Report No. 33400-ECA Public Expenditure Policies in Southeast Europe November 2007 Poverty Reduction and Economic Management Unit Europe and Central Asia Region Document of the World Bank FISCAL YEAR January 1 – December 31 CURRENCY AND EQUIVALENTS (As of November 2, 2007) Albania: Albanian Lek (US$1=LEK 84.42 ) Bosnia and Herzegovina: Convertible marka (US$1= BAM 1.3527) Croatia: Croatian kuna (US$1= HRK 5.1370) FYR Macedonia: Macedonian denar (US$1= 42.4027 Denar) Serbia: Serbian dinar (US$1= RSD 53.5744) Kosovo, Montenegro and Slovenia: euro (US$1 = 0.6922) ACRONYMS AND ABBREVIATIONS BH Bosnia and Herzegovina NMS New Member States of the EU CEB Central Europe and the Baltics NOE Non-Observed Economy CEE Central and Eastern Europe OECD Organisation for Economic Cooperation and Development DRGs Diagnostic Related Groups PAYG Pay-As-You-Go EBF Extra-Budgetary Fund PISA Programme for International Student Assessment EBRD European Bank for Reconstruction and PEIR Public Expenditure and Institutional Review Development ECA Europe and Central Asia PSMAL Public Sector Management Adjustment Loan EPS Electricity Company RS Republika Sprska (an entity of BH) EU European Union SEE Southeast Europe EU8 The countries that acceded to the EU in SEEREM Southeast Europe Regional Energy Market 2008 (excluding Cyprus and Malta) EU7 The EU8 excluding Slovenia SDR Standardized Death Rate FBH Federation of Bosnia and Herzegovina SFRY Socialist federal Republic of Yugoslavia (an entity of BH) FDI Foreign Direct Investment SME Small and Medium Enterprises GDP Gross Domestic Product TB Tuberculosis GFS Government Finance Statistics USD US Dollar IBRD International Bank for Reconstruction VAT Value Added Tax and Development IDA International Development Association NBY National Bank of Yugoslavia Vice President: Shigeo Katsu (ECAVP) Country Director: Orsalia Kalantzopoulos (ECCU4) Anand Seth (ECCU5) Sector Director: Cheryl W. Gray (ECSPE) Sector Manager: Bernard Funck (ECSPE) Task Manager: Satu Kähkönen (ECSPE) Ivailo Izvorski (ECSPE) ACKNOWLEDGEMENTS This study was prepared, under the direction of Bernard Funck, by a team lead by Satu Kähkönen and Ivailo Izvorski. The responsibilities of the team members were as follows: Satu Kähkönen, Ivailo Izvorski and Sanja Madzarevic-Sujster (public spending); Anita Schwarz and Snjezana Plevko (pensions); Snjezana Plevko and Anna-Marie Vilamovska (social protection); Sarbani Chakraborty and Caryn Bredenkamp, with Michele Gragnolatti (health); Peter Darvas, Mamta Murthi and Toby Linden (education); Tony Verheijen (public service) and Ivailo Izvorski (infrastructure). Juan Carlos Ginarte contributed to the analysis of the size of the public sector. Mismake Galatis was responsible for processing of the document. World Bank country economists for the SEE countries helped gather the data. Kyle Peters and Bill Dorotinsky were peer reviewers. Hermann von Gersdorff, Tracey Lane, Erjon Luci, Richard Martin Humphreys, Ali Mansoor, Felix Martin, Alia Moubayed, Evgenij Najdov, Gary Reid, Kanthan Shankar and Danijela Vukajlovic-Grba provided helpful comments on the report. TABLE OF CONTENTS ACKNOWLEDGEMENTS EXECUTIVE SUMMARY......................................................................................................................... I 1. FISCAL DEVELOPMENTS AND GOVERNMENT SPENDING ...............................................1 I. Introduction...................................................................................................................................1 II. Government Revenues ..................................................................................................................5 III. Level And Trend of General Government Expenditures ..............................................................5 IV. Expenditure Composition .............................................................................................................8 A. Economic Classification..........................................................................................................8 B. Functional Classification.......................................................................................................10 V. Lessons Learned from the EU Members.....................................................................................12 VI. Conclusion ..................................................................................................................................13 2. ENSURING SOCIAL WELFARE: SOCIAL PROTECTION ....................................................15 I. Public Social Protection Systems................................................................................................15 A. Social Protection Expenditures .............................................................................................16 B. Social Protection Revenues ...................................................................................................17 II. Reforming the Pension System ...................................................................................................17 A. Key Issues .............................................................................................................................18 B. Pension System Design .........................................................................................................23 C. Looking to the Future............................................................................................................24 III. Reforming Social Welfare Benefits ............................................................................................27 IV. Reforming Labor-Market Support Programs ..............................................................................28 V. Conclusions and Recommendations ...........................................................................................29 3. HEALTH ...........................................................................................................................................31 I. The Institutional Setup................................................................................................................31 II. Health Expenditures....................................................................................................................32 III. Health Outcomes and Utilization of Healthcare .........................................................................34 IV. Health Financing and Expenditures Policies...............................................................................36 A. Health Financing Policies .....................................................................................................36 B. Expenditure Policies..............................................................................................................37 V. Conclusions and Recommendations ...........................................................................................41 4. EDUCATION....................................................................................................................................43 I. Education Expenditures and Sources of Financing.....................................................................43 II. Education Outcomes ...................................................................................................................44 III. Institutional and Policy Context..................................................................................................47 A. Challenges Related to the Content, Quality and Structure of education ...............................47 B. Challenges in Education Administration and Management ..................................................48 C. Budgeting and Financing Arrangements ...............................................................................49 IV. Efficiency issues .........................................................................................................................50 V. Equity Issues ...............................................................................................................................52 VI. Conclusions and Recommendations ...........................................................................................53 5. SIZE, STRUCTURE, AND REMUNERATION OF THE PUBLIC SERVICE ........................57 I. Comparative Analysis of Public Sector Wage Bills ...................................................................57 II. Trends In Public Sector Staffing And Wage Levels ...................................................................58 A. Public Sector Employment....................................................................................................58 B. Wage Levels..........................................................................................................................59 C. Sector and Local Government Wage Bills ............................................................................60 III. Wage Setting Mechanism and Pay Structure..............................................................................61 A. Compression Ratios ..............................................................................................................63 B. Politicization..........................................................................................................................63 IV. Ongoing Reforms........................................................................................................................63 A. Rationalization of the Pay System ........................................................................................64 B. Introduction of Performance-Based Elements to Pay ...........................................................65 C. Decompression and Depoliticization.....................................................................................65 D. Staff Reduction .....................................................................................................................66 E. Improving Public Sector Effectiveness through Administrative Reforms ............................66 V. Conclusions and Recommendations ...........................................................................................67 6. INFRASTRUCTURE: FISCAL SPACE AND PUBLIC INESTMENT MANAGEMENT ......69 I. Infrastructure Gaps......................................................................................................................70 II. Macroeconomic Considerations..................................................................................................71 III. Fiscal Space ................................................................................................................................74 IV. Public Investment Planning ........................................................................................................76 V. Conclusions and Recommendations ...........................................................................................78 Tables Table 1.1: But Income Levels in Most SEE Countries Are Low .................................................................1 Table 1.2: The Countries of Southeast Europe: Relations with the EU .......................................................3 Table 1.3: SEE, Decomposing Fiscal Adjustment, 2000-05\ .......................................................................3 Table 1.4: General Government Debt ..........................................................................................................4 Table 1.5: Fiscal Developments in SEE, 1997-2006....................................................................................6 Table 1.6: General Government Interest Payments and Primary Expenditures, 1997-2006........................7 Table 1.7: General Government Expenditures, Economic Classification ....................................................9 Table 1.8: General Government Expenditures, Functional Classification .................................................10 Table 1.9: Expenditure Adjustments in Select EU Countries in 1990s......................................................13 Table 2.1: Public Revenues and Expenditures on Social Protection..........................................................16 Table 2.2: Public Expenditures on Social Protection: A Breakdown, 2005...............................................17 Table 2.3: Payroll Contribution Rates in SEE and Selected CEE Countries, 2006 ...................................18 Table 2.4: Basic Features of Reformed PAYG Systems in SEE................................................................21 Table 2.5: Second Pillar Introduction.........................................................................................................23 Table 3.1: Sources of Healthcare Financing, 2005.....................................................................................32 Table 3.2: Health Expenditure in SEE, 2004 .............................................................................................33 Table 3.3: Health Spending as Percentage of Total Government Expenditures ........................................33 Table 3.4: Health Outcomes in the SEE Countries, 1992-2005 .................................................................35 Table 3.5: Social Insurance Contribution Rates for Health........................................................................37 Table 3.6: Health Service Delivery Data from SEE countries, 2005 .........................................................38 Table 3.7: Health Insurance Fund Expenditures by Level of Care (percentage), 2002..............................38 Table 3.8: Paying Health Care Providers ...................................................................................................39 Table 4.1: Public Spending on Education, 2004 ........................................................................................44 Table 4.2: PISA Reading Scores: Albania, Macedonia and Comparators .................................................45 Table 4.3: Enrollment Rates in SEE and Selected Comparators, 2004......................................................45 Table 4.4: Public Unit Costs by Level of Education ..................................................................................50 Table 4.5: Number of Students in Basic Education ...................................................................................51 Table 4.6: Student Teacher Ratios in SEE Countries and Selected Comparators......................................51 Table 4.7: Population Projections, 5-19 year olds (In thousands)..............................................................52 Table 4.8: Participation Rate in Education in Macedonia by Ethnic Group ..............................................53 Table 5.1: Public Sector Wage Bill ............................................................................................................58 Table 5.2: Public Sector Employment........................................................................................................59 Table 5.3: BH, Croatia and Serbia and Montenegro: Total Net Compensation, 2004...............................60 Table 5.4: A Breakdown of Wage Expenditures, 2004..............................................................................60 Table 5.5: Civil Service Salary Compression Ratios .................................................................................63 Table 6.1: Density of Road Infrastructure (latest observation available, 1997-2003)................................70 Table 6.2: SEE: Fixed Investment and Government Capital Spending, 2000-2006 ..................................72 Figures Figure 1.1: Convergence with the EU is ......................................................................................................1 Figure 1.2: SEE Macroeconomic Developments .........................................................................................2 Figure 1.3: Primary Expenditure and Income per Capita.............................................................................8 Figure 1.4: Actual and Predicted Levels of General Government Expenditures .........................................8 Figure 1.5: Public Sector Wages and Salaries..............................................................................................9 Figure 1.6: Public Administration Performance...........................................................................................9 Figure 1.7: NMS Fiscal Adjustment, 2000-05 ...........................................................................................12 Figure 2.1: SEE Population Aged 65 Years of Older.................................................................................19 Figure 2.2: OECD Population Aged 65 Years of Older.............................................................................19 Figure 2.3: Dependency Rates....................................................................................................................19 Figure 2.4: Old Age Dependency Ratios, 2005-2025 .................................................................................19 Figure 2.5: Contribution Rates ...................................................................................................................20 Figure 2.6: Share of Pensioners that Retired Under the Age of 65 ............................................................20 Figure 2.7: Percent of the Elderly Receiving Pensions Now and in the Future .........................................22 Figure 4.1: Student Performance in Reading and GDP per Capita, PISA 2000.........................................44 Figure 6.1: Road Network Quality .............................................................................................................70 Figure 6.2: Problems Doing Business, SEE and Comparisons, 2005 .........................................................71 Boxes Box 3.1: Improving Governance in the Health Sector in Macedonia ........................................................40 Box 5.1: Politicization of Civil Service......................................................................................................64 Box 5.2: Public Administration Reform in Montenegro ............................................................................67 EXECUTIVE SUMMARY 1. The countries of Southeastern Europe (SEE) and the province of Kosovo (thereafter, Kosovo) have undergone a significant, though difficult, transition over the past decade.1 Series of conflicts in the region have complicated the transition process, which in the initial phase focused on macroeconomic stabilization and reconstruction. Helped by macroeconomic stability and efforts in advancing structural reforms, real GDP growth in region has picked up this century, averaging 5.8 percent during 2000-2006, although it has still lagged the pace in both the more advanced reformers and the average for emerging Asia. In addition to boosting living standards, the SEE countries’ shared aspiration to advance EU integration has been increasingly shaping the reform agenda. 2. Fiscal adjustment has been an integral part of the broader progress in macroeconomic stabilization and advancement in structural reforms to date. Expenditure cuts have helped reduce fiscal deficits and trim spending relative to GDP in most countries in the region, helped by introductions of treasury systems, enactment of procurement and audit laws and other fiscal reforms. Reflecting expenditure cuts, the size of government declined in all countries except in Macedonia, Serbia and the province of Kosovo this decade (in Serbia due to the shift of revenue and spending assignments from the federal level of government of Serbia and Montenegro to the government of Serbia, and in Kosovo reflecting the surge in donor-financed spending and formalization of off-budget outlays). On average, revenues rose modestly relative to GDP in the SEE (except in Serbia and Kosovo), as sizable reductions in direct tax rates were more than offset by offset improved collection efforts, hikes in excise tax rates and the introduction of a VAT in most countries. Fiscal deficits fell everywhere except in Serbia and Kosovo. 3. Progress has been substantial in fiscal consolidation, but in several of the countries in the region (notably, Bosnia and Herzegovina and Croatia) the government’s presence in the economy remains oversized, larger than among the new member states of the EU. Outcomes, by contrast, remain poor than warranted by the level of spending. At the same time, costs related to advancing EU integration, the need to complete reforms in key sectors and the imperative to tackle the issues of ageing and rapidly declining populations in most countries, are expected to generate further pressures for public spending. Creating the fiscal space for addressing such pressures would require a further reduction in existing spending, given little scope (or wisdom, for that matter) for increasing tax rates. Moreover, most countries labor under onerous payroll contribution rates that need to be cut to curb and then reverse the pervasive informalization of economic activity and employment. 4. All told, reductions in existing spending are needed to lower the government burden on the economy, tackle future spending pressures and help reduce payroll contribution rates. Spending reductions need to be carried out within a framework of expenditure restructuring, that would boost the efficiency of spending, help ensure that fiscal policy supports economic growth and living standards, while addressing issues of equity and access to public services. 5. This report reviews the level, composition and outcomes of government spending and distills some of the lessons that emerge from efforts by the SEE countries in reforming expenditure policies. The 1 This report focuses on Albania, Bosnia and Herzegovina, Croatia, FYR Macedonia, Montenegro, Serbia and the province of Kosovo. Kosovo is currently under the administration of the United Nations Interim Administration Mission in Kosovo (UNMIK) according to the terms of UN Security Council Resolution 1244 (1999). On November 10, 2005, the UN Security Council endorsed the proposal of the UN Secretary General to appoint former Finnish president Martti Ahtisaari as his special envoy for the future status process for Kosovo. Given that this process is ongoing, this report does not pre-judge any aspect of its outcome, and the analysis and recommendations contained herein are explicitly contingent on this process. i report identifies key remaining challenges and proposes a broad menu of options in further reforms of sectors that account for the largest shares of public spending across the countries, and where reforms are likely to have significant budgetary implications. This has motivated the choice of sectors discussed in the report: social protection (including pensions), health, education, public administration and infrastructure. Social Protection 6. Generous social protection systems in the SEE countries are a key legacy of the past. In the Socialist Federal Republic of Yugoslavia (SFRY) and in Albania, similarly to the other countries of central and eastern Europe, there was nearly universal coverage by social insurance (for health, old-age and disability), and eligibility for social benefits and unemployment relief. Nearly universal social protection coverage was accompanied with nearly universal formal employment, resulting in the choice of payroll contributions to finance the insurance schemes. These ”cradle-to-grave” protection models quickly proved fiscally unsustainable at the start of transition as formal sector employment plummeted due to the collapse of linkages between the constituent countries and enterprises both between and within them, the start of privatization, and a pickup in informal activity. Wars and ethnic conflicts added to the economic dislocation of transition and led to increasing demands for social protection spending to accommodate the social and economic re-integration of war veterans, invalids, refugees and displaced persons. 7. Governments of the SEE countries adopted different strategies to halt the sharp rise of social protection spending. These included accumulation of deficits, sharp divergence between legally mandated benefits and actual payments, and the reduction of the scope and size of benefits. Efforts thus far have borne fruit, helping limit government deficits, but further adjustment and profound change to social protection systems would be needed in most countries to assure the efficiency of spending, reduce the government burden on the economy while improving equity, and assure fiscal sustainability. 8. With pensions typically accounting for the largest share of social protection outlays, pension reform has been a key to sustainability of public finances. Parametric reforms of the pay-as-you go systems so far have helped improve sustainability, but difficult issues remain to be addressed, related to low coverage, high and rising ratios of retirees to contributors, and relatively generous pensions. Looking forward, a more comprehensive reform of pensions, accompanied by a redesign of the overall system of financing through payroll contributions, would be needed to improve their performance and fiscal sustainability of programs. Key measures on the revenue side include the need to improve compliance and collection of social contributions, both through improvements in record-keeping and the administration of the collection processes and through intensified enforcement efforts. Additional specific measures to consider include the need to raise the retirement age (an increase to 65 years is either accomplished or on its way to be completed in most countries except Macedonia) and equalize the age for men and women (only in Kosovo and in the Federation entity in BH are the ages equalized). A move toward inflation indexation is also well advised (Albania and in Kosovo for the basic pension indexation is fully to inflation, and Serbia is moving toward full inflation indexation starting in 2009). 9. On the expenditure side, there is significant room to improve cost-effectiveness of social protection programs. Coordination, targeting, eligibility conditions, and finance allocation for existing social welfare programs could be improved to streamline and reduce expenditures. This could be done for example by reducing administrative expenditures of programs. Coordination and consolidation of programs and reallocation of funds between them would also be needed to reduce fragmentation, overlaps, and waste of resources. 10. Efforts on the revenue and expenditure side for the PAYGO pillar should be supplemented by efforts on the regulation of the funded pillars where these have been introduced. In particular, regulators ii should assess whether fees levied on second-pillar funds are not too high and work with fund managers to bring them down. The authorities in Macedonia are moving in the direction of reducing second-pillar fees with a view to boosting return. Health 11. Public spending on health relative to GDP is higher in SEE (except Albania) compared to other countries at similar income levels, while outcomes are poorer. On average, overall health spending (both public and private) exceeds even the EU15 average. With the exception of Albania and Montenegro that spend as much the EU15 (6.8 percent of GDP), all other SEE countries spends substantially more at about 8 percent. In Albania, private spending is disproportionately high within overall health expenditures. Utilization of primary and preventive healthcare is typically poor, as are inpatient admission rates, the latter with the result that available resource, including hospitals, beds and staff are under-utilized. These outcomes owe much to the institutional setup, in most countries inherited from the former Yugoslavia, including the health insurance and financing arrangements and the delivery of care. Challenges are set to multiply in the coming years, necessitating determined reform efforts. 12. To contain the already high public health spending and health fund arrears, SEE health financing systems would need to be made more efficient and provide incentives for cost control. Given the already high levels of payroll contributions for health, there is no scope to increase them further; indeed, the governments in the region would be well advised to ponder options to bring contribution rates down. In Albania, general government revenues finance three-fourths of public health spending despite the existence of a mandatory contribution scheme, the most extreme example in the region of the effect of the pervasive informalization of the economy on public finances. 13. Hence, a key avenue of reforms in SEE (except in Kosovo, where a contributory scheme does not exist) focuses on strengthening contributions collection, reducing evasion and improving cost recovery. SEE experience suggests that measures in the following four areas are central to increasing cost recovery. Firstly, there would be a need to improve administrative efficiency in contribution collection. Secondly, to enhance collections, there would be a need to reduce contribution exemptions. Thirdly, co-payments for health services would warrant expansion, while ensuring that the vulnerable groups (the poor, people with certain chronic conditions) are protected. International experience indicates that co-payments can help reduce unnecessary use of services, and increase the use of primary health care while reducing hospital care. Fourthly, it would be important to ensure that voluntary supplemental health insurance is provided only by the private sector or does not cover co-payments. Both in Croatia and Slovenia, the health insurance fund introduced a voluntary supplemental insurance a few years ago to cover co- payments. However, this move re-introduced the problem of excessive health service use, which co- payments were to mitigate in the first place, and adverse selection, since the insurance was purchased by beneficiaries with highest medical costs. The Slovenian government is now looking for ways to reform the system, while the Serbian government, learning from its neighbors’ experience, has excluded the supplemental insurance coverage of co-payments. 14. A second key avenue for reforms would continue to focus on health expenditures. Measures introduced that need to be deepened include the need to enhance the efficiency of health spending, including by introduction payment mechanisms that provide incentives for cost control at hospitals such as Diagnostic Related Groups (DRGs). Efforts in this area have been advanced in Albania and Serbia. Since the pre-requisite for the proper functioning of the system is good hospital and cost data, the first step in the reform process in other SEE countries would need to address that. 15. Further, rationalization of generous health benefit packages would also be warranted in SEE. In Serbia, the 2005 Health Insurance Law removed a number of procedures, including plastic surgery, form iii the benefits package as a first step toward a more affordable set of essential services. The government in Macedonia is also embarking on a reform of the benefits package. While this is a politically difficult exercise, it could be launched by removing or reducing the scope of non-health benefits. 16. Finally, revamping the system of pharmaceutical procurement will be a key to curbing outlays on drugs, another key cost driver. Options include purchasing drugs through international competitive bids (ICB), or at least introducing reference pricing or strengthening of the existing reference pricing system, monitoring and regulating the positive list of drugs covered by health insurance funds, encouraging the use of generics, and streamlining of the system of drug registration and marketing authorization. Reforms along these lines have been introduced in several countries in the region. In Serbia, a new Law on Medicinal Products mandated the establishment of an independent Drug Agency, which together with harmonization of licensing practices with those in the EU and the introduction of new operational guidelines for pharmaceutical procurement, has helped reduce drug prices substantially. The government in Macedonia is carrying out its second ICB, and improved management of the process has more than doubled the number of firms that have pre-qualified compared with a cancelled earlier round. The authorities in Albania, Croatia and Macedonia have moved to tighten the positive list of drugs. Education 17. Education indicators in most SEE countries suggest that, on average, the stock and quality of human capital, a key determinant of economic growth and living standards, lags most counties in the EU, including the new member states. Results from the Programme for International Student Assessment (PISA), Albania, Macedonia and Serbia and Montenegro (the only countries in the region covered by the assessments in 2001 for the first two and 2003 for Serbia and Montenegro) scored well-below most other countries, with roughly 90 percent of the students scoring below level 3 (compared with 67 percent in Bulgaria and 47 percent in Poland). In terms of enrollment, with the exception of Albania at the primary level, enrollment rates in the SEE lag the average for Eastern Europe and Central Asia and are well-below the rates among the NMS. Despite progress in some countries over the last several years, notably in educational attainment and enrollment, these results confirm the need for deeper reforms to improve education systems and enable students to graduate prepared for the challenges of today’s globalize and increasingly knowledge-driven economies. Persistent problems and new challenges that should be tackled include the need to ensure equitable access, improve overall coverage, and strengthen the relevance of non-compulsory education. 18. The education systems in SEE harbor structural inconsistencies and inefficiencies caused by inadequate response by education management to emerging social and economic challenges. Most importantly, systems have not adjusted to demographic changes (including shrinking student-age population), migration, and evolving ethnic structures. Further, rigid allocation schemes and ill-defined (or often missing) decentralization frameworks create disparities among schools and localities and offer perverse incentives to continue inefficient service delivery. In addition to deficiencies in primary and secondary education, the quality and performance of higher education is sharply declining as a result of fragmented administration and muddled financing. 19. The persistence of these problems often stems from lack of policy and planning to adjust outdated content and structure. Averaging 4.5 percent of GDP, public expenditures on education in SEE are higher than both the amounts spent by Chile, Thailand and Korea at present (emerging markets with higher GDP per capita) and by Japan in the 1970s, indicating more money is not necessarily the answer. In most other countries, the efficiency of resource use would need to be increased by education policy changes, adjustments in management and administration, and expansion of private financing (fees and involvement of the private sector). Expectations that SEE countries improve their education systems to accelerate convergence to EU standards reinforce these pressures to reform. iv 20. Given the centrality of education to achieving higher growth rates in real GDP and higher standards of living, the governments of the SEE countries are invited to reflect on a series of recommendations. These include the need to improve accountability by, inter alia, giving sub-national governments authority over reducing or increasing teaching staff and initiating restructuring efficiency improvements and restructuring of schools. Such a shift in authority should be accompanied by a shift to capitation financing (financing based primarily on the number of students) and targeted improvement in quality and relevance of curricula (including by setting standards at the central government level). The governments will be well advised to address equity and disparity concerns, both within the levels of education, ethnic groups of geographic regions. Public Wages and Public Administration 21. Overall government wage outlays in the SEE are larger than among the EU15 on average, and for some countries in the region by a large margin. With wages and salaries the single largest non- discretionary item in countries’ budgets and with quality of public services crucially dependent on the quality of the public administration, rationalizing public wages is a key fiscal concern. Employment and pay reform would also be essential to help align public sector institutions to EU norms and facilitate policy coherence. High public sector wage bills are due to a large extent to excessive employment in police, defense, education and healthcare. Only in BH, excessive wage bills also reflect higher wage rates for lower-grade employees at the state level and in the Federation. In addition, wage bills reflect structural characteristics of the labor market, including rigid labor markets and inflexible bargaining procedures. 22. To rationalize the wage bill, some SEE countries have embarked on a reform of civil service and public sector employment and wage policies. Demobilization is ongoing, in particular in BH. Measurable progress has been made in Serbia in the rationalization of civil service pay systems and decompressing salaries. A phased salary decompression has also been launched in Macedonia and Montenegro, albeit at a slower pace, given the overall fiscal envelope. Reforms in Serbia have included streamlining of allowances, which in most SEE countries constitute a significant share of total pay and contradict the EU principles of equal pay for equal work and transparency of pay systems. Almost all countries have established legal bases for development of performance appraisal system, which is essential to create incentives for performance. However, implementation would need to follow, and the linkage between the pay and performance forged. These reforms have been accompanied to a varying degree by efforts to depoliticize the civil service. In most SEE counties, top civil service posts previously held by career officials have been politicized, which is impeding the development of stable and professional civil service by reducing incentives for talented staff to remain in the service. A number of countries, including BH, Montenegro, Macedonia, Albania, Croatia, and Serbia have lately taken steps towards depoliticization of the civil service, starting with legislative changes. 