Report No. 48604-LT Lithuania Social Sectors Public Expenditure Review Prepared for the Republic of Lithuania by the World Bank Group June 2009 Human Development Department Europe and Central Asia Region Document of the World Bank ACKNOWLEDGEMENTS This report was prepared by a World Bank team including Cristobal Ridao-Can0 (team leader), Agnes Couffinhal (Health), Asta Zviniene (Pensions), Oleksiy Ivaschenko (Social Protection), Michael Mertaugh (Education), Svetlana I.Proskurovsk (Public Administration), Valerie 0. Herzberg (Macroeconomics) under the overall guidance of Gordon Betcherman, Tamar M. Atinc, Thomas Laursen and Orsalia Kalantzopoulos. In addition, Nina Todorova Budina (IMF) prepared the section on fiscal sustainability and risks for this report under the guidance of Catriona Purfield (IMF Mission Chief for Lithuania). The World Bank would like to thank Nina, Catriona and the IMF for this fruitful collaboration and understands the findings and recommendations do not necessarily represent the view of the IMF. Charles Griffin, Emil Tesliuc and Kaspar Richter provided very useful comments. The team would also like to thank officials at the Ministry of Finance, Ministry of Interior, Civil Service Department, Ministry of Education, Lithuanian State Science and Studies Foundation, Education Department of Vilnius Municipality, Ministry of Health, Ministry of Social Security and Labor, Lithuania labor exchange, SODFU, State Patient Fund, State Health care Accreditation Agency, Bank of Lithuania, Prime Minister Office, President Office, Government Chancellery, Customs Department, Statistical Office for their hospitality and for very useful discussions and background information for this study. The report greatly benefitedfrom the commentsprovidedby the Government on an earlier draft. ... 111 TABLE OF CONTENTS EXECUTIVE SUMMARY 1 1 15 2.. .................................................................................................. MACROECONOMIC OUTLOOK AND RISKS .................................................... ................................................... 20 20 The Economic Crisis andthe ShortfallinTax Revenue ........................................... The Role of Fiscal Policy inManaging Volatility..................................................... OVERALL FISCAL SCENARIOSAND RISKS Fiscal Sustainability and Risks to the Baseline Scenario .......................................... 22 Reducingthe Fiscal Deficit: PossibleRevenueMeasures......................................... 23 26 3. ......30 Assessment of Public Administration........................................................................ CIVIL SERVICE, PUBLIC ADMINISTRATION AND THE WAGE BILL 30 Proposedaction for improvingthe efficiency ofpublic administration.................... 34 4. SOCIAL PROTECTION ........................................................................................... 38 Overall social protection expenditures. composition and benefits .......................................................................................................... ............................ 38 42 Overall SODRA's finances ....................................................................................... Social insurance 42 44 Disabilitypensions andcompensations for nursing and care.................................... SODRA's Old-Age Pensions..................................................................................... 49 51 Maternity leave benefits ............................................................................................ 53 Pensionsaccumulation funds..................................................................................... State pensions............................................................................................................ 52 Sickness benefits ........................................................................................................ 55 Unemployment benefits............................................................................................. 56 .......................................................................................................... Improvingthe financing and monitoring of SA benefits........................................... Increasing social challenges faced by the SA system as a result of the crisis ...........57 Social assistance 57 60 61 Child Benefits............................................................................................................ Social Benefit Program (SBP) ................................................................................... 63 65 5. Active Labor MarketPrograms(ALMP) .................................................................. HEALTH ..................................................................................................................... 66 The hospital System.................................................................................................. Overview ofthe health systemand its performance................................................. -66 -70 73 6 . Primary Care and HumanResources......................................................................... EDUCATION .............................................................................................................. Higher Education....................................................................................................... 83 Primary and Secondary Education ............................................................................ 76 76 Scienceand Technology............................................................................................ 84 iv LIST OFTABLES Table 2: Non-performing loans ofthe banking system..................................................... Table 1:Macroeconomic indicators .................................................................................. 15 16 Table 3: Possible composition of fiscal adjustment .......................................................... 27 29 Table 6: The composition ofthe pubic wage bill and changes ......................................... Table 5: VAT Revenue Productivity inEuropean Countries............................................ Table 4: General GovernmentOperations in2007............................................................ 28 Table 7: Social insurancebenefits increaseddramatically in2008................................... 33 Table 8: The bulk of SP benefits go to social insurance, little left for M T programs.......39 39 Table 10: Disability Benefit Ciaims, per 1000by Age Group, 1999................................ Table 9: Coverage and benefit generosity of SP benefits.................................................. 41 Table 11: State spendingon pensions, mlnLt,2008-2009............................................... 50 52 Table 12: Estimatedfiscal savings from options 1and2 .................................................. Table 13: Estimatedimpact of the crisis on household welfare ........................................ 55 57 Table 14: Targeting performance andcost-effectiveness of SP programs........................ Table 16: Estimatedadditional SBP costs in2009 under alternative coverage rates........62 Table 15: Size, coverage andtargeting performance of SBP over time ............................ 59 63 Table 18: CT Scan and MRIsper millionpopulation ....................................................... Table 17: Healthstatus inLithuaniaand European countries........................................... 69 Table 19: Average class sizes and studendteacherratios.................................................. 72 81 V LIST OF FIGURES Figure 1:Quarterly change inreal GDP (seasonally adjusted) ..................................................... 15 Figure2: Bank loansto the non-financial sector........................................................................... Figure 3: Capital inflows into the banking sector.......................................................................... 16 17 Figure 5: Financial leverage o f enterprises inthe Baltic states ..................................................... Figure 4: Profitability and financial leverage inthe corporate sector .2007 and 2008.................16 17 Figure 6: House price expectations for the next twelve months.................................................... Figure 7: Household debt over GDP ............................................................................................. 18 18 19 Figure9: Real effective exchangerates......................................................................................... Figure 8: Unit labor costs inHongKong -1998-2003 ................................................................... Figure 10:GDP growth and capital inflows.................................................................................. 19 Figure 11: Public Spendingduring the Absorption Boom: Composition and Trends...................20 21 Figure 12: Trends inPublic Sector RevenueGrowth and GeneralGovernment Deficit (%GDP) . ....................................................................................................................................................... 21 Figure 13: Output Gap, Structural Deficitand Inflation ................................................................ Figure 14: Baseline fiscal strategy and impact on fiscal sustainability ......................................... 22 Figure 16: Compositionof the public wage bill in2009 ............................................................... Figure 15: Wages of public sector employees and comparable private sector employees............25 32 Figure 17: SP expenditures are low comparedto most EUcountries ........................................... 32 38 42 Figure 19: Cumulative growth ratesof SODRA's revenues and expenditures ............................. Figure 18: Evolution of SODRA's membership........................................................................... 43 Figure21:Accrual rates in selectedOECD countries................................................................... Figure 20: Revenueand Expenditure of SODRA.......................................................................... 43 Figure 22: Incentives to contribute ................................................................................................ 44 45 45 Figure24: Evolution ofparametersdeterminingpension calculation........................................... Figure 23: Cumulative growth of pensions .................................................................................. 46 Figure 25: Alternative structure o f pension components............................................................... 47 47 Figure27: Distribution of elderly populationby benefittype and activity status......................... Figure 26: Age of Retirement inselectedOECD countries........................................................... 48 Figure28: Projected SODRA's Pension Deficits.......................................................................... 49 Figure 30: Family Policy Expensesas a % of GDP andtotal fertility rates.................................. Figure 29: Number of compensations for care by age ofthe beneficiary...................................... 50 54 Figure31: Healthexpenditure inLithuaniaand comparable European.countries (2007) .............66 Figure 32: Public and private health spending in Lithuania (2004-2007), current prices .............67 Figure33: Increaseinhealth workers salaries compared with other variables (2004-08) ............67 Figure 34: Trends inLife Expectancy ........................................................................................... Figure 36: Admission rates inLithuania and other European countries (2007) ........................... Figure 35: Hospital infrastructure in Lithuaniaand other European countries (2007) ..................69 71 71 Figure37: Number of physicians per capita inLithuaniaand Europe (2007)............................... Figure 38: Share of pharmaceutical intotal health expenditure (circa 2005) ................................ 74 75 76 Figure40: Real Gross MonthlyEarnings by Sector, 2004 and 2008 ............................................ Figure 39: Public Expenditureon Education (% of GDP), 2006................................................... 77 Figure41: Student-teacherratios for primary and secondary education (2006) Figure 42: % Change inenrollments and teachers, primary education inNMS, 1999-2006 ........79 ........................... 79 Figure43: % Change inenrollments and teachers, general secondaryeducation, NMS, 1999-2006 ....................................................................................................................................................... Figure 44: Projected school-age population by age group (thousands) ......................................... 79 Figure 46: Summary Innovation Index (SII) and per-capita Public Spendingon R&D................82 Figure45: Savings from reducingthe numberofteachers (relative to 2008) (million LT) ..........80 85 vi Executive Summary After a decade of strong growth and remarkable economic and social progress,Lithuaniais now facingitsworst economic crisis since independence.GDP contracted by close to 10% quarter on quarter in 2009 (quarter 1) and it is projected to contract by 11% in2009.' A key transmission mechanism o fthe global crisis to Lithuania was the large dependency on foreign capital to finance growth, as the combination of limitedaccess to foreign capital andbad credit caused domestic credit to plummet. A number of macroeconomic risks could exacerbate the downturn even further: (i)the required increase in savings to deleverage balance sheets will depress consumption and investment; (ii) rising credit defaults, including those in neighboring countries since the banking sector i s mostly foreign-owned, will reduce further credit availability; (iii)and the economy-wide adjustment through the internal price mechanisms could leadto a negative spiral o f lower prices leading to lower spending and this to lower prices andso on. Despite planned counter-cyclical measures, the actual fiscal policy implementedin the period leadingup to the crisis amplified the impact of capital flows on the economy and raised public spending to unsustainable levels. Public spendingrose by 50% between end-2006 and 2008, following generous public wage and social benefit increases, which grew in real terms by nearly 28% and 39%, respectively. The result was a four-fold increase inthe structural deficit duringthe boomperiod-from 1.6% o fpotential output in2005 to a record higho f 7% o f potential output in2008. After the crisis hit at the end of 2008, the new Governmenttook rapidaction and introduced a large upfront fiscal adjustment for 2009 in an attempt to regain short term fiscal sustainability, and in light o f the emerging financing constraints and the commitment to support the currency board regime and to join the eurozone. A large revenue shortfall, combined with the full-year impact o f last year's spending increases, resulted in a general government deficit o f 4% o f full-year GDP by end-April. Since December 2008, the Government has introduced fiscal adjustment measures for about 7% o f GDP and plans to implement further measures inJune. This report shows that reduced tax compliance can exacerbate the revenue shortfall and some o f this shortfall may be permanent. Even with the additionalfiscal adjustmentproposedfor June it will be hard to meet the Maastricht deficit criterion in the medium term, and too much adjustment can exacerbate the crisis. If the economic recession continues, the fiscal deficit could reach 6.5% and 7% of GDP in 2009 and 2010, respectively.2 The fiscal deficit i s projected to improve in 2011 as the economy recovers. However, this report shows that while there i s room for further cuts in spending over the short run, 1EuropeanCommission-Spring Forecast2009. This is based on the hypothetical scenario of GDP contracting by 14.5% and 4% in 2009 and 2010, respectively, which is around the predictions of the Bank of Lithuania. These estimates also take into the adjustment measures taken so far and the one proposedfor June 2009, totaling 8.1% of GDP. I permanently lower revenues and lower potential growth imply that without further structural reforms, the fiscal deficit will remain at about 4% o f GDP in 2013 and 2014. Lithuania is faced with the unenviable task o f having to improve the structural fiscal position while avoiding contributing to a downward economic spiral. Additional efficiency-enhancingstructural reforms will be needed to regain fiscal sustainabilityandto ensure compliancewith the Maastrichtcriteria. Giventhe rapid growth insocial spending and the public sector wage bill (the two largest categories o f public spending) in recent years, and the need to align public spending with permanently lower revenue, lowering the fiscal deficit over the medium run will require both short term cuts in spending and medium term structural expenditure reforms. And while the fiscal adjustment over the short- and medium term should be mostly expenditure-based, there may also be room for some revenue measures. In this context, this report aims to provide technical assistance to the Government of Lithuaniain preparationfor the June 2009 budget amendment and beyond.This Public ExpenditureReview (PER) includes a number of recommendations on short run fiscal consolidation (mainly through spending cuts) as well as efficiency- enhancing reforms that could eventually generate fiscal savings and reduce the fiscal deficit over themmediumterm. In the context o f increased social needs arising from the crisis, this report also tries to identify the changes needed in social assistance benefits to better protect the poor and vulnerable during the crisis and beyond in a cost-efficient way. In agreement with the Government, social sectors (social protection, health and education) have been chosen as the focus o f analysis, given the room for short and medium term adjustments and the large implications o f those on the overall budget. The PER also analyses the public wage bill and public administration. Finally, macroeconomic risks as well as the overall fiscal scenario and risks (including an exploration o f revenue measures that could bring fiscal savings over 2009-10) are also analyzed to provide the context for the subsequent sectoral sections. Relativesize ofsocial sectors and the Dublicwage bill in2007 %ofTS %ofGDP Public wage bill 28.5 10 Social sectors Social protection Health Education 14.9 Source: StatisticalOffice. TS refers to total public expenditures. Social sector spendingincludeswages. The rest of the executive summary elaborates on the proposed short term and medium term fiscal consolidation policies as well as the proposed changes in social assistance to better protect the poor and vulnerable duringthe crisis. The table at the end o f the executive summary contains a summary o f these proposed measures. On social protection, the Government may consider reforms to ensure the long run sustainability o f contributory pensions by gradually raising the retirement age; tying the base and earning- related pension components to objective parameters and subjecting them to rules-based 2 indexation; and credibly restoring the full flow o f contributions to the pension accumulation funds by the end o f 2010. The Government could also consider reducing the maternity benefitreceipt period, currently the most generous inthe EUand the single most important relative contributor to SODRA's deficit, to 1 year. And further targeting o f child benefits for children o f ages 0-3 and lowering the eligibility threshold could generate significant fiscal savings. The introduction o f a unique payment system with a single beneficiary registry could significantly reduce transaction costs, allow for a more timely response to the increased demand for benefits and strengthen the control over benefits. To improve the protection o f the poor during the crisis, the Government could consider meeting the additional demand for the Social Benefit Program (SBP), by increasing resources allocated to the program to reach up to 30% o f the poor. This measure could be complemented by a temporary extension o f the unemployment benefit receipt by 3 months and shifting the focus o f active labor market programs (ALMPs) towards direct job creation programs. To prevent the escalation o f public sector wages observed in recent times, the Government could consider increasing central monitoring and authority over public pay and employment, including the introduction o f a treasury payroll module. Lithuania has a good public administration with adequate institutions and procedures in place, but there are significant efficiency gains from streamlining the structure o f public administration institutions, including the consolidation o f government institutions under line ministers. On health, progress achieved since the transition i s remarkable, but there is considerable room for improving the efficiency and performance o f the health system, which mostly requires medium term reforms. Downsizing hospital infrastructure and adapting it to the needs o f the population would generate significant savings and contribute to improving patients' safety. The success o f this reform requires strong political consensus among central and local authorities. The current economic crisis provides a unique opportunity to reach this consensus and to start pursuing the hospital restructuring agenda. A key component o f the strategy to reduce the use o f hospital services it to have primary care providers provide more services than they currently do. Accreditation and quality assurance mechanisms for health facilities need to be strengthened. On education, more efficient utilization o f teachers could yield major budget savings. Amending the per-student financing formula for general education so that declining student populations are matched by proportional reductions in the number o f teachers, and gradually increasing class size to regional norms could yield budget savings o f about 8 billion LT by 2020. Greater reliance on fee-financing o f higher education and targeted budget support could generate significant savings and provide the necessary funding for improving the quality o f higher education while ensuring equitable access. And financing science and technology based on competitive funding would generate savings from the rationalizationo f research institutes and lead to higher research quality. 3 The public sector wage billand the public administration The public wage billhas been growing rapidly in recent times, particularly in 2008, following sharp increases in public wages. Between 2004 and 2008, the public wage bill grew by 50% in real terms (19% in 2008), and accounted for 29% o f the increase in public expenditures in 2008. Most o f this increase is due to rising wages (60% between 2004 and2008). All public sector wages grew at similar rates but wages o f doctors more than doubled inreal terms, and all public sector wages grew faster in 2008 thancomparable private sector wages. In2008 public sector wages were on average 15% lower thancomparable private sector wages. Monthlywages of public sector employeesand comparable privatesector employees 400 I1 * 8 350 -Public sector 300 01 - Public administration and 250 z 3u -0 .-8 -defence School teachers $ $) 200 150 - University education N vi h Doctors 'I m -.s 0 0 -Non N 0 -employees manual private ~ 2004 2005 2006 2007 2008 Source: Labor Force Surveys 2004-2008 The escalation of the public wage bill is partly driven by political pressures and prompted by the lack of a reliable system for managing and monitoring public sector pay policies. While improving public pay may bejustified in some cases, across- the-board increases (as opposed to targeted increases tied to performance) are not a cost- effective way to attract, retain and motivate public employees. The escalation o f public sector wages inrecent times was also drivenby political pressures. While the Ministry o f labor and Social Security (MSSL) is responsible for ensuring balanced public sector pay policies, political pressures result in some public sector groups being favored over others. More so, the decentralized management of the public sector payroll without central monitoring creates incentives for expandingpersonnel expenditure Civil servants have so far borne most of the burden o f adjustment, but if the Government wishes to undertake any significant reduction in the wage bill the adjustment would have to involve other public employees and explore alternative sources of savings. Further cuts in civil service pay risk undermining the attractiveness and stability o f the public administration. At the same time, the pay o f other public employees increased even in 2009. Given that civil servants account for a small share o f 4 the total wage bill, any significant reduction inthe wage bill would need to involve other public employees. The Government could also consider other sources o f savings: structural consolidation, reduction o f scope o f non-essential functions and tasks, temporary freezes on hiring, temporary withholding o f pay increases related to seniority (lengthof service), discontinuing paymentofbonuses andadditional pay for expansion of the scope ofwork due to substitutionoftemporary absent employees. Lithuania has a good public administration with adequate institutions and procedures in place, but there is substantial for improvingefficiency and quality. During the previous decade the Government focused its effort on designing rules and building new systems, which has produced reasonably well-performing institutions. The emphasis should now shift toward improving the efficiency and quality o f the public administration. Fiscal tightening presents an opportunity to identify inefficiencies, reduce unnecessary expenditure, and streamline procedures. To this end, the World Bank recommends the following actions: a. Improving the performance focus of the public sector by linking budgeting to priority policy results; strengthening systems for monitoring, evaluating and reporting o f results; and increasingly involving citizens' feedback on service quality. b. Increasing central monitoring and authority over public pay and employment policies by articulating and strengthening oversight roles of the Government Secretariat on policy outcomes, the Civil Service Department on staffing, the Ministry o f Finance (Treasury) on monitoring payroll management, and the Ministry o f Labor and Social Protection on overall labor remunerationpolicy. C. Streamlining the institutionalsystem of government bodies to align it with the policy framework by consolidating government institutions under line ministers to improve coherence and accountability for policy results, eliminating redundant structures, improving business processes and making more use o f Information and Communication Technologies. Social Protection Social protection (SP) expenditures increasedsubstantially in 2008 (34% in real terms), particularly maternity benefits. Social insurance benefits accounted for most o f this increase: pensions account for 57% o f the total increase (mostly due to the increase inthe base pension) while maternity leave benefits increased by 127%. The bulk o f social protection benefits go to social insurance (SI), particularly pensions and maternity. Social assistance (SA) benefits account for 14% o f total SP expenditures, and means-tested (MT) benefits for the poor and vulnerable only account for 14% o f SA. The continuous deteriorationof SODRA finances over the short and medium terms require significant adjustments in benefits. SODRA finances started deteriorating quickly in the last quarter o f 2008 due to the contraction o f the economy, and the significant increases in social insurance benefits. The deficit at the end of 2008 used up all the reserves accumulated inprevious years. Without the cushion o f reserves, SODRA's finances are deteriorating even further in 2009. Going forward, the number o f 5 beneficiaries i s expected to increase in the next few years while the number o f contributors is expected to decline. So far SODRA has been able to cope but it will not be able to hold for much longer given the rapid deterioration o f the economy and the structural nature o f the deficit. Social security contribution rates in Lithuania are on the high side compared to other EU countries, so the bulk of the adjustment will have to come from cuts and reforms insocial insurance benefits. Social Insurance Contributorypensions The pension system provides reasonably adequate benefits but retirement ages are low, the benefit formula is subject to a lot of discretion and generates disincentivesto contribute, makingthe system unsustainableover the longrun.The share o f the basic pension has been increasing, generating redistribution within the system and decreasing incentives to contribute. Lithuania also has an extensive system o f state pension supplements that would benefit from better monitoring, integration and rationalization. It has also instituted a system of pension accumulation funds that could provide a substantial portion o f retirement benefits inthe future. However, this system i s being threatened by reduced contributions to these funds to help pay for SODRA's pensions and the lack o f a credible strategy to ensure that this reduction will be allowed to expire. The Government could consider the following actions to address these concerns: a. Gradual increase in retirement ages for both genders up to age 65. Retirement age could increase by 2 months per year for men and 4 months per year for women, which would generate savings o f 65mln Lt in 2010 and 13Omln Lt in 2011 alone in additionto contributing to higher future pensions. b. Base the earnings-relatedpensioncomponent on the statistical average insured wage (rather than the discretionary insured income parameter) and subject it to a fiscally sustainable rule-based indexation. These changes could be introduced in 2010 and would generate savings in the long run, contribute to sustainable pension growth and improve transparency, incentives to contribute and intergenerational fairness. C. Make the base pension (currently at 360Lt) equal to the value of the basket of basic goods (currently 350Lt) that is calculated every year by the Department of Statistics. This would generate savings o f 120mln Lt in2010 and additional 40mln Lt per year for each 1% o f real wage growth compared to the 2009 level. As a medium term reform, the Government could consider transforming the base pension and the social pension into a single demogrant financed from the State budget. d. Restore the full flow of contributionsto the pension accumulation funds by the end of2010. The best strategy to make this credible is to commit to structural pension reforms to make them sustainable over the longrun. e. Rationalizethe system of state pensionsupplements by shifting the administration o fthese benefits from line ministries to SODRA. 