and procedures. 21 Figure2.3: Compositionof Public Investmentin 2.13 Of total infrastructure spending, Moldova (percentof GDP) the gas network program accounts for the largest share. The gas network project (also known as the "gasification program" or the effort to connect rural households to the natural gas network), 3.5 3.0 in fact, has accounted for about half of 2 5 the central government's investment 2.0 1.5 spending and dominates planned central 1.0 0.5 government investment program through 0 0 2010. As a result, there have been few 2W2 2W3 2004 2005 investment opportunities for other Sources: MTEF (2006) and World Bank staff estimates sectors. Investment in the health sector has been mostly in the form of maintenance or repair o f existing facilities. There has been little or non-existent investmentinthe transportation sector and as a result the road network has deteriorated with a substantial portion of the network classified as in poor condition. Starting with the 2007 budget, however, i s projected to allocate more resources toward this sector. 2.14 Third, local budgets finance about a third of all investments. Domestic spending is increasingly dominated by local government spending. Local government spending has grown rapidly in recent years to account for most of domestic spending by 2002. Financingof capital investments from local budgets i s not predictable, as they are basically not planned by local budgets. When state investment budget falls short, local authoritiesdirect part of their resources towards investment^.^ Not surprisingly, the medium term- budget over the 2007 to 2009 period excludes localbudgets' plannedinvestments. 111.PublicInvestmentPlanningandManagement 2.15 Public investment budgeting in Moldova has a relatively short history. In the 1990s, Moldova developed two Public Investment Programs (PIPS), with very disappointing results. Because they were mainly a tool for drawing external assistance, they were developed independent of the budget process and with no clear connectionto the strategic priorities of the government. Some of the key weaknesses ofthe PIPprocess are summarizedinBox 2. 2.16 There are notable improvements inthe budgetprocess in general over the last few years, including inparticular, inthe area of investment planning and management. First, the investmentbudget, coveringbothdomestic andexternally financed investment, has been graduallyincorporatedinto the state budget with the budgetannex providing details ofthe relevantprojects. 9Local budgets are observed to "underestimate" own-revenues. During the budget rectification process, local governments then direct "extra" resources to investment purposes. 22 Box 2: Moldova's Experience with PIPs, 1994-1997 Current practices and weaknesses in investment budget management mirror many shortcomings in Moldova's previous two Public Investment Programs (PIPs). The first PIP was prepared in 1994 and covered the 1995 to1997 period. It included 19 investment projects amounting to some $169 million. The second PIP was prepared in 1997 and covered the 1998 to 2000 period. The program consisted of some 36 projects amounting to $1.3 billion o f which close to 40 percent o f budgeted resources were allocated to the transport sector. Projects included in the PIP were typically proposed by line ministries and were examined by the Ministry o f Economy. They were then submitted for approval to the Inter-Ministerial Committee on External Relations. The notable weaknesses o f the PIPs include the following: First, the PIP representedone component in a fragmented investment budget. Like many o f the PIPs initiated in transition economies in the 1990s and in many developing countries before then, these programs were mainly a vehicle for soliciting external financing. The Ministry o f Finance had limited involvement in the PIP budget process and the PIP excluded projects that were completely financed by domestic resources. Instead, the PIPs covered only projects with external financing. Second, the portfolio o f investments was often too large and overambitious given available financing. The first PIP was prepared from a list o f about 4,000 public investment "objects", many o f which were inherited from the Soviet period. The number o f "objects" was subsequently trimmed, reducing the cost o f the program from an original $1.9 billion to $169 million, but overtime in the PIP process appeared to have quadrupled in size, expanding its sectoral scope and significantly exceeding available financing. Third, many of the projects appearedto have been poorly selected, with littleregard for their economic value and with little reflection on the proper public and private roles in the provision o f these investments. A review o f the first PIP revealed a number of projects that were private and commercial in nature, such as a cement factory at Rezina." The expansion o f the gas network to connect rural households, proposed in 1994-95, was found to be o f little economic value. In contrast, few resources appear to have been allocated to the rehabilitation o f existing public assets, such as roads. Fourth, the PIPs appear to have had a poor implementation record. Of the 19 projects listed in the 1995-1997 program, only one project was started by August 1997, with another project tendered around that same period. The Government's review o f the experience with the first PIP blamed "insufficient capacity to complete the preparatory phase o f the new projects," among other reasons." Little is known about the implementation record o f the second PIP. However, it appears that the second PIP secured little o f the financing request for implementationand was subsequently discontinued with the election o f the new government in2001. 0 Second, the MTEFhas also facilitated improvements inthe annual budget process. Four MTEFshavebeenprepared since 2002 andhas provided greater discipline inthe annual budget process and improvements in sectoral planning and allocation within the overall resource envelope, as reflected inthe budget proposals submitted by line ministries." I OSee World Bank (1996) Moldova Public Expenditure Review. I1 Ministry of Economy & Reforms (1997) Public Investment Programme (Chisinau: Republic of Moldova Ministry ofEconomy & Reforms). 12 These weaknesses are described in the PFM project document and they include the lack o f political commitment to the MTEFto make it more bindingon the annual budget exercise, the quality o f macroeconomic 23 0 There are efforts to plan public investments more strategically. As mentioned, an MTEF technical annex was recently prepared on public investment.l3 2.17 Nonetheless several weaknesses remain, as documentedbelow. 2.18 Budget presentation. Public investment projects are not integrated into a fully consolidated general government investment budget, as o f the most recent Annual Budget Law (2007), reflecting the broader fragmentation o f the general government budget. There i s no consolidated general government investment budget. While budget information on investment expenditures exists separately for donor-financed spending, special funds and means, the National Social Insurance Fund and the Health Insurance Fund budgets, and local government budgets, they are not consolidated for presentational purposes. Investment spending i s thus also currently included in several different places within the annual budget document. Projects financed by external credits and grants are listed in an annex to the budget. Domestically financed investments at the central government level are usually included as a budget line within the budgets of various line ministries. 2.19 Budget preparation. The investment budget process is best understood withinthe context o f the annual budget process. The MTEF guides the annual budget process and provides the basis for the overall annual resources envelope and expenditure ceilings that in turn determine annual budget negotiations. It i s not, however, submitted to the Parliament for discussion or approval, although it was sent for information inthe latest round. While the MTEF provides broad guidance to the strategic allocation o f resources, the basis for allocation investment resources more narrowly i s not explicit. 2.20 The procedures for selecting public investment projects, whether across sectors or within sectors, are not clear and there is no evidence that the analysis o f expected benefits and costs o f proposed projects (whether based on economic returns, poverty impact, or some other measurable goal) factor into project screening, selection and pri~ritization.'~Public investment i s also hampered by the persistent overhang o f uncompleted projects. The Government's list o f ongoing investment projects is derived from a list o f uncompleted projects, some o f which date back to the Soviet period. Of the total list o f 449 projects (or `objects') drawn up in 1999, the Government selected 251 projects and abandoned the rest. It is estimated that that the 198 projects abandoned would have required 3.3 billion lei to complete in2000 prices, or 20 percent o f GDP. In2005, the list o fprojects was further trimmed down to 201. 2.2 1 Multiple decision-making authorities. The MTEF is preparedjointly by several agencies including the Ministry o f Economy, Ministryo f Finance, National Bank o f Moldova and various line ministries. The Ministry o f Economy takes the lead in preparing the macroeconomic forecasts. For the annual budget process as a whole, the Ministryo f Finance i s the lead agency in budget planning and budget preparation. It also leads the intersectoral working group that coordinates the sectoral allocation o f resources. For the 2005 and 2006 annual budgets, the forecasting, and the lack of technical capacity to assess the immediate and future recurrent costs o f policy intervention. 13A capital investment program was reportedly prepared in 2005 by the Capital Investment Directorate (CID) within the Ministryo f Finance but the document has not beenmade publicly available. l4Seealso IMF(2004) Report onthe Observance ofStandardsandCodes (ROSC) FiscalTransparencyModule (Washington: IMFFiscal Affairs Department). 24 Capital Investments Division o f the Ministry o f Finance led the formulation o f the capital investment program within the MTEF process. For the 2007 annual budget, a unit within the Ministry of Economy was leading efforts to introduce a more strategic setting o f priorities in public investment across sectors. A technical annex on investment spending for the 2007-2009 MTEFhas been prepared by the interministerial working group ledby the Ministryo f Economy and Trade with participation o f other ministries, including the Ministry o f Finance (see the new MTEF), with proposed expenditure ceilings for the relevant sectors. Defining the intra-sectoral priorities within these ceilings i s an exercise to be carried out duringthe budget cycle. 2.22 TheRole of Parliament. Significant reallocation o f investment resources i s carried out by Parliament, including the deletion o f proposed projects and the addition o f new projects. Prior to 2005, the investment budget law was passed separately by Parliament and included major revisions to the draft investment budget submitted by the Government. Line ministries report final allocation o f investment resources that bear little resemblance to their original proposal. The process o f mid-year budget rectification, and the significant amount o f resources involved (initial revenue projections are typically conservative and, inrecent years, budget rectifications increase actual expenditures to accommodate significantly higher actual revenues) leads to additional reallocation o f resources, unrelated to the stated sectoral allocation indicated by the MTEF (see for example Table 2.1). In particular, while education appears to be the key priority area for investment within the EGPRSP, the energy sector (the `gasification' project), in practice, has dominatedthe investment budget. 2.23 While the local government appears to have assumed a greater role in Moldova's public investments, it i s not clear how local government budgets are prepared with respect to the aggregate investment budget. It appears that local government spending functions as a residual or a source o f funding for bridging financing gaps in the proposed investment program o f selected ministries. 2.24 Budget Execution and Monitoring. After the annual budget law i s approved by Parliament, the Ministry o f Finance distributes the allocation limits. Line ministries and spending agencies enter into contracts for procuring goods and services based on the appropriations; these contracts are in turn registered and verified by the Agency for Regional Development. With respect to investment projects, the work i s verified by the Ministry o f Finance in consultation with line ministries, spending agencies and local government units.Treasury verifies that request for payments submitted by line ministries and spending agencies are consistent with appropriations and the availability o f funds. Payment orders are then processed by Treasury and transfers are made between accounts. Local government units conduct their operations through territorial treasury branches. 2.25 Main responsible agencies. For domestically-financed investments, the Ministry o f Finance and its various directorates i s the lead agency for managing budget execution and monitoring. With respect to externally financed investment projects, such projects, untilrecently, were not fully integrated into the budget monitoring processes. These projects, typically managed by Project Implementation Units (PIU), followed their own accounting and reporting practices (reflecting in part donor-specific practices in accounting and procurement) and were outside the treasury and mainstream fiscal operations. Efforts are underway to integrate these projects more fully into state budget monitoring. PIUs are now required to report activities to the Public Debt Directorate (PDD) withinthe Ministryo f Finance, on a monthly basis. The PDD inturn reports to the treasury. There have been proposals to prepare comprehensive reports on budget execution, includingby functional classification. 25 Table 2.1: Investment Budget Allocation 2005 EGPRSP MTEF 2005- BudgetLaw Mid-Year Actual Budget Ratio of costs 2007 Budget EGPRSP (coveredby document Rectification Budgetto the MTEF) Actual (InthousandMDL) Sector Education 57,200 28,644 18,659 52,839 68,300 0.84 Health 6,180 8,75 1 10,100 14,310 9,680 0.64 Energy 5,740 53,920 45,588 157,649 248,339 0.02 (InpercentofGDP) Sector Education 0.16 0.08 0.05 0.14 0.19 0.84 Health 0.02 0.02 0.03 0.04 0.03 0.67 Energy 0.02 0.15 0.12 0.43 0.68 0.03 Sources:NationalauthoritiesandWorld Bank staff estimates. Weaknesses inInvestmentBudget Management 2.26 As reviewed in the previous section, there have been many notable improvements in investment budgeting, due in large part to the introduction o f the MTEF process. However, as illustrated by the brief review o f the experience with PIPS,several weaknesses in the PIP budget process reflect weaknesses that persist in current budget practices. The review further illustrates how the focus o f the PIP process on securing external financing and its separation from the rest o f domestic budget processes may have significantly limited its usefulness. Rather than reviving the PIP process, however, the Government should be exploring opportunities for strengthening current procedures, with the objective o f facilitating the realistic planning and programming o f the investment budget within the MTEF framework. The following are some persistent challenges ininvestment budgetingand opportunities for improvement: 2.27 There is no clear mapping of the investment program to EGPRSP priorities. The EGPRSP i s an ambitious plan for sustaining growth and poverty reduction and undertaking the necessary structural reforms to meet the needs o f a market economy. While the public spending envisioned by this development plan exceeds available financing, it does lay out a strategic direction for the allocation o f scarce government resources. Inpractice, however, there i s no clear alignment between these stated priorities and budget planning. Inparticular, inthe case o f public investments, weaknesses in the project selection process as well as multiple decision makers involved in the budget process have typically led to the final sectoral allocation o f resources that are radically different from that intendedby the EGPRSP (see for example Table 2.1). While the fundamental issues regarding the strategic allocation o f resources are not specific to the investment budget process, improving the efficient allocation o f scarce public investment 26 resources i s critical in light of massive investment and infrastructure upgrading needs of Moldova. This will require effort along the following fionts: Box 3: Investment Budget Preparation: SomeExamples In South Korea, the authorities recently introduced the concept of preliminary feasibility studies (PFS) to promote efficiency inthe selection o f major infrastructure investment projects. Based on the new initiative, every large-scale project expected exceed US$50 million is subject to a PFS, under the direction o f the Ministry o f Planning and Budget (MPB). The PFS consists o f two sections: an economic assessment utilizing cost-benefit analysis and a policy analysis, covering social issues that are difficult to quantify. The policy analysis section, in turn, is subject to two different types of assessment, one based on "common criteria" (regional development impact, regional economic impact, environmental impact assessment, and others) and "project-specific criteria" based on the recommendations o f the PFS Review Committee. To synthesize the results o f the economic and policy analyses, MPB uses a multi-criteria decision-making process, with the results o f the economic analysis typically more heavily weighted (65 percent) than the policy analysis (35 percent). Between 1999 and 2004, some 209 projects were evaluated o f which about 129 were passed. For these projects, budgets were allocated for more detailed feasibility study and basic design. Proposals have been made to scale up this process, primarily applied to large construction programs, to other areas such as subsidyprograms. In Chile, a comprehensive evaluation of public investment projects is led by the Budget Department of the Ministry of Finance. Public investment project proposals are submitted by line ministries, which are then assessed and prioritized by the Department. Chile has some 30 years o f experience in the use o f economic analysis in public investment screening and appraisal and in the systematic evaluation o f public spending, more broadly, It is considered by many to be a good practice example expenditure management. Some 60 percent o f public spending i s evaluated in a systematic way. Cost-benefit analysis o f public investment began in the 1970s and has been sustained and improved over time. Chile is also reported to have been making an effort to strengthen ex post monitoring and evaluation. In Latvia, based on a review of its investment planning and management inthe 199Os, the Bank recommended the development and articulation o f clear priorities and strategies to drive the allocationo f investment resources. The review also recommended the rigorous application o f cost-benefit analysis to all projects included in the public investment program. A preliminary assessment was also carried out, during the review process, o f the share o f high-priority projects (based on expected impact) and low-priority project in the existing public investment budget.They also compared these broad shares with available financing. In the U.S., the General Accountability Office provides some useful guidelines for assessing proposed investments and their projected effects, The present a series o f questions meant to address (1) the productivity- enhancing potential o f a project; (2) the value o f an investment; (3) whether the design contributes or hampers a project's effectiveness; and (4) how to create the conditions for effectively evaluating the actual results o f an investment. Sources: Jay-Hyung Kim (2005) Developing and Managing a Public Investment Program in Korea (Seoul: The Korea Development Institute); World Bank (1994); U S General Accounting Office (1993); IMF(2005) Public Investment and Fiscal Policy. 0 First, there i s the fundamental needto clarify the Government's priorities given multiple planning documents: In addition to EGPRSP, there are the EU Action Plan, the Moldovan Village Program and sector-specific strategic plans. While there i s plenty o f overlap in these planning documents, the multiplicity of planning documents also suggests that the list of strategic priorities may be overambitious. 0 Second, a related point, many of these p l a i i n g documents and list o f priorities are not fully costed; not surprisingly, there is no full financing for these priorities. The Government thus needs to clarify, given limited domestic and external resources, what are the priorities for spending and for investment expenditure more narrowly. The 27 alternative-allocating inadequate financing to many investment projects-is not efficient. As such, investment projects are taking several times longer to complete due to inadequate funding provision. These projects, therefore, consume valuable public resources without providing timely economic benefits. 2.28 Third, submitting the MTEF to Parliament for information and discussion early in the budget preparation process may support greater fiscal discipline and closer alignment o f the Government's stated priorities and the final allocation o f resources. The use o f economic analysis o f the costs and benefits on investment projects may provide a clearer ranking o f investment priorities to support the Parliament's decision-making process. Grouping smaller objects within wider projects, as suggested earlier, may also limit space for the micro-management o f the investment budget, as currently practiced. 2.29 The investment budget remains fragmented (i-e., external versus domestic, central versus local). Despite recent gains inmedium-term planning o f investments and presenting more fully all investment projects in the annual budget law, there are a few challenges remaining. A complete list o f investment projects, integrating domestically and externally financed projects, needs to be compiled and aggregated within major programs and sectors. Such a comprehensive list can then be reviewed for consistency with the broader government strategy. l5A consolidated general government budget on investment will facilitate improvements inpolicy and planning. 2.30 The budget preparation process does not involve systematic procedures for project screening and project selection. Project screening includes the initial project identification stage. This involves the careful initial screening ofthe project concept and scope, and its alignment with the Government's stated priorities, before the proposed project advances into more detailed preparation and appraisal stages (including detailed feasibility studies, cost-benefit analyses, social and environmental impact analyses, where appropriate and an assessment o f the financial viability o f the proposed projects based on these studies).16 Project selection then selects for fhding, within the MTEFhudget process, a subset o f sound projects from among those previously screened, given that public resources are insufficient to finance every sound, proposed project. 2.3 1 InMoldova, some documented rationale is required for proposedprojects, but otherwise no systematic economic evaluation, cost-benefit analysis or feasibility studies o f project proposals are carried out although systematic assessments are required by the organic budget law. As a result, there i s no technical assessment o f lower-cost alternatives to the proposed public investment projects. Recognizing that the technical capacity required for conducting meaningful cost-benefit analyses can be quite demanding, an emphasis on the project identification stage- where the project concept and alignment with strategic priorities are assessed-may go a long way toward ensuring a better use public resources inMoldova. 2.32 There are also clear opportunities for greater efficiency and fiscal savings, where policy problems meant to be addressed by proposed investment projects may lend themselves to non- engineering solutions or lower-cost alternatives such as public-private partnerships or regulatory reforms to encourage greater private sector participation. In some cases, ensuring the sustained financing o f non-wage, recurrent maintenance costs may be more economically viable that new 15 Such an aggregation has previously been used in other countries leading to a substantial reallocation of resources across sectors and across programs, within sectors. 16 See Groom (2005) for a more in-depth discussion o f various stages o f the project cycle and how they are relatedto the budget process. 28 investment spending, inthe long-run. InMoldova, however, many public assets have deteriorated in the absence of steady spending on maintenance. Inother cases, other instruments for public intervention may be more appropriate. For example, experience with public infrastructure investments or investment inutilities inother economies intransition suggest that, in some cases, a policy o f deregulation or price liberalization intransportation or the increase in energy prices to full cost-recovery-to create incentives for private sector investment-may make more economic sense than the public provision o f infrastructure." These are o f course related to the more fundamental review o f the proper public and private roles in capital spending. In a number o f economies in transition, such a review has helped weed out a number o f proposed projects that were primarily commercial in nature and leftover from the central planning period.18 A similar review in Moldova may be productive and a normative framework for public investment, illustrated inBox 3, may be helpful. 