23. Looking forward, these reforms need to be brought to completion to enable the authorities to attract, develop and retain a professional civil service without overburdening the taxpayers. Rationalization of pay determination, decompression of salaries and depoliticization of the civil service need to be accompanied by determined efforts to reduce public employment. A careful review of organizational structures of core ministries and agencies would be warranted to identify possibilities for reallocation and rationalization of staff, while phasing out entities engaged in market and commercial functions to reduce/contain the size of the state administration. So far, for political reasons, instead of targeted reductions, SEE governments have mostly resorted to across the board staff cuts, which have tended to have limited impact are, most likely, more harmful than useful to the quality and coherence of state administration. v Infrastructure: Fiscal Space and Public Investment Management 24. Following recent sizable investments in public infrastructure by Croatia and Slovenia, the other countries of SEE have started to plan for large outlays on infrastructure. The priority SEE governments have awarded to building new motorways, railroads and energy generating capacity is motivated by perceptions about the poor state of infrastructure and efforts to accelerate convergence to Europe. Strengthened fiscal positions and more favorable international environment, recent market turmoil notwithstanding, have also given support to such plans. Active involvement by multilateral and bilateral institutions in the planning and sometimes the execution of some of these plans has also been a factor, while private foreign investors are yet to become more actively involved. 25. This short chapter lays out some of the key issues to consider when planning and evaluating new investments in infrastructure. Public provision of infrastructure depends crucially on finding budgetary room for new spending without jeopardizing fiscal sustainability or macroeconomic stability (“fiscal space”). The scope for making fiscal space available for additional infrastructure spending by relaxing substantially deficit targets or boosting substantially the level of government spending are rather limited in most but not all SEE countries, given the large size of government in most and high levels of government and external debt in some. Outside a few special cases that may, potentially, sustain somewhat larger deficits, fiscal room for additional infrastructure investment can be only found in reducing or reprioritizing existing expenditures or involving the private sector. Along with these considerations, it is important for governments to reassess whether infrastructure is the most binding constraint to development and whether, given the stage of development, efforts should not also target other areas. 26. Any potential new borrowing, moreover, needs to be undertaken only after careful debt sustainability analysis. Addressing the absorptive capacities of the governments to handle new projects also needs to be enhanced. Given the small size of SEE countries, cross-country cooperation appears crucial for almost all projects, notably motorways, railroads and electricity generating capacity. 27. Private sector participation in infrastructure is important, and it can lessen the direct fiscal burden. However, this does not reduce the need to have a high quality project within a strong sectoral framework. Failing this, the implicit contingent liabilities of the government in partnering with the private sector are likely to translate into actual expenditure and bailouts of unsatisfactory projects. This means that the suggested caution on financing new government expenditure for investment applies equally to extending guarantees, implicit or explicit, for encouraging private sector participation. vi Fiscal Developments and Government Spending I. INTRODUCTION 1.1 The countries of Southeast Europe (SEE) have experienced a broad economic recovery since the end of the regional conflicts last century.2 Incomes have risen in most countries, but with the exception of Albania, Slovenia and Croatia, levels of GDP per capita have yet to reach the levels of the early 1990s. Recovery in Slovenia, one of the EU new member states (NMS), has been particularly rapid after a shallow transition recession, helping Slovenia maintain by far the highest income per capita among both the SEE and the NMS countries (equivalent to 90 percent of the EU average in PPP in 2006 rising from 80 percent in 2004)3 (Figure 1.1). Croatia’s per capita income is also high, but the other SEE countries are substantially poorer (Table 1.1). Figure 0.1: Convergence with the EU is Table 0.1: But Income Levels in Most advancing … SEE Countries Are Low GDP per capita at PPS, EU25=100 Atlas GNI per capita, 2006 (Slovenia=100) 120 NMS 51.2 100 1999 2002 Croatia 49.4 80 2004 Serbia 20.7 60 Montenegro 20.4 40 BH 18.0 20 Albania 15.7 0 Macedonia 16.2 EU15 EU7 Croatia Slovenia Source: World Bank GDI, 2006. Note: EU7 = EU8 excluding Slovenia. 1.2 Real GDP growth among the SEE countries averaged 5 percent a year over the last five years, but has lagged the more advanced reformers in the Baltics and had trailed the 6¼ percent pace of expansion in Southeast Asia on average (Figure 1.2). Supported by a surge in bank credit, large remittances and accelerating capital spending on imported machinery and equipment, current account deficits widened substantially from 2000 through 2006. External imbalances also reflect a relatively low level of exports and pose policy challenges. Inflows of foreign direct investment have risen strongly, but have not covered fully current account deficits, increasing reliance on foreign borrowing. With most SEE countries having graduated from World Bank IDA financing (except BH and Kosovo), concessional credits are available largely from the EBRD and the EIB – and these will have a substantially smaller concessional element than IDA credits. Increasingly larger shares of borrowing from abroad are accounted for by loans from foreign banks to their SEE subsidiaries, in turn financing rapid domestic credit expansion that has added to imports and helped widen external deficits. 2 This study focuses on Albania, Bosnia and Herzegovina, Croatia, FYR Macedonia, Serbia and Montenegro. The province of Kosovo is included in the study subject to data availability. Slovenia is included in some of the discussions for comparison. 3 NMS includes the eight countries that joined the EU in 2004 (Czech Republic, Estonia, Latvia, Lithuania, Poland, Hungary, Slovak Republic, and Slovenia, also known as the EU8), Bulgaria and Romania. 1 Figure 0.2: SEE Macroeconomic Developments Real GDP growth has been robust, but The share of the private sector is rising, but is still unspectacular (2000-06 average, in percent) below comparator countries (share in GDP) 9 80 8 70 7 60 6 50 5 40 4 30 3 20 2 10 1 0 0 SaM BH Cro fMA Rom Bul Alb Svk Hun Est SEE Baltics E Asia S Asia 2002 2005 Source: EBRD Transition Reports Inflation has remained moderate, despite Current account deficits have widened one-off factors (in percent) substantially (in percent of GDP) … 14 30 12 25 10 20 8 15 6 4 10 2 5 0 0 Alb BH BG Cro fYRM Rom Serb Mont Tur Alb BH BG Cro fYRM Rom Serb Mont Tur 2005 2006 2000 2006 … reflecting low levels of exports (in percent of FDI has often not covered fully external GDP) deficits (2006, in percent of GDP) 35 80 30 70 25 M e rc h a n d is e e x p o rts , % G D P 60 20 50 15 40 30 10 20 5 10 0 0 Alb BH BG Cro fYRM Rom Serb Mont Tur Alb Tur SaM BH Rom fMA Cro BG Hun Svk Ire CAD FDI Sources: National authorities, IMF, EBRD and World Bank. 1.3 EU integration has been a major driving force of reforms in the region since the turn of the century. Similarly to the disparity in incomes per capita, there is disparity in terms of the pace of EU integration, with Slovenia an EU member since 2004 (and a member of the eurozone from the start of 2007), Croatia advanced in its accession negotiations, and Macedonia is an EU candidate (Table 1.2). Albania signed a Stabilization and Association Agreement (SAA) with the EU in 2006, Montenegro initialed the agreement in early 2007, and Serbia is in the process of negotiations. Bosnia and Herzegovina (BH) has finalized the technical negotiations but has yet to complete a number of political 2 requirements. The SAA provides for closer ties with the EU through gradual implementation of a free trade area and advancing reforms to help speed the ultimate adoption of EU standards.4 Table 0.2: The Countries of Southeast Europe: Relations with the EU Albania Singed a SAA on June 12, 2006. Technical negotiations on the SAA completed. Agreement on police reform, enactment of Bosnia and a law on public broadcasting and full cooperation with the International Crime Tribunal for Herzegovina the former Yugoslavia (ICTY) needed to sign the SAA. Croatia SAA in force from February 1, 2005. Accession negotiations started on October 3, 2005. Macedonia SAA in force from April 21, 2004. EU candidate country as of mid-December 2005. Montenegro SAA initialed on March 15, 2007. Serbia Negotiations on SAA ongoing. Slovenia EU member from May 1, 2004. Source: The European Commission. 1.4 The size of government declined in all countries this century except in Macedonia and Serbia, and the province of Kosovo (in Serbia reflecting the shifting of spending by the federal state of Serbia and Montenegro to the government of Serbia, and in Kosovo due to the formalization of spending that was off-budget in earlier years and a surge in outlays financed by foreign assistance). Expenditure reductions equivalent to 0.6 percent of GDP a year on average helped reduce the average fiscal deficit in SEE to 1.6 percent of GDP on average in 2005 from 4.2 percent in 2000 (Table 1.3). Average general government expenditures fell to 44.1 percent of GDP in 2005 from 47.4 percent in 2000, a substantial reduction. Revenues fell by 0.7 percent of GDP on average over the same period, as sizable reductions in direct tax rates offset the positive effects on revenues from improved collection efforts, the introduction of a VAT in most countries this century and increases in excise taxes. Table 0.3: SEE, Decomposing Fiscal Adjustment, 2000-05 (In percent of GDP, SEE averages, weighted by nominal GDP) change, 2000 2001 2002 2003 2004 2005 2000-2005 Revenues 43.2 42.8 44 43.5 43.5 42.5 -0.7 Expenditures 47.4 47.3 47.3 46.6 45.8 44.1 -3.3 Primary Expenditures 45.6 45.6 45.5 44.8 44 42.4 -3.2 Interest Payments 1.8 1.7 1.8 1.8 1.8 1.7 -0.1 Fiscal Deficit 4.2 4.6 3.3 3.1 2.2 1.6 -2.6 Sources: Selected Public Expenditure Reviews; IMF staff reports and Bank staff estimates. 1.5 Even relative to adjusted GDP, however, government spending in the region is 4-5 percent of GDP larger on average than in faster growing emerging market economies or in countries with similar levels of GDP per capita. The discussion in this report, and the ratios to GDP, use officially-reported national account data. While most countries in the region have incorporated adjustments for the size of the non-observed economy (NOE) in national accounts, Montenegro and Serbia have not and in others the included adjustments are much smaller than suggested by third-party estimates of the NOE (BH, Macedonia and Serbia). Including adjustments for the NOE for Montenegro and Kosovo, and boosting the adjustments for BH, Macedonia and Serbia to reflect more fully third-party estimates of the NOE, 4 The Stabilization and Association Process (SAp) is the EU’s policy for the countries of Western Balkans. The SAA is the final stage of that policy, creating a contractual relationship between a country and the EU. The SAp is accompanied by generous financial assistance, now provided under the Instrument for Pre-Accession (IPA). The IPA replaced the so-called Community Assistance for Reconstruction, Development and Stabilization (CARDS). 3 would result in general government spending amounting to about 41 percent of GDP in 2005, or 2.5 percent of GDP lower than relative to officially-estimated GDP. 1.6 Expenditure reduction and rationalization notwithstanding, key fiscal issues remain to be resolved. These include the still large size of governments in many of the SEE countries and the increased rigidity of spending (with larger shares accounted for by public wages and social transfers, categories that can be reduced only gradually). The composition and efficiency of spending, moreover, is often seen as not conducive to growth, a considerable concern given the countries’ aspirations to catch up quickly with EU standards of living. Further rationalization and reduction of expenditures, together with tighter fiscal policies, should help limit external imbalances, provide room for the authorities to tackle spending pressures that are set to arise going forward and help assure debt sustainability. 1.7 Fiscal pressures going forward are likely to reflect in part the direct fiscal cost of EU integration resulting from the harmonization of standards and institutions with the acquis communautaire.5 In addition, pressures are likely to emerge as a result of: (i) past policy commitments on trade liberalization and decentralization; (ii) ageing population; (iii) still pending reforms of social welfare and education, enterprise restructuring, and public outlays on infrastructure; and (iv) public administration reforms that aim to attract highly qualified personnel to guarantee a merit-based and depoliticized civil service. In addition, some SEE countries, such as BH, have considerable amounts of domestic debt and contingent liabilities (including unsettled claims for arrears of budgetary and extra budgetary beneficiaries, old foreign currency savings, war damages and restitution). 1.8 Given the limited scope (and Table 0.4: General Government Debt wisdom) for further increasing the (In percent of GDP) tax burden because of the already 1995 … 2004 2005 2006 detrimental effects of high taxation Albania 35.3 56.6 56.7 55.7 on growth and private sector BH … 27.7 28.1 25.1 development, fiscal space to tackle with domestic claims 1/ … … 48.6 43.2 the above-mentioned fiscal pressures would need to be found by reducing Croatia 2/ 30.3 43.2 43.8 41.0 existing expenditures. Reducing Macedonia 48.0 36.6 39.9 34.7 existing spending also needs to Montenegro … 53.0 44.5 38.3 Serbia … 53.7 46.4 39.0 accommodate cuts in high marginal Slovenia 18.8 28.8 28.0 28.2 payroll contribution rates and the need to sustain or accelerate fiscal SEE 33.1 51.2 adjustment to assure debt EU7 31.7 39.5 38.6 38.1 sustainability. EU15 3/ 70.9 64.7 Notes: 1/ Includes yet to be formalized claims for government arrears, frozen 1.9 Government debt levels in currency deposits and war claims; 2/ Does not include government guarantees; most SEE countries have been 3/ Data for 1997. Sources: World Bank, Eurostat. reduced in the last decade, reflecting limited borrowing needs as a result of privatization receipts and narrower fiscal deficits, and debt workouts with the London and Paris Clubs earlier this decade or in the late 1990s (Table 1.4). Government debt increased in Albania, by contrast, and at about 56 percent of GDP over the last several years is the highest in the region. Debt has also increased in Slovenia. With privatization largely completed, deeper fiscal adjustment is needed in countries with relatively high debt levels, including Albania. Deeper fiscal adjustment will likely be needed in Croatia, given the growing risk that 5 Referring to the body of EU law and institutional provisions. The cost of implementing the acquis refers to the co- financing costs for most projects financed by the pre-accession funds, and the cost of establishing the necessary additional administrative infrastructure. 4 government guarantees to the shipyards may well be called, as well as other guarantees extended by the State Development Bank. II. GOVERNMENT REVENUES 1.10 General government revenues rose by 1.2 percent of GDP this century, amounting to 43.1 percent of GDP in 2005, contributing to somewhat less than half of the average fiscal adjustment in all countries. The introduction of value-added taxes that replaced Yugoslav-era sales taxes has helped strengthen revenues, as has improved revenue collection. This new room for maneuvering has been used to correct sharp cuts in direct taxes. Sizable informal economies in the region are both a cause and a consequence of higher than optimal social security contribution rates. The size of the NOEs among the SEE countries average about 25 percent of GDP, and a large part of these would yield revenues if formalized. Similarly, undeclared household income is estimated to amount to about 25 percent of declared income in the region with direct implication for personal income taxes and social security contribution revenues. 1.11 Measured relative to GDP, revenues in the SEE countries are higher than in countries with similar levels of income per capita, broadly similar to the levels in the EU7, though lower by about 6 percentage points than in the EU15. The SEE countries can be divided into two groups: one including countries with relatively low revenue generation (Albania and Kosovo), and another with high spending and tax burden (BH, Croatia, Slovenia, and Serbia and Montenegro). In Kosovo, the newly established fiscal system is evolving: revenues have risen since 2002 (to a large extend due to a surge in foreign grants), and public expenditure policy has been improving. Macedonia falls in between the two groups. 1.12 Revenues from income taxes are low in the region relative to GDP, reflecting large reductions in rates over the last several years and pervasive evasion. In Albania, BH and Macedonia, for example, income tax revenues amount to just 3.5 percent of GDP and in Serbia to about 6 percent of GDP, compared with 8 percent in the EU on average. Despite low personal income tax rates, however, relatively large social security contributions have boosted the wedge between labor costs and take-home pay, contributing to high informality and unemployment. Aside from distortionary social security taxes, the bulk of the tax revenues in SEE are generated from less distortionary indirect taxes. A VAT was introduced in most countries this century (in Slovenia in the late 1990s) and has contributed to boosting revenues as collection has been streamlined. III. LEVEL AND TREND OF GENERAL GOVERNMENT EXPENDITURES 1.13 General government expenditures in the SEE countries fell from 46 percent of regional GDP in 2000 to 44.4 percent in 2005, a moderate downward adjustment on average (Table 1.5). Spending fell most in BH (by about 13 percent of GDP), reflecting a sharp drop in defense spending and outlays on reconstruction funded by foreign assistance, and a drop in interest outlays. Expenditures were also reduced sizably in Croatia and Albania, in the former country reflecting cuts in public wages and transfers in an effort to reverse a large fiscal expansion in the 1990s. In Albania, two-thirds of the reduction in expenditures was accounted for by a drop in interest outlays. In contrast to these developments, government expenditures rose in Serbia, reflecting the assumption of functions previously carried by the federal government of Serbia and Montenegro, starting in 2002. Expenditure also rose in Kosovo, reflecting the surge in donor-financed expenditures and formalization of some previously off-budget spending. Expenditure consolidation in Macedonia has helped reverse most of the sharp increase in spending relative to GDP in 2001 that reflected a sharp contraction in economic activity as a result of the 2001 conflict. 5 1.14 Separating government expenditures into interest outlays and non-interest (primary) spending confirms that only in Albania was a sharp reduction in interest payments instrumental for reducing overall government spending (Table 1.6). The reduction in interest outlays was also large in BH, but their contribution to the overall reduction in spending was rather moderate. The reduction in interest expenditures (and, as shown above, government debt) in BH, Serbia and Montenegro has benefited from debt restructuring with the London and Paris Clubs. Table 0.5: Fiscal Developments in SEE, 1997-2006 (In percent of GDP, regional averages are weighted by nominal GDP) 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Revenues … … … 41.9 41.6 42.6 42.6 43.1 43.1 … Albania 16.9 23.3 22.7 23.8 23.7 24.7 24.5 24.1 24.0 25.2 BH 56.0 46.9 50.5 47.6 44.0 42.1 42.0 49.6 50.8 47.0 Croatia 48.3 51.8 48.8 47.2 45.7 45.8 45.5 45.6 45.4 46.0 Kosovo … … … … … … … 27.0 28.2 30.8 Macedonia 34.8 33.3 35.4 36.6 37.0 38.2 38.4 36.5 35.5 33.7 Montenegro … … … … 38.0 35.1 35.0 36.4 37.7 … Serbia … … … 32.4 35.8 40.1 40.7 41.5 41.4 41.1 Slovenia 42.2 43.1 43.7 44.3 44.8 45.5 45.3 45.1 45.6 44.8 Expenditures … … … 46.0 46.0 46.3 46.4 45.5 44.4 … Albania 29.2 32.5 34.9 31.9 31.6 31.4 29.0 29.2 27.4 28.3 BH 55.4 50.4 57.7 53.5 46.9 48.3 46.8 49.7 49.5 44.2 Croatia 50.7 54.2 56.6 53.4 52.0 49.8 51.3 50.3 48.5 48.4 Kosovo … … … … … … … 33.1 31.3 27.3 Macedonia 35.1 35.0 35.4 34.3 43.6 42.2 38.5 36.1 35.3 34.3 Montenegro … … … … 42.8 38.9 39.9 39.9 40.4 … Serbia … … … 32.6 36.2 43.4 43.7 41.4 40.6 42.5 Slovenia 43.4 43.8 44.4 48.1 48.9 48.0 48.0 47.4 47.0 46.3 Fiscal Balance … … … -4.1 -4.4 -3.8 -3.9 -2.4 -1.3 … Albania -12.3 -9.2 -12.2 -8.2 -7.9 -6.6 -4.5 -5.1 -3.4 -3.1 BH 0.1 -3.6 -7.2 -6.0 -2.9 -6.2 -4.9 0.0 1.3 2.8 Croatia -2.4 -2.4 -7.8 -6.2 -6.3 -4.0 -5.8 -4.6 -3.1 -2.4 Kosovo … … … … … … … -6.1 -3.1 3.6 Macedonia -0.4 -1.7 0.0 2.4 -6.6 -3.9 -0.1 0.4 0.2 -0.6 Montenegro … … … … … … -4.9 -3.5 -2.6 … Serbia … … … -0.2 -0.4 -3.4 -3.0 0.0 0.8 -1.5 Slovenia -1.2 -0.7 -0.7 -3.8 -4.1 -2.5 -2.7 -2.3 -1.4 -1.5 Note: The averages do not include Kosovo. SaM is presented as included for reference only: the figures for SaM are for comparison only. Sources: National authorities, the IMF, and Bank staff estimates. 1.15 Government spending in most SEE countries, however, remains higher than among countries with similar levels of income per capita, higher than in countries with similar real GDP growth rates, and higher than in fast growing economies (Table 1.6). At about 44.5 percent of GDP on average, outlays are higher than the 40 percent average among the NMS and are well above averages for ECA as a whole, East Asia and Latin America. Even relative to GDP adjusted more fully for the size of the NOE, spending is about 3 percent of GDP larger than in countries with similar levels of income per capita, or countries growing faster. It is telling that government spending even in the country with the lowest level of 6 spending in Europe and SEE (Albania), is higher than in Chile, a country with income per capita that is twice as high as Albania’s. Moreover, Albania’s government spending is only marginally lower than the average for the countries of Latin American and the Caribbean (29.6 percent of GDP) and East Asia and the Pacific (29 percent). On a more disaggregated level, government spending in BH and Croatia is about 5 percent of GDP and 7 percent of GDP, respectively, higher than in comparator countries, while in Albania it is about 5 percent of GDP lower and in Macedonia is broadly similar to that in comparators.6 Table 0.6: General Government Interest Payments and Primary Expenditures, 1997-2006 (In percent of GDP, weighted by nominal GDP) 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Primary Expenditures … … … 44.3 44.3 44.6 44.7 43.8 42.7 … Albania 23.6 24.0 27.5 26.1 27.4 27.5 24.6 25.5 24.9 25.7 BH 54.5 49.7 56.7 52.4 45.9 47.4 46.1 49.1 48.9 43.7 Croatia 49.2 52.7 55.0 51.5 49.8 47.7 49.2 48.2 46.2 46.1 Kosovo … … … … … … … 33.1 31.3 27.3 Macedonia 33.0 33.1 33.9 32.5 41.7 40.7 37.4 35.2 34.4 33.1 Serbia 32.0 35.5 42.6 42.7 40.2 39.2 41.1 Slovenia 42.2 42.5 43.0 46.7 47.4 46.2 46.1 45.7 45.7 45.0 Interest Payments … … … 1.7 1.7 1.7 1.8 1.7 1.7 … Albania 5.6 8.4 7.5 5.8 4.2 3.9 4.4 3.7 3.1 2.8 BH 0.9 0.7 1.1 1.2 1.0 0.9 0.7 0.6 0.6 0.6 Croatia 1.5 1.5 1.7 1.9 2.2 2.1 2.1 2.1 2.3 2.2 Kosovo … … … 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Macedonia 2.2 1.9 1.5 1.8 1.9 1.5 1.1 0.9 0.9 1.0 Serbia 0.6 0.7 0.9 1.0 1.2 1.4 1.4 Slovenia 1.2 1.3 1.4 1.4 1.5 1.8 1.9 1.7 1.3 1.3 Sources: National authorities, the IMF, and Bank staff estimates. 1.16 Government spending is also high when comparing with countries with similar real GDP growth rates. Figure 1.3 plots the actual and predicted values of government spending from a regression for 38 countries over 1994-2003 (Serbia, Montenegro and Kosovo are not included because of lack of data).7 Actual and predicted spending is the same for countries falling on the diagonal. For countries under the diagonal, actual expenditures are higher than predicted. While actual expenditure levels in Albania, Macedonia and Slovenia are close to the model’s predictions (but still larger than the predictions), spending in Croatia and Bosnia and Herzegovina is substantially above predicted (by as much as 15 percentage points) (Figure 1.4). 6 See Bosnia and Herzegovina, 2006, Public Expenditure and Institutional Review, The World Bank. 7 The observations were selected by putting together a dataset containing only countries whose per capita GDP growth rates (on a purchasing power basis) were at least 2.5 percent (or one-quarter standard deviation more that the average growth rate in about 130 countries). Average expenditure ratios in 2000 to 2004 were regressed on real per capita GDP in 2000 (in logs) and on regional dummy variables. The predicted expenditure ratios are calculated from the coefficients of this regression. 7 1.17 All told, there is Figure 0.3: Primary Expenditure and Income per Capita substantial room for some (Average 2000-2005, percent of GDP and US dollars per capita) countries to reduce government 60.0 expenditures relative to GDP BH 55.0 without jeopardizing growth Croatia prospects. Noninterest Expenditures (In percent of GNI) 50.0 Slovenia 45.0 SaM IV. EXPENDITURE 40.0 MAK COMPOSITION 35.0 30.0 1.18 High level of Albania government spending has been 25.0 accompanied in the region by 20.0 poor outcomes. The challenge 15.0 of reducing the government burden on the economy needs to 10.0 0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000 be tackled together with GNI per capita (In US$, average 2000-2005) restructuring spending to increase its growth impact and improving the stance of fiscal policy to reduce external vulnerabilities. 1.19 There is a wide disparity in the region, but the structure of expenditures in several Figure 0.4: Actual and Predicted Levels of General countries is seen as not supportive of growth Government Expenditures (notably BH and Macedonia). Of key concerns is the share of wages, social transfers and subsidies. In addition to helping retool fiscal policy to be growth-enhancing, examining the composition of public spending should help identify the scope of possible expenditure retrenchment and restructuring. A. Economic Classification 1.20 On average, current expenditures accounted for broadly similar shares of overall general government spending in the SEE and in the EU7 (Table 1.7). Nonetheless, government capital spending in the EU7, as well as Slovenia, is much lower than in the SEE as a share of GDP, thanks to higher private spending and better quality of public infrastructure. Within the SEE, Kosovo, Albania, BH and Croatia dedicate larger shares of spending to capital projects than the SEE average. Within current spending, SEE countries spend more on public wages and transfers; these two categories, together with capital outlays, account for all the difference in spending levels between the SEE and EU7. Examining the composition of spending gives interesting insights and should help design reform strategies going forward. 8 Table 0.7: General Government Expenditures, Economic Classification (In percent of GDP; data for 2004 unless indicated otherwise) Croatia SEE EU7 Albania BH Kosovo Macedonia SaM Slovenia (2006) Average1/ Average Total expenditures 29.2 29.2 48.0 33.1 36.1 45.4 47.4 45.8 40.3 Current expenditure 22.8 40.6 39.7 25.6 33.0 42.7 14.5 40.7 36.5 o/w: Wages and salaries 7.0 13.0 9.8 8.1 8.3 10.3 11.3 10.6 7.8 Interest payments 3.7 0.6 2.2 0.0 0.9 1.4 1.7 1.8 1.7 Subsidies and transfers 9.7 19.1 18.7 9.0 18.9 22.1 18.6 19.6 18.8 Others 2/ 2.4 7.9 0.0 0.0 4.9 8.9 12.9 8.7 8.2 Capital expenditure 6.4 7.5 8.3 7.5 3.1 2.7 2.9 5.1 3.8 Memoranda: Current spending in % of total expenditures 78.1 84.4 83.2 77.3 91.4 94.1 93.9 88.9 90.6 1/ Weighted by nominal GDP. 2/Including clearance of arrears, structural, statistical discrepancies. For Macedonia including also refugee-related expenditures. Source: National authorities and the IMF estimates 1.21 Wages and salaries: Except in Albania, (and in Macedonia and Serbia if measured relative Figure 0.5: Public Sector Wages and Salaries to GDP more fully adjusted for the size of the non- 14.0 Percent of GDP 12.0 observed economy) outlays on public wages are 10.0 8.0 well above the EU7 average, and close to the EU15 6.0 average of 10.5 percent of GDP. After BH, which 4.0 2.0 spends the equivalent of 13 percent of GDP for 0.0 public wages, Slovenia also dedicates a large share a ia vo ti a 1/ ia ni ) en 03 on a ro so ba oa in 15 ov 20 eg ed Ko of government expenditures to wages (Figure 1.5). Al Cr ov EU Sl ten 7( ac eg M EU on rz He M Even in BH, however, the 2004 outcome is lower & & ia ia rb sn than in earlier years, reflecting army Se Bo demobilization. Source: National authorities, IMF and the World Bank. 1.22 High public wage outlays in most SEE countries have not translated into better public administration in SEE (except in Slovenia) compared with the EU7 or, as is to be expected, the EU15 (Figure 1.6). The corruption perception index compiled by Transparency International ranks Figure 0.6: Public Administration Performance Slovenia 28th, Croatia 69th, and all other SEE Composite Governance Indicators countries roughly much more poorly placed.8 In EU-15 the World Bank’s Doing Business Survey for EU-7 2007, Slovenia is ranked highest at only the 61st Slovenia SCiG place.9 These results underlie the dual challenge Macedonia 2000 faced by the SEE countries: improve the Croatia 2004 effectiveness of public administration and B&H Albania strengthen its capacity while reducing costs. 0 20 40 60 80 100 These challenges are discussed in more detail in Country’s percentile rank the following chapter. Source: WB Governance Matters 2004 Note: Includes Voice and Accountability, Political Stability, Government Effectiveness, Rule of Law, Control of Corruption, Regulatory Quality 1.23 Subsidies and current transfers: Bona fide subsidies to enterprises within this category are much smaller than transfers to households, but 8 Transparency International, 2006, Corruption Perception Index, www.transparency.org. 9 Doing Business, 2007, The World Bank. 9 the expansion of subsidies since the start of the century to cover costs of enterprise restructuring and for agriculture has been pronounced. Subsidies and current transfers are at the high end in most countries of former Yugoslavia, while they account for just half the regional average in Albania and Kosovo. 1.24 Capital expenditures: Capital spending in Albania, Kosovo, BH and Croatia is much larger than the SEE average, reflecting for the first three countries disbursements of generous international assistance and outlays on post-war reconstruction. Croatia’s public outlays also reflect accelerated and ambitious motorway construction (amounting to about 6 percent of GDP in 2004). While capital is one of the key factors that generate economic growth, there is a need to exercise caution as to the government contribution to capital stock and the pace of its accumulation. To limit the government burden on the economy, the SEE countries would need to consider ways to increase the involvement of the private sector, but such an involvement needs to be preceded by a substantially improved institutional framework, including for concessions and Public Private Partnerships (PPPs). The capacity needs to be developed to evaluate investment projects and allow these to proceed only in line with the fiscal envelope, under a well-prioritized strategy and in line with international best practice. Further, governments should work to ensure cost recovery from users, where feasible: for example, appropriate pricing of utility services can help to control the demand for services, reduce the investment needs, and help to recover costs. B. Functional Classification 1.25 The functional classification of government spending reveals similarities between the SEE and EU7 countries, with defense, social security and welfare, health, and education accounting for the bulk of public expenditure (Table 1.8). The following paragraphs present some stylized facts of the comparisons, some of which are explored in more detail in the following chapters. Table 0.8: General Government Expenditures, Functional Classification (In percent of GDP; data for 2003 unless indicated otherwise) BH Croatia Kosovo Macedonia Serbia and Slovenia SEE EU7 Albania (2004) (2006) (2004) (2002) Montenegro (2004) (Mean) (Mean) Total expenditures 28.9 52.6 48.0 33.1 41.8 45.8 44.0 46.6 40.4 General public services 4.0 7.7 3.2 7.3 6.1 4.6 8.5 4.8 4.7 Defense 1.1 2.4 1.4 0.8 2.3 2.7 1.4 2.0 1.7 Public order and safety 1.7 4.8 2.5 3.5 2.9 2.4 2.0 2.8 3.2 Education 3.1 6.2 4.7 4.8 3.7 3.5 5.9 4.5 6.4 Health 2.3 7.0 6.2 3.8 5.8 6.2 6.8 5.6 5.4 Social security and welfare 7.2 15.1 15.9 4.6 12.7 17.9 18.3 13.5 13.2 Housing and community amenity 1.7 … 3.6 0.9 0.3 1.9 … 1.3 2.0 Recreational, cultural 0.4 … 1.4 0.4 0.6 0.6 … 0.7 1.2 All economic services 3.1 4.3 6.5 7.0 3.6 4.2 3.5 5.4 5.3 Other, including interest and statistical adjustment 1.3 5.1 2.6 0.0 1.2 1.8 -2.4 1.8 -7.9 Sources: National authorities, the IMF, and Bank staff estimates. 1.26 Defense: At about 2 percent of GDP, the level of defense outlays in the SEE is similar to that in the EU7 but somewhat higher than among the EU15 on average. Public expenditures on defense in SEE have been much reduced from substantially higher levels in the late 1990s and early this century. During 1996-2004, the SEE countries spent on average over 4 percent of GDP on defense, reflecting the wars of 10 Yugoslavia’s disintegration. Croatia, BH, Serbia and Montenegro have made considerable progress in reducing and rationalizing defense expenditures. However, governments continue to feel under pressure to increase defense spending to upgrade their armed forces in line with their commitments to NATO ahead of potential NATO entry. 1.27 Social Protection: The SEE countries face the challenge of ensuring fiscally affordable, equitable and efficient social services. Spending on pensions, health, social assistance, and child allowances alone amounts to 15.2 percent of GDP on average and at one-third of expenditures constitutes the largest component of SEE budgets. Pensions and health are the largest spending categories within social protection, accounting for about one-half and one-third, respectively, of overall social spending. 1.28 Social protection expenditures among the SEE countries are lower than among the new member states of the EU (NMS) in health but are lower for pensions and broadly similar for social assistance and unemployment protection. Rapid ageing and declining populations, except in Albania and Kosovo, are likely to put pressure on pensions systems going forward. Chapter 2 analyzes these issues in detail and offers several policy recommendations. 1.29 Health: Public expenditure on health care is higher in SEE compared to other countries with similar levels of income per capita and with the NMS (except in Albania). Ageing population, rapid development of new medical new technologies and changing epidemiological profiles continue to exert upward pressure on health care costs. Also, most SEE governments have launched decentralization processes in the health sector, aiming at transferring ownership of primary and secondary care facilities to local governments and making them responsible for their maintenance, while funding actual service provision from the central budget. Despite recent progress, sectoral restructuring and organizational reforms in SEE have lagged behind those in EU8. As result, the sector continues to face a series of challenges: (i) worsened infrastructure owing to neglected maintenance; (ii) increasing arrears and waiting lists for services; and (iii) poor quality of service delivery. 1.30 Education: Spending on education as a share of GDP is in some SEE countries comparable to EU levels, while in others it is significantly below the EU average. However, even if the education spending as a share of GDP had remained relatively constant, the demographic trends in SEE imply that education spending per student has been and is increasing in many countries. At the same time, education sector outcomes in SEE are poor. Completion rates among tertiary students, for instance, are half of the OECD rates. There is a need to explore ways to increase efficiency of education spending and increase the quality of education to develop a highly skilled and adaptable labor force. 11 V. LESSONS LEARNED FROM THE EU MEMBERS 1.31 EU members have gone through several episodes of fiscal adjustment that Figure 0.7: NMS Fiscal Adjustment, 2000-05 offer valuable lessons to the SEE (In percent of GDP; expenditures and interest on the left axis; countries. Key among them is the fiscal balance right axis) conclusion that EU accession should not lead, per se, to fiscal expansion; quite the 45 -8 opposite has been true for the NMS 44 Interest Primary Expenditure Balance -7 43 Expenditures, % GDP (Figure 1.7). The 10 NMS (excluding 42 -6 Deficit, % GDP Bulgaria and Romania) reduced primary 41 -5 40 -4 expenditures to 39.6 percent of GDP in 39 -3 2005 from 41.9 percent in 2000, while the 38 -2 average fiscal deficit was cut from nearly 37 Balance -1 36 4 percent of GDP to 1.8 percent over the 35 0 same period. The bulk of expenditure 2000 2001 2002 2003 2004 2005 reduction reflected cuts in current non- wage outlays, but with the exception of Source: Eurostat; and World Bank staff estimates. wages, cuts were in most major expenditure categories. Wage outlays rose marginally as public administration capacity was strengthened as EU accession was advanced. 