6 Maternity leave benefits Maternityleavebenefitsmorethan doubledin2008, as the periodfor receipt of maternity/paternitybenefitwas extendedto 2 years, makingit the most generous among EU countries. Maternity benefits experienced the largest proportional increase among all social protection programs in 2008. The benefit receipt period i s the most generous among EU countries and effective replacement rates are higher than stipulated by law, as beneficiaries artificially inflate earning before leaving work. The cost o f the current program interms o f fiscal sustainability andwork disincentives far outweighs any potential benefit in terms o f fertility and child care. The Government could consider the following options: a. Option 1. Revert the benefit receipt period to what it was in 2007 (1 year after the birth of the child) at 100% replacement rate (higher than the 2007 rate). This would still leave maternity benefits in Lithuania among the most generous inthe EU. b. Option 2. Inaddition to option 1, limit the total leave period o f the maternity benefit (currently 70 days before birth and 56 days after) and the maternity/paternity benefit (upto 1year after the birtho fthe childwith the proposed amendment) to 52 weeks (1 year) to be distributed betweenpre-birthand after birthby the beneficiary. Fiscal savings in 2009 Fiscal savings in2010 Option 1 266 millionL T 355 million LT Option 1+ Option 2 445 million LT 534 million L T Unemployment benefits (UB) As the Governmentdiscusses the new draft law on unemploymentinsurance, the Government should refrain from any cuts in unemployment benefits. This is because (i)by all measures, these benefits are not excessive (benefits are among the lowest in EU while the benefit receipt period i s standard among EU countries); and (ii) UBprovide akey social safety net to households to cope withthe crisis. The Government could consider a temporary increase in the benefit receipt period,as other EUcountries are doing. Extending, the unemployment benefit period by 3 months would cost an additional 141million LT in2009. SocialAssistance 19. The social challengesfaced by the SA system are increasingas a result of the crisis. Reduced incomes and wealth will have a significant impact on household welfare. It is estimated that the number of people from households with per capita consumption 7 below the poverty line (350 LT month) could increase by 208,000 (48.6% increase), bringingthe total numbero fpoor people to 636,000 (18.9% ofthe population) in2009.3 Means-tested SA programs are the main (potential) safety net for the poor and vulnerable to cope with the crisis. UB do not cover individuals who become unemployed but do not have sufficient insured employment history, nor the increasing numbers o f working poor. Active Labor Market Programs (ALMP), particularly direct job creation programs (public works and subsidized employment in the private sector) can also provide complementary assistance. But current SA benefits are too low and not adequately targeted to the poor, resulting in a small overall impact on poverty and vulnerability. About 25% o f the poor do not receive any kind o f social assistance, and the main benefit for the poor, the SBP, only covers 6.7% o f the poor. Most benefits go to non-poor households (64%), reflecting the lack o f means-testing o f most SA benefits. And the value o f SA benefits relative to the consumption level o f poor beneficiaries i s low (23%), reflecting the small expenditure on SA programs. Inadequate targeting o f the poor and low relative SA benefits result inlow overall impact on poverty and highefficiency cost. Improving theJinancing and monitoring of SA benefits The Government could consider bringing all SA programs under a central unique payment system with a single beneficiary registry that will pay all SA benefits directly to beneficiaries (instead o f payments being channeled through municipalities). The planning, administration o f benefits and monitoring o f beneficiaries would still remain under local governments. Such a system would not reduce accountability o f municipalities, as under the current system municipalities have very little control and discretion over SA funds, and would bringthe following benefits: Reduced transaction costs inthe financing o f benefits. 0 Muchfaster response to the increased demand for SA benefits duringthe crisis. More control over SA benefits through a single beneficiary registry. An alternative to the creation o f a unique payment system for SA benefits only would be to use SODRA, which would reduce costs and improve control over all S I and SA benefits andbeneficiaries, compensating it for the increased administrative costs. Social Benefit Program (SBP) The SBP is the benefit of last resort for the poor-good design and implementation features result in good targeting performance and adequate benefits. About 66% o f the beneficiaries are below the poverty line (the best targeting performance o f any SP program) and benefits account for 29% o f the consumption o f the poor. Povertysimulations are basedonthe assumptiono f a 10%across-the-boardreduction inlabor income, 8 But the program has very low coverage among the poor (6.7%), resulting in a low overall impact on poverty. This low coverage may be partly explained by the introduction in2004 o f the value o f household assets as an additional eligibility criterion. The introduction o f the more generous social pension in 2006 also made some households with poor elderly ineligible. More likely, however, low coverage i s due to low demand, but this i s likely to change with the crisis, as more families fall below the poverty line and poor households face more limited income-generating opportunities. The Government could consider the following actions to better protect the poor and vulnerable duringthe crisis: a. Raising the benefit to 100%o f the difference between the SSI and household income per capita. b. Improving or relaxing (at least temporarily during the crisis) the household assets eligibility criteria, given the little value added in terms o f targeting performance relative to its cost interms o f reduced coverage o f the poor. c. Speedingup the application and verification process or even relaxing the verification process temporarily (beneficiaries would still be subject to reassessment). d. Meeting the additional demand for SBP by expanding program resources to cover up to30% of the poor. Under a desirable and realistic coverage rate assumption (30%), the additional SBP spending in 2009 would be about 57 million LT. Child benefits (CB) The CB is the largest SA program for families and children. I t has the largest coverage of the poor (70%) but most of the benefits go to the non poor (80%). This reflects the universal nature o f this benefit, which also results in very high efficiency costs. The crisis management plan o f December 2008 limited child benefits for children aged 3 and older to families with per capita family income less than3 times the SSI. With no change inthe program, the increase in spending resulting from the increased number o f families falling below 3*SSI in2009 would be 19.4 million LT for 2009.4 While the CB does not have a poverty-reduction objective, granting benefits to wealthy families for whom these benefits do not affect their decisions to have children or to invest in their well-being results in large efficiency costs. Evenseeing this benefit as a right for all children, it is hard to justify the financing o f such a right for all children ina context of limitedfiscal space and increasing social needs. The Government could consider the following actions: a. Limiting the benefits for children under 3 to families with per capita family income i s less than 3 times SSI. Assuming all families that qualify apply for the benefit and that the change would be implemented from July 1, savings between July and December o f 2009 would be about 22 million LT (42 million on an annualized basis). Not all qualifying families will apply for the benefit, so this is an upperbound estimate. 9 b. Lowering the income eligibility threshold to 2 times the SSI. Assuming this applies to all children and the change i s implemented from July 1, savings between July and December o f 2009 would be around 89 million LT (178 million on an annualized basis). c. Subject applicants to the same questions about income and information verification process as for SBP applicants to ensure effective targetingo f benefits. Health Progress achieved since the transition is remarkable, but there is considerable room for improving the efficiency and performance of the health system, which mostly requires medium term reforms. By international standards, the amount Lithuania spends on health appears reasonable. But health outcomes, such as life expectancy and mortality from heart diseases and external causes (e.g. alcohol and tobacco), are poor compared with countries that spend similar amounts, and have not been improving in recent times despite public health expenditures increasing by 85% between 2004 and 2008. This suggests that there i s considerable room for improving efficiency in the health system, which still relies too heavily on hospital-based care instead of more cost-effective prevention and primary care. Moreover, Lithuanians are generally dissatisfied with the health system which they consider to be prone to corruption. Salaries of health workers, particularly doctors, have increased very rapidly and are significantly higher than the salaries of other public employees and comparableprivate employees-but any adjustmentwould be hard to implement. A 10% reduction o f physicians' payroll would save about 60m LTL on a full year basis, but this might be hard to implement because of the complex transmission mechanisms of salary cuts: the State Patient Fund(SPF) purchases services and decisions on salary levels are made at the facility level. Reducing the price o f health services (which would reduce outlays) may not automatically translate into reduction o f salaries, and other critical inputs to the production o f health services may suffer inthe process. Since public health expenditures increased much less than salaries in the past few years, the budget share available for this input has already been decreasing. This could result inpeople having to purchase drugs and medical consumables that should be available free o f charge at the facility. If such an option were pursued, since outpatient care needs to be further developed, efforts should be made to protect services whose production needs to be encouraged. In any case, future salary increases should be tied to the performance o f healthworkers. Restructuring the hospital infrastructure. Despite significant progress since independence, the number o f hospital beds and hospitals relative to the size o f the population and the inpatient admission rate are among the highest in the EU. And the number of births per year in each obstetric bed is too low to ensure adequate safety for patients. Downsizing hospital infrastructure and adapting it to the needs o f the population would bring about savings and contribute to improving patients' safety. This requires a hospital master planwhich i s based on an analysis o fthe population needs and reasonable 10 access criteria to emergency and acute specialized services. The plan would establish a clear hierarchy o f the levels o f care (secondary and tertiary) each hospital i s expected and allowed to provide and define referral protocols and systems. Departments and facilities which do not provide enough services and for which accessible alternatives exist would be closed. In addition, equipments could be allocated (and if necessary redistributed) to ensure hospitals can fulfill their redefined mission. The success o f this reform requires strong political consensus among central and local authorities. The current economic crisis provides a unique opportunity to reach this consensus and to start pursuing the hospital restructuring agenda. Strengthening the accountability of health facilities. Quality management and control are weak, and health facilities are not effectively accountable for the quality o f the services they provide. Hospital are funded on the basis o f a list o f services and pre-set prices, but those do not necessarily reflect costs. Accreditation and quality assurance mechanisms could be put in place and health technology assessed. Information about performance, including but not limited to quality, should be collected, and broadly disseminated. The Government may also consider allowing payments and contracting and tying those to performance. Inparticular, the hospitalpayment system could be revisedto induce hospitals to use inputs more efficiently (including using fewer doctors) and to improve performance. And the SPF could also be allowed to selectively contract providers, and/or to not purchase all the services an institution i s licensed to produce. Accelerating efforts to strengthen primary care. Currently, only about 25% o f physicians work in primary care settings. A key component o f the strategy to reduce the use o f hospital services it to have primary care providers provide more services than they currently do. Primary care providers tend to refer patients to higher level o f care more than they should. One possible policy is to link the pay o f primary care providers to performance interms o f coverage o fkey preventive services. Developing a genuinely selective new health policy, with a strong M&E framework, and clearly linked to health priorities (prevention o f non-communicable diseases). It could include a solid human resources component to better prepare the systemto cater to the needs o f the ageing population. The process o f elaborating this new policy could be supported by broad consultations. Education There may be room to reduce teacher salaries but this will face strong political resistance. Average teacher salaries at the end o f 2008 were still below those o f other public-sector employees (particularly doctors) and comparable private sector employees, although this difference was reduced in 2009 by the 10% increase inteacher salaries and the reductions in civil service pay. Since teacher salaries account for the largest share o f the public wage bill (34%), any significant reduction inthe public-sector wage bill would need to include teacher salaries. Some approaches to reduced salary outlays for teachers could be less contentious than an outright reduction in base salary. Latvia, for example, has reduced salary expenditures for teachers by eliminating 11 supplementary payments for some categories o f non-teaching activities, and proposes further expenditure reductions through an increase from 21 hours to 27 hours in the number o f student contact hours that comprise a full teaching load. Reductions already made in the 2009 education budget pose risks for education quality. In many municipalities, this reduction could be absorbed without adverse educational consequencesby ajudicious reduction inthe number o f teachers or a reduction in paid overtime teaching hours. However, indications from Vilnius municipality suggest that local governments will opt instead for reducing the already- small amounts provided for non-salary education inputs (e.g. textbooks, teaching materials, in-service teacher training, and internet connectivity), which have already been reduced from 6% in 2007 to 3% in 2009 to help make room for the teacher salary increases. To ensure adequate financing o f these crucial inputs for education quality, recommended that the balance o f salary and non-salary elements be restored to 6% o f the student basket, which could be financed by an offsetting 3% reduction in the salary portion o fthe student basket. More efficient utilization of teachers could yield major budget savings. Studenuteacher ratios are lower than in most o f the New Member States (NMS) and much lower than inmost OECD countries. And the number o f teachers has adjusted little in response to the declining general education enrollments in Lithuania. Declining school-age populations will offer the potential for additional budget savings in the medium term. The preferred approach for reducing education expenditures in general education is to reduce teacher numbers by changing the per-student funding formula. There are currently too many teachers and too many small classes in general education. Several changes are proposedinthe per-student financing formula. base the adjustment for size-related cost differences on non-discretionary factors that affect class sizes and unit costs, rather than discretionary differences such as class size and school size; 0 gradually and progressively raise the class size and teaching load assumptions on which the formula i s based in order to motivate improved efficiency in teacher deployment, with the long-term goal o f raising studentlteacher ratios to OECD averages; and 0 allow municipalities to spend salary savings under the formula for non-salary activities to improve teaching and learning. These measures could yield budget savings of about 8 billion LT by 2020. Amending the per-student financing formula for general education so that declining student populations are matched by proportional reductions inthe number o fteachers (Le. keeping unit costs constant at 2008 levels) would generate savings of 44 million LT in 2009 (assuming the changes are introduced at the start o f school year 2009-10) and 275 million LT in 2010. Savings are predicted to increase each year until 2020 when they peak at 853 million LT. Gradually increasing class size by 1% each year would generate additional savings o f 600 million LT by 2020. Some o f these savings are expected to be invested innon-salary activities to improve teaching andlearning. 12 Central financing and management of secondary vocational schools generates significant inefficiencies and disarticulation with general secondary education. Vocational schools are not attracting as many students as gymnasia and general secondary schools, and have higher unit costs and lower staffing efficiency. In order to promote better efficiency and better labor-market outcomes o f vocational schools, it is recommended that management of general secondary schools and vocational secondary schools be consolidated, and that both types o f schools be managed by municipalities under the same process and subject to the same performance criteria. Municipalities should be free to convert ineffective vocational schools to other uses that better meet local needs. This system of budget financing in higher education consumes a large amount of resources. It covers about half of total higher education enrollments that could otherwise be used to improve higher education quality (expenditure per student i s among the lowest on Europe); it is inefficient (because it finances a number o f students who have the capacity to pay and the arbitrary manner in selecting specific fields and institutions for support); and inequitable (because tuition, scholarships and loans are exclusively allocated on the basis o fperformance, not needs). The Ministry of Education and Science has proposed a higher education reform that would address most of these deficiencies-it i s recommended that these proposed reforms be approved and implemented. Significant additional budget savings could be achieved through increased reliance on student fees, broader access to student loans, and targeted budget support. Public spending on science and technology (S&T) institutes is high, but resultsin terms of innovationare below those of most other countries,which is part due to the financing system. Four-fifths of the S&T resources (managed by MES) are used to support the core costs o f research institutes, which were inherited from the Soviet period. Only 20% o f budget financing is used competitively to support specific R&D products. This approach to financing science and technology development embodies several important deficiencies: (i) it is supply-driven; (ii) provides weak incentive for it improved research productivity and relevance; and (iii)it prevents universities from playing the dynamic role that they should play inscience and technology development. It is recommendedthat budget financing for S&T development be changed to a competitive, internationally peer-reviewed approach. This would improve the efficiency and accountability o f public support for S&T and would improve the quality and relevance o f research products. This approach could be used to support solicited contract research on topics o f agreed national priority and high-quality, unsolicited grant proposals. The same approach could support institutional contracts for specified research and technology products. Financing research products based on proposal quality should lead to a consolidation o f research capacity among research institutes and universities. i s also recommended to reduce the number o f research centers to be supported under "Sunrise Valleys" proiect from five to two, and use the remaining funds to support competitive research under this approach. 13 Restorationoffull flow ofcontributionsto the I2010 I I IImprovedcredibility ofthe secondpillar andfuture pension Limit total maternityleave periodto 1 year July2009 445 534 Lower the eligibility threshold of child benefitto 2 July 2009 89 Additional savingseachyear dependingonhow many times the SSI and includechildrenunder 3 families fall below2*SSI Introduce acentraluniaueuavmentsvstem with a 2010 Reducetransaction costs: muchfaster resuonseto demand: singlebeneficiaryregistryfor all SA benefits I I , Iand more controlover SA beneficiariesandbenefits Note: Numbers inbracketsrepresentadditionalspending.Numberswithout bracketsrepresentfiscal savings 14 1. Macroeconomic outlook and risks 1. After a period of strong growth following EU accession, Lithuania is now facing its worst economic crisis since independence. GDP contracted by close to 10% quarter on quarter in 2009 (quarter 1) and projections for the remainder of the year and 2010 indicate ongoing weakness (Figure 1 and Table 1). Data for Q4 and partial high frequency data indicate that the decline has been mainly explained by falling domestic investment andexports, with consumption growth also weakening. ,Figure 1: Quarterly change in real GDP Table 1: Macroeconomic indicators (seasonally adjusted) Annual percentagechange (unless indicatedotherwise) 2005 2006 2007 2008 m 2010 4 1 Percentagechangeon previousquarter RealGDP 7 8 7 8 8 9 3 1 1 41 I Employment 2 5 1.8 2.8 45 -71 -24 Unitlabourcosts 6 101 103 106 .7 66 Currentaccounfbalance%oiGDP -71 1 0 4 -151 1 2 2 -19 0 7 Source: European Commission-Spring Forecast -12 J 2004 9 1 2004 P4 2005 Q3 2006 Q2 2007 Q1 2007 Q4 2008 Q3 2009 Source: Statistical Office 2. The severity of the contraction can be explained by a sharp fall in external and domestic demand andtightening credit conditions: Financial markets. Following EU accession, growth was fueled by massive capital inflows, which peaked at almost 23% of GDP in mid 2007.5 These inflows fueled domestic credit and created excess demand, which resulted in significant overheating and large current account deficits. With the sharp deterioration inthe global economy in 2008 (quarter 4), the collapse of capital inflows and the bursting of the property market bubble, banks started to tighten credit conditions, resulting innegative capital inflows into the banking sector inearly 2009 (Figure 2 and 3). Non-performing loans started rising inthe fourth quarter of 2008 (Table 2). 