2.33 First, wide disparities still exist inbudget drafts submitted by line ministries and in draft budget submitted to Parliament, despite the reported greater realism inthe budget submissions o f line ministries and spending agencies. This could be an indication o f the persistent o f unrealistic budget planning and overambitious list o f strategic priorities. 2.34 Building the technical and institutional capacity to facilitate the creation o f an environment for evidence-based, strategic policymaking will involve, o f course, non-trivial efforts. But the payoffs, with respect to reductions inthe waste o f resources and improvements in the efficiency o f government spending, are well worth the cost. 2.35 While the Government has taken steps to introduce greater predictability in the budget process, many weaknesses remain. There are a number o f opportunities for introducing greater discipline inthe investment budget process: 2.36 Second, a more comprehensive assessment o f the fiscal impact o f new public investment will be useful, including its longer-term impact on public resources. There are currently no systematic efforts to estimate the implied future recurrent expenditure requirements o f new investment projects, the so-called "r coefficients" or the ratio o f incremental recurrent costs to total investment, for both domestically financed investments as well as externally financed investments. These coefficients reflect the amount o f maintenance spending, repairs, and personnel expenditures required for the efficient functioning o f a completed investment project. In some countries, the computation of these coefficients has served to demonstrate that some investment projects were unsustainable inthe long run. Table 2.2 reports indicative r-coefficients by sector for selected World Bank-supported investment projects across countries. The coejkients reflect in large part sign$cant diferences in economic environments across countries and they should be interpreted as illustrative ranges. They suggest that the incremental recurrent spending requirements can be substantial. For example, a 0.03 r-coefficient suggests that a 40 million MDL investment projects requires additional recurrent spending each year o f about 1.2 million MDL. Similarly, there are no efforts to explicitly identify the co-financing requirements o f externally-financed investment. Such efforts, when introduced, may promote greater predictability and transparency inthe budget process. 17 Christian Von Hirschhausen (1999) "What Infrastructure Policies for Post-Socialist Eastern Europe? Lessons from the Public Investment Programmes (PIP) in the Baltic Countries," Europe-Asia Studies, Vol. 51, No. 3: 4 17-432. Sanjay Pradhan (1997) "Evaluating Public Spending," World Bank Discussion Paper No. 323. 29 Box 4: A Frameworkfor Public Investment The role o f Government in investments i s essentially about the rationale for state intervention in the national economy and the appropriate instrument for government intervention. There are two fundamental rationales for government interventions: market failures (and thus the need to provide public goods and services that promote social and economic development) and redistribution (and thus the need to protect the poor and vulnerable and promote greater equity). While there may be a clear rationale for government intervention, however, the government need not be the sole provide of public services and goods such as basic infrastructure, education and health services. Many examples o f successful government intervention involved partnerships with private providers and with markets, more generally. These considerations lead to a decision tree illustrated below. intervention? Marketfailure, redishibution No rationale There is arationale I What is the right instrument?I I I I I i Outsourcing Subsidy Regulation Public provision -7 I What are the fiscal costs? The challenge for any government i s thus to know when to intervene, when not to intervene, and how best to intervene. Source:World Bank(2001) Russia: TowardsImproving theEficiency of Public Investment Expenditures. 'able 2.2: R-Coefficients:World BankProjects Number Average of r-coefticient 2.37 Third, there are no systematic efforts to Projects assess expected future revenues (e.g., user fees) Agriculture 22 0.047 from investment projects. Where investment Education 17 0.074 projects may be largely donor-driven, making Energy 14 0.013 explicit the net long-term financing cost (and Environment 12 0.017 benefits) o f donor-financed investments may Health 15 0.030 clarify whether these investments are indeed Telecommunications 3 0.003 Transportation 15 0.025 consistent with the government's priorities and UrbanDevelopment 11 0.017 fiscal constraints. Water supply 12 0.044 sanitation 2.38 Clari5 responsibilities and lead agencies. Average (unweighted) 123 0.030 As a critical component o f the budget planning Average (weighted) 123 0.035 process, the primary responsibility for investment Source: Hood and other (2002) planning and budgeting should rest within the Ministryo f Finance. In addition, realistic medium- 30 term planning and budgeting for the investment budget also needs to take place within the MTEF process, a process overseen by the Ministryo f Finance. Furthermore, for fully taking into account the recurrent cost implications o f investment decisions, investment budget processes built around single lead agency would make sense, and the Ministry o f Finance again has a comparative advantage inassessing the future recurrent budget requirement o f new investments. 2.39 However, as currently implemented in Moldova, investment planning's lead agency i s not defined clearly. For example, the Ministryo f Economy recently took the lead inpreparing the first MTEF technical annex on the investment budget. The Government needs to designate the lead agency in investment planning and budgeting, and define more clearly the supporting role o f other agencies and ministries. One possible arrangement would be for the Ministry o f Economy to develop an overall strategy for public investment and identify the priority areas for public investment, while the Ministry o f Finance i s assigned to oversee the public investment management and financing procedures within the MTEF framework. l9 2.40 Efforts to integrate externally-jinanced investments in mainstream government monitoring are welcome. The steps taken to integrate infomation on the budget execution o f externally-financed investment into the treasury system should be sustained. The Government should also explore opportunities for improving budget execution and monitoring processes, such as the ex post evaluation o f investments, which can then feed back into more realistic investment budget planningprocesses. IV. Conclusions 2.41 Spending on some sub-sectors o f the infrastructure remains low. For upgrading infiastructure and sustaining economic growth, Moldova needs to invest more. Deficiencies in public infrastructure represent a considerable constraint to economic growth inMoldova. 2.42 Many improvements have been carried out in recent years in the budget process in general and also for the investment budget process more specifically. Despite these gains, some weaknesses exist in the investment budget management system. The following are the main policy recommendations: Align budget allocations with strategic priorities as articulated in the EGPRS. A greater consistency between stated priorities and actual budget allocations i s needed. While the MTEF has introduced greater budget discipline, the MTEF process by itself cannot be expected to allocate resources strategically A few difficult political choices will have to be made including clarifying the Government's main priorities in the context o f multiple planning documents and trimming an ambitious list of priorities to match the Government's capabilities and resources. The framework described in Box 3 may be useful for clarifying the rationale for government intervention and the appropriate instrument for such intervention. Introduce systematic project screening and project selection to ensure the more eficient use of public resources. Project screening can draw from cost-benefit analysis, environmental impact analysis, or poverty impact analysis, as appropriate. Not every project needs to be subjected to rigorous analysis, and not all projects lend ''We thank Andrew Birdand Simon Groomfor raisingthesepoints. 31 themselves to cost-benefit analysis and cross-sectoral comparisons. An emphasis on the identification stage, clarifying a project's concept and assessing its broad consistency with the Government's objectives may go a long way toward ensuring better use o fresources. ClarifL the lead agency in managing the investment budget. There are persistent overlaps in the functions o f selected units within the Ministry o f Economy and Ministry of Finance. Given current arrangements, institutional mandates, and technical capacity in Moldova, there are benefits to clearly designating the Ministry o f Finance as the lead agency in investment programming and management while letting the Ministry o f Economy take the lead in develop an overall strategy for public investment and identifying priority areas. Calculate the forward fiscal costs of investment projects. The required recurrent spending o f new investments may be substantial. These should be reflected in the calculations o f costs and benefits o f investment projects, ina transparent manner. Consider gradually setting up the institutions for ex post evaluation and monitoring. Lessons o f implementation, when properly documented, can then feed into subsequent investment planning and management. 32 3. SAVING THE ROAD NETWORK A. THEDETERIORATIONOFMOLDOVA'S ASSETVALUE ROAD 3.1 Moldova has a road network totaling about 16.800 km, o f which 22% are classified as National Roads and the remainder of 78 % as Local Roads. Considering the size of the country and its population, Moldova's road network size i s roughly adequate, with little or no need for expansion. It has been built up gradually over the past seven decades through massive public investments. If Moldova had no roads and today's network needed to be constructed from scratch, it would cost at least US$ 13 billion. This figure shows the huge magnitude o f investment and sacrifice which was made inthe past inorder to be able to reap the benefits from today's road network. 3.2 However, Moldova's road network i s not in good condition. About 58 percent of National roads are classified as poor. For local roads the situation i s worse, with 75 percent estimated to be in poor condition. Only 10 percent of National roads and about 5 percent of Localroads are ingood condition. For any country, and especially for a European country, this i s a catastrophic situation. It i s the direct result o f severe neglect of the roadnetwork, mostly during the past 15 years. Very little resources have been spent for road maintenance and rehabilitation since 1991. About 400 kmof formerly pavedroads have lost their pavement and have revertedto unpaved gravel or earthroads. 3.3 The lack of roadmaintenance and rehabilitation inthe past 15 years has ledto a massive physical deterioration and therefore to a heavy loss o f road network asset value. The present asset value o f the Moldovan road network i s only about US$ 8 billion; instead of the US$ 12 billion it would be if the network was well-maintained. This means that the loss o f road network asset value which resultedfrom insufficient maintenance andrehabilitation i s a shocking US$ 3.6 billion, equivalent to 1.4 times the entire GDP of Moldova. B. "SAVINGMONEY'' BYNOTMAINTAININGROADS? 3.4 The severe reductioninGovernment spending for roads after 1990were imposedinlarge part by the dramatic collapse of economic activity in Moldova, in the aftermath o f the disintegration o f the Soviet Union. On the other hand, in the political decision-making on Government spending over the past years, roads were clearly not a priority. Road spending thus became a victim o f a true lack of resources and of policy decisions alike. Whatever the reasons were, it is certain that decision-makers felt that road maintenance expenditures where o f lesser importance and needed funding for roads was redirected to meet other priorities. How much money should the Government have been spending on roads in order for them to be maintained adequately? Table 3.1 shows the difference (year by year) between the amounts which would have beenneededinorder to maintainthe road network, and the actual amounts spent. The result i s that the total amount "redirected" during the period 1991 - 2006 amounted to about US$ 1.47 billion. This calculation i s based for one part on the actual expenditures presented in the tables for the period 1997 - 2006, and on rough estimates made by the Bank for the previous period between 1991 and 1996. While Moldova redirectedan amount of about US$ 1.47 billion to other priorities, the roadnetwork gradually deteriorated, and with it the RoadNetworkAsset Value. 33 3.5 As shown earlier, the road asset value deterioration between 1991 and 2006 was in the order o f US$ 3.63 billion. It i s thus simple to conclude that the Net Loss to Moldova was in the order o f US$ 2.16 billion, which i s equivalent to about 72 percent o f Moldova's GDP o f the year 2005. For every one $ redirected inroad spending, Moldova lost about $ 2.47 inroad asset value. 3.6 The loss o f road asset value i s however only one o f the consequences o f the deficient maintenance o f Moldova's roads. While the roads were gradually deteriorating, the cost of operating vehicles on those roads also went up. Badroads shorten the life o f cars and trucks, and impose more repairs, spare parts, fuel, etc. An estimate o f the additional road user costs imposed by the deteriorating road condition has been carried out and i s shown on Table 3.2. The table compares how much money road users would have spent in Vehicle Operating Costs (VOC) between 1991 and 2006 if the road had been well-maintained, with the actual situation o f higher VOC due to the steadily deteriorating road network. The result i s that due to deficient road maintenance, road users in Moldova have spend about US$ 1.8 billion or 30 percent more than they would have spent if the road network had been maintained adequately. This amount i s equivalent to about 60 percent o f Moldova's GDP o f 2005. By reducing competitiveness o f Moldova business, higher VOC also undermine private sector investment and, hence, undermine future growth prospects. 3.7 Insummary, for every one US$ "saved" bynot maintainingthe roads adequately between 1991 and 2006, the induced loss was about US$ 3.70, composed o f (i) additional Vehicle the Operating Costs road users had to pay o f US$ 1.23, plus the deterioration o f road network asset value o f US$2.47. c. ECONOMICEFFECTS OFMOLDOVA'S DETERIORATEDROADS 3.8 The effects o f Moldova's deteriorated road network are described well in the transport section o fthe World Bank's (2003) Moldova Trade Diagnostic Study: 0 Located strategically, Moldova i s becoming increasingly important as a future border between the European Union and Eastern Europe. Moldova could become a transport hub for the region, were it not for its deteriorating infrastructure and highlogistics costs in comparison to neighboring countries. Moldova's economy i s largely based on agriculture and agro-industry, which depend on a solid road infrastructure, well functioning road transport and adequate storage facilities. Many farmers are unable to commercialize their production due to poor road access between domestic market places, storage facilities and the settlements. 0 The poor quality o f the road infrastructure severely affects mobility o f goods and people especially inrural and poor areas. It is also a significant impediment to expanded trade by imposing additional costs on exporters and importers: over 30 per cent o f responding firms in ExporterAmporter survey indicated that Moldovan transport infrastructure had become more o f an impediment to their operations duringthe past five years. Transport costs can affect economic growth in several ways. First, higher transport costs reduce rents earned from the exports o f primary products, lowering an economy's savings available for investments. They pushup import prices o f capital goods, directly reducing real investments. 34 sk # al, countries axth higher tra de. Those countriesa as, O f access ~ a r ~ and~ social and a d ~ i n i s ~ a services, or ro transport goods produced by the e s ~ i ~ e household to locaf ~ a r k e for t ~ s e l l i ~them there. If the road network and ore s ~ e c ~ ~ ~ a ~ l y i ~ access roads to villages are in bad condition, transport operators charge h~gher~ r a n ~ prices p o ~ than they o t h e r ~ ~ ~would. This af'fects poor househol~smuch more. than non-~oor~ o ~ s c h o ~ d s . s e A househoId survey carried out in 2003 tn rural c o ~ m u n i ~ j in$Moldova showed that poor roads e and ~ a n s ~ o ~ aw i~s~considered the worst prclblcm faccd by c o ~ ~ ~ ~ nand i e s o i i i ~~illages'~. 3.10 The single largest ~ o v e ~ n revenue~ source from the Road Sector has ~ a d ~ t ~ ~ n a l ~ y ~ e n been the^^^^ excise fax. In2005 the total a ~ o u collected ~ ~ o u the n ~ g hfuel excise tax was 415.2 ~ ~ lLeil (US$ ~32.95 ~ ~ f l i butnonly~a relativcly srtiall percentage of the fuel excise tax 1s ~ n o ~ a c ~ u a ~al~ocatedto the road sector ( I S percent in 2005). r ~ y ~ ~ ~~ ~ ~oroad~~ s .~i#have ~~ ~ o ~ ~ risen sharply over tfic fast few years because of the increase in the number of vehicles ~ ~ p o ~ e d ~ in particular of passenger .8 rn1llton Le1 (US$ 3 3 , u mllllon), which \vas that year. However, none of art duties tor roa Othcr ~ ~ ~ ~ revenues from the roadsec the road sector ~ ~ theoRoad~ Fund, ~ F@rcJ.I: ~ h ~ ~ w w ~ S ~ ~ c lto 1n18~ ~ lLe1(US$o9.36 million). u ~ ~ ~ n sector can be separa~edinto three c ~ t c ~ o r ~ e ~ . First,there is s~endIngrough the RoadFund 45% which is entirely used for the ~ a ~ n t e n ~ ~ c e and repair of the network of ~ a t ~ oroads, ~ a l Second, there are a ~ ~ o c a ~ ~ ofor s the n n ~ a ~ n t e n aofnlocal roads. Finally, there are ~ ~ ~ o c aforiroad ~ n v e $ ~although~In~actual ~ ~ ~ o n ~ e ~ , practice these havc been almost n o n ~ ~ x iduring ~ s ~ ~ nthe past years. 3.12 ~ o l d o ~ atotal annuat road scctor spcnding varied between a low US!$ 6.45 rnifhon in ' s 2001 and a ~ a ~ i n of about~USS 23 million in 1997, The tendency since 2001 i s towards an ~ u n increase in road s ~ e n d i n ~In . 2005, thc latest year for which actual figures are a~ai~able,the 20~ r o - Project~ Design: Povcrty and Social Analysisrs far ~ ~ o ~ Road Transport ~ e#t r u e FluProject, ~ ~ s r ~ ~ ~~ ~ ~ Final fteport, ECQRYS, July2003 37 Government spent an amount o f US$ 14.44 million in the road sector, which i s equivalent to 19 percent o f the resources it has collected from road users through the various forms o f taxation, fees and duties. 3.13 It is estimated that because o f the poor overall condition o f Moldova's roads, a sustained spending level o f approximately US$ 135 million per year would be necessary to fully stop the degradation o f Moldova's road network and to carry out backlog maintenance works. This necessary annual amount would correspond (i) 1.1percent o f the road network asset value, and to (ii) about4percentoftheGDP. Forafullrecovery ofthenetwork, therequiredannual to spending.levels would be even higher, because o f the need to carry out heavy rehabilitation works on parts o f the network. 3.14 H o w do the specific GDP-related indicators for Moldova compare to generally accepted benchmark figures for those countries which do maintain their road networks adequately? Statistics compiled by the International Road Federation (IRF) show that industrialized countries, which in most cases maintain their roads adequately, (i) spend on average about 1.6 percent o f their GDP on roads, and (ii) spend between 2 and 2.5 percent o f the road network asset value on road maintenance and rehabilitation. Conditions in Moldova are however very different from those inindustrializedcountries, with much lower indicators for overall wealth expressedthrough GDP per capita, traffic level on roads, funds collected by the Government from road users, etc. Inaddition, Moldovaneeds to recover from many years of severe under-spending on roads. For the preparation o f a future financing strategy for Moldova's roads, it would thus be o f little help to simply apply the benchmark figures o f developed countries. 3.15 The analysis o f the present situation o f Moldova's road infrastructure sector reveals a situation which can only be described as an acute crisis. The question i s how to overcome this Moldova's entire road network to an "as new" condition would require about US$ 4.3 billion - - crisis and bring about a turnaround of the road network. A full and immediate recovery o f which i s o f course not realistic, because Moldova i s unable to mobilize and afford such a level o f financing. A more realistic approach would be to gradually rehabilitate destroyed road sections over a medium-term (10-year) period, while at the same time maintaining all other roads as well as possible. It i s almost certain that Moldova will not only need to raise more domestic funding through road user fees, but also substantial external loans, credits or grants in order to finance the needed investments. In any case, the recovery o f Moldova's Road infrastructure sector will require at least two elements, which are: (i)financing plan to mobilize additional resources, and a (ii) institutional changes. 3.16 Any financing plan to be defined with the objective o f improving Moldova's road network will first have to satis6 the annual needs for road maintenance (routine and periodic). It i s estimated that around US$ 95 million which would need to be allocated every year to this task just to avoid further deterioration o fthe roadnetwork. Moreresources will be requiredto finance backlog maintenance works21 and heavy rehabilitation. To achieve a clear improvement in the road network within a 10-year period, it i s estimated that some $160 million a year will need to be spent. In 2005, the Government allocated $14.4 million for road maintenance and rehabilitation. This implies that there i s nearly a $146 million shortfall between the resources needed and those actually allocated and spent for roads in Moldova. The most recent MTEF 21 Backlog Maintenance works are those works carried out to offset the lack o f maintenance in earlier years. Backlog Maintenance mostly consists o f spot repairs and light rehabilitation works. If Backlog maintenance i s not carried out within a reasonable time, road conditions deteriorate so much that heavy rehabilitation becomes necessary. 38 foresees only minor increases inbudget allocations for road maintenance, of about 20 million Lei per year, which i s equivalent to a rate of growth of about 12 percent. It i s clear that the planned level ofroad expenditureswould result ina further decline of the roadnetwork. 3.17 A general principle for road financing is that road users should pay at least for the long- term maintenance cost for roads. This should also be the guidingprinciple for any road financing plan for Moldova. One possible source of financing with the best potential and lowest collection cost would be a fuel levy included in the price of gasoline and diesel, to be collected directly by the RoadMaintenance Fund. Box 5: TheExperience with Second Generation Road Funds A large body of evidence has emerged over the past several decades demonstrating clearly that in road systems across countries, road maintenance i s often underfunded.Because road users may be identifiedand because they can and are levied selected taxes (e.g., gasoline tax), road funds were set up in many countries, earmarking arrangements for road maintenance or construction. These arrangements were typically opposed by macroeconomists, on grounds that such arrangements can undermine fiscal flexibility and efficient management o f government resources. Inaddition, inpractice, many o f these first road funds failed to guarantee an appropriate level o f road maintenance or construction. Inresponse, over the last decade or so, the so-called "second-generation roadfunds" have emerged, with an emphasis on creating a transparent, predictable, and accountable framework for road maintenance. In principle, a second-generation road fimd involves roadusers inthe management o f roads, secures sufficient financing for road maintenance each year, ensures all parties are aware o f their responsibilities; establishes clear accountability rules; and identifies charging instruments that are clearly related to road use, separable from other taxes, and administratively simpleto administer. This promising approach can serve as an efficient means of managing road systems. Macroeconomists also welcome second-generation funds, "provided the right conditions are met," including sufficient capacity, management by a board free from producer and contractor interests, and others (Potter 1997). However, because second-generation road funds are relatively new, only preliminary appraisals and evaluations currently exist. A recent assessmento f road funds inAfrica (Gwilliam and Kumar 2003) suggests that such funds have not underminedfiscal flexibility and have infact improved roadmanagement. Nonetheless, the assessment strongly recommends keeping road funds under review, such as through technical and efficiency audits. Sources: Gwilliam, Ken and Ajay Kumar, 2003, "How Effective Are Second Generation Road Funds? A Preliminary Appraisal," World Bank Research Observer, Vol. 18, No. 1 (Spring): 113 128; Potter, Barry, 1997, "Dedicated Road Funds:A Preliminary View o f a World Bank Initiative," IMFWorking Paper 97/7; Schlavo Campo, Salvatore and Daniel Tommasi, 1999, Managing Government Expenditure (Manila: Asian Development Bank). 