1.32 Fiscal adjustments among the older members of the EU are richer in lessons and more diverse (Table 1.9). The most expenditure reductions were in the following areas: • Social security and welfare payments: In most Nordic countries eligibility criteria for unemployment benefits were tightened and the pension system was thoroughly reformed in the 1990s. • Wages and salaries: Cuts in the government employment accompanied by moderate increases in public sector wages in an environment of high economic growth were central to fiscal consolidation in Ireland in 1988-89. Irish government adopted schemes for early retirement of public employees, while the rapidly expanding economy absorbed the released employees in the private sector. The wage moderation was largely accepted after the government carried out broad tax cuts and thus boosted after-tax incomes. • Health and education: Ireland and Sweden successfully reformed their healthcare and education systems, respectively. • Subsidies and transfers: Subsidies were cut in most countries in the 1990s. • Defense: Most countries managed to achieve some cost reductions in this area in the 1990s. 12 Table 0.9: Expenditure Adjustments in Select EU Countries in 1990s (In percent of GDP) Ireland Denmark Finland Spain Sweden Netherlands UK 1988-89 1995-97 1984-85 1996-97 1994-97 1994-97 1994-97 1994-96 1994-97 Expenditures -11.4 -6.2 -5.2 -3.8 -10.9 -6.6 -8.5 -5.2 -5.9 Defense -0.1 -0.2 -0.5 -0.1 -0.2 -0.3 -0.4 -0.4 -0.9 Education -1.3 -0.7 -0.8 0.3 -1.5 -0.5 -0.9 -0.6 -0.1 Health -1.3 -0.5 0.0 -0.1 -0.2 -0.5 0.6 -0.2 -0.2 Social security and Welfare -2.5 -2.1 -1.7 -2.5 -8.0 -2.6 -5.8 -2.2 -2.3 Housing and other Amenities -1.6 -0.3 -0.6 0.0 -0.3 -0.1 -1.9 -1.1 -0.3 Economic services -2.9 -0.3 -0.8 -0.1 -1.9 -1.6 -3.7 -0.1 -0.9 Interest payments -1.1 -1.8 1.3 -0.8 0.9 -0.3 1.0 -0.5 0.4 Current expenditures -10.8 -5.7 -4.7 -3.5 -10.8 -5.5 -7.6 -4.5 -4.4 Goods and services -1.8 -1.3 -1.1 -0.2 -1.1 -1.8 0.0 -0.2 -1.2 Wages and salaries -1.0 -1.1 -0.8 -0.2 -1.0 -1.1 -0.3 0.0 -2.0 Others -0.8 -0.2 -0.3 0.0 -0.1 -0.6 -0.6 -0.2 1.0 Interest payments -1.1 -1.8 1.3 -0.8 0.9 -0.3 -0.6 -0.5 0.4 Subsidies and transfers -7.9 -2.7 -4.8 -2.5 -10.6 -3.4 -8.6 -3.8 -3.6 o/w: Subsidies n.a. n.a. n.a. -0.1 -6.0 -0.4 -3.4 0.5 -0.3 Capital expenditures -0.6 -0.4 -0.5 -0.3 -0.2 -1.2 -0.8 -0.6 -1.5 Source: IMF (2003), Hungary: Selected Issues, IMF Country Report No. 03/125. VI. CONCLUSION 1.33 Economic recovery in the SEE countries has been accompanied by commendable fiscal adjustment in most states that has, together with efforts in advancing structural reforms and propitious global conditions this century, lifted living standards considerably. More remains to be done in the SEE countries, nonetheless, to improve the conditions for private sector-led growth and fine-tune the size, quality and efficiency of governments. The rest of this study will explore in further detail government spending in SEE and expenditure policies in selected areas. The focus is on areas that account for the largest share of public spending across countries and are the key to ensuring fiscal and social sustainability over the longer term. The analysis works to distill important lessons that emerge efforts to reform expenditure policies and consolidate public spending, and highlight challenges and options for the way forward in each area. 1.34 Chapters 2 and 3 examine the current situation and challenges ahead in social protection, pensions and health. In these areas, there is strong popular dissatisfaction with the quality of service provision, equity and efficiency, while fiscal pressures are accumulating. The chapters examine options for reforms in social insurance and transfers. 1.35 Education received much less attention in the past than social transfers when analyzing public expenditures. This neglect, however, has not been justified given the primacy of education and human capital for growth and living standards. Urgent efforts are needed to revamp education policies and spending, especially in light of the long lags most changes need. Chapter 4 examines these issues in detail. Finally, Chapter 5 analyzes the sizeable public sector wage bills in SEE, and review the size, structure, and remuneration of the public service and discusses possible reform options. 13 14 Ensuring Social Welfare: Social Protection 2.1 Generous social protection systems in the SEE countries are a key legacy of the past. In the Socialist Federal Republic of Yugoslavia (SFRY) and in Albania, similarly to the other countries of central and eastern Europe, there was nearly universal coverage by social insurance (for health, old-age and disability), and eligibility for social benefits and unemployment relief. Nearly universal social protection coverage was accompanied with nearly universal formal employment, resulting in the choice of payroll contributions to finance the insurance schemes. 2.2 These social protection models quickly proved fiscally unsustainable at the start of transition as formal sector employment plummeted due to the collapse of linkages between the constituent countries and enterprises both between and within them, the start of privatization, and a pickup in informal activity.10 Wars and ethnic conflicts added to the economic dislocation of transition and led to increasing demands for social protection spending to accommodate the social and economic re-integration of war veterans, invalids, refugees and displaced persons. Soaring unemployment and the emergence of new pockets of poverty posed additional demands for increased social transfers. The demographic transformation that took place at the same time -- falling or stagnating birth rates and steadily increasing longevity along with spells of emigration affecting mostly those aged 20 to 35—led to a sharp increase in old-age dependency ratios in all SEE countries, aggravating further the financial standing of their social protection systems. 2.3 The governments of the SEE countries adopted different strategies to halt the sharp rise of social protection spending. These included accumulation of deficits, sharp divergence between legally mandated benefits and actual payments, and the reduction of the scope and size of benefits. Understanding that without change these “cradle-to-grave” social protection systems were unaffordable became the main motivator for governments to consider either partial or comprehensive reforms. Efforts thus far have borne fruit, helping limit government deficits, but further adjustment and profound change to social protection systems would be needed in most countries to assure the efficiency of spending, reduce the government burden on the economy while improving equity, and assure fiscal sustainability. 2.4 This chapter reviews public expenditure policies on social protection in the SEE countries during the last decade, outlining the efforts taken to balance the efficiency, effectiveness and affordability of pension, social assistance, and labor-market programs. The chapter also examines the relative effectiveness of particular reform measures taken in different countries and their applicability to others, highlighting options for the way forward. I. PUBLIC SOCIAL PROTECTION SYSTEMS 2.5 Although social contribution rates remain high in the SEE countries, the gap between social security revenues and expenditures widened in the decade through the early 2000s (Table 2.1). This contrasts with the experience of the CEE countries that faced similar transition challenges (including rising unemployment, vulnerability, and poverty), where the social protection deficit has been shrinking over time. 10 In mid 1980s social protection spending accounted for 33 percent of total government spending in SFRY. 15 2.6 While the delay in restructuring of social protection spending in SEE can be partly attributed to wars and ethnic conflicts, the fact that these expenditures have continued to rise in the past few years suggests that other factors have also been fueling the rise. In 1996-2001, the gap between revenues and expenditures was on average 33.5 percent of total government revenues in SEE countries, while the CEE average was 16.2 percent. Further, in 2002, the deficit had expanded to 53 percent in SEE, but had fallen by two percentage points in CEE. A. Social Protection Expenditures 2.7 Of SEE countries, only Macedonia and Albania reduced social protection spending through 2002. After reaching a peak of eight percent of GDP in 1996 and 1997, Albania has recorded a steady decrease in the share of GDP devoted to public social protection. Similarly, Macedonia managed to reduce its social protection expenditures from the peak 13.6 percent in 1999 to 12.7 percent of GDP in 2002. All other countries in the region have ether retained the level of expenditures (for example, BH at 15 percent of GDP, and Slovenia at 18 percent) or increased spending. In Croatia, social protection expenditures peaked at 17.9 percent of GDP in 2001 after rising from 12.3 percent in 199, but subsequently fell to 17.9 percent in 2002. Table 0.1: Public Revenues and Expenditures on Social Protection (In percent of GDP) Revenues Expenditures Balance 1996-2001 2002 1996-2001 2002 1996-2001 2002 SEE Albania 3.7 3.8 7.8 7.0 -4.1 -3.2 BH 12.1 14.3 15.4 16.1 -3.3 -1.8 Croatia 13.8 12.3 17.4 17.9 -3.6 -6.0 Macedonia 11.3 10.6 12.8 12.7 -1.5 -2.1 Serbia and Montenegro 11.4 10.7 13.1 … -1.7 … Slovenia 13.8 13 17.9 18.3 -4.1 -5.3 CEE Bulgaria 7.5 7.3 11.7 13.4 -4.2 -6.1 Czech Republic 14.7 13.9 13.8 13.8 0.9 0.1 Estonia 10.6 12.3 … 11.2 … 1.1 Hungary 11.9 12.9 14.2 15.7 -2.3 -2.8 Latvia 10.1 10.2 14.5 12.4 -4.4 -2.2 Lithuania 7.6 6.8 10.7 10.1 -3.1 -3.3 Poland 11.3 11.5 19.5 … -8.2 … Romania 9.3 10.7 10.2 9.9 -0.9 0.8 Slovak Republic 12.7 12.7 12 13.8 0.7 -1.1 Notes: a/ For Albania 1996-2000; b/ For Serbia and Montenegro expenditures only for 2000 and 2001; c/ For BH expenditures for 2000 and 2001 Sources: Selected World Bank Public Expenditure Reviews and IMF documents. 2.8 Measured relative to officially-estimated GDP, outlays on social protection, including all contributory and general revenue-funded social transfers are smaller than among the NMS. If GDP in Macedonia, BH and Serbia is adjusted more fully to reflect the true size of the non-observed economy, outlays are likely to be broadly similar in the SEE and the NMS on average (Table 2.2). Among the SEE countries, Croatia, Serbia and Montenegro have the largest transfers relative to GDP, well above the NMS average, while Albania and Kosovo have substantially smaller outlays. 16 Table 0.2: Public Expenditures on Social Protection: A Breakdown, 2005 (In percent of GDP) Kosovo Serbia SEE Albania BH Croatia Macedonia Montenegro NMS (2006) (2004) (Mean) Contributory 1/ 7.8 13.1 18.0 4.9 16.9 13.1 15.2 Pensions 5.4 6.8 11.5 2.8 8.8 7.0 9.9 Health 2.3 5.7 6.1 2.1 5.7 5.0 6.2 5.4 4.7 Unemployment 0.1 0.6 0.4 0.0 2.4 0.7 0.6 Non-contributory 2/ 1.2 3.7 2.2 2.2 1.5 2.5 2.4 Total 6.7 16.8 20.2 12.0 18.4 24.6 20.8 15.2 17.6 Sources: National authorities, the IMF, and Bank staff estimates. 1/ In Kosovo the basic pension, health insurance, and employment assistance are non-contributory. 2/ Includes social assistance, child and maternity benefits, benefits for refugees and veterans, and others. B. Social Protection Revenues 2.9 Social protection revenues have not kept pace with expenditure growth. The main source of revenues remain social security contributions collected through payroll taxes, the rates on which are higher in SEE than in most OECD countries (Table 2.3). High payroll contribution rates have contributed to keep the tax wedge elevated and (especially relevant for countries at such levels of income per capita) have heightened the informalization of the economy and resulted in a pervasive contribution evasion with perverse effects compared to the intended consequences. In addition, several countries (including Macedonia) have minimum threshold bases for paying contributions, which add to the wedge for lower- wage or part-time earners. 2.10 When comparing the average revenues in 1996-2001 to 2002, only Albania and BH appear to have achieved an increase in collected revenues. This is somewhat surprising since Albania recorded the biggest drop in the number of those who pay contributions in the early 1990s. Also, interestingly, Albania managed to increase collections while reducing contribution rates. Social security contributions in Croatia fell moderately from about 14 percent of GDP on average in 1995-1997 to 12.1 percent in 2002. Similarly, in Slovenia revenues from social security contributions dropped from 16.5 percent of GDP in 1995 to 13 percent in 2002. These observations suggest that contribution rates and collections might follow a Laffer curve, but that would require further examination. 2.11 The fact that in the past few years expenditures have continued to rise and revenues have remained insufficient to meet them suggests that reforms in the area of public social protection have not been comprehensive enough. The following sections review the reform experience of SEE countries in the three separate areas of social protection--pensions, social assistance and welfare, and labor-market policies. II. REFORMING THE PENSION SYSTEM11 2.12 Pension outlays in the SEE countries account for a large share of public spending and, similarly to other ECA countries, face rising concerns about fiscal sustainability. Public pension outlays are higher than in most other countries at similar or even more elevated levels of income per capita, reflecting, primarily, the ongoing rapid ageing and declining of populations in most SEE countries and overly generous pension systems. Other key issues include high contribution rates that have contributed to large 11 This section draws heavily on Schwarz, Anita, 2007, “Pension Systems ad Labor Costs in Southeast Europe.” The section focuses on the SEE countries and Kosovo, excluding Croatia and Slovenia. 17 informality and low rate of participation in the system among today’s workers, the latter implying that when these workers retire, a lower percentage of them will be eligible for pensions than today. Thus, the countries face a future with many elderly receiving no benefits from the pension system with adverse social and political consequences, high contribution rates which continue to elevate labor costs, and potentially mounting fiscal costs as well. Table 0.3: Payroll Contribution Rates in SEE and Selected CEE Countries, 2006 Personal Social Security Contributions Income Health Other Social Tax Pensions Unemployment Total Insurance Contributions Rate W E W E W E W E W E T SEE countries Albania 5-30 9.0 14.9 1.7 1.7 -- 2.0 -- 3.6 11.2 21.7 32.9 BH Federation 10-25 17.0 7.0 13.0 4.0 2.0 0.5 -- -- 32.0 11.5 43.5 RS 22 24.0 -- 15.0 -- 1.0 -- -- -- 40.0 -- 40.0 Croatia 15-45 20.0 -- -- 15.0 - 1.7 -- 0.5 20.0 17.2 37.2 Macedonia 15-18 -- 21.2 -- 9.2 - 1.6 -- - - 32.0 32.0 Serbia 10-15 11.0 11.0 6.2 6.2 0.8 0.8 -- -- 17.9 17.9 35.8 1.95- 18.05- 37.55- Montenegro 17-25 12.0 9.6 7.0 6.0 0.5 0.5 -- 2.25 19.5 18.35 37.85 Slovenia 17-50 15.5 8.9 6.27 6.98 0.1 0.1 0.1 0.6 22.1 15.9 38.0 CEE countries Bulgaria 10-24 8.1 15.0 2.1 3.9 1.1 2.0 1.2 2.8 12.4 23.6 36.0 Czech Republic 15-32 6.5 19.5 4.5 9.0 0.4 3.2 1.1 3.3 12.5 35.0 47.5 Poland 19-40 9.8 9.8 9.0 0.0 2.5 -- 6.5 6.5 27.7 20.6 48.3 Romania 16 9.5 19.8 6.5 7.0 1.0 2.5 -- -- 17.0 29.3 46.3 Note: W=employee; E = employer; T = total 2.13 Fiscal sustainability is also an issue due to the flexible retirement provisions inherited from the past, a large number of legal exemptions from paying contributions, and pervasive evasion. At present, in all SEE countries the share of the elderly receiving benefits is greater than the share of those in working age making contributions. Adding to the incentives for early retirement, some of the countries created further special provisions in the context of enterprise restructuring over the past decade that added to undermine the sustainability of pension systems. Pervasive evasion reflects the large and, in several counties, perhaps increasing size of the informal economy and large unemployment.12 Moreover, in countries affected by wars, an increased number of survivor pension beneficiaries and relatively young disability pension beneficiaries put an additional pressure on pension systems. A. Key Issues 2.14 Demographic trends: There are substantial differences in the age profile among the SEE countries (Figure 2.1). Bosnia, Macedonia, and Montenegro are comparable demographically to the younger countries in the OECD, including Ireland and the US (Figure 2.2). Serbia is comparable to the majority of European countries, which are much older. Kosovo and Albania are much younger demographically and more comparable to the younger countries of the developing world. 12 For example, in Albania and Kosovo the informal economy is thought to account for about 40 percent of the overall economic activity and in BH and Macedonia for 30 percent of more. 18 Figure 0.1: SEE Population Aged 65 Years of Figure 0.2: OECD Population Aged 65 Years Older of Older Percentage of Population aged 65 and older Percentage of Population aged 65 and older 20 25 20 15 15 10 10 5 5 0 0 Gr m d No rk ly y Ice d Po tria d Fr in Sw gal Ca S UK itz d Ge ece Be n De da ce Fi y lan an l an l an Sw nlan I ta e u a a U a an na ed lgi nm rw r tu Sp s rm e er Ir e Au Albania Bosnia Kosovo Macedonia Montenegro Serbia Source: World Bank Development Indicators. Source: World Bank Development Indicators. 2.15 Dependency rates: Despite their general youthfulness, the SEE countries still face pension problems arising from a number of historical developments. Before the start of transition, all individuals of working age, male and female, were required to work if able and thus, most elderly have rights in the pension systems and are receiving public pensions (in contrast to much of continental Europe that had lower labor force participation among women). As a result, the percentage of elderly receiving pensions is higher than in comparable West European countries, while retirement ages tend to be even lower. Adding to this burden, most of these countries inherited relatively lax regulations on early retirement for many occupations. Further, corporate restructuring and conflicts have also resulted in an increase in early retirement. 2.16 All told, the ratio of pensioners to workers has been steadily increasing (Figure 2.3). In all countries, the system dependency rates (the number of beneficiaries relative to the number of contributors) far exceed the population dependency rates (the ratio of those older than 65 years to the ratio of those 15-64 years of age) because the number of beneficiaries is significantly higher than what pure demographics would suggest for all the reasons cited above and the number of contributors is significantly lower than what is demographically possible because of the shrinking employment in the formal sector coupled with the growing informal sector. In Western Europe and other OECD countries, the system dependency rate is typically less than twice the population dependency rates. In the SEE, it ranges from a little over three times the population dependency rates in countries like Serbia and Bosnia to over seven times the population dependency rate in Albania. Projections indicate that old-age dependency ratios in SEE would increase to 23.5 percent by 2025 from 13 percent in Albania and 37.4 percent from 21 percent in Slovenia over the same period (Figure 2.4). Figure 0.3: Dependency Rates Figure 0.4: Old Age Dependency Ratios, 2005-2025 System Dependency Rates Compared to Population Dependency Rates 70 90.0% 60 80.0% 50 70.0% 40 2025 60.0% 30 2005 50.0% Population Dep 40.0% System Dep 20 30.0% 10 20.0% 0 10.0% ia tia nia H M a on i Bi an Si oa ve 0.0% ed b Cr o Al ac Sl Albania Bosnia Kosovo Macedonia Montenegro Serbia M Source: World Bank Development Indicators. Source: U.S. Census Bureau, International Data Base, April 2005. 19 2.17 High contribution rates: All SEE countries except Kosovo have high mandatory contribution rates for pensions compared to their West European counterparts (Figure 2.5). High contribution rates are a key contributor to the informalization of the economy, ultimately resulting in a substantial loss of revenues. Concerned by the growth in the informal sector, some governments have lowered contribution rates: Albania, for example, cut pension contribution rates to 23.9 percent of gross wages in September 2006 from 29.9 percent. Cutting contribution rates without improving enforcement efforts are unlikely to be very effective, however. Figure 0.5: Contribution Rates Figure 0.6: Share of Pensioners that Retired (In percent of gross wage) Under the Age of 6513 Contribution Rates as a Percentage of Gross Wage Percentage of Pensioners Below the Age of 65 30 60% 25 50% 20 40% 15 30% 10 20% 5 10% 0 0% Albania Bosnia- FBIH Bosnia - RS Kosovo Macedonia Montenegro Serbia Albania Bosnia Kosovo Macedonia Montenegro Serbia Source: World Bank Development Indicators Source: World Bank databases. 2.18 Low retirement ages: Other characteristics of the pension systems which exacerbate the inherited problems include the prevalence of de facto low retirement ages. In most of the countries, retirement ages for men are rising to 65 or have already reached 65, with the exception of Macedonia where men can retire at age 64. Women, on the other hand, are typically allowed to retire earlier than men (only in Kosovo and in the Federation entity in BH are retirement ages equalized for men and women, although the difference in Macedonia has been reduced to two years from the typical five). There is no economic rationale for a lower retirement age for women, who live longer than men and therefore, from an actuarial perspective, would rather need to accumulate more pension resources to help offset the longer retirement period. But even more problematic is the prevalence of early retirement for both sexes. As a result, the number of pensioners under the age of 65 years is very high (Figure 2.6). 2.19 Benefit Design: The PAYG parametric reforms in all countries have aimed at reducing pension spending and increasing contribution compliance. Expenditure consolidation measures have included increases (unification for men and women in some cases) of the retirement age to extend the contribution periods and reduction of pension benefits in the case of early retirement. Also, changes in the retirement benefit formula have been introduced, increasing the reference period on which benefits are based, and strengthening the link between contributions paid and benefits. 2.20 Further, indexation of pensions solely to nominal wages has proven fiscally unsustainable, and indexation formulas based on varied price-wage composites have been introduced in some countries to make pension spending fiscally sustainable over time. Equalization of retirement provisions for previously different types of contributors has also been done. Table 2.4 summarizes the basic features of reformed PAYG systems across SEE. 13 The past data on Kosovo are incomplete. There may be people below the age of 65 who had previously been granted pensions and may still be receiving them. However, the current policy does not provide pensions to those younger than 65. 20 Table 0.4: Basic Features of Reformed PAYG Systems in SEE Old-Age Pension 1/ Pension Country Contrib. Accrual Rate Men Women Indexation Period Albania 65 60 35 2.1% 100% inflation BH: Federation 65 65 20 2.25% for first 20 years; 1.5% 100 % wage, 40 for next 20 aggregate revenue 65 60 20 2.25% for first 20 years; 1.5% BH: RS constraint 40M/35W for next 20 40; defined contribution – Basic pension: Kosovo 65 65 depends on the interest rate inflation 2.33% (2.6%) for years prior to 80% CPI + 20% 2003; 1.8% (2.05%) for those in Wage the first pillar only; 0.75% and Macedonia 64 62 15 0.86% for those who switch to the second pillar plus defined contribution pension Rising Rising 40M/35W 1.42% of average wage in 50% CPI+50% to 65 to 60 December 2003; indexed to wage Montenegro 50% inflation and 50% wage growth thereafter Rising Rising 40M/35W 1.6% of average wage in 50% CPI+50% to 65 to 60 December 2003; indexed to wage , moving to Serbia 50% inflation and 50% wage 100% CPI from growth thereafter 2009 1/The conditions for a full old-age pension combine the age and length of the contributions period. For example, in Slovenia a woman fulfills conditions for a full old-age pension at age 54, if she has 35 years of contribution history; at age 60 with a 20-year contribution history; or at age 62 with a 15-year contribution history. 2.21 Pension reforms notwithstanding, there are still features in all pension systems (except Kosovo and Macedonia) providing workers with disincentives to contribute. In Albania, the key factor is the maximum pension which cannot exceed twice the minimum pension, while the maximum wage subject to contribution is five times the minimum wage, resulting in a compressed pension distribution relative to the wage distribution. In Serbia and in Montenegro, the problem is a bit different, but the impact in the long run will be similar. Both of these countries have adopted the point system which awards workers points based on their contributions in a given year. The number of points together with the value of the point at the time of retirement then determines the pension. Since the points earned are determined both by the duration of contribution and the level of contribution, the point system in theory closely links contributions and benefits. However, the value of the point in Serbia set at the time of the reform was originally being indexed 50 percent to inflation and 50 percent to average wage growth. Subsequent reforms undertaken to stabilize the huge losses of the Serbian pension system (about 5 percent of GDP) are moving the indexation purely to inflation. The result of this indexation is that while wages and the contributions paid with respect to those wages grow over time, the pensions will not grow in real terms, reducing the pension relative to average wage in the medium term. Montenegro has a similar point system, but is maintaining the indexation of the general point to a 50-50 mix of inflation and average wage growth. 2.22 Bosnia and Herzegovina will also end up de-linking contributions and benefits, but for different reasons. Both entities in BH have constrained their pension expenditures to the revenues available. The system thus far links contributions and benefits, but unless reformed, given population ageing, the number of contributors relative to the number of pensioners will be reduced over time, reducing pensions. The Macedonian system provides a much tighter link between contributions and benefits. Initially, the minimum pension was on the high side, at 41 percent of the average wage in 2000, undermining incentives to contribute on the basis of wages somewhat higher than minimum wage. These disincentives 21 should be reduced over time should the authorities keep the planned indexation to 80 percent inflation and 20 percent wage growth. In addition, the Macedonians put a portion of their contribution in an individual investment account where individuals will earn returns proportional to their contributions. 2.23 The overall system in Kosovo is perhaps less related to contributions than the Macedonian in that a basic pension is given to all inhabitants of Kosovo aged 65 and above, financed out of general revenue, and irrespective of their earnings history. In addition, personal contributions are fully put in individual pension accounts which are invested abroad and earn rates of return. As with the Macedonian individual accounts, investment earnings will vary year by year, but will always be proportional to the contributions invested. This defined contribution part of the Kosovo pension is completely tied to contributions, but the basic pension part is completely unrelated to contributions, resulting in a combination that is not fully tied to contributions. 2.24 Indexation: Indexation is another parameter in the SEE pension systems that tend to raise the costs, making it difficult to reduce contribution rates. Most countries in the OECD index pensions post- retirement only to inflation, as a means of lowering the pension expenditures. Most SEE countries are more generous and link pension benefits in part to wage growth. Serbia is the only country which legally indexes to inflation only.14 2.25 Coverage: In addition to the Figure 0.7: Percent of the Elderly Receiving Pensions demographic and pension design issues, Now and in the Future the countries will soon face a new set of (In percent of all pensioners) problems. Currently, many of the elderly Percentage of Elderly Receiving Pensions Now and in the Future are receiving some type of pension from the public system. As noted before, with 160% the growth of the informal sector in the 140% labor market, many of today’s working age population are not contributing to the 120% pension system and will not have rights to 100% a pension when they retire. As a result, the 2005 80% pension coverage of the elderly is expected 2075 to fall in the longer run, creating additional 60% demands on the government, that of 40% providing some type of old age assistance 20% to the large groups of elderly who have no 0% means of support in old age. The expected Albania Bosnia Serbia change in future coverage among the elderly for three of the countries for which Source: World Bank estimates, PROST model. projections have been completed using the World Bank’s pension model, PROST, is shown in Figure 2.7. In all three countries, there will be declines in pension coverage among the elderly in the future, with the largest declines expected in Albania, although the unusually high current coverage suggests some unintentional duplicate pensions being paid as well as pensions paid to elderly living outside of Albania who are not counted in the Albanian population. Such a reduction in coverage will have substantial social and political implications unless addressed. 14 Even in the Serbian case, full inflation indexation will not be reached until 2009 from the current mix of inflation and wage growth. Moreover, there is growing discussion of reversing the indexation away from inflation after the latest elections in Serbia. The authorities in Macedonia have also proposed to reverse the intention to move to an indexation to 80 percent inflation and 20 percent nominal wage growth to one more in favor or wage growth. 22 2.26 Government transfers to the pension funds: As a result of the above-mentioned trends, the pension systems in all SEE countries have experienced a rapid expansion of pension expenditures that has not been matched by an increase in contributions. The resulting shortfalls have typically been financed by general budget transfers; the extent of these transfers has been so large that it is hard to call public pension systems in the region “contributory.” For example, together with a relatively high pension contribution rate of 20 percent of gross wages in Croatia and 21.2 percent in Macedonia, contributions finance only about 60 percent of pension expenditures in both countries. In Albania (contribution rate of 30 percent of gross wages), contributions finance 70 percent of relatively modest pension outlays. 2.27 Funded pensions: Croatia, Macedonia, and Kosovo have also introduced mandatory defined contribution pension schemes (“second pillar”). Table 2.5 highlights the key features of the second pillar in these countries. Although it is yet too early to make an assessment of the impacts of these reforms on the overall sustainability of the pension system, issues that have emerged include insufficient diversification (some funds invest a much larger fraction of their portfolios domestically, predominantly in government securities, and outcome that in some countries reflects legal limits) and high fees. The experience of these countries should be studied carefully by the other SEE countries before introducing second-pillar pensions. Table 0.5: Second Pillar Introduction Contribution to Country Starting Date First Pillar Switching Strategy Second Pillar Mandatory for younger than 40 years, Croatia January 2002 PAYG DB 5 percent voluntary for those 40-50 years old Kosovo January 2002 Minimum 10 percent Mandatory Mandatory for new entrants in the labor Macedonia January 2006 PAYG DB 7.4 percent market after 2003. B. Pension System Design 2.28 Before proceeding to a discussion of possible reform options, it is worthwhile reviewing the relationship between pension systems and payroll taxes. Pensions are different from almost all other types of public expenditures. They are not targeted toward the poor; typically, richer individuals receive higher pensions. They do not provide positive externalities to others like education expenditures do where even though individuals benefit from education, the society as a whole also benefits from having an educated population. Pensions are even different from other payroll tax financed programs. For example, health insurance which is often financed through payroll taxes is a true insurance program in that no one knows in advance who will need substantial health care and who will need less. Pensions are not really insurance where benefits are rarely paid out, but more a mechanism to prepare oneself for a certain future inability to work. The only uncertainty is the cause of that future inability to work and the timing. 2.29 The above discussion suggests that pension systems are a blend between a savings mechanism, an insurance mechanism, and a poverty alleviation mechanism. The need for such a blend arises from the main objectives of a pension system: (1) the alleviation of poverty in old age, and (2) the replacement of labor income when the individual is no longer able to work due to old age, disability, or death. In an ideal system, the first and second objectives could be separately financed, with government financing for the first and the second left entirely to the discretion of the individual. The Kosovo system is an example of such a split, although the saving for the second objective is mandated by the government, and the savings are managed by a public institution. Typically, however, the second objective may not be left to the individual or even privately financeable. People may be subject to myopia, and neglect to save when 23 young and only realize the need for saving when they are too old to accumulate sufficient savings. People may also realize that the government will not allow them to be poor in old age and will provide some benefits, providing little incentive for individuals to give up consumption during their working years to provide for themselves in old age, thereby causing a moral hazard for the government. 2.30 In light of these factors, countries have historically turned to the government as the only institution capable of providing a secure savings mechanism and of intermediating between generations. Because the government was serving as an intermediary rather than directly providing pensions as a form of government expenditures, the logical conclusion was that the government accepted contributions out of payroll and hold them much as a bank would to finance the promised pension when the individual retired. Today, many countries in the world use payroll taxes from workers to finance pensions for today’s retirees, with programs covering at least some part of the labor force. Once these systems began, politically powerful groups began to lobby for various increases in pensions and reductions in eligibility conditions. These benefits were virtually costless for the politician to grant during the time when there were few pensioners and many contributors. They were in fact costly for future generations, but this additional cost would not become apparent for several generations. As a result, these pension plans around the world became actuarially unbalanced from both political demands as well as the increases in life expectancy which should have led to parametric adjustments to maintain actuarial balance, but which rarely took place. 2.31 The chief drawback of the publicly provided pension programs is their vulnerability to the inevitable aging of the population. Programs which are viable and beneficial when the population is young become fiscally unsustainable when the population begins to age. The cost of maintaining their sustainability falls largely on labor markets since payroll taxes are the main source of financing for the pension expenditures. As a result, many countries are now resorting to pre-funding some or all of their pension liabilities, by accumulating reserves in their publicly managed pension system as is the case in Sweden, Ireland, and the United States, or by gradually moving toward individually funded pension accounts as in the case of much of Latin America and some of the larger countries in Central Europe such as Hungary and Poland as well as in the case of Australia. C. Looking to the Future 2.32 Social Pensions. Given the moderate level of contribution coverage for the near term, contributory pensions are unlikely to provide old age security for all of the next generation of elderly as they reach retirement age. The government will need to supplement with some form of old age assistance. How much old age assistance and to whom needs to be decided on an individual country basis. This old age assistance can be provided in the form of a universal pension available to all those above a certain age, as currently the case in Kosovo, but also in many other countries such as New Zealand. The age at which such a pension is given is often higher than the normal retirement age, providing some incentive for individuals to contribute if they are able, and the amount is generally lower than the minimum contributory pension for the same reason. Offering such a benefit may also relieve countries from the pressure to offer contribution credit for those workers whose enterprises did not remit contributions during wartimes and the transition to a market economy. From a fiscal perspective, the amount and age can be determined based on fiscal resources rather than on some previously acquired right, making the social pension fiscally sustainable. Administratively, this type of pension is relatively to manage since it only requires information on ages or birthdates. From a social perspective, the pension does provide needed old age support, but can be costly because it protects the entire elderly population including those who have additional means of support. Kosovo has opted for this type of support since there were no contributory pensions for any of the elderly. 