'thegross external debt amountedto 72% of GDP inthe third quarter of 2008, as bank creditto the private sector more thandouble inthe last 5 years 15 0 Exports. Exports have fallen by 28% in February 2009 year on year, partly on account o f severe weakness inEUandRussiandemand.6 Table 2: Non-performing loans of the banking Figure 2: Bank loans to the non-financial sector system (comparedto arespectiveloanportfolio, %) P.rs.ntaw shanw on a p a r moo Total Business Mortgage Consumer loans Loans Loans Loans 06/30/08 2.45 2.82 1.47 3.03 09/30/08 2.65 2.95 1.75 3.70 12/13/08 4.55 5.64 1.94 5.23 03/31/09 8.23 10.45 3.23 8.80 Note: NF'Ls are the sum o f overdue (non-impairedloans) andimpairedloans.This definition is broaderthanthe one 3 0 0 . traditionally employed (e.g. loans overdue 90 days or more only). 2 0 0 . Figure 3: Capital inflows into the banking sector EUR million 800 1 600 400 200 0 -200 I -600 Source: Bank ofLithuania 3. And the downturn could be more protracted than anticipated because the required increase in savings to deleverage balance sheets will depress consumption. Indebtedness, although still less important than inthe neighboring Baltic economies, has increased considerably in recent years and may need to be unwound. With little prospect o f rising incomes or higher inflation, increased savings will be necessary for balance sheets to deleverage. This could depress consumption and investment for some time to Exportsto Russiamade up 16%ofLithuanian exports in2008-this share declined to 11% inearly 2009. Exportsto the EUaccount for 70% oftotal Lithuanian exports. 16 come. In some parts of the corporate sector (e.g. wholesale, retail, fishing, hotel and restaurants), profitability declined in 2008, while financial leverage-though still in aggregate well below levels observed inLatvia-was either high (over 150% of GDP) or rising (Figure 4 and 5). Household debt doubled relative to GDP between2006 and 2008 (Figure 6). About 10% o f indebted households in 2007 had a housing loan. Meanwhile, house prices declined by 25% year on year in March 2009, and household and banking surveys indicate that a majority of respondents expect further declines in the next 12 months to come (Figure 7). This suggests that more andmore householdsare likely to fall into negative equity, encouraging savings further and depressinggrowth. Figure 4: Profitability and financial leverage in1 Figure 5: Financial leverage of enterprises in the corporate sector- 2007 and2008 the Baltic states Pcengges Percentages UI 250 831 200 150 100 50 2002 - 2003 2004 2005 2006 2007 2008 Lbhuania I Latvia Sources Depadment of Statistics, Central Bureau of Statistics of Esbnia Latvia, Statistics Agancy of Estonia and Bank of Lithuania Note: Leverageis definedas companydebt to equity 4. The behavior of local banks will also be affected by rising credit defaults, including those in neighboring countries, which could also prolong the recession. The banking sector i s mostly foreign-owned. Exposures are concentrated intwo Swedish bankswith large operations inall three Baltic States and also inthe Ukraine. These banks will face credit risk pressures not just on their exposures in the Baltic States (roughly worth 19% of GDP), but also in their home country and elsewhere. The Swedish economy is expected to enter recession in 2009. The European Commission in its latest Spring Forecast (May 2009) expects unemployment in Swedento rise from 6.2% in2008 to 10.4% in,2010. 5. InApril 2009, the rating agency Fitchdowngraded the Swedish banks and their Baltic subsidiaries on account of the deteriorating credit outlook. According to its estimations, one bank inparticular would face considerable pressure on its capital base if the downturn turns out to be sharper than anti~ipated.~ such an event, the credit In squeeze could take longer, prolonging the downturn. Moreover, the Lithuanianeconomy See WesternEuropeanBanks' exposureto EasternEuropeandthe CIS -FitchRatings, April 2009 17 is directly exposed to spillovers o f liquidity risks faced by Swedish banks: everything else equal, given the constraints posed by the currency board, failure o f foreign banks to rollover local banks' large external debt (for example in a situation where they themselves faced a sudden increased demand for liquidity) would negatively affect FX reserves and hence domestic money supply, further shrinkingthe domestic economy. IS Banks, March2008 Source: Bank o fLithuania 6. And the economy-wide adjustment through the internal price mechanisms could generate negative feedback loops. Unlike the Czech Republic, whose currency has depreciated sharply recently-helping to lower the real effective exchange rate, in Lithuania-as in the other Baltic states4omestic wage and price inflation need to be lower compared with that o f its competitors (Figure 9). But this could lead to a deflationary spiral, where low or negative inflation discourages consumption and investment, placing further downward pressure on prices. The example o f Hong Kong, a small open economy with a currency board, illustrates these risks. It was hit by a range of severe internal and external shocks inthe late 1990s. Unit labor costs fell by 10-15% and remained at the lower levels for a prolonged period (Figure 8). Domestic consumption remained depressed for many years, as households struggled with negative equity and falling house prices.8 7. Finally, although constraints posed by the exchange rate regime and the desire to rebuild credibility require strict fiscal policy, overly restrictive policies could prolong the economic downturn. In April 2009, the government adopted a supplementary budget with expenditure reducing measures o f around 4% o f GDP. Still, ~~ ~~ ~~ ~~~~ Jao Y.C: TheAsian Financial Crisis andthe Ordeal of HongKong, QuorumBooks, 2001 18 the European Commission in its latest Spring Forecast expects the fiscal deficit to widen to over 5% of GDP in2009. The government now plans an additional adjustment inJune to bringthe deficit down to around 5% of GDP. While it will be important to strengthen Lithuania's structural fiscal position and ensure that expenditures are well targeted, fiscal policies must avoid contributing to a downward economic spiral. Figure 8: Unit labor costs in HongKong - Figure 9: Real effective exchange rates 1998-2003 11M y 1998.1W 170 thzrrl2.16 Unit labcureost CzechRepulc I6O 150 (0aQ1=1o0) I 110 Uhuanla i io4 80 90I rn Source: Datastream 19 2. Overall fiscal scenarios and risks' The Role of FiscalPolicy inManaging Volatility 8. Managing capital flow volatility has proven to be the main challenge facing the Lithuanian economy. Following the EU accession, rapid growth was fueled by a surge incapital inflows, which peaked at almost 23% o f GDP inmid-2007 and remained highin2008. These massive inflows fueled domestic credit and created excess demand, which resulted in overheating and macroeconomic imbalances. With the sharp deterioration in the global economy in the last quarter o f 2008, a near sudden stop in capital inflows and the burstingo f the property market bubble have brought an abrupt end to the foreign financed credit boom and caused a rapid adjustment in macroeconomic imbalances (Figure 10). Figure 10: GDP growth and capital inflows - 15 Lithuaniafollowed the Baltics.intorecession in QN ... _ _ 35 - Capital Inflows 10 30 (Four-quarter cumlative, percq\` of GDP) 30 25 . * 5 li 25 0 20 -5 - 15 -10 - - - Lithuania -Latvia 10 \ I -15 V-Lithuania-Btonia -- - - -Latvia1 -CE -20 I * I 1 -15 -10 I a 0 a a w a 8 a0 E Source: Statistical Office, Bank of Lithuania. The average o f CEE countries includes Czech Rep., Hungry and Poland. 9. Despite planned counter-cyclical measures, the actual fiscal policy has been pro-cyclical during the capital flow boom, leaving a legacy of unsustainablepublic spending (Figure 11). Spending rose by 50% between end-2006 and 2008 following generous public wage and social benefit increases, which grew in real terms by nearly 28% and 39%, respectively. Spending on pension benefit grew on average close to 30% per annum in2006-2008, reflecting ad hoc increases in `non-contributory' base pensions. Maternity benefit spendingmore than doubled between 2007-2008 to almost 1% o f GDP, which reflected the extension inmaternity leave to 2 years and an effective replacement rate in excess o f 100%. Spending on sickness benefits grew nearly 40% between 2007- 2008 reflecting strong wage growth (average benefits rose by 53% in between 2006 and 2008) and a 14% increase inclaimants due to lax certification. 9 This section was prepared by IMF staff Nina Todorova Budina, assisted by a larger IMF team, for the World Bank Public Expenditure Review. The findings and recommendationsdo not necessarily represent the views ofthe IMF. 20 Figure11:Public Spendingduring the Absorption Boom: Compositionand Trends. 36 224 Corrposition of Expenditure Trends in Social insurance Benefits, 20 04-09 6 00 5 00 -Unemployment 400 - 4 00 300 4 300 2 00 2 00 Wages I benifits Social 100 100 Q Goods and services IInvestment 2004 2 0 0 5 2006 2007 2008 2009 I Psubsidies (budget) Source: Statistical Office; Ministry o f Finance; and IMF staff estimates. 10. The fiscal expansion was masked by cyclically-high fiscal revenuesand large windfall receipts related to domestic absorption boom." For example, rapid growth in domestic demand has resulted in windfall VAT receipts o f about 1 percent o f GDP. Similarly, rapid wage growth, well inexcess o f the productivity growth resulted in a high share o f labor in GDP. This impliedwindfall receipts and high effective rates of PIT and social security contributions as well. Strong revenue growth and the presence o f windfall receipts kept the general government deficit low until 2007 (Figure 12). But when the economy slowed down in 2008, the headline fiscal deficit widened to 3.2% o f GDP, despite the buoyant fiscal revenue. The social security (SODRA) deficit, caused by the growth in social benefits in 2008, represented about one third o f the general government deficit in2008 and was financed by exhausting accumulated reserves. Figure 12: Trends in Public Sector RevenueGrowthand GeneralGovernmentDeficit(% GDP). 30 . 30 Revenues. Direct and Indirect Taxes General Government Fi~cal Deflcit (Annual percentage change) (Percent of GDP) 25 - - -Revenue, total 25 Indirect taxes 0 0 20 - --- - - -Direct taxes - 20 -1 -1 15 - - 15 10 - - 10 -2 -2 5 -3 -3 / " - 0 -5 Source: Statistical Office; Ministry o f Finance; and IMFstaff estimates. lo Forthe revenueestimatesduringthe absorptionboomsee Kangur (2009), "Impact ofthe Colapseofthe AbsorptionBoomon Revenue", mimeo, Washington DC. 21 11. Going forward, the fiscal management of capital flow volatility needs to be improved. The new fiscal rules framework introduced only became effective for the 2009 budgetplanning process. The previous budgetary framework lacked comprehensive fiscal rules and was not effective in controlling spending during the boom-expenditure grew rapidly, passively following rapid revenue growth. The result was a four-fold increase in the structural deficit during the boom period-from 1.6% o f potential output in 2005 to a record high of 7% of potential output in 2008 (Figure 13). The failure to implementplannedcountercyclical fiscal measures generated positive fiscal impulse and exacerbated domestic demand pressures during the boom period. Strengthening fiscal institutions to promote fiscal discipline and savings in good times can create a cushion against future downturns. Figure 13: Output Gap, StructuralDeficit and Inflation Cyclically-adjusted deficit, percent of potentialoutput Source: Statistical Office; Ministry of Finance; and IMF staff estimates. The Economic Crisis and the Shortfall in Tax Revenue 12. As the crisis hit, revenues associated with the credit boom dissipated and a large revenue shortfall emerged in 2009. Tax revenue collections of the central government were down by 27% year-on-year inJanuary-March 2009, far inexcess o f the nominal drop in the tax bases (nominal GDP and private consumption decreased by 15.6% and by 8.6% year-on-year, respectively). All types o f revenues have been affected, including social security contributions. The largest drop so far i s inVAT revenue, despite the increase in the standard VAT rate from 18% to 19% and the elimination of most exemptions in January 2009. Revenue shortfalls are common during downturns (Box 1). With the end o f the absorption boom, the large windfall receipts from VAT (nearly 1 percent o f GDP) will cease and will likely result inpermanently lower VAT tax receipts. Furthermore, as a result of the recent PIT tax cut and the downward adjustment inwages, PIT collections can be expected to decline by 0.6-0.8 percentage points in an effective PIT rate vis-a-vis GDP. See Kangur (2009), mimeo, Washington DC. 22 13. There is evidence of worsening tax compliance in a downturn, as firms and consumers facing cash flow problems delay tax payments. In particular, the VAT compliance had begun to deteriorate earlier in 2008 before the recession truly deepened, suggesting pre-existing problems in administration-an assessment also supported by rising stock o f arrears. The compliance ratio (the ratio of the net VAT proceeds over the maximum collectable amount at the headline rate), an indicator o f the efficiency o f collection was very high, especially towards the peak o f the boom - at about 70% compared to an EU average o f 63%- while VAT collection i s difficult in small transit countries. However, stocks o f unpaid VAT refunds o f 1% o f GDP in Lithuania boosted this ratio. These refunds are likely to be more quickly claimed by firms as the downturn deepens andthis should negatively impact headline VAT collections. Fiscal Sustainability and Risks to the Baseline Scenario 14. In2009, as the crisis hit with full force, sizeable expenditure commitments could not be met with the much lower fiscal revenue and a large upfront adjustment was introduced to regain short term fiscal sustainability. Without adjustment measures, the general government deficit would have reached 15% o f GDP. Since the beginningo f the year, the authorities implemented discretionary fiscal adjustment measures for about 6.5% o f GDP and planto implement further measures inJune this year. 23 Box 1. Economic Downturnsand Revenue Shortfalls: Lessons for Lithuania The global financial and economic crisis presents a major challenge for tax revenues in many countries. Three out o f four high profile crises o f the 1990s suffered revenue losses (See Table 1). Turkey, for example, suffered a large revenue loss (4 percent o f GDP) in the year following the 2001 financial crisis. In four out o f six countries with IMF programs in Europe and former Soviet Union, there have been sharp revenue shortfalls in 2008 and 2009 relative to original program revenue targets. Most o f these countries are also suffering severe economic downturns and there have been significant downward revisions to macroeconomic parameters. Furthermore, large structural changes in the economy (e.g. collapsing import and consumption and "disappearance" o f key economic sectors o f the economy) have led to a decline even disappearance o f tax bases duringrecession. DevelopmentsinGovernmentRevenueduringCrisis (3ses 1997 1998 1999 2000 Russia(Jul 98) 30.8 34.6 - 32.9 36.2 1998 1999 2000 2001 Brazil(Nov. 98) 28.1 30.5 28.9 30.2 1994 1995 1996 1997 Mexico @ec. 94) 11.3 9.3 8.9 9.8 2000 2001 2002 2003 Turkey(Feb0 1) 30.4 33.0 28.7 28.2 Revenue shortfalls are generally larger than what can be explained by deviations in macroeconomicparameters. A recent FAD board paper on fiscal policy in crisis presents econometric evidence that links tax collections to asset prices. Structural change (e.g., increase in demand for large VAT credits outstanding) or worsening compliance with the tax laws exacerbate the revenue loss from the crisis. To help mitigate the compliance decline and associated revenue loss, tax agencies are encouraged to develop a tax compliance strategy for the crisis by expanding assistance to taxpayers, refocusing enforcement on emerging compliance risks, enacting legislative reforms that facilitate administration, and improving communication programs. So far most of crisis countries have been forced to take expenditure measures but some implemented revenue measures as well. Evidence based on a sample o f 66 successful fiscal adjustments, defined as no or small reversal inthe fvst three years, show that on average, expenditure measures comprise 76 percent o f total fiscal adjustment.' However, some countries also implemented emergency stop gap measures (e.g. rate increases) in response to widening fiscal deficits. These could be used ifjustified by fiscal sustainability, tax system efficiency and fairness reasons (e.g. getting marginal tax rates right) or structural reasons (e.g. impact on labor market incentives could be improved, problems in the social security addressed). However, these revenue-increasing measures should use the opportunity to implement more fundamental structural reforms. Source: FAD, IMF 24 15. Even assuming additional fiscal adjustment, it will be hard to meet the Maastricht deficit criterion in the medium term. Under a hypothetical scenario in which GDP contracts by 14.5% and 4% in 2009 and 2010, respectively (which i s around the projections made by the Bank o f Lithuania), the fiscal deficit i s likely to reach 6.5% and 7% of GDP in 2009 and 2010, respectively.12 The fiscal deficit is projected to improve in 2011 as the economy recovers. (Figure 14). However, permanently lower revenues and lower potential growth imply that without further structural reforms, fiscal deficit will remain at about 4% o f GDP in 2013 and 2014, well above the Maastricht deficit threshold o f 3% o f GDP. 16. Such a fiscal strategy is also not risk-free in terms of fiscal sustainability. With fiscal deficits withinthe 6 %-7% of GDP range, public debt is projectedto increase to 40% in2011. Beyond that, public debt will continue to increase, but will still be below 50% in 2014. However, given the uncertainty surrounding key macro variables in projecting the debt dynamics, stochastic simulations show that the maximum likely debt limit in2012 is already well above the Maastrichtdebt-to-GDP criterion o f 60% of GDP. Beyond 2012, public debt would rise rapidly and there i s more than 10% chance that public debt will breach the Maastricht debt-to-GDP criterion o f 60% o f GDP by the end o fthe projection period (Figure 5). Figure 14: Baseline fiscalstrategy and impact on fiscal sustainability Fiscal Deficit(Percent of GDP) Distribution of Public sector debt-bGDP ratio 8 5 ~ - - - m95-975 n9095 c31090 c 3 2 5-5 06 2007 2008 2009 2010 2011 -Baseline 2008 2009 2010 2011 2012 2013 2014 Source: Statistical Office; Ministry o f Finance; and IMF staff estimates. 17. Additional efficiency-enhancingstructural reforms will be needed to regain fiscal sustainabilityand to ensure compliancewith the Maastricht criteria. Given the rapid growth insocial spending and the public sector wage bill (the two largest categories o f public spending) during the absorption boom, and the need to align public spending with permanently lower public sector revenue, fiscal adjustment should mostly rely on structural expenditure reforms for fiscal savings. For example, reforms to bring pension benefits in line with contributions would be needed to eliminate the deficit o f the social security (SODRA). Furthermore, fiscal risks from contingent liabilities in banks, '*These estimates incorporate fiscal measures already implemented, and an additional adjustment inJune for 1.5% o f GDP. 25 SODRA, local governments and guarantees related to SMEs and PPPs may further undermine fiscal sustainability. Fiscal institutions should therefore align incentives and capacity to monitor, disclose and manage risks and integrate them in fiscal policy analysis. Finally, going forward, changes in fiscal rules could be considered, to allow a role o f fiscal policy insmoothing volatility. 18. While the fiscal adjustment should be mostly expenditure-based, there may be room for achieving fiscal savings through some revenue measures. A full assessment o f the scope for additional revenue measures would require an overview o f the tax system. The next section provides some preliminary ideas/options for revenue measuresthat could bring fiscal savings over 2009-10. Reducing the Fiscal Deficit: Possible Revenue Measures 19. There may be room for fiscal savings through some revenue measures. A tentative list o f possible revenue measures affecting VAT and property taxes could yield up to 1% of GDP on a full-year basis, or 0.5% for the remainder o f 2009 ifimplemented as of Jul 1,2009 (Table 3). Other options could include user fees and the (politically very difficult) reversal PIT tax cut. Both indirect and direct tax collections are below the average for the Baltic countries and the sample o f nine new EUmembers for which data i s available. Enhancing monitoring o f revenue collections and developing a tax compliance strategy for the crisis could help minimize revenue risks. 20. Value-added tax (VAT). Lithuania's standard VAT rate remains modest by international standards, despite the recent 1 percentage point increase to 19%. This could be increased to 21% without exceeding the standard rates now levied inLatvia or Poland. It is estimated that the full year effect o f a 2 percentage point increase o f the standard VAT rate (from 19 to 21%) would affect about 0.6% o f GDP. This calculation incorporates estimated declines in the tax compliance ratio and the tax base (consumption) o f 1% as a result o f the 2 percentage point increase in the VAT rate. However, the fiscal savings may be overestimated if tax compliance weakens during the crisis. The estimate does not account for the cyclical impact o fthe economic downturn on the VAT receipts. Data from 2007 also point to inefficiencies in VAT collections in Lithuania (Table 5). The phasin out o f concessional VAT rates will help raise efficiency butmore stepsmay be needed.18 21. Personalincometax (PIT). Following tax cuts over the past decade, including in the 2009 budget, the PIT standard rate is now 15% (20% for dividend income). Abstracting from the inclusion (until this year) o f health insurance contributions in PIT, revenues from PIT (as a percentage o f GDP) have been below the average for the Baltics and for the sample o f nine new EU members even before the most recent tax cut. A reversal o f the most recent cut in PIT rates, i.e. increasing the flat tax rate from 15 to 13 While it is not possible to quantify near-term revenue gains that could be achieved ftom improvements in tax administration; the Fund's Fiscal Affairs Department has considerable experience in improving VAT administration in other countries in the region and could be available to give technical assistance, if requested. 26 18%, could yield fiscal savings o f about 0.8% o f GDP on a full-year basis.14 Such a measure, however, would also have some negative effect on consumption. 22. Property taxes. International comparisons show that in 2007, Lithuania's property tax revenue as a share o f GDP was the second lowest amongst the sample o f 9 new EU members. While recognizing that increasing property taxes in the midst o f falling real estate prices may not help economic recovery, an increase in revenue from property taxes may provide some scope for fiscal savings inthe medium term. Therefore, the authorities may consider ways to bringproperty tax revenues closer to its average for the region, which could yield a full year effect o f 0.4% o f GDP. However, in view of current constraints, it is assumed that the scope o f such an increase would be limited in 2009 and can perhaps result from a small increase intransaction taxedduties which may yield about 0.1% o f GDP. Table3: Possiblecompositionof fiscal adjustment IFull-year Impact in savings 2009 (Yoo fGDP) Possible Rewnue Measures (1+3) 1.o 0.5 Indirect Taxes VAT (1) Increase ofthe VAT rate by additional 2 percentage points to 21% 0.6 0.3 Direct Taxes PIT (2) Reverse income taxdecline in09 budget (e.g. through a reversal ofthe recent PIT taxcut) 0.8 ... RealEstate and Property Taxes (3) Increase revenue inline with the average for the new EUmember countries 0.4 0.2 Other revenue (fees, dividends) Possible increase in fees, other (4) 0.5 0.3 14In Table 3 such a measure is not assumed to be effective in 2009, but some impact in 2009 may be possible if the authorities are able to overcome any legal and/or logistical administrative obstacles to increasing the PIT rate inthe middleo f the year (or applying a surcharge). 27 m m e, c z .: . . . . . . . . . . . . . . : .. .. .. .. .. .. .. .. .. .. .. .. .. N j p i i j j j j i i j i 2 a 6 P P 3 3c 8 '4 : : : : : : : : : : : : : : : . . . . . . . . . . . . . . . ' 4 5 : j j j j j j jj z x 0 r r ;; P N P P O A EP +a z ai Table 5: VAT RevenueProductivityinEuropeanCountries Current Current Standard Other Positive Total VAT rewnue RevenueProductivity 21 VAT Rateon PUced) Rates (In percentof (Inpercent Based on consumption) of GDP) Consunption GDP Year Albania 20.0 8.7 8.3 0.437 0.415 2006 Austria 31 20.0 10.0; 12.0;16.0 10.6 7.7 0.529 0.385 2006 Belarus 18.0 10.0 17.4 12.3 0.964 0.682 2006 Belgium31 21.0 6.0; 12.0 9.6 7.3 0.457 0.349 2005 Bosnia and Hewegovina 17.0 13.0 14.2 0.767 0.835 2006 Bulgaria 31 20.0 13.6 11.9 0.678 0.593 2006 Croatia 22.0 18.3 13.9 0.830 0.632 2006 CYPNS3141 15.0 5.0; 8.0 12.6 10.4 0.840 0.691 2006 Czech Republic 3/ 19.0 5.00 9.5 6.7 0.499 0.351 2006 Denmark 31 25.0 13.7 10.3 0.548 0.410 2006 Estonia 31 18.0 5.00 12.7 9.1 0.703 0.506 2006 Finland 31 22.0 8.0; 17.0 11.8 8.7 0.538 0.394 2006 France31 19.6 2.1; 5.5 9.4 7.6 0.477 0.386 2006 Gemany 3/ 16.0 7.00 8.3 6.4 0.518 0.399 2006 Greece 31 19.0 4.5; 9.0 7.9 6.0 0.418 0.318 2005 Hungary 31 20.0 5.0; 15.0 13.1 10.0 0.654 0.501 2006 Iceknd 24.5 14.00 13.6 11.5 0.553 0.468 2005 Ireland 31 21.0 4.8; 10.0; 13.5 12.7 7.8 0.606 0.371 2006 Isael51 15.5 11.9 9.6 0.768 0.619 2006 Italy 31 20.0 4.0; 10.0 8.0 6.3 0.398 0.315 2006 Latvia 31 18.0 5.00 10.2 8.3 0.566 0.463 2006 Lithuania 31 18.0 5.0; 9.0 9.0 7.5 0.497 0.417 2006 Luxembourg 31 15.0 3.0; 6.0; 12.0 10.6 5.6 0.708 0.372 2006 Macedonia 18.0 5.0 9.0 8.8 0.503 0.487 2006 Malta 31 18.0 5.0 4.1 3.4 0.226 0.188 2006 Moldova 20.0 5.0; 8.0 11.9 13.8 0.596 0.692 2006 Montengro 17.0 12.2 12.7 0.716 0.747 2006 Netherlands 31 19.0 6.00 10.2 7.6 0.537 0.398 2006 Netherlands Antllles 5.0 3.0 6.6 5.3 1.319 1.055 2005 Norway 25.0 8.0; 13.0 13.3 8.0 0.532 0.322 2006 Poland 31 22.0 3.0; 7.0 9.5 7.7 0.430 0.351 2005 Portugal 31 21.0 5.0; 12.0 9.5 8.0 0.451 0.382 2004 Romania 31 19.0 7.00 9.4 7.8 0.496 0.413 2006 Russia 18.0 10.00 8.5 5.6 0,475 0.311 2006 Serbia 18.0 11.6 10.6 0.645 0.589 2006 Slovak Republic 31 19.0 9.8 7.6 0.517 0.397 2006 Slovenia 31 20.0 8.50 12.2 9.1 0.612 0.455 2006 Spain31 16.0 4.0; 7.0 8.4 6.4 0.528 0.401 2006 Sweden31 25.0 6.0; 12.0 12.5 9.4 0.499 0,376 2006 Switzerland 7.6 2.4; 3.61 5.57 4.01 0.733 0.527 2006 Tukey71 18.0 1.0; 8.0 9.0 7.2 0.502 0.399 2006 Ukraine 20.0 11.9 9.3 0.593 0.463 2006 UnitedKingdom31 17.50 5.00 11.91 10.46 0.681 0.598 2006 Unweigkedaverage (All) 18.76 10.77 8.60 0.594 0.475 (Excluding European Unioncounaies) 18.74 11.70 9.72 0.635 0.527 (Excluding OECD corntries) 17.8 11.16 9.36 0.659 0.551 Sources: IMF, Country documents; GovernmentFinance Statistics (IMF); World Economic Outlook (WEO); Revenue Statistics Database(0ECD); National Account Database (OECD); International Bureau of Fiscal Documentation (IBFD); Taxation and Investment in Central and East European Countries (IBFD); Corporate Taxes, Worldwide summaries(PricewaterhouseCoopers);and Fund staff estimates. Figures refer to General Government; Z/Revenue productivity = Total VAT revenue as percentage of consumption or GDP, divided by the VAT standard rate; 3/ European Union countries; 4/ Central government; 5/ State Budget; 6/ The reduced 3.6 percent rate applies to the supply of accommodation; 7/ InTurkey 26 percentand40 percentrates apply to luxury goods. 