3-18 Given sufficient resources, at the end of the 1O-year period would be a much improved overall condition of Moldova's road network, which would be able to sustain a growing economy. About 60 percent o f the entire road network and about 85 percent of National Roads would be in fair or good condition. The remaining portion of poor roads would be on the less traveled part of the network. Most roads would be in a maintainable condition, and the trend would be towards further gradual improvement. Average Vehicle operation costs (VOC) to be borne by road users would be significantly reduced. Most importantly, road maintenance financing would be assured at an acceptable level and could be entirely financed by road users through road user charges. 39 E. CONCLUSION 3.19 Inaddition to the need to greatly augment funding for roads, there are two important institutional changes which would appear necessary. These are (i) the reform o f the existing road maintenance financing mechanism, and (ii) the contracting modalities for road maintenance works. 3.20 One lesson which can clearly be learned from Moldova's past experience with road funding is that there is a need to separate o f financing for road maintenance (through earmarked resources) from those for new road investments (through the general budget). Duringthe past 15 years, successive Governments and the Parliament have not been able to agree on a level o f road maintenance expenditures anywhere near the required levels. This failure o f political will suggests that there i s a need to consider replacing the existing Road Fund with a so-called "Second-generation RoadMaintenance Fund(RMF)". 3.21 While the lack o f funding for road maintenance i s clearly and by far the biggest problem in,Moldova's road sector today, solving the fimding problem alone would be insufficient. A second problem i s the present system o f awarding and supervising road maintenance contracts. This impact o f this problem is less apparent today, because o f the very low level of resources for road maintenance, but will certainly become very important once there i s more money available for road maintenance in Moldova. The present system o f sole-source contracts with the state- controlled local district road maintenance firms has several deficiencies which may result in a lower efficiency in the use o f resources. It would therefore seem that in parallel with the reform o f the funding mechanism for road maintenance, the introduction o f competitive bidding for all road maintenance contracts could bringimportant benefits to Moldova. 40 4. MANAGING CIVIL SERVICE PAY REFORM A. INTRODUCTION 4.1 Moldova's public sector wage bill in percent o f total government expenditures i s among the highest in the region, including several advanced OECD economies. The public sector wage bill, however, is programmed to continue rising through 2009 to keep pace with rising overall wage levels in the economy. Inrecent years, the stable level o f public employment and rising level o f public wages has made it difficult to increase other spending, especially operation and maintenance expenditures as well as capital expenditures. As public sector wages are projected to rise and outpace economic growth in the short-term, the rising wage bill will likely create significant fiscal and macroeconomic pressures and continue crowding out essential spending. 4.2 However, the design o f the public sector wage system also has significant implications on the quality and efficient function o f government. Despite the recent, rapid increase in public sector wages, compensation levels in the public sector are still very low when compared with the private sector. In addition to low pay, the compression o f salaries in the civil service remains quite high. The composition o f public sector compensations i s also very complex and lacking in transparency. Recognizing this, the Government has articulated a program ("Modernization o f the Country - Welfare o f the People") that envisages a system o f remunerationthat i s able to attract and retain qualified personnel, fosters greater professionalisminthe civil service, and strengthen the efficient functioning o f government. 4.3 The fundamental challenge then i s to strike the right balance between a public wage bill that i s fiscally affordable and a remunerationsystem that i s able to attract qualified and motivated workers to the public sector. This chapter reviews recent developments inpublic sector pay and identifies opportunities for reducing complexity and promoting greater transparency in the pay system. It focuses on the civil service remuneration system rather than on the broader public sector remuneration system that covers public sector workers in the education, health, and other sectors. While the civil service accounts for a relatively small share o f total public sector employment, the focus on civil service i s pay i s driven in large part by the ongoing central public administration reformprogram. Inaddition, sector-specific employment issues are covered inthe respective chapters.22 Finally, because the civil service remuneration system serves as a benchmark for the rest o f the public sector, the issues discussed in the chapter are effectively crosscutting. B. RECENTDEVELOPMENTS INPUBLIC SECTORPAY 4.4 Public sector salaries consume close to a third o f total government resources. In2005, the wage bill accounted for some 29 percent o f total government spending, representing about 10.8 percent o f the GDP. Across sectors, there are some variations inthe share o f wages and salaries in 22For example, the education sector employs 71 percent o f total public sector employment. By any international comparison, the education sector i s overstaffed. The structural reforms required in the education sector to rationalize government expenditures are discussed in Chapter 6. 41 4.6 i t s a rcsuIt ofrapid wage increases, there has been a su at growth in public sector . wages and all othcr ions of wage @OU'tft In by 2009+" The pricc of ry o f Financc. It may mean eases in ~ a ~ ~ ~ an n a n ~ e ~ impacts the level of p~bIicsector p ~ r f ~ ~ ~~ ~n cc ~~e.salaries~ will requisc somc a s i ~ r a ~ ~ o ~ a ~of~~~ a~ t i~~ nl o y ~ ~ ~ i ~ . Figure4.1: Wages and Salaries in Percent of Total ~ o ~~ x ~~ ~ ~~ d2003t u r~e ~ ~ e ~ t . . .- ..... . __ ..-..-_ _- . - . ...- -... 42 4.7 Moldova's public sector wage bill i s currently among the highest inthe region. Figure 4.1 compares Moldova's personnel expenditures with those o f selected OECD and NIS comparator countries. On average, wages and salaries in percent o f total government expenditure amount to about 21 percent, nearly a third lower than that o f Moldova.24Moldova's wage bill was larger than all other countries, except Denmark. C. CIWL SERVICE REMUNERATION SYSTEM 4.8 Rapid wage increases in the public sector were driven, in part, by the need to keep pace with rising aggregate wage increases. Indeed, while public sector wages have increased sharply in recent years, workers in non-agricultural sectors o f the economy have experienced even higher increases in averages salaries (Figure 4.2). Compensation levels inthe public sector are still very low when compared with the private sector. While civil servants are underpaid compared to the private sector at all levels, the relative under compensation o f middle and senior management i s quite extreme. At the lower end o f senior management, for example, the private sector compensation i s more than six times the civil servant ~alary.~'In addition to low pay, the compression o f salaries inthe civil service remains quite high. As a result o f l o w and compressed compensation levels, the public sector in Moldova has had a very difficult time in attracting and retaining skilled employees. Figure4.2: Index of Monthly Salary (2002=100) I "V Industry I40- 120- Public administratio1 100- 80 4 I 2002 2003 2004 Sources: National authorities and World Bank staff estimates 4.9 Recognizing the need to improve the remuneration o f public sector employees to attract and retain highly qualified staff, the Government initiated a functional review of the central public administration in 2005, covering 28 ministries and other central government agencies. The broad objective was to identify opportunities for reducing non-essential staffing, streamlining government processes, and promoting public sector productivity. The functional review should 24For cross-country comparability, the chart uses 2003 data (which inthe case of Moldova includes the wages of all public sector workers, including public health employees). 25In 2004, over 59 percent of the public servants in the public administration earned between 750 and 2000 M DL. A third of public servants received between 100 and 750 MDL, while only 8 per cent - those in senior management - earned over MDL 2000. In comparison, about a third of financial sector employees in the private sector earned as much, or more than, senior public sector managers. 43 determine if staff redundancy could provide sufficient savings so as to allow for improvement of remuneration conditions in the reformed public sector. The early indications reveal functional overlaps, inefficiencies, and duplication o f functions among subordinate agencies. Apart from inefficient use o f public resources, maintaining fragmented and weakly coordination organizations with the very similar functions create confusion among consumers and economic agents, and dilute accountability for policy implementation. Therefore, institutional restructuring could (i) produce savings and (ii) improve technical efficiency o f the public administration. 4.10 The initial functional review revealed that elimination o f functional overlaps and inconsistencies in functional allocation could lead modest personnel cuts. However, personnel cuts have not been fully implemented inthe wake o f the second stage o f the functional review in 2006. The analysis o f the second and broader in scope functional review, covering subordinated and deconcentrated services, i s expected to reveal the relevance o f functions, budgets and staffing in five major sectors. Redundancy is expected to be identified through better alignment of mandates, functions, organizational structures and staffing. 4.11 In addition, the World Bank with the financial support from the Department for International Development o f the United Kingdom (DFID) recently conducted a study o f civil service pay. The study reviewed weaknesses in the present personnel management system and deficiencies in the pay structure. The sheer difficulty o f obtaining data on salaries during the course o f the study suggests serious inadequacies in the payroll management system. More broadly, the absence o f centralized timely and accurate information on actual individual earnings o f public servants i s not conducive to effective public expenditure management. With the public wage bill representing a third o f total government expenditures, the absence o f efficient monitoring and control mechanisms in public sector pay weakens overall government fiscal control. 4.12 The review o f civil service pay indicates the following: 4.13 First, the structure o f civil service pay in Moldova is complex and, as a result, not sufficiently transparent. The total (take home) remuneration o f civil servants consists o f three components: basic pay related to grade, pay supplements, and performance-related bonus. Prior to the enforcement o f the new public sector remunerationlaw, the basic pay accounted for less than a third o f total pay, on average. The variable component o f total pay, consisting o f supplements and bonuses, was large, not transparent, and unpredictable. The unpredictability i s driven by the lack o f formal criteria for performance bonuses, substantial managerial discretion in awarding bonuses, the absence o f systematic prior or post review o f salary levels, and budget contingency. However, the situation has been improved somewhat by the new law, mainly by raising basic pay levels to about 50 percent o ftotal civil service pay, as indicated a pay analysis o f some 900 posts. 4.14 Second, considerable discretion, the lack o f uniform criteria for setting the wage bill for budget organizations, and the lack o f independent oversight provide little assurance that civil servants receive equal pay for equal work. Incontrast to international (OECD) practice, where total pay closely reflects the position in the civil service, and where the base pay i s consistent with systematic assessment o f the value o fjobs based onjob descriptions, inMoldova, individual qualities o f employees weigh heavily in the determination o f total pay. The supplements for seniority, rank, and the length o f public service, along with discretionary monthly bonuses represent a substantial share o f the total pay. This effectively distorts the link betweenjob value and total pay. OECD experience suggests that the base pay, related to grade, should reach 80-90 percent o f total pay. Furthermore, the underlying principle in international practice is the matching o f civil service grade with job complexity, as specified in the job description, rather 44 than personal a t ~ ~ bof ~ cjobhol ~ the s related 'to the formal a s ~ e ~ of skillsnand p e r ~ o ~ ~thus ~distor~~n~incent~~es s ~ ~ ~ ~ n e , the for p r o ~ e s s ~growth,~~ o ~ ~ ~ f ~ to l l as sum^ i ~ more respons~bili~~ ~ n d ~ a k i n g that are ~ ~ ~ ~ s and tasks more complex. The pay system therefore faik to pro~~de~ c efor~ im~ro~~ingthe ~ ~ u of a i t ~ ~ ~ ~ ~ ~ ~ the public a d m ~ ~ i ~ s ~ a ~ i o ~ . 4.15 In ~ ~ o ~ djob~~al. a s ~ ~IS ~onlyain~di~ore~crelated to the job c ~ n ~ e~'os~~i#ns o ~ ~ ~ ~ n ~ , in ~ o l d o ~ a n ser~iccare catcgortzed by rank and class. There arc three ranks, each sub~i~ridcd civil ile classes reflect ~ n d ~ ~ i d u a ~ s 1n two ccneaf frcahon, pay andjob ~ o n ~ ~ n ~ are not related, e r n ~ l o y ~feels d i ~ s ~ t with~the dpay and are not ~ ~ ~ i to~ be more d ~ ~ s ~ a ~ e ~ r o ~ u cIn iltght e ~~ n t e ~ a ~experience, ~ ~ # l d oshould consrder i n ~ ~ ~ uofc a~ i o ~ ~ ~ ~ of i ~ ~ a l ~ a u n i ~ posi~~on o ~ rad^^^ system ~ ~ ntok s~andardiz~djob d c s ~ ~ that,~in~turn,~d~e ts e ~ ~ ~ e ~ d p ~ base pay. A ~e~hodo~ogyassessing job values based on objective and ~easura~~e of criteria shouldbe ~ ~ ~ rto~support this eStandardiza~iQn. ~ o ~ d 4.16 Third, salaries in the civit service r ~ m asi ~ ~ ~ i com~ressed,The deco~p~e~sion ~ ~ a ~ ~ ~ y ratio, calculated by div o~~ression ratio i e ~ c ~ ~ ~ e ~ a t ~ ~ n and~ ~tn ~ ~ n the civil senkc. 45 This has adverse effects on the capacity building prospects of the public administration. In addition, an important consequence o f salary compression at lower levels i s the inability o f senior public sector managers to rely fully on staff for high quality work. As a result, public sector managers inMoldova do much o f what morejunior staff would ordinarily do inOECD and other comparator countries. Little delegation o f responsibility takes place in Moldovan public administration. This triggers a vicious circle o f overwork by senior civil servants, the lack o f attention to more strategic issues at senior levels and dissatisfaction with routine tasks at junior levels, and constrained opportunities for knowledge and experience transfer down the hierarchy. 4.18 Finally, performance bonuses as currently implemented inthe civil service pay system do not generate better performance. Inpractice, additional bonuses can be paid monthly based on managerial assessment (mostly undocumented) on civil servant's performance over the past month. There are no formal criteria in place regarding performance appraisal, or the size and frequency o f awarding performance bonuses. The funding for such allowance-type bonuses i s from savings drawn from vacancies within the organization. 4.19 Managers exercise significant discretion in awarding bonuses. Bonuses, which are not linked to formal performance appraisal, are expected as a more or less permanent pay supplement by all civil servants regardless o f actual performance. This sense o f entitlement creates a serious disincentive to improve performance. In contrast, robust performance management systems in other countries ensure that performance i s planned, monitored, formally assessed and rewarded, usually on annual basis. In Moldova, a process called "attestation" comes close to what may be considered performance appraisal. It i s conducted once every three years to decide on the promotion to the next class or rank. The new class designation then leads a permanent supplement to the base pay, independent o f actual performance 4.20 It is o f course expected that efficient performance management systems take time to develop. They require clear rules, effective personnel management systems, internal controls and some degree o f central oversight by the civil service managing body. Based on experience in OECD countries, the role o f performance management should not be overstated. Transparent pay systems, proper job grading, and clear rules o f career progression linked to performance that create the conditions for attracting and retaining civil servants, rather than a system o f ad hoc bonuses exercised without clear rules and oversights. The priority for Moldova should be to improve the current civil service pay system and the next section identifies opportunities for doing so. D. OPTIONSFORCIVIL SERVICE PAYREFORM 4.21 Insum, the current civil service pay system is complex, not transparent, uncompetitive, and does not provide the right incentives for greater public sector productivity. The Government has problems assessing reliable individual pay data for the analysis and design o f a better pay system. The centralized payroll management, which in principle should ensure horizontal compatibility and transparency o f pay, cannot be implemented due to the design o f the current system, which allows significant discretion. Job descriptions are not structured around clearly identifiable and measurable criteria linked with the job value. The culture o f effective performance management system i s absent. Multiple supplements and bonuses do not provide right incentives for professional growth and improved performance. Therefore, without radical changes o f the civil service pay system, it i s unlikely that the quality o f the public administration can be improved. 46 4.22 To make civil service pay motivating, it i s not sufficient to increase salary levels. It i s necessary to achieve a direct link between the pay level on the one hand and employee effort, qualifications, and outcomes, on the other hand. Increasing wage dispersion (or pay decompression) and consolidating multiple salary supplements into the base pay while keeping constant the overall wage bill, will be the key elements of a good reform program. There are a number of steps that can be taken, following other countries that have successfully instituted effective remunerationsystems: 4.23 First, to realign the present pay and grading practice with a more effective civil service pay, an approach shouldbe developed that i s founded on the followingprinciples: regular guaranteed monthly pay i s based on the agreed pay for the job (or position), rather than i s being "assembled" from a number of pay elements which have little relevance to thejob; the pay system is designed around a clear job grading structure and recognized methodology and process for grading jobs; the job grading structure allows for a greater degree of career progression through promotion to higher non-managerialjobs; job grade-related pay ranges (or scales) provide for regular pay progression based on length of service and allow for accelerated progression to reflect individual performance while inthejob; includes limited discretion for additional performance-related bonuses within common fundingarrangementsacrossthe service; and at a later stage performance appraisal system i s developed to underpin the performance reward and career development process. 4.24 Second, because the design of a new pay system i s an extensive technical exercise, it should be undertaken with guidance of experienced specialist in designing pay systems. The recommended sequence ofpay systemreformi s describedbelow: A. Preparingjob descriptions for all typical positions within the civil service. The job descriptions could be createdby a trained group ofjob analysts, based on a structuredjob questionnaires, covering such elements as: jobpurpose; keyresponsibilities; main duties; responsibility for job organization, staff management and resources; 0 decision-making; contacts and representation; and 0 required knowledge, skills and experience. B. Designingjob evaluation system, which guides how to assess the value (or level) of certainjob factors. C. Evaluating sample jobs and grouping jobs according to the value and weight o fjob factors to establish parameter for a grading structure. D. Determiningparameters ofthe grading structure: requirednumber of grades, relevant number of points within the grade to allow for a gradual pay progression and for distinguishing between concrete jobs under the same title; the appropriate size o f steps 47 from one point to another within the grade; the degree o f overlap o f the adjacent grades, etc. E. Slotting intypicaljobs into the grading structure. F. Conducting a pay determination simulation exercise to determine pay levels corresponding to typicaljobs within the realistic civil service wage bill volume. G. Conducting the reclassification o fjobs across the public administration and creation o f the job database. Further validation o f the pay grading system through a simulation exercise. H. Drafting a normative act determining the pay system and the rules o f career progression linked to performance appraisal. I.Trainingofpersonnelmanagersinassessingjobs. J. Developingrules andprocedures for performance management. 4.25 The proposed approach will require abandoning the old tradition o f length o f service supplements as not the length o f service but competences to perform a job will determine what position a civil service can take and what pay to receive. In the new system, monthly bonus payments will be eliminated. Instead, a formal annual performance appraisal will lead to a revision o f civil servant's grade and pay level. Currently administered pay supplements for "high intensity o f job" should cease, but intensive conditions o f a job will be reflected in the job description and result in a higher base pay level (e.g., assigning a higher point within the grade). Thus, the suggested measures will leadto better alignment o fpay withjobs and the realization o f the principle: equal pay for equal job. When the salary will take account o f factors important to the job, rather than some personal attributes, the pay system will be perceived as more just. Moreover, by rewarding competences and performance, the new pay system will provide incentives for professional development, initiative andbetter performance. 4.26 The amount o f work involved in developing a new pay system i s considerable. It may probably require at least two years in order to conduct all the necessary steps to develop and implement analytically justifiable pay system. It i s also important that personnel management system i s being further developed to support effective execution o fjob grading, career planning, performance appraisal and career management withinthe new pay system 48 5. INCREASING THE EFFICIENCY AND QUALITY OF HEALTHCARE SERVICES A. INTRODUCTION 5.1 Health sector indicators deteriorated rapidly during the 1990s and public health care expenditures fell to only 2.9 percent o f GDP by 2000. In addition, health care providers were experiencing high levels o f debt which undermined the functioning o f the health system and created the conditions for increasing levels o f informal payments. While donor and bilateral financing partly eased the burden, satisfaction with the system was low and 33 percent o f all people that were illwere unable to access health services for a lack o f money. 