24 2.33 An alternative form of old age assistance consists of more selective targeting of recipients which is less costly from the fiscal perspective and more effective at reaching those most vulnerable to poverty. In OECD countries, this type of pension is typically available on a means-and-asset-tested basis. However, the administration of a means and assets test is not costless and can lead to efforts to hide resources to gain eligibility to the old age assistance. An alternative lower cost approach might be to simply test the individuals: those individuals receiving a contributory pension will not be eligible for this type of old age assistance. Given the lower level of benefits and the possibly higher age for eligibility, most individuals will not change their contribution pattern just to receive this pension. The issue of rural workers who do not make full contributions, which is a problem in some countries, such as Albania and Serbia, could be resolved by requiring full contributions for a contributory pension. Those who are unable or unwilling to make such complete contributions could receive the old age assistance benefit instead. Given that about half the labor force is contributing in Macedonia, Serbia, and Montenegro, this targeted old age assistance may be sufficient to provide old age support for those who will not be receiving public support in the future. In Albania, a much lower percentage is actually contributing fully, and in Kosovo, the decision has already been taken for a universal basic pension. 2.34 Contributory pensions. Assuming that most countries will retain a contributory system for at least some segments of the population, the contributory system needs to become self-financing as soon as feasible to avoid the negative redistributions involved with a government financed pension system from which only some people are eligible to collect pensions. The historically generous benefits available at relatively young ages are simply not going to be affordable going forward. While this was clear in the early parts of the transition to a market economy and the elderly more or less accepted lower pensions, now that the economies have begun to recover, political pressure to make the systems generous again is beginning to reappear. Countries need to look at their demographics, choose a desired benefit level and then determine the contribution rate required to support the chosen benefit level. If that contribution rate is not affordable, adjustments need to be made in the benefit level and adequately communicated to the population. Conversely, countries could choose an affordable contribution rate and then determine what portion of income will be publicly replaceable by the pension system and what portion will need to be covered on a voluntary basis by additional savings. 2.35 In order to achieve the best combination of contribution rates and benefit rates possible given the current demographics, countries should try to realign their systems with international norms as quickly as feasible. The two primary elements of pension system design which remain outside international norms are the retirement age and indexation. 2.36 Not only does the legal retirement age, particularly for women, need to rise, but the large number of new early retirees also needs to be stopped. Individuals could be allowed to retire early, but with actuarially reduced pensions. 2.37 In terms of indexation, the countries need to move toward inflation indexation as quickly as possible. As the economies recover, most of the countries have experienced relatively rapid wage growth which is then translated into rapid growth of pension expenditures through formulas which continue to include wage growth in the adjustment of pensions already in progress. The inclusion of wage growth helps pensioners share in the current wage growth, but is simply not affordable given the demographics and the large number of pensioners relative to contributors. While indexation post-retirement should be linked to inflation only, those countries whose benefits are expected to fall rapidly have chosen to link other parameters of the pension system to the post-retirement indexation. It is the linkage of the other parameters to post-retirement indexation which causes the sharp fall in benefits. In the case of Albania, the maximum pension is indexed to inflation which causes future benefits to fall. In the cases of Serbia and of Montenegro, the value of the general point is linked to post-retirement indexation which results in 25 freezing real pension values at their 2009 levels for all eternity in the case of Serbia. The consequences in Montenegro are not as drastic, but similar in nature. 2.38 Finally, special care needs to be taken to avoid creating compliance disincentives in the pension system and to dismantling those that already exist. In middle income economies, employers and workers always have the option of informal labor markets. Access to pensions and health insurance can be seen as an advantage of formality. But if the systems are designed to finance existing pensioners only and not to provide reasonable benefits for future workers, formality will be seen as penalizing workers and employers rather than providing an advantage. 2.39 Funded pensions. In the longer run, countries may want to consider moving part of their mandatory contributions toward a funded system, where contributions are invested in individual accounts and earn a certain rate of return. Croatia, Kosovo and Macedonia have introduced second pillar funded pensions, as have many countries in the broader region (Poland, Hungary, the Slovak Republic, Bulgaria, Estonia, Lithuania, Latvia, and Kazakhstan). While a long term move in this direction will be beneficial for these countries, in the short and medium run, they are constrained by transition costs. 2.40 One approach could be gradual. As benefits fall and fiscal space appears, a portion of the contributions could be diverted toward funded pensions. These funded pensions would then provide a supplement to the future reduced level of public pensions. However, if policymakers want to pursue this type of strategy, they need to resist political pressure to counteract the falling level of public pensions. Policymakers need a clear long run strategy and need to articulate this to the population at large. 2.41 Key policy decisions when introducing funded pensions need to consider whether the funded schemes are: (i) publicly or privately managed, and (ii) mandatory or voluntary. For example, in the U.S., plans provided under the 401K section are voluntary, privately managed, offered by companies that make additional contributions to those made by the worker, and both employees and employers benefit from tax breaks. On the public vs. private split, there is some evidence that government management of funded schemes has produced returns that are inferior to privately managed funds, largely reflecting the fact that funded publicly-managed schemes tend to invest based on objectives in addition to those related to pension provision.15 The discussion in the next two sections assumes choices are along the line of privately managed and mandatory funded pensions. 2.42 Even if additional fiscal resources were available to support the pension system, the countries often do not have sufficient secure financial market instruments in which to invest the pension contributions currently. As a result, Kosovo has mandated that all of the savings are to be invested abroad. Finally, sufficient supervisory and regulatory capacity to properly secure the long term savings does not yet exist in some cases. The limited regulatory capacity has first focused on bank supervision and regulation and is gradually moving to other financial instruments. All of these factors suggest that a move to funded pensions needs to be considered as a long-run strategy rather than an immediate measure, but careful country-specific analysis will be necessary to determine the appropriate policy in each case. 2.43 In fact, the future decline in pension levels relative to average wages in most of the countries, aside from Macedonia, provides the opportunity to achieve transition to a funded system at relatively minor costs. As contribution revenues continue to rise with wage growth, but pension expenditures remain relatively constant in real terms, the contributory pension could be transformed into a social pension or a flat supplement to one’s own savings and the bulk of the contributory system moved toward a funded system. The extent of the contribution devoted to funded accounts would depend on the country’s own philosophy regarding private vs. public sector responsibility, with a total contribution to 15 For a detailed discussion, see www.worldbank.org/pensions. 26 the funded system signifying full reliance on the private sector while a small shift in contribution signifies diversification with major reliance on the government. The extent to which this decline in publicly provided pensions should be allowed to occur depends on the country’s choice of public-private mix. But governments need to make this choice to avoid short run political considerations from moving away from what would be the optimal long run combination for the country. III. REFORMING SOCIAL WELFARE BENEFITS 2.44 Social welfare programs have multiple, and often competing goals: to combat poverty through provision of benefits to people whose resources fall short of socially recognized needs standards; to support economic efficiency, including work incentives; to maintain fairness between recipients and non- recipients of assistance; and to keep program costs affordable. Balancing benefit adequacy, work incentives, equity, and program affordability unavoidably involves tradeoffs. Efficiency and equity objectives are not always synonymous, and adequate coverage for vulnerable groups can imply unsustainably high fiscal costs. The challenge for governments is to find a balance among various objectives, while keeping the fiscal costs under control. 2.45 No SEE country has yet managed to comprehensively reform social welfare policy and make it well-targeted, coherent, abuse-free, and fiscally sustainable. Although social benefits absorb a far smaller share of national GDP than pensions, there is a lot of room to increase cost-efficiency and targeting of social welfare programs. 2.46 Social welfare programs in all SEE countries are sub-optimally targeted (as are such programs in countries with similar levels of income per capita in many parts of the world) both in terms of their poverty alleviation impact and cost-effectiveness (with respect to the mix of programs and allocation of resources between them). Almost all countries have programs targeted to the poor that under-perform either due to insufficient funding or loose eligibility criteria. In Croatia, social welfare programs are estimated to cover only ten percent of the poverty gap of the lowest income decile. However, targeting is questionable: only half of the spending reaches the bottom income quintile of the population, and half of social beneficiaries are able-bodied unemployed. The social welfare program (Ndhime Ekonomike) in Albania has the best poverty targeting in the region, but even that could be improved. Coverage has improved from 25 percent in 2002 of all poor to 32.5 percent in 2005, but the extreme poor accounted for only 12 percent of beneficiaries. Moreover, in 2005, about a 64 percent of those receiving benefits were not poor, compared with 57 percent in 2002. This outcome most likely has reflected slower recertification process that has not matched the process of households moving out of poverty in Albania. 2.47 Another common problem in SEE social welfare systems is “double-dipping”-- that is, a household or an individual receiving benefits from more than one source for the same purpose. Fragmented and weak monitoring and evaluation systems, and eligibility criteria make this possible. No provisions are made for families receiving veteran benefits to be excluded from other social assistance programs. 2.48 Further, all SEE countries have numerous parallel social welfare programs, which has resulted in an overly complicated social safety net that covers narrow risks, and duplicates coverage or leaves some groups uncovered. For example, about eight percent of the population in Serbia and Montenegro receives more than one social transfer: most commonly a pension and a social assistance payment.16 16 Serbia and Montenegro Poverty Assessment, 2003, The World Bank. 27 2.49 At the same time, about 22 percent of the poor do not receive any social welfare transfers. This is partly due to the multiple objectives, target groups, targeting mechanisms, source of financing, and implementing agencies involved in the social safety net administration. 2.50 In all SEE countries, social welfare programs need re-assessment and consolidation to ensure their affordability and value for money, and a reallocation of resources within social welfare programs is needed to enhance the targeting. Resources would need to be allocated to specific poverty-targeted benefits aimed at the lowest income group, as opposed to “categorical” benefits. Since the locality of residence is strongly associated with poverty, programs aimed at poor communities would be the most efficient way to improve targeting. There is also a need to re-assess the balance between cash benefits and services to ensure the allocation reflects community needs and priorities. 2.51 While consolidation and reallocation would be a step forward, it alone may not be enough. In the case of Serbia and Montenegro, for example, the exclusion errors are partly due to limited public information about the existence of some programs. Governance of all social assistance programs would need strengthening, including improving coordination between the programs and development of a sound monitoring and evaluation system. 2.52 War veteran benefits also suffer from targeting problems. In BH, for example, where spending on these programs amounts to about 2.8 percent of GDP, benefits are not based on need; benefits are awarded on the basis of functional disability rather than inability to work and the system is overwhelmed by beneficiaries with relatively low levels of disability. All in all, poorly targeted war veteran support has a tendency to distort labor markets and crowd out other social welfare spending. Croatia is a case in point. These programs could benefit from refocusing on the most vulnerable with an ambitious effort to prevent “double dipping”, all politically challenging recommendations. In addition, the authorities should consider shifting from simple cash benefits to conditional transfers and an effort to support veterans’ reintegration into economic life. Such an approach could mitigate the political pressures that are likely to arise if these benefits were just reduced. IV. REFORMING LABOR-MARKET SUPPORT PROGRAMS 2.53 Unemployment is a major challenge in all SEE countries and labor market reforms are called for to increase labor market efficiency. While reforms of the employment protection legislation and the wage bargaining system are considered to be key for improving labor market performance social protection policies aim to mitigate the adverse impact of unemployment through select labor market support programs.17 2.54 The labor market support programs in SEE tend to focus heavily on unemployment benefits. All SEE governments, except Kosovo, provide such benefits.18 Although those benefits have not generally been found to cause rigidities in the labor market or to distort the labor supply, how these benefits impact the labor market hinges on their design. For example, in Serbia and Montenegro the current structure of unemployment benefits is ill-suited to the needs of the economy. The key shortcomings with the current benefit structure are: (i) overly generous size and duration of benefits by regional standards; (ii) limited guidance on the implementation of cost-effective active labor market policies; and (iii) no incentives for re-entry to the labor market. In the Federation of BH, a heavy emphasis on active labor market programs has had little impact on improving job prospects for the unemployed as these programs, until recently, 17 Rutkowski, Jan. Does Strict Employment Protection Discourage Job Creation? Evidence from Croatia. 18 Benefit depends on the average wage the recipient received prior to termination of employment. 28 focused on providing subsidized credits to enterprises, as well as wage subsidies, overall contributing to supporting unprofitable enterprises and practices that destroyed value. 2.55 The growth of informal economies in SEE has created a sharp discrepancy between real and registered unemployment rates and resulted in fraudulent claims for unemployment benefits, increasing the cost of these programs. Registered unemployment in BH is higher than 45 percent, while results from the first Labor Force Survey (LFS) in 2006 recorded a rate of 31 percent. The situation is similar in the other countries, with the gap substantially smaller in Croatia, Serbia and Slovenia. In BH and Macedonia, for example, those that are registered as unemployed automatically receive health insurance coverage. As a consequence, people employed in the informal sector register as unemployed in order to gain access to unemployment benefits, social assistance, and health insurance. In Albania, to reduce fraudulent claims, the government launched work-testing pilots and published the lists of families receiving social welfare benefits with good results. 2.56 Kosovo presents an interesting case of short-term alternatives to unemployment benefits. Preliminary estimates show that the introduction of a meaningful unemployment benefit in Kosovo would exceed the country’s fiscal capacity, given the high level of unemployment. Moreover, it would necessitate increasing payroll taxation, which would impair the flexibility of the labor market, and likely push some firms from the formal to the informal economy. Therefore, instead of rolling out unemployment benefits, Kosovo is considering of strengthening the poverty benefit, which the analysis indicates would be a cost-efficient short-term measure to mitigate adverse impacts of unemployment and cover the unemployed poor. However, in the longer run other solutions would need to be developed. 2.57 In all SEE countries, there would, however, be a need to encourage the development of labor market programs that would cost-effectively address the high level of unemployment. This could include: (i) strengthening the advisory services through public and private networks of providers; (ii) strengthening the focus on the active labor policies; (iii) providing special assistance to selected target groups and promoting social development partnerships at the local level; and (iv) intensifying activities that aim at reducing the informal sector. V. CONCLUSIONS AND RECOMMENDATIONS 2.58 Reforming social protection systems is important to help ease the fiscal burden while providing adequate income in old age and a reasonable level of protection for the most vulnerable members of society. SEE countries advanced reforms in this area, but a lot remains to be done to improve the effectiveness of social protection spending in many areas, particularly by retooling targeting and reducing the large number of programs. 2.59 Geographic and ethnic characteristics of SEE have generated complex social and economic problems that have contributed to keeping the pace of reforms relatively slow compared with the best reformers in emerging Europe. Instead of a radical approach, SEE countries have often chosen to pursue a piece-meal reform of their social welfare systems, and countries that have undertaken comprehensive reforms did so only as the last resort after the piece-meal reforms failed. The hesitance for more radical reforms can be attributed partly to other urgent priorities, such as nation-building and post-conflict recovery, and partly to the presence of politically strong vested interests (such as pensioners and war veterans) in maintaining the status quo. Regardless, SEE governments have been less successful than their CEE counterparts in explaining the rationale for social protection reforms to the public, which is an area needing attention. 29 2.60 Looking forward, a more comprehensive reform of social protection programs would be needed in SEE to improve the performance and fiscal sustainability of programs. This will require measures both on the revenue and expenditure side. 2.61 On the revenue side, the efforts would need to concentrate on improving compliance and collection of social contributions, both through improvements in record-keeping and the administration of the collection processes and through intensified enforcement efforts. As mentioned earlier, payroll taxes in SEE are already high, and further increases would not be advisable. By contrast, to stimulate formal private sector activity and economic growth, decreasing the tax wedge would be recommended. However, given the already high level of social protection spending, without either reduction in expenditures and increased collection of contributions, tax cuts are not feasible. Additionally, specific measures to consider include: • Raise the retirement age, equalize the retirement age for men and women • Reduce the weight of wage increase in indexation and move toward a full inflation indexation • Reduce exemptions from paying contributions. • Bring accrual rates to more reasonable levels • Consider carefully pension design and options for reducing pension contribution rates 2.62 On the expenditure side, there is significant room to improve cost-effectiveness of social protection programs. Coordination, targeting, eligibility conditions, and finance allocation for existing social welfare programs could be improved to streamline and reduce expenditures. This could be done for example by reducing administrative expenditures of programs. Coordination and consolidation of programs, and reallocation of funds between them would also be needed to reduce fragmentation, overlaps, and waste of resources. Improving the targeting of programs is one of the priorities. This could be achieved by revising the eligibility criteria so that the targeting of benefits is improved. 2.63 Efforts on the revenue and expenditure side for the PAYGO pillar should be supplemented by efforts on the regulation of the funded pillars where these have been introduced. In particular, regulators should assess whether fees levied on second-pillar funds are not too high and work with fund managers to bring them down. 30 Health 3.1 The fiscal sustainability and effectiveness of public spending on health merits examination since significant resources are allocated to health care in all SEE countries, but these resources are not used as efficiently as they could be. In the years ahead, moreover, SEE countries are likely to face the additional pressures of aging populations, an increased prevalence of non-communicable diseases that are expensive to diagnose and treat, and growing provider and consumer expectations of better quality care and expensive medical technology. 3.2 Improving the performance of the health sector will require ambitious government actions to improve both the efficiency and equity of spending. Measures will be needed to strengthen both the allocative efficiency of public funds (that is, using public resources for achieving the government goals in healthcare) and the technical efficiency (maximizing the value for money allocated). The chapter recommends that the authorities consider several priority avenues for reform. These include the need to strengthen contribution collection, including by sharply reducing exemptions from paying contributions and co-payments, the need to revise and substantially tighten the benefits package for social health insurance, and tighten procurement practices, especially for pharmaceuticals. Recommendations also include the need to adopt modern provider payment and performance management systems. I. THE INSTITUTIONAL SETUP 3.3 Health Insurance and Financing: All SEE countries, except, in practice, Albania,19 have publicly financed and administered health insurance programs. These programs are based on the concept of social insurance in which citizens are expected to contribute according to their ability to pay, and receive basic health services.20 The health financing systems of the countries of the former Yugoslavia are similar to those in most EU15 countries, but the institutional framework for delivering health insurance is different and more complex. This section will focus on the countries of the former Yugoslavia and not on Albania. 3.4 There is a purchaser-provider split in SEE health systems, and health insurance funds (acting as the purchaser) contract with public and private providers to deliver health services (the public-contract model). Unlike countries such as Germany and the Netherlands, which have a multi-payer system with public and private health insurance companies competing for the provision of health insurance, the single- payer system21 predominates in SEE. Extra budgetary health insurance funds (HIFs) are responsible for overseeing and implementing compulsory and voluntary health insurance and about 90-95 percent of all public financing for health flows through these funds. The private health insurance market is almost non- 19 In Albania, the social insurance and purchaser–provider split model is still evolving. The Ministry of Health exercises considerable control as a financier and provider of health services. Health services are directly financed through the budget based on inputs (e.g. salaries), local governments administer primary health care, and the health insurance fund (a quasi-autonomous public agency) is slowly being granted control over health finances and contracting. 20 Social insurance is not a new concept in SEE. The health system of SFRY was unique in Eastern Europe because it was historically funded from compulsory social insurance contributions rather than the budget. 21 Even if the health insurance system is entirely financed from public funds, there may be a single-payer or multiple payers in the system. Single-payer systems have many advantages since the administrative costs of managing health insurance is low and the purchasing power of health insurance funds is strengthened vis-à-vis providers. 31 existent in SEE (with the exception of Serbia). In those countries where supplementary health insurance is available, it is provided by the public health insurance fund. 3.5 Sources of Health Financing: In all SEE, except in Albania and Kosovo, social insurance contributions (by employers, employees, the self-employed, pensioners, the unemployed, and other groups) are the main source of health financing (see Table 3.1). While contribution rates vary significantly across categories of insured people and countries, they tend to be highest for salaried people. Contributions for vulnerable groups and funding for special programs (such as public health programs) are made through the general budget. Out-of-pocket payments, either in the form of co-payments for services included in the benefit package or for services that lie outside the benefit package or for people not covered by insurance, constitute a large share of total health expenditure in some countries, such as Albania and Bosnia and Herzegovina, leaving households particularly vulnerable to health shocks. Table 0.1: Sources of Healthcare Financing, 2005 (In percent of total) Public Private General Out-of-pocket Private SHI revenues payments insurance NGOs Albania 10.8 32.7 56.4 0.0 0.1 Bosnia and Herzegovina 46.6 2.1 51.2 0.0 0.0 Croatia 71.9 9.5 17.5 1.1 0.0 Macedonia, FYR of 67.9 2.7 29.4 0.0 0.0 Serbia and Montenegro 52.6 16.9 27.6 2.9 0.0 Slovenia 70.4 6.6 10.0 13.0 0.0 Note: The data for Kosovo are from 2004 and figures do not show the 2.1% of donor funding received in that year. Sources: WHO NHA database, World Bank Public Expenditure Reviews. 3.6 Health Delivery Arrangements: Health care delivery systems in the SEE countries are less diverse than in EU15, where the provision of health services is largely private.22 The systems in SEE consist of an extensive network of public health facilities: ambulantas (small primary health care facilities scattered throughout the country), public pharmacies, dom zdravlja (larger primary health care centers with diagnostic facilities and specialists), hospitals, clinical centers and tertiary health facilities typically located in larger cities. 3.7 Private practice is permitted by law, and private pharmacies, dental practices, and primary care health and specialist doctors are becoming common. Some SEE countries (for example, Macedonia) have passed laws to entirely privatize specific sectors (such as pharmacies, dental practices, and primary care). In Croatia, a large percentage of medical doctors are private, but in rural areas the population is largely dependent on the public sector. II. HEALTH EXPENDITURES 3.8 Overall health spending in SEE is high relative to GDP compared to countries with similar levels of income per capita. In 2004, total spending on health in SEE was on par with that of many EU members states, and in most cases exceeded the EU12 average (Table 3.2). However, with the exception of Slovenia, even high expenditures relative to GDP still translate into fairly small per capita amounts in 22 An exception among EU countries is the UK which as a public integrated model, where the government is the financier and provider of health services. Many changes have been introduced in the UK system in the past decade (including a purchaser-provider split and autonomous providers). 32 absolute terms, even when PPP-adjusted (for Croatia, health spending in dollars per capita is only one- half of the EU15 average, and in the remaining countries is about one-fourth of the EU15 average). Table 0.2: Health Expenditure in SEE, 2004 (In percent of GDP unless indicated otherwise) Total Public Private Per capita Per capita US$ $PPP Albania 6.7 3 3.8 157 409 BH 8.3 4.1 4.2 198 359 Croatia 7.7 6.2 2.1 609 897 Kosovo 7.8 3.9 2.9 96 -- Macedonia 8 5.7 2.3 212 411 Montenegro 6.6 6.3 2.4 … … Serbia 10 7.3 3.1 … … Slovenia 8.7 6.6 2.1 1,438 1,760 EU12 average 6.8 4.9 1.9 461 871 EU15 average 8.6 6.4 2.2 2,045 1,980 Sources: WDI Database, 2007 for total, public, private and current US$ figures, and WHO HAF-DB 2007 for PPP$ figures for Albania, Macedonia, Bosnia and Herzegovina, Croatia, Slovenia and EU aggregates; World Bank estimates for Serbia and Kosovo.. 3.9 The share of public health expenditure tends to be lower than in Table 0.3: Health Spending as Percentage of Total comparator countries, however, suggesting Government Expenditures that private out-of-pocket expenditures 1998 1999 2000 2001 2002 account for a much higher share of total expenditures (Table 3.3). Moreover, Albania 7.6 9.2 9.8 7.3 -- BH -- -- 14.0 13.0 13.2 available data on private spending typically Croatia 13.7 14.9 16.0 15.9 14.4 underestimates health expenditures since Macedonia 20.4 16.2 15.8 13.2 14.0 they are based on official co-payments Slovenia 14.3 14.0 13.5 14.6 14.7 reported to health insurance funds.23 Poverty Note: Statistic for Kosovo was 11.5 in 2004; Serbia and Montenegro assessments based on survey results suggest averaged 14.0 between 2000 and 2002; EU15 averaged 12.9 between that patients sometimes pay three as much as 1996 and 2002 Sources: Public Expenditure and Institutional Reviews, WHO-Euro officially reported (or about 2-4 percent of data, 2000-2004, IMF Government Finance Statistics GDP). If these estimates are used, total health spending in the SEE countries would amount to 6-12 percent of GDP, with the top of the range being very high by international standards, in some cases almost as high as in the United States. 3.10 There is evidence of some spending containment. With the exception of Albania and Slovenia, the rate of increase of health expenditure has been slower than the rate of growth of the economy, resulting in a modest decline in spending relative to GDP over time. This is in contrast to developments in the EU: from 1990 to 2004, total health expenditures rose by 1.6 percent of GDP in both the EU15 and EU12. Developments in the EU are no justification to increase spending on health in SEE, however: 23 For example, the Serbia Poverty Assessment (World Bank, 2003) showed that outpatients in state institutions, including primary care clinics, pay on average YUD 617 per month for a doctor’s visit, prescriptions, and diagnostic tests. Those who visit private doctors paid on average YUD 2,158 YUD per month, while those admitted to hospitals paid on average YUD 9,752 for the hospital care, including drugs, diagnostic tests, and procedures. These sums are higher than official co-payments (which amount to 1.3 percent of the revenue of the health insurance fund, according to 2003 budget estimates). Official co-payments are set at YUD 20 for a medical examination or a laboratory test, YUD 30 for an x-ray, and YUD 50 for an ultrasound. 33 government outlays on health in SEE are similar as a share of overall spending as in the EU15, despite lower income. 3.11 In many SEE countries, the health insurance fund annually incurs a sizeable deficit. For example, in Croatia and Macedonia health insurance funds have been building arrears each year as their expenditures outstrip revenues. In the past, Serbia and Macedonia tried to deal with the arrears by a set of measures that even a priori were known to be inefficient and ultimately led to further buildup of arrears: (i) delaying payments to suppliers of medical goods and services, especially pharmaceutical companies; (ii) delaying payments to health care providers; (iii) maintaining artificially low reimbursement rates to health care providers, or setting contractual payments at low levels, leading to implicit rationing of services by providers and informal payments, and thereby generating access barriers. In Croatia and Serbia too, the government has bailed out the health insurance fund, and cleared the arrears to ensure continued provision of health services. However, all these measures are temporary fixes that are not addressing the underlying causes of deficits, but rather creating additional economic problems. 3.12 Recently, countries such as Macedonia have aggressively tackled the problem of health insurance fund arrears by setting hard budget ceilings for health care institutions and implementing strict financial reporting and monitoring systems. Additional measures have included competitive procurement of pharmaceuticals which has reduced the prices of drugs on the outpatient list by 30 percent. The chronic deficit problem is being addressed through benefits package reforms and improved revenue collection. III. HEALTH OUTCOMES AND UTILIZATION OF HEALTHCARE 3.13 Despite relatively high spending, health outcomes in SEE trail those of the EU15 (Table 3.4). Although life expectancy has been steadily improving and is currently over 70 years in all SEE countries, under-five mortality in all countries (except Croatia and Slovenia) is still high compared to EU15. The standardized death rates (SDR) from all circulatory diseases, as well as from ischemic heart disease, are much higher than the EU15 average. Of communicable diseases, tuberculosis (TB) is a problem across the SEE region, but especially among internally displaced people and other vulnerable groups. Compared to the EU15 average of 10 per 100,000 in 2004, the tuberculosis incidence was 22 in Albania, 30 in Macedonia, 48 in Serbia, 48 in Kosovo and 53 in BH. The HIV/AIDS prevalence rate is, by contrast, low in all SEE countries. However, this may reflect weak surveillance systems, since many structural factors that could fuel an HIV/AIDS epidemic are present in SEE.24 3.10 Inpatient admission rates are low with the result that available resources, such as hospitals, beds and staff, are under-utilized. Annual inpatient admission rates range from 8.7 per 100 people in Albania to 12 per 100 people in Serbia, compared to 17 in Croatia, 18 in Slovenia, 17 in the EU and 21 in the EU12. With the exception of Montenegro, the number of outpatient contacts per person per year is low compared to the average in EU countries of around 7: in Albania, the average person utilizes outpatient care only about 1.5 times per year and in Bosnia and Herzegovina around 3.3 times per year. A variety of demand and supply side factors may be responsible for low utilization of health services, including the cost and quality of health services. In Bosnia and Herzegovina, in 2003, hospitalization could cost E100, which was almost half of the monthly budget of a poor household.25 In Macedonia, consumers’ perception of the quality of health sector services is poor. 24 These factors include unemployment among youth, mobile populations, human and drug trafficking. 25 Bosnia and Herzegovina Poverty Assessment, World Bank 2003. 