29 3. PublicAdministration and PublicWage Bill PublicAdministration 23. Lithuania is a regionalleader in developin its civil service among the new EUmembers. According to the 2009 SIGMAreport,' Lithuania achieved the highest fit with the European principles o fpublic administration and continued civil service reforms after its accession to the EU. The main challenge now, exacerbated by the economic downturn, i s to improve the performance o f the civil service and achieve better value for money, while preserving the merit system that upholds professionalism and ethics in the public administration. 24. Lithuaniahas a crediblesystem of strategicplanningthat informs allocation of public resources through the budget,but there is no monitoringsystem for policy outcomes. This enables the Government to make better informed decisions on core policy issues and presents a good foundation for developing performance informed budgeting. However, a system for monitoring policy outcomes i s not yet inplace, which hampers accountability o fpublic officials and government organizations. 25. The decentralizedinstitutionalsystem functions well but there are too many agencies reporting directly to the government and limited use of Information and CommunicationTechnology(ICT) hampers the integrationof governmentservices. The institutional system is built around the principle o f vertical hierarchy with decentralization o f tasks to agencies reporting to ministries and decentralization o f service delivery function to local authorities. The large number o f agencies reporting directly to the government (as opposed to sector ministries) creates problems with management and accountability. The poor use o f ICT limits the integration o f government services, resulting in inefficiencies and lower-standard services to citizens and businesses. 26. The size of the civil service is not too big but it has increased significantly recently.According to the Law on Civil Service of 2002, the civil service inLithuania is narrowly defined, covering 30,144 administrative personnel at the central and local level with responsibilities for management or government policies and execution of state budget. Additionally, almost 35,000 officials in police, customs, diplomatic service, and national defense form a statutory civil service, which is governed by special laws. In early 2009 the total civil service employed 1.4% o f the total labor force (3.1% including statutory civil servants). The size o f the general civil service has increased by 40% 15Meyer-Sahling, J.-H. (2009)Post-Accession Sustainability of Civil Service Reform in Central and Eastern Europe. Paris: OECD-SIGMA Publications. 30 between 2002 and 2008 from 19,000 to over 30,000 persons.16 This increase is partly explained by the direct impact o f EUmembership, which requiredthat some functions be performed by civil servants under a stricter legal regime thanthat o f a Labor Code. The public wage bill 27. The public wage billhas been growing rapidly inrecenttimes, particularly in 2008, following sharp increases in public wages. Between 2004 and 2008, the public wage bill grew by 50% in real terms (19% in 2008), and accounted for 29% of the increase in public expenditures in 2008. Most o f this increase is due to rising public wages (60% between 2004 and 2008). All public sector wages grew at similar rates but wages o f doctors more than doubled inreal terms, and all public sector wages grew faster in 2008 than comparable private sector wages (Figure 15). In2008, doctors were by far the best paid public employees, followed by public administration officials, and both groups had significantly higher wages than non-manual private sector workers. University staff had wages similar to non-manual private sector workers, while teacher salaries were significantly below. Since teachers are the main group o f public employees (Figure 16)," overall, the average public sector wage was 15% lower than the average comparable private sector wage.l8 28. The evolution of public sector pay has been partly motivated by increased EU labor mobility. Freedom of movement of persons in the EU created pressures for leveling salary levels with the EUaverages for core professional groups, such as qualified doctors. However, with increasing pay levels in the health sector, other occupational groups, such as teachers, used their union to push for higher pay.19 Also, despite the increases inthe salaries of doctors health services and outcomes have not improved. And despite the high salaries relative to comparable private sector workers, there were over 600 vacancies inthe public administration before the freeze imposed inDecember 2008. While improving public sector pay may in some cases be justified, increasing salaries across-the-board (as opposed to targeted increases tied to performance) i s not the most cost-effective policy to attract, retain and motivate public employees. 29. The escalation of the public wage bill is partly driven by political pressures and prompted by the lack of a reliable system for managing and monitoring public sector pay policies. While improving public pay may be justified in some cases, across- the-board increases (as opposed to targeted increases tied to performance) are not a cost- effective way to attract, retain and motivate public employees. And the escalation o f l6The highest increasestook place inthe systems of interior (265%), labor and social protection (187%), agriculture (174%), and courts and prosecution(1 14%). Municipal agencies added 38% ofnew civil servantsto atotal of 6,488. This public wage bill reported in Figure 16 only includes employees paid from the state and municipal budgets. It does not include those paid by SODRAS (e.g. health staff) or employees from public and state enterprises. '*In order to identify whose wage is above the market rate an analysis of wage differentials taking into account difference inworker andjob characteristics would be needed. Also, judges, usually a well-paid group to compensatefor corruption risks inthejustice system, have been given significant pay increases, settingthe high standardfor other sectors to follow 31 public sector wages in recent times was also driven by political pressures. While the Ministryo flabor and Social Security (MSSL) is responsible for ensuring balanced public sector pay policies, political pressures result in some public sector groups being favored over others. And the decentralized management o f the public sector payroll without central monitoring creates incentives for expanding personnel expenditure. Figure 15: Wages of public sector employeesand comparable privatesector employees _-_- __.""--.,-*-- - - _f_ ~ _ " -_- ~ _ l Y lx x1 - 1I XXII_-x__ 400 -- ~ x-"l_ *-_ ~ __x_ E-------- sp- I/ I I 350 publicsector M I 2 300 aJ - Public administration and 1 5 - I defence = -.E 3 250 - 3 School teachers * .- " & g 4: 0 200 ' 2 150 University education - N vi .- s 100 L - Doctors I Q 0 m I o 50 Non manual private 1 N I 0 employees --- _- - ~~~-~2005 - 2004 2006 2007 - - -2008- ~~--~~~-----~ - ~~-~~--~-~-~--~.--~--~~ ~ ~ Y x y _ _ ~ L ^__I x. Source: Labor Force Surveys 2004-2008 Figure 16: Composition of the publicwage bill in2009 I Political appointees QPoliticalappointees .Members of Parliament Civil Servants 0Judges DCivil Servants .Statutory Civil Servants OCivilSev at Local Government w/o teachers .Military 4% OState employees w/o teachers .Teachers I I 32 30. In 2009 the Government has cut the public administration wage bill as part of the fiscal adjustment, but further adjustments in the wage bill will also have to consider other public employees and alternative sources of savings. The 2009 budget appropriations for salaries were reduced for some categories o f public officials 31. Table 6): political appointees (17%), general civil servants (lo%), and the military (6%). The appropriation for statutory civil servants has remained practically unchanged despite their average pay being roughly the same as that o f general civil servants.20The April 2009 budget amendment included a reduction inpay levels for civil servants o f rank 15 and above o f 10% on average. At the same time, the appropriation for educational employees increased in 2009 by LTL 0.5 billion (or 28%), accounting for 36% o fthe total public sector appropriation for salaries at level Iand I1in2009.21 32. Giventhat civil service only accounts for 27% o f the total public wage bill and the wages o f other public employees have even increased in 2009 in some cases, any significant reduction inthe wage bill would have to involve other public employees and explore alternative sources o f savings. To generate additional savings, the Government could also consider structural consolidation, reduction o f scope o f non-essential functions and tasks, temporary freezes on hiring, temporary withholding of pay increases related to seniority (length o f service), discontinuing payment o f bonuses and additional pay for expansion o fthe scope o f work due to substitution o ftemporary absent employees. Table 6: The compositionof the pubic wage bill and changes 2007 2008 2009 Change 09/08 Level I Political appointees 70,799 79,670 65,739 -17% Members o fParliament 8,497 9,596 20,875 118% Judges 56,788 62,639 77,170 23% Civil Servants 653,529 853,454 772,046 -10% Statutory Civil Servants 730,141 934,617 925,815 -1% Civil Serv. at Local Govt. 202,568 263,339 256,387 -3% Military 281,817 314,372 295,796 -6% Total at levelI 2,004, 139 2,517,687 2,413,828 -4% Level I1 StateBudgetEmployees 1,288,814 1,560,141 1,562,305 0.14% o f which teaching staff d a 254,34 1 217,505 -14.48% w/o teaching staff d a 1,305,800 1,344,800 2.99% Municipalemployees 2,179,294 2,606,754 3,212,707 23.25% o f which teachers 1,4 12,320 1,665,63 1 2,249,024 35.03% without teachers 766,974 941,123 963,683 2.40% Total inc. teachers 3,468,108 4,166,895 4,775,012 14.59% Total teachers d a 1,919,972 2,466,529 28.47% other employees 2,246,923 2,308,483 2.74% Level I+II publicwage bill Total 5,472,247 6,684,581 7,188,841 7.54% Source: Ministry o f Finance. The public wage bill reportedhere includes employees paid from the state and municipal budgets. It excludes those paidby SODRAS (e.g. health staff) or employees from public and state enterprises. Level Irefers to state officials performing core administrative functions o f the 2oThe size ofthe civil service wage bill was reduced by 10%at the central level and by 3% at the local level (without social insurance contributions) (Ministry o f Finance). 21No disaggregated information is available for health workers. 33 state and local governments, while level I1 covers public employees (including teachers). The total state budgetis 26.9 billion LTL including the EUfunds (5.46 billionLT). 33. Government plans to further adjust public expenditure and improve business opportunities in the short runbased on the recommendations from various commissions. The Government Crisis Management Plan (December 2008) envisaged expenditure and revenue adjustments, as well as measures to improve business environment by reducing administrative barriers. Two types of commissions have been set up: a Sunset Commission and a Sunrise Commission. The former was tasked with trimming the public expenditure through structural adjustments: revising government programs, identifying institutional redundancies and inefficiencies, developing a methodology for review of budget programs with the aim of improving efficiency and cutting expenditure. The tasks of the Sunrise Commission are aimed at improving the growth prospectsthrough deregulation and government efficiency. Both commissions are expected to provide a pathway for immediate and medium-term efficiency gains for encouraging growth. To look at further potential structural adjustments within sectors, ministries have set up their Sunset Commissions to review the sources of savings and improve efficiency. Their results are expected by June, the period after the Presidential electionswhich would coincide with changes inthe Government. Proposedactionfor improvingthe efficiency of public administrationand the controlover the publicwage bill Civil service andpublic administration institutions 34. Improving the skills of the civil service. Enhancing civil service performance requires improving the professional qualifications in some parts of the civil service. To do that, an analysis of a requiredskill mix for the attainment of organizational objectives could be carried out, along with the preparation (and subsequent central monitoring) of a planto improve the skill mix through targetedtraining, better staff allocation, revisedjob descriptionsand performance appraisalaligned with organizational goals. 35. Introducing a strategic staffing policy. The staffing levels should be better aligned with the scope of work, to avoid misbalances in the workload across the civil service. The Government could consider introducing a strategic staffing policy, aligning skills and staffing levels with organizational priority objectives. In the process of the review of current functions, tasks, and structures to identify work areas that are no longer priorities being conducted by the Sunset Commissions, it is important to pay attention to the relevance of job descriptions to the organizational tasks. Staffing levels should be reviewed accordingly to matchthemto the business demands of each organizational unit. 36. Strengthening the mechanisms for central supervision of the civil service to ensure the effectiveness and unity of the civil service. The Civil Service Department (CSD) under the Minister of Interior is responsible for setting the methodological norms, drafting legislation, and maintaining civil service statistics, while the implementation of the civil service policy i s a responsibility of organizational managers (Le. staffing levels, 34 qualification requirements and individual salary levels). While formally the government sets the maximum staffing levels for institutions based on proposals o f managers, the justification presented by managers i s not subject to scrutiny by the CSD or the Ministry o f Finance. This may leadto organizational expansion, fueled by the in-builtincentives to generate wage bill savings for the allocation o fbonuses. 37. Thus, central supervision and monitoring should be conductedto ensure that (i) an appropriate level o f civil service employment i s maintained; (ii) descriptions reflect job organizational needs; (iii) civil servants' performance i s realistically assessed, and (iv) incentives for improved professionalism and performance are in place. And to maintain the quality o fpublic administration, the government needs to ensure proper legal regime, monitor merit based recruitment and promotion, enforce civil service discipline, protect professional independence o f civil servants, uphold ethical standards, monitor the performance o f the civil service as a whole and assure career and professional development opportunities for the civil service. All this requires a stronger central civil service management function. 38. Strengthening central monitoring of employment and wage bill to improve efficiency of public expenditure. The CSD maintains a database that can be used for a more rigorous monitoring o f staffing levels in the central public administration and suggesting action regarding vacant positions. Currently, the Law on Civil Service does not give the CSD authority to monitor staffing levels. However, it i s recommended that the centralized civil service establishment control be exercised by the CSD. This would include the review by the CSD o f the staffing plans, their relevance to the business needs, adequacy o f civil service post grading, compliance with the professional development and promotion criteria. While approval o f maximum staffing by the Government shall remain, the CSD should be asked to review the proposals for introducing changes and providing an opinion on the appropriateness o fproposals to change staffing. 39. Monitoring the adequacy of civil servant grading- based on job responsibilities.Whenunconstrained, managers will have incentives to assign higherjob grades than might be justified by the job description. This i s a source o f wage bill inflation. The CSD has identified, through its database, instances when the salaries for supposedly lower grade civil servants were higher than determined by law. Also the State Audit when reviewingthe annexes for justification of institutional wage bills found cases where calculations did not comply with methodology, developed by the Ministry o f Finance in2003, for determining the size o f the wage bill. Inthe context o f weak central control and absence o f externally imposed financial restraints, staffing may not reflect actual institutional requirements, resulting inlower efficiency o fthe civil service. 40. Authorizing the CSD to inspect the compliance of organizations with personnelpolicies. Based on risk assessment and using information in the civil service database, the CSD should have a right to review personnel arrangements in civil service organizations. This would be compatible with the role o f monitoring the implementation o f provisions of the civil service legislation. It would also facilitate uniform application o fthe civil service rules throughout the public administration. 35 41. The Government should protect the professionalism of the civil service. The proposed amendments o f the Civil Service Law, which envisage that, as o f July 1 2009, some senior positions in ministries (currently under secretaries) would become political, may lead to weakening civil service professionalism. The potential negative impact o f this measure could come from the possible loss o f the senior professional staff with management experience in ministries. Moreover, introducing the Vice Ministers to replace the state secretaries will mean that merit principles for selection o f senior managers may not be as relevant, and the continuity o f the public administration could suffer as a result o f the loss o f management capacity incurred during likely changes in management with every new government. 42. The consolidation of public administration through reducing the number of central agencies directly subordinated to the Government is a good step. Currently 20 agencies directly reporting to the Prime Minister are being considered for inclusion into ministerial structures. The rationale for this consolidation stems from the Constitution, which allocates political responsibility for management o f the executive to the members o f government, or to ministries, each with a specific policy mandate. But beyond the legal consideration, there i s an important consideration o f improving accountability for policy design and attainment o f policy goals. As political appointees, ministers bear all the responsibility for effective management o f the policy sector assigned to them. This responsibility i s met through managing public administration institutions operating within the sector and ensuring their accountability to the minister for achieved results. If this accountability line is broken and institutions are not accountable to the sector minister, the basis for ministerial accountability is incomplete, potentially leading to government inefficiencies. And it i s easier to allocate resources to priority areas ifpolicy implementation is consolidated under a single political leadership. Strategicplanning system 43. Introducing a robust system of monitoring policy implementation is likely to improve the accountability and government performance. While providing important directions and facilitating better allocation o f resources, the system o f centrally coordinated strategic planning can have a stronger impact on government performance if policy implementation is systematically monitored, and evaluated. Without monitoring core performance indicators it i s difficult for the public administration to stay focused on outcomes. The budget process should also be driven by plans based on a system of performance targets, which set the management and accountability framework for public organizations. 44. The Government Secretariat has prepared an EC-funded project to improve the system of monitoring of policy results. This project would support a system o f recording and monitoring indicators o f the core government programs in an IT-supported environment. Currently the government manages about 700 programs through some 50 main budget appropriators. Developing methods o f program evaluations (ex ante and ex post) and training would be also a useful component o f the proposed initiative. These 36 plans should be supported as they are likely to improve information for policy management based on well defined indicators. The outcome o fthe project i s likely to be a significant enhancement o f government performance inthe mediumterm. Having clearly stipulated and measurable program indicators which are monitored will help identify expenditure inefficiencies and assess public sector performance in a more objective way. Moreover, budget allocations may be better aligned with planned results. Realistically, the results o f such initiative will be discernible inthe medium to long term. However, in combination with quality management, introduction o f performance management may eventually place Lithuania in the group o f effective public sector performers within OECD. 45. Strengthening policy and budgeting links. The Government Chancellery is increasingly taking a proactive role in improving public sector governance through promoting more efficient public administration and effective resource allocation. The strategic partnership between the Government Chancellery and the Ministry o f Finance should be maintained and further strengthened to better support the government with setting policy priorities, ensuring their adequate funding and implementation. The importance o f this partnership is determined by the immediate impact o f these institutions on government decisionmaking. For the policy-budget link to be successful, the Ministry o f Interior and the Civil Service Department should also play a proactive role in actual monitoring o f the effectiveness o f institutional structures and the performance o f the civil servants. ThePublic wage bill 46. Strengtheningthe overall payrollmanagement.The government does not have a reliable system for managing and monitoring public sector pay policies given the decentralization o f pay policy decisions and the absence o f central monitoring o f the application o f this policy. Given the significant size o f the public sector payroll (9.3% o f GDP), it i s crucial to manage it effectively, especially when it comes to the need for macro-fiscal adjustment. Decentralized management o f the public sector payroll without central monitoring creates incentives for expanding personnel expenditure. The solution that countries have found effective i s introducing a public employee database and a payroll module in the central treasury. These measures give the government tools for analyzing implementation o f the pay policy and making better informed decisions based on actual data analysis. 47. It is recommendedthat a single central entity in the government be assigned with the responsibilityfor pay policy and public employee database. This measure would help maintain coherent, uniform pay policy and avoid uncoordinated increases o f pay levels for particular occupational groups. Additionally it would be desirable to introduce a payroll module in the treasury to monitor remuneration expenditure. However, none o f this would work without the political will to keep the wage bill under control. 37 4. Social protection Overall social protection expenditures, composition and benefits 48. Social protection (SP) expenditures (as a percentage of GDP and per capita) are low compared to most other EU countries in 2006, the latest year for which comparable data i s available (Figure 17). SP expenditures decreased slightly between 2000 and 2006 in line with most other EU countries. The composition o f SP benefits i s similar to the EUaverage, withpensions makinga bit morethan halfo ftotal benefits. Figure 17: SP expenditures are low compared to most EUcountries I 30 1 I 8,000 7,000 25 6,000 20 5,000 815 - VI 4,000n. * W m 0 %lo 3,000'E 3 L 2,000 5 1,000 0 0 I EU27 EU15 Bulgaria Czech Lithuania Latvia Estonia Hungary Poland Romania Slovenia Slovakia Republic - _ _ Source:ESSPROS database. PPS refers to purchasingparity standards. 49. But SP benefits have increased dramatically in 2008 (34% in real terms), particularly maternity benefits (Table 7). The SP system in Lithuania comprises two pillars: (i)social insurance (SI) (contributory) and (ii)social assistance (SA) (non- contributory). Although all benefits increased, social insurance benefits accounted for almost all o f this increase (81%). While pensions account for 57% o f the total increase, maternity leave benefits increased by 127%between 2007 and 2008. 50. The bulk of social protection benefits go to social insurance, particularly pensions and maternity (Table 8). Contributory pensions (old-age, disability and survivors) are the largest SI program (SO%), followed by maternity and sickness benefits. Old-age pensions have an earnings-unrelated component (tied to the base pension, 360 LT) and an earnings-related part, while sickness and maternity benefits are tied to earnings. Unemployment insurance has a fixed component (tied to the State Supported Income, SSI) and an earnings-related component. SA benefits only account for 14% o f total expenditures. Non means-tested benefits (NMT) are granted to families and 38 individuals irrespective o f their socio-economic level, and account for 86% of total SA spending. Non-contributory pensions are the largest social assistance benefit (the benefit is 90% o f the base pension).22 The largest social assistance benefit for families and children are child benefits (CB). Up until 2009 CB covered all children up to age 18 (benefits are tied to the Minimum Living Standard, MLS). From 2009, CB for children aged 3 and older are limitedto families with per capita income less than 3 times the SSI. Realterms, 2003 prices (million LTL) A s a O/O of GDP 2007 2008 Change Contri3ution 2007 2008 Total SP benefrts 7,455 10,005 34 100 8.7 10.5 S A benefrts 953 1,436 51 19 1.1 1.5 MT 132 201 52 3 0.2 0.2 NMT 821 1,235 50 16 1.o 1.3 S I benefrts 6,502 8,569 32 81 7.5 9.0 Pensions 5,357 6,820' 27 57 6.2 7.2 Sickness 442 568 29 5 0.5 0.6 Maternity 384 873 127 19 0.4 0.9 Other 319 307 -4 0 0.4 0.3 Table 8: The bulk of SP benefits go to social insurance, little left for MT programs (SP expenditures and composition, beneficiaries and average benefits, 2008) MillionLT Shares Beneficiaries Average I3 Social protection 11,705 100.0 Social insurance 10,025 85.6 Pensions 7,980 79.6 944,550 786 Sickness 665 6.6 ,053 70 Maternity 1.02 1 10.2 ,111 2,140 Unemployment insurance 306 3.1 23,700 591 Work injuries anddiseases 53 0.5 6,250 82 Social assistance 1,680 14.4 NMT 1,445 86.0 Childbenefis 482 33.3 624217 64 Social pensions 820 56.7 153,862 395 Other 143 9.9 h/iT 235 14.0 Social benefrt 79 33.6 37,300 176 Housingandutilityallowances 46 19.5 102,800 62 School meals Other 2,174 800 State Supported Income (SSI) 350 MinimumLivingStandard(MLS) 130 Source:MSSL.Unless otherwise noted, benefits are per person, per month. Beneficiaries for pensions includeold-age, disability and survivor, but average benefits refer to old-age only (disability benefit = 622 LT; survivor benefit = 84LT). Sickness and work injury benefits are per day, and the latter refers to accidents at work only. Number of beneficiaries and average benefit for maternity refers to the matemity/patemitybenefit. Number of unemployment insurancebeneficiaries is estimated. 