5.2 At the end o f the 1990s Moldova embarked on a health sector reform which focused on strengthening primary health care (PHC) and restructuring the hospital sector. Notable progress has been observed in introducing health insurance, transforming hospitals from budgetary institutions to self-governing entities, and strengthening primary health care. Public health expenditures have recovered to 4.3 percent o f GDP by 2005 and per capita health spending figures are now at the highest levels experienced since independence. end o f the decade and beginning o f the new millennium. Health outcomes have improved - in 5.3 All these factors helped the system inreversingthe unfavorabletrends experienced at the terms o f a broad range o f health indicators, financial protection o f the population, responsiveness o f the system to the needs o f the population and their overall satisfaction. Between 2000 and 2005, for example, infantmortality rates decreased by 32 percent (from 18 infant deaths per 1000 live births to 12 in 2005). Over roughly the same period, maternal mortality decreased by 60 percent (from 43.9 maternal deaths per 100,000 live births in 2001 to 18.6 in2005). However, mortality and morbidity rates inMoldova are still far from the EUregional rates. Infant mortality rates (IMR) are 2.5 times higher in Moldova than in the EU, while maternal mortality rates are 4 times higher, TB incidence rates are 11times higher and HIV-AIDS incidence rates are 1.3 times higher. Significant efforts will be required ifMoldova i s to approach EUrates inthe mediumand long term. 5.4 Despite these significant achievements, the hospital sector continues to be burdened by overcapacity. Informal payments still constitute an important source o f revenue inthe system, and clinical practices have been only slightly modified to promote more cost-effective care and quality based initiatives. Further investment in the health care system will require additional reforms to improve the functioning and cost-effectiveness o f the system. Intra-sectoralchanges in expenditure allocation will be needed to achieve the fiscal space to build upon recent successes. This chapter concentrates on a number o f selected issues related to health sector performance, with an emphasis on improving the efficiency o f health expenditures. It lays out the strategic directions for health reform that can be integrated into the budget over the short and medium term. 49 TabXe 5.1: Canrp cnditures in 2005 161.3 4.7?4 Donors 259.3 7.6% Science and Innavaaon Mcdical education Totai Direct Health Expenditures 3,335. I Toral Health-Retated Expenditures(THE) 3,412.5 1000/0 50 nearly 2 tmcs largcr than formal co-payn~cn~~. Finally, with the low public shrtsr:of s p ~ n d onn ~ ~ drugs ~ p ~ bsector l ~ c spend in^ r ~ ~ r eonly~ 20 percent of total endin in^ on drugs), the p s ~ ~ $ purch~s~~ h ~ ~ ~ c e ur~pr~sen~sbulk of private sector e x p ~ ~ ~oni ~~ u r~~ ~ ~ of ~ i c a ~the s a ~ ~ costs are the major c o ~ ~ ~ ~tobthetlackr of ~ ~ n ~ p ~ ioat l~ ~ t i o ~ ~ ~ o ~ c u.orks at the rayon ~ a t ~ ~the~ bulkl of ys t~a f ~ ~1sgcantposed of Nurses (33 percent of total), ~ ~ l o ~ byI o ~ ~ ~ ~ a~tendantsand other workers (39 percent) and Doctors (1 8 ~ ~ r c Whilet by i~~ t e ~ a ~ j o ~ ~ ~ ~ n ~ cornpar~sons~r e ~ ~ n c r a t ~ othe sector 1s still low ~ in n a ~r n ~r ~a.age~~oflaround $60), real ~ ~ t e~ wages inthe sector have grown ~ a ~ i (42~percent between 2003 and 2005). d y 5.9 The loner share HI drug spending can be c~plainedby the r e c ~ acute shortage of ~ ~ ~ n ~ drugs at i ~ - p a f~a ~ i~l jn~most~~: ~ e ~ possessonly few very basic~centrallyt ~ ~ i r ~c ~drugsel d ~ ~ a ~ ~ 51 5.10 Moldova has made important progress in the hospital consolidation process by reducing the total number o f facilities from 265 hospitals in 1995 to 65 in 2002. The consolidation o f facilities has been a key step in reducing the fixed costs in the system. Despite this significant reduction in infrastructure, very little resources are devoted to infrastructure maintenance and equipment purchases. Most o f the remaining hospitals have exceeded their expected lifespan. While international parameters indicate that full depreciation o f a hospital i s between 25 to 33 years, the average age o f a typical Moldovan facility i s around 45 years. A recent survey o f health facilities in three rayons revealed that one-third o f the evaluated units had damaged roofs, doors and windows, one fourth had broken floors and almost halfhad damaged furniture. D. HEALTH SECTOR PERFORMANCE 5.11 The provision o f universal health insurance coverage is one o f the key objectives o f the Moldovan Health Insurance system. With health insurance premiums currently only covering one-fourth o f total public health expenditures, the financial sustainability o f the health system i s threatened by the weak participation rate among productive groups o f the population. As o f 2005, 74.8 percent o f the Moldovan population was covered by mandatory health insurance. While children under 18 years old and pensioners were reported fully covered by health insurance, less than half o f the 25 to 44 year age group was affiliated to the MHI. Among the 25 percent o f the population that i s uninsured, nearly 70 percent (or about 17 percent o f all workers) are formally employed (largely in the agricultural sector) with the remainder being unemployed. The larger cities or municipalities have much higher coverage rates due to the higher rate observed in Chisinau, where 85 percent o f the population is covered. Differences across the various Rayons are significant. The top 10 RayonsMunicipalities have on average 81 percent o f their population covered by health insurance, while the 10 RayonsMunicipalities with the lowest rates showed an average rate o f 63 percent. Increasing the participation o f payroll contributions in total NHIC revenues, thereby lessening dependence from budget transfers, is one o f the key challenges o f the system inthe coming years. 5.12 Access to Cure.The evidence provided by different sources suggests that access to health care services in Moldova i s satisfactory and that the MHI system has improved access o f the insured population. In Moldova, 97.1 percent o f households live within five km o f the nearest health facility, 93.5 percent o f households need less than an hour to reach a health service provider and 87 percent o f households are within five km and need less than one hour to reach primary health care facilities. The health insurance "removed or significantly reduced" barriers and allowed insured persons to have better access to care. Evidence from household surveys shows that insured people not only receive higher family doctor (FD) coverage (87 percent o f the insured are covered by FD services versus 48 percent o f the non-insured) but, as a result, they make more intensive use o f FD services (3.2 visits per year to FD versus 2.8 visits among non- insured). 5.13 Quality of Care. While access has improved, analysis o f quality o f health care indicators suggests that the Moldovan healthcare system performs poorly. In-patient mortality rates in Moldovan hospitals grew by 10 percent between 2002-2005. Coverage rates are very low in cholesterol measurement, mammography and flu vaccination, and moderate in blood pressure measurement. For all these quality indicators, Moldova lags behind international benchmarks. Over 80 percent o f the Moldovan population believes that the quality of the health care i s a problem and that the sector requires considerable changes to improve its performance. Some additional problems persist, such as the absence o f clinical management approaches (protocols, integral pathways, etc), an inadequate payment system to promote quality o f care and unmet demand for family doctor services inapproximately 30 percent o f the Rayons. 52 Table 5.2: Selected Quality and Health Care Indicators YOMeasured YOMeasuredBlood YOreceivedFlu YOMammography Cholesterol in Blood Pressure Vaccinein last 2 years in last 2 years Moldova 8.8 69 16.6 13.3 Benchmark 75 55-90 70 70 5.14 Resource Efficiency. While public health expenditures have increased significantly since 2000, the overall efficiency o f the sector has not improved correspondingly (see Table 5.3). Hospital efficiency can be measured from three perspectives. The first deals with productivity, as measured by the number o f discharges per doctor. In 2005, Moldovan hospitals reported an average 58 discharges per doctor with little improvement in the last few years. The second measure relates to resource utilization and average length o f stay (ALOS). The average occupancy rate in Moldovan hospitals was 66 percent in 2005, considered moderate to low by international standards. Some 44 percent o f hospitals have excess bed availability. While the average length o f stay has declined in the last few years, they are some 17 percent higher than in the EU. The existing payment mechanism has increased incentives towards longer hospital stays. A final measure considers the perspective o f cost-containment, measured by cost per discharge. Over the last few years, the cost per discharge has been increasing, the result o f increasing budgets and declining discharges. On balance, poor performance i s observed inproductivity and cost-containment indicators. There have been little incremental gains in productivity while all cost-related indicatorshave increased over time due to the combination o f increased budgets with reduced number o fbeds and hospital activity. 5.15 There i s a wide range o f areas where efficiency gains can be achieved by implementing specific cost-control policies. Some key areas are: 1. Utilitvmanagement: utilities represent around 8 percent o f total hospital expenditures. In best-practice systems, utilities represent around 5 percent o f the budget. 2. Ambulatory surgery: international practices suggest that countries should increase ambulatory surgeries as a way to reduce overall costs. In some countries, such as the U S A and Canada, ambulatory surgeries now represent more than 50 percent o f total surgeries. 3. Consolidation o f central services like sterilization and laundry: hospitals can move from the current practice o f supplying services like laundry and security to outsource them by contractingprivate firms. 4. HosDital consolidation: consolidatinghospitals represents a significant source o f savings by reducing fixed costs, especially infrastructure-related ones. Some estimates suggest that in Chisinau area alone, 1000-1250 hospital beds (13% o f current) could be closedby improved management procedures without affecting level o f services provided. This would enable the closing o f 4 hospital buildings, with associated saving in infrastructure related costs such as utilities. 53 Table 5.3: EfficiencyIndicators of the Moldovan Hospital Sector Productivity Discharges per physician Small progress; productivity rates increased around 1 percent in 2003- 2005. Unequal performance by hospital category and by Rayon. Hospital Occupancy Occupancyrate Moderate-to-low average rate (66 percent) but declining trend in 2002- 2004; small recovery in 2005. Wide dispersion o f occupancy rates among Republican hospitals. Occupancy rates in Moldova among the lowest in Europe. Intensity o f use of hospital B dturnover rate (BTR) Increasing trend in the BTR from 25 beds to 27 patients per bed per year. Using LASSO analysis, 40 percent o f the hospitals perform well and 44 percent o f the hospitals have excess bed availability. Among hospitals not performing well, 43 percent have excess beds due to lack o f demand for their services while the remaining 57 percent have problems with excess hospitalizations. Cost-containment Cost per discharge Costs per discharge and costs per bed- Cost per bed-day day (and variations, like costs o f food per bed-day) increased inreal terms as a result o f combining increasing budgets with decreasing beds. Significant variations exist by hospital type in any o f the cost categories evaluated. Length Average Lengtho f Stay Declining trend in 2002-2005 (ALOS (ALOS) fell 14 percent), although small increase in 2005. ALOS in Moldova still higher than EUaverage. E. RECOMMENDATIONS 5.16 Many o f the issues considered require long-term consideration, but there are some problems that need immediate attention because they either take a long period to be hlly implemented or because the severity o f the situation affects key outcomes o f the system. Among the most important issues that need immediate action are: A. HosDital efficiencv and consolidation: Since about 40 percent o f the health care facilities belong to other Ministries/Departments, these facilities have stayed outside the current reform effort.26 It i s critical that the Government adopts a strategy for 26Based on statistics from Ministry o f Health, there were in 2004 a total o f 107 inpatient health care facilities in Moldova, o f which 65 were municipal, rayonal and republican and 42 were categorized as "other ministries and departments". 54 restructuringhonsolidating these facilities and to define how freed resources would be allocated. For instance, given that infrastructure conditions of most hospitals are unsafe; buildingsare old, and equipmentis far from the besttechnology currentlyavailable, those could be areas of interest for future investments. As part o f a restructuringplan, the M o H should pay attention to generate adequate incentives to enhance efficiency through the selection of the best performers, leavingbehind less efficient providers. B. Cost control:public health expenditures have been growing significantly duringthe last years despite the fact that hospital activity decreased. As a result of both increasing budgets and reduced hospital activity, cost-related indicators increased. This situation points to the need to control expenses through improved resource allocation mechanisms and to emphasize the quality rather than quantity o f spending. As part o f a cost-control strategy, the government should pay attention to the exceptional growth of the National Programs and analyze productivity, cost-effectiveness, and appropriateness of programs. C. Fiscalsustainabilitvand coverage:The largest part health expenditures comes from the central budget, therefore it i s critical to ensure the fiscal sustainability o f these funds. There i s a need to increase the collection of mandatory health insurance contributions to and include more working age people as contributors in order to reduce the level of budgettransfers to NHIC. Despitethe highrate o f coverage achieved injust two years, there are 12 Rayons where coverage rates fall significantly below the rest o f the country. Increasing coverage inthose Rayons shouldbe a short-term policy goal giventhe positive outcomes in terms of access and equity for the population. Moreover, the expansion of health insurance to cover the poorest population plus the introduction of mechanisms to convert informal payments into formal co-payments, and the introduction of governance mechanisms (through community awareness on co-payment structure) could further reduce the burdenof out-of-pocket payments. D. Oualitv of healthservices: The development of programs to improve quality and focus providers on user satisfaction has lagged behindthe changes in financing and provision. Quality indicators, both at hospital and primary health care levels, deteriorated over time or were far from matching international standards. Some additional problems persist such as the absence of clinical management approaches (protocols, integral pathways, etc), an inadequate payment system to promote quality o f care, and unmet demand for FD services inapproximately 30% of the Rayons. 5.17 There is still considerable opportunity to further consolidate services and introduce efficiency oriented measures at the proiider level. At the same time, innovations need to focus on improvements in quality and the development o f a culture oriented to satisfy the needs and expectations of patients. Policy options along this line include the preparation o f a hospital consolidationhestructuringstrategy, continuation o f efforts to enhance primary care facilities and ensure availability o f basic medical equipment inall centers as well as the strengthening o f health management to focus on performanceh-esults and to learn to develop management structures in hospitals. 55 6. ENHANCINGTHE EFFICIENCY OF RESOURCEUSE INEDUCATION A. INTRODUCTION 6.1 Education spending in Moldova i s broadly comparable to other countries in the ECA (Europe and Central Asia) region. However, the sector suffers from a number of weaknesses, includinglow enrollment rates, persistent problems with the quality of services, and institutional arrangementsthat do not facilitate the most efficient use o f public resources. 6.2 First, Moldovan enrollment rates lag behindmany of its neighbors. At the crucial pre- primary level, the enrollment rate among 3-6 year olds, after falling duringthe 199Os, has revived recently and, at 62 percent in2004, it was close to the average inRussia, Ukraine and Belarus but ' Europe. The basic education (primary and lower secondary - grades 1to 9) gross enrollment rate still well below the average in the EU member and candidate countries of Central and Eastern was 94 percent in2004 - lower than rates inthe EUmember and candidate countries of CEE, but higher than in other western CIS countries. At the upper secondary level (grades 10 to 12), however, the overall enrollment rate of 45 percent i s not much more than half that achieved by CEE EU members and well below that in the other comparator countries. This reflects the collapse o f vocational/technical education (obsolete and no longer attractive to students and their parents), where enrollment now represents only 16 percent of the age group, and a relatively slow expansion of general secondary education over the past ten years to an enrollment rate o f 28 percent in 2004. Inter-country comparisons o f higher education enrollment rates are made difficult by definitional problems, but Moldova's enrollment rate o f 28 percent i s less than half that inthe CEE EUmembers. There are also inequities in access to education. As far as gender i s concerned, boys, rather than girls, seem to suffer from inequity in access to the higher levels of education, but data on enrollment rates by age group reveal inequities between income groups and between rural and urban areas, which are particularly acute in the 3-6, 16-18 and 19-25 age groups. 6.3 Second, recent data on learning outcomes suggest that there are still quality problems at lower secondary level, but the relatively good results of the TIMSS 2003 grade 4 tests suggest that there i s hope for improvements arising from the reforms of curriculum, textbooks, teacher training and assessment that are already in train. The TIMSS report showed that the shortage of resources for teaching materials and equipment i s still acute. Secondary vocational schools are in particular trouble, and the quality of higher education has deteriorated considerably in recent years.27 27The 2005 World Bank note on Education ("Moldova - Education policy note - analysis in support o f improvements inquality, equity, and efficiency inthe education sector.") notes that the draft strategy paper (Ministry of Education 2004) recognizes that quality of higher education (including teacher training) has deteriorated considerably inrecent years, owing partly to the big fall in the real value o f allocations from the government budget and partly to internal factors. These include: the continued orientation o f the curriculumtowards memorization; the prevalence o f inefficient teaching techniques; the absence o f criteria and mechanisms for objective assessment; andthe marginal use o f information technology. 57 6.4 Third, there is a crisis inpublic finance o f education, reflected most dramatically in the increasing number o f schools that receive less total hding from the state than the amount that they need to pay staff salaries. The low level o f teachers' salaries has an adverse effect on quality o f learning outcomes at all levels, making it difficult to attract good young teachers into the profession. This problem i s due primarily to inefficiency in the use o f the sector's budget. With funds allocated to schools on the basis o f specific guidelines (on classroom hours per week, minimum class sizes, and others) established by the Ministry Education together with the Ministry of Finance the flexibility o f local governments and school managers in using the resources i s severely limited. 6.5 Inlight ofthese challenges, issuesrelatedto the allocationofresources withinthe sector as well as those o f technical efficiency are dominant. A first step in any movement towards a more efficient education system i s to define exactly what the state undertakes to provide free to Moldovan residents in the way o f education services. What i s the `basic package' that they can expect from their government? In Moldova, however, where education at all levels and o f all types has traditionally been regarded as a public good, many o f these and related issues are currently unsettled, vague or subject to unpredictable change, as reflected in conflicting legislation.28And while there has been some progress with educational reform, the draft law on education i s ambivalent on further reforms necessary for achieving greater efficiency in the sector.29No mention i s made o f network optimization nor o f reform o f vocational education. Moreover, any argument over the appropriate level o f funding for education i s pre-empted by the statement that `the primacy o f education i s guaranteed by the government through priority funding o f the education system at the level o f 10 percent o f the Gross Domestic Product' (Framework Law, Art. 45). 6.6 This chapter explores the opportunities for fiscal savings in the sector, while identifying measures to help ensure that there i s no deterioration inthe level and quality o f services provided and that there are adverse poverty and distributional consequences. It focuses on a number o f areas where increased efficiency o f resource use can be achieved, including the rationalization o f the school network (class/school size and high level o f non-teaching staff), and estimates their likely fiscal impact. The extent to which decentralization o f authority to autonomous schools will help address the allocation o f inefficient education expenditures i s also explored. B.COSTSOFEDUCATION 6.7 Education expenditures currently amount to over 7 percent o f GDP, steadily rising from about 5.7 percent o f GDP in 2001. Expenditure on education i s projected to fall to below 6 percent by 2009. Because o f GDP growth, the education budget will still be increasing, but only by about 4 percent in real terms over the whole 2004-2008 period. Over the same period, the number o f school-age children i s expected to fall - by 9 percent in the case o f 3-6 year olds, 24 percent for 7-15 year olds and 18 percent for 16-18 year olds - and the number o f 19-24 year olds i s expected to rise by only 6 percent3'. On the face o f it, these demographic trends may help the search for fiscal space inthe education budget. 28 For example, the draft education law says that `public education i s free o f charge', but also that `institutions o f public secondary vocational education, post-secondary vocational education, and higher education are allowed to charge fees for educational services' (Framework Law, Art. 7). 29 World Bank 2005: section 111. 30 These are unofficial projections based on estimates o f population b y age group. Official census-based projections are not yet available. 58 6.8 The proportion of public educational expenditure that goes on personnel (salaries, social contributions and medical insurance) has increased since 2002, to 67 percent of current and 61 percent o f total expenditure in 2004. Other important categories o f expenditure are heating and fuel (6 percent of the total in 2004), electricity and gas (5 percent) and food (5 percent). Capital expenditure (including purchase of equipment and repairs to buildings as well as construction) accounted for about 9 percent of total spending in 2004 (about the same as the OECD average), but only 10 percent of the total could be regarded as directly relevant to quality of education (teaching materials, books and magazines, maintenance, repairs and purchase of equipment, professional retraining, ITworks and repairs to buildings). 6.9 Primary and secondary general schooling absorbs most o f the education budget. This includesprimary schools (grades 1-4), gymnasia (grades 1-9), general secondary schools (grades 10-11) and lycea or high schools (grades 10-12), which cost 897 million lei - more than half of total public spending on education in 2004. Personnel expenditure accounts for 70 percent of the sub-sector total, with teachers alone accounting for around 53 percent o f the total, managers 5 percent and other non-teaching staff 13 percent3' Capital expenditure accounted for 9 percent of sub-sector spending, as did quality-related expenditure. Electricity, gas, heating and fuel together consumed 11 percent of the budget. No publicly-funded stipends or scholarships are awarded at this level. Obvious areas in which to look for fiscal savings in primary and secondary general education are personnel expenditure, where student/ teacher ratios and the ratio between non- teaching and teaching staffneed to be reviewed, and spending on heatingand fuel. 6.10 The next largest area of public education expenditure is pre-schools, on which 277 million lei, 16 percent o f the total, was spent in 2004. At this level, personnel expenditure represents only 51 percent o f total spending, while the provision of food (13 percent) i s a relatively large item. Secondary vocational education, including both vocational and trade schools, i s now a relatively small sub-sector (4 percent of the total education budget): over a quarter o f its budget goes on stipends to students (received by 83 percent of them, predominantly from disadvantaged backgrounds), 40 percent of whom are accommodated in hostels. Students pay for food from their stipends. Other institutions o f secondary education include some expensive institutions, the general and the special-needs boarding schools (internats), and accounts for 8 percent of total education expenditure: food and other items associated with fully- subsidized boarding (heating, fuel and other utilities) feature disproportionately inthe budgets of these institutions. A relatively small proportion of total education spending goes to colleges (3 percent) and universities (6 percent) which now rely heavily on extra-budgetary revenue, particularly fees. Quite a large proportion o f public expenditure at these levels (10 percent for colleges, 16 percent for universities) goes on scholarships/ stipends, received by 74 percent of college and 14 percent o f university students. In general, exploration o f fiscal space in these institutions should focus not only on personnel spendingbut also on expenditures associated with boarding and on scholarships/ stipends. 6.11 Also relevant to fiscal saving possibilities i s the amount of public spending per student at each level. Of the educational institutions for which unit cost figures are available, the most expensive per student to the taxpayer are the secondary vocational schools. They spent 2,943 lei of public money per student in 2004 - exceeding pre-schools by 16 percent, colleges by 25 percent, general schools by 76 percent and universitiesby 185 percent. 31 Langton2006: Appendix 8. 59 c. FINANCING SYSTEM FOR PUBLIC EDUCATION 6.12 Most public educational institutions are the responsibility o f local governments, the main exceptions being most vocational schools and all colleges and universities, which are funded directly from the state budget. However, a large part o f the finance for locally managed institutions comes in the form o f budget transfers from higher levels. The Ministry o f Education i s involved inthe process only as a recipient and supplier o f information. The Ministry o fFinance drafts the budget, and monthly transfers are made from the state to the territorial treasuries32. 