34 Table 0.4: Health Outcomes in the SEE Countries, 1992-2005 1992 1995 2000 2001 2002 2003 2004 2005 Life expectancy at birth (years), 1992-2005 Albania 73.2 74.9 74.9 77.2 76.4 75.8 76.2 72.0 BH … … … … 74 … … … Kosovo … … … … 69 … … … Macedonia 71.3 72.2 73.4 73.8 73.3 73.5 … … Montenegro 75.0 74.9 75.2 75.3 74.0 73.8 74.2 74.2 Serbia … … 71.7 72.4 72.4 72.5 72.8 72.9 Croatia 71.8 73.3 73 74.7 74.9 74.7 75.7 75.4 Slovenia 73.7 74.8 76.3 76.5 76.7 76.5 77.3 77.6 EU 75.6 76.1 77.5 77.8 77.9 77.9 78.3 78.3 EU15 77.0 77.5 78.7 79.0 79.1 79.1 79.5 … EU12 70.8 71.2 73.0 73.3 73.4 73.5 73.9 73.9 Under-five mortality rate per 1,000 live births, 1992-2005 Albania 45.3 33.2 16.8 16.0 15.8 14.1 12.4 --- BH … 19.0 17 … 18 … 15.0 … Kosovo … … … … 69.0 … … … Macedonia 34.4 25.4 13.7 13.7 11.9 13.0 … … Montenegro 15.6 13.9 12.6 15.7 11.4 12.1 9.5 11.0 Serbia … … 12.7 11.9 11.7 10.5 9.4 9.2 Croatia 13.6 10.2 8.4 9.0 8.2 7.3 7.1 6.6 Slovenia 10.3 6.6 5.5 4.7 4.9 4.8 4.7 5.3 EU 10.7 9.0 7.1 6.9 6.6 6.5 6.3 6.3 EU15 … 8.3 6.8 5.7 5.7 5.5 5.4 5.3 EU12 18.7 16.7 12.6 12.0 11.5 10.9 10.7 10.5 Standardized death rate (SDR), diseases of circulatory system per 100,000, 1992-2005 1992 1995 2000 2001 2002 2003 2004 2005 Albania 384.4 403.0 518.9 367.6 397.9 439.9 419.4 … Croatia 500.2 488.8 572.7 486.2 482.5 499.6 419.0 438.8 Macedonia 586.3 603.8 582.2 557.3 613.6 599.1 … … Slovenia 418.8 371 315.0 298.4 290.5 295.3 277.0 288.0 EU 380.0 359.5 308.2 300.4 297.4 294.9 279.4 279.4 EU15 308.1 288.0 245.6 238.7 235.7 233.5 221.0 … EU12 629.6 610.9 533.0 523.4 523.9 522.2 496.9 498.7 Standardized death rate (SDR), ischemic heart disease, all ages per 100,000, 1992-2005 Albania 73.4 83.3 125.2 86.5 108.2 127.8 121.4 … Macedonia 107.9 106.5 109.7 110.8 113.0 112.3 … … Croatia 175.9 183.6 201.1 162.8 159.5 186.9 154.5 167.9 Slovenia 141.6 111.6 105.1 100.2 89.4 94.4 82.5 80.2 EU 143.5 137.2 119.0 115.3 112.9 111.0 105.1 105.9 EU15 124.8 117.2 97.5 93.9 92.0 90.1 85.0 … EU12 208.3 207.5 196.4 192.7 190.1 188.4 180.1 184.4 Note: SDR data are not available for Bosnia and Herzegovina, Kosovo, Serbia and Montenegro Sources: WHO HFA database 2007, PEIRs and World Bank HNP Statistics. 35 3.14 Owing to inadequate financing of maintenance and capital investments, health care institutions are in various stages of disrepair in some countries, and need basic renovations and re-equipping to provide a minimal standard of health services. For example, in Serbia, only one-third of hospitals have functioning sterilization systems, and 75 percent of the medical equipment in health care facilities is over ten years old. This has negative implications for the clinical quality of care and patient satisfaction, and suggests that intra-sector spending allocation may need adjustment. IV. HEALTH FINANCING AND EXPENDITURES POLICIES 3.15 Current health financing and expenditure policies are contributing to the health sector fiscal deficits and poor performance and will need to be addressed to improve the sustainability and effectiveness of health spending. A. Health Financing Policies 3.16 To contain already high public spending on health and to limit health fund arrears and government bailouts, the health financing system will need to be made more efficient and provide incentives for cost control. Countries relying primarily on social insurance contributions for health care financing (BH, Croatia, Macedonia, Serbia and Montenegro) face a limited revenue base for health insurance owing to the use of the payroll as a base to fund social insurance contributions, low formal employment (reflecting considerable informal activity and high unemployment), and the exemption of sizeable segments of the population from paying health insurance contributions. Contribution rates for health, moreover, are very high compared with the EU15, Bulgaria and Romania. 3.17 At the same time, under-collection of contributions from private and public entities is common, and will need to be addressed. For example, the Macedonian health insurance fund estimated that unpaid contributions amounted to 2.5 billion Macedonian denars in 2003. Because entitlement to benefits is not clearly linked to contributions in practice, it has been particularly difficult to collect contributions from the self-employed. Transfers from the pension and unemployment funds and the general budget for vulnerable groups have often been erratic, and pension and unemployment funds are in arrears to health insurance funds. Formal sector social insurance contributions (mainly from the public sector) and co- payments for services have become the only predictable source of health financing. 3.18 Recognizing the problems on the revenue side, some SEE countries have implemented or are in the process of implementing policies to strengthen the health financing system. For example, Croatia has (i) reduced the payroll tax contribution rate for health insurance (from 9 percent to 7 percent for employees) to stimulate job growth; (ii) established central and local government contributions to social health insurance; and (iii) introduced co-payments for selected services (higher rates for hospital and specialist services, diagnostic services and prescription drugs) to control demand. 3.19 Unfortunately, these policies have not yet had the desired impact on the fiscal sustainability of the HIF, partly because voluntary supplementary health insurance (offered by the health insurance fund) was often introduced at the same time to cover the co-payments. This re-introduced the problem of excessive health service use which co-payments were intended to mitigate in the first place. It also increased the administrative costs of the health insurance fund, and created a problem of adverse selection with fiscal consequences since supplementary health insurance is more likely to be purchased by those beneficiaries with the highest medical costs. Also, a large share of the population is exempted from co-payments. Slovenia, for example, has voluntary supplementary health insurance covering co-payments, and, like Croatia, is currently thinking of reforming it since, instead of helping, it has ended up exacerbating the fiscal problems of the health sector. 36 Table 0.5: Social Insurance Contribution Rates for Health Year Salaried Self-employed Non-active and contribution- started (Employer: exempt groups Employee share) Albania 1995 3.4%of payroll 3-7% of statutory minimum Central budget (1.7:1.7) wage (urban or rural) Bosnia and Herzegovina Federation 1997 17 % payroll 15% of cadastre revenue Central budget, pensions and (4:13) unemployment funds RS 1999 15 % 15% of cadastre revenue Central budget, pensions and (0:15) unemployment funds Macedonia 1991 9.2% payroll 9.2% of income Central budget (9.2; 0) Montenegro 1993 15% payroll 13.5% of main wage Pensioners: 19% of net (7.5:7.5) pension. Unemployed: 9.7% of unemployment benefit, but in practice 7.5% Others: central budget Serbia 1992 15.92% 14.4% of net wage Pensioners: 12.3% of net payroll Farmers: 4% of property tax pensions. Others: central (7.95:7.97) budget, equal to 15.95% of average wages. Croatia 1993 15% payroll 18% of income 18% of gross pension and (15%:0) Farmers: 15% of income if in other benefits plus central VAT system, or 7.5% of budget and county budgets income estimated based on land ownership Slovenia 1993 13.25% 13.25% of income Central budget. payroll Sources: Health Insurance Funds B. Expenditure Policies 3.20 On the expenditure side, a number of factors are contributing to high health costs and limited effectiveness of health expenditures. First, health care delivery in SEE countries is still largely dominated by specialist and hospital-based care with weak preventive and primary health care (PHC). This is reflected in the expenditure patterns of health insurance funds (Table 3.6). In addition, the health care delivery network is fragmented, which leads to inefficiencies in the utilization of health services. In the EU, it is increasingly recognized that controlling costs without sacrificing access and quality of care requires better integration of primary care with specialist services and improved coordination across different levels of care. Many countries have established gatekeeper functions, which manage the use of specialist services. For example, in Slovenia, government’s reform efforts focus on strengthening primary care and rationalizing secondary and tertiary care. 37 Table 0.6: Health Service Delivery Data from SEE countries, 2005 Bed occupancy rates Average Length Hospitals PHC units Hospital beds (In percent, acute care of Stay per per 100,000 per 100,000 hospitals) (In days) 100,000 people people people Albania … 6.4 1.6 76.55 297 BH … 9.6 0.95 30.28 304 Serbia 69 10.6 1.38 … 593 Montenegro 69.7 9.8 1.93 29.52 418 Macedonia, FYR 55.5 11.1 2.65 91.05 470 Croatia 88.1 10.3 1.76 74.54 545 Slovenia 70.1 7.1 1.45 3.2 483 EU 75.9 9.3 3.04 … 585 EU15 … 9.6 3.19 … 571 EU12 74 8.03 2.63 62.22 644 Note: Data are for 2005, except for the EU-15 data which are for 2004 Source: WHO HFA-DB, 2007 Table 0.7: Health Insurance Fund Expenditures by Level of Care (percentage), 2002 Primary Secondary Prescription Capital and Tertiary Drugs Investments Bosnia and Herzegovina Federation 31 38 7 5 RS 27 51 7 2 Croatia 16 45 13 … Macedonia 37 50 13 .36 Montenegro 30 39 12 4 Serbia 36 26 12 .30 EU15 … 38 … … Source: Health in Transition Reports, PEIRs, OECD Health Data The data on the EU15 refers to expenditures on in-patient care. 3.21 While spending on hospitals is high, the productivity of hospitals in all SEE countries (except Slovenia) is low compared to the EU15 (Table 3.7). Hospital occupancy rates (except in Croatia and Slovenia) are lower than in the EU15, and there is excess capacity in terms of the number of beds and staff. For example, a hospital planning study in Serbia showed that 17,000 beds could be eliminated without any negative impact on the access and quality of health services. As will be discussed below, the low productivity of hospitals is linked to the type of payment mechanism for hospitals (Table 3.8), which is input-based (salaries, utilities, consumables) and not linked to outputs. 3.22 Furthermore, health benefit packages are generous. This is a SFRY legacy. In SFRY, health care was a constitutionally protected entitlement of all citizens and benefits were comprehensive, including health-related benefits (that also covered treatment abroad) and generous non-health benefits such as funeral expenses and sick leave. For example, in Croatia, after the first 42 days of sick leave, beneficiaries are entitled to up to one year of sick leave paid by the health insurance fund. There is also anecdotal evidence from Croatia and Macedonia of widespread misuse of sick leave to temporarily hide redundant employees. In Serbia, the 2005 Health Insurance Law removed plastic surgery, funeral expenses and dentistry for adults of the benefits package as a first step to introducing a more affordable set of essential services. 3.23 A large share of total expenditure is on drugs. The reasons for high drug pending include: (i) non-competitive pharmaceutical procurement; (ii) absence of reference pricing for drugs; (ii) no generic 38 substitution policies; (iii) limited supply-side regulation of the pharmaceutical industry through price- 26 fixing; and (iv) limited prescription control. For example, a study in Macedonia found that the health insurance fund was paying 50 percent more for drugs than the neighboring countries. In Serbia, a study found that the health insurance fund could have saved 25 percent if drugs had been procured using competitive procurement practices. 3.24 Current systems of paying providers, and especially hospitals, neither encourage cost containment nor provide incentives for better performance. In SEE, line budgeting is still the most common provider payment mechanism for hospital care. In this system, the health insurance fund pays providers according to line items, such as wages and salaries, utilities, medicines and other supplies, on the basis of an annual contract negotiated between the health insurance fund and each provider. Because the budget in one year is largely determined by the costs of inputs in the previous year, hospital management has an incentive to utilize the entire budget each year rather than to contain expenditures. In addition, since the budget is related to historical patterns, and not to the number and the severity of the cases treated, there is no incentive to treat difficult cases, but rather an incentive to refer these cases to higher levels of care. This means that primary care providers do not act as effective gate-keepers to higher levels of care. Table 0.8: Paying Health Care Providers Primary care Outpatient specialist care Inpatient care Albania Capitation Salary based on national scale Budgets based on staff salaries and other inputs BH HIF allocations based on staff salaries and other inputs Croatia Capitation Fee for service based on point system Global budgets with point system for paying providers Macedonia Capitation Fee for service based on point system HIF allocations based on staff salaries and other inputs Serbia and Capitation Fee for service HIF allocations based on staff Montenegro salaries and other inputs Slovenia Capitation + fee Fee-for-service based on point Global budget + DRGs- hospitals for service system have limited autonomy over allocating fixed and variable costs Sources: World Bank Public Expenditure Reviews, Health Financing Studies and European Observatory Health in Transition Reports. 26 In Macedonia and Serbia, if a public pharmacy is unable to fulfill a prescription, patients can have the prescription filled by a private pharmacy and obtain a re-fund from the health insurance fund. Since health insurance funds procure drugs for supply to public pharmacies, this duplicates costs. 39 Box 0.1: Improving Governance in the Health Sector in Macedonia Various studies conducted in Macedonia highlight severe governance failures in the health sector. For example, despite high public spending on the sector (5-6 percent of GDP), citizens pay substantial out-of-pocket payments, mostly on an informal basis. A few years ago, prior to the introduction of competitive tendering, the Health Insurance Fund (HIF) was paying 30 percent higher prices for drugs than neighboring countries. The public generally views the health sector as corrupt and considers its services to be poor. Poor governance in the health sector results in limited value for money for health services and a lack of confidence of the public in the health sector. Addressing governance in the health sector has been a key emphasis of World Bank lending and non- lending services during the previous CAS period. Encouraged by the Bank’s dialogue, and as part of the first Development Policy Loan, the HIF initiated competitive procedures for purchasing of drugs for the first time in 2005. As a result of this intervention, many foreign manufacturers were registered to sell drugs in the Macedonian market. All the newly-registered manufacturers are GMP (Good Manufacturing Practices) certified. This has contributed to the HIF purchasing better quality drugs as well as increasing competition in the health market. As a result of competitive procurement, the HIF reduced expenditures on outpatient drugs by 30 percent. Cost savings were the greatest for those drugs for which multiple suppliers competed, as compared to drugs for which there were single suppliers. A key focus of the World Bank Health Sector Management project as well as Development Policy Loans is to improve transparency and accountability of the HIF and health care institutions (HCIs). The HIF Board, an important mechanism for improving governance of the HIF, has undergone reforms. The number of board members was reduced, selection of Board members by the Parliament was stopped, Board members were required to sign conflict of interest clauses and the number of doctors on the HIF Board was balanced by members of other groups. The minutes of Board meetings are now published on the HIF website, increasing transparency and accountability to the public. Formal training of Board members is also being supported through the Bank-financed project. The HIF now requires HCIs to regularly produce monthly financial reports, which are also posted on the HIF website, and HCI managers have also been trained in budget and general management. The plans are now for HIF to monitor key performance indicators (e.g. admission rate, referral rate, patient satisfaction) among HCIs and also report these on the HIF and MOH website. The current government is committed to addressing poor governance in the health sector. Recently, the MoH and the HIF adopted changes in the legal and regulatory framework aimed at improving the governance in the health sector and, to that end, have endorsed the “four eyes” principle, by introducing two directors for health care institutions, with a view to improving control and accountability of the HCIs’ management structures. Provisions prohibiting conflict of interest situations in the health sector have been systematically implemented in recently amended laws and regulations governing this sector. Under the upcoming CPS period, there are plans to continue to strengthen governance in the health sector by aggressively addressing the problem of informal payments, changing the legal framework for health care institutions so that they are more accountable to MOH, HIF and the public for results, as well as continue to build the capacity of HIF as a transparent purchaser of health services. The Ministry of Health and other institutions’ (e.g. Drug Bureau) capacity for regulation and system oversight will also be strengthened 3.25 Governance of the HIFs and the health sectors is weak across SEE and across all levels of the health system. In all SEE countries, the capacities of ministries in policy-formulation, priority-setting, regulation, and monitoring and evaluation of the sector need to be strengthened substantially, given the implications of these functions for the efficiency and effectiveness of health spending. In most SEE countries, health insurance funds are quasi-autonomous extra-budgetary funds that enjoy considerable freedom in the management of health sector funds. Collection and analysis of information, and its use to promote providers’ compliance with contracts is weak. Overall, financial management systems are often neither well-developed nor synchronized with the overall public financial management systems. To address this problem, several countries have moved or are moving health insurance funds into a single Treasury account. In 2002, the Croatian health insurance fund was consolidated under the Treasury 40 account in order to improve fiscal discipline and debt management. The health insurance fund in Macedonia has also been brought into the Treasury, and similar measures have been taken in Serbia and Montenegro. V. CONCLUSIONS AND RECOMMENDATIONS 3.26 The countries of SEE share common problems in the health sector, which will need to be addressed to improve the fiscal sustainability of health spending.27 Given the already high levels of payroll contributions for health, there is no scope to increase payroll contributions further. In fact, SEE countries will need to reduce payroll tax contributions to stimulate job creation in the formal sector. Hence, the reform measures will need to focus on strengthening contribution collections and cost- recovery. Priorities on the health financing side should include the following: • Improve administrative efficiency in contribution collection: SEE countries will need to reduce the administrative complexity of contribution collection (e.g. social contributions collected by multiple agencies and the complicated transfers of funds between agencies), and strengthen the capacities of health insurance funds to effectively purchase and manage health expenditures. The Slovenian health insurance fund has centralized contribution collection in the Public Revenue Office, and Croatia has centralized its contribution collection (for health and social programs). Centralized systems have also been introduced in Serbia and Macedonia. These changes have helped to reduce fraud and evasion, and significantly improve collections. • Reduce contribution exemptions: To enhance revenue collection, there is a need to reduce exemptions, and improve the collection of contributions from the self-employed and farmers. Linking the contribution rate to the ability to pay is also needed. An evaluation of exemption categories will need to be done as part of an overall review of social benefits. • Expand co-payments: International experience shows that co-payments can help reduce the unnecessary use of services (especially expensive outpatient specialist care and pharmaceuticals), and increase the use of primary health care while reducing the burden on hospital care. SEE countries would particularly need to reduce the generous exemptions from making co-payments, and introduce co-payments for specific categories of medical services and products, while ensuring that vulnerable groups (such as the poor and people with chronic conditions) are protected. 3.27 A series of changes are also needed on the health expenditure side to address the intra-sector spending allocation and to help improve the efficiency and quality of the health system. Among these, the authorities are invited to consider the following set of measures: • Strengthen payment mechanisms to improve efficiency of resource use: Given that hospitals are a major cost driver in the health system, there is a need to implement payment mechanisms that provide incentives for cost control. The monitoring of hospital performance 27 The challenges that Albania faces with respect to health financing are slightly different. Given the level of GDP per capita, and large informal and agricultural sectors, social insurance contributions are not expected to become the primary source of health financing in the medium-term. While there is a need to develop realistic targets for public health spending in the government’s medium-term budget framework, emphasis needs to be on improving the effectiveness of health spending through better targeting on priority programs and vulnerable groups, and implementing payment mechanisms that encourage efficiency, transparency and accountability in resource use. 41 based on a few indicators of financial management, clinical quality, and patient satisfaction will also be needed. One option would be to implement global budgets for hospitals. While global budgets are a good first step in imposing cost controls, they provide little incentive for hospitals to improve the efficiency of their operations. A complex alternative defines prices and volumes of care (case-based payment or diagnostic-related group/DRGs). Some SEE countries, such as Croatia, are already implementing a version of case-based payments, but these efforts should be strengthened and expanded In other SEE countries, DRGs could be implemented when enabling conditions are in place (such as hospital cost data, health information systems, health insurance fund and health care institution capacity in coding and using codes for billing). Implementation of case-based payments would need to be combined with hospital autonomy over budget allocation and staffing, and profit retention. Payment mechanisms for other care providers could also be strengthened; for example, group practice could be encouraged for primary care. • Rationalize the health benefit packages: Reforming the benefit package is a politically difficult exercise, but could be launched by removing or reducing the scope of non-health benefits. Benefits to consider for removal or rationalization include funeral, transportation, sick leave and maternity leave benefits. Internal controls that prevent the misuse of non- health benefits would also be needed. Concurrent with a review of the existing benefit package, voluntary health insurance could be introduced. • Initiate reforms of pharmaceutical procurement: To control drug expenditures, SEE countries would benefit from purchasing drugs through international competitive tenders, or at least introducing reference pricing or strengthening the existing reference pricing system, as well as monitoring and regulation the positive list of drugs covered by health insurance funds, and streamlining the system of drug registration and marketing authorization. For example, Macedonia introduced international competitive tenders for a selected list of drugs and lowered its drug expenditures after it was found that the health insurance fund was paying 50 percent more for drugs than neighboring countries. Over the medium to long run, allowing generic substitution of drugs by pharmacists, and strengthening rational drug prescribing and utilization is recommended. Also, a regional pharmaceutical purchasing group could be formed with neighboring countries to improve purchasing power. • Strengthen the public health and health promotion system: Given the disease burden in SEE countries (associated with lifestyle choices such as smoking, alcohol consumption, and diet), greater emphasis is needed on public health and health promotion systems. For example, the implementation of anti-smoking programs in Croatia contributed to a significant decrease in the prevalence of smoking, especially among the youth. • Strengthen health sector capacity to implement reforms: To successfully implement all the above reforms, the capacities of ministries of health in policy formulation, monitoring and evaluation need to be built. Health insurance funds need to improve their capacity for economic evaluation, quality of care review, actuarial analysis, and financial management. Health care institutions’ capacity in planning and management also require attention. 42 EDUCATION 4. 1 Education indicators in most SEE countries suggest that, on average, the stock and quality of human capital, a key determinant of economic growth and living standards, lags most countries in the EU, including the new member states. Despite progress on some measures in some countries over the last several years, notably in educational attainment and enrollment, deeper reforms are needed to improve the effectiveness of education policies and spending and enable students to graduate prepared for the challenges of today’s globalized and increasingly knowledge-driven economies. Persistent problems and new challenges that should be tackled include the need to ensure equitable access, improve overall coverage (especially in secondary and higher education), and strengthen the relevance of non-compulsory education. 4. 2 The education systems in SEE harbor structural inconsistencies and inefficiencies caused by inadequate government responses to emerging social and economic challenges. Most importantly, systems have not adjusted to demographic changes (including shrinking student-age populations), migration, and evolving ethnic structures. Further, rigid allocation schemes and ill-defined (or often missing) decentralization frameworks have created disparities among schools and localities and offer perverse incentives to continue inefficient service delivery. In addition to deficiencies in primary and secondary education, the quality and performance of higher education is sharply declining as a result of fragmented administration and muddled financing. 4. 3 The persistence of these problems often stems from lack of policy and planning to adjust outdated content and structure. Public expenditures on education in SEE, though lower relative to GDP than in the EU or among the OECD countries on average, are in line with the levels observed in many fast growing economies at comparable levels of income. Increasing spending in the sector is not necessarily the answer, except perhaps in Albania. In most other countries, the efficiency of resource use would need to be increased by education policy changes, adjustments in management and administration, and expansion of private financing (fees and involvement of the private sector). While education and training reform is only a small part of the acquis, it is central to the countries’ future and living standards and deserves the authorities’ undivided attention. I. EDUCATION EXPENDITURES AND SOURCES OF FINANCING 4. 4 Public spending on education averages 4.5 percent of GDP in SEE and is lower than in the NMS or the OECD on average (Table 4.1). From a historical perspective, however, public outlays are somewhat higher than the amounts spent by Japan in the 1970s, as well as the Philippines, South Korea and Thailand in the 1990s, indicating that overall government expenditures are not low. (The latter countries also relied heavily on private financing of non-compulsory education, in a similar way the SEE countries approach the tertiary level.) The education system in Albania appears under-funded with spending at about 3 percent of GDP, but is relatively efficient in its use of inputs. 4. 5 In addition to public funds, education is financed also through non-budgetary resources (including income generated by schools and special allocations for non-educational or research activities in education institutions) and direct private spending on education (fees and voluntary contributions by parents). For the education systems altogether, all the non-budgetary resources may add up to 20-30 percent to the budgeted expenditures or more (with the expenditures overwhelmingly at the higher education level). In Macedonia, for example, private expenditures on education amounted to 1.4 percent of GDP in 2005, and are higher by about half than the average for the EU15 and by about 15 percent than the average for the NMS. Given the experience of countries outside Europe, private spending, especially at the tertiary level, has been helpful in 43 improving education outcomes and should be encouraged, subject to changes that foster increased transparency and efficiency of that spending. This is especially true in the context of tight budget envelopes. II. EDUCATION OUTCOMES 4. 6 Despite the level of spending, the stock of human capital in many SEE countries is Table 0.1: Public Spending on Education, 2004 (In percent of general government spending and GDP) rather low, as measured by educational In percent of In percent of attainment, average years or schooling and Country government spending GDP enrollment rates. While levels of skill and Albania 10.4 3.0 knowledge are generally correlated with the BH 12.0 4.7 level of GDP per capita, learning outcomes in Croatia 9.3 4.6 most of the SEE countries are in general lower Macedonia 11.5 4.0 than would have been predicted by their levels Montenegro 14.3 5.7 of income (Figure 4.1). This is corroborated by Slovenia 11.9 5.1 the results of the Programme for International SEE average 11.5 4.5 Student Assessment (PISA), an assessment by the OECD administered for 15-year olds. PISA Hungary 10.3 5.5 included Albania and Macedonia in its 2001 Japan 10.6 3.6 study and Serbia and Montenegro in the 2003 OECD average 12.9 5.4 study.28 According to these assessments, Sources: Trans MONEE, PEIRs of individual countries, Education Albania, Macedonia and Serbia and at a Glance, OECD Montenegro scored well-below most other countries, with roughly 90 percent of the students scoring below level 3 (1 being the lowest and 5 the highest); only individuals who attain a score equal to or higher than 3 are considered able to functions adequately in a modern workplace (Table 4.2). Moreover, the SEE countries have lower enrollment rates for these age groups (see below), so those students with the poorest educational records do not even participate in the PISA assessment and the overall scores of SEE countries are inflated (although still very low in absolute terms. Figure 0.1: Student Performance in Reading and GDP per Capita, PISA 2000 Source: Figure 3.9, p.56. Expanding Opportunities and Building Competencies for Young People: A New Agenda for Secondary Education, World Bank, 2005. 28 PISA was also conducted in 2006, but the results will not be available until December 2007. 44 Table 0.2: PISA Reading Scores: Albania, Macedonia and Comparators Percent scoring Percent scoring at levels 3-5 Country < level 3 Total % levels 3-5 % level 3 % levels 4-5 Albania 90.9 9.1 7.7 1.4 FYR Macedonia 87.0 13.0 11.1 1.9 Bulgaria 67.3 32.7 21.5 11.2 Latvia 56.9 43.1 25.2 17.9 Russian Federation 56.7 43.3 26.9 16.4 Poland 47.4 52.6 28.2 24.4 Hungary 47.7 52.3 28.8 23.5 Czech Republic 42.3 57.7 30.9 26.8 EU15 40.1 59.9 28.8 31.2 Sources: OECD, 2001 and World Bank staff calculations. 4. 7 In most SEE counties, the average years of schooling individuals have attained lag the EU15 and NMS, but are often higher than in countries in other parts of the world with similar income levels. Albania’s average number of years attained has increased substantially from 8.3 years in 2002 to 9.6 years in 2005, and Macedonia’s has risen marginally to 10.3 years over the same period. These results compare with 12 years in the EU on average. Besides the years of schooling, however, the amount of compulsory instructuion hours is also important. In Macedonia and Croatia, for example, students at the primary level receive substantially fewer compulsory instruction hours (ahout 432 a year in the former and 473 in the latter vs. 781 in the EU15 and 583 in the NMS), reflecting a combination of double-shift schools, mandated limited number of classes, and the thinking of local psychologists that helps limit instructional hours. Double-shift schools are prevalent in many SEE countries and have limited scope for instructional time to be in line with the EU average (in every country, moreover, there are schools with even three shifts). Eliminating double-shifts schools, however, would require substantial capital investment. In the short-run, as a result, countries need to pursue other more cost-effective ways to improve quality. 4. 8 With the exception of Slovenia Table 0.3: Enrollment Rates in SEE and Selected Comparators, 2004 (and Albania for primary Pre-primary Primary Secondary Tertiary education), enrollment Albania 49.5 105.7 78.0 19.3 rates in the SEE BH 4.3 92.8 72.6 24.2 countries lag the ECA Croatia 47.9 95.6 88.3 41.8 average and are well Macedonia 32.1 97.7 84.1 28.0 Serbia and Montenegro 44.0 96.5 55.0 23.9 below the rates in the Slovenia 78.7 99.4 99.8 73.7 NMS (Table 4.3). At the pre-primary level, in Bulgaria 77.8 104.8 102.1 41.1 contrast to the times of Czech Republic 107.4 101.9 95.7 43.2 the former Yugoslavia Hungary 81.1 97.9 96.5 59.6 when pre-schools were ECA 50.5 101.5 90.5 48.9 linked to the place of Note: All data on a gross basis except for BH, where they are on a net basis. Data for Serbia employment, the pre- and Montenegro are for 2001. school system today has Sources: World Development Indicators, 2006 and Trans MONEE, 2003. severely declined along with the large numbers of state-owned and public enterprises. The remaining pre-schools are largely concentrated in more affluent urban areas. Limited access to pre-primary education puts the poorest families at a disadvantage in terms of their readiness for the subsequent levels of education, diminishing opportunities for poorer students to achieve at least an average level of education attainment. 4. 9 Providing better access to pre-primary education for underprivileged children would promote overall equity and help improve education performance. All governments in the region have initiated 45 reforms to make some period of pre-primary education compulsory or extend compulsory education from eight to nine years. These initiatives, however, must do more than simply extend existing arrangements: instead, they have to be implemented to ensure that public resources are used first to ensure equal access in the rural areas. There are cost-effective policy options to increase access in the early years of education as the student-age population is declining sharply in most countries. A lot of schools, especially in rural areas, have excess capacities in classrooms, and often in staff, that can be transformed for this purpose. 4. 10 Nearly full enrollment has been achieved in basic (primary) education (most SEE countries have recently made this years 1 to 9, although most of the data in this report is from the time when it was 1 to 8). There has been little change over the last several years in terms of aggregate enrollment. In Croatia, Serbia, and Montenegro, earlier shortcomings in enrollment rates appear to have been fully addressed. However, officially-estimated aggregate enrollment rates are often inconsistent with international comparisons and household surveys. This suggests that countries have persistent problems with reliable information, which makes planning, adjustment of capacities to demographic or migration trends, and monitoring of enrollments of the at-risk population difficult. 4. 11 It is also important that policy makers look beyond aggregate enrollment numbers as these figures may mask several problems. For example, ethnic and political conflicts in many countries in the region have resulted in periods of discontinuity in many children’s basic education. Focusing on aggregate gross enrollment rates (and even net rates) can mask this, as some student groups catch up with schooling, while others either miss education altogether or fail to complete basic education. Household surveys from the region in the last decade offer some indirect evidence: a significant portion of students start compulsory education and complete various grades with delay (caused by grade repetition), whereas there is a persistent 10-15 percent of the population over compulsory age (15 to 18 years olds) which did not complete compulsory education. 4. 12 Disaggregated household survey data from some SEE countries show that drop-out (lack of completion) is concentrated in the later years of basic education. This means that drop-out rates for grades 5-8 are probably understated in government statistics (which cover the whole of primary education and regularly show a drop-out rate about one to three percent). Many of the affected population groups belong to marginalized communities (such as the Roma, internally displaced populations or the extremely poor). 4. 13 Evidence from qualitative work in the lower-income SEE countries (such as Albania) suggests that although children are enrolled, they may be attending school irregularly for a variety of reasons, including the poor physical condition of classrooms, limited access to schools in the winter, and the need for earn income. The scope of the problem is difficult to measure, as schools do not reliably report absent students. 4. 14 In most SEE countries access and equity worsen considerably at the secondary level (typically years 9 to 12). Secondary education is not yet compulsory in most countries, though Macedonia will make it so from September 2008. In most countries large gaps appear with respect to whether students live in urban or rural areas. In some rural and mountainous areas, there are no secondary schools within reasonable distance of students, and no transportation assistance schemes are available to facilitate access. Although secondary education could be provided by private schools in addition to public schools, no SEE country has meaningful private sector engagement in upper secondary education. 