22This benefit was introduced in2006 and currently covers 4% of all pensioners, but coverage is expected to grow to 15% of all elderly inthe next 2 decades due to reduced coverage inSODRA's scheme. 39 51. The social benefit program (SBP) is the core social safety net for the poor, but the program is very small to make a difference. Means-tested benefits only account for 14% of SA benefits (below below the EU27 average, 20% in2005) and 0.2% of GDP in 2008, down from 29% of total SA and 0.3% of GDP in 2004. MNT benefits include cash benefits and social services (benefits in kind) rendered subject to family's (or single individual's) income andproperty (inthe case o f SBP). The largest cash benefit program for the poor i s the SBP, which follows similar guaranteed minimum income programs (GMI) in other EU countries, and compensates families that meet the income, property and employment criteria for 90% o f the difference between their income per capita and the SSI. Program benefits, however, only account for 0.1% o f GDP and only cover a small percentage o f poor families (more on this later). Inaddition to this benefit, there are other auxiliary tested social assistance programs, such as compensations for heating, hot and cold water costs, free o f charge catering at schools for children from poor families. 52. The SP system covers (directly and indirectly) most of the population, being coverage greater for social assistance and among rural residents and the poor. The analysis o f the Household Budget Survey (HBS) for 2008 shows that 80% o f the population lives in a household where at least one member received at least one social protection benefit (Table 9). Coverage i s greater for social assistance (54%) and among rural residents (87%) and the poor (97%), reflecting the welfare-redistributive role o f SP benefits. Although the coverage o f SA among the poor compares well with other new member states,2315% o f the poor do not receive any social assistance. And the non-poor are also well-covered by SA, reflecting the lack o f means-testing o f most SA benefits.24 Old-age pensions and child benefits have the greatest coverage among social insurance and social assistance benefits, respectively. The benefit o f last resort for the poor, the SBP, only covers 6.7% o f the poor.2s 53. Social insurance benefits are significantly higher (relative to household consumption) than social assistance, which accounts for a small share of consumption among the poor. Although SA has greater coverage, the value o f SA benefits relative to the consumption level of beneficiaries i s significantly smaller (6%) than those of social insurance benefits (57%), reflecting the small expenditure on SA programs. Not surprisingly, relative benefits are higher for the poor than the non-poor, buteven among the poor SA benefits are only 23% of consumption. ~ ~~~~ 23 Nguyen, Sundaram, and Lindert (2009) "The Redistributive Impact of Social Protection Systems in EuropeandCentral Asia", mimeo, the WorldBank. 24 The percentage of total SA benefits kept by the poorest quintile of the population(40%) compares well withthat inother new member states(Nguyen, Sundaram, andLindert,2009). 25The SBP coverage amongthe poorestquintile ofthe population(4.7%) is amongthe lowestof suchtype ofprograms(GuranteedMinimumIncome)innew member states (Nguyen, Sundaram, andLindert, 2009). 40 Social insurance Overall SODRA's finances 54. During the past few years the number of SODRA's beneficiaries remained relatively stable while the number of contributors increased steadily (Figure 18). This was due to the sharp decline in fertility rates during 1941-1945, which translated into fewer retirees reaching retirement ages in 2001-2007, and the increase in the retirement age. However, during this period increased number o f disability claims dampened the temporary effect o f these tail winds. Meanwhile, the number o f contributors has been steadily increasing due to the continued coverage expansion which, combined with a relatively stable beneficiary population, should have generated at least temporary surpluses inSODRA's budget duringthe last decade. Figure 18: Evolution of SODRA's membership. Survivors Disability Old Age __ +Contributors 2001 2002 2003 2004 2005 2006 2007 2008 Source: MSSL data. 55. SODRA's revenues did increase but expenditures have slightly outpaced that growth as benefit levels were generously indexed and new entitlements were created (Figure 21). Maternity benefits increased by almost 3 times the revenue growth and reached 8% o f total expenditure in 2008. The base component of pensions also increased sharply in2008. These developments resulted inan expansion o f SODRA's spendingfrom 8.5% o f GDP during 2002-2005 to 11.3% of GDP in2008. 56. SODRA finances started deteriorating quickly in the last quarter of 2008 due to the contraction o f the economy, and the significant increases in social insurance benefits (particularly maternity benefits) in 2008 (Figure 20). The deficit at the end o f 2008 (1.19 billion) used up all the reserves accumulated in previous years. Without the cushion o f reserves, SODRA's finances are deteriorating even further in 2009 (the revenue shortfall between December 2008 and March2009 i s about 600 million LT). The estimated deficit for 2009 (under a 10.5% GDP contraction) is about 2 billion LT. 26Revenuesinclude state budget transfers to partially cover SODRA's transfers to pension accumulation funds. 42 57. Going forward, the number of beneficiaries is expected to increase in the next few years while the number of contributors is expected to decline. Large cohorts o f post war baby boomers are expected to retire in the next few years. And the sharp reduction infertility from 1991 untilrecently will result infewer numbers o f 18 year olds joining labor market starting from 2009. Coverage expansion experienced during the booming years i s not likely to continue. Therefore, SODRA i s expected to experience much slower revenue growth and stronger fiscal pressures in the coming years and decades. Given that such a highproportion o f SODRA's spending is related to pensions, and especially to old age pension, these developments will likely necessitate some revisions inthis group o f expenditures. 58. The developments require SODRA to adjust its finances by lowering expenditures. So far SODRA has been able to cope by borrowing at relatively low rates and very short maturity and temporary reduction o f contributions to the pension accumulation funds, but it will not be able to hold for much longer given the rapid deterioration o f the economy and the structural nature o f the deficit. Social security contribution rates in Lithuania (33.7% o f gross earnings) are on the high side compared to other EU countries.27 While there i s room to increase revenue collection through improved administration, the bulk of the adjustment will have to be through cuts and reforms insocial insurance benefits. Figure 20: Revenue and Expenditure of SODRA Lt, mln 15 - - - - . -- Maternity benefits ~ _ " I " I_ ~ 10 -* Transfersto pensionfunds 5 Other expenditure I 0 - - Disability and survivor 2002 2003 2004 2005 2006 2007 2008 pensions Source: MSSL data. 27The MISSOC database: .http://ec.euro~a.eu/social/main.~sp?catId=443&lan~ld=en 43 SODRA's Old-AgePensions 59. SODRA is the main provider o f pensions to current retirees, financing this spending on a pay-as-you-go basis by a contribution o f 26.3% o f gross wage o f current contributors. Since 2004 contributors have a right to transfer part o f their contribution to private pension accumulation accounts. At the end o f 2008 this redirected contribution was equal to 5.5% o f contributor's gross wage and 69% o f all contributors have chosen this option (new entrants are choosing this option at a rate o f 90%). SODRA is being partially compensated for this loss o f contributions from the State budget. Initially 50% o f the lost revenue was compensated, although recent deterioration o f SODRA's fiscal position ledto 100% compensation for the year 2009. Benefit adequacy and structure 60. Average pension for the full career worker in Lithuania stands at 43% o f average insured wage,28 which given an average length o f service o f 36 years results in an effective accrual rate o f 1.2%. This i s in line with average in the OECD (Figure 21). However, OECD countries combine these accrual rates with higher retirement ages and less generous indexation provisions. Inaddition to the main pension, 45% o f pensioners receive another benefit, among which 31% receive survivor's pension and 12% get additional deprived person's pen~ion.~' Figure 21: Accrual rates in selected OECD countries. UK Canada Japan Norway Germany H W i I Switzerland Iceland Sweden Korea Belgium Finland Netherlands Lux France Austria Italy Portugal Malta Spain Turkey 0 0.5 1 1.5 2 2.5 3 3.5 Source: EdwardWhitehouse, "Pensions at a Glance". 61. The large base component of the SODRA pension discourages contributions of those who plan to have shorter careers and of high earners. SODRA's old age pension i s comprised o f two main parts: a base component that depends on length o f service o f the beneficiary and earnings-related component that is dependent on both length o f service and past income. Since base component stands at relatively high 55% o f 28Fullcareer worker inLithuania is definedas anyone with more than 30 years ofcontribution 29Deprived person's pension i s paid to victims o f wars, occupations, survivors o f forced labor camps, etc. 44 average pension (57% o f minimum wage), this structure generates a large redistribution from high income to low income workers. It also introduces significant incentives to evade the system or to under-declare wages. 62. The incentive structure faced by minimumand average wage workers i s presented inFigure 22. A person who never contributes to SODRA is entitledto the social pension currently equal to 324Lt. On the other hand, a person who contributes for 15 years i s entitled to 198Lt o f the base component and 169Lt/6OLt o f the earnings related component for average wage/minimwn wage workers, for a total o f 367Lt/258Lt, which i s slightly abovehelow the non-contributory social pension. In fact, the SODRA pension o f minimum wage worker contributing up to 19 years would just even the non- contributory pension. Recent changes in pension benefits have increased the weight o f the base component at the expense o f the earnings related part (Figure 23). Figure 22: Incentives to contribute 1,200 Total contributory pension 1,000 for averagewage worker(2257Lt) 800 Total contributory pension 600 for min. wage worker (800Lt) 400 Base component 200 --Noncontributory pension, (324Lt) 0 5 10 15 20 25 30 35 40 63. To understand the mechanism through which the weights o f the two pension components continue to shift consider the pension calculation formula.30First, pension 30Total Pension= EarningsRelatedComponent + Base component, where Earnings RelatedComponent = Pension rights * insured income, Pension rights = average life time coefficient, and Coefficient = (individual wagexinsured income). 45 rights are accrued in a form o f coefficients which are derived each year by dividing individual's contributory wage by insured income which i s a parameter approved in a discretionary manner as part o f SODRA's budget planningprocess. At retirement average life-time coefficient i s derived and then multiplied each year by the same insured income parameter to calculate the amount o f earnings related component. Base component level i s also approved ina similar discretionary way. 64. Recent attempts to introduce more redistribution into the system by increasing Base component have, under budget constraint, led to a suppression o f the insured income growth (Figure 24).31It also impliesthat inrecent years accrued pensionrights in a form o f coefficients have been growing explosively. Inaddition to being unsustainable, this also undermines the structure of the pension system by valuing most recent contributions more than those o f previous years and further discouraging long careers (e.g. pensioncontributions made in2007 are valued 20% more thanthose made in2001). Figure 24: Evolution of parameters determining pension calculation. " insured income / av. Insured wage ---__.- I insured income / av. Economywide wage **-tc- basic pension/average *- pension ~ Ixx-l_II ^I 2002 2003 2004 2005 2006 2007 2008 2009 Source: MSSL data. Proposed actions 65. Pension rights should be calculated based on statistical average insured wage. Linking the decisions on how to calculate accrued rights and how to index pensions to the same parameter i s not optimal and should be abandoned as soon as possible. Pension rights should be calculated based on statistical average insured wage (as originally conceived in 1994 and practiced until the year 2001. The decision not to index pensions fully to wage growth (currently implicit by the suppression o f the insured income growth parameter) should be openly discussed and made explicit. 66. The base pension should be tied to the cost of a basic basket of goods and services (currently 350 LT), which is calculated every year by the Statistical Office. This would tie the base pension to some objective measure o f basic needs, control its growth and improve incentives to contribute (by increasing the weight o f the earnings-related component in total pension wealth as wages continue to grow). This change would 3' Basedon labor force survey data, the average economy-widewage is 23% higherthan average wage insuredby SODRA. 46 generate savings o f 120mln Lt inthe first year and additional 40mln Lt per year for each 1%ofreal wage growth compared to the 2009 level. 67. A more ambitious option is to further replace the base pension and the social pension with demogrant, i.e. non-contributory pension to all citizens o f Lithuania. This would leave the earnings-related component a sole determinant of contributory pension. Figure 25 shows this could be designed in such a way that the total amount o f average pension i s unaffected (currently average pension corresponds to the average career o f 36 years). The demogrant would be financed from the State budget with appropriate transfer o f resources from SODRA to State budget. Figure 25: Alternative structure of pension components. 1,200 Total contributory pension for averagewage worker(2257Lt) 1,000 Total contributory pension for min. 800 wage worker (800Lt) Base component 600 -+a- Noncontributory pension, (324Lt) 400 -- Total contributory pension for min. 200 --- wage worker (800Lt) -+Noncontributory pension, (324Lt) Total contributory pension for 0 5 10 15 20 25 30 35 40 averagewage worker(2257Lt) Source: Staff calculations Retirement ages 68. Statutory retirement ages are too low. Statutory retirement age currently stands at 62.5 for men and 60 for women. This i s much lower than prevalent retirement age o f 65 in most developed OECD countries represented in Figure 26. The difference i s especially stark for women who, even after adjusting for differences in life expectancy, spend on average 2 years longer inretirement inLithuania compared to women inOECD country sample. Women also live on average 6.8 years longer inretirement thanmen. Figure 26: Age of Retirement in selected OECD countries. c3 Male Source: EdwardWhitehouse, "Pensions at a Glance", 47 69. Only around 50% of Lithuanian population completely retires at the time of reaching retirement age (Figure 27). Many others continue to work combining wage and retirement income. Average reported wage at age of 65 stands at around 2000Lt for men and 1500Lt for women, which is still significantly above the average pension of 831Lt at the beginning o f 2009. This suggests that a significant portion o f pre-retirement cohorts would not be affected by an increase inretirement ages were to be raised. 70. While this report is against early retirement provisions in general, the 2004 early retirement option for the long-term unemployed is well-designed. People who have been unemployed for more than 1 year can retire early provided that they are not more than 5 years younger than the statutory retirement age. The scheme stipulates a permanent reduction inpensions by 0.4% for every full month remaining untilretirement (which i s actuarially fair), and requires at least a 30 year-long career. This provision would also serve as a safety net against the increase inretirement ages. Fieure 27: Distribution of elderlv r>or>ulationbv benefit tvoe and activitv status. 100% 80% 80% ... 60% 60% 40% 40% 20% 0% - 55 56 57 58 59 60 61 62 63 64 65 55 56 57 58 59 60 61 62 63 64 65 contributors disabled, not working working pensioners and disabled pensioners not working Source: MSSL data. Proposed actions 71. Start the process of gradual increase in retirement ages equalizing eligibility conditions for both genders. Retirement age could increase by 2 months per year for men and 4 months per year for women, which would generate savings o f 65mln Lt the first year and 130mlnLtthe second year inaddition to contributing to higher future pensions. Long term sustainability of SODRA 72. Long term projection o f SODRA's pension sub-scheme finances was prepared based on 2008 data and the macroeconomic projections discussed in Section 1. The results are presented in Figure 28. The numbers exclude other spending categories that are already experiencing deficits such as sickness and maternity program, but provide an indication o f fiscal pressures that SODRA i s going to face inthe next few decades. 48 _ _- SODFU'sPensionDeficits - - - -x I I t I llll_ __I II__I-I-_^II^ " I %ofGDP 73. Figure 28 shows that the economic crisis i s generating additional short term fiscal stress on the pension system. Measures recently adopted by SODRA (Le. increased contribution rates for those working without the labor contract, postponed repayment o f SODRA's debts to pensioners who have been underpaid during 2002-2005; and temporary reduction o f contributions to the pension accumulation funds) are already taken into account inthese simulations. Those measures are clearly not sufficient to offset the effects o fthe macroeconomic shock or longer term structural fiscal problems. 74. While some adjustment of pensions to the declining insured wages is needed:* reducingpensionsdrastically in times of severe economic contractionmay be counterproductive.Giventhat pensioners represent a big segment of the population and have a highpropensity to consume, reducing their income might introduce a further pressure to the economy. Pensioners also represent population with lowest incomes often no longer able to work to supplement their incomes which makes them especially vulnerable to unexpected reduction inincome. 75. Proposed action: A preferred approachwould be to minimize reductions in pensions while at the same time committingto structural pensionreforms,including a gradual increase inretirement age and a switch to rules-based indexation, which would place a muchhigher weight on inflation rather than wage indexation. Disability pensions and compensationsfor nursingand care 76. Although Lithuania has made considerable progress in stabilizing and even reversing the growth o f disability rates in recent years, the number of disability pensionersstill remains very high compared to the average of 14 OECD countries presented inTable 10. These rates are especially highat pre-retirement ages reaching 30% o f the population aged 59 to 60. Growth innumbers o f disability recipients has stabilized in recent years inpart due to the creation o f a new Agency for Assessment o f Loss o f Working Capacity, which is charged with certifying disability for the working age population. 32 In the end pensions have closely followed wage increases inrecent years. SODRA's budget for 2009 does not foresee any pension increases, despite the decrease ininsured wages. 49 Table 10: DisabilityBenefit Claims, per 1000 by Age Group, 1999 Number of disabled per 1000 20-34 35-44 45-54 55-59 Poland 13 61 182 262 Norway 18 54 120 219 Netherlands 32 63 120 192 Sweden 15 37 60 171 Denmark 11 30 83 156 UnitedKingdom 31 52 89 144 Portugal 11 24 70 175 Switzerland 19 33 55 86 Australia 21 36 64 131 Spain 6 17 40 80 UnitedStates 21 35 56 93 Austria 3 11 40 188 Germany 4 14 39 106 Mexico 0 2 7 17 OECDAverage (14) 15 33 73 144 1 Lithuania, Lithuania, 2005 11 46 113 233 2009 8 43 108 226 Soui-ce: OECDdatabase onprograms for disabledpersons 77. A new benefit to compensate for nursing and care of the disabled and old has been introduced in recent years and the number of beneficiaries is growing at the alarming rate due to lack of proper controls. The benefit i s funded from the State budget and administered by Municipalities. The system lacks adequate controls, especially for the beneficiaries above retirement age for whom the eligibility for the benefit i s certified solely by health system professionals (doctor recommendations have to be reviewed by the Agency for Assessment o f Loss o f Working Capacity for those younger than retirement age). The number beneficiaries more than doubles immediately after reaching retirement age (Figure 29). This indicates a muchmore lenient certification process for older people. Total spending on nursingand care programs i s 15OmlnLt. Figure 29: Number of compensations for care by age of the beneficiary. Source: M S S L data. 50 Proposedactions 78. Strengtheningthe monitoringand controlsof the benefitcertificationprocess by having the Agency oversee the certification of benefits. The Agency certifying disability for the working age population has introduced impressive monitoring and control methods and has proven to be able to control the growth o f disability benefits. It would be a natural candidate to oversee the certification o f nursing and care benefits for the whole population. However, it would need to be compensated for the increased administrative costs 79. Carry on the reductionin the number of pensionsupplements that disabled population can be eligible for proposed by the Ministry of Social Security and Labor.Currently disabled can be eligible for 1) care 2) nursing 3) help and 4) transport supplements and the proposal would limit the eligibility to one benefit only. This i s estimated to save 20 mln Lt annually. Disabled and old age pensioners are also eligible for transport subsidies administered by the Ministry o f Tran~portation~~which would also merit a review to assess the efficiency, targeting and possible overlap with other programs like pension supplements for transport mentioned above. State pensions 80. State pensions are rapidly increasing and there is little control over them. State pensions are financed directly from the State budget supplementing SODRA's pensions and account for 0.34% o f GDP.34These pensions are awarded to servicemen (administered by Ministries o f Defense and Interior), victims o f wars, occupations, forced labor camps, etc. (administered by SODRA). Scientists, people who have provided exceptional service to the state, renowned artists, judges and professors emeritus o f state universities also qualify for state pensions which can be administered either by SODRA or by the corresponding line Ministry. Current spending on state pensions for servicemen i s not large {Table ll), the schemes are relatively new and are likely to growth but substantially. Currently the active insuredpopulation of these schemes constitutes 3% o f SODRA's active membership and covers 6% o f SODRA's insured wage bill, but this i s likely to increase given that retirement age in these schemes is often lower than the statutory retirement age. 81. Giventhat servicemenare already insuredby SODRA it is unclearwhy these additionalinsurance schemes are needed.One legitimate purpose would be to insure servicemen for the period between early retirement and the statutory retirement age. However, this i s not how these schemes function, introducing inefficient kinks in retirement benefit stream for these individuals. While it may legitimate to compensate for difficult work conditions, it i s not clear why compensation should be deferred. 33Childrenand students are also eligible for State transport subsidies. 34Thisestimate excludes transfers to SODRA related to the introduction of the pension accumulation accounts and state spending on social security contributions for unemployed and mothers of young children. 51 Pension supplements for judges, prosecutors, prison guard, military, police, etc: 128 Pension sumlements for victims / demivedDersons f 1) 191 Pension supplements for scientists and others (2) 58 Pension contributions from state budget for unemployed and mothers o f children <3 99 Refundto SODRA for transfers to pensionaccumulation funds (100% for 2009) 442 Total 918 Proposed actions 82. Schemes for servicemen should be taken outside of the control of line Ministries. One option would be to create independent funded occupational schemes for servicemen, which would be regulated similarly to the other pension accumulation funds, with contributions paid by the employing line Ministry. Alternatively, administration o f these schemes could be shifted to SODRA which has much greater capacity to monitor performance o fthe schemes and forecast future funding needs. 83. I t is not clear why compensations to victims, scientists and others should be associated to retirement age. Beneficiaries o f compensations to victims constitute 12% o f all pensioners and their household income i s 9% higher than that o f other households. While people inthis category have undoubtedly suffered from these different tragedies it i s not clear why compensation for this suffering i s associated with retirement age. If it i s meant to compensate for the lost ability to work or lost years o f contribution then the fact that these people are on average better off would contradict this purpose.36Rewards and recognitions for scientist suffer from the same lack o f justified association to retirement age. These accomplishments should be better rewarded through appropriate wage structure (inthe case o f scientists), recognition grants, lump sum awards, etc. Pensionsaccumulationfunds 84. Although no significant pension payments are expected from pension accumulation accounts until this component o f the pension system matures, this scheme i s expected to provide at least 25% of future old age pension benefits. 