6.13 The amount transferred from the state budget to the ruions (for onward transmission to mayors' offices and thence to schools) i s determined by the number o f students33.Table 6.1 shows the norms per student (excluding funding o f school meals and the organization o f examinations) that have been used since 2003 and projected to 2008. As can be seen, norms are higher for pre-schools than for general schools and highest o f all for the internats - boarding schools for disadvantaged children and those with special needs. Table 6.1: Allocationof StateFundsto ATU budgets: Normsper Student, 2003-2008 (lei) 2003 2004 2005 2006 2007 2008 Pre-schools 1021 1367 1565 2642 3026 3166 Primary schools, gymnasia & high 963 1151 1232 1725 1816 1895 schools Complementary (extra-curricular) 3963 96 100 144 156 161 institutions Family-type orphanages 5441 6707 6707 6847 7249 7326 Generalboarding gymnasia (internats) 4584 5315 8342 10900 14650 Special-needs boarding gymnasia 6515 8313 11389 14182 18171 Source: Ministryof Finance. 6.14 While these are simple formulae for allocatingpublic money to schools, a school director i s not allowed to use her allocation flexibly, thus constraining the efficient use o f resources. The amount that a school director must spend on salaries i s determined by the number o f teaching hours inher school, which is, inturn, determined by the number of classes (however few students they may contain). Inthe increasing number o f schools with very small classes, the salary bill (based on the number o f classes) thus exceeds the total allocation (based on the number o f students); inmany schools, it leaves little or no public money for furniture, materials, repairs, etc. The formula funding system i s generating signals but they cannot be acted on. In 2005, only about 59 percent o f schools' budget needs are estimated to have been covered by these transfers from the state budget (CASE Moldova: Table 1.2). Funds have to be supplemented at the village level by local taxes, extra-budgetary revenue or transfers from other budget headings. 32 See CASE Moldova (2006) for an excellent description of the financing system. 33 The only adjustmentto the formula i s to give a higher coefficient to Chisinau in each case, to reflect the higher cost of utilities. See CASE Moldova (2006: 43) for a more detailed description of the flow of funds from the centre to the administrative territorial units. 60 6.15 Schools vary in the extent to which they enjoy managerial autonomy. In some cases, directors, working closely with the local mayor's office or with raion financial officers, agree and set annual budgets, appoint non-teaching and (less often) teaching staff, distribute teaching workloads, etc. Inother cases, directors are kept out o f the budget-setting process. In all cases, responsibility for managing funds lies with the local authority, the number o f classes (and hence of teachers) can only be changed with permission and the number and type of non-teaching staff are governed by norms (Langton 2006: 11). 6.16 A new Government project will initiate (via pilot schemes) a move towards decentralization of authority to fully autonomous schools, in which school directors have flexibility in the use o f budgets, and school bank accounts, with a managerial role for school councils and with full involvement of communities in school affairs. This would be accompanied by a broadening of the funding formula to include some additional factors. A pre-project study (Langton 2006: 18) has suggested, as possible candidates for inclusion, in addition to student numbers, age-weighting to give more money to younger students, a measure of poverty in the area, the state o f school buildings,and school size, but the debate on this has hardly begun. D.SCOPEFORIMPROVEMENTS INEFFICIENCY 6.17 What scope i s there for improvements in efficiency? Some orders o f magnitude can be estimated. In the next two sections, existing weakness will first be identified then options for more efficient use o f resources will be specified, with a view to outlining a possible `basic package' for the education sector. Then, the MTEF will be used to explore their impact and implications for fiscal space. 6.18 First, many schools have very low numbers of students, small classes and low student/ teacher ratios, as Table 6.2 shows. The figures are particularly low, in comparison with OECD averages, for gymnasia, which account for some 23 percent o f enrollment in general schools. Moreover, there are wide variations between schools inthese indicators. For instance, Paiu (2004) showed that, o f the 15 gymnasia in Edinetjudet in 2003, five had student/ teacher ratios o f between 5.1 and 6.4, five o f between 9.4 and 9.8 and five of between 13.8 and 13.9; while, in Bgltijudet, class sizes varied from 5.0-9.9 (11 classes) to 10.0-14.9 (20 classes) to 15.0-19.9 (24 classes) to 20.0-24.9 (28 classes) to 25 or more (3 classes). As for school size in rural areas, nearly two thirds of gymnasia have 200 students or less, while 14percent have fewer than 100: in urban areas the comparable figures are 21 and 7 percent. At secondary general and lyceum level, 38 percent o f rural schools have fewer than 401 students, 14 percent fewer than 281, compared with 18and 9 percent respectivelyinurbanareas. 6.19 The optimization of the school network in general education (from primary to upper secondary) thus represents an opportunity for generating substantial fiscal savings. As the school- age population continues to fall, it becomes more urgent to focus on schools and classes that are below optimal size. Political realities have to be faced, however. It i s highly unlikely, given the importance of the village school to the community that any government would decide to close small schools in villages that already have them. A suggested model that could be explored is to retain pre-schools and primary schools (with multi-grade teaching) in these villages (where possible sharing a building) and to optimize at the secondary level, including the secondary classes shifted from these small villages. Inthe absence o f the full school-mapping information, it i s impossible to estimate accurately the possible savings from such a move. However, it can be assumed (particularly if school autonomy allows directors to hire and use teachers flexibly) that it would allow an increase in student/FTE teacher ratios to around OECD levels (16.6 in primary and 13.6 in secondary schools). This would be offset by a needto provide transport for students 61 ingrades 5 andhigherto the nearestschool, involvingthe hireof vehicles with drivers rather than purchase, and by the cost of training teachers in multi-grade teaching methods and other re- structuringcosts. Table 6.2: SelectedEducationIndicators by level,generalschools, 2005 Secondary Primary Gymnasia Lycea General Total Schools Schools Student/FTE 15.3 1.4 3.1 13.0 12.7 Teacher ratio Students per 22.5 8.5 24.6 22.4 22.2 class Students per 147.3 87.6 641.6 444.8 358.9 institution Number of 1.6 1.1 2.2 1.8 1.6 classesper grade Source: Langton(2006: Appendix 8). 6.20 Second, non-teaching staff represent a disproportionate share of total staff, about 37 percent of total staff in general schools, compared with the OECD average of 27 percent. Norms for non-teaching staff, which are laid down centrally, could be relaxed in autonomous schools. Reducingnon-teaching staff levels down to the OECD averagerepresent a further opportunity for generating fiscal savings. 6.21 Third, heating costs are inflated by the poor insulation of most school buildings, the duration o f heating (sometimes 24 hours a day, seven days a week during term time), and the large amount of space utilized per student (Tibi et al., 2002: 35ff,). While the number of students ineach school is falling, the size of school buildingsremainsthe same: between 2001 and 2003, the proportion of school space utilized in primary and secondary schools fell from 85 to 79 percent, and it i s expected to fall further to 54 percent by 2007 and 50 percent by 2010 (Paiu, 2004). While the scope for reducing heating costs i s reduced by the rising price o f energy in international markets, various options and measures for containing heating costs can be explored. These measures include the following: where costhenefit analysis justifies it, equipping schools connected to district heating with meters that have valves, and modifying fixtures to reduce heat loss; merging schools to make more efficient use o f premises; changing the school year to allow long vacations duringwinter; and, inthe longer run, reviewing the choice of fuel and technology and progressivelyreplacing old schools inpoor condition with new buildingsdesigned for energy effi~iency.~~ 6.22 Fourth, textbooks, which had been excluded for the past few years from the free basic package for general school students in favor of a rental scheme, are once again beingprovided free o f charge to all primary grade students.There i s a case for reviewing this decision. The virtue of the textbook rental scheme and associated extra-budgetary textbook fund was not only that they saved public expenditure: they also ensured the availability o f high-quality textbooks to 34See Tibi et al.: 36-37. 62 students. A selective subsidy to enable children from disadvantaged families to obtain textbooks i s certainly justified on equity grounds but a blanket subsidy to all primary students i s likely to threaten quality by reducing the number of good textbooks available and to squeeze other components of school budgets. 6.23 Finally, opportunities for fiscal savings exist at various levels of education, but they need to be offset by social and distributional considerations. The following i s a summary o f key opportunities at each level, while qualifying the scope and limits for fiscal savings: Pre-schools absorb an unusually large proportion o f total education expenditure (16 percent, equivalent to 1.1 percent of GDP compared with 0.5 percent on average in OECD countries) and their unit costs are 52 percent higher than those o f general schools. In response to this, Tibi et al. (2002: chapter 8) recommended full cost recovery for pre-school children between the ages of one and five, but free pre-school education for six-year-olds in general secondary schools. However, the equity implications of this recommendation have to be taken into account. Early childhood development i s particularly important to the future academic achievement of children from disadvantaged families: continued subsidy o f pre-schooling for younger children from such families would bejustified on these grounds. There is a major question about the viability o f vocational schools intheir traditional role of providing ready-to-work recruits with specialized skills for the economy's enterprises. They are chronically under funded inrelation to this objective and, even so, are the most expensive and least efficient secondary educational institutions. Although male vocational school graduates do better in the labor market than their general school counterparts, this i s not the case for females. As secondary students increasingly seek routes to higher education, enrollment invocational schools is steadily falling, equipment and curricula are outdated, and there i s a shortage o f competent technical teachers. Re-equipment would be too costly. Inthe past, also, they were used as a repository for the less able, on the assumption that such children had reached the limits of their academic absorptive capacity by their early teen-age years. That assumption has been increasingly questioned: the benefits of avoiding premature allocation of students to academic and vocational streams are acknowledged, particularly since a sizeable proportion o f those classified as "less able" tends to include students from disadvantaged backgrounds with unrecognized potential. Training (wherever possible, in- plant and on-the-job and financed as far as possible by beneficiaries), after a high-quality general education has beencompleted, i s a more promising alternative model. Another area inwhich a change inmodel would be useful(and would save taxpayers' money) i s that o f the general and special-needs boarding schools (internuts). In general, the educational, developmental and future-employability needs of many o f the children in such schools are more likely to be fulfilled by their inclusion in mainstream schooling systems than by isolating them in separate institutions. Others could be more efficiently served by local day placement facilities. In higher education, a "dual-track" system of fees and stipends provides inadequate academic funding and i s inequitable. Fee levels vary by subject (inMoldova State University they range from 2,800 to 7,000 lei per year, dependingprimarily on the ratio of applications to places), but budget students (accounting for 51 percent o f the total in public colleges and 24 percent in public universities in 2005/6) do not pay fees. Stipends are paid to budget students, who are chosen primarily on academic merit. An alternative package i s worth 63 exploring, along the following lines: First, all students at public higher education institutions could be required to pay fees (Tibi et al. suggested a fee level o f between 50 and 70 percent o f average unit costs). Second, in place o f existing stipends and fee remission, grants, to cover fees and living expenses fully or partially, would be made available by the Ministry o f Labor and Social Protection only to students (with adequate qualifications, o f course) from disadvantaged backgrounds. Universities could still use their own money to give some merit- based scholarships, but they would not get allocations for this purpose E. FISCAL SPACE? 6.24 The budgetary envelope for education is provided by the Medium Term Expenditure Framework (MTEF) for 2007-2009, and projections made for the purposes o f this chapter break down the MTEF totals further by category o f expenditure and combine them with statistics on enrollment, staffing, etc. The driving force behind the projections is the expectation, already mentioned, that the number o f 3-6 year olds will fall in2004-2009 by 9 percent, 7-15 year olds by 24 percent and 16-18 year olds by 18 percent - and that the number o f 19-24 year olds will rise by only 6 percent. At the same time, it is assumed that progress will be made towards the enrollment rate targets o f the Economic Growth and Poverty Reduction Strategy (EGPRSP), and the Education for All (EFA) National Action Plan. Otherwise, it i s assumed at each level that stafflstudent ratios, the number o f institutions and the distribution o f expenditure between different categories will remain unchanged. 6.25 Onthese assumptions, expenditure on publicly owned pre-schools is set to increase by 92 percent over the period. A steady increase inenrollmentrate to 80 percent by 2009 will more than offset the fall innumbers inthis age group, yielding an increase inthe number o f pupils and staff over the period o f 6 percent. Total expenditure per student and average personnel expenditure per staff member will thus increase, on these assumptions, by around 80 percent - compared with the 54 percent rate o f inflation projected by the MTEF. 6.26 The second category o f institutions i s public general schools teaching at primary, lower secondary and upper secondary levels. The MTEF lumps all secondary schools together, including also vocational and boarding schools. For the purposes o f these projections it has been assumed that the split o f expenditure between different types o f secondary schools remains unchanged over the period. A rise in enrollment rate i s more than offset by the fall in population in this age group, so total expenditure per student and average personnel expenditure per staff member rises by 68 percent - a relatively small increase inreal terms. 6.27 The MTEF projects an increase in extra-budgetary funding o f only 20 percent over the period. On the assumption that its distribution between levels does not change, secondarv vocational education i s projected to show a 59 percent increase in expenditures. With no increase in enrollment rate expected, the number o f students and staff members is projected to fall by about 18 percent - to the benefit o f total expenditure per student and average personnel expenditure per staff member, which will increaseby 95 percent. 6.28 Public colleges are also expected to experience a faster rate o f increase inpublic than in total expenditure. With enrollment rate unchanged, the number o f students i s expected to increase by 6 percent, with the result that total expenditure per student and personnel expenditure per staff member will rise by 56 percent - about the same as the rate o f inflation. 6.29 The picture in public universities is even worse, suggesting a near-crisis in financing at this level. Total expenditure is projectedto increase by only 35 percent, in spite o f a 124percent 64 rise in public spending. A 6 percent increase in the number o f students implies a rise in total expenditure per student and average personnel expenditure per staff member o f only 28 percent - well below the inflation rate. 6.30 These projections can be used to illustrate the rough order o f magnitude o f the potential savings of various efficiency-boosting options specified as part o f a `basic package' in the previous section, as follows (Figure6.1 shows the results for 2009): if public spending on pre-school pupils aged five or below were cut by 70 percent (assuming that 30 percent3' of pupils, from the poorest households or otherwise disadvantaged, would still be subsidized), and free food for those over the age of five were also confined to the least advantaged 30 percent, while the pupil/ teacher ratio was raised to the average OECD level, fiscal savings of 211 million lei would be made by 2009; in general schools, if student/ teacher and student/ non-teacher ratios were reduced to OECD levels, through a combination of relocating secondary classes from schools that are below an efficient size (while allowing primary classesto remain where they are, with multi-grade teaching if necessary) and the abolition o f strict norms for the number of teaching and non-teaching staff in autonomous schools, fiscal savings totaling 277 million lei would result; 0 there would be some offsetting increases in expenditure, estimated at 40 million lei for hiringbuseswith drivers (assuming that a 20-seater bus couldbehired for around 20,000 lei per year to make two journeys of up to 10 kmper school-day and that 40,000 children would be riding to school) and around 101million lei in other extra current expenditure on severance packages for n~n-teachers~~,re-training teachers, and making other adjustments to school premises and arrangements; renovation of school heating and insulation systems (preferably donor-funded) and a more efficient use of space in merged schools are assumed to be necessary to hold heatingcosts at current levels rather than to yield any savings; 0 ifvocationalschoolswere closedandtheir studentswere absorbedintothe general school nearestto them, an annual saving of some 64 million lei would result, reflecting the much lower unit costs to the budget of the general day school compared with the vocational boarding school, but a one-off severance package (the cost of which i s not included in Figure 6.1) would beneeded for vocational school staff; 0 ifthe budget for state-funded stipends inhigher education were cut by, say, 70 percent, and they were awarded to students from disadvantaged backgrounds rather than mainly on grounds of academic merit,public expenditure would fall by about 27 million lei; 0 if, at the same time, all higher education students (except those from disadvantaged backgrounds) were required to pay fees, rather than just the contract students, both colleges and universities would benefit from substantial increases in fee income, but the government would incur some costs insetting up and guaranteeing a loan system. 35The povertyrate in 2004 was estimatedat 27.8 per cent (Republic of Moldova2006a: Table 2). 36Naturalwastage would probablybe enoughto reducestudent/ teacher ratios to the desired level. 65 Figure 6.1: FiscalSpace in Education: PossibleSaving & Extra Spending, 2009 -250 -200 -150 -100 -50 0 50 100 150 million lei Source: Author's Calculations. F. CONCLUSIONS 6.31 Increasing the efficiency o f education sector spending can yield net potential savings o f about 437 million lei, a substantial amount out o f the total projected 2009 education budget. However, there are at least two grounds for caution (in addition to the very rough basis o f the calculations) intreating this amount as pure fiscal space. . 6.32 The first i s the crisis in the general school teaching profession. Many observers have reported the concerns o f staff inraion offices and schools about the future o f the profession. They point out that many teachers are near the end o f their working lives and relatively few new teachers are coming forward to replace them. They relate this directly to the low salary levels for teachers, discussed above. Special incentives for young recruits are only part o f the answer to this problem, which requires a re-shaping o f salary structures for the whole profession. Our fiscal space calculations retain the assumptions o f our base projections as far as personnel expenditure per teacher i s concerned - a 68 percent increase on average between 2004 and 2009. This would implynot only a slight increase inreal terms over the period (since prices are expectedto rise by 54 percent) but also a dramatic further falling behind other wage earners in the economy: the latest MTEF estimates that the average wage will rise by 159 percent over the period. As with the rest of the public sector, higher teachers' wages would be an important element o f attracting and retaining qualified teaching staff. 6.33 The second reason for caution in the fiscal space discussion i s the very l o w proportion of the publicly funded budgets o f educational institutions available for quality improvement, i.e. expenditure on teaching materials, books and magazines, maintenance, repairs and purchase of 66 equipment, professional retraining, IT works and repairs to buildings (for instance, 8 percent in general schools). Higher spending on quality-related inputsi s needed. 6.34 The magnitude o f net potential savings in the education sector suggests that there is a significant scope for improved efficiency in the sector. Before resources are ultimately allocated back into the education sector, care must be taken to ensure that limited public resources are being efficiently used. The preceding analysis suggests a challenging reform agenda and a number o f opportunities for greater efficiency at various levels of education concerning staffing, resource use, and input mix. 67 7. PROMOTINGTHE FISCAL SUSTAINABILITY OF THE PENSION SYSTEM A. INTRODUCTION 7.1 During the recession that followed independence in Moldova, the pension system inherited from the Soviet period came close to collapsing. To strengthen the fiscal sustainability o f the system, a reformprogram was initiated in 1998. Pensions entitlements earned before 1998 were revaluated (and reduced), and a new formula was introduced for the pensions rights associated to the contributions paid after 1998. To reduce uncertainties in the revenue- expenditure gap, increase transparency, and improve equity, the reform program mandated that pension payments be made based on actual paid contributions rather than on stated wages, supported by an appropriate information technology (IT) infrastructure containing electronic records o f individual contributions. Retirement age was also set to increase gradually. 7.2 While a number o f short-term objectives have been met by the reformprogram (including pension payments provided on time, in-cash rather than in-kind, and without the need for budget transfers), the reformprogram remains incomplete. A hold was put on the retirement age increase in2001, the pension formula has been revised several times, and the National Social Insurance House (NSM) record system has not been completed. As a result, although the NSIH has enjoyed a favorable economic and demographic environment for 4 years now, it was in deficit at the end o f 2005. The medium and long-term financial stability o f the system requires that the reform program be completed. From a fiscal perspective, the fundamental challenge i s that o f ensuring that pension benefits are fully financed by available resources within the pension system. It requires eliminating unfunded liabilities and imbalances inthe system, to reduce fiscal pressures and preserve the fiscal space required by the Government to carry out its development program. 7.3 This chapter focuses on fiscal issues related to social insurance programs and in particular, the pension system.37It looks at the long-term financial viability o f the pension system as well as the challenges in completing the reform program while taking into account the likely poverty and distributional consequences. In particular, it focuses on the following key issues: relatively l o w levels o f pension benefits; the substantial redistribution from non-agricultural workers to agricultural workers and the strategy to unify the pension systems for all categories o f retirees; and the financial stability o f the pension system, given current and future demographic trends, large emigration flows, and projected macroeconomic and fiscal developments. It concludes by assessing the likelihood o f generating substantial reserves and the institutional requirements for managing suchreserves well. 37 This chapter thus focuses primarily on the social insurance program o f the government, representing 87 percent o f all the NSIH social transfers. However, while the distinction between social insurance and social assistance programs is clear in principle (i.e., social insurance programs aim to smooth individuals' income shocks due to temporary or permanent loss o f earning capacity, and not directly linked to alleviating poverty), NSIH social insurance and social assistance programs are not financially independent. As such, this chapter presents also some NSIH aggregate financial indicators, though the in-depth analysis o f current social assistance programs and the possible impact o f their reform are presentedin Chapter 8. 