4. 15 At the tertiary level, enrollment rates across SEE have grown slowly over the past decade and lag significantly other European countries. The enrollment rate in Croatia reached 43 percent in 2005, while the rate in most other countries is almost half as low. 46 III. INSTITUTIONAL AND POLICY CONTEXT A. Challenges Related to the Content, Quality and Structure of education 4. 16 Curriculum and education structures especially for grades 5-12 are still mostly focused on the transmission of subject or field-specific factual information rather than on general skills and capabilities to apply knowledge, as demanded by the global economy and labor markets. Most countries have initiated reforms focusing on improving curriculum standards and study programs, though too often this has been through the introduction of additional subjects and material to an already over-loaded curriculum, rather than on re-assessing what students learn and how they are taught. An important tool for change might become national qualification frameworks, especially where these cover both secondary and higher education. Qualification frameworks present learning outcomes (i.e. statements of what a learner can do and is expected to know after a certain period of learning) in a systematic way thus facilitating flexible learning paths, recognition and curriculum reform. They can help shifting the focus from detailed curricula to a more flexible, student-centered approach. 4. 17 Secondary enrollments remain concentrated on vocational and technical schools, except in Albania, constraining students’ choice of schools and programs. In Croatia and Macedonia, the enrollment in vocational/technical schools has even increased. In Albania that share has decrease, contributing to the overall shrinkage of the upper secondary system. Such programs, while they may have matched the needs of a rigid socialist economy during the former SFRY period, are less relevant in the fast changing global economy. Evidence of Serbia shows that secondary education graduates appear to be no more successful in the labor market than those with just primary education; in fact, secondary graduates have the highest unemployment rates. Their wages are only about two-thirds of the wages of those with university degrees, and are not much higher than the wages of those with a primary education.29 4. 18 Narrowly-specialized and practical courses of technical and vocational education also make these programs expensive. The vocational programs are about 30 percent more expensive than general secondary education in SEE, even though equipment is typically obsolete as maintenance or improvements have been deferred. The return on vocational education in BH, for instance, is zero percent, whereas the return on general education is 4 percent.30 High unemployment in the region means that any job-specific skills learnt in vocational courses may be quickly lost. A better secondary education strategy would provide skills that are needed for life-long learning, and create opportunities and incentives for the private sector to engage in training, and alleviate the burden away from public education. Throughout Europe, educational policies adjust curriculum by delaying specialization, providing general skills up to the 10th grade, and reform the technical content to also transmitting general skills applicable in a wide range of jobs. 4. 19 One of the more successful aspects of education reform in SEE has been the development of new education standards and reliable methods for assessing the performance of students against those standards. These were started by international agencies and donors, but increasingly governments themselves are taking over their financing. In addition, SEE countries have, somewhat sporadically, participated in international assessments conducted in recent years.31 Each country is also introducing a new examination at the end of secondary education (matura): Albania has already successfully introduced it and Macedonia will run it for the first time in 2008. These examinations will have an important role beyond the recording of 29 Serbia and Montenegro: An Agenda for Economic Growth and Employment, World Bank, 2004. 30 Papps, Ivy and Jon Burton, 2003, “Bosnia and Herzegovina, Skill Gaps and Training Needs - Current Status and Policy Implications: Qualitative Study No. 2”, Durham, Birks Sinclair and Associates. 31 Macedonia and Albania participated in PISA in 2001 and Serbia in 2003 and 2006. Macedonia (2000) and Serbia (2004) also participated in the Third International Mathematics and Science Survey and Macedonia in the Progress in International Reading Literacy Study in 2002. 47 performance; they will sharply reduce the opportunities for corruption in entrance to higher education. 4. 20 The PISA assessments also highlight the challenges ahead in education financing. Currently, resources are used to buy the “wrong product” in terms of quality. PISA sought to measure how young adults at the age of 15 use their knowledge and skills to meet real-life challenges rather than the extent to which they have mastered a specific school curriculum. High input standards in the SEE countries (especially staffing) motivate schools to overestimate their staffing needs and focus less on performance outcomes, especially since staffing and wages are often paid by the central government. Meanwhile, there are almost no public resources for upgrading and modernizing pedagogy or upgrading teaching skills. Furthermore, such standards may reinforce disparities between rural and urban areas, ethnic groups, and along poverty lines. Qualified teachers are less inclined to teach in poor communities, and small schools in sparsely inhabited localities are frequently short of adequately qualified teachers. Moreover, differences in non-teaching expenditures between localities with different ethnic mixes are often large (estimated at about 40 percent in Macedonia). PISA results showed that such disparities cause significant declines in overall education system performance. 4. 21 In higher education, all SEE countries have signed on to the Bologna process which reforms the structure of degree programs and introduces new mechanisms for quality assurance. However, these countries’ progress is significantly behind the rest of Europe. While some universities in each country have moved towards a structure of 4 years for an undergraduate degree, divided into courses credits, this has not resulted in a re-appraisal of what students need to learn and how they are taught. Croatia and Macedonia have ratified the Lisbon Recognition Convention, though in BH different universities within the country do not mutually recognize each other’s credits. Finally, in the long term, SEE countries will need more balanced higher education systems, which cater for diverse – including diverse regional - needs in their countries may include a strengthened polytechnic sector and specialized higher education institutions. There are almost no institutions (only Croatia has non-university type institutions), degrees, and capacities for shorter programs. B. Challenges in Education Administration and Management 4. 22 In most SEE countries, there have been several significant changes in education administration, but with inconsistent results. After the dissolution of the SFRY, many of the newly emerging countries got rid of the Yugoslav system of self-management to accommodate the newly established national authorities’ efforts to gain control over what is being taught. However, centralized education administration resulted in an overly politicized system in which school directors and local administrators were regularly hired and fired based on political considerations. Financial resources for the schools, in particular for capital improvements, were also allocated based untransparently. 4. 23 Since the end of the 1990s, all countries, especially those with significant ethnic minorities, have initiated some form of decentralization. This has meant various forms of changing authority: for example, de-concentration of authority to the municipal level in Macedonia starting in 2005, and the delegation of authority to cantonal levels in the Federation of BH in 1996. In Macedonia, as well as in Croatia in 2001, decentralization included the devolution of responsibilities for financing of maintenance and capital expenditures to municipalities, and selection of school principals to school boards. Macedonia has begun moving toward a fuller devolution of school finances to the municipal level, and the authorities intend schools to be responsible for paying teacher salaries starting in late 2007 or early 2008. The modalities of this transition are still being worked out, however. 4. 24 The underlying reasons for decentralization have been political or ethnic, and the reforms have not properly taken into account the financing needs and effect on efficiency, equity, access, and on quality of education services. In fact, they have often worsened disparities between groups. For example, funds for 48 salaries in Macedonia will initially not be allocated on a per pupil basis; as of late 2007, they are set to continue to reflect historical allocations. While this may make sense in the short term to reduce disruption, delaying these difficult changes will likely worsened disparities between groups. In the long term, improving efficiency may contribute to boost access and quality, if savings are invested in developing education services for better performance and effectiveness. 4. 25 In universities, the most acute problem with respect to public funding is the fragmentation of universities into separate faculties, with separate administration, accounts, budgets, revenues, and staff.32 Each faculty typically teaches the full range of courses necessary for their own degree programs. Fragmentation is the source of major inefficiencies and problems of quality and relevance. Fragmented faculties (organizations within a university in charge of a specific branch of training) have hardly any incentives to adjust their programs, engage in innovations, or increase the number of graduates given the shortage of public places. Existing methods of public financing and the general practice of user fees reinforce the faculties’ disinclination to initiate change. In the last couple of years, all SEE countries have passed laws which mandate an integrated university structure. However, SEE policy makers should be aware that laws intended to integrate institutions take many forms. Some, for example in Slovenia, were not intended to lead to full integration. Whereas the reform process was envisaged as a step-by-step approach, only the initial steps were ever taken. SEE governments must acted decisively and in a determined fashion to successfully implement the necessary deep reforms. The University of Tuzla in BH is the rare case of a succesful integration, where graduation rates are up and costs per graduate are down. C. Budgeting and Financing Arrangements 4. 26 In all SEE countries, the central government’s education budget formula assigns priority to staffing needs, and fails to provide incentives to pursue cost reduction through efficiency measures. The budgeting procedure first determines the number of teachers and other staff required to operate the school and then determines the wage bill for this staff. Staffing requirements are defined by coefficients that are applied to the number of class sections, students in the classes, special staff assigned to the school, the curriculum being offered, and the maintenance needs for the buildings and grounds. Whereas the budget formula and associated regulations on class size are apparently rigid, they are in practice frequently ignored, allowing schools to manipulate funding by opening new classrooms. A study in Serbia looked at those primary schools with around 100 students in a given year and found that they created between 4 and 6 classes for these students, resulting in quite different allocations of staff (and resources). 4. 27 School principals have no incentive to use resources more efficiently, since any savings generated (for example from reducing the number of classes or saving on heating bills) are lost to government. Also, if schools are forced by demand to operate overcrowded classes, as is more common in urban areas, no remedial measures (that is, additional equipment or teaching assistants) can be introduced. Disparities between class sizes are significant across the region. Some rural schools in Macedonia or Montenegro are kept open for only a few students (in Montenegro ten schools have one students each, in Macedonia three schools have less then five students); though Montenegro has recently planned a consolidation of the school network. Simplification of the budget formula by emphasizing the number of students being taught would eliminate these perverse incentives, and encourage schools to find ways to increase efficiency, while providing a larger share of funds to overcrowded schools. 4. 28 Unlike in general education, SEE countries appear to have neither a stable budgeting formula in higher education, nor stable regulations for financing of faculties and universities. Instead, these are often negotiated faculty by faculty. Budgets are input-based, like those in general education, perpetuating 32 Macedonia is to a degree an exception, as the newest public institutions were established as integrated bodies. However, the largest and most prestigious institutions remain a confederation of independent faculties. 49 inefficiencies and de-linking the provision of courses from student demand: moreover universities have every incentive to delay the graduation of students to increase income from fees. In most countries, neither public funds nor significant private financing are detailed in the budget (at most an overall public allocation is recorded). Private funds, generally tuition fees and other payments by students, but also extracurricular activities such as rents, are not accounted for, even though evidence indicates that as much as half of the revenues can come from non-state sources (in Macedonia up to 60 percent). 4. 29 Faculty members are employed by individual faculties and not the central university administration. Since a full-time teaching load is a few hours per week, large numbers of professors teach in more than one institution and, since each faculty within a university is considered a separate institution, teaching staff can draw more than one full-time salary within one university. Budgets are planned based on the number of new students to be enrolled on a full-time basis, although the incidence of students staying enrolled but without making academic progress (either because they do not wish to graduate and become unemployed or professors do not provide opportunities to pass examinations) is very high across the region, making budgets very arbitrary. A study of the University of Sarajevo in BH found that only three percent of students graduated on time. In this respect, the absence of any control over part-time students, who simply remain enrolled notwithstanding their progress, and any mechanisms for converting part-time students into full-time equivalents, are major problems, weakening the performance and the planning capacity of the system. IV. EFFICIENCY ISSUES 4. 30 Allocation between levels of education: While the SEE countries allocate the largest share of their education budget to primary Table 0.4: Public Unit Costs by Level of Education education, expenditure per student (In Percent of Public Unit Cost for Primary Education) is far greater for tertiary education. Pre-school Primary Secondary Higher Table 4.4 shows the public unit OECD average 80 100 134 199 costs at different levels of Albania .. 100 116 338 education compared to the public BH .. 100 132 232 unit costs for basic education. This Serbia 146 100 109 224 comparison eliminates the effect of Macedonia 167 100 119 204 variations in GDP and also in the Croatia 118 100 98 150 overall student numbers. Source: Public expenditure reviews, statistical reports from individual countries, Compared to OECD, in Serbia and Education at a Glance, OECD Macedonia preschool education unit costs are high relative to basic education33. Further analysis is needed to assess whether these unit costs reflect overly high standards (which make services scarce and inequitable), or low enrollment rates. The unit cost of secondary education in SEE is generally lower than in OECD, but significantly higher for higher education. This tendency to overspend on higher education needs attention. Relative to enrollment, higher education receives more resources than the other sub-sectors: more than a quarter of the education budget is allocated to university students that account only for 5-8 percent of total students. Secondary education in SEE receives slightly more, but enrolls almost three times as many students on average. In all, these unit cost comparisons suggest that there is a need to revisit intra-sector allocation of education resources in SEE. 4. 31 Allocation between expenditure categories: The greatest share of the primary education budget (about 77 percent in BH, more than 80 percent in Macedonia and Croatia) is for wages and salaries. In practice it means that under fiscal pressures the first priority is to pay the teachers and other staff, and funds 33 Data on enrollment and especially financing of pre-primary education is sparse, since it is often not considered to be part of the education system and hence not under the Ministry of Education. Often, pre-primary education provision is fully decentralized. 50 often get depleted before materials, teacher training and capital investment can be provided. However, as mentioned, the large allocation to salaries is also based on perverse incentives for municipalities/schools to increase the quantity of teachers and classrooms rather than focus on performance (and pay fewer teachers at higher levels). The present system of allocations, regulations and incentives leave little for repairs and capital investments, and deferring necessary repairs and investments often contribute to declining efficiency and effectiveness. Recommendations to move to capitation financing, together with sufficient decentralization to make staffing decisions, may help alleviate some of these pressures. The experience of ECA countries that have introduced some form of capitation financing (Czech Republic, Hungary, Poland, Lithuania and the Slovak Republic) demonstrate, however, that the inability of local governments to dismiss staff threatens the success of capitation financing, preventing countries from taking advantage of the decline in school cohorts to restructure education spending and boost education outcomes. 4. 32 While the share of wages and salaries within the total education budget is high, individual teacher salaries are relatively low, reflecting high staffing levels. While accurate comparative data is not available, data from individual countries suggest that salaries of teachers are often lower than average salaries of other public employees. Salaries for primary school teachers in grades 1-4 typically average 70-75 percent of the average salary in the SEE countries; for grades 5-8 about 80-85 percent, and at the tertiary level about 90 percent. Typically, salary structures are compressed and differentiated only by length of experience and status. There is no recognition of excellence nor are there incentives for performance or training. All this makes it hard for the teaching profession to attract highly qualified young people. 4. 33 Adjusting education Table 0.5: Number of Students in Basic Education (In thousands) services to demand (technical 1997 1998 1999 2000 2001 2002 2003 efficiency): Throughout the region, the school-age population Albania 560 559 553 543 535 523 503 and the number of students have BH 390 396 403 .. 383 363 367 been declining (Table 4.5) and Croatia 423 418 413 405 400 395 393 are likely to continue on a Macedonia 259 257 254 248 244 237 .. downward trend. Moreover, these overall declines conceal Serbia 763 745 731 711 696 680 667 significant internal migration Source: Data collection from Ministries of Education trends. While the number of students declines in rural schools, urban schools almost always teach in two and sometimes in three shifts. Staffing levels continue to be defined centrally and have not been adjusted in the last decade. As a result, although there is a general decline in the average size of schools and there is evidence of fragmentation of the school system into very small rural schools and large, overcrowded urban schools. 4. 34 Student-teacher ratios (STR) have been systematically declining in all countries Table 0.6: Student Teacher Ratios in SEE Countries over the last decade, except in Albania. and Selected Comparators STRs compare favorably with ratios in the Country Primary Secondary Tertiary EU and the OECD on average (Table 4.6). Albania (2004-2005) 18.0 -- 16.0 Averages in the SEE countries conceal large BH (2003-2004) 17.5 16.9 -- disparities across rural and urban schools, Croatia (2004-2005) 14.7 11.5 -- however. In rural areas, it is not rare for Macedonia (2005-2006) 16.0 16.0 16.9 schools with 50 students or less to have four Serbia and Montenegro 36.5 -- -- or even eight classrooms, each class requires NMS (2003-2004) 14.0 13.3 15.3 at least one teacher for lower grades and a EU15 (2003-2004) 13.5 11.8 16.1 number of part-time teachers specialized in OECD average (2003-2004) 16.5 13.6 14.9 different disciplines from grade 5 to 8 or 9. Sources: Trans MONEE and World Development Indicators, 2006. The evidence, for example from national assessments in Serbia, is that these schools teaching only grades 1- 51 4 offer the worst quality education. In large urban schools, classroom size reaches 40 or even 50, and teachers in classrooms with significant numbers of students from minority populations often have poorer qualifications. Population estimates project a decline over the next ten years of between 13-26 percent of the 15-19 year olds in the population (the population group expected to represent the demand for schooling from preschools to entry into higher education). It is likely that the decline will especially affect the already sparsely inhabited areas, which would call for more tangible efforts to realign capacities (especially staff numbers with demographic changes) (Table 4.7). 4. 35 The task of improving internal efficiency is Table 0.7: Population Projections, 5-19 year essential to enable countries to free up resources for olds (In thousands) increasing enrollments and improving quality. 2015/ However, this task is both technically and politically Country 2000 2005 2010 2015 2000 difficult. The main focus should be on the use of BH 859 777 672 628 73% primary school facilities and teachers as this is where Croatia 840 755 685 660 79% enrollment rates are already high but absolute numbers Macedonia 482 454 429 413 86% of pupils will fall. Effective plans to reallocate Albania 929 890 822 783 84% physical, human and financial resources will require Serbia 2273 2160 2035 1979 87% changes in the core regulations: in input standards, Source: Calculated based on population projections by the roles and responsibilities, and incentive structures. World Bank. First, funds would need to be allocated through a transparent formula based primarily on student numbers; this will expose the more inefficient facilities and create incentives for closure. Second, countries will have to decide whether the central or local governments should take the lead in rationalizing facilities. There are no easy answers here. The central government is distanced from local politics but also does not have the detailed information to make good choices. Whoever is responsible should also have authority over large enough student populations, changed regulations in terms of input standards, and incentives to initiate reallocations especially between salary and non-salary expenditures. Therefore in countries with significant numbers of small local jurisdictions, the central government will likely need to take the lead. Third, experience in developed countries suggests that school principals, given the right incentives, can use staff and other resources flexibly to improve service delivery (including coverage, completion, and student performance). 4. 36 Another efficiency concern is related to low on-time completion and graduation rates at different education levels. No SEE country has accurate records on full-time student numbers or capacity to track students through their school and university careers, but some anecdotal information on a country basis exist. Available information indicates that less than 40 percent of the students graduated on time in Macedonia and one-third in Croatia. A tracking of students initiated in BH concluded that less than 10 percent of students graduated on time. Students in the region typically study 7-9 years to complete a four- year program. This has significant impact on higher education’s overall performance, cost, and the labor market. Fixing this inefficiency would make space for expansion of tertiary education, since if students were to graduate sooner, more could be admitted. V. EQUITY ISSUES 4. 37 Equity problems in education persist in SEE. There are few programs which specifically target students from disadvantaged backgrounds and help them stay and succeed. Surveys and social assessments in the region repeatedly show that children belonging to poor or extremely poor households, those with unemployed fathers, living in rural communities, and belonging to ethnic minorities (especially Roma) are more likely to have irregular daily school attendance and less likely to complete compulsory schooling. If more than one or two of the above factors are present, there is a higher chance of non-completion than that of completion. (By some estimates, only one-third of Roma complete basic education in Croatia, 52 Macedonia, and Serbia.) 4. 38 Selectivity of education systems starts with entrance examinations to secondary schools and continues to streaming students to vocational programs. Children with poor family backgrounds have difficulties achieving adequate grades and performing well in examinations, and often end up in programs with lower academic quality and lacking opportunities to transition to higher levels of education. International assessments show that the countries participating from SEE have not only shown low overall performance, but also have had high standard deviation--evidence of large number of students with very low scores. Results from Macedonia show that these students are dominantly coming from ethnic minority groups. Overcrowded urban schools in poorer districts are forced to have class sizes reaching 35 to 45 children and in three or even four shifts, which allows very short teaching hours or time on tasks. Pupils going to school in three shifts may spend as little as three hours; those in four shifts may spend as little as 2 hours and 10 minutes in classes per day. 4. 39 As a consequence, ethnic minorities (Table 4.8), and families with low social and economic status reap fewer benefits from post-compulsory education. While the poorest quintiles take a relatively large share of the benefits from basic education Table 0.8: Participation Rate in Education in (given the relatively high number of children Macedonia by Ethnic Group among the poor), the participation rate becomes In percent of quite strongly regressive in secondary and Basic Secondary Tertiary population tertiary education. In Albania, the poorest 20 Macedonian 58.1 77.8 89.2 66.6 percent of the population represents only 10 Albanian 30.7 16.5 4.9 22.7 percent of those participating in upper Turkish 4.3 1.7 1.2 4.0 secondary education and only 2.5 percent of Roma 3.1 0.5 0.3 2.2 those participating at the tertiary level. In BH Vlachs 0.2 0.2 0.9 0.4 and Macedonia, the benefit incidence analyses Serbs 1.2 1.4 1.8 2.1 showed similar patterns in terms of under- Others 2.5 1.8 1.6 1.9 Source: Macedonia Central Statistical Office. representation of the poorest in post- compulsory education. Incidentally, in these three countries, girls living in rural areas and with Muslim religious background represent a significant portion of those who rarely continue their education beyond the compulsory level. VI. CONCLUSIONS AND RECOMMENDATIONS 4. 40 Governments of the SEE countries are invited to reflect on the following recommendations: • Improve accountability: Ministries of Education would need to develop analytical capacities that link enrollment and financial data with information about quality and effectiveness to enhance accountability of education spending. Regular quality assessments take only 0.4-0.6 percent of total education budgets in other European countries, and these assessments can help save many times that with the results they generate. • Funding to follow students: Financing should be based mainly on the number of students (capitation financing) instead of the number of classrooms or teachers. Such a financing system should help better follow internal migration patterns and demographic changes. In decentralized systems, formula-based financing is used to allocate resources from the central budget to sub-national education authorities, and central governments need to monitor that local authorities allocate adequate resources further to schools. Also, incentives need to be established to ensure schools and education authorities strive for 100 percent net enrollment, including full completion of compulsory education. For per-student financing to work, 53 governments need to have the discretion to switch funds between wage and non-wage recurrent categories. In the short-term, this should provide an incentive to close small schools and reallocate the funds saved. However, there are risks that savings may be used to increase the number of teachers or other staff rather than increase spending on inputs vital for improving quality. Setting up national standards is likely to help regulate these issues. • Tackle equity and disparity concerns: To finance expansion of pre-primary and secondary education, there would need to be reallocations of education funds within the existing budget ceilings from primary and tertiary education to pre-primary and secondary, while the unit cost of pre-primary education would need to be reduced. Part of public education resources needs to target those of the population that are at risk. Governments would need to carefully monitor the completion rates among these groups, especially girls from ethnic minorities and the poor in rural areas. Compensation funds would need to be established in decentralized systems to correct disparities between regions and municipalities. • Target improvements in quality and relevance of curriculum: Both compulsory and non- compulsory education would need to prepare students for uncertain labor markets, provide them with general skills and capabilities to succeed and be able to update their skills during their entire lives. Curricular standards would need to be made more realistic and adjusted to the demands of lifelong learning. Pedagogical innovation, school-based initiatives, and in- service training are the keys to reforming compulsory education and would need to be complemented with structural reform at the upper secondary level. At both levels, stronger focus on skills and competencies should be a pre-condition of curriculum and structural reform. • Improve higher education performance: Higher education needs to improve its performance, while the per-student public allocations will have to decrease so that the public budgets can focus more on primary and secondary education. This will require the increase of admissions and the decrease in the average number of years needed for graduation. These outcomes may be achieved only through a number of institutional and policy reforms: Universities and other higher education institutions are only autonomous in terms of their academic competence. To the extent that they are publicly financed, they need to be made more accountable and transparent. Quality and efficiency requires that there is an overarching university strategy and that the rector and his management team are able to act on these priorities. This means, at a minimum, that they and not individual faculties are responsible for budget planning, reallocation of resources between departments, and hiring and firing of staff. The European credit transfer and accumulation system needs to be fully implemented and the quality of university institutions needs to be assured through independent accreditation. All these activities would benefit from regional cooperation schemes. Public financing would need to be linked to the number of full-time students enrolled. Over the medium term, such a per-student financing formula creates incentives for on-time graduation. Research financing also needs to be made accountable and competitive. The student support schemes that governments may establish would need to be equitable, motivate students to improve their performance and graduate on time, and facilitate increased private financing. 54 These schemes could include student loan options, and need-based scholarships which are linked to the credits pursued and achieved by the applicant. Staff salaries and career advancement need to be made transparent and better linked to performance. 4. 41 The reforms discussed in this chapter are politically difficult, but they are part of an agenda that needs to be advanced to help boost the quality of education outputs among the SEE countries. Given the centrality of education outcomes and quality for economic growth and the distance most SEE countries need to cover to catch up with the more developed EU countries, pondering and acting on the education reform agenda should be a priority for the countries in the region. 55 56 Size, Structure, and Remuneration of The Public Service 5.1 Strengthening the quality of public administration while reducing its burden on the economy is a key challenge for many SEE countries. The challenge is reinforced by the need to augment or create new administrative capacity needed to help meet EU standards. Despite substantial efforts in most countries over the past few years, public wages relative to GDP are higher than in most relevant comparator countries Outcomes related to the public administration often leave much to be desired, however, suggesting that improving the structure and efficiency of the public service, and strengthening the value of money spent on public wages, would need to be among the key priorities of the SEE governments. 5.2 This chapter discusses the size, structure, and remuneration of the public sector in SEE countries. It starts with a discussion of overall wage bill levels and a comparison of public sector salary and wage expenditures. This is followed by a discussion of public sector employment and wage levels. A brief analysis of wage expenditures in key sectors and at the level of the local government follows, along with a summary on ongoing reform efforts and challenges ahead. I. COMPARATIVE ANALYSIS OF PUBLIC SECTOR WAGE BILLS 5.3 The wage bill in SEE is large on average despite substantial progress in reducing wage outlays over the last several years (Table 5.1).34 Excluding wages of healthcare professionals (which in most SEE countries are financed via lump-sum transfers from extra budgetary funds to hospitals), public wage outlays amounted to about 10 percent of GDP, compared with about 8 percent among the NMS. The wage bill in both the SEE and the NMS countries will increase if the wages for health professionals are included. In some countries, including Serbia and Montenegro, wage outlays will be higher still if salaries for temporary workers are properly counted as wage outlays rather than as outlays on goods and services. All told, nonetheless, there are substantial variations across the SEE countries. Albania spends less than the NMS on public wages, while Bosnia, Croatia and Montenegro are at the other extreme, with wage outlays well-above the NMS average. 5.4 The results of such cross-country comparisons are similar when considering wage outlays relative to overall government spending. Wage outlays are about 4.5 percent of GDP higher in the SEE than the NMS on average, with shares particularly high for Montenegro (28 percent) and BH (27 percent). This indicates risks to government effectiveness as wage outlays squeeze non-wage expenditures, including maintenance and operations and capital expenditures. 34 Non-uniform methods of data gathering by governments, different interpretations of what expenditures are included under the public sector wage bill and the existence of various in-kind benefits make comparisons across countries difficult. 