85. The economic crisis has severely reduced the values of assets in pension portfolios. Average losses inbalanced portfolios in2008 were 19%, but many individual portfolios, especially aggressive ones, have suffered even higher losses. However, given the very small weight o f this pensioninthe overall pension entitlements, a personretiring today would have lost less than 1%o f the overall pensionwealth due to this crisis. ''Excludes ~~ social disability and old age pensions and pension supplements for care and nursing, all of which are administeredby municipalities. 36 Also fact that practicallyall current retiree populationhave lived through at least one war also raises the question whether the 88% of the pensioners that are currently not receiving this supplement cannot claim havingbeenthe victims. 52 86. Average returns since inception of the pension accumulation funds (PAF) stand at negative 0.5% but are expected to grow. But over the coming decades returns achieved by these funds will likely be significant and positive (the average return before the crisis was 5.7%). Therefore, younger generations are likely to have sufficient time to compensate for the 2008 losses and to augment their portfolios at reasonably highrates o f return. This in contrast with negative real returns from PAYG (without Government subsidies), given the expected sharp demographic contraction. Thus, PAF will be increasingly important source o f retirement income o f future pensioners. 87. In an attempt to improve SODRA's finances the contribution rates to PAF have been temporarily reduced from 5.5% to 3% which i s scheduled to expire on June 30, 2009. A further reduction o f contribution rate to 2% i s scheduled from July 1, 2009 untilthe end of the year 2010. Prolongedperiods of reduced contribution could severely reduce accumulations inpensionaccounts and damage the fee base o f pension funds, thus jeopardizing the stability o f the industry.37 Proposedactions 88. The Government should find a way to credibly ensure that the temporary reduction in contribution rates will be allowed to expire. Inthe absence o f structural reforms, SODRA's budget will not be balanced by the end o f 2010 (with or without contributions from the PAF). The best way to credibly ensure that the temporary reduction incontribution rates will be allowed to expire would be to commit to structural sustainability-enhancing reforms to SODRA's benefits. 89. It is also important that additional regulation is developed for the PAF to further strengthen it. One o f such enhancements currently being considered by supervisory authority i s an introduction o f the life cycle funds which would ensure that in pre-retirement years accumulated pension funds get invested in increasingly less risky assetsto better protect these portfolios from catastrophic losses close to retirement. Maternity leave benefits 90. Maternity leave benefits increased dramatically in 2008, as the period for receipt of maternity/paternity benefit was extended to 2 years. Nominal spending increased from 437 billion in 2007 to 1 billion in 2008, the largest proportional increase among all social protection programs, accounting for 19% o f the total increase in SP expenditures, and 50% o f SODRA's deficit at the end o f 2008 (including a 41% deficit relative to notional contribution for maternity and sickness benefits). 91. Effective replacement rates are higher than stipulated by law, as beneficiaries inflate earning before leaving work. Although replacement rates are 100% o f the compensatory wage in the first year and 85% in the second year, effective rates are higher as, in many cases, beneficiaries agree with employers to get extra 37The expectationthat the reductionmight be prolongedwould havethe same effect, makingrestorationof the systemmuchmore difficult inthe future. 53 payments just before leaving work. This creates a clear disincentive to go back to work. The average monthly benefit (2,140 LT) i s slightly higher than the average wage in2008. 92. The benefit receipt period i s the most generous among EU countries. Maternity benefits include maternity (70 days before birth and 56 days after), paternity (one month after delivery), and maternity/paternity (up to2 years after birth for the parent that remains at home with the child). The pre-birthbenefit receipt period (70 days) and, particularly, the after-birth period (2 years) are the most generous inthe EU.38 93. The cost of the current program in terms of fiscal sustainability and w o r k disincentives far outweighs any potential benefit in terms of fertility and child care. The current benefit cannot be sustained in the current fiscal scenario or even over the long run, as social security contributions are already on the high side among EU countries. One alleged justification for the increased maternity benefits i s lack o f child- care facilities. But the market for those cannot develop without demand, and more appropriate instruments can be used to assist parents with child care cost. The main argument for the increased benefits is the alleged positive effect on fertility and thus long-term sustainability o f pensions. The literature shows that the effect o f pronatalist policies i s negligible if any. Countries spend various proportions of GDP on family policies the resulting differences inbirthrates are very modest ( 94. Figure 30). Figure30: Family Policy Expenses as a % of GDP and total fertility rates Source: InstituteofDemography, MoscowHSE.Countriesrankedby family policy expenses 1991-2001. Proposed actions a. Option 1. Revert the benefit receipt period to what it was in 2007 (1 year after the birthofthe child) at 100% replacement rate (higher thanthe 2007 rate, since the last 6 months were paid at 85%). This would still leave maternity benefits in Lithuania 38The MISSOC database includes detailed description of social protection system in all EU countries, compiled by national focal points and coordinated by DG Employment, Social Affairs, and Equal Opportunities.httu://ec.euroua.eu/sociaVain.i sp?catld=443&landd=en 54 among the most generous in the EU. Table 12 shows the estimated fiscal savings if the change i s effective from July 1,2009.39 b. Option 2. Inaddition to option 1, limit the total leave period o f the maternity benefit (currently 70 days before birth and 56 days after) and the maternity/paternity benefit (up to 1year after the birthofthe childwiththe proposed amendment) to 52 weeks (1 year) to be distributedbetween pre-birthand after birthby the beneficiary. ODtion 1 266 million LT 355 million LT 94. I t is recommended to carry on the proposed increased in the period over which the average wage is calculated for the purpose o f computing the compensatory wage to 12 months. This would reduce the effect o f any increase in the wage o f the beneficiary before leaving work, and reduce the incentive for beneficiaries and firms to agree on such increases. This would bringeffective replacement rates down 1OO%.40 Sickness benefits 95. Expenditures have increased in recent years but smoothly. This increase is due to the increase in wages, and thus average benefits (53% in nominal terms between 2006 and 2008), while the number o f beneficiaries increased by about 14%. 96. Benefits are generous but comparable to other EUcountries. The replacement rate (85%) i s about average in the EU and the maximum benefit period (122 days) i s generous but on the low side compared to other EUcountries (MISSOC database). 97. Sick days claimed at comparable to other EU countries, but there is some anecdotal evidence of doctor over-certification. Data from the World Health Organization (WHO) shows that the number o f absent days fiom work due to illness (9 days per employee per year in 2007), is about the EU average. However, anecdotal evidence suggests that there seems to be over-certification o f sickness by doctors (i.e. beneficiaries and doctors collude to certify sickness when sickness is not warranted). 98. And sickness benefits have doubled over what was planned in the budget for the period January-March of 2009. There is little objective basis to explain this 39 Assumptions: The change i s effective fiom July 1, 2009, and affects new mothers as well as motherdfathers already on leave by July 2009. The estimates are upper bounds as unemployment claims from individuals returning to work as a result of the policy change are not factored in. Even if they all became unemployed, there would still be substantial fiscal savings because average maternity benefits are substantiallylargerthanaverage unemploymentbenefits. 40 It is also recommended to eliminate the differential pay between the maternity benefit and the beneficiary's wage if he or she decides to return to work during the f i s t year after the birth of the child. This differentialshouldbe minimalifabove actionsuccessfillybringsreplacementrates downto 100%. 55 significant increase. Workers who are about to lose their jobs may claim sick leave inan attempt to have greater social insurance coverage after they lose theirjobs. Proposed actions 99. The World Bank supports the proposed April amendment measure to make the first 3 days o f sickness unpaid in order to reduce the incentive o f potential beneficiaries to request sickness when sickness i s not warranted. The Parliament, however, approved instead a reduction in replacement rates after the second day o f sickness. In addition to this measure, more control is needed over the increasing number o f sickness benefit claims observedinthe last few months. Unemployment benefits 100. Expenditures on unemployment insurance have only increased modestly between 2006 and 2008 (about 10% in nominal terms, 2.5% in real terms). This is because unemployment benefits (UB) are loosely tied to wages and the unemployment rate remainedbasically the same between 2006 (5.7%) and 2008 (5.85). 41 101. Benefits are among the lowest in EU while the benefit receipt period (between 6 and 9 months) is standard among EU countries. The average benefit in 2008 was only 58% o f the minimum wage (21% relative to the average wage in the economy), representing 28% o f total consumption expenditures o f beneficiaries (44% among poor beneficiaries). 102. The projected increase in UB in2009 needs to be factored in the estimation of the fiscal gap for 2009. The projected increase in unemployment for 2009 (13.5%) over what was previously forecasted (4.2%) would generate expenditures over and above plannedexpenditures o f about 234 million LT. 103. In absence of discretionary measures, UB are a key social safety net available to households to cope with the crisis, by providing temporary income relief to individualslosing theirjobs, provided they have sufficient insuredemployment history. Proposedactions As the government discusses the new draft law on unemployment insurance, the Government should refrain from any cut in unemployment benefits. This i s because (i)by all measures, these benefits are not excessive; and (ii) provides a UB key social safety net to households to cope with the crisis. The Government could consider a temporary increase in the benefit receipt period, as other EU countries are doing. Extendingthe unemplovment benefit period by 3 monthswould cost anadditional 141millionLT in2009. 41Benefits are equal to the SSI plus 40% of the average, over the 36 months prior to the unemployment spell, of the ratioof insuredwages to an artificially determinedreference insuredwage. 56 Social assistance Increasing social challenges faced by the SA system as a result of the crisis 104. The economic crisis is affecting households mainly through reduced employment (the unemployment rate is predicted to be 13.5% in 2009) and labor income (a predicted 20.9% decrease in nominal gross wages in 2009). Household wealth is also being affected by the reduced value o f financial and real assets (particularly housing), as well as the increase cost of debt. While poorer households are more likely to be affectedthrough reduced wages and employment, the impact through the value of assets and cost o f debt is more likely to affect wealthier household^.^^ And the ability o f household (particularly the poor) to cope withthese wealth shocks is limitedbythe credit squeeze. 105. These shocks to incomes and wealth will have a significant impact on household welfare, as measured by consumption expenditures. Using the 2008 Household Budget Survey (HBS), and assuming a 10% across-the-board reduction in labor income, it is estimated that the number o f people from households with per capita consumption below the poverty line (350 LT month, which i s equal to the SSI) will increase by 208,000 (48.6% increase) from 428,066 in2008 (12.7% o f the population) to 635,979 (18.9% o f the population) in 2009. The impact on consumption is significantly larger than on income because, relative to income, there were many households in 2008 with consumption levels right above the poverty line, andthus vulnerable to poverty. 2008 2009 Change Numbers 'YO Numbers 'YO Absolute 'YO Consumption Individualsfrom HHwith PCE < SSI 428 12.7 636 18.9 208 48.6 Individualsfrom HHwith PCE <2*SSI 1.803 53.9 2.085 62 282 15.7 Individualsfrom HHwith PCE < 3*SSI 2,669 79.31 2,809 83.51 140 5.3 Income Individualsfrom HHwith PCI < SSI II 207 6.1I 254 7.5 I 47 22.7 Individualsfrom HHwith PCI <2*SSI 1,125 33.5 1,305 38.8 180 16 Individualsfrom HHwith PCI <3*SSI 2,239 66.6 2,411 71.7 171 7.6 106. Inthe context o f increased social needs arising from the crisis, UB do not cover individuals who become unemployed but do not have sufficient insured employment history, nor the increasing numbers of working poor. Infact Table 14 shows that in2008 most UB went to non-poor households and their impact on poverty was negligible. 42 Data from the 2008 BHS shows that the ratio of average household per capita income to average household per capita consumption increases as the household gets wealthier, being less than 1 among the richest 20% of households, indicating net borrowing among these households. 57 107. Means-tested SA programs, particularly the Social Benefit Program (SBP), are the main (potential) safety net for the poor and vulnerable to cope with the crisis. Active Labor Market Programs (ALMP), particularly direct job creation programs (public works and subsidized employment in the private sector) can also provide complementary assistance. Inthe context o f increased social needs arising from the crisis, the following questions are analyzed below: a. Are means-tested SA programs and ALMP ready to respond to the increased social challenges? b. What changes are needed in these programs to better protect the poor and vulnerable duringthe crisis and beyond ina cost-efficient way? i. Efficiency-enhancingreformsintheSAsystem ii. Efficiency-enhancingreformsinsomeprograms iii. Improvementandexpansionofexistingprogramswithprovensuccessin protecting the poor and vulnerable 108. SA benefits are too low and not adequately targeted to the poor, resulting in a small overall impact on poverty and ~ulnerability.~~ About 15% o f the poor do not receive any kind o f social assistance, and the benefit o f last resort for the poor, the SBP, only covers 6.7% o f the poor. And most benefits go to non-poor households (64%), reflecting the lack o f means-testing o f most SA benefits.44 The value o f SA benefits relative to the consumption level o f poor beneficiaries is low (23%), reflecting the small expenditure on SA programs. As a result o f the inadequate targeting o f the poor and low benefits relative to their needs, the impact o f SA benefits on poverty i s small (the poverty rate would increase from 12.7% to 16% if there were not SA benefits). And this small impact i s achieved at a relatively high cost (the poverty gap reduction for each LT spent on SA is only 0.25). 109. The overall results on SA mask differences among SA benefits. Child benefits have the best coverage o f the poor (70%), while the SB i s the best targeted to the poor (67% o f the beneficiaries are poor). The child benefit also has the largest impact on poverty but SB i s the most cost-effective inreducing poverty.45 43Although not all SA programs aim to reduce poverty, looking at the distribution of beneficiaries and benefitsbetweenpoor and non-poor households, as well as cost-effectivenessof these programs interms o f "overty provide a useful benchmark Infact the distribution of SA benefitsis moreregressivethanthat of SI benefits, which is the opposite of what is found inmost countries. ''The pregnancy grant has a higher cost-benefit ratio but the sample i s very small to really make a statistically meaningful inference. 58 Improvingthe financingand monitoringof SA benefits 110. Currentlylocalgovernmentsdo the planningof SA benefits;they administer benefits and channel paymentsto beneficiaries.On the basis of approved allocations for each benefit, funds are transferred from MSSLto each local government on a monthly basis. These funds are earmarked for each benefit.46 Local governments administer the benefits (processing o f applications, verification and monitoring) and their administrative costs are financed separately by MSSL. MSSL monitors the implementation o f SA benefits by local governments, through quarterly reviews and selective inspections, and has authority to discontinue payments to local government ifirregularities are found. 111. The increasedunanticipateddemandfor SA benefitswill put the current SA financingsystem under stress. As the crisis increasesthe number of households that fall below the poverty line, and thus the demand for SA benefits, local governments can initially reallocate funds across benefits, but eventually they will have to request additional funds, which has to be first approved by MSSL and then by MOF. This can create some delay in payments to eligible households and possibly some rationing o f beneficiaries and benefits by local governments. 112. There is no central databaseof SA beneficiariesand benefits, and thus little overall controlover who gets what. Information on beneficiaries and benefits i s kept at the local government level, and in many cases this information is only available for each program separately. 113. Different benefits use different application forms and verification systems even when the information being collected and verified is the same. For example, applicants to the SBP have to go through a detailed form asking for different sources o f income, and this information i s subject to a strict verification process and thenreassessed regularly. In contrast, applicants to child benefits for children 3 years o f age and older only have to report their total income per capita, which i s not subject to any verification or reassessment. And yet both benefits are administered by the same municipalities. 114. The SSI level, which is used as a reference for determining eligibility and benefits of means-testedprograms, is determined and revised discretionally rather thanbeingtied to some objectivemeasure ofbasic needs. In2008 and2009, however, the SSI happens to be equal to the poverty line, which measures the cost o f providing a basic basket o f goods and services and it i s recalculated every year by the Statistical Office. Proposed actions 115. Bringing all SA programs under a central unique payment system with a single beneficiary registry that will pay all SA benefits directly to beneficiaries (instead o f payments being channeled through municipalities). The planning, administration o f benefits and monitoring o f beneficiaries would still remain under local governments. A 46(local governmentshave some flexibility to transfer funds across SA benefits at the end of eachquarter. 60 number o f EU member states are implementing such a unique payment system for SA benefits (e.g. France, UK, Romania, Bulgaria). Such a system would not reduce accountability o f municipalities, as under the current system municipalities have very little control and discretion over SA funds. This unique payment system would bring about the following benefits: a. Reduced transaction costs in the financing o f benefits, including savings from supervision o f municipal finances. b. Much faster response to the increased demand for SA benefits during the crisis, preventingthe rationing o f benefits and/or beneficiaries. c. More control over SA beneficiaries and benefits through a single beneficiary registry. 116. An alternative to the creation o f a unique payment system for SA benefits only would be to use SODRA. Using SODRA's infrastructure would reduce costs relative to having a separate payment system. It would also improve control over all SI and SA benefits and beneficiaries. SODRA would have to be compensated for the increased administrative costs. 117. Having a common set of questions, including questions on family size, composition and income (for income-tested benefits) for all application forms for SA benefits. 118. Using the same system for verifying and reassessinghouseholdincome for all means-tested benefits. Given the need to respond quickly to the increased demand for means-tested SA benefits, the Government may consider speeding up the verification process or relaxing it temporarily. 119. Making the SSI equal to the poverty line, which is recalculated every year by the Statistical Office, rather than being determined discretionally. This will increase control over benefits and ensure adequate coverage o f the poor and adequacy o f benefits to needs. Social Benefit Program (SBP) 120. The SBP is the socialsafety net of last resort for poor households.This benefit mimics the guaranteed minimum income programs (GMI) in most EU countries. It compensates families that meet the income, property and employment criteria for 90% o f the difference between the SSI and household income per capita (for each family member). 121. The program has good design features. The targeting mechanism combines income-testing with testing based on the value o f household assets, which addresses the problem of income underreporting among individuals working in the informal economy. The program also provides strong links with the labor market: adult and able family members must be either employed or, if unemployed, they have to be registered in the labor exchange office, and either be ready to take on ajob offer or participate inALMP. 61 122. SBP also has good implementation features. Applicants have to go through a comprehensive application form, after which information is carefully reviewed and verified and benefits are granted. And beneficiaries are reassessed every 3 months. 123. As a result, the program is well targeted and benefits compare well to needs of the poor. About 66% of the beneficiaries are below the poverty line, the best targeting performance of any SP programy4'and benefits account for 29% of the consumption of the poor. It has a large impact on poverty among beneficiaries at a relatively low cost. In fact the SBP is the most cost-effective SA program inreducing poverty. 124. But the program has very low coverage among the poor (6.7%), resulting in a low overall impact on poverty. Understanding the reasons behindthis low coverage is essential in order to identify how to improve and expand the SBP to meet the increasing needs of the poor. 125. This low coverage may be partly explainedby the introduction in2004 of the value of household assets as an additional eligibility criterion, as some (mostly elderly) poor owning a home were no longer eligible to receive benefits. The number of beneficiaries and expenditures fell substantially in2004. The program went from 0.2% of GDP in2000 to 0.1% in2004, where it has remaineduntil 2008 (Table 15). Only Latvia had a smaller program relative to GDP in 2004. The targeting performance of the program (% of beneficiaries from the poorest 20% of households) increased from 60% in 2000 to 73% in 2008. However, this came at the expense of a proportionally larger decrease inthe coverage ofthe poor (from 11%in2000 to 4.7% in2008). Tab time Source: Based on the analysis of HouseholdBudget Surveys (2000, 2004, 2008). Coverage is the YOof beneficiaries amongthe poorest20% of households.Targetingis the %of beneficiaries from the poorest20% ofhouseholds. 126. But coverage was low even before the 2004 reform-low coverage is more likely due to low demand, but this is likely to change with the crisis. Many poor families may have decided not to apply inthe past given the long application process, the strict eligibility rules (particularly activation), and a cash amount they could get from somewhere else. However, the situation i s now different, as more families fall below the poverty line andpoor households face more limitedincome-generatingopportunities. Proposedactions a. Ensuringtimely financing of benefits through a uniquepayment system 47Interms of the percentageof total benefits that go to the poorest quintile of the population(59%), the SBP is among the best performing GMI programs among new member states (Nguyen, Sundaram, and Lindert(2009). 62 b. MakingSSI equal to the poverty line from now on. C. Raising the benefit to 100% o f the difference betweenthe SSI and household income per capita. d. Improving or relaxing (at least temporarily during the crisis) the household assets eligibility criteria, given the little value added in terms o f targeting performance o f relative to its cost interms o f reduced coverage ofthe poor e. Speeding up the application and verification process or even relaxing the verification process temporarily. The Government could also consider doing the reassessment every 6 to 12 months instead o f every 3 months. f. Meetingthe additional demandfor SBP, targeting 30% of the poor. Itis possible to establish some benchmarks for the additional SBP expenditures in 2009 resulting from the crisis under alternative assumptions about the coverage of the program among families with income lower than SSI (Table 16).48The lower bound estimate measures the additional cost from the increased number o f households whose incomes fall below SSI (assuming they also meet the house value and employment eligibility criteria) and the increased benefits o f existing beneficiaries (because of reduced incomes), assuming the same coverage o f the poor as in 2008 (17%).49 Alternative estimates assume larger coverage rates among existing and new poor. Under a desirable and realistic coverage rate assumption (30%), the additional SBP spendingin2004 would be 57 million LT. Table 16: EstimatedadditionalSBP costs in2009 underalternativecoverage rates I Cowrage of families below SSI 17% 30% 40% 50% Additional costs 16,835,499 57,219,831 88,100,647 118,981,462 Child Benefits 127. The child benefit (CB) is the largest SA program for families and children in terms of beneficiaries and expenditures. The child benefit is equal to 70% o f the MinimumLiving Standard (MLS) for childrenunder 3 and 40% ofthe MLS for children aged 3 and older.50Benefits are significantly larger for families with 3 or more children (110% of MLS for children under 3 and 40% for children 3+), but these families only account for 2.8% o f all families with children and are significantly poorer.51Children older than 18 can continue receiving the benefits untilage 24 provided they are full-time 48It will not be possible to move the current low coverage situation to 100% coverage all poor eligible households. It is also not possibleto predicthow many qualifying families will claimthe benefit.And some ofthe benefits will go to non-poorfamilies. 