69 B. PENSIONINCOME 7.4 Inthe 199O's, the social insurance system was incrisis. In 1998 and 1999, most of the pension benefits remained unpaid and, only 40% o f the elderly entitled to pensionreceived some benefit3*. In such conditions, pensions clearly could not be increased. In 2000, the average pension represented no more than 22% o f the average wage and 38% o f the poverty line. The recent pension increases thus correspond to a logical recuperation o f meaningful pension income levels. With an average pensionjust above the poverty line, the levels obtained at the end o f 2005 still do not appear particularly high. 7.5 Figure 7.1 compares the average pension and the minimumpensions to the average wage and the poverty line. There are several types o f minimum pensions in Moldova, only the 2 principal ones are represented in Figure 7.1. One i s paid to the old-age retirees from the agricultural sector and the other i s paid to the old-age retirees from the non-agricultural sector. The other minimumpension i s paid to the disabled and the survivors. They are lower than these two levels because the insured have contributed for shorter periods, or because the survivors are only entitled to a portion o f the deceased insured pension (50% in the case o f widows). In the general system (excluding privileged pensions) disabled pensions are on average 15% lower than old-age pensions, and survivor pensions are 39% lower. Figure 7.1: Pension BenefitsLevels incomparisonto the AverageWage and the Poverty Line MDL per month 14W 1200 1000 BW BW 4W 200 - - I) 1888 -AverageMinimum - - - -1808 2000 2W1 2W2 2W3 2W4 2W5 Pension Pensionall Sectors except Agriculture Pension MinimumAgriculture PovertyLine ---~t Wage Average Sources:NSMdata, SCERS poverty report. 7.6 Social assistance programs, by contrast, aim to alleviate current poverty. They are direct or indirect transfers mostly funded by the State budget. InMoldova, this aid consists o f monthly benefits, social services and subsidies (primarily in energy and housing) to specific vulnerable groups. Ideally, the recipients o f these transfers should be only poor households, and a reform to improve the targeting i s under study. 38P. Caste1"Moldova: Public andPrivateSocial SafetyNet", PovertyAssessment backgroundpaper 2004. 70 7.7 The average pension in2005 was more or less equal to the poverty line, but about 17.2% o f the current pensions were still below such level, at the beginning o f 2006.39The increase of the pension will certainly improve many pensioners welfare situation but may not lift all the pensioners out of poverty. Poor pensioners tend to live inextended rural families and eliminating poverty among them would require higher transfers to push all their families' members out of poverty. Such action i s clearly relevant to social assistance programs. Some attention could be given, however, to the minimumpensionpaid to the disabled and the survivors. Simulations show that most of the pensions could be in 2009 above the poverty line, but still 4.4% could remain under. About 78% of the recipients of these very little pensions are working-age permanent disabled persons. As this population i s usually among the groups with high risk of poverty, revising the minimumbenefits paidto the disabled and survivors seems advisable. 7.8 Since 2001, pension benefits have been increased through two types o f interventions. Most of the increases have been done through general pension increases. In 2002 and 2004, however, the increaseswere obtained through the revaluation of the pensions rights associatedto the years worked before 1998. This choice probably reflects concerns about the low level of pension income at retirement. Figure7.2: PensionIncomeat Retirement by 7.9 Reassessingpension rights in order to Pensioners Category increase pension, similar to exercises 1200 conducted in 2002 and 2004, i s not efficient or 1000 transparent. It leads to cumbersome 800 600 recalculations of pension benefits, and pension 400 increases that vary across recipients. Given 200 these inconveniences, policymakers' choice 0 has probably been motivated by the following additional consideration, namely the change in pension formula. Thus, besides increasing the level of current pensions, the recalculation of the pre-1998 pension rights leads also to the Source:NSIHdata. increase of the levels of pension at retirement, so the measure also raises future retirees' pensions. 7.10 Despite these measures, pensions at retirement are still rather low in comparison to the current wages. Current retirees (about 30,000 persons retire per year) are also likely still dissatisfied with the level of pension income at retirement. First, the disparity o f pension income among the new pensioners i s large (see Figure 7.2). Second, replacement rates (the ratio between pension and the retiree's income at the end of hisiher working period) are very low for some pensioners' categories. 7.11 The income drop is particularly high for the retirees of the manufacturing sector. Estimates show that, in 2006, pension income at retirement could replace just 22% o f their last income. Estimates also indicate that this replacement rate could fall further to as low as 19.3% in 2008. The reason o f the large income drop at retirement i s that, inthe calculation o f the pension rightsearned after 1998, workers' wages are not adjusted inthe computation o f the averagewage. 39This does not necessarily mean that recipients o f these pensions live in poverty. Many elderly live in extended families, about 7% are still working, some obtain return from their savings (rents etc.), while others receive social assistancebenefit. Only a detailed analysis o f the results o f the household survey o f 2005 would indicate if pensioners are relatively poorer than the rest o f the population. 71 Table 7.1 illustrates this point. It compares, depending on different rates o f wage growth, the average wage and the pension obtained by new retirees. The results indicate that in a context o f modest wage growth o f 2 percent per year, for example, the average wage on which the pension i s calculated represents 69.8% o f the last wage. Pension replaces 37.7 % o f the last income and given that pensions are partially indexed to the nominal wage growth, the average replacement rate duringthe overall retirement period i s equal to 40%. This replacement rate i s comparable to international standards. Replacement rates significantly decrease, however, in a context o f strong wage growth because the wages earned at the beginning o f the working career pulls down the overall average. With a nominal wage o f 8 percent per year, the replacement rate at retirement drops to 17.4%. Table 7.1: Average Wage andReplacementRates Note: Average wage and replacement rate at retirement computed for a contributory period o f40 years, average replacement rate computed over a retirement period of 25 years, assuming real wage growth equal to half of the nominal wage growth. Sources: NSIHdata 7.12 Because wages have increased rapidly since 1998 (by about 25% per year), the average wage computed for pension since 1998 has fallen sharply inrelation to the current wage. It was o f around 69% in 2002, it could be at 48% in 2008. Without changes in the computation o f the average wage inthe pensionformula, replacement rates at retirement will keep falling at a result. 7.13 If economic growth is sustained, and wages keep rising, it is doubtful that such an outcome receives political and social support. Replacement rates can be rather low, but the gap between the system implicit financial return and the average return o f other private savings opportunities cannot be too large. Policy makers should, therefore, analyze what upgrading strategy would be the best for Moldova. To be fiscally sound, the envisaged changes must affect only the new generations ofpensioners, so the upgrade o f the system increases pension expenditures, but gradually. The size o f the upgrade depends, moreover, on the system's capacity to collect revenues. This issue and the overall fiscal impact o f upgrading the system are analyzed inthe next sections. c. TOWARDSUNIFIED A CONTRIBUTORYSYSTEM 7.14 Being universal, almost all workers, wage employed, farmers and self-employed must be insured to NSIH. Assuming that the level and the structure o f employment remained roughly between 2004 and 2005, the comparison o f the number o f contributors reportedby NSIH with the figures published by the National Statistic Bureau (BNS) indicate that NSMcovers about 93.6 % of the working population (Table 7.2). All these reported contributors, however, do not participate on the same basis. The amount o f contribution largely varies depending on the type 72 and sector o f employment: wage employed contribute in proportion to their wages; farmers and self-employed pay a flat amount per year and subsistencefarmers likely contributenothing. 7.15 As explained inprevious reports, farmers contribute very little: on average, about 179 lei per year per person in 2005. Three quarter of self-employed workers also contribute very little: 180 lei per year per person in 2005. Incomparison, the contribution due from a worker paid the minimumwage of 100 lei per month amounts to 348 lei per year (=.29* 100*12; employee's and employer's shares). Farmers, self-employed and wage-employed are, however, entitled at retirement to the same minimumpension(20% lower inthe case o f the farmers). The low amount of contribution paidby self-employed workers and farmers in comparison to the wage-employed appears, therefore, largelyunfair. Table 7.2: Employment and Numberof Contributors '' Self-employed 164940 101323 Self-employed flat rate 23845 Self-employed with permit 77465 Land owners (renting) 13 Farmers 167324 304227 Subsistence farmers 119534 Total employment 1316042 Notes: 1/ BNS andNSIH havedifferent definitions of employmentcategories. 7.16 In Government decision No. 1219 from October 23, 2006, the government approved a draft strategy to reform the pension system in agricultural sector. Under the new framework, the workers from the agricultural sector who are wage employed should be applied the same contribution rate than the other wage employed. Farmers and self-employed should contribute a minimumconsistent with the minimumpensionbenefit they will beentitled to at retirement. Such strategy should significantly increase NSIH revenues. The contributions due from self-employed workers and farmers (excluding the subsistence farmers) could amount to about 1.3% of GDP (about 560 lei million in 2006) and those due from wage-employed in the agricultural could represent about 0.5% o f GDP (about 230 lei million in 2006). These changes in contributions payments, however, are gradually introduced because many o f these contributors are poor or work inlagging enterprises. 7.17 In the case of farmers, in 2006, the flat amount of contribution was set at a level equivalent to 26% of the officially desired minimumlevel (480 o f 1815 lei per year). There i s no readily available information on farmers; average income in 2005, but if the income distribution among farmers was similar to the income distribution among the wage-employed in the 73 agriculture sector, this flat amount o f contribution would still appear excessive for about 20 percent of farmers4'. Inthe case of self-employed workers, in 2005, the payment of the desired minimumlevel of contribution was required of about a quarter of the self-employed p~pulation~~. The rest of the self-employed continue to pay very little amounts (equal to 30% of the monthly fees they pay to the State). 7.18 The resultingimbalance between farmers and self-employed workers' paid contributions and future pension entitlements roughly represents NSM implicit subsidies of 247 million lei in the case of the farmers and 138 million lei inthe case of the self-employed, a total of about 0.8% of GDP usingdata for 2006.42 7.19 The government has not articulated a strategy for the integration of all the self-employed workers and farmers into the new contribution framework. In the case of farmers, this process will probably take several years. Simulations show that still in 2015, a flat contribution above 53% of the desired minimumlevel will still be unaffordable for some 20% o f farmers43. 7.20 Supporting the participation of poor workers in social insurance i s an effective and efficient pro-poor policy, as long as the policy i s transparent and does not disproportionately benefit workers who do not require public support. Inthe absence of support, poor and near poor workers cannot contribute to a pension system on a regular basis. As a result, very few are able to contributeby themselves the number o f years necessary to finance a minimumpension. Ifinstead, duringtheir spell of poverty, these workers receive some financial support to enter or pursuetheir participation into the pension system, they have more chances to accumulate pension rights. In the long-term, such policy enhances the overall population welfare and reduces the number of future elderlywith no or insufficient old-age income support. 7.21 Since poverty rates are extremely high among farmers (53.6% in the first semester of 2005 compared to 28.5% for the total population), setting a subsidized uniqueflat rate for farmers seems at this stage reasonable and efficient.44However, subsidizing more than three quarter of the self-employed seems inefficient, since the population of self-employed workers have the lowest poverty rate (17.6% in 2005). Leakages45are likely important inthat group. The payment o f the desired minimumcontribution shouldbe enforced rapidly inthat sector. 7.22 With respect to workers from the agricultural sector who are wage employed, enterprises were asked in 2004 to pay a contribution rate of 20% (against 27% in the other sectors). The introduction of a contribution rate on wages in2004 has significantly increased the labor costs of many enterprises in the sector. I t might have led to some retrenchments, and certainly numerous case of financial squeezing. As the 2004 Social report observes, most o f the new arrears in contributions in 2004 are due from agricultural sector enterprises that could not comply with the new regulations. Instead of suspending the progressive convergence o f the contribution rate, the 40For these farmers the flat amount o f contribution would represent more than 10% o ftheir income. 4'Based on NSIH 2005 number o f contributors, 83858 among 101323self-employed paid a flat contribution amount. 42This is the difference between desired minimum and paid contributions. More precise estimate would compute the system's subsidies based on the present value o f the expected amount of the contributions paid, and the number andthe level o fthe future payments o fthe minimumbenefit. 43For these farmers the flat amount o f contribution would represent more than 15% o f their income 44The cost o f screeningthe applicants would outweigh the benefitsfrom reducing the leakage. 45The leakage i s the amount o f implicit subsidies that are not directed to the poor but to better-off. 74 government responded with the introduction of an explicit subsidy equal to 4 percentage points of the contribution rate. In2006, this subsidy could amount to about 32 million lei. Providing such a blanket subsidy to all the agricultural enterprises was probably unnecessary, and therefore not fiscally and socially efficient. The particular strong wage growth inthe agricultural sector in2004 and 2005 (23.8 and 19.7%) reflects important productivity gain and suggests that not all the agricultural enterprises currently face financial difficulties to comply with the new regulation. 7.23 To summarize, the government strategy to unify the contributory system should increase the system revenues inthe medium-term. The new regulations should be introduced gradually in the case of farmers, becausemany are poor, but could be implemented more quickly in the case of self-employed workers and wage employedworkers o f the agricultural sector. D.SETTINGUPTHEPENSIONFUND'S REGULATIONSAND SUPERVISORYINSTITUTIONS 7.24 The pension fund is not financially independent of other NSIH programs. While expenditures by types o f programs are easily distinguished, revenues and current administrative costs are distributed across programs. The budget law mentions some redistribution, but it i s theoretical and does not have practical implications. As a result, the deficit, the surplus as well as the reserves are produced at the NSM aggregate level. Table 7.3 presents the NSM overall financial balance and an approximation of the financial balance of the pension system that does not include the administrative costs and the amount of reserves and outstanding debt in contributionsthat couldbe attributed to the fund. 7.25 These figures show that after three consecutive years of surplus, NSIH ran into deficit in 2005 because of the pension fund. This deficit i s the result of two government actions: policy makers implemented a huge pension increase in November 2004, while decreasing revenues by reducing the contribution rate by 1percentage point, and thus the revenues from contribution by about 0.2% of GDP. 7.26 More generally, these results indicate that the recent policy of pension increases had reached its financial limit. Although wages and employment in the wage employed sector have steadily increased over the period, contributions in relation to expenditures have increased at a much slower pace. In2004 and 2005, wages in the administration have increase by about +14% per year and, wages inthe industryhave increased by 18.6%. Figures on employment in 2005 are not available, but employment (agriculture sector excluded) has increased in 2004 by 1.3%. Consequently, revenues from contribution have increasedby 0.5 percentage points of GDP (0.7 if contribution rate had not been cut). Pension expenditures, however, in the same period have grown by 1.7 percentage points. This i s clearly not sustainable. 7.27 The short-term financial position o f the NSMis, however, not alarming. The institution reports reserves of about 1.6% of GDP and the pension system relies very little on State budget transfers. O f the 566 million lei o f social benefits financed by the State budget in 2005, only 50 million accrued to pensions. These payments finance higher pensions to civil servants and privilegedpensioners (about 11500 beneficiaries or 2% o fthe total pensionexpenditures). 75 Table 7.3: PensionFund FinancialResults2002-2005 Source: NSM. Notes: 1/ NSMsocialinsuranceandsocial assistance program are not financiallyindependent.Revenues, surplus, deficit, reservesandoutstandingdebts are accumulatedat the House level. 2/ Estimatesbasedon the budgetlawindicationsabout the share contributionratedestinedto pensionprogram. 3/ Government mid-termframework. 7.28 The Government's mid-term Figure 7.3: Elderlyinpercentof Working-age Population financial position of the NSM over the period 2006-2008 will improve. The current demographic trends are favorable. Despite the interruption in the increase of retirement age in 2001, 20% I the total number o f pensioners has decreased in the past years. In the 15% - medium-term as larger cohorts from the postwar baby boom reach retirement 1 0 % - 1 , , , , , , , , , , , , , , , I , , / I I I I I I , I age, this demographic trend i s expected 2004 2009 2014 2019 2024 2029 , to reverse. and the- share- of elderlv in - men above 65, women above 60 .- .. -.-I ....- .-. - I I man inrl wnmnn mhn\rn rntirnmnnt inn ___I_ -- ------J --- ~ ~ I I , W , I OII" llvlllrll C4"""G , r r l l r , , , u l a r O y r relation to the working-age population should steadily increase. If, as planned, the government resumes in 2008 the increase o f the retirement age (until it reaches 65 for men and 60 for women in 2015), the upturn could be delayed until 2015 (see Figure 7.3 basedonNSMpensionmodel). 76 7.29 This favorable trend is reinforced by the relatively older age of the current pensioners. According to the NSIH database, old-age and disabled pensioners were on average 75 years old. This surprisingresult is partly due to the low number o frelatively young old-age pensioners less than 60 in the case of the women, and less than 65 inthe case of the men. They represent about 45% of the population 55 years and older but only 19% of the older pensioners. Many workers have probably delayed retirement because of the large income drop it implies. From a fiscal point of view, this observation means that little savings can be expected from reviving the programmed increase inretirement age. Such increase is, however, important because it guarantees that inthe future when pension levels at retirement improve, workers will continue to retire only after they reach the age of 65 and 60. 7.30 Based on these expected demographic trends, and given the expected results of the government strategy to increase revenues from contributions, Table 7.4 presents simulations of the pensionsystemfinancial balance inthe next 25 years. The system is constantly insurplus. 7.31 Obviously, these medium-termresults rest on many assumptions and unknowns about the economic environment, the contributory capacity o f the new workers generations and the profile of the future cohorts of pensioner^^^. Table 7.4: Medium-termPensionSystemFinancialBalance Sources: NSMdata andpensionmodel and author estimations. Notes: I/Estimatesbasedon the budget law indicationsaboutthe share contributionrate destinedto pensionprograms. 2/ Administration costs andother relatedbenefits (funeral, etc.) excluded. 7.32 The consequencesof the current large emigration flows are not expected to be significant in the short-term. The major impact will probably be in the long-term. Individuals currently leaving the country to work abroad are not likely to be contributing to the pension system at the moment. According to a MONEE study, half of the persons who live the country are less than 30, and many are from rural areas. Wage employment rate among these groups is rather low. In the long-term few returningmigrants will be entitledto pensions because the system requires at least 20 years o f contribution. Inthe long-term, the decline inthe number of births could accelerate the aging o f the population after 2020. This will translate into severe resource limitations for the pension system. Accordingly, the share of the revenues from contributions in the baseline scenario stabilizes around 9% of GDP inthe long-term. ~~ ~~~~ 46Simulations usingthe NSIHdatabase on contributors would produce more precise estimates of future retirees' pensions at retirement. 77 7.33 The assumption that the current trends in wages and employment will continue may appear optimistic. However, variants show that for the system to be in deficit in 2010, wage per year (instead o f + .6% in the baseline) starting 2006, and nominal wages growth rate would employment inthe administrationand the non-agricultural sector would have to decrease by 0.7% have to be lower than 10.8% per year (instead o f + 18% inthe baseline). The possibility of such a scenario seems unlikely since from 2000 until2004 wages (agriculture sector included) have been increasing inaverage by 26% per year and employment (agriculture excluded) has been growing by 1.1%per year. 7.34 On the expenditure side, the assumption o f strong wage growth leads to a relative fall in pensions inrelation to wages. As explained indetail in section 1, under the current pension law, a retiree's recorded wages or contributions are simply averaged in the computation o f pension at retirement. In a context o f strong recent wage growth, pension levels appear to be quite low in relation to worker's last earnings. Because this situation i s most likely unsustainable in the medium and long-term, Table 7.5 presents the results o f a variant, in which the system is upgraded and average pensions at retirement steadily increaseu. The results indicate that the system could keep producing surplus untilabout 2030. Table 7.5: Medium-termPensionSystemFinancialBalance Gradual increaseof pensionsat retirement Sources: NSMdata andpensionmodeland author estimations. 7.35 At a pace o f an equivalent o f 1 to 1.5 percentage points o f GDP per year, NSIH should accumulate a substantial level o f reserves. Instead o f producing surplus,policy makers could opt to reduce the contribution rate. Given the strong wage growth, this choice does not appear, however, pressing. Because o f the enormous investment needs inMoldova, these resources could make, by contrast, a valuable contribution to economic growth ifadequately invested. 7.36 Like that o f East European countries with mature pension systems, the contribution rate inMoldovais high.Thisrate couldbe progressively loweredto 23% around 2010, butbecause o f the probable rapid aging o f the population due to emigration it should have to be increasedagain a decade later. Because o f the strong wage growth, cutting the contribution rate does not seem urgent. International experience indicates that lowering contribution rates translates principally into higher wage increases. In some cases, it favors the employment o f unskilled workers, but does not appear to increase total employment significantly. 41 Pensionsat retirementare computedon adjustedpast wages combining50% inflationand 50 % wage growthThis method i s appliedto all the new retirees, civil servantsincluded. 78 7.37 There is no obvious need for policy makers to stimulate higher wage increase over and above the wage increases fueled by large capital inflows and domestic economic growth. By contrast, there i s a large need for higher fiscal savings inMoldova and growing pension reserves mightbe a convenient source of long-termfinancing. Insuch an environment, the pensionfund is forced to buy only non-tradable government bonds (like in the US) or it i s authorized to manage theportfolio givenpre-established government mandatesandrestrictions. 