57 II. TRENDS IN PUBLIC SECTOR STAFFING AND WAGE LEVELS 5.5 In most SEE countries high wage outlays appears to be driven mainly by Table 0.1: Public Sector Wage Bill overstaffing (most pronounced in the In percent of GDP sectors of defense, security, police and in Without With In percent of health health government spending some cases the sub-national level) rather than high wage rates. Over-staffing is Albania 6.3 -- 22.1 BH 10.0 12.7 26.9 due to over-extended public sector Croatia 10.0 12.5 20.5 functions and inefficiency in the delivery Macedonia 8.3 10.2 23.5 of public services. In some instances, Montenegro 10.6 13.4 28.1 however, higher wage rates are the key Serbia 8.5 … 17.8 factor, as is the case in BH for lower-skill Kosovo 8.1 … 24.6 employees (at the levels of the state and in the Federation entity). Furthermore, SEE 9.0 11.5 23.5 high public sector wage bills in SEE NMS 7.9 … 19.2 countries reflect a number of additional EU25 … 9.0 22.3 Note: Excludes wages in the health sector, which amount to about factors, including powerful public trade 3 percent of GDP on average. Numbers for Serbia are relative to GDP unions and sometimes rigid labor adjusted for the non observed economy (the only country in the region markets. Efforts to control the wage bill where official statistics do not include an adjustment). Data for Croatia, applied in the form of general wage Macedonia and Montenegro are for 2005, all others for 2004. freezes have been only partially Source: National authorities, WB estimates. successful, however, and the system for authorizing and monitoring civil service posts is either not formalized or does not function well in most SEE countries. A. Public Sector Employment 5.6 As a share of overall employment, public sector employment is high in SEE. Relative to overall employment, in Kosovo and BH, public sector employment amounts to 35 percent and 27 percent, respectively. This high share in total employment reflects in part low labor force participation rates in the region (amounting to 50-60 percent, the lowest in Europe), high unemployment, and the early retirement of a large number of workers as part of enterprise restructuring. Moreover, building new government structures after the wars of the past decade have added to public employment, most visibly in BH.35 Spending on security and defense (largely on soldiers and policemen) has also been on the high side, with outlays in BH double the rate in the NMS. Easing of tensions has helped reduce defense and public order employment only gradually. The full impact of the savings from demobilization could be realized only over the medium term: in the short run, severance payments and the costs of retraining the demobilized soldiers will need to be accommodated. 5.7 The high degree of politicization of public administrations in the region has also driven up public sector employment, as many governments have opted to establish new ministries and provide jobs in the public administration for political supporters and patrons. Overall employment growth has been modest in the region, as economic growth has been driven largely by productivity gains that have continued to involve substantial labor shedding in the former state-owned enterprises. As a result, unemployment in 35 In BH, the Dayton peace agreement that ended the 1992-95 war created complex administrative structures which led to a significant duplication of functions and excessive public sector employment. The implementation of the Peace Agreement in Macedonia calls for greater participation of ethnic minorities in the public sector, which has increased public employment. The complex political structure has also contributed to high public sector employment in Serbia and Montenegro. 58 the region has been elevated, ranging from 11.2 percent in Croatia to 20.9 percent in Serbia and to more than 30 percent in BH, Macedonia and Kosovo (Slovenia is an exception with an unemployment rate below 5 percent). Under these circumstances, the public sector has at times assumed the role of employer of last resort, and there was less adjustment in public sector employment than in the enterprise sector.36 5.8 Government employment as a share of the population in the SEE countries is not excessive compared to the CEE average (Table 5.2).37 Croatia, with government employment amounting to about 6 percent of the population is at the high end of the SEE countries, but is still lower than the 6.5 percent average for the CEE (1996-2000). Government employment represents less than 4.5 percent of the total population in BH and Macedonia, about 4 percent in Slovenia, and 3 percent in Albania. Nonetheless, central government employment as a share of the total population is above CEE levels in some SEE countries (Macedonia, Croatia, and Slovenia). Overstaffing in the security sector is an issue for all post- conflict countries, most notably BH, Macedonia and to a degree in Serbia. Employment in health and education is also high in most SEE countries, matching the internationally high levels in CEE countries. B. Wage Levels 5.9 Similar to the OECD and EU countries, public service pay in most SEE countries lags behind the private sector (Croatia, Macedonia, Serbia and Montenegro).38 In Serbia, the difference between the public and private sector pay is estimated at 12 percent for low-skilled employees, rising up to 40 percent for higher-skilled employees of those holding higher education degrees.39 In Macedonia, average private sector pay levels surpassed the public sector levels in 2000, with the difference increasing steadily over time. In Croatia, the average gap is about 15-20 percent, but for high skilled staff public sector pay lags private sector salary levels by up to 50 percent (adjusted for the level of education). Typically wage rates for public sector positions amount to 85-90 percent of comparable private sector positions in most OECD countries, with the difference reflecting the higher level of job security in the public sector. Table 0.2: Public Sector Employment (In percent of the total population) Total public Central Education Health Sub- Police employment Government national Albania 4.5 0.1 1.5 0.9 0.2 0.5 BH 4.0 0.2 0.4 1.0 1.8 0.6 Croatia 2/ 5.9 1.4 1.5 1.3 0.7 1.0 Macedonia 4.4 2.1 1.0 0.8 0.1 0.4 Kosovo 3.8 1.0 1.5 0.7 0.3 0.2 Slovenia 3.9 1.4 1.3 0.9 0.2 0.2 Montenegro 6.7 … … … … … CEE 6.5 0.5 1.4 1.0 0.5 0.3 Source: World Bank Administrative and Civil Service Thematic Group, Puma database. 2/ Croatia: 2005 (PEIR, 2007); police refers to uniformed services. 36 For example, in BH employment in the economic sector relative to employment in the non-economic sector has fallen from pre-war levels of 5.6 to post-war levels of just around 2.8. 37 General government employment in the CEE amounts to about 6.5 percent of the population (1996-2000) and is high by international standards, although is lower than the EU average of 8.9 percent. Decomposition of the numbers shows that high employment in the CEE is driven by overstaffed health and education sectors. 38 Pay rates in the private sector in the SEE countries are likely to be higher than reported, and hence reported public and private sector wage differentials can be misleading. Also, when assessing pay differentials, differing staffing profiles in the civil service and private sector would need to be taken into account. For example, the public sector tends to employ relatively more staff with higher education. 39 These numbers are based on reported wages levels. The actual wage level difference is likely to be higher. 59 5.10 In contrast to the other SEE countries, at the state level in Bosnia and Herzegovina, where most of the growth in staffing is expected over the next years, salaries exceed private sector pay for three out of six public service posts. Only at higher skill levels is the difference in favor of the private sector, with the gap approaching 25 percent. While there has been substantial convergence in wage levels between the Federation and the RS, there are sizable gaps across the state and entity level governments, with salaries at the state level still substantially higher despite a nominal wage freeze since 2001. All told, total net compensation (including both wages and allowances) in BH at the state level and in the Federation is higher than in Serbia and Croatia in purchasing power parity terms (Table 5.3). The disparity is most striking for lower-grade employees. Table 0.3: BH, Croatia and Serbia and Montenegro: Total Net Compensation, 2004 (In euros per month and in US dollars PPP) Job Title BH/State FBH Croatia SaM BH/State FBH Croatia SaM In KM per month In US dollars PPP Secretary General 864 536 1584 851 3,588 2,227 2,963 2,192 Assistant Minister 717 510 1584 851 2,977 2,117 2,963 2,192 Senior advisor 594 488 735 257 2,467 2,027 1,376 663 Senior associate 557 483 629 184 2,314 2,006 1,176 473 Associate 520 472 579 179 2,161 1,962 1,083 461 Clerk 447 339 455 160 1,856 1,410 851 411 Source: Price Water house Coopers, BH: Final Report, October 2005, for DFID. C. Sector and Local Government Wage Bills 5.11 As in most ECA countries, health and education account for the largest share of wage outlays (about 50-60 percent), with an approximately equal split between the two. These two sectors are overstaffed, with respect to both professional and support staff (Table 5.4). Over-staffing results from a combination of politicized hiring practices, low workload norms compared to OECD, and structural weaknesses, such as high numbers of support staff (see this report’s chapters on health and education). In contrast to the EU25, most SEE countries also tend to have a low level of local government employment, reflecting slower progress in advancing decentralization. BH is a key exception to this rule, as public administration in BH is more decentralized than most EU countries. 5.12 There are large efficiency gains that could be reaped in health and Table 0.4: A Breakdown of Wage Expenditures, 2004 (In percent of GDP) education through a more optimal use of Local Health Education professional staff, especially teachers and Governments primary care doctors. Spending on wages BH … 2.7 3.5 in the education sector in BH, Croatia, Croatia 0.9 2.5 2.9 Macedonia, and Serbia is Macedonia 0.3 1.9 3.0 disproportionately high compared to Montenegro … 2.8 4.3 spending on equipment and material, Serbia … 2.6 2.1 crowding out investments and having an Sources: Selected World Bank Public Expenditure Reviews. adverse impact on the quality of education. The situation in the health sector is similar, though relatively low salaries have until recently been mitigating the overstaffing problem, especially in Serbia and Macedonia. 60 5.13 Local government wage bills are low in SEE, with the exception of BH where the complex administrative structure at the sub-national level is driving up costs. In other SEE countries, the share of local government wage in GDP ranges from 0.3 percent in Macedonia to 0.9 percent in Croatia. These numbers reflect the fact that staff in health and education is counted as employees of either the central government or extra-budgetary funds. Similarly, local utility companies are owned by the local authorities, but their staff is not included in the local government wage bill. 5.14 In countries going through further decentralization, such as Macedonia, it would be important to assess and track the impact of decentralization on both the local and on the overall public sector wage bill. The potential for wage bill expansion can be significant, given the need for balanced ethnic representation in the administration and strengthening capacity in local administrations. Also, it would be important to ensure wages and salaries at the local and central levels are properly aligned. For example, in Croatia salaries at local level can be higher than salaries paid to senior officials and politicians at the central level. Linking central and local government salaries, as was done in Poland, or capping may be needed to keep expenditures in check. III. WAGE SETTING MECHANISM AND PAY STRUCTURE 5.15 Reform efforts to reduce the size of the public sector wage bill while decompressing salary scales need to take into account the method of wage determination and disparities among various groups. Public and civil service systems remain strongly influenced by powerful trade unions. Moreover, paternalistic traditions in the region remain strong, including the view that the government needs to provide employment opportunities. These factors have combined to make advancing already difficult reforms even more challenging, with adverse consequences for fiscal costs and the effectiveness of public service delivery. 5.16 All SEE countries have a similar civil service wage system, which is highly fragmented, non- transparent, and ill-suited to the demands of modern market economies. After a formula set by a government decree has established the minimum public sector pay, a coefficient system is used to calculate the base pay, and multipliers based on the rank and position are applied to determine individual salary levels. Salary levels are thus determined primarily by characteristics of individual civil servants and their rank, which in turn is determined by years of service and the level of education. Neither the type of the job nor performance significantly affects pay levels. While a system of this nature can bring benefits in a highly politicized environment, it is not well-suited to attract, nurture and retain a high quality civil service. 5.17 In addition to the base pay, civil servants benefit from several allowances. For example, all countries have an allowance for years of service; the allowance in some instances accounts for as much as 35 percent of total pay. Allowances for meals and transportation remain common, though several countries have started integrating these into the base pay. Specific allowances are usually provided for management positions, academic degrees, participation in government commissions, and work under difficult conditions; some of these are granted at the discretion of individual ministers. This practice distorts pay levels and contradicts two key EU principles that underpin effective and competitive pay systems: firstly, equal pay for equal work; and secondly, transparency. The trend in EU countries is to eliminate allowances and shift to performance-based rewards. For example, in Austria, a comprehensive overhaul of the pay system has eliminated most bonuses and allowances over the last eight years, and replaced them with a performance-based bonus.40 In Lithuania, the ongoing reform of the civil service 40 For a comprehensive overview of the status of OECD pay-for-performance practices, see: OECD, 2004, Performance-Related Pay Policies for government Employees, Main Trends in OECD Member Countries, Paris. 61 pay system also aims to reduce non-performance related allowances and increase flexibility in salary bands. These are trends that SEE countries may wish to consider when reviewing their current systems.41 5.18 Even though most SEE countries have legislation governing public sector wages, its effect has been distorted by the ability of individual institutions to adopt their own rules and regulations on pay of employees. This ability has been exacerbated by the existence of decentralized collective bargaining processes.42 Line ministries thus have considerable discretion regarding the setting of individual salaries and the composition and deployment of their workforce, which both reduces the coherence of the civil service and violates the European principle of ‘equal pay for equal work.’ The majority of countries do not have centralized recruitment and career management systems for the central civil service, and some even lack systems for monitoring the implementation of civil service legislation. This has led to inconsistent pay levels across the central state administration. For example, in many countries the Ministries of Interior, Defense, Customs, and Revenue Administration have managed to provide higher salaries for their employees through adoption of various by-laws and allowances. 5.19 Decentralized collective bargaining has also led to increasing discrepancies in actual wage levels since wage increases vary by sector and the base pay is also set by sector. Decentralized collective bargaining is advantageous for large groups of public sector workers, such as health professionals and teachers, but disadvantageous for central government civil servants that have lower bargaining power compared to other public sector workers. In Croatia and BH health professionals and teachers have in recent years succeeded in obtaining significantly larger salary increases than other public sector workers. Capacity development in the public administration has been hindered by these distortions, since wages of central government civil servants are often not only low in comparison to private sector wages, but also to other public sector workers and equivalent positions in public sector enterprises.43 Further, where local governments have discretion over salary levels, salaries of managers at local level may be higher than at the central government level (as in Croatia). 5.20 Turnover in public administration is constrained by the scarcity of appealing private sector opportunities or attractive but unsustainable public sector salaries. In Macedonia, the reasons generally cited for the low turnover include high risk and uncertainty in the private sector, as well as high overall unemployment. Turnover is also low for lower-grade employees at the state and Federation levels in BH, as salaries for these employees are higher than for private sector comparators. There are, however, isolated cases of high turnover, especially among the young – and skilled – who are discouraged by the strong reliance on years of service as a key determinant of salary and promotion.44 41 The integration of allowances into the base pay affects civil service salary compression ratios. 42 For example, in Croatia the Law on Salaries (2001) did not have significant impact on public sector wage harmonization, as institutions and sectors circumvented the law by adopting their own secondary legislation. 43 As is the case in Serbia, Montenegro, Macedonia, and the Republika Srpska in BH. 44 Both excessively low and high turnover have disadvantages: the former means there are no opportunities to upgrade skills by attracting adequate staff, while the latter means loss of experienced staff and institutional memory. 62 A. Compression Ratios 5.21 Civil service salary structures throughout SEE are compressed, except in Table 0.5: Civil Service Salary Compression Ratios Albania 3.7 Croatia and Serbia as of late. The ratio of the highest to the lowest salary is typically in the BH (state) 2.8 range 3-4, compared with about 8 in OECD on Croatia 7.6 average (Table 5.5). The salary structure in Macedonia 4.0 BH is the most compressed in Europe with a Kosovo 4.0 ratio of 2.8 (and the compression ratios in the Montenegro 5.1 two entities are somewhat smaller). The Serbia 7.0 compression ratios in SEE are also lower than France 5.7 in most NMS countries, which themselves suffer from compressed wage structures. Netherlands 8.2 OECD 8.0 5.22 Low compression ratios prevent United Kingdom 9.8 attracting and retaining a qualified civil service, Note: Ratio of the highest to the lowest civil service salary. but, unless accompanied by reductions in staff Sources: Selected World Bank Public Expenditure Reviews. levels, decompression will entail substantial increases in already high wage outlays. Low compression ratios also discourage staff from assuming higher level responsibilities. Given this, it is encouraging to note that several countries (for example, Macedonia) are now attempting to align their compression ratios with EU, thus creating conditions to retain qualified staff in the civil service. The most radical example of introducing decompression measures is Serbia, where changes in coefficients and the base pay raised the compression ratio from 1:4.9 to 1:7 in January 2005, and were accompanied by a ten percent reduction in staff levels in July 2005. The decompression measures are expected to be enshrined in new salary legislation in 2006. B. Politicization 5.23 The high degree of politicization of civil service posts is another key factor that limits career progression in SEE (Box 5.1). The more pervasive the system of political appointments, the fewer the opportunities for talented staff to reach a management level position, which takes away incentives to stay in the civil service and contributes turnover. IV. ONGOING REFORMS 5.24 Recognizing that deficiencies in pay systems are among the main impediments in attracting and retaining high quality professionals needed to build a professional civil service, many SEE governments have launched reforms aiming to make the compensation more attractive to skilled workers, while containing fiscal costs. Owing to fiscal constraints, large salary increases are impossible, and hence decompression needs to be accompanied by rationalization of the overall pay system and the size of the public service. The pay reform agenda, together with challenging depoliticization reforms, are being advanced by many NMS as well, providing valuable lessons to the SEE countries. 5.25 The reform agenda in SEE is concentrated in four areas: (i) rationalization of pay systems; (ii) de- compression and depoliticization of civil service; (iii) staff reductions in specific areas; and, (iv) moving to a performance-based reward system. 63 Box 0.1: Politicization of Civil Service In most SEE states, top civil service posts previously held by career officials have been de-facto politicized. A high degree of politicization impedes development of a stable and professional civil service, especially if no professional standards for political appointees are established, and reduces incentives for talented staff to remain in the civil service, as promotions depend largely on political affiliation. Ironically, former Yugoslav civil service systems were considered among the least politicized among pre-transition countries, as values of professionalism and impartiality were retained to a certain degree even under communism. In the 1990s, these systems were reversed mainly due to political polarization brought by the violent break-up of SFRY. States rapidly moved to expand positions held by political appointees, first at the Assistant-Minister level (previously the highest substantive civil service position), and later at the Secretary-General level (previously the highest management position). Croatia went furthest in codifying politicization in the Law on State Officials, which made all positions at the Assistant Minister level and above subject to political appointment. The aim was to raise salaries for top officials without raising salaries of the broader civil service, but in practice the law has adversely affected civil service professional standards, as officials are appointed based on political loyalty. In Serbia, similar arrangements are in place, though they are not formally codified and not linked to higher salaries. The new Law on Civil Service and Law on State Administration in Croatia and Serbia, respectively, will reverse these processes and reintegrate Assistant Ministers and Secretaries into the civil service. Appointments to top positions in Montenegro are made for five years (renewable) only, but brought back under the civil service regime, with the new Law on Civil Service that was adopted in April 2004. Implementation of new civil service laws in Macedonia and Albania also include attempts to reverse the currently high degree of politicization in state administration. A. Rationalization of the Pay System 5.26 SEE countries have made some progress in rationalization of civil service pay systems. The new Law on Salaries in Montenegro has reduced the number of allowance categories to 2-3, and the share of allowances in total pay by about five percent. The remaining allowances include the years of service and work in the field allowances, of which in particular the first one continues to be significant. The new law also introduced a performance allowance, which, if properly implemented, could alter the nature of the pay system. Tax reforms in Serbia have also led to a gradual integration of most allowances into the base pay, though transport, academic degree, and several other allowances remain. No performance-based element has been introduced in promotion or pay decisions, though a performance assessment system is being piloted.45 Croatia rationalized some allowances under the 2001 Law on Salaries. However, owing to entitlements under the labor code and reflected in civil service employment conditions, civil servants still receive a large number of additional allowances that constitute a significant element of the pay. An ongoing salary system review is expected to create the basis for further rationalization of allowances and their integration into the base pay, though. 5.27 However, little overall progress is made across SEE in eliminating the often generous years of service allowance, which is detrimental to younger staff motivation and performance, and costly in budgetary terms. This is unsurprising because the allowance is typically linked to labor code provisions, and hence most difficult to change. Therefore, its gradual elimination should be part of a larger reform of the grading system, and introduction of a performance-related pay, both of which would be expected to help retain young talented staff in public administration. 45 Performance assessment is envisaged by the law, but has been carried out rarely and ineffectively, as managers are not trained to conduct assessments and concerned it might create divisions among the staff. There is no official performance element in pay, but ministries can reward performance using the materials budget. Especially the tax and customs administration have taken the advantage of this opportunity. 64 B. Introduction of Performance-Based Elements to Pay 5.28 Most SEE countries have the legal basis for a performance-based appraisal system. However, owing to weaknesses in human resource management, performance appraisal systems are not applied on a consistent basis and linked to the pay. In some countries, there is also a reluctance to introduce performance-based pay. Decision-makers are concerned that performance measurement in the public sector cannot be done objectively, and that in a highly politicized and ethnically sensitive environment it is unlikely to succeed.46 5.29 Montenegro has introduced a performance-related element into the pay under the new Civil Service Law and Law on Salaries. Macedonia also intends to introduce a performance-based element in its pay system, as does BH. Croatia and Serbia have established a legal basis for development of performance appraisal systems through recently adopted Civil Service Laws, and in Serbia pilots are underway to test appraisal mechanisms. 5.30 Well-managed performance assessment systems could provide a basis for further innovative management approaches: for example, for the use of management contracts for top level staff, as in Latvia in the 1990s.47 However, this approach is controversial in systems where transparency is low and politicization high. The Montenegro Law on Salaries includes a softer version of the Latvian approach. C. Decompression and Depoliticization 5.31 The need for decompression of civil service salaries varies across the region. While in Croatia (and to an extent in Serbia) the compression ratio is broadly aligned with the OECD average, in all other states advancing decompression efforts would be called for to ensure increased attractiveness of top level positions. Decompression, however, tends to be difficult and costly. Labor unions tend to resist decompression and demand compensation for lower paid officials, while increasing pay for top officials would increase the civil service wage bill, unless salaries of lower level officials are reduced accordingly. A phased introduction of decompression, which is being implemented in Macedonia and Montenegro48, is one possible solution to this problem. Another option is to tie pay reform to systemization reform, as has been done in Serbia.49 5.32 The clarification of the division of labor and authority between elected politicians and appointed officials, and depoliticization of top management in the administration, should be another core element of the civil service reform. The pay decompression would need to be tied to depoliticization of the civil service. Without depoliticization, decompression measures will not have their intended effect, as benefits will go to political appointees, selected without proper merit-based 46 However, experience from OECD and EU states (e.g. Austria, the Netherlands, the UK), where the introduction of performance appraisal systems often raised similar fears, shows that as long as a system is well designed and those who implement it receive the necessary training, such risks are limited. 47 Experiments with bonus systems, such as the system of management contracts in Latvia, have generally failed to bring results and have instead raised concerns about their secrecy and perceived subjectivity. Latvia has operated a system of management contracts (between ministers and senior officials) since 1997. The contracts are secret, both in terms of the contents and the level of bonuses, which are known to be significant. This has created resentment in the population over what are perceived to be inappropriate reward levels. More incremental processes, as in Lithuania over a five-year horizon, for example, have shown better results. 48 The Montenegro Law on Salaries includes a clause that requires the government to re-visit compression levels on a regular basis. 49 The government of Serbia increased base-pay levels for certain positions (senior management and those involved in drafting laws) from the start of 2005, thus implementing selective decompression. 65 procedures. Therefore, establishing a clear link between decompression and depoliticization is crucial. This remains an area where progress in SEE has been lacking, though initial steps have been taken in some SEE countries. D. Staff Reduction 5.33 As mentioned before, core ministries in most SEE countries remain relatively small compared to OECD countries. In some countries, including BH and Macedonia, there is serious understaffing in specific core functions (e.g. financial management and control), which affects policy-making capacity. Hence, pay reform is unlikely to be funded by reducing the size of the core civil service, but would need to rely on staff reductions in other overstaffed areas in public administration, most notably security, health, and education. Financing civil service pay reform by reducing health, education, and other social expenditures will be politically challenging, however. 5.34 Reducing staff in the security sector (police and defense) has been advanced in some countries, such as BH, due to demobilization after the 1992-95 war. Besides demobilization, governments in SEE have mostly resorted to across-the-board staff cuts rather than targeted reductions. Such across-the-board cuts have tended to have a limited impact and to be potentially harmful to the quality and coherence of state administration. Serbia has reduced the size of its civil service by ten percent. The government of Montenegro also announced a reduction of public sector employment in 2005 by about ten percent. A careful review of the organizational structures of the core ministries and other institutions in SEE is warranted to consider possibilities for reallocation of staff to core institutions, while phasing out entities engaged in market and commercial functions to reduce and contain the size of the state administration. Similarly to across-the-board cuts, cuts by attrition are unlikely to be successful because all budgetary users bear the same burden. E. Improving Public Sector Effectiveness through Administrative Reforms 5.35 Establishing an effective public sector requires a well-selected and fairly paid labor force. The integrated approach to pay and employment reform requires a human resource base that operates with a clearly defined set of functions and responsibilities at each government level. Outlining the full set of measures and policy actions in this area goes beyond the scope of this report. It is worth noting, however, that reform measures include, inter alia, a clear decision-making process, good coordination among different government levels, reforms in public expenditure management, improved accountability systems. An example of a public administration reform package applied in SEE is presented in Box 5.2. 66 Box 0.2: Public Administration Reform in Montenegro50 In March 2003 the government of Montenegro approved a seven-year Public Administration Reform strategy, which sets out key objectives of public administration reform and puts forward an ambitious structural and legislative reform program for the reorganization and reorientation of the state administration. The main priority is the delegation of competencies from higher to lower administrative levels, which should enhance flexibility of the administrative system and leave senior servants greater latitude in performing their tasks and functions. Administrative reform will be conducted in three stages: the first in 2003-2004, the second in 2004-2006, and the third in 2007-2009. The first stage concentrates on the development of legal acts, while the next two stages are devoted to the implementation of the strategy and adoption of new rules and procedures. So far, the government has endorsed an action plan to implement the strategy, which has been fully incorporated into the government program. A set of key public administration laws, such as the Law on the State Administration, the Law on Ombudsman, and the Law on Inspection Control, were approved by Parliament in June 2003. Furthermore, the Law on Administrative Procedure and the Law on Administrative Dispute were enacted in November 2003. A further set of laws, including the Law on Civil Service and the Law on Salaries was adopted in late April 2004. In 2005, the Civil Service Agency began implementing a performance appraisal system, with the ultimate goal of linking performance to salary increases for all civil servants. Strengthening the institutional capacity and especially accountability system in Montenegro is of crucial importance for carrying out the second stage of reforms. The inherited Montenegrin tradition of top-down management requires clearly established lines of authority, supported by strong accountability mechanisms and institutions to control the legality and acceptability of administrative conduct. An important step forward in strengthening accountability has been made with the adoption of the Law on Courts in 2002, which foresees the establishment of two new courts, the Administrative Court and the Appellate Court, by mid 2004. Furthermore, the creation of the Ombudsman in October 2003 and the Civil Service Agency in mid-2004 should together provide the basis for reducing abuse of office and corruption in Montenegro, and also create a more responsive and client-oriented administration. V. CONCLUSIONS AND RECOMMENDATIONS 5.36 Spending on public sector wages in SEE is moderately higher than among the NMS, both as a share of GDP and total public spending. Public employment is not excessive relative to the size of the population, but is large relative to the total labor force. Employment is larger than in comparator countries in public security and education. Overall, the size of the public sector wage bill is most bloated in Bosnia and Herzegovina, Croatia and Montenegro. Reducing the size of the public wage bill is a key challenge for most SEE governments, but especially for those three, going forward. 5.37 There is a need to strengthen governments’ capacity to monitor and control public sector employment levels and wage costs. There is also a need for comprehensive pay and incentive structure reforms. All SEE countries use a form of coefficient system for civil service wage determination, which links the pay level, rank, and qualifications. Generous bonuses and allowances are common, compression ratios are low, politicization still high, and modern performance management tools and mechanisms are sparsely used. The pay system is static in nature, and does not meet the requirements of a modern pay and benefits system. 50 See “Serbia and Montenegro Public Administration Development: Creating the Conditions for Effective Economic and Social Reform”, 2005, Policy Note, Poverty reduction and Economic management Unit, Europe and Central Asia, World Bank. Also see, Montenegro: A Public Expenditure and Institutional Review, 2006, World Bank. 67 5.38 Several broad lessons and challenges for the short to medium-term could be highlighted: • Apart from the still high spending on security and armed forces, large wage bills in sectors such as education and health are the key contributors to the high overall wage bills in the SEE. • In general, wage levels in the public sector are not excessive in SEE. However, public sector wages are high compared to the private sector at least in BH for lower-grade employees. • Rationalization of the wage bill should be given a high priority to help assure fiscal sustainability and release resources for growth-enhancing activities. This will require a reform of civil service and more generally public sector employment, and revamping of wage policies. These reforms should help align public institutions with the EU acquis and facilitate policy coherence. 5.39 The reform measures may be grouped into four broad areas: • Rationalize the pay system, in particular streamlining and incorporating into the wage still existing allowances: Rationalization of the system would bring transparency and predictability to the wage setting-mechanism, and allow for a consistent and focused approach to the public sector wage policy. Since many of the allowances are enshrined in the labor law, its revision would need to be part of the reform. • De-compress salary scales and advance depoliticization of the civil service: Decompression of salaries and reduction of politicization are essential for attracting skilled workers, providing incentives for staff to seek increased levels of responsibility, and decrease turnover levels in the administration. • Reduce staff: Changing systematization and staff planning systems would be needed to rationalize staff numbers in particular in administrative functions and social sectors, thus allowing for rightsizing and eliminating historically developed distortions. This could be accompanied by identification of non-core services that can be contracted-out without affecting the quality of the public service. Possibilities to offset increases in local government employment associated with ongoing devolution of functions from central to local governments by reductions in central government would need to be considered. • Move towards merit-based hiring and pay system, with the understanding that the SEE countries are a long way from satisfying the conditions necessary for a performance-based pay system to operate reasonably well.: To create incentives for performance, there is a need to link performance and career advancement, and use of instruments such as accelerated promotion, fast track career development,51 and performance based rewards.52 While performance based reward systems are most difficult to apply to civil service jobs, there is ample experience of functioning systems in OECD and other countries. In public service sectors, where outputs and quality can better be measured, linking career progression and pay levels to performance should be relatively easy. 51 As used in the UK, Ireland and Slovakia. 