49The HBSdoes not have enoughinformationto check the house value and employment eligibility criteria. 50The MLS is the cost of a subsistencebasket of goods calculatedby the StatisticalOffice every year. 5'About 82%ofthese families are below2 times the SSI in2008, relativeto 27% of families with less than 3 children. 63 students. Otherwise, the child benefit is not attached to any conditionality such as school participation. 128. Child benefitshave the bestcoverage of the poor (70%) among SA programs, but most of the benefits go to the non poor (SO%), reflecting the universal nature of this benefit. While benefits are low relative to consumption, the large coverage of the program among the poor explains the larger impact on poverty compared to other SA programs. However, since the benefits are also given to children innonpoor families, the impact on poverty is achieved at a highefficient cost (i.e. low cost-benefit ratio). 129. The December 2008 crisis management plan limited child benefits for children aged 3 and older to families with per capita family income less than 3*SSI (1,050 LT per month), covering 64% o f all families with children. Families applying for benefits for children aged 3 and older only have to self-report per capita family income, and this information i s not subject to any verification, nor reassessment. Benefits for children under 3 remain universal. A number of new member states have targeted child benefit programs, including Czech Republic, Slovenia andPoland. 130. Additional CB expenditures resulting from the crisis. With no change, and assuming that all qualifying families apply for the benefit, the increase in spending resulting from the increased number of families falling below 3*SSI in 2009 could be about 19 million LT for 2009. 131. While the CB does not have a poverty-reductionobjective, granting benefits to wealthyfamilies for whom these benefits do not affect their decisions to have children or to invest intheir well-beingresultsin large efficiency costs. Evenseeing this benefit as a right for all children, it i s hardto justify the financing o f such a right for all children inacontext oflimited fiscal spaceandincreasing socialneeds. Proposed actions a. Limiting the benefits for children under 3 to families with per capita family income less than 3 times SSI. Assuming all families that qualify apply for the benefit and that the change would be implemented from July 1, savings from this measure between July and December o f 2009 would be about 22 million LT (42 million on an annualized basis). b. Lowering the income eligibility threshold to 2 times the SSI. Assuming this applies to children o f all ages and the change i s implemented from July 1, savings between July and December o f 2009 would be around 89 million LT (178 millionon anannualized basis). C. Limitingthe benefits to childrenupto age 18 d. Subjecting applicants to the same questions about income and information verification process as for SBP applicants. 64 Active Labor Market Programs(ALMP) 132. Since 2004 the Government has taken efforts to link SA benefits to participation in ALMP. ALMP are mainly targeted to the long-run unemployed. The increased demand for UB has limited the scope for financing ALMP out o f the employment fund. In 2009 18% o f spending on ALMP will be financed from the employment fund and78% from the European Social Fund(ESF). 133. ALMP include skills training, subsidized employment in private firms, job rotation (temporary employment to cover temporary leave o f employees), public works, as well as programs for the disabled (subsidized employment and vocational training). The total budget in 2009 for these programs is about 300 million LT. Most funds go to skills training relative to direct-job creation programs (public works, subsidized employment inprivate firms). Proposed actions a. ALMP are a good complement to SA to meet the increased demand for income relief while increasing the employability o fbeneficiaries. b. The Government needs to ensure ALMP are available for SBP beneficiaries, as ALMP participationis one ofthe eligibility criterion for SBP benefits. C. Given the low demand for labor during the crisis and need for temporary income relief, the Government may want to consider shifting the focus o f ALMP spending towards direct-job creationprograms. d. The Government may want to consider employing more public works beneficiaries in public infrastructure projects, as this provides value-added in terms o f public infrastructures andthus economic growth. 65 5. Health Overview of the healthsystem and its performance 134. Over the past 10 years, the health system has moved away from an integrated towards a contractual where universal insurance i s provided to the population by the State Patient Fund(SPF), which pools more than 80% o f total heath expenditure and purchases services from providers. The Ministryof Health still runs 19 heathfacilities but its primaryfunction i s the general supervision o f the system. 135. Total health expenditure remains in line with Lithuania's income level. In 2007, Lithuania spent about spent 6.3% of GDP onhealth or PPP$l,115. Figure 31 shows that total health expenditure, as a percentage o f GDP and in absolute terms, remains comparable to countries inthe region inthe same income range. Figure31: Health -expenditurein Lithuaniaand comparableEuropeancountries(2007) _ _ _ _ _ - - 9 2,500 8 - 7 2,000 E5 n. W ijb xm w 5 1,500 & n - 0 4 .-Ea!= 2 c. sr c. 3 W W P 1,000 E W -W 3 5 -W m r 2 - com 500 c. co 0 Macedonia Romania Poland Estonia Lithuania Latvia Slovakia Hungary Czech Slovenia (S8,SlO) ($10.980) ($15,330)($18.810) ($17,180) ($16,890) ($19,340) ($17,210) Republic ($26,640) ($22,020) ...I ....I. ^_^_._x - "- I I _^__" Note: Inparenthesis below the country S name, GNIper capita $PPP 2007, source: DDB, WorldBank. Health .- expenditure datafrom WHONHA (httD://www.wlzo.int/nha/en:). 136. Health expenditure in Lithuania has been increasingrapidly over the past few years, particularly doctors' salaries. Data from the National Health Accounts shows that total health expenditure increased by more than 70% between 2004 and 2007 when it reached 6 billion LTL (Figure 32). The largest increase corresponds to public expenditures which increased by 85% over that period.53 Data from the Ministry o f 52 An integratedhealthmodel i s one in which coverage of the population (the health insurance function) is provided by the Ministry ofHealth which produces health care in a vertically integratedsystem. 53 Data on private expenditureis likely to be under-estimated. 66 Finance shows that public expenditure continued to increase rapidly in 2008 and that, if the April 2009 budget amendment i s adopted, it should stabilize around 5b LTL, twice its 2004 level. Public health expenditure between 2004 and 2008 grew more rapidly than the GDP (Figure 33). Health workers salaries, particularly physicians', grew more rapidly than the cost o f other health inputs. The average net monthly earnings o f doctors increased by 120% in real terms between 2004 and 2008, by far the largest increase among other public sector employees and equivalent private employees, making doctors the best paid professional^.^^ Figure32: Publicand private healthspendinginLithuania (2004-2007), currentprices 7 Private (+42%) 6 H Public (+85%) 2004 2005 2006 2007 2008(preitm.)2003(plan) Sources: 2004-07: National Health Accounts, Statistikos Departamentas. 2008 and 2009: Public expenditure only MoF, no estimate availablefor private expenditure. Figure33: Increasein healthworkerssalaries comparedwith other variables(2004-08) 300 280 -Doctors salaries (LFS) .Doctors 260 salaries (SPF) All health personel (LFS) 240 -Public health expenditure 220 GDP (nominal) 200 180 160 140 120 100 2004 2005 2006 2007 2008 Sources Labor Force Survey, State Patient Fund 137. Lithuania lags behind comparable countries in terms of health outcomes, which have not improved in recent times. Life expectancy, which used to be above that 54The earnings o f civil servants went up by 40%, teachers by 60% and non-manual private employees by 50%. In2008, the salary o f doctors is 70% higher than the average salary o f public sector employees (30% higher thanthe salary o f civil servant, 50% higher than university staff) and 40% higher than non-manual private employees. In2004, the salary o f doctors was below that o f civil servants and at the same level as that o f university staff and non-manual private employees. 67 o f other Baltic countries, i s now the lowest in the region and it continues to decline (Figure 34). According to data from the World Health Organization (WHO), the incidence of tuberculosis i s 70% higher thanthe average inthe new EUcountries. When compared with the same group of countries, maternal and child mortality are slightly lower; mortality by cancer i s about average, but mortality from diseases of the circulatory system, and inparticular ischemic heart diseases, i s the worst among the countries inthe region ( 138. Table 17). Mortality by external causes i s also staggeringly high and in particular Lithuania's suicide rate i s the highest among all o f Europe and Central Asia. Alcohol related mortality i s more than twice that the average o f new EUmembers, and mortality induced by smoking i s the highest among the countries compared. Many o f these premature deaths could be avoided through public health interventions, prevention, and early detection and treatment in primary care settings, which are relatively under- developed. This also suggests that Lithuania should and could get better health outcomes given its current investment inhealth. 68 Figure34: Trends in Life Expectancy 16 ~ Finland --Po Iand 68 - ELI members post 2004 ~ Estonia -Latvia --Lithuania 66 1999 2000 2001 2002 2003 2004 2005 2006 2001 Source European Healthfor all Database. httu /hww.eurowho int/HFADB Table 17: Health status inLithuania and Europeancountries(2007 or latest available) j . . , : : ; " y . 1 l i l l i l . l : l i . ~ ~ ~ I I :Standardized~ Death Rate( er 100,000 : : : : i / E ~ , Alcbhol Smoking Maternal related related deaths per at birth disease causes causes I00000 post 2004 74 200 468 178 63 991 373 7.3 EU I 79 I 15 I 176I 252 II 96 I 40 I 641 225 II 5.8 Source: EuropeanHealth for all Database. httu://www.euro,who.int/HFADB.The figures for TB include relapses. 139. Lithuanians are dissatisfied with the health system, particularly its level of corruption. In 2003, more than one in four adults declared that they were dissatisfied with the health system.55Data from the 2007 EU-SILC survey suggests that Lithuanians are in fact less satisfied with their health system than other European citizens: 10% of them declare that some of their medical needs have not been met inthe past 12 months (the average for EU-25 i s 6%), and waiting lists are the most frequent reason alluded.56In 2002, 64% of participants ina social audit conducted inthe Baltic countries reported that corruption inhealth services was very high inLithuania (compared to 45% inLatvia and 43% in Estonia). Although people are generally reluctant to provide such information, 8% of interviewees declared they had made an unofficial payment in the health sector 55Bankauskaite and Saarelma (2003). Why are people dissatisfied with medical care services in Lithuania? A qualitative study using responses to open-ended questions International Journal for Quality in Health Care 15:23-029. 56Eurostat, Perception o f health and access to health care inthe EU-25 in2007, Statistics in focus 212009 69 (compared to 3% in Latvia and less than 1% in Estonia).57In 2008, 31% o f a sample o f residents said that health services were very corrupt. Health facilities were the four most frequently cited public institutions where residents said they had paid a bribe.58 140. Overall, this information suggests that the rapid increase in public health expenditure over the past few years, driven by increases in health workers salaries, has not translated into improvements o f the health system's responsiveness to patients' needs andhealthoutcomes. The Hospital System 141. Despite significant progress since independence, Lithuania's hospital infrastructureremains oversizedand needs to be better adaptedto the needs of the population.There were 12.4 hospital beds per thousand population in 1990 and 8.1 in 2007, which i s one o f the highest rates in the region (Figure 35). There are nearly 5 hospitals per 100,000 population, which i s twice the average o f new EUmembers. Fifty- three hospitals have less than 50 beds and an additional 25 have less than a hundred. Evensmall hospitals incur significant fixed costs. Moreover, hospitals with less than 200 beds are generally considered to be too small to provide a full range o f general hospital services. Distances in Lithuania are not very large, and there i s significant scope to consolidate hospital infrastructure without compromising access. Further restructuring would also be an opportunity to adapt services to the needs of the population. For instance, although most tuberculosis patients can be treated outside o f the hospital, one specialized bed i s available for 2 patients in Lithuania. On average, 20 births occur per year ineach obstetric bed, which i s too low to ensure adequate safety for patients.59 57 Cockcroft, Andersson, et al. (2008) An inter-country comparison o f unofficial payments: results of a health sector social audit inthe Baltic States, B M C Health Services Research 8:15. 58 When asked where they paid a bribe, 20% responded "in a town and regional hospital", 16% "in a clinic", 15% in a "national hospital", and 14% "a state health care institution". The 5' institution cited by 11% ofresidents was traffic police. Lithuanian Map o f Corruption, Transparency International, 2009 59 This estimate is based on official statistics dividing number of live births by the number of "obstetrics and gynecology beds" (http://db1.stat.gov.lt/statbanMSelectTable/OmradeO.asp?PLanguage=l).According to the Ministry o f Health the number is 60 deliveries per obstetric bed. 70 Figure35: Hospitalinfrastructure inLithuania andother European countries (2007) 142. Aggregate data suggest that hospital productivity i s at a reasonable level, but more patients should be treated outside of the hospital-Lithuania has one of the highest inpatient admission rates in Europe. The average length o f stay for acute care was approximately 6.6 days in 2007,60close to the OECD average in 2005 (6.9 days). 61 Bed occupancy was 77% in2007, which i s the average for EUcountries in2006. But the hospital system suffers from a number o f deficiencies. More patients should be treated on an outpatient basis-Lithuania has one o fthe highest inpatient admission rates inEurope. Since 2006, efforts have been made by the State Patient Fund (SPF) to encourage the development o f day surgery, and results appear encouraging: the number of inpatient stays (excluding short stays) decreasedby 11% in2006 and 1.2% in200862.These efforts would need to be supported by a reduction in the size o f the hospital infrastructure and the strengthening o f primary care as an alternative to cater to patients needs. Figure36: Admission rates inLithuania and other European countries (2007) 25 Source.European Healthfor all Database. http //www euro who.mi/HFADB 143. A second set of concerns about the hospital system broadly relates to a lack o f transparency and poor governance. Most hospitals are non-profit entities owned by 60 '' LithuanianHealth Informationcenter OECD Health Data 2008 - Frequently RequestedData, http://www.oecd.org/ Patientsfund, Overview of2007 - http://www.vlk.lt/vlk/en/?l=welcome 71 Lithuania CzechRep. Hungary Poland Mexico MRI 3.2 3.8 2.6 1.9 1.4 CT scan 14.2 13.1 7.2 9.2 3.6 144. Quality managementand controlare weak. Health institutions report information about whether they meet some input norms, based on which they receive an administrative license to provide a range o f services from the Accreditation Agency. Although the Agency has been in place for over 10 years, the legal framework or processes for accreditation are still not in place. About 30 hospitals have sought, at their own initiative, an I S 0 ~ertification,~~ received no financial support or specific but recognition. Other institutions dealing with quality o f care include the medical audit inspection unit, which investigates patients' complaints and promotes patients rights. Facilities are supposed to report infections and various adverse events but reporting i s patchy and inaccurate. The Ministryo f Health issues orders describing standards o f care as well as prescription guidelines, and the State Patient Fundperforms audits to monitor compliance, but it i s unclear whether the scope and frequency o f these audits are genuinely conducive to developing a quality assurance culture in the system. No Health Technology Assessment system i s inplace to ensure that investment decisions are based on needs (including considerations o f physical access) and cost-effective or to monitor the safety and adequate use o f high-cost equipment. 145. Overall, some elements of quality assurance and control exist but the system is fragmented and ineffective. Hospitals and, more generally, health facilities are not accountable for the quality o f the services they provide. More broadly, there i s no evidence that the existing information about the performance o f facilities i s compiled, disseminated or effectively used for decision making. 146. The Government can consider revisingthe hospitalpayment system needs and strengtheningstrategicpurchasing.Hospitals are paid by the SPF on a "per case" basis and also receive payment for specific services or procedures provided. There is a consensus that both the list o f cases/services and their associated prices, which in many cases do not accurately reflect the costs, leave room for the hospitals to "play the 63The most frequent illustration ofthat point was that an order ofthe Ministry to close down maternities in which fewer than 300 deliveries ayear take placehas beensuccessfullyresisted. 64International Organizationfor Standardization. 72 system". The hospital financing system should designed to induce hospital to use inputs more efficiently (including using fewer doctors) and to improve performance. The SPF plans to move towards a Diagnostic Related Group Payment system, which in theory generates incentives for hospitals to combine resources more efficiently, and envisages combining it with some element o f global budget to undermine the incentives hospitals to induce a high number o f cases. Rebasing and improving the payment system is a necessary but not sufficient condition to empower the SPF to strategically purchase services. The SPF could also be allowed to selectively contract providers, and/or to not purchase all the services an institution is licensed to produce. 147. Advancing hospital reforms and restoring public trust in the system will requiresustained politicalcommitmentand concertedefforts from the Ministryand the SPF. The hospital restructuring agenda needs to be pursed,with lower level facilities providing only a limited range o f services that respond to the patient needs in a safe environment. The provider payment system would need to be revised and incentives maintained to increase the use o f same-day surgery. A strong push i s also needed to increase the accountability o f hospitals to patients needs and improve the quality o f services. Accreditation could get started as soon as possible and the hospital payment system could reward facilities which have undergone the process. The information already available about various dimensions o f hospital performance (including financial) would need to be collected, analyzed and broadly disseminated, including though the media, at regular intervals and explicitly used for decision making. Additional information could also be collected though exit surveys, citizens' reports cards, etc. Eventually, contracting, payments and the hiring o f hospital managers could be tied to hospital performance. Primary Care and HumanResources 148. Efforts to strengthenprimary care should be accelerated. Currently, about 25% o f physicians work in primary care settings (it i s 50% in France).65 Over the years, the number of physicians formally trained as general practitioners has increased rapidly and they tend to work more frequently in solo-and/or private practices. In 2007, 23% o f patients were registered with a private primary care provider under contract with the State Patient Fund. Overall, however, private practice remains limited: while 23% o f all physicians declare some private activity, only 8% o f them declare that their mainjob i s in private practice.66Primary care facilities receive a risk adjusted capitation per registered patient (which represents 85% o f primary care payments in 2008), fees for preventive services (9%) and a bonus for the achievement o f performance indicators (6%). 149. As a key part o f the strategy to reduce the use o f hospital services, primary care providers should provide more services than they currently do. Primary care providers tend to refer patients to higher level o f care more than they should. Different reasons are 65The number for France is computed as the sume of GPs, pediatricians and gynecologists who work in private practive as a proportion of total working physicians. (data from Conseil National de l'Ordre des Medecins). Dataprovided by the Department of Statistics. 73 invoked: lack incentives to treat patients, equipment, capacity and/or authority to provide more comprehensive services. The State Patient Fundreports that it plans to introduce an element o f performance in the remuneration o f primary care providers. This i s a step in the right direction. Performance could include the coverage o f key preventive services and incentives to limit referrals. 150. A key priority for the mid to long term should be to devise a human resource strategy for the health sector. Currently, the number o f physician per capita in Lithuania is 60% higher than the average inthe new EU countries (Figure 37). Data also suggests that the physicians' skill mix i s not adapted: the pediatrician to child ratio i s 242 (in the USA and France it is around and an obstetrician in Lithuania oversees around 40 births per year which very low. As a result the productivity o f currently employed doctors i s low. And yet their salaries are not tied to performance and have more than doubled since 2004. In addition, there are only 1.8 nurses per physician in Lithuania. A ratio o f 2:l is considered a minimum (the average in OECD countries i s 3:1)68and 4: 1 i s more satisfactory for cost-effective and quality care. 151. Suboptimal number o f doctors, pay level and structure, skills mix, and ratio o f nurses to doctors are due to a number o f factors, including the absence o f selective contracting (managerial autonomy to adjust staff numbers and the skill mix), a hospital financing system that does not induce increased efficiency and performance, the lack financial accountability o f facilities, and a pay structure that is not tied to performance. The Ministry, incollaboration with training institutions, could prepare a human resource development plan to ensure that the health workforce i s better able to respond to the futureneedsof anageing population. Figure 37: Number of physiciansper capita inLithuania and Europe (2007) 450 400 Numberof Physicians 350 per 100000 300 250 2w 150 - 1w 50 0 Source. EuropeanHealthfor all Database http //www euro who.int/HFADB 67Per 100,000 children under 14. Freed et a1 (2004) Relation of per capita income and gross domestic product to the supply and distribution ofpediatriciansinthe United States.Jun;144(6):723-8. http:llfiordiliji.sourceoecd.org/pd~health2007~fiel8107052e-4-3.pdf 74 152. Preliminary evidence suggests that pharmaceutical expenditure is high. Expenditure on pharmaceuticals i s not publishedor readily available, but some evidence suggests that it is relatively high. For instance, in OECD countries the pharmaceutical market represents on average 1% o f GDP,69and inLithuania 1.7%. Although the data for Lithuania is an appr~ximation,~'Figure 38 shows that Lithuania has the highest share health expenditure on drugs across a number o f EU countries. The SPF estimates that Lithuania's expenditure is the highest among the Baltic countries and the amount it reimburses on pharmaceuticals i s growing at fast pace (about 16% between 2006 and 2007).71 Some cost-control policies are in place (drugs are reimbursed based on a reference price, physicians are supposed prescribe generics, etc), but their effectiveness has not been assessed. A detailed analysis o f the pharmaceutical market, pricing and purchasing, reimbursement mechanisms, prescription and retail practices, as well as patients' consumption behavior would be required to identify the underlying causes and to elaborate credible cost-saving options for the public sector. Additionally, a better control o f pharmaceutical expenditure i s likely to directly improve population welfare: both in Latvia and Estonia, out-of-pocket expenditure on drugs has a significant impoverishing impact on household^.^^ Figure38: Share of pharmaceuticalintotal healthexpenditure (circa 2005) 35 30 25 20 15 - 10 5 ." 0 Source. European Healthfor all Database httu./iwww euro who intiHFADB 69http:/lwww.oecd.orgldataoecd/20/60141566367.pdf 70"medical goods dispensedto outpatient" (NHA aggregate) which includes items such as prostheses and medical goods and may not include hospital drugs. 71State Patient Fund: Overview 2007. http:l/www.vlk.lt/vlk/paglfiles/2007~VLK~overview.pdf 72XuK e et al. (2009)Access to health care andthe financial burdenofout-of-pocket healthpayments in Latvia. Technical Brieffor policy makers, WHO http:/lwww.who.int/health~f~ancingldocuments/pb~e~O9~1 -oopslat.pdf 75 6. Education 153. Lithuania spends less than the EU average on education. In 2007, total education spending in Lithuania amounted to 4.9% o f GDP, versus 5.1% for the EU as a whole (Figure 39). Expressed in terms o f average expenditure per student, Lithuania's education spending i s further below the EU mean -amounting to 21.3% o f per-capita GDP versus 30.8% for the EUas a whole.73 Figure39: PublicExpenditureon Education(% of GDP),2006 Source: EUROSTATdatabase Primary and SecondaryEducation 154. Budget outlays for education can be significantly reduced without threatening education outcomes, but it will take time to achieve these budget savings. Since the education sector is intrinsically labor intensive, any significant savings will require sizable reductions in either teacher salaries or teacher numbers. Teacher numbers can be reduced without major disruption o f the teaching process only ifnumbers are reduced at the beginningo f the school year. Even if budget-reduction measures could be put in place by September, 2009, they would yield savings only for the last four months o f 2009. 