7.38 The major risk associatedwith the buildingof pension fundreserves is mismanagement. Moral hazard and agency problems typically lead to situations where funds are not always investedin the best interest o f the insured.Inthe case o f a pension fund the beneficiaries, that is the future pensioners and the future generations of taxpayers, cannot closely participate in the management of the trust fund. The government does have influence over the funds investment strategy but it has also more than one objective and time horizons. The government goals (such as therestructuringofabankruptedenterprise inan abandonedregion, socialhousingetc.) mightnot be in conflict with preserving the long-term solvability o f the pension system if "they are designed to correct market failures, address specific social or historical needs or capture some externalities". The fact that these social and economic goals can also easily serve political purposes (e.g., free riding of some politicians or government agencies) creates additional complications. The problems of corruption, shirking, and overall monitoring o f the trustees performance are also not negligible. 7.39 There are no obvious solutions to these problems. It i s advisable to compose a Board of Trustees that reflects the diverse groups and balance the different conflicts o f interest. The most successful experiences seem to be associated with pension fund tripartite boards with representatives from government employees and employers. Setting monitoring rules based on * "behavioral" controls-adequate procedure o f disclosure, presentations o f investment strategies, use of auditing etc.-are also recommended. Finally, transparent rules for decisions on public investment and expenditures are also certainly helpful. 7.40 Clear regulations and supervision rules on what NSMcan and cannot do with its reserves need therefore to be rapidly developed. The design o f institutional bodies to monitor NSM decisions and holdit accountable needs also to be created. E.CONCLUSIONS 7.41 ThepensionsysteminMoldovacame close to collapsinginthe late 1990sand as aresult, a major reform program was initiated to strengthen the system's long-term fiscal sustainability while promoting transparency and greater equity. The reform program remains incomplete and the systemfaces anumber ofcritical challenges: First,pensionincomes at retirement arevery low in comparison to current wages. Second, the participation in the pension system of a many contributors, including farmers and self-employed workers, i s largely subsidized. 7.42 In the first case, the strategy to upgrade the system has not yet been presented. Policymakers should analyze what upgrading strategy would be the best for Moldova. To be fiscally sound, the envisaged changes must affect only the new generations o f pensioners, so the upgrade of the system increases pension expenditures, but gradually. The size o f the upgrade depends, moreover, on the system's capacity to collect revenues. In the second case, the government has developed a draft strategy to reform the pension system inagricultural sector and i s moving towards the unification o f the contributory system. Becausepoverty rates are extremely high among farmers, a gradual phasing in of new regulations should be considered. Self- employedworkers, more than three quarters of whom are subsidized, have the lowest poverty 79 rate. Because leakages are likely in this group, a more rapid enforcement o f new regulations, including the payment o fthe desired minimumcontribution, i s desirable. 7.43 With respect to the system's long-term financial sustainability, prospects are generally favorable but a number o f important risks remain. Over the last four years, the National Social Insurance House (NSIH) has enjoyed a positive economic and demographic environment. Although the NSIH experienced a deficit at the end o f 2005, the pension system i s currently on sounder financial footing with accumulated reserves o f about 1.6 percent o f GDP. Simulations also suggest that low pension benefits and favorable demographics trends should effectively allow N S M to keep producing surplus and accumulating reserves inthe medium-term. However, as the dependency ratio deteriorates, the long-term financial viability o f the system will depend on how the reserves are managed and secured for the next generation. Setting up the regulations for this purpose i s critical, given that moral hazard and institutional problems typically lead to situations where funds are mismanaged. The most successful experiences are associated with pension fimd tripartite boards, representing diverse groups and interests. 80 8. SOCIAL ASSISTANCE BENEFITSINMOLDOVA A. INTRODUCTION 8.1 Before independence, Moldova's comprehensive and centralised social welfare system was principally designed to assist those unable to work (illor disabled) and children inneed. The structure of the social welfare system was characterised by an emphasis on categorical benefits and little attention to household welfare as work was guaranteed to all and poverty was a relatively small phenomenon. In this context there was little need to distinguishbetween social insurance and social assistance given that the state owned all means o f production and was expected to cater for all individuals. 8.2 The collapse o f the command economy necessitated reform of Moldova's social welfare system. One of the first reforms was to nominally split the social protection system into two separate branches including social insurance and social assistance. The economic crises of the 1990s saw a significant reduction in the income maintenance capacity o f social assistance benefits due to hyperinflation, the suppression o f some benefits or some categories of beneficiaries and the growth of arrears inpayments. By the end of the 199Os, the social assistance system had a very marginal role in protecting households from poverty. With resumption of economic growth and Government resources, spending on social assistance began to increase. The also Government began to re-establish some elements a "Soviet-type" system o f social assistance with the reintroduction of a number of measures designed to increase categorical benefitsparticularly for the disabled and the war veterans and whichresultedina general increase insocial assistancetransfers. 8.3 The backdrop of such reforms was an exasperation of the already evident mismatch between the scope of social assistance outlined in government stated intentions and government actions. On the one hand, the Government o f Moldova claims that social assistance is a system designedto assist the poor and the vulnerable which is in line with the scope of social assistance in most countries worldwide and in the region. On the other hand, as it will be argued in this chapter, the social assistance system i s de facto a categorical system where the poor are not a privilegedtarget category. Relying on categorical benefits i s not a problemper se if there i s a clear distinction between those benefits to be granted to special categories of citizens inneed such as the disabled and those benefits to be granted to poor families based on means tests. Most countries inthe world rely heavily on categorical cash benefits48,but Moldova stands out as one o f the few countries that rely exclusively on categories while claiming that the system is designed for the poor. It i s the total mismatch between stated intentions and real actions that makes of Moldova an exceptional case. 8.4 This chapter makes a summary evaluation of social assistance benefits in Moldova, assesses the financial viability and sustainability of the system, provides some possible future scenarios and proposes possible reforms compatible with the Government reformagenda. 48 In a study of 47 countries, Coady, Grosh and Hoddinott (2002) find that out of 87 types of cash benefits provided, 68% are categorical, about 26% are means tested or proxy means tested and the rest are community based assessed. However, the study considered all cash benefits under the social protection systems and notjust social assistancebenefits which are generally means tested. 81 B. SOCIALASSISTANCEEXPENDITURES 8.5 There are several clussificutions o f social assistance benefits in Moldova used by different organizations including the MOF, the Statistical Bureau, the MHSP and the NSIH. As we collected data from all these organizations, we first reconciled these different classifications into one homogeneous list o f benefits which include the following nine items4' 1. Utilities compensations 2. Child indemnities 3. War veterans allowances 4. Social allowances 5. Death grants 6. Chernobyl compensations 7. Care takers allowances for disabled 8. Transport compensations for disabled 9. Material assistance 8.6 The state budget pays for most o f the benefits with the exceptions o f transport compensations for the disabled which i s paid out o f budgets o f Administrative Territorial Units (ATU or local budgets) and material assistance which is mostly financed from the Republican Fund and local funds. Inreality, transport compensations for disabled are financed by the local authorities that are partly financed by the central state budget and material assistance i s financed by the Social Support Fund(SSF) o f the population which also receives some contributions from the state budget. The administration o f social assistance benefits i s mostly handled by the NSM (centrally and locally) with the exceptions o f the transport compensations for disabled managed by the MHSP and administered locally by the local public administrations and material assistance which i s controlled centrally by the SSF and administered locally by the local public administrations. 8.7 Social assistance benefits are dl categorical with only two exceptions, child benefits for children in age 3-16 and material assistance. However, both these benefits are not means tested benefits as such but they are categorical benefits accompanied by means tested filters. Also, most benefits cover both the insured and the non insured persons and the distinction between social insurance and social assistance i s unclear interms o fwhich benefit falls under what system and in terms o f budget subdivisions. These are two features o f the system evidently in contrast with the basic principles o f social protection. 8.8 In2005, total expenditure on social assistance benefits was approximately 538 millions lei equivalent to 1.7% o f GDP and 4.8% o f public e~penditure.~~ As a rule o f thumb, expenditure on social assistance inpercentage o f GDP i s typically between one and two percent and Moldova fits into this range. The largest benefit i s compensations for utilities payments representing about 37% o f total expenditure and this i s followed by child benefits (21%), war veterans allowances (13%), social allowances (8%) material assistance (8%), transport compensations for disabled (5%), Chemobyl compensations (3%), death grants (3%) and care-takers allowances for disabled 49 A list ofkey codes to reconcile the various classifications is provided inVolume I1of this report. 50 Note that these are the figures taken from the state budget and provided by the NSIH. The NSIH also produce expenditure reports based on administrative records contained in the NSIH database which are somewhat different from the state budget data. As these last figures are provided by category of beneficiaries, these are the figures usedfor the macro assessment o f individual benefits discussed further inthe chapter. 82 (2%). Expenditurehas increased for all benefits and inreal terms between 2000 and 2005 and the largest benefits are among those that increased the fastest. Expenditure on child benefits have increased by 335%, utilities compensations by 136% and social allowances by 73% during the period while war veterans allowances increasedby about 33% since their introductionin2001. 8.9 Efficiency of Social Assistance Spending. Administrative costs for social assistance are not very highby world standards but they are highrelatively to the administration o f pensions. In 2005, expenditures for administration o fN S M constituted51.8 million. lei in administrative costs o f which about half was for salaries, 20% for goods and services, almost 13% for social contributions and 9.4% for maintenance with the rest being distributed equally among other items including medical insurance, travels, interests, capital investments and purchase o f fixed assets. If we split these costs between social insurance and social assistance according to the number of beneficiaries, the administration o f social assistance benefits consumed about 46.5% o f this budget which i s equivalent to about 5% o f the total expenditure on social assistance benefits managed by the N S M . The annual administrative cost per beneficiary for social assistance can be estimated at 46 lei for 2005. These are not great amounts and far smaller that comparable figures for OECD countries but it should be noted that social assistance consumes much more than the administration o f pensions relatively to the respective budgets. The number o f beneficiaries in these two branches is similar but pensions are much higher than social assistance benefits on average. 8.10 Disabled, war veterans and families with children are the largest groups o f beneficiaries. All benefits have multiple categories o f beneficiaries and several benefits include the same categories o f beneficiaries such as the disabled, children or war veterans which appear under several benefits. Ifwe group all benefits by type o f beneficiaries we see that the disabled are the most numerous and receive by far the largest share o f total expenditure. This group i s followed by children, pensioners and elderly and war veterans. Although many o f the benefits for the disabled are not specifically targeted for children or war veterans these two categories account for a good share o f the benefits paid to the disabled which makes children and war veterans the largest categories o f beneficiaries together with other adult disabled. The poor as a category appear only once across all social benefits, under material assistance, and the share o f beneficiaries i s very small. 8.11 Social assistance benefits are a relatively small source o f income for all households and only a marginally more important source for poor households. According to the Household Budget Surveys (HBSs) the main source o f earnings for households i s wages followed by social insurance (pensions) and by personal transfers from friends and relatives, including remittances. Agricultural income comes infourth place and this i s followed by finance which includes various types o f financial income. Social assistance benefits come only in seventh place in terms o f size. In2001, they contributed for only 1.3% of total income for all households on average and 2.4% for poor households and these shares increased by 2004 to 2.1% and 5.1% respectively. Therefore, even in 2004, these benefits were not a very relevant source o f income for poor households. Micro data also confirm that the largest social benefits received by households are utilities compensations, child benefits, war veterans' allowances and social allowances in this order. In 2004, these four types o f benefits accounted together for over 90% o f all social assistance benefitsreceivedby households. 8.12 Coverage has expanded for all benefits between 2001 and 2004. The coverage o f utilities compensations has significantly increased by more than three folds for households and by almost five times for individuals. By 2004, 15.3% of households and 6.4% o f individuals received at least one compensation for utilities payment. Among utilities compensations, the largest coverage 83 o f the population i s by gas and electricity compensations. Child benefits (or benefits for families with children) coverage also increased for both households and individuals reaching 4.3% and 1.7% o f the population respectively in 2004. Benefits for children in age 3-16 are those with the largest coverage among child benefits. War veterans, social allowances and all other benefits bundled together also increased in coverage several folds. In 2004, social allowances covered about 3.2% o fhouseholds andwar veterans' allowances about 2.9%. 8.13 The increase in coverage has been accompanied by a mixed performance in terms of average benefit in real terms which, on average and for all benefits, has declined between 2001 and 2004. Average benefits paid varied significantly in real terms between 2001 and 2004 and across type o f benefits. Utilities compensations decreased from an average o f 107 lei per month in 2001 to 50 lei in2004 and war veterans' allowances from 313 lei to 139 lei. Instead child benefits increased inreal terms from 68 lei per month to 98 lei and so did social allowances from 54 lei to 86 lei. 8.14 Some people and some households benefit from multiple transfers and this phenomenon has increased inrecent years. Social benefits are not mutually exclusive by design. Most benefits are categorical and some persons can fall in several categories. For instance, the disabled and the war veterans are included as categories o f beneficiaries in almost all types o f benefits. In2001, only 11% o f households' beneficiaries and 8.5% o f individuals were receiving more than one benefit but this share increasedto 28.6% for households and to 24.9% for individuals by 2004. 8.15 The information available from the HBSs does not allow for a proper evaluation o f targeting. A rough evaluation o f targeting was only possible for child benefits where it i s evident that these benefits suffer from both leakage and under-coverage. For example, we estimated a leakage rate for benefits for children below the age o f three o f 41% in 2004 and a corresponding under-coverage rate o f 68%. The respective rates for children in age 3-16 were 39% and 90% respectively. Adjustments to the HBSs questionnaires would be needed to make a more accurate evaluation o f targeting but if these figures are confirmed these are very highleakage and under- coverage rates by international standards. 8.16 Equity of Social Assistance. If social benefits were to be distributed completely randomly we should expect to see a similar share o f benefits reaching each quintile, which i s about 20% o f total expenditure for social assistance per quintile. Inthis case the system would be welfare improving and distributional neutral. If, instead, benefits were distributed according to household income, with lower income households receiving the greater transfers, we should expect to see lower quintiles receiving a relatively higher share o f benefits. Such a distribution o f resources would be a progressive type o f distribution. If, on the other hand, richer quintiles receive the greatest share o f social benefits, then the system would be regressive, it would distribute public resources infavour o f the rich. 84 85 Figure 8.2: Distributionof SocialBenefitsby IncomeQuintile 100% 80% 60% 40% 30% 1 3 H u H 20% 20% 10% 0% 0% 2001 2002 2M13 2M34 Utilities Chiid Warvst Socaii Caredis Transpdis Malass Source:HBSs 2001-2004: Rightpanelrefers to 2004. 8.18 These findings are generally supported by parametric estimates on the household probability o f receiving benefits, with some caveat^.^' Utilities compensations are biased towards rural and larger households and towards households with older age members. The probability o f receiving child benefits i s affected positively by household size and negatively by the age o f the head o f the household. These effects are both expected given that child benefits are provided for children and given that children tend to be concentrated in younger households. However, controlling for these variables shows that child benefits are in fact progressive with poorer households being more likely to benefit than richer households. War veterans allowances are biased towards households where the head o f the household i s older which i s expected given the nature o f the benefit. However, despite controlling for this fact, the benefit i s highly regressive with upper quintiles being more likely to receive this benefit. For social allowances, household size seems to matter the most inthe allocation o f this benefit. Larger households are more likely to get this benefit. Also this benefit i s still regressive after controlling for household characteristics although the significance o f this regressivity is less marked than for war veterans. Inconclusion, parametric and non parametric estimates concord in establishing that all benefits are regressive with the only exceptionof child benefits which regressivityturns into progressivity once we control for householdcharacteristics. 8.19 The poverty incidence o f social assistance benefits i s small on average but significant for the very poor. As we already know from various studies made on poverty in Moldova, all the Foster-Greer-Thorbecke (FGT) poverty measures have significantly declined between 2001 and 2004.52In2001, the impact o f social benefits on poverty was very marginal, estimated at 0.6% o f the headcount ratio and increasing to 2% for the poverty gap ratio and 3.7% for the severity o f poverty ratio. Looking at the very poor, these impacts are larger between 1.1% for the poverty ratio and 5.6% for the severity o f poverty index. The impact i s small but increases significantly 51 Resultsofthe parametric estimates are availablein Volume I1of this report. 52See for example: GOM (2005) Economic Growth and Poverty Reduction Paper, Government of Moldova, wired at http://www.scers.md; MET (2004) Poverty and Policy Impact Report, Ministry of Economyand Trade; World Bank (2004) Recession, Recovery and Poverty in Moldova, Report No. 28024-MD and World Bank (2004) Moldova, Poverty Reduction Strategy Paper and Joint World Bank-IMF Staff Assessment, Report No. 29648. 86 amongst the poorest. Moreover, the elasticity of poverty incidence to social benefits increases over time. By 2004, social benefitsmanaged to reduce the headcount ratio for the poor by 4%, the poverty gap by 9.3% and the severity of poverty index by 15.2%. For the very poor these changes are much greater, 6.8% for the headcount ratio, 13.8% for the poverty gap and 23% for the severity of poverty. Therefore, social benefits for the extremely poor (the poorest among the very poor) have a positive and significant effect on welfare. Table 8.1: FGT PovertyMeasuresWith andWithout Social AssistanceBenefits Poor Very poor 2001 with without diff % with without diff % Headcount ratio 68.5 68.9 -0.6 54.4 55.1 -1.1 Povertygap ratio 29.1 29.7 -2.0 20.0 20.7 -3.2 Severityof poverty 15.6 16.2 -3.7 9.7 10.3 -5.6 2002 Headcount ratio 53.9 54.9 -1.8 39.8 40.9 -2.7 Povertygap ratio 20.0 21.1 -5.2 12.5 13.6 -7.8 Severityof poverty 9.7 10.6 -8.8 5.4 6.3 -13.5 2003 Headcount ratio 43.0 44.5 -3.5 26.5 28.4 -6.8 Povertygap ratio 13.1 14.3 -8.6 6.8 7.8 -13.6 Severityof poverty 5.5 6.4 -14.6 2.5 3.3 -23.9 2004 Headcountratio 40.9 42.6 -4.0 26.0 27.9 -6.8 Povertygap ratio 12.6 13.9 -9.3 7.0 8.2 -13.8 Severityof poverty 5.4 6.4 -15.2 2.8 3.6 -23.0 Source: HBSs 2001-2004. 8.20 We have also regressed, the poor and the very poor (as dummies) against the main four types of benefits (also as dummies). This particular procedure, allow us to see what benefits are more likely to accrue to the poor and to the very poor controlling for other benefits, or in other words, controlling for the multiple benefits effect already described. Among social benefits, utilities compensations and war veterans' allowances tend to be pro-rich while child benefits and social allowances tend to be pro-p~or.'~ In2001 and as compared to the non poor, the poor were less likely to receive utilities compensations and war veterans allowances and more likely to receive childbenefits and social allowances. The same was true for the very poor. Throughout the years, the negative likelihood o f receiving utilities compensations and war veterans' allowances for the poor disappears while it i s maintained the positive likelihood for child benefits. For the very poor, the negative bias disappears for war veterans and is weakened but still significant for utilities compensations while the positive biases for child benefits and social allowances are maintained. 8.21 In substance, the overall impact of social benefits on poverty is rather small but not negligible for the very poor. We observed in previous sections that the share o f income constituted by social assistance benefits i s small for all quintiles but also that it was larger for lower quintiles. Results on the poverty incidence suggest that the relatively small changes in income determined by social benefits are sufficient to close a certain share o f the poverty gap for s3 Note that we are not testing single benefits regressivity and progressivity here as we are controlling for all benefits together. 87 the very poor. The equity analysis and the poverty regressions presented concord in establishing that child benefits and social allowances are the two benefits that are more likely to contribute to thisphenomenon. c. OPTIONS FORREFORM 8.22 As discussed in previous sections, expenditure on social assistance benefits (including total expenditure as a percentage o f GDP and government expenditure, expenditure per beneficiary and administrative cost) i s not very large by any standards and cannot be expected to address the poverty problem alone. In2001, the estimated cost to fill the poverty gap54was about 3,264 m. lei per year (272 m. lei per month) for the poor and 1,756 m. lei per year (146 m. lei per month) for the very poor. These amounts decreased between 2001 and 2004 as the poverty ratios declined and in 2004 the gap for the poor was down to 1,780 m. for the poor and 785 m. for the very poor. In 2004, these shares were equivalent to 3.3 and 1.5 times social assistance expenditure, to 5.6% and 2.5% o f GDP and to 15.8% and 7% o f govenunent expenditure respectively which are far greater amounts than the existing social assistance budget. 