52 A majority of EU and OECD states have experimented over the last two decades with the introduction of such mechanisms in previously rigid career systems in order to build civil service systems more suited to the current global competition. 68 Infrastructure: FIscal space and public inestment management 53 6.1 Following recent sizable investments in public infrastructure by Croatia and Slovenia, the other countries of SEE have started to plan for large outlays on infrastructure. These plans are motivated by the priority SEE governments have awarded to building new motorways, railroads and energy generating capacity and rehabilitating existing facilities, reflecting perceptions about the poor state of infrastructure and efforts to accelerate convergence to Europe.54 Strengthened fiscal positions and more favorable international environment, recent market turmoil notwithstanding, have also given support to such plans. Active involvement by multilateral and bilateral institutions in the planning and sometimes the execution of some of these plans has also been a factor, while private foreign investors are yet to become more actively involved. 6.2 This short chapter lays out some of the key issues to consider when planning and evaluating new investments in infrastructure. Public provision of infrastructure depends crucially on finding budgetary room for new spending without jeopardizing fiscal sustainability or macroeconomic stability (“fiscal space”). The scope for making fiscal space available for additional infrastructure spending by relaxing substantially deficit targets or boosting substantially the level of government spending are rather limited in most but not all SEE countries, given the large size of government in most and high levels of government and external debt in some. Nonetheless, some countries may potentially sustain larger fiscal deficits, provided investments are well prioritized, rigorously evaluated and monitored and help boost productive capacity. 6.3 Outside these special cases, however, fiscal room for additional infrastructure investment can be only found in reducing or reprioritizing existing expenditures or involving the private sector. Along with these considerations, it is important for governments to reassess whether infrastructure is the most binding constraint to development and whether, given the stage of development, efforts should not also target other areas. 6.4 Any potential new borrowing, moreover, needs to be undertaken only after careful debt sustainability analysis. Addressing the absorptive capacities of the governments to handle new projects also needs to be enhanced. Given the small size of SEE countries, cross-country cooperation appears crucial for almost all projects, notably motorways, railroads and electricity generating capacity. 6.5 Private sector participation in infrastructure is important, and it can lessen the direct fiscal burden. However, this does not reduce the need to have a high quality project within a strong sectoral framework. Failing this, the implicit contingent liabilities of the government in partnering with the private sector are likely to translate into actual expenditure and bailouts of unsatisfactory projects. This means that the suggested caution on financing new government expenditure for investment applies equally to extending guarantees, implicit or explicit, for encouraging private sector participation. 53 This chapter is prepared by Ivailo Izvorski and has borrowed from a note prepared by Sanjay Kathuria (“Fiscal Space for Infrastructure Borrowing in SEE – A Suggested Approach,” 2006). 54 This chapter focuses on the SEE countries excluding Croatia and Slovenia. 69 I. INFRASTRUCTURE GAPS 6.6 Infrastructure gaps in the SEE countries are claimed to be large, but hard numbers on them are hard to come by. In terms of roads, for example, countries in the region have fairly well developed networks, but – with the exception of Croatia and Slovenia - these are well short of the levels of coverage among some of the EU new member states and suffer from heterogeneous quality (Table 6.1). The quality of roads is rather poor, however (especially of local ones), particularly in Albania, Macedonia and Serbia (Figure 6.1). Table 0.1: Density of Road Infrastructure Figure 0.1: Road Network Quality (latest observation available, 1997-2003) 80% (km/1000 (km/1,000 70% sq km) People) 60% FYRM FYR Macedonia 518 6.4 Croatia 506 6.4 50% Croatia BH 427 5.6 40% Bosnia SaM 494 4.8 30% Kosovo 783 4.2 20% Albania Albania 657 3.5 10% 0% Hungary 1,733 15.7 Czech Republic 1,646 12.5 Good Fair Poor High income: OECD 1,340 17.3 Upper middle income 1,076 9.2 Source: BCEOM, 2005; FNRR, 2005, World Bank PEIRs for ECA 580 8.6 Kosovo and Bosnia and Herzegovina. Source: WDI and IEF databases. 6.7 In most countries in the region, users are indeed of the view that the quality of infrastructure is poor, but with the exception of electricity in Albania and Kosovo, infrastructure is not rated among the ten obstacles to doing business in SEE (according to the BEEPS 2002 and 2005) (Figure 6.2). About half of the firms in SEE consider macroeconomic stability, uncertainty about regulatory practices, tax rates, contract violations, anti-competitive practices, cost of financing, corruption and functioning of the judiciary are seen as obstacles to doing business. On average, less than a fourth can say the same about transportation, electricity and communications. These results are similar in all countries, except in Albania and Kosovo, where half and three-fourths of the firms, respectively, consider electricity an obstacle to doing business. 6.8 Estimates of road maintenance backlogs prepared by national governments are large, but they have to be considered in light of the simplifications and other assumptions used in the calculations. In Macedonia, for example, the estimated cost of clearing the maintenance backlog on the road network would amount to 2 percent of GDP a year for five years. The estimated costs in BH are about 1 percent of GDP a year for five years. A number of qualifications to these estimates are important. Firstly, these estimates reflect the authorities’ assumption that the road network be returned to the existing design standards (which may now not be appropriate in the context of changed travel patterns and increasing demand for road transport. Further, these estimates crucially depend on national categorizations of roads. Finally, the estimates do not take into account the cost of closing roads and the cost of congestion. All in all, these estimates are a useful guide, but need to be interpreted cautiously. 70 6.9 Foreign investor perceptions of Figure 0.2: Problems Doing Business, SEE and the quality and availability of Comparisons, 2005 infrastructure are important for attracting (Percent of firms indicating a problem) foreign direct investment, but yet again, investors consider macroeconomic Co nt r act vio lat io ns stability, competitive markets, smaller A nt i-co mpet it iv e informal economies and country size as pr act ices o f o t her s more, or at least equally, important.55 Or ganis ed Potential foreign investors also seem to cr ime/maf ia focus on the availability of S t r eet cr ime, t hef t infrastructure, rather than their costs in and dis o r der influencing decisions.56 C o r r upt io n 6.10 These considerations F unct io ning o f t he notwithstanding, improved physical and judiciar y human capital are key drivers of M acr o eco no mic economic growth. Given that such ins t abilit y investments take a lot of time to plan U ncer t aint y abo ut r egulat o r y po licies and execute before contributing to economic growth, however, S kills and educat io n o f wo r ker s governments will be prudent to look further into the future as they tackle L abo ur r egulat io ns current challenges. B us ines s licens ing and per mit s II. MACROECONOMIC C us t o ms and t r ade CONSIDERATIONS r egulat io ns T ax adminis t r at io n 6.11 The perceived poor quality of infrastructure reflects to a large extent T ax r at es underinvestment in maintenance, stagnating or declining government T it le o r leas ing o f land investment in the context of underdeveloped capacity by the private A cces s t o land sector to contribute to infrastructure. The problems have been particularly acute in T r ans po r t at io n countries where the decline or stagnation of government capital spending was not E lect r icit y offset by larger investment by the private sector (see below). Albania is a T eleco mmunicat io ns case in point: government capital spending fell by 1.6 percent of GDP on C o s t o f f inancing average during 2003-2006 compared with 2000-2002, accounting for almost A cces s t o f inancing all the decline in (overall) fixed S E E -2 00 5 0% 2 5% 5 0% 7 5% 10 0% investment (Table 6.2). Government C IS -2 00 5 capital spending was little changed over E U 8-2 0 05 E C A -2 00 5 55 McKinsey Global Institute, 2004, “The Truth about Foreign Direct Investment in Emerging Markets,” in McKinsey Quarterly, Washington DC. 56 IMF, 2003, “Foreign Direct Investment in Emerging Market Countries,” Report of the Working Group of the Capital Markets Consultative Group, Washington DC. 71 these periods in BH, Macedonia and Serbia, and only in Serbia did the private sector boost investment (with only a small part going to infrastructure). With the exception of Croatia and Slovenia, fixed investment among the other SEE countries averaged 21 percent of GDP during 2000-2006, with little variation across the region. Fixed investment averaged 26 percent over the same period in Croatia and 27 percent among the new EU member states. Table 0.2: SEE: Fixed Investment and Government Capital Spending, 2000-2006 (In percent of GDP) Averages 2000 2001 2002 2003 2004 2005 2006 2000-02 2003-06 Albania Fixed investment 24.7 27.5 24.5 23.4 23.8 23.6 25.0 25.6 24.0 Gov’t capital spending 6.6 7.3 6.6 4.5 5.2 4.8 5.6 6.9 5.0 Croatia Fixed investment 20.2 23.9 29.1 31.1 30.6 31.0 32.8 24.4 31.4 Gov’t capital spending 5.8 5.8 6.7 8.6 9.2 7.4 8.7 6.1 8.5 BH Fixed investment 21.2 19.4 18.7 18.8 19.4 22.0 16.2 19.8 19.1 Gov’t capital spending 9.4 8.0 7.9 10.3 9.5 9.1 5.9 8.4 8.7 Montenegro Fixed investment 21.4 23.0 17.6 16.6 20.6 22.1 31.7 20.7 22.8 Gov’t capital spending 4.0 3.4 2.2 5.1 4.1 5.2 5.6 3.2 5.0 FYR Macedonia Fixed investment 22.3 19.1 20.6 20.0 21.4 20.7 20.9 20.6 20.7 Gov’t capital spending 2.7 3.5 3.6 4.1 3.1 3.6 2.9 3.3 3.4 Serbia Fixed investment 8.4 15.9 17.2 22.6 23.6 22.5 21.2 13.9 22.5 Gov’t capital spending 2.8 2.2 3.5 3.0 2.7 2.9 4.4 2.8 3.2 Sources: National authorities and World Bank staff estimates. 6.12 The macroeconomic situation in all SEE countries has now admittedly improved, but key vulnerabilities remain that needs to be taken into consideration when charting a public investment strategy. Macroeconomic vulnerabilities include rather high public debt levels (Albania), high external debt (Croatia), large general government expenditures relative to GDP (BH, Croatia and Serbia), and high and rising current account deficits in all countries except Macedonia (see Chapter 1 for details on fiscal deficits and debt). Moreover, all SEE countries except Slovenia have oversized informal economies, reflecting, in large part, high payroll contribution rates. 6.13 Current account deficits are high in most countries (except Macedonia) and while financing has been readily forthcoming so far, the turmoil in international financing markets in 2007 has been a reminder that capital inflows are subject to substantial fluctuations and “sudden stops”. The current account deficits of the SEE countries have been on upward trend since the turn of the century, with the region’s current account deficit widening from 4.1 percent of regional GDP in 2000 to about 9 percent in 72 2006, compared with little change on average for all countries of emerging Europe. Current account deficits rose from 2000 among all SEE countries except Macedonia, where the current account deficit actually narrowed in 2005 and 2006. 6.14 Underlying the high current account deficits are relatively low levels of exports of goods and nonfactor services (GNFS). Exports of GNFS in 2005 amounted to 39 percent of regional GDP, compared with 67 percent in Hungary (Table 2). In Albania, an extreme on the downside, for example, exports per capita in euro terms amounted to only 8 percent of the level in Hungary, illustrating the poorly developed export capacity. These observations, however, need to be considered in the context in rapidly increasing exports in most countries except Macedonia, at paces faster than in Hungary. Interestingly, the meager increase in exports in Macedonia is to be compared with the very low level of the current account deficit in that country, indicating either a low confidence by foreign investors (and thus small capital inflows), low domestic demand for foreign saving, or both. Inflows of foreign direct investment rose by about 2 percent of regional GDP from 2000 until 2006, much less than the region’s current account deficits widened. Inflows of FDI, as a result, covered just two-thirds of the regional deficit in 2006, down from almost full coverage in 2000. 6.15 Dependence on foreign grants, both current and capital, and on remittances needs to be monitored carefully. Macedonia, BH and Albania depend to a very substantial degree on remittances (estimated at 18, 16 and 15 percent of GDP, respectively) that crucially depend on confidence in the banking system, as developments in the early part of this century will testify. 6.16 Additional capital spending, which in the SEE countries would spill to a fair share into imports, could contribute to further widening already oversized current account deficits. There imports will automatically be financed with foreign borrowing, as part of the most likely package of financing of infrastructure, but their contribution to external debt needs to be evaluated carefully, lest they lead to concerns about macroeconomic stability and debt sustainability. Gross external financing needs are already high in most countries and are expected to increase in many. The recorded gross financing needs, which include the current account deficit, amortization due on external debt and increase in reserves, averaged 20 percent of GDP for the SEE countries in 2005, varying from about 10 percent in Albania to about 28 percent in BH. Overall, maintaining and dealing with large capital inflows averaging 20 percent of GDP presents a constant challenge to policy makers. 6.17 Tackling macroeconomic vulnerabilities will require fiscal prudence and continued fiscal consolidation, together with efforts to tighten financial sector supervision and advance structural reforms. Reducing government expenditures relative to GDP and maintaining prudent fiscal positions is crucial for helping strengthen investor confidence, boost inflows of foreign direct investment and accelerate growth in output, exports and employment. Managing these challenges will be compounded by the need to tackle fiscal pressures that include (budgetary) costs related to EU integration (all countries), contingent liabilities (such as completing property restitution and formalizing claims for frozen currency deposits and war claims, especially in BH), and the costs related to population ageing and corporate restructuring (all countries). Further, when considering debt sustainability, the present value (NPV) of debt is significantly lower than the face value for most of the countries, but given the likely hardening of terms of new borrowing, the burden on incremental debt will still be higher than it has been thus far. 6.18 Given macroeconomic concerns and assessments of key obstacles to doing business and growth, plans to boost infrastructure spending need therefore to be assessed carefully. While improved infrastructure can, in principle, have a significant positive impact on economic growth, governments need to consider whether this is among the most binding constraints and subject proposals to rigorous evaluations of technical and economic merit (see next section for more detail). Moreover, additional capital spending needs to be considered within the overall fiscal envelope and should not jeopardize debt 73 and macroeconomic sustainability. Fiscal deficits could be modestly increased in some countries. In any case, the increases that will not jeopardize fiscal sustainability are far smaller than the perceived or identified need to boost infrastructure spending, and are thus no substitute for expenditure (re)prioritization. 6.19 Given these macroeconomic constraints, the pace at which countries will be able to fill perceived infrastructure gaps will depend critically on their capacity (i) to find fiscal space within existing resource envelopes; and (ii) to program public investment towards economic priorities, the points to which we now turn. III. FISCAL SPACE 6.20 Fiscal space is defined as the “availability of budgetary room that allows a government to provide resources for a desired purpose without jeopardizing the sustainability of its financial position.”57 Such budgetary room can be created by, inter alia, one or more of the following elements: increasing revenues, increasing fiscal deficits, reducing existing spending, or shifting the provision of certain goods and services, in part or in whole, to the private sector. The latter could involve privatization or different arrangements such as public private partnerships (PPPs), government guarantees and others. 6.21 Creating fiscal space for investment in infrastructure addresses only one part of the broader issue on how to increase overall infrastructure spending in the economy. The other part refers to the development of a private sector, including foreign-directed invested companies, able and willing to undertake investments in infrastructure on its own or together with the government. Boosting private sector participation depends crucially on the maturity of markets and institutions, including security of property rights and contracts, and prospects for the economy.58 6.22 Creating fiscal space through increasing tax rates or fees is likely to be problematic for most of the SEE countries, given the large government burden on the economy and the large informal sectors in most. There is substantial room, however, for improving tax administration and collection efforts. Further, foreign grants are set to increase substantially as EU integration is advanced. 6.23 Creating fiscal space by relaxing deficit targets is rather limited for most SEE countries, given the large size of government spending in most and high levels of deficits and government and external debt in some. Clearly, the answer depends on the source of financing (whether borrowing or privatization receipts) and debt sustainability. In Macedonia, the government increased the medium-term deficit target under the IMF-supported program to 1.5 percent of GDP from 2008 onward from about 1 percent in 2007, but this should not jeopardize fiscal sustainability given improved growth prospects and further progress in advancing structural reforms. Additional issues to consider before increasing deficit targets are discussed in the following sections. 6.24 Prima facie, the situation would look as follows on a country-by-country basis: a. Albania’s significant fiscal deficit and the need to start formalizing the oversized informal economy suggests that spending should not be increased relative to GDP. Albania’s capital expenditures, even after the decrease to 4.5 percent of GDP on average over the last several years from about 7 percent in 1999-2002, are still rather high. 57 Heller, Peter, 2005, “Fiscal Space: What It Is and How to Get It,” in Finance and Development, The International Monetary Fund, Washington, DC. 58 IMF, 2005, “Public Investment and Fiscal Policy – Lessons from the Pilot Country Studies,” SM/05/118, International Monetary Fund, Washington, DC. 74 Government capital outlays are projected to increase to around 7 percent of GDP by 2009, but whether this is realized remains to be seen given existing absorption capacity. For Albania, fresh proposed infrastructure expenditure should be weighed against existing planned capital expenditures, and substitute such expenditures if found to have higher returns. This argument to cap capital expenditures at currently forecast or planned levels is strengthened by the relatively high debt to GDP ratio (see earlier section), shortcomings in prioritization and implementation of public investments, the significant fiscal deficit, and the weak revenue generation capacity in Albania. The government will be well advised not to boost public investment as a share of GDP until public investment management is strengthened considerably. b. At current levels, Macedonia’s expenditures relative to GDP are fiscally sustainable and broadly similar to the average of countries with similar levels of income per capita. In terms of fiscal space, government plans to increase the fiscal deficit to about 1.5 percent of GDP a year over the medium term suggest a case could be made for high quality additional infrastructure spending, which would require well-designed projects as well as a good regulatory framework. (Part of the increase is accounted for by tax cuts and hikes in current spending, however.) Inadequate implementation and expenditure management capacity, however, has left government capital spending well below budgeted levels over the last several years. Strengthening capacity is needed to reduce the risks of inefficient government capital expenditure. Further, although Macedonia’s balance of payments is not currently considered vulnerable, the current account deficit is projected to start increasing in 2008. Dependence on remittances remains high and statistical data regarding the remittances is still spotty. c. In comparison, BH and Croatia both have high levels of government spending and high capital outlays by the government. In Bosnia and Herzegovina, capital spending by the government averaged an already high 8 percent of GDP over the last several years, and further increases need to take into account the role of the private sector. The government appears intent on using privatization receipts and the surge in revenues due to the introduction of the VAT at a higher-than-neutral rate to finance large infrastructure spending without reducing existing spending. (Privatization receipts amounted to about $1 billion in 2007, thanks to privatizations in the RS, and may well amount to as much in 2008 and 2009 combined should privatization in the FBH entity proceed as planned). These additional expenditures will add to the need to settle the large stock of domestic claims, handle property restitution and strengthen the state level of government to advance EU integration. Should these plans be carried out, the size of the government in BH will increase, ultimately with detrimental impact on growth in output and employment. d. In Croatia, government capital expenditures at about 7 percent of GDP a year are dominated by spending on a massive motorway construction program (2001-2008) that consumes the equivalent of 4.2 percent of GDP a year. Similarly to BH, the government has carried out sizable fiscal consolidation over the last several years, but at about 49 percent of GDP, the size of government is among the highest in the region and is certainly higher than warranted from a growth and efficiency point of view. Further consolidation is planned, to be helped by the winding down of public spending on infrastructure, but progress will be complicated by the need to meet EU accession requirements (that carry short-term costs to the budget). 75 6.25 Expenditure rationalization, including reduction in wasteful spending, is therefore generally the most promising path to create fiscal space for infrastructure, especially given concerns in most about the poor outcomes of government spending. Details recommendations on expenditure rationalization in key sectors in the economy are presented in the World Bank Public Expenditure and Institutional Reviews. 6.26 Fiscal space could be created by prioritizing the government’s participation in the provision of key services. Planned public investments with low rates of return could be dropped. Government and public expectations of the needs of the country and the economy could be revised, as these needs are sometimes based on ambitious of catching up with more advanced countries in the region and the EU that do not take into account the substantial time needed for convergence. 6.27 Provision of other services that used to be in the purview of the public sector in the past could be privatized, such as telecoms. Nonetheless, there are strong concerns in some countries (for example, BH and Macedonia) that privatization, specifically of telecoms, has not been accompanied with liberalization, with the result that competition has remained curbed, costs high and access limited. There has been little progress in privatizing the electricity sector; this contrasts with well advanced privatizations in most new EU member states. In several countries, including Macedonia, the government continues to subsidize the electricity sector to help cover the losses that have resulted from keeping regulated prices well-below market levels and, in some cases, below recovery costs. Government subsidies are also large to the railroads in most countries. Issues are also raised by the limited scope for local utilities (including those providing water and sewage) to charge prices that ensure cost recovery. Progress in resolving these issues, including by moving toward cost recovery and ensuring transparency of remaining subsidies, would go some way toward creating fiscal space for other infrastructure spending. 6.28 An established legal framework for engaging the private sector is crucial, including laws and institutions for concessions and PPPs, as well as the framework and the experience of managing unsolicited proposals, an issue that is especially pertinent in SEE. Using PPPs to involve the private sector could reduce present fiscal costs, but unless well designed, these may well bring substantial costs in the future. The over-riding objective should be to protect the fiscal stability by limiting the government’s fiscal exposure wherever possible. Other objectives include the need to shift the risk allocation away from taxpayers and toward providers and users, and the need to improve the efficacy of agencies involved in financing of infrastructure.59 IV. PUBLIC INVESTMENT PLANNING 6.29 High quality infrastructure investments will be crucial for accelerating and sustaining strong output growth, while poorly designed investments will be a drag on the budget and further hamper the economies’ productive capacity. Building the capacity to ensure that projects are selected through a careful evaluation, within well-prioritized strategies and taking into account regional dimensions and integration, will be essential for bringing to fruition high quality infrastructure investment. These issues are tightly linked with the need to carefully consider the fiscal and macroeconomic constraints. 6.30 Capacities for public investment planning are rudimentary in most SEE countries (excluding Slovenia and Croatia). Countries typically lack a systematic approach to new infrastructure, with sectoral ministries submitting expenditure needs to the ministry of finance that is ultimately responsible for preparing public investment programs. Strategies for the road sector (and transport sector more generally), one of the most important, are prepared in almost all countries with external assistance, but the 59 Budina, Nina et al., 2007, “Public-Private Partnerships in the New EU Member States: Managing Fiscal Risks,” World Bank Working Paper No. 114, The World Bank, Washington DC. 76 processes to update them are not yet established. In addition, the use of asset management systems to prepare multi-year plans for rehabilitation and maintenance in the road sector on a regular basis is absent in most countries. Moreover, the use of formal techniques for project evaluation is underdeveloped and rarely used. Last but not least, countries that have benefited from generous foreign assistance – and have relied on donors’ systems in providing infrastructure maintenance and rehabilitation – should lose no time in upgrading their own country systems to help evaluate, plan and finance infrastructure investment (Albania, BH and the province of Kosovo). 6.31 The need for stronger and comprehensive databases of existing and newly approved projects, as well as project disbursements, cannot be overemphasized. This is especially important for Albania, BH and Kosovo. In Albania, a weak information base at the MoF on ongoing public investment projects has undermined realistic expenditure planning and execution. There is no single register of approved investment projects in which annual allocations and expenditures are recorded against the total estimated cost of each project. In BH, expenditures on foreign-funded projects are not even consolidated in entity budgets and even current efforts to prepare consolidated general government statistics (spearheaded by the central bank) do not include such outlays. The issue of incomplete recording in public budgets is also important in Albania. 6.32 Croatia appears to be the most advanced country in the region (save for Slovenia) in terms of planning for public investment, notably road infrastructure. The current institutional setting has been instrumental in doubling the length of the motorway network from 2001 to 2004 to more than 800km in what has become a regional example of good program implementation and high quality output. Nonetheless, there is an opportunity to further reduce the operating costs and improve the planning process. The overall operating cost accounted for 2.9 percent of the overall road outlays during 2001- 2004 and this is projected to increase to 3.8 percent on average in 2005-2008. The government’s political decision to invest in the motorway network lead to the subsequent reshaping of the road infrastructure institutions, driven by the need to implement this program while adopting different financing mechanisms. 6.33 Planning for road infrastructure development in Croatia is a comprehensive process, but does not involve any economic appraisal techniques (a key weakness in all SEE countries). There are three phases of the planning process. First, the ministry of transport prepares a strategy for the development of public roads (based on the strategic plans of the Croatian Motorways, Croatian roads and the counties); the strategy is adopted by the government and subsequently by the parliament. Second, the ministry prepares a four-year program based on the adopted strategy. (So far, there have been two programs, for 2001-2004 and 2005-2008.) Finally, the Croatian Motorways, the Croatian roads and the County Roads Administrations prepare one-year implementation plans for the construction and maintenance of public roads. A key weakness of the program, as emphasized above, is the lack of formal economic appraisal techniques (including cost-benefit analysis) to review and prioritize the investment program. 6.34 In Albania, the government has taken initial steps to strengthen public investment management by transferring the responsibility for this function from the Ministry of Economy to the Ministry of Finance, and setting a new Public Investment Management Department in the General Directorate of Budget. Priority next steps need to include, as in all other countries, the development of procedures for project identification, appraisal, and approval. Introducing a common process for both domestically and externally financed investment projects would ensure that all public investment is subject to a common appraisal requirement regardless of the source of financing. New measures should ensure that public investment planning and procedures are fully integrated within the medium-term expenditure plan. New draft procedures applied for the supplementary budget 2006 and 2007 go in that direction. 77 6.35 In Montenegro, until 2007, capital expenditure plans were with a one-year horizon and little heed is paid to the medium-term budgetary needs for completing capital projects. The 2007 budget is the first that included capital budgeting. Amendments to the budget law have provided the legal framework for the introduction of the capital budget, and have been underpinned by the creation of a capital budgeting unit in the ministry of finance to coordinate capital works with the Directorate of Public Roads and the Directorate of Public Works. Spending units are not required to submit a multi-year capital budget to the ministry of finance and the two above-mentioned directorates for all capital projects. Key issues for the government to resolve going forward include the need to strengthen the institutional capacity of all spending units so capital projects proposed are based on sound economic, social and environmental analysis. Selected projects, thereafter, need to fit the overall strategic goals of the government and the available fiscal envelope. 6.36 In Macedonia, the planning and budgetary process for public infrastructure, road transport in particular, has been gradually improving but capacity remains poor. Deficiencies that undermine the efficiency of transport expenditures include: (i) the absence of a formal sector policy and strategy, consistent with the needs and the fiscal envelope available to the government; (ii) the absence of a process that contributes to the formulation and revision of a sound medium-term sector strategy by the Ministry of Transport and Communications (MOTC); (iii) weaknesses in project identification and assessment both within the MOTC and Road Fund (FNRR); and (iv) limitations in the management of the assets in the sector by FNRR and Makedonija Pat (the public company in charge of road construction and maintenance). The limited use of formal techniques of economic appraisal in project identification and prioritization is a key weakness in the budgetary process in Macedonia and in most other SEE countries. In Macedonia, emphasis has been placed on the development of the Helsinki corridors VIII and X without a clear action plan for phasing of investments. 6.37 Public investment planning in BH is multi-annual, but is fragmented, carried out separately by the state and the two entity governments. In the transport sector at present, the state-level MOCT, together with the two entity ministries and with the external support from the World Bank and the EBRD, is working to develop a country-wide transport sector policy and strategy. This process should also include the strengthening of capacity in the public sector institutions to facilitate: (i) the revision of a sound sector strategy; (ii) the use of formal project appraisal to assist in the identification and prioritization of projects; and (iii) the operational establishment of robust asset management systems in the operational agencies. The authorities are well-advised to accord greater responsibilities to the state-level MOCT for the policy development and harmonization, planning, and prioritizing in the sector, for national and international transport issues. This would remove some of the burred lines of responsibility in the sector at present, particularly in relation to the development of the strategic road infrastructure. V. CONCLUSIONS AND RECOMMENDATIONS 6.38 Governments need to continuously and carefully re-assess the key constraints to stronger economic growth and improved living standards. Given the costs, time lags and large macroeconomic effects involved, a forward-looking approach is crucial. In this regard, governments need to re-evaluate their approach to government capital spending, and especially the planned large increases in outlays for motorways and other transportation works. The core issues include whether maintenance should be favored over new construction and to what extent, or whether motorways need to have priority over local roads. The following recommendations are intended to guide such re-assessment: • Increase the share spent on maintenance within existing funds. Within the road sector, most importantly, be mindful of the costs of closing down public roads and of the trade-offs with new construction. Review the existing categorization of roads – that usually drives 78 estimates of maintenance backlogs and helps direct outlays – by linking resource allocation to road classification and demand. • Introduce competitive tendering to reduce maintenance outlays. • Strengthen capacities of the ministries of transport and public road funds to improve the prioritization of road maintenance and new investment. • Develop infrastructure sector policy and strategy consistent with the fiscal envelope available to the authorities, and establish a process to regularly revise these to inform public investment programs and the annual budget preparation. • Strengthen or create databases (registers) of investment projects and consolidate all investment expenditures into government budgets, whether financed domestically or from foreign loans/grants. Such databases of registers need to be integrated within the public expenditure planning system and reflected in medium-term budget plans. • Strengthen public investment management. Projects included in public investment programs, need to undergo a rigorous evaluation on technical, economic and budgetary ground. • Closely monitor projects in execution and after execution. Often, projects tend to be profitable at the feasibility stage or at the start of execution, but problems can appear quickly. • Carry out open and transparent competitive bidding for all projects to ensure that best value is obtained for the public resources available. • Develop the legal framework for private participation in the provision of infrastructure, including by strengthening laws on concessions and public private partnerships. 79