155. There may be room to reduce teacher salaries but this will face strong political resistance. Average teacher salaries were lower than salaries o f other public- sector employees in 2004. Teacher salaries increased in line with other public sector salaries (except for doctors) and non-manual private sector employees (except in 2008 when teacher salaries increased by 46%) between 2004 and 2008. Average teacher salaries at the end o f 2008 were still below those o f other public-sector employees (particularly doctors) and o f non-manual workers in the private sector (Figure 40), 73EUROSTATdatabase 76 although this difference was reduced by the 10% increase in teacher salaries and the reductions incivil service pay in2009. Since teacher salaries account for the largest share o f the public wage bill (34%), if the Government wishes to undertake any significant reduction inthe wage bill through wage cuts the adjustment would have to involve other public employees, including teachers. Figure40: Real Gross Monthly Earningsby Sector, 2004 and 2008 I 2004 __ I2008 Source: HouseholdLabor-Force Survey Data, 4" Quarter, 2008. 156. Although any form o f salary reduction for teachers would be politically sensitive, some approaches to reduced salary outlays for teachers could be less contentious than an outright reduction in base salary. Latvia, for example, has reduced salary expenditures for teachers by eliminating supplementary payments for some categories o f non-teaching activities, and proposes further expenditure reductions through an increase from 21 hours to 27 hours in the number o f student contact hours that comprise a full teaching load. 157. Reductions already made in the 2009 education budget pose risks for education quality. Reductions o f 57 LTL million (amounting to 1% o f the total education budget) have already been made in2009. The largest element o f this reduction i s a 30 million LTL reduction in the "student basket" capitation transfer to local governments for primary and secondary general education. In many municipalities, this reduction could be absorbed without adverse educational consequences by a judicious reduction in the number o f teachers or a reduction in paid overtime teaching hours. But indications from Vilnius municipality suggest that local governments will instead opt for reducing the already-small amounts provided for non-salary education inputs such as textbooks, teaching materials, in-service teacher training, and internet connectivity. 158. A strength of Lithuania's per-student financing formula is that it includes specific allocations to meet schools' needs for non-salary recurrent inputs to the teaching process that are crucial to teaching effectiveness and learning outcomes. The portion o f the student basket transfer that i s earmarked for these non-salary expenditures was reduced from 6% in 2007 to 3% in 2009 to help make room for the teacher salary increases. The recent cuts inthe student basket and the contraction o f municipal governments' revenues to supplement these resources pose an additional threat to education quality. To ensure 77 adequate financing o f non-salary recurrent inputsto education, it is recommended that the balance o f salary and non-salary elements inthe student basket be restored 6% o f student basket, which could be financed by an offsetting 3% reduction inthe salary portion o f the student basket. 159. M o r e efficient utilization of teachers could yield major budget savings in the medium to long term. Small class sizes and low studentheacher ratios (STR) are an indication o f low efficiency o f teacher utilization in Lithuania. STRs are lower than in most o f the New Member States (NMS) (Figure 41), and much lower than inmost OECD countries. Another perspective on the efficiency o f teacher use i s the elasticity o f teacher supply, the change in teacher numbers in response to changes in enrollment. All o f the N M S (except Estonia) showed greater flexibility than Lithuania in shedding teachers in response to declining primary-school enrollments between 1999 and 2006 (Figure 42). The elasticity o f teacher supply in Lithuania over the period 1999-2006 was 0.54,74 significantly lower than that for Poland (0.87). 160. Where teacher utilization is too low, improved efficiency (in terms o f higher STR) requires an elasticity ofteacher supply greater than 1.O when enrollments decrease, and less than 1.0 when enrollments are increasing. In general secondary education, the number o f teachers in Lithuania fell by 40%, as much as enrollments, between 1999 and 2006 (Figure 43). Inother NMS, general secondary enrollments increased over the same period. Slovakia, Romania, and Latvia improved efficiency by expanding the number o f teachers proportionally less than enrollments, while the Czech Republic and Estonia actually reduced teacher numbers. 161. Demographic change will offer the potential for additional budget savings in the medium term. Negative demographic growth has led to declining school-age populations since the 1970s. This decline will continue for at least the next 15 years. Most o f the decline will occur in the age groups for middle schooling (grades 5 through 10, ages 11 to 16), secondary schooling (grades 11 and 12, ages 17 and 18) and, especially, higher education (bachelors degree, ages 19 through 22) (Figure 44). These cohorts are projected to shrink by more than 200,000 over the next 15 years. This contraction will inevitably lead to reduced enrollments. Smaller enrollments should enable major medium-to-long-term budget savings. Assuming enrollments decline in direct proportion to the shrinkage o f school-age cohorts, teacher numbers decline in direct proportion to the decline in enrollments, and per-student costs remain constant in real terms, total annual budget requirements would be 1.2 billion LTs less infifteen years than they are today. A more likely (and desirable) scenario is that shrinking enrollments will leadto budget savings andeducation quality improvements at all levels. 74I.e., the number of teachers declinedby 54% as much as the decline inenrollments in primary schooling during that period 78 Figure 41: Student-teacher.ratiosfor primary and secondary education (2006) 35.0 30.0 25 0 20 0 15 0 100 5.0 0 0 0Vocati0nal Source: UNESCOfor NMS; others from OECDEducation at a Glance, 2007. Figure 42: YOChange in enrollments and teachers, primary education in NMS,1999-2006 %Ain Teachers D%Ain Enrollments -50.0% -40.0% -30.0% -20.0% -10.0% 0.0% 10.0% Source: UNESCO Institute for Statistics database Figure 43: YOChange in enrollments and teachers, general secondary education, NMS,1999-2006 -20.0 -10.0 0.0 10.0 20.0 30.0 40.0 % A in Teachers 61 % A in EnrollmentsI _____ Source: UNESCO Institute for Statistics database. 79 Figure 44: Projected school-age population by age group (thousands) 300 250 -*-11 200 to 16 150 100 50 I I Source: United Nations Population Divisionwebsite, medium variant projection. 162. T h e preferred approach for reducing education expenditures in general education is to reduce teacher numbers. There are currently too many teachers and too many small classes ingeneral education. Futureshrinkage o f school-age cohorts will tend to make this problem more acute. The introduction o f per-student financing was intended to address this problem. It didnot achieve this result becauseo ftwo fundamentalreasons: The formula provides strong incentives for keeping schools small, which inevitably leads to uneconomically small class sizes.75 In some cases, there i s no practical alternative to small schools, when there are no nearby schools that could be consolidated into a larger, more efficient school. Consolidation i s often not possible or desirable for four-year primary schools in rural areas. But in other cases, consolidation o f small schools is possible and would lead to improvements in efficiency and quality.76 The incentives for small school size and small class size are not the only reasons behind the low studenuteacher ratios in Lithuania. Average school size i s quite highinurbanareas (484 students in2008/09), implying a norm o f 20 students per class under the current financing formula. Even inrural areas, average school size (151 students per school in 2008/09) conforms to a class-size norm o f 15 students per class. Both figures are far above the studenuteacher ratio o f 10.1 in general secondary education and 9.6 for vocational secondary education. 0 The teaching load for teachers is relatively small (18 contact periods o f 50 minutes each per week). This modest teaching load inflates the number o f teachers which are 75For example, the financing formula assumes an average class size ofjust 10 students per class in basic education schools (grades 1-10) with 130 or fewer students. It provides 50% more per student under the formula than schools with 131to 300 studentsandtwice as muchper student as schoolswith more than 300 students. Thus, a basic educationschool with 130 studentswould see its total budget allocation reduced by one-third if it accepted one additional student. The problem of too-small classes i s not common in larger schools, where the financing formula assumes more reasonable average class size of 25 students for basic education schools with more than 600 students and for complete 12-grade schools with more than 700 students. 76Student assessment results in Lithuania show that students in very small schools usually perform at a lower levelthan students inlarger schools. 80 necessary to deliver the approved curriculum and leads to studentheacher ratios which are much smaller than average class sizes.77 And the gap between studentkeacher ratios and average class sizes is larger in Lithuania than the OECD average (Table 19), implying that teaching loads in Lithuania are less than in most OECD countries, or the planning of teacher deployment within schools is less efficient, or both. Lithuania OECD Source: OECD figures from Education at a Glance, 2007; Lithuania S/T ratios from UNESCO Institute of Statistics database, average class sizes are derived from the financing formula norms corresponding to average school sizes for urban and rural areas. Several changes areproposed in theper-studentJinancing formula. Base the adjustment for size-related cost differences on non-discretionary (or exogenous) factors that affect class sizes and unit costs, rather than discretionary (or endogenous) differences such as class size and school size. For example, a cost adjustment based on population density or rural-urban location would reflect the intrinsically higher costs of provision in areas with dispersed population without discouraging consolidation of classes and schools where it makes sense to do so. Keeping unit costs constant (and thus the current student-teacher ratio), adjusting the formula so that reduced student populations are matched by proportional reductions in teachers would generate savings (relative to 2008) of 44 million LT in 2009 (assuming the reform starts with the 2009-10 school year), 275 million L T in 2010, 408 million L T in 2011, and increasing savings thereafter until 2020 when savings peak at 853 (Figure 45). Gradually raise the class size and teaching load assumptions on which the formula is based in order to motivate improved efficiency in teacher deployment, with the long-term goal of raising STR to OECD averages (16.7 for primary schooling and 13.4 for secondary schooling). An annual 1% increase inthe STR over 16 years would produce STR around the OECD average and generate savings on 44 million LT in 2009, 281 million L T in 2010, 421 million L T in 2011, and increasing savings thereafter until 2020 when savings peak at 963 million. 77 Typically, studentkeacher ratios are quite close to average class sizes for the initial four grades o f classroom teaching in primary schooling, where a single teacher usually teaches all subjects to a single class throughout the school day. StudenUteacher ratios are considerably lower than average class sizes in higher grades, where teachers usually teach a single subject and need more preparation and follow-up time than inthe initial grades 81 Allow municipalitiesto spend salary savings for non-salary activitiesto improve teachingand learning. ._l--l.-.-_.._-.- 45: Savings from reducingthe number of.teachers (relativeto ZOOS) (million.LT) Figure ... .... .................... .......... .- .. ..... . .................................................................... l~l^....'...l.~.~.~ - - Constant STR Annual 1% in STR ............................. ................... I IL 2009 2010 2011 2015 2020 2024 __ ------ . ..l-^-.l.lllll_. .__.._l._.__l._.._I__..." . ... ...... ..... ...I__I ....... ................. Source: Staff calculations based on population projections by the Statistical Office and data from the Ministry o Education and Science. Savings in 2009 are annualized. 163. Teachers in multi-grade classes need more preparation and support. For many small schools, multi-grade classes (in which a single teacher teaches a class o f students comprising more than one grade level) are a practical necessity and will continue to play an important role as Lithuania moves toward greater efficiency in education provision. About 2,200 o f the 20,000 primary-school classes in Lithuania currently operate on a multi-grade basis. Most multi-grade classes are in small, rural schools with below-average learning achievement. Student assessmentresults in Lithuania confirm the common perceptionthat multi-grade instruction i s not as effective as conventional single- grade instruction. But international experience confirms that multi-grade instruction can lead to learning outcomes that are as good as or even better than those o f conventional single-grade classes if appropriate support i s provided7' (specialized training and materials to help teachers teach effectively in that environment). Currently, no specialized training or materials are available to teachers. It i s recommended that the Government seeks support from the European Social Fundto develop and deliver teacher training programs and support materials for effective multi-grade teaching. 164. Central financing and management of secondary vocational schools generates significant inefficiencies and disarticulation with general secondary education. Vocational schools are not attracting as many students as gymnasia and general secondary schools, and have higher unit costs and lower staffing efficiency. Central management o f vocational schools generates inefficiency, leads to an undesirable disconnect between general and vocational education, and reduces the impetus and '' See, e.g. Angela W. Little, ed. Education for All and Multigrade Teaching: Challenges and Opportunities, Springer Verlag, 2006, and G. Psacharopoulos, C Rojas, and E. Velez, "Achievement Evaluation of Colombia's Escuela Nueva: I s Multigrade the Answer?", Comparative Education Review, Vol. 37, No. 3, 1993. 82 opportunity to use those resources for maximum benefit o f the community. Municipalities tend to see vocational schools as the responsibility o f others. In order to promote better efficiency and better labor-market outcomes o f vocational schools, it is recommended that management o f general secondary schools and vocational secondary schools be consolidated, and that both types o f schools be managed by municipalities under the same process and subject to the same performance criteria. Municipalities should be free to convert ineffective vocational schools to other uses that better meet local needs, including general secondary education and life-long-learning programs. Higher Education 165. Enrollmentin higher educationis higher than any other Europeancountry, butthere are serious concerns abouthighereducationquality.Symptoms include: expenditure per student i s among the lowest on Europe; 0 many o f the most capable students leave the country to study abroad; 0 indicators o f research output are among the lowest inEurope (including NMS); 0 no Lithuanian university i s among the top-ranked 500 universities inthe world; 0 most recent enrollment growth i s through extramural and part-time students; and 0 supply ofhigher educationplaces exceeds demand. 166. Budget financing in higher education currently supports a designated number of students in specific fields and specific institutions.Selection of fields and institutions for budget support i s made through a non-transparent process reflecting, inter alia, the political connections o f individual deans and rectors. Eligibility for budget- financed places i s based solely on applicants' matura (end-of-secondary) scores. Students inbudget-financed places mustmaintainanaverage score of 8.0 each semesterinorder to retain budget financing. Incoming students with scores below the eligibility threshold and budget-supported students who lose their eligibility due to low performance pay tuition fees that are intended to recover the full cost o f provision. "Extramural" and evening students comprise 46% o f total higher education enrollments. Many o f these students are working while they pursue their studies. Because extramural and evening students pay full fees but do not impose the same demands on university facilities and faculty as regular students, their fees ineffect subsidize regular students who inmost cases do not pay fees. Scholarships and student loans are largely limited to budget-financed students. 167. Because the number o f students supported by the budget i s ultimately an arbitrary decision, large budget savings could in principle be achieved by reducing budget support for university places. However, the Government's discretion in reducing budget support for higher education i s constrained by the Constitutional provision that students with "good" performance have the right to free higher education. A recent ruling o f the Constitutional Court established the Ministry o f Education and Science's right to define what constitutes "good" performance for this purpose. Currently, budget financing covers about half o f total higher education enrollments, but a much higher proportion (about 80%) o f full-time, regular students. Furtherambiguity arises because about half o f all budget-financed students pay nominal fees o f 520 LTs per semester. This practice 83 neither respects the Constitutional principle o f free higher education nor significantly contributes to the actual cost o f provision. 168. This system of budget financing is inefficient because it finances a number of students who have the capacity to pay and inequitable because it is not based on needs. Support is provided to many students and would be pursuing the same programs in the absence o f budget financing. It is arbitrary and non-transparent inthe manner in which it determines the number o f students who are to receive budget support and in selecting specific fields and institutions for support. And it i s inequitable because it finances tuition, scholarships and loans exclusively on the basis o fperformance. 169. The Ministry of Education and Science has proposed a comprehensive reform of higher education that would address most of these deficiencies. Key elements o f the reform are a new legal status which gives universities more autonomy, a new system o f university management by professional Boards, a move from financing places in specific universities to a demand-driven system o f voucher financing for bachelor-level studies, clearer standards for university accreditation, and broader access to scholarships and a new student loan program managed by private banks. These proposed changes would provide important long-term benefits of improved quality and efficiency. Iti s recommended that these proposed reforms be approved and implemented. 170. Significant additional budget savings could be achieved through increased reliance on student fees, broader access to student loans, and targeted budget support. The analysis o f returns to higher education shows that individuals' investments in higher education are justified by their higher lifetime earnings. Thus, an appropriate use o f budget financing is to ensure access o f qualified needy student^.^' Measures to do so would need to be judiciously designed in view of the Constitutional provision on higher education and the recent ruling on that subject by the Constitutional Court. Broader access to student loans and increased limits on student loan amounts would ensure that all qualified student can access higher education. Science and Technology 171. Public spending on science and technology (S&T) institutes is high, but results in terms of innovation are below those of most other countries (Figure 46Error! Reference source not found.). The private sector and the university sector play an unusually small role in research and development (R&D) in Lithuania. Private R&D spending accounts for only 0.2% o f GDP in Lithuania (the EU average is 1.2%). State funding o f S&T development i s equally divided between the Ministry o f Economy and Ministry o f Education and Science (MES). Four-fifths o f the S&T resources (managed by MES) are used to support the core costs o f research institutes, which were inherited from the Soviet period. Only 20% o f budget financing i s used competitively to support specific R&D products. This financing system has important deficiencies: ''Eligibilityforbudgetfinance couldbe basedon needsandmeritby first definingthe eligiblepopulation interms ofneeds, thenselectingstudentsamongthis groupbasedonmerit. 84 0 This supply-driven approach i s inconsistent with Lithuania's demand-driven approach for financing primary and secondary education programs. The 20% competitive element inthe allocation formula provides a weak incentive for improved researchproductivity andrelevance; and 0 By concentrating public financing on research institutes that are separate from universities, it prevents universities from playing a dynamic role in science and technology development-science faculty and graduate students could work on cutting-edge applications related to their fields o f study. It also fails to exploit the economies o f scale universities could offer by using lower-cost staff (graduate students) and facilities incarrying out research. Figure46: Summary InnovationIndex (SII) and per-capitaPublicSpendingonR&D CH DE UK * g esr E t*2 * 9 LU * AT NI I ** ES * * * I L I------- I -_- -----I I I I "T'"- 0 50 Public R&D spending per capita, 2005 PPP US$ 100 150 200 250 300 I * sii-2007 Fittedvalues 1 `ozwce: Alex Ivaschenko, "Lithuania's Research, Development and Innovation System - Benchmarking & EffectivenessAnalysis," World Bank Draft Working Paper, February, 2009. 172. The international best practice is to support all S&T development through a competitive, internationally peer-reviewed process, with the bulk o f the resources usedto support specific research products in areas o f national priority. An example is the competitive R&D model in the United States, which has helped finance the research o f more than 170 Nobel laureates. Some long-term institutional support i s provided under this model. But institutional support is limited to specific research areas of national importance, and institutional contracts are competitively awarded with continued support conditional uponmeeting annual performance benchmarks. 173. The EU-supported "Sunrise Valleys" project, which proposes to finance five regional research institutes, could further deepen the supply-driven model. As the consultant reviews o fthese proposals confirmed, all five proposals focus on infrastructure development and pay too little attention to the role o f private partners. Only two o f the five proposals appeared viable interms o f research content. As in higher education, the 85 question o f whether and how much budget financing should support S&T development i s ultimately an arbitrary decision. Additional budget savings could be achieved by reducing the overall level o f budget support for science and technology development and relying more heavily on commercial sponsorship. 174. I t is recommended that budget financing for S&T development be changed from the current supply-driven approach to a competitive, internationally peer- reviewed approach. This would improve the efficiency and accountability o f public support for S&T and would improve the quality and relevance o f research products. This approach could be used to support solicited contract research on topics o f agreed national priority and high-quality, unsolicited grant proposals. The same approach could support institutional contracts for specified research and technology products. Financing research products based on proposal quality should lead to a consolidation o f research capacity among research institutes and universities. This process would help strengthen capacity in the stronger and more dynamic research institutes and universities. It would also provide an objective mechanism for closing the weaker research institutes and moving to a more modern framework for science and technology development. It is also recommended to reduce the number o f research centers to be supported under "Sunrise Valleys" pro-iect from five to two, and use the remaining funds to supuort competitive research under this approach. 86 MAP SECTION 21°E 22°E 23°E 24°E 25°E 26°E L A T V I A To To Pamoali Pamoali To To To To To To To To Auce Auce Auce Auce Jelgava Jelgava Priekule Priekule To To To To Barbele Barbele LITHUANIA Baltic Mazeikiai Naujoji Bauska Bauska Akmene To To Skuodas Akmene Joniskis Sea Rucava Rucava Vieksniai Seda Venta Musa Likenai Birzai Linkuva Suvainiskis To To 56°N Sventoji Akniste Akniste Darbenai S I A U L I A I Telsiai Kursenai Pasvalys Pandelys 56°N Joniskelis Palanga Pakruojis To To Plunge Kretinga T E L S I A I Virvycia Siauliai PA N E V E Z I O Rokiskis Daugavpils Daugavpils Vabalninkas To To Gargzdai Radviliskis Lev uo Kupiskis Daugavpils Daugavpils Klaipeda Minija Rietavas Uzventis Seduva Subacius Zarasai Panevezys Svedasai Kelme Priekule Kurshskiy Tytuvenai Nevezis U T E N O S Snieckus Anyksciai Zaliv KLAIPEDOS Utena Silale Skaudviles Silute Dubysa Gudziunai Kavarskas Raseiniai Ignalina To Kedainiai Rybachiy Taurage Susve Ukmerge Moletai TA U R A G E S Svencioneliai Pagegia K A U N O Jurbarkas Nemunas Sventoji Svencionys 55°N Jonava 55°N To To Sirvintos To To Pabrade Chernyakhovsk Chernyakhovsk Sakiai Lyntupy yntupy V I L N I A U S Kaunas Kaisiadorys RUSSIAN FEDERATION RUSSIAN FEDERATION Kudirkos Garliava Neris BELARUS BELARUS Naumiestis MARIJAMPOLES Prienai To To Nemunas Ziezmarai Naujoji VILNIUS To To Vilnia Kobylnik Kobylnik LITHUANIA Trakai Chernyakhovsk Chernyakhovsk Vilkaviskis Jieznas To To Smorgan Smorgan Marijampole Juozapines/ Rudiskes Kalnas Merkys (292 m) To To Oshmyany Oshmyany SELECTED CITIES AND TOWNS Kalvarija Sestokai Alytus Jasiuni Daugai COUNTY (APSKRITIS) CAPITALS Salcininkai This map was produced by the Map Design Unit of The World Bank. NATIONAL CAPITAL The boundaries, colors, denominations and any other information To To Lazdijai A L Y TA U S shown on this map do not imply, on the part of The World Bank Suwalki Suwalki RIVERS Varena To To Group, any judgment on the legal status of any territory, or any Merzine Voronovo oronovo endorsement or acceptance of such boundaries. Eisiskes MAIN ROADS To To Augustow Augustow To To 54°N 54°N Druskininkai Radun Radun 0 20 40 60 Kilometers DECEMBER RAILROADS IBRD COUNTY (APSKRITIS) BOUNDARIES POLAND POLAND To To To To Vasilishki asilishki 0 20 40 Miles To To Hrodna Hrodna 33437 2004 INTERNATIONAL BOUNDARIES Hrodna Hrodna 23° E 24°E 25°E 26°E 27°E