8.23 There is, however, a great mismatchbetween what the system i s supposed to do - support the poor and the vulnerable - and what the system actually does - support selected categories of beneficiaries irrespective o f income. Therefore, we have little ground to argue that social assistance expenditure should be simply increased but plenty of elements to argue that the administrationand delivery o f social assistance benefits has to be improved before any increase in budget expenditure can be considered. The question i s how this should be done and what the expected outcomes of such changes are. Table 8.2: PovertyGap 2001 2002 2003 2004 Poverty gap (m. lei) Poor 3264 2348 1721 1780 Very poor 1756 1151 691 785 % of GDP Poor 17.1 10.4 6.2 5.6 Very poor 9.2 5.1 2.5 2.5 % of Government Expenditure Poor 54.3 30.4 18.8 15.8 Very poor 29.2 14.9 7.6 7.0 Average poverty gap per capita (lei) Poor 899 649 476 494 Very poor 484 318 191 218 Average povertygap per poor (lei) Poor 1313 1204 1108 1207 Very poor 889 799 721 837 Source: HBSs 2001-2004. 8.24 A couple o f simulations may help to better understand the possible implications of reforms. We can see what would happen if the system i s kept as it is with the forecasted changes inmacroeconomic variables (scenario 1) and what we shouldexpect insteadfromaredistribution exercise (scenario 2). Forecasts are based on a common set o f basic indicators taken from the 54The poverty gap i s the cumulated distance that separates each poor from the poverty line. This gap is therefore an estimate o f the resources required to bring each poor up to the poverty line (simulating the elimination of poverty altogether). 88 Medium Term Expenditure Review (MTEF) 2005-2008 and 2006-2009 including estimates o f GDP, budget expenditure and inflation. 8.25 In Scenario 1, no reforms will be implemented and the social assistance system will continue as it was in 2005. Expenditure projections are made by simply interpolating past expenditure by type o f benefit and based on the basic forecasting indicators. Under this scenario all benefits will tend to decline as a percentage o f GDP and as a percentage o f Government expenditure with the result that total expenditure will significantly decline from 5% to 3.7% o f government expenditure between 2003 and 2009. In such scenario the problems described with the system in terms o f targeting, equity, efficiency and poverty incidence will persist and the reducedrelative expenditure is likely to reduce even further the impact on the very poor. Figure8.3: ExpenditureForecast(Scenario 1) X GDP 2.000 I+Childinsured) benefits (insured and man 1.800 Waf veteram dlowances 1,600 +Socid allowances I 1.400 II1.200j --+-Death grarts A 1.000I/\ 3w 4-Cherncbgcmpersations 0.8004 / I, 0 600 2w +Caretakers allowancesfw disabled 0400 0 200 --Materialxwstnce 0 000 OW 2000 2W1 ZOO2 2003 20W 2005 20C6 2007 2008 M09 2WO 2CQ1 20322M3 2W 2005 2006 2007 2008 2009 - -Told Sources: MediumTermExpenditureReview 2005-2008,2006-2009 andPoverty ReductionStrategyPaper,October 2004. 8.26 In Scenario 2, real expenditure is kept constant inreal terms at the 2004 level and the government manages to improve targeting and redistribute all social benefits given to the top quintile to the bottom quintile. The redistribution i s achieved by spreading the total amount previously given to the top quintile over all persons belonging to the bottom quintile in equal shares (people in the first quintile were all very poor in 2004). Inthis case, there i s no effect on the poverty headcount index, neither for the poor nor for the very poor. This measure would not move any poor out o f poverty or any very poor out o f deep poverty. This is due to the fact that in 2004 the total amount available to redistribute from the top quintile was only about 5 m. lei per month (0.19% o f GDP in annual terms) which was an average o f about 7 lei per capita. Transferring these amounts to each person inthe bottom quintile does not bring anyone above the poverty line or above the extreme poverty line because all people in the first quintile are below the extreme poverty line by more than 7 lei. W e have instead a marginal effect on the poverty gap and on severity o f poverty indexes because these two indexes consider the distance between each poor from the poverty line. This distance i s reduced for those o f the first quintile and these indexes are also reduced. 8.27 Itis evident that the existingbudget for social assistance has a limited potential for lifting all the poor out o f poverty. In the best possible scenario with perfect targeting, full vertical efficiency and all benefits means tested only, the 2004 social assistance budget could have filled 89 the gap for about two thirds o f the very poor only. However, this would have implied a reallocation o f benefits from the non extreme poor to the extreme poor which would have moved some o f the existing poor into deep poverty reducing significantly the impact on extreme poverty. Moreover, perfect targeting would imply a perfect knowledge o f household income which i s where the Government o f Moldova i s very weak. Therefore, solutions to improving the social assistance system are more complex, have to be found in a combination o f multiple actions and are necessarily long term. D. POLICYRECOMMENDATIONS 8.28 We look first at what the government intends to do, we then see whether it can be done and on the basis o f this assessment, we propose alternative or complementary solutions. The focus is on the financial efficiency and poverty alleviation capacity o f the system 8.29 The Government o f Moldova (GOM) has repeatedly stated the necessity to reform the social assistance system in various social reports over the years. Many o f the shortcomings that we have highlighted in this chapter have been already identified by ministerial staff and reported on different occasions in formal and informal meetings and reports. The Government has clearly understood that social benefits are not achieving the main objective they are designed for: support the poor and the vulnerable. 8.30 Inorder to reformthe social assistance system, the GOMwent through a number o f key steps. In December 2003, an inter-ministerial commission charged with implementing the new social protection strategy was established and in December 2004 the Government approved the Economic Growth and Poverty Reduction Strategy Paper (EGPRSP) which outlined a number o f objectives and the long-term strategy for social assistance reforms. The 2004 Social report published in 2005 explained more in detail the main features o f the reform and this plan was further developed and finally approved with Government resolution No. 151-15311221 "On Streamlining the Social Assistance System", dated November 1l',2005. Box 4 shows the main guidelines o fthe government reform strategy. 8.3 1 The adoption o f the new system should require the implementation o f four main steps: A first step where the appropriate filters and scoring mechanism are selected and put into law; a second step with the implementation o f a pilot project aimed at testing the efficiency o f the filters in selected areas; a third step for introducingthe new system at the national level; and a fourth step when these filters will be assessed and revised on a regular basis. Step one and two are already intheir implementation phase while step three i s expected to be implemented in2007. 8.32 On the surface, the government reform strategy is sound and would be able to address many o f the shortcomings o f the social assistance system that we have identified in this chapter. However, some o f the elements that will cover the gap between intentions and actions have yet to be defined insome important areas. 90 Box 6: The GovernmentReform Strategy A. Longterm pillars (EGPRSP) 1. Monitoring individualprogramsin order to improve eligibilitycriteria and rationalisetypes, forms and durationof benefits; 2. Provision of social benefits to the poor which will require the development and approval of poverty criteria and a household welfare evaluationmechanism; 3. Creation of a separatesocial assistance budget; 4. Developmentof a unified database managementsystem for social assistancebeneficiaries: 5. Revisionof the legalframework; 6. Improvementsin the institutionalcoordinationmechanisms. B. Medium term actions (EGPRSP) 1. Monitoringof the social assistance programscarried out in the framework of the EGPRSPimplementation; 2. Gradual introductionof a means tested system; 3. Legislation,povertyassessmentand householdwelfare assessmentmechanismsdesignedwith external assistance; 4. Separationof the social assistancebudgetfrom the state social insurancebudget; 5. Institution of a single databasefor social assistance. C. Specific recommendations (Govt. resolution No. 151453112211 1. Householdwelfare should be assessedwith a mixed system made of means-testedmechanismsaccompaniedby proxy-means filters. 2. Well-being should be assessedbased on income and non income criteria looking at living standards in their complexity. 3. Urban and rural areas should be assessed using differentcriteria. 4. The SubsistenceMinimumestablishedby law annually should be used as a benchmarkfor establishingbenefits' size. 5. Household incomeshould be made of all sources whether in cash or in kind includingsocial services provisions. 6. Benefits earned for special state merits should be separated from benefits earned on the basis of low income. This should be achieved by reducing the number of benefits and categories of beneficiaries and by reducing the possibilityof earning multiple benefits. 7. a. Household incomeassessmentshould be initially made annually and later, if possible,every six months. Two pillars of the new system are the income assessment and the improvement of the accessibility to social assistance. This means that the role of local authorities in the implementationof the new system is essential. The new figure of social assistant introduced by law in 2003 should be strengthened. 9. The role and functions of the social assistant should be clarified and a new questionnaire for social assistance should be developed. The database for social assistance beneficiariesshould be unified and for this purpose it is necessaryto implement two steps: 1) Separation of responsibilitiesbetween social assistance and family protection departments, who will be charged with the materialassessmentof the families, and the NSlH who will be charged with keeping a separateand unified database of social assistance beneficiaries;2) The social assistance and family protection units will take full responsibilityfor the provision, payment and tracking of social service provision. 10. The changes proposedwill require changes in the finance mechanisms and allocations of resources.These changes should be implementedin the frameworkof the MediumTerm ExpenditureFramework(MTEF). 11. The new system will be tested with a pilot project. A. Revisionof thelegalframework and transition to a means testedsystem 8.33 There i s not much in the government reform agenda on the legislative changes that need to be introducedto reformthe system. Yet, this is essential and preliminary to most o f the reforms suggested. Some actions such as defining the methods for means-testing can be implemented with government resolutions but some others such as changes inthe benefits structure will require new legislation. Moreover, the introduction o f new legislation will require a clear vision o f how the social assistance benefits system should look like after the reforms. 8.34 The reform strategy i s also silent on the categories and benefits that need to be suppressed in order to move from a categorical to a means tested system. As it stands, it would seem that means tested filters will accompany each category o f benefit but that all categories and all benefits will remain even if the strategy calls for a unification o f benefits. It i s also clear from the various exchanges we had on this topic with government officials that, so far, there has been little discussion on which benefits and categories should be discontinued and when. This i s a problem also because the question o f multiple benefits beneficiaries cannot be addressed without the streamlining o f existing benefits. 91 8.35 The lack o f clarity about legal changes and about the shift from a categorical to a means- tested system may hide a more substantial political obstacle. It i s understandable that announcing the elimination o f categorical benefits i s not a politically sound move and that people will need a great deal o f information about the shift from one system to another to accept such changes. And even then many people who are not poor and are currently benefiting from social transfers will resist such changes. For this reason, it i s not guaranteed that parliament will endorse the refoms. The current government has been elected and supported by former communists and most o f the changes made in the social assistance sphere in the course o f the first mandate have been in the direction o f stepping back towards a Soviet-type system o f categorical benefits focused on war veterans and the disabled. This i s what clearly emerged from this study. It i s hard to see how parliament will be easily convinced about the necessity to shift from a categorical to a means tested benefits. Inother words, while ministerial staff and even some ministries may be convinced o f the necessity o f the reforms, these same reforms may be dumped by parliament when submitted as laws. Rather than sidelining the problem, the solution would be to draft a calendar for change which clearly indicates which benefits and categories should be discontinued and when and bringthis agenda inthe public domain as soon as possible. B. Administrative division between social insurance and social assistance 8.36 The reform strategy includes the separation o f the social insurance and social assistance budgets and the separation o f the two database systems. As already described, at present the N S M manages social insurance and social assistance jointly. This reform i s long overdue and would make social assistance consistent with its scope. However, while the N S M i s an established and experienced body there i s no equivalent for social assistance precisely because social assistance i s largely managed by the N S M . The institutional capacity needs to be established and developed. 8.37 Creating a new social assistance agency i s costly and the Government itself has suggested keeping the administration o f social assistance under the responsibility o f the N S M . Inthis case, the NSJH becomes a social protection rather than a social insurance institution and needs to operate a clear distinction between social insurance and social assistance in-house by creating separate departments with separate budgets and data management systems. This i s perhaps the most feasible and least costly solution at present. It would also facilitate the needed division between the insured and non-insuredfor children and disabled benefits. Keeping the insured and the non insured under the same agency certainly simplifies the monitoring and coordination o f these benefits. 8.38 It i s also suggested that benefits eamed for special state merits should be separated from means tested benefits. We argued that this was a problem and the government agrees but there are no suggestions on alternative management solutions. If war veterans' benefits are separated from social assistance they can either stand on their own and be administered separately or they have to go under social insurance. Given that there i s no ground to justify the presence o f war veterans' benefits under social insurance, the most sensible option i s to keep the management o f these benefits separate from both social assistance and social insurance. The same argument applies to benefits designed for extraordinary calamities which will have to stand on their own under a separate `humanitarian' chapter. Such changes are mostly simple nominal reallocations and do not really require additional finance or extra administrative costs provided that the N S M continues to administer all benefits. If this i s the case, then the scope o f the agency should be redefined as social protection rather than social insurance agency and the budget and administrative subdivisions should be clarified. Any alternative which will require the establishment o f new administrative bodies i s likely to be more expensive. 92 C. Develop a means-tested system of welfare evaluation 8.39 This is one area where the government agenda is more specific. The November 2005 resolution explains in detail and compares various technical options for evaluating welfare and it i s suggested to use a combination of means and proxy-means testing together with the required documentation. There is a government regulation (No. 1084, 4* October 2004) which explains how to assess household and individual income. This will need to be updated but some o f the main guidelinesare inplace. There i s also an on going pilot project which is testing with means- tests filters applied on utilities compensations. The results o f this test are expected to guide the government with further reforms. This i s perhaps the area o f social assistance reforms where the government i s more advanced with. If implemented, the proposed reform i s expected to improve significantly equity and vertical efficiency. D. Enhancement of the role of local administrations 8.40 Shifting from a categorical to a means tested system requires the strengthening o f local capacity in the functions o f income assessment and delivery of benefits. This strengtheningi s in the reform agenda and includes the increase of responsibility of the new figure of social assistant introduced in 2003 and the separation of the social assistance and family protection units at the local level. The family protection units are expected to carry out the household welfare assessment while the social assistance units are expected to administer the benefits. These reforms are certainly needed and will require an initial investment in training and in the administrative reorganization at the local level. What i s still missingi s the detailed schedule and costing o f these activities. E. Improve monitoring and evaluation of social assistancebenefits 8.41 Inorder to be able to properly monitor social assistancebenefits some additional actions need to be outlined after the determination of the new set of benefits that should substitute the current structure. First, the NSM, Statistical Bureau and MHSP should agree on a single terminology of benefits which i s also compliant with existing legislation and the state budget. Second, the HBSs questionnaire needs to be adjusted so that questions are able to identify each possible category of beneficiary and each possible type of benefits. Third, ajoint monitoring and evaluation unit between the statistical bureau, the MHSPand the NSMshould be established and trained to use the HBSs and macro data for such evaluation. As a result of these measures a monitoring and evaluationreport should be publishedon a regular basis. F. Other issues 8.42 The sequence of reforms is also important and the legislative changes that will be necessaryto reduce the number o f benefits and to shift toward an income based system should be prior to the administrative changes which require a lengthy phase o f adjustments o f the organizations involved. Again, the political obstacles described should not be undervalued in the process of reforms and the existing political knots (if any) should be untied before the administrative reforms are undertaken. 8.43 The government also faces a complex financial dilemma. On the one hand, even if the government will manage to keep social benefits expenditure at the same level of 2004 as a percentage o f government expenditure (and achieve in this way a real growth of expenditure of over 64%) and even if the government i s extremely good in improving targeting and distribution, the social benefit budget will still be small. It was shown how a simple redistribution of social 93 benefits from the top to the bottom quintile will not affect the poverty headcount. We should not expect social benefits, under any possible scenario, to be a major force for poverty reduction in the medium term unless the government budget increases expenditure very significantly. On the other hand, expenditure should not be increased until the government manages to improve targeting and distribution. Increasing expenditure with the existing distributive system will increaserelative poverty, not decrease it. 8.44 These considerations suggest that solutions are complex and long-term. Redistribution cannot occur with drastic measures but needs to be aJine tuning exercise. Simply canceling benefits to the rich i s not feasible in the short-term and it i s irrelevant to give micro amounts to the poor. We suggest instead a two-tier long-term approach: 0 The focus should be first and foremost on the verypoor and the amounts should be adequate to close the gap with the extreme poverty line. This i s because this group i s the most disadvantagedby definition and because we showed that the poverty rates o f this group are relatively more sensitive to changes in social assistance benefits. Activities aimed at improving the distribution o f benefits should be focused on increasing the transfers to the very poor by adjusting first the benefits with the worst performance. Utilities compensations and material assistance should be the benefits to reform first followed by social allowances and child benefits in this order. The objective here i s to lift the very poor out o f deep poverty. The cost o f this activity has been estimated in Table 8.1 and this cost could be covered in part via redistribution and inpart via the estimated growth inreal expenditure. 0 The next focus group should the poor who are not extreme poor. For this group, the focus should be on the transition toward the formal social insurance system. In a sentence, active labor market policies instead o f social assistance. This i s costly but could be financed partly out o f the real increase in social assistance expenditure keeping this expenditure constant as a percentage o f Government expenditure and partly out o f the budget for ALMPs. These policies should be specially tailored for the poor with means-tests and linked to social assistance benefits for example by offering to social assistance beneficiaries the alternative o f a public work. Public works, social funds, work for food schemes and micro-credits for start-up businesses would be some o f the possible schemes to utilize. This policy should achieve a double objective: stop the current outflow o f people from social insurance and free financial resources to be reallocated to the very poor. An estimate o f costs and benefits o f this proposal should be made once the specificities o f the ALMPs have been defined. 8.45 In the meantime, there are other short and medium term actions that may prove to be effective from a financial perspective. One i s a thorough evaluation o f the disability assessment procedures. The disabled are the largest category o f beneficiaries and the single category with the largest number o f benefits. Being diagnosed with a disability gives automatic access to a range o f benefits and the diagnosis i s often an arbitrary decision based on medical evidence. It i s no mystery that disability commissions worldwide are prone to corruption, collusion or mismanagement. This does not have to be the case inMoldova but reviewing procedures and test existing disabled on a random basis i s a useful tool to make sure that the system i s healthy and there i s no waste o fresources. 8.46 A second grey area is material assistance. This is the largest benefit in terms o f beneficiaries and the thirdinterms o f expenditure. It i s mostly extra-budgetary and for this reason 94 reforms o f this benefit will not lead to automatic state budget savings. However, revenues are growing, they are property o f the state and it i s expected that these funds shouldbe well spent. At present, the central management o f the benefitwhich i s the social support fund of the population i s very weak and very little transparent while there i s a wealth o f data at the local level which have never been mined. Inthis case, a thorough analysis will have to start with the data collection and with an assessment of national financial procedures and local administrative practices. This exercise may lead to findings that can help to improve the financial administration of the fund. Also, given the experience accumulated by the local staff in visiting households and compiling the income based questionnaire, this benefit is the ideal candidate for experimenting with a full means tested system. E.CONCLUSION 8.47 Insum, the government reform agendaprovides a good platform for reforms but should be complemented with a number o f short and medium-term actions to achieve the proposed long- term objectives as follows: Short-termactions Define the future structure of social assistance benefits (number of benefits, categories, etc); Define a more specific calendar for the discontinuation o f existing categorical benefits and the transition to a means-testedsystem; Draft the legal changesrequiredand open this draft to public debate; Evaluationof the disability assessment procedures; Review of the RepublicanSocial Support Fundand LocalFunds. Medium-termactions Submit legislative reforms to parliament; Reorganise the NSIH into a Social Protection agency with administrative and budget subdivisions for social insurance and social assistance according to the new social assistance structure; Implementthe reforms of individual benefits; Develop a systemof monitoring and evaluation of social assistancebenefits with changes inthe HBSsandthe establishment ofamonitoringandevaluation group; Design active labor market programs for the poor linked to social assistance benefits and prepare a cost benefit analysis of such proposal. Long-termobjectives Progressive redistributionof the existingbenefits infavour o f the very poor; Progressive transition of social